TRAF3 as a regulator of T lymphocyte activation

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Summer 2017 TRAF3 as a regulator of T lymphocyte activation Alicia M. Wallis University of Iowa Copyright 2017 Alicia M. Wallis This dissertation is available at Iowa Research Online: Recommended Citation Wallis, Alicia M.. "TRAF3 as a regulator of T lymphocyte activation." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Immunology of Infectious Disease Commons

2 TRAF3 as a Regulator of T Lymphocyte Activation by Alicia M. Wallis A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Immunology in the Graduate College of The University of Iowa August 2017 Thesis Supervisor: Professor Gail Bishop

3 Copyright by Alicia M. Wallis 2017 All Rights Reserved

4 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL This is to certify that the Ph.D. thesis of PH.D. THESIS Alicia M. Wallis has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Immunology at the August 2017 graduation. Thesis Committee: Gail Bishop, Thesis Supervisor Jon Houtman John Colgan Kris DeMali Scott Lieberman

5 To my family who saw the best in me when others did not, never stopped pushing me when I was discouraged or knowing how far I could go when I lost my path. Your support has gotten me to this point and to the person I am today. I am stronger because of you. ii

6 Everything stinks till it s finished ~Dr. Seuss iii

7 ACKNOWLEDGEMENTS First I would like to thank my mentor, Dr. Gail Bishop. Her guidance and encouragement have chiseled away my imperfections, sculpting me into a better scientist. Her dedication over the last five years has driven me further than I ever could have imagined. It has been my pleasure being a part of the Bishop lab, a community she has developed that is full of support and respect. Secondly, I would like to thank my co-mentor and committee chair, Dr. Jon Houtman. His knowledge of T cells has helped bring my research to the next level. I am also appreciative of the teaching opportunities he made available to me. His enthusiasm and banter have made these five years go by quickly. I would also like to thank my committee members, Drs. John Colgan, Kris DeMali, Scott Lieberman and Lyse Norian. Their guidance during committee meetings has been instrumental in my science and progression as a graduate student. I have had many great discussions with them and hope to continue to have these. One cannot go too far without thanking members of the Bishop and Houtman labs, both current and former. A special thanks to Dr. Zuoan Yi, a brilliant mind who helped me get my project off the ground and taught me the philosophy of a good scientist. Dr. Wai Wai Lin is an amazing scientist who developed my skills in my early years and continues to help me despite the distance between us. Dr. Laurie Stunz is a faithful friend who helped guide me personally and professionally. I will miss our caffeine runs and discussions. Dr. Bruce Hostager has wit and intelligence like no other. His banter made my long days bright. Dr. Mahmood Bilal is an all-around great guy whose help was essential to the advancement of my research. I thank Emma Hornick, who was a true iv

8 friend and the best support a person could get. I must also thank Emma Hornick, Isaac Jensen, Ellie Wallace and Laura Mallinger, I had the pleasure of mentoring each one of these amazing minds and their work helped generate some of the data presented in this dissertation. Lastly, I would like to thank my family. By far I must thank my mother, Regina Wallis, through her support and encouragement I gained my strength and perseverance. A huge thank you goes out to my sisters Melinda Clark and Kayla Wallis. Even at a young age Melinda taught me how to be a student and was the best role model a younger sister could have. When I was wound too tight, it was my younger sister Kayla who reminded me to cut loose and enjoy life. To my nephews Dean and Dillon Pickle Clark, you have been the best distractions. And finally, to my father Michael Wallis, while you were not here physically I know you were here in spirit. I owe my love of science, pure stubbornness and uncanny wit to you. v

9 ABSTRACT T cells are an essential component of the adaptive immune system, which evolved to facilitate development of long-term, effective protection against infectious diseases. Upon activation, T cells play an important role in clearing infections, and especially, in preventing establishment of subsequent infections with the same pathogen. Because this is such a powerful response, it must be tightly regulated. Our lab has long been interested in how signaling molecules regulate the function of T and B lymphocytes. Our prior studies stimulated an interest in the signaling adapter molecule, Tumor necrosis factor receptor (TNFR)-associated factor 3 (TRAF3). Our group previously produced a T cellconditional (CD4-Cre) TRAF3 -/- mouse, which demonstrated that TRAF3 unexpectedly plays an important positive role in T cell functions, including providing help for B cell responses, protection from infectious pathogens, cytokine production and proliferation. After TCR engagement, TRAF3 associates with the T Cell Receptor (TCR)/CD28 complex. These data identified a new role for TRAF3 in T cell activation. There are three signals that are required for full T cell activation. The three types of receptors that deliver these signals are the TCR, co-stimulatory receptors and cytokine receptors. This dissertation explores the regulatory role of TRAF3 in the 3 signals required for T cellsactivation. In signal 1, TRAF3 enhances TCR signaling by regulating the localization of the TCR inhibitors, PTPase non-receptor type 22 (PTPN22) and the c-src kinase (Csk). Our lab previously reported that recruitment of TRAF3 to the TCR complex requires co-stimulation of CD28, the primary receptor for signal 2. In this dissertation, we show that TRAF3 associates with the Linker of Activated T cells (LAT) complex, demonstrating preference for distinct LAT-associated proteins. For delivery of vi

10 signal 3, T cells require stimulation of a cytokine receptor, such as IFN R, for differentiation of a T cell to an effector cell. Upon IFN stimulation, TRAF3 inhibits IFN R-induced early molecular events, which results in the regulation of both canonical and non-canonical IFN R signaling pathways. The results presented in this dissertation highlight the dynamic roles of TRAF3 as a regulator of T cell activation, by regulating multiple T cell signaling pathways. vii

11 PUBLIC ABSTRACT T cells are an important component of the human immune system. T cell activation is tightly regulated to optimize the ability to fight foreign invaders entering the body, while avoiding autoimmunity. However, if T cell regulation is disrupted, such as transplant rejection and autoimmunity can occur. It is thus essential that we understand how T cells are activated. The main function of a protein called TRAF3 is to facilitate protein interactions, which enhance their functions. TRAF3 promotes T cell activation via several different mechanisms. T cells require the activation of three different signals for activation. Signal 1 requires the activation of the T cell receptor (TCR) by foreign antigens. TRAF3 enhances TCR signals by preventing negative regulatory molecules from access to the TCR. The CD28 receptor, signal 2, enhances the signaling power of signal 1 and is required for TRAF3-induced enhancement of T cell activation and function. Our results demonstrate that TRAF3 associates with a group of signaling proteins that interact with CD28. Signal 3 is delivered via receptors for soluble mediators called cytokines, for full activation and specialized functions. TRAF3 restrains the delivery of signal 3, by associating with and inhibiting signaling by the receptor for the cytokine Type 1 Interferon. Together, these three signaling pathways are required for a normal T cell response to occur. The information we gained for how TRAF3 regulates each pathway provides a better understanding of how our immune system is optimized, essential for the design of immunotherapies. viii

12 TABLE OF CONTENTS LIST OF FIGURES... xi LIST OF ABBREVIATIONS... xiii CHAPTER I: INTRODUCTION... 1 TRAF3...1 T cell TRAF3...3 Three signal model of T cell activation... 6 Signal 1: TCR signaling... 7 Early TCR activation...11 Lck Inhibitors of early TCR signaling...12 Csk PTPN Dissertation objectives for signal Signal 2: CD28 activation TCR/CD28-induced SMAC formation...20 LAT complex...21 Dissertation objective for signal Signal 3: cytokine stimulation IFN R1 signaling pathway...24 Canonical type 1 IFN signaling...25 Non-canonical type 1 IFN signaling...26 Tonic IFN R signaling...28 TRAF3 in type 1 IFN production and IFN R signaling...29 PTPN22 in IFN R signaling...31 Dissertation objective for signal Overall dissertation objective CHAPTER II: MATERIALS AND METHODS Mice Cell lines Viral RNA constructs, cell culture and transduction Production of CRISPR subclones Western blot analysis Immunoprecipitation Whole cell lysates Cell fractionation HEK 293T cell transfection Cell numbers, cytokine and protein production with prolonged type 1 IFN stimulation Flow Cytometry Total Internal Reflection Fluorescence (TIRF) microscopy Quantitative PCR ix

13 Statistical analyses CHAPTER III: TRAF3 ENHANCES TCR SIGNALING BY REGULATING THE INHIBITORS CSK AND PTPN Introduction Results The impact of TRAF3 upon early TCR signaling events...51 TRAF3 association with Csk in T cells...52 Csk regulation by TRAF TRAF3 regulation of PTPN22-Csk association...55 TRAF3 regulation of PTPN22 localization...56 Conclusions CHAPTER IV: TRAF3 ASSOCIATION WITH THE LAT COMPLEX Introduction Results TRAF3 association with the LAT complex...75 TRAF3 association with wild-type LAT...76 Association of Grb2 family members with TRAF Membrane localization of TRAF3 and Grb2 family proteins...77 TRAF3 and Grb2 association with the TCR/CD28 complex...79 Conclusions CHAPTER V: TRAF3 INHIBITION OF BOTH CANONICAL AND NON- CANONICAL IFN R1 SIGNALING PATHWAYS IN T LYMPHOCYTES Introduction Results TRAF3 inhibition of canonical IFN R1 signaling...90 The impact of TRAF3 and PTPN22 on T cell IFN R activation...92 TRAF3 associates with and inhibits IFN R activation...93 The role of TRAF3 in canonical IFN R-mediated transcriptional regulation in T cells...94 TRAF3 regulation of the non-canonical pathways of IFN R signaling Conclusions CHAPTER VI: DISCUSSION TRAF3 in regulation of signal 1 (TCR) TRAF3 in regulation of Signal 2 (CD28) TRAF3 in regulation of signal 3 (IFN R) Closing remarks REFERENCES x

14 LIST OF FIGURES Figure 1. TCR signaling model. 10 Figure 2. TCR signaling inhibitors. 18 Figure 3. The LAT signalosome. 23 Figure 4. IFN R canonical and non-canonical signaling. 33 Figure 5. Protein levels of TRAF3. 60 Figure 6. Requirement for TRAF3 in Src kinase activation by the TCR complex. 61 Figure 7. TCR signaling in crtraf3 -/- T cells. 63 Figure 8. Loss of TRAF3 decreases TCR signaling and CD3ζ activation. 63 Figure 9. TRAF3 association with Csk in T cells. 64 Figure 10. Role of TRAF3 in the regulation and localization of Csk. 65 Figure 11. Csk expression in T-traf3 -/- T cells. 67 Figure 12. TRAF3, PTPN22 and Csk association. 68 Figure 13. Csk and PTPN22 association in TRAF3-deficient activated T cells. 69 Figure 14. TRAF3 regulation of PTPN22 localization and association with membrane Csk. T cells were stimulated for indicated times via CD3/CD Figure 15. Membrane TRAF3 clustering is unaffected by the loss of PTPN Figure 16. TRAF3 regulation of Csk and PTPN22 in TCR/CD28 signaling. 72 Figure 17. Predicted model of TRAF3 in the LAT signalosome. 82 Figure 18. LAT association with TRAF3 in HEK293T cell transfections. 83 Figure 19. TRAF3 association with Gab family members in HEK293T cell transfections. 84 xi

15 Figure 20. TRAF3 and Gads membrane clustering upon TCR/CD28 stimulation. 85 Figure 21. TCR/CD28-induced membrane clustering of Grb2 and TRAF3. 86 Figure 22. Grb2 and TRAF3 competition for TCR association. 87 Figure 23. Predicted model of TRAF3 regulation of canonical IFN R1 signaling via JAK Figure 24. Requirement for TRAF3 in type 1 IFN-induced STAT1 activation. 101 Figure 25. Levels of IFN R signaling proteins in TRAF3 deficient T cells. 102 Figure 26. Production and selection of crptpn22 -/- clones. 103 Figure 27. STAT1 activation in the absence of TRAF3 or PTPN Figure 28. TRAF3 association with and regulation of the IFN R1 complex. 105 Figure 29. mrna expression of IFN R1 and IFN response genes in TRAF3- deficient T cells. 106 Figure 30. PD-1 surface expression by CD4 + and CD8 + T cells. 107 Figure 31. Predicted model of TRAF3 regulation of non-canonical IFN R1 signaling via JAK Figure 32. Inhibition by TRAF3 of early type 1 IFN-induced Erk activation. 109 Figure 33. Figure 34. Erk activation and proliferation in the absence of TRAF3 and PTPN Predicted model of TRAF3 regulation of the non-canonical IFN R1 signaling pathway via JAK1 and PI3K. 111 Figure 35. Regulatory role of TRAF3 in type 1 IFN-induced Akt activation. 112 Figure 36. Akt activation in the absence of TRAF3 and PTPN xii

16 LIST OF ABBREVIATIONS Abs Antibodies APC Antigen presenting cell Ca 2+ Calcium Csk C-terminal Src kinase EBV Epstein Barr virus Erk Extracellular signal-regulated kinase Gads Grb2-related adaptor downstream of Shc GAS Gamma-activated sequence Grb2 Growth factor receptor-bound 2 IFN Interferon IFN R Interferon alpha receptor IL Interleukin inkt Invariant natural killer T cells IRF9 Interferon regulatory factor 9 ISGF3 IFN-stimulated gene factor 3 ISRE IFN-stimulated response element ITAM Immunoreceptor tyrosine-based activation motif Jak Janus kinase LAT Linker for activation of T cells Lck Lymphocyte-specific protein tyrosine kinase LMC Littermate control LMP1 Latent membrane protein 1 MAPK Mitogen-activated protein kinases MHC Major histocompatibility complex NFAT Nuclear factor of activated T cells NF- B Nuclear factor of kappa B PI3K Phosphatidylinositol-3-kinase PLC Phospholipase C PTK Protein tyrosine kinase PTPases Protein tyrosine phosphatases PTPN22 Protein tyrosine phosphatase non-receptor type 22 SH Src-homology SLP76 SH2-domain-containing leukocyte protein of 76kD SMAC Supramolecular activation complexes SOS Son of sevenless homology STAT Signal transducer and activator of transcription TCR T cell antigen receptor Tcm T central memory Tem T effector memory Th T helper TIRF Total Internal Reflection Fluorescence TLR Toll-like receptors TRAF Tumor necrosis factor receptor (TNF-R) associated factor xiii

17 Treg Regulatory T cell T-traf3 -/- Deletion of Traf3 in T cells Tyk2 Tyrosine kinase 2 WT Wild-type Y Tyrosine Zap70 Zeta-chain-associated protein kinase 70 xiv

18 CHAPTER I: INTRODUCTION TRAF3 The Tumor-necrosis factor receptor (TNF-R) associated factors (TRAF)s are members of a family of adapter proteins. TRAF members are characterized as having similar protein structures, which include a TRAF homology domain at the C-terminus, the N-terminal really interesting gene (RING) and Zinc (Zn) finger domains [1-5]. There are a few exceptions to this general homology in structure; TRAF1 lacks the RING and Zn finger domains, and TRAF7, the most atypical member of the family, lacks the entire TRAFC domain, instead containing a series of WD repeats [6]. The TRAF family of adaptor proteins regulate signaling downstream of immune receptors, being named for their association with TNF-R family members. It was subsequently found that TRAFs also regulate signaling via the T cell antigen receptor (TCR), cytokine receptors, Toll-like receptors (TLR) and other innate immune receptors, such as RIG-I. Due to the high structural similarity among the TRAF members, as well as their ability to heteromultimerize, it is common for multiple TRAF proteins to regulate a single signaling pathway. For example, TCR signaling is regulated by both TRAF3 and TRAF6 [6, 7], and CD40 signaling is regulated by TRAFs 1, 2, 3, 5 and 6 [8]. Although different TRAF proteins can regulate the same signaling pathway, each can have different roles as an inhibitor or enhancer depending upon their protein binding partners. Similarly, the function of TRAF proteins is cell type dependent. For example, in B cells TRAF3 inhibits homeostatic survival, but it doesn t do so in any other cell type examined to date, including T cells, dendritic cells, and macrophages [7, 9-11]. Each TRAF family member has both cell type and receptor-specific roles. 1

19 TRAF3 was first discovered in the mid-1990s, when it was identified as a RING and Zn finger containing protein similar in structure to the already identified TRAF1 and TRAF2 proteins [12-14]. TRAF3 was first shown to associate with the intracellular cytoplasmic domain of CD40, and subsequently with the Epstein Barr virus (EBV)- encoded Latent Membrane Protein 1 (LMP1), a functional CD40 mimic [12, 13, 15, 16]. A conventional Traf3 -/- mouse strain was developed in the late 1990s, displaying early neonatal lethality that precluded detailed functional studies of the strain itself. Adoptive transfer experiments using cells from this mouse led to a recipient phenotype of impaired T cell-dependent humoral responses, low blood glucose, and high levels of corticosteroids [17]. While this mouse suggested role(s) for TRAF3 in neonatal development and in T-dependent B cell responses, its early lethality hindered its use to further understand TRAF3-regulated immune responses. Further advancements in in vivo models of TRAF3 functions were delayed until the technology became available to delete genes in specific cell types. To determine the cell type specific roles of TRAF3 and to avoid the early neonatal lethality caused by lack of TRAF3, the approach of conditional gene deletion was used to develop mice in which the Traf3 gene was surrounded by LoxP sites [9, 18]. First produced by the Bishop Lab, this strain was initially bred to mice expressing the Cre recombinase gene driven by the B cell-specific CD19 promoter [9]. Subsequently, the Traf3 flox/flox mouse was bred to the CD4-Cre transgenic mouse for the deletion of Traf3 in T cells (T-traf3 -/- ), at the double positive stage in T cell development [7], to the CD11c-Cre mouse to delete Traf3 in dendritic cells, and the Lys-Cre mouse to remove Traf3 from macrophages [7, 10, 11]. Using the T-traf3 -/- mouse model, together with TRAF3-deficient human T cell lines, we explored the roles of TRAF3 in T cells. 2

20 T cell TRAF3 Our lab initially utilized the T-traf3 -/- mouse model to identify a potential role for TRAF3 in T cell biology. Traf3 deletion in T cells did not recapitulate the TRAF3- deficient B cell phenotype. Unlike the B-traf3 -/- mouse, all immune cell numbers in the T-traf3 -/- mouse are similar to wild-type littermate control levels [7, 9]. However, these mice display a profound defect in their in vivo responses to infection with an intracellular pathogen, or to immunization with a model T-cell-dependent antigen (Ag) [7]. It is thus clear that their T cells do not function effectively. As there are a number of TNF-R superfamily receptors that co-stimulate the T cell response to Ag (e.g. CD30, CD137, CD120), it was initially thought that the phenotype results from compromised function of these receptors. However, isolated TRAF3-deficient resting T cells from these mice (depleted of T regulatory cells, Treg) unexpectedly show markedly defective in vitro responses to TCR + CD28 stimulation, with reduced proliferative responses and greatly reduced cytokine production. These surprising results led to the identification of TRAF3 as a TCR/CD28 associating protein. Stimulation of T cells with anti-cd3 and anti-cd28 antibodies (Abs) recruits TRAF3 to the TCR/CD28 complex, while stimulation of either receptor alone does not result in TRAF3 association. Analysis of TCR signaling events revealed reduced activation of TCR signaling proteins, including Zeta-chain-associated protein kinase 70 (Zap70), Linker of Activated T cells (LAT), and Extracellular signalregulated kinase (Erk) [7]. These initial findings identified a new role for TRAF3 in T cell activation and function. Further research into the regulatory roles of TRAF3 in T cells revealed additional complexity. 3

21 While the loss of TRAF3 does not alter numbers of mature conventional CD4 + and CD8 + T cells, the effects of TRAF3 deficiency differ in the various T cell subsets. Cytokine production after TCR/CD28 stimulation has more profound defects in CD4 + T cells depleted of Tregs (CD4 + Treg - ) than CD8 + T cells [7]. The loss of TRAF3 in CD4 + T cells decreases the number of naïve T cells compared to the littermate control mouse (LMC), with a reciprocal increase in the effecter/memory CD4 + T cell population. Interestingly, this same phenotype is not seen in CD8 + T cells, where the loss of TRAF3 does not affect naïve or T effector memory (Tem) numbers. Rather, the number of T central memory (Tcm) CD8 + T cells is significantly reduced. The reduction in CD8 + Tcm is the result of impaired signaling by IL-15, a cytokine required for CD8 + Tcm homeostasis [19]. TRAF3 thus displays different regulatory roles in CD4 + and CD8 + conventional T cells. Additional exploration reveals that at least two other subsets of T cells examined to date are regulated by TRAF3. Initial studies of the T-traf3 -/- mouse unexpectedly indicated a 2-3 fold increase in CD4 + Foxp3 + Tregs. The amplified Treg numbers are a consequence of increased thymic Treg development, not from improved survival or proliferation, nor from an elevated number of inducible Treg cells. During Treg development, the absence of TRAF3 does not alter the number of Treg precursors, but mature Treg numbers increase following treatment of precursors with IL-2, reaching numbers 2-3-fold greater than that of the LMC respective populations. Experiments based upon this finding demonstrated that loss of TRAF3 in T cells enhances IL-2 receptor signaling, which not only drives increased production of mature Tregs but stimulates their earlier development [20]. 4

22 Deletion of TRAF3 at the CD4 + CD8 + stage does not affect conventional T cell numbers, yet increases mature Treg numbers, and surprisingly drastically reduces the numbers of invariant natural killer T cells (inkt). Decreased TCR function in the absence of TRAF3 halts the development of inkt cells at developmental stage 2. In the T-traf3 -/- mouse, a ~10-fold decrease in both the percentage and number of inkt cells is observed in the liver and thymus, with a 2-fold decrease in the spleen. The transition of inkt cells from stage 1 to 2 in development requires TCR-induced upregulation of the transcription factor T-bet and the IL-2 chain (CD122). In TRAF3-deficient T cells, the inkt transition from stage 1 to 2 does not upregulate T-bet. The addition of T-bet to TRAF3-deficient inkt cells via retroviral transduction rescues the developmental defect, increasing inkt numbers. The lack of TRAF3 impairs IL-15 mediated T-bet and CD122 production, preventing the expansion and survival of inkt cells [21]. The focus of this dissertation is to define the various roles of TRAF3 in T cell activation and function. TRAF3 functions as a regulator of many different immune receptors. Thus, I predict that TRAF3 regulates each of the 3 signals required for full T cell activation: those provided by the TCR, costimulatory receptors, and cytokine receptors. My findings, presented in this dissertation, address the molecular mechanisms by which TRAF3 enhances early TCR-mediated signals, associates with the TCR/CD28 complex and regulates the type 1 interferon alpha receptor (IFN R) in T cell signaling. The 3-signal model of T cell activation is discussed below. 5

23 Three signal model of T cell activation The three signal model of T cell activation requires stimulation of three different receptors on the T cell by antigen presenting cells (APC). Without induction of all three types of signals, proper T cell activation and function will not occur [22, 23]. In the case of signal 1, the primary mode of T cell activation occurs when the TCR, expressed on the T cell surface, comes into contact with its cognate peptide antigen, loaded into the peptide-binding pocket of the MHC molecules expressed by an APC [24]. The interaction between the TCR and its antigen causes the phosphorylation and activation of the immunoreceptor tyrosine-based activation motifs (ITAMs) present in the cytoplasmic domains of the CD3 subunits of the TCR complex, the earliest event in T cell activation [25-27]. Induction of the TCR-mediated signaling cascade ultimately leads to transcriptional activation via several pathways, including those mediated by nuclear factor of kappa B (NF- B), nuclear factor of activated T cells (NFAT), and various transcription factors downstream of mitogen-activated protein kinases (MAPK) [28]. These three pathways regulate TCR-induced increases in cell proliferation, calcium influx, metabolism, differentiation, and cytokine production [23]. T cells require engagement of a co-stimulatory receptor, signal 2, together with induction of TCR signaling for enhanced proliferation, cytokine production, cell survival, and altered cellular metabolism to occur [23]. While there are several costimulatory receptors, including CD28 and various TNF-R family members, the more commonly studied enhancer of TCR signaling is CD28. Activation of the CD28 receptor occurs upon association with the ligands CD80 and CD86 (previously known as B7-1 and B7-2) expressed on the APC [29]. The CD28 receptor induces the activation of phosphatidyl 6

24 inositol 3-kinase (PI3K), and together the TCR and CD28 converge at PI3K to enhance downstream signaling pathways [30]. TCR and CD28 signaling are required for normal T cell functions, including cytokine production and proliferation, but a third signal is required to direct T cell differentiation. Considerable data have identified a third signal, cytokine receptor stimulation, as being required for T cell activation and polarization. Cytokine receptor stimulation drives the T cell to differentiate into specific phenotypic subsets with specialized functional roles. Macrophages and dendritic cells release cytokines that are effective in aiding T cell survival, and establishment of effector and memory populations [31]. In CD4 + T cells, cytokine stimulation drives the production of transcription factors that differentiate the T cell into one of three T helper (Th) effector cell subtypes: Th1, Th2 or Th17 [28], or into inhibitory Treg cells [32]. Unlike CD4 + T cells, cytokine stimulation of CD8 + T cells does not lead to known polarization into different subtypes. Rather, CD8 + T cells require cytokine stimulation for the induction of their survival, effector functions, and the production of a memory population [33, 34]. With activation of both TCR/CD28 receptors (signal 1/2) but in the absence of cytokine receptor stimulation (signal 3) T cells cannot perform specific T cell effector functions, making signal 3 essential for proper function in both CD4 + and CD8 + T cells. Signal 1: TCR signaling In the early 1980s, T cell biologists first identified the proteins that make up the TCR complex [35]. Initially the TCR was described as a heterodimer, which consists of an acidic α-chain and a basic β-chain. The genes encoding these chains consist of 7

25 multiple segments which undergo rearrangement during development, causing variability in the antigen binding sites. At the cell membrane, TCR chains form a complex with a set of proteins collectively referred to as CD3 (initially called T3). Both the TCR itself and CD3 are required for TCR complex-induced IL-2 production, calcium (Ca 2+ ) influx and proliferation [35-37]. During this time period, researchers also identified the involvement of the enzyme phospholipase C (PLC) 1 in TCR-induced Ca 2+ influx into the cytoplasm from the endoplasmic reticulum. PLC 1 initiates Ca 2+ influx by catalyzing the hydrolysis of membrane-bound phosphatidylinositol-4,5-bisphosphate (PIP2) to inositol-1,4,5-trisphosphate (IP3). Exploration of TCR-induced phosphoinositide breakdown in T cells identified a role for PLC in TCR signaling, though the exact PLC isoform involved was initially unknown [38]. In the early 1990s, many immunologically relevant receptors (collectively referred to as the Ig gene superfamily) were studied, revealing the importance of tyrosinecontaining signaling motifs in their cytoplasmic domains. The cytoplasmic domains of the TCR and chains do not have protein tyrosine kinase (PTK) activity themselves, but the CD3 complex proteins contain tyrosine residues. This suggested that the activation of the TCR was regulated by both PTKs and their antagonists the protein tyrosine phosphatases (PTPases). A T-cell specific PTK discovered in 1985 called Lck, a Src-family member, was shown to have a major impact in TCR signaling [39]. In the 1990s Lck was discovered to associate with the co-receptors CD4 and CD8 on T cells and transmit phosphorylation-based signals to the TCR complex [40, 41]. Congruently, researchers also identified the TCR subunit CD3ζ and the phosphorylation of its ITAMs 8

26 [26, 42, 43]. Phosphorylation of the ITAMs was predicted to recruit proteins to the TCR, inducing downstream activation. One of the first TCR-associating proteins to be identified was the PTK ζ-chainassociated protein of 70kD (Zap70) [44]. Zap70 recruitment to the TCR requires Lck phosphorylation of the TCR ITAMs, and Lck is also required for the phosphorylation and subsequent activation of Zap70 [44]. The identification of Zap70 led to the thorough exploration of its role phosphorylating other key TCR proteins, including linker for activation of T cells (LAT), PLC-γ1 and SH2-domain-containing leukocyte protein of 76kD (SLP76), following TCR induction [45-47]. Another major breakthrough in characterizing TCR-mediated signal transduction came with the usage of immunoprecipitation to reveal TCR-associated proteins. Immunoprecipitation led to the initial identification of LAT, a transmembrane associated adaptor protein, which upon TCR activation nucleates a signaling complex [46-48]. The LAT complex regulates the activation of multiple downstream signaling pathways that are responsible for a majority of currently known TCR-induced T cell functions [49], to be discussed in more depth later in this document (Figure 1). 9

27 Figure 1. TCR signaling model. T cell activation occurs when the TCR, expressed on the T cell surface, comes into contact with its cognate peptide antigen, loaded into the peptide-binding pocket of the MHC molecules expressed by an APC. Full T cell activation occurs with additional stimulation through the co-stimulatory receptor CD28, which occurs by interacting with the APC expressing ligands CD80 and CD86. The interaction between the TCR and its antigen causes Lck-induced activating phosphorylation of the ITAMs present in the cytoplasmic domains of the CD3 subunits of the TCR complex. Phosphorylated ITAMs recruit Zap70, which is phosphorylated and activated by Lck. Activated Zap70 phosphorylates the adapter protein LAT, inducing the formation of the LAT complex through the recruitment of Grb2. Multiple Grb2 molecules associate with SOS1, forming the LAT complex structure. Gads association with LAT allows for the sequestration of SLP-76. PLC- 1is then recruited to the LAT complex through an association with SLP-76 and LAT which then associates with PLC- 1. TCR induction increases T cell metabolism, Ca 2+ influx, cytoskeletal rearrangement and cytokine production. 10

28 Early TCR activation Lck Lck was first discovered in 1985 and was subsequently identified as the key kinase required for initial TCR activation. The Src-homology (SH)2 and SH3 domains of Lck are essential for regulating cellular localization [50] and activation of its catalytic domain [51]. In T cells, Lck is found in the plasma membrane, associated with the receptors CD4 and CD8 [52]. The c-src kinase (Csk) phosphorylates Y505 of Lck, resulting in the reduction of Lck catalytic activity [53]. Phosphorylation of Y505 induces a conformational change in Lck, that causes an interaction between the polyproline type II helix in the SH2 kinase linker region and the SH3 domain, to prevent Lck activation by blocking phosphorylation of the activating residue Y394 [54]. Lck activation occurs after dephosphorylation of Y505 by the phosphotyrosine PTPase CD45 [53] and the autophosphorylation of Y394 [55]. Activation of Lck is controlled by equilibrium between phosphorylation and dephosphorylation of both the activating site (Y394) and the inhibitory site (Y505) [56]. Phosphorylation of the TCR complex CD3 ITAMs is performed by Lck quickly after TCR activation. To initiate a quick response, the T cell maintains a mixed pool of activated (Y394) and inhibited (Y505) Lck in the resting T cell [57]. With a pool of pre-activated Lck, induction of TCR signaling is extremely fast and sustained by the TCR-induced increase in Lck Y394 [56]. Four major forms of Lck have been identified based on Lck phosphorylation patterns. In unstimulated human CD4 + T cells, 14% of Lck is classified as closed-inactive (phosphorylated at Y505), 48% is primed (unphosphorylated), 17% is py394 active (phosphorylated at Y394) and 21% is 11

29 DPho-active (phosphorylated at both Y394 and Y505) [56]. Unlike primary human T cells, the frequently-used immortalized Jurkat human T cell line contains almost equal amounts of each major form of Lck [56]. This dynamic pool of Lck activation is required for the quick response to TCR stimulation and subsequent prolonged signaling. Tight regulation of Lck is required for controlling TCR signaling intensity and duration. Inhibitors of early TCR signaling In TCR signaling, the homeostatic balance required for early TCR induction and inhibition is regulated by two different mechanisms, PTKs and PTPases. Below, I address the effects of the PTK Csk and the PTPase (PTPN22) on regulating TCR signaling intensity through the phosphorylation and de-phosphorylation of TCR signaling proteins, specifically Lck (Figure 2). Csk Partanen et al were the first to identify the Csk gene [58]. Analysis of homology between the human gene sequence and that of rat showed 98% conservation of the amino acid sequence for Csk. Due to the high conservation of Csk between species, researchers hypothesized that Csk may be important for cellular regulation and/or metabolism [59]. Characterization of Csk led to the discovery of its inhibitory role as a PTK that downregulates the Src family PTKs. Csk inhibits the Src family kinases, including Lck, by phosphorylating the inhibitory tyrosine amino acid located at the target kinase s carboxyl terminus [53, 60, 61]. To further explore the role of Csk, conventional Csk -/- mice were developed. Homozygous embryos lacking Csk died at developmental days 9 and 10 [62, 12

30 63]. Furthermore, examination of the fibroblasts isolated from the Csk-deficient embryos showed an increase in the activation of the Src family kinases [50]. Structurally, Csk is similar to other Src family kinases in containing three domains, the SH2, SH3 and catalytic domains. The association between Csk and Src kinases requires both the Csk SH2 and SH3 domains, although all three domains are essential for Csk-driven inhibition of the Src kinases [61, 64]. Csk is expressed in a wide range of cell types, but for the purpose of this dissertation, the role of Csk in T cells will be the focus of discussion. Csk is required for T cell signaling and is one of the most well-known TCR signaling inhibitors. Inhibition of Csk with a small molecule drug inhibitor increases and prolongs TCR activation, leading to an increase in T cell proliferation. Furthermore, removal of Csk increases the activation of Lck by Y394 phosphorylation, although the phosphorylation of Y505 does not decrease in a reciprocal manner [65]. For complete TCR signaling to occur, the inhibitory effects of Csk on Lck must be prevented. To allow full TCR activation, Csk translocates away from the membrane to the cytoplasm, allowing for increased Lck activation. In T cells, Csk is mostly localized to two subcellular compartments, with ~80% of Csk found in the cytoplasm, while the remaining 20% is confined to the plasma membrane [53]. In human T cells, 2 minutes after TCR stimulation, the majority of Csk has translocated out of the plasma membrane. Inhibition of TCR signaling occurs with the return of Csk to the membrane between 3-5 minutes post-tcr stimulation [66]. Over-expression of a Csk mutant lacking the catalytic domain in unstimulated Jurkat T cells increases phosphorylation of the TCR- CD3-ζ subunit and downstream activation, prompting the spontaneous activation of the 13

31 IL-2 promoter [66]. The movement of Csk in and out of the plasma membrane is essential for the regulation of duration and strength of TCR activation. While the studies discussed above elucidated a role for Csk in TCR signaling, the mechanisms for Csk association with the plasma membrane and translocation to the cytoplasm are still being clarified. Initially, phosphorylation of the transmembrane adapter protein phosphoprotein associated with glycosphingolipid-enriched microdomain (PAG) at Y317 was thought to sequester Csk in the membrane of resting T cells [67-69]. The TCR-induced dephosphorylation of PAG was thought to release Csk for translocation to the cytoplasm. Recently this idea has been contested, with the production and characterization of PAG-deficient T cells. These T cells do not recapitulate the phenotype displayed in Csk-deficient T cells [70, 71]. This led researchers to suggest that PAG is not the only protein with which membrane Csk can associate. Several other transmembrane anchor proteins, including Lck-interesting molecule (LIME) and vascular endothelial cadherin (VE-cadherin), have been identified as Csk-associating proteins [69, 72, 73]. While the protein that removes Csk from the membrane is unknown, the protein that brings Csk from the cytoplasm back to the membrane to inhibit TCR activation has been identified as Focal adhesion kinase (FAK). In a FAK-deficient transformed human T cell line, loss of FAK decreases Csk expression levels at the membrane and TCR complex [74]. The regulation of Csk localization and function during TCR signaling has been extensively studied, but modifications to Csk itself have been mostly unexplored. Not long after the discovery of Csk, it was determined that Csk can be phosphorylated on both tyrosine and serine sites. Little is known about the effect tyrosine phosphorylation 14

32 has on Csk. Phosphorylation of Csk at S364 by camp-dependent protein kinase (PKA) inhibits the catalytic activity of Csk [75]. Therefore, the measurement of Csk S364 phosphorylation is a tool for determining Csk activity levels in the cell. PTPN22 PTPase non-receptor type 22 (PTPN22; also known as Lyp and PEP) was first cloned in 1992 from mouse splenic cdna [76]. The PTPN22 protein is expressed almost exclusively in hematopoietic cells, with robust expression in the thymus and spleen [76]. Structurally, PTPN22 contains an N terminal phosphatase domain, 5 C-terminal PESTcontaining sequences and a 300-amino acid interdomain that separates the two [76]. Human and mouse PTPN22 protein sequences share high conservation in the catalytic and C-terminal domains, although the conservation between species is reduced in the interdomain [77]. The PEST sequence is often indicative of proteins with fast protein degradation rates [76], although pulse-chase experiments have shown an increased halflife of PTPN22, greater than 5 hours [78]. Localization of proteins to different cellular compartments allows a wide range of cellular functions to occur. Cell localization of PTPN22 has been at times controversial. Initial research observed PTPN22 localization predominately in the nucleus of HeLa transformed epithelial cells [79]. This is not surprising, as the P4 domain of PTPN22 contains a potential nuclear localization sequence [79]. However, other researchers utilized an epitope-tagged PTPN22 to detect PTPN22 localization to the cytoplasm and peri-nuclear compartments [80]. Despite this discrepancy, the data strongly suggest that PTPN22 localization occurs in the cytoplasm with a smaller fraction in the nucleus. 15

33 Research on PTPN22 has increased in the last 10 years, with the discovery of the human PTPN22-C1858T missense SNP in exon 14, which leads to the amino acid change R620W. Expression of PTPN22 R620W is associated with an increased predisposition to autoimmune diseases, including type 1 diabetes, rheumatoid arthritis and systemic lupus erythematosus [81-83]. Predominately, the PTPN22 R620W SNP is found in Caucasians [84]. Due to the PTPN22 R620W association with autoimmune diseases, the function of PTPN22 has been largely explored for its potential roles in immune cells, including T cell activation and function. PTPN22 plays an inhibitory role in the TCR signaling pathway. Thymocytes and effector/memory T cells isolated from the Ptpn22 -/- mice display increased TCR activation, following anti-cd3 and anti-cd28 Ab stimulation. These T cells display an increase in Lck Y394 phosphorylation, proliferation and Ca 2+ influx [85-87]. The inhibitory role of PTPN22 in TCR signaling occurs through the dephosphorylation of key TCR signaling proteins, including CD3ζ, Lck and Zap70 [77]. PTPN22 dephosphorylates the activating sites of both Lck and Zap70, Y394 and Y493 respectively, and not their regulatory sites, Lck Y505 and Zap70 Y319 [88]. In PTPN22- deficient T cells, Lck phosphorylation increases after TCR stimulation [85]. A similar increase in the phosphorylation of both Lck and CD3ζ after TCR induction was observed in human T cells treated with a PTPN22 chemical inhibitor [89]. Minor alterations in PTPN22 can result in drastic changes to TCR signaling, as is seen in T cells expressing the mutant PTPN22-R620W. Expression of PTPN22-R620W in primary T cells results in an approximately 50% increase in PTPN22 catalytic activity, increasing the inhibitory effect of PTPN22 on TCR signaling [89]. 16

34 More recent research has been conducted to determine the mechanism by which PTPN22 inhibits TCR signaling. These studies led to the discovery of an association between Csk and PTPN22. It has been estimated by Vang et al. that 50% of PTPN22 in resting T cells is associated with Csk, while only 6% of total Csk is associated with PTPN22 [89]. For optimal association in T cells, the amino acid R620 of PTPN22 and the SH3 domain of Csk are required [89]. T cells expressing the point mutant PTPN22- R620W display a decrease in Csk:PTPN22 association and subsequently less effective at inhibiting TCR signaling [89]. Interestingly, T cell association between Csk and PTPN22 occurs in the cytoplasmic region, and upon TCR activation the association dissipates. After PTPN22 disassociates from Csk, PTPN22 translocates to the plasma membrane where it inhibits TCR signaling by dephosphorylating Lck [89]. In PTPN22-sufficient Jurkat T cells transfected with PTPN22-R620W more WT PTPN22 associates with the plasma membrane compared to PTPN22-R620W [89]. This suggests that recruitment of PTPN22 to the membrane contributes to the inhibition of TCR signaling. Dissertation objectives for signal 1 TCR signaling is the first and most fundamental signal for T cell activation. Activation of the TCR is antigen-specific, inducing clonal expansion of a T cell population specifically designed to combat a foreign invader. Tight regulation of TCR signaling is required for proper T cell activation and function. TRAF3 is a newlyrecognized regulator of this process. Thus it is essential that we understand how TRAF3 regulates the TCR signaling pathway. 17

35 Figure 2. TCR signaling inhibitors. TCR signaling is inhibited through two different mechanisms. The first mechanism is inhibition by the kinase Csk. The plasma membrane-associating Csk regulates TCR activation by sequestering Lck away from the TCR, preventing Lck-induced phosphorylation of the TCR-ITAMs. Importantly, Csk phosphorylates Lck at Y505, the inhibitory site, causing a conformational change in Lck that prevents Lck phosphorylation at the activating site, Y394. The second TCR inhibitor is the phosphatase PTPN22. After TCR signaling, PTPN22 translocates to the membrane from the cytoplasm. At the membrane, PTPN22 dephosphorylates the TCR/CD3-ITAMs and the activating site Y394 on Lck. Dephosphorylation of these TCR signaling proteins prevents prolonged TCR activation. 18

36 Signal 2: CD28 activation T cells activated through the TCR alone (signal 1), without additional signals, become anergic and unresponsive to further TCR stimulation. To avoid the anergic state, the T cell must be activated through both the TCR and a co-stimulatory receptor, such as CD28 and/or TNFR family members (e.g. OX40, 4-1BB, TNFR2, CD30) [90]. Costimulation enhances the TCR signaling pathway, with CD28 being the most effective costimulator of TCR signaling [23]. CD28 activation occurs through its association with the CD80 and CD86 ligands on the APC [29]. Stimulation of CD28 induces cellular proliferation and survival, cytokine production, and increased metabolism [30]. Additionally, CD28 signaling converges with the TCR signaling pathway at PLC- 1, a protein activated by the LAT complex, to enhance TCR signals [91]. CD28 activation recruits the p85 subunit of PI3K to the CD28 cytoplasmic domain, allowing subsequent recruitment of p110, the catalytic subunit of PI3K [29]. Following this event, PI3K converts membrane-associated phosphatidylinositol 4,5- bisphophate (PIP2) to phosphatidylinositol 3,4,5-trisphophate (PIP3), providing a docking site for 3-phosphoinositide-dependent protein kinase 1 (PDK1). PDK1 can now recruit and activate the kinase Akt and various downstream Akt-induced proteins, some of which include the transcription factors NF- B and NFAT, and the glucose transporter GLUT1. Cell survival and proliferation occurs after Akt activation, via NF- B translocation to the nucleus and initiation of transcription of pro-survival genes [30]. Another downstream target of Akt is glycogen-synthase kinase 3 (GSK-3), the nuclear exporter of the transcriptional regulator NFAT. Inhibition of GSK-3 prolongs NFAT 19

37 nuclear localization, increasing the production of IL-2 [92]. Akt activation after TCR/CD28 stimulation also increases the expression level of the glucose transporter GLUT1, promoting cellular uptake of glucose and increased glycolysis [93, 94]. In a previous study in our lab, TRAF3 was found to associate with the TCR/CD28 complex upon stimulation with both anti-cd3 and anti-cd28 stimulatory Abs, but not with stimulation of either receptor alone [7]. Thus, TRAF3-mediated enhancement of TCR signaling requires CD28 stimulation via several possible mechanisms: 1) TRAF3 associates with CD28 directly after TCR/CD28 stimulation, and/or 2) TRAF3 interacts with the LAT complex, and TCR/CD28 stimulation allows for LAT complex association with TCR/CD28. One objective of experiments presented in this dissertation was to explore the potential association between TRAF3 and the LAT complex. TCR/CD28-induced SMAC formation In the late 1980s, researchers identified the formation of protein clusters surrounding the TCR after association with antigenic peptide-mhc complex on the surface of the APC [95-97]. Further research identified structural characteristics that make up the supramolecular activation complexes (SMAC; also known as an immunological synapse (IS) [98]), including the central region (c-smac) which is surround by the peripheral ring (p-smac) [95]. The c-smac consists of TCR signaling proteins, including but not limited to CD28, Lck and LAT [95, 99, 100]. The furthest ring away from the TCR/CD28 and c-smac is the p-smac, which consists of integrin Intercellular adhesion molecule-1 (ICAM-1), Talin (a cytoskeletal-bound protein) and the co-receptor CD2 [95, 99]. The formation of the SMAC gives stability to the TCR:MHC 20

38 association, promoting full T cell activation, including TCR signaling [98, 99, 101, 102]. Formation of SMACs requires TCR-MHC association, as well as co-stimulation. Blocking co-stimulation alters the structural integrity of the SMAC by decreasing its size and the amounts of proteins that cluster, including the LAT complex [103]. LAT complex LAT is an essential transmembrane adapter protein required for early T cell development and TCR signaling [104]. Inducible deletion of LAT from mature T cells inhibits TCR-induced proliferation and long term survival [105]. As an adapter protein, phosphorylated LAT is capable of directly binding several SH2 domain-containing proteins, including Growth factor receptor-bound protein 2 (Grb2), Grb2-related adaptor protein downstream of SHC (Gads), Grb2-related adapter protein (Grap), PLC- 1, and indirectly binding PI3K [49, ]. TCR-induced phosphorylation of LAT occurs predominately by Zap70, although Lck and the PTK Itk have been shown to play a role as well [48, ]. Grb2 and Gads are members of the Grb2 adaptor protein family, whose members enhance downstream signaling pathways, including the LAT complex [115]. Grb2 can bind to the phosphorylation sites Y171, Y191 and Y226 on LAT to recruit the guanine nucleotide exchange factor Son of sevenless homology (SOS) 1 and the E3-ubiquitin ligase c-casitas B-Lineage Lymphoma (Cbl) Proto-Oncogene [49]. Multiple Grb2 proteins associate with SOS1, which provides the lattice for formation of the LAT signaling microcluster [116]. LAT microcluster formation in turn draws in more signaling proteins that are required for the initiation of Ca 2+ influx and cytokine 21

39 production [116]. Similar to Grb2, Gads also associates with the phosphorylation sites Y171 and Y191 on LAT. The recruitment of SLP-76 (SH2 domain-containing leukocyte protein of 76 kda) to Gads allows SLP-76 to induce phosphorylation of PLC- [117]. PLC-γ1 is an important activator of TCR-induced phosphoinositide breakdown and influx of Ca 2+ [35, ] (Figure 3). To enhance signaling intensity, CD28 signaling converges with the TCR signaling pathway at PLC- 1, downstream of the LAT complex, to enhance Ca 2+ influx and IL-2 production [91]. Dissertation objective for signal 2 Induced TRAF3 association with the TCR/CD28 complex requires both TCR and CD28 signaling (signals 1 and 2). The objective in this dissertation is to determine the mechanism by which CD28 signaling drives TRAF3 recruitment to the TCR/CD28 complex. Our working hypothesis is that TRAF3 associates with the LAT complex in the resting T cell. TCR/CD28-induced SMAC formation results in the clustering of LAT in close proximity to the TCR/CD28 receptors. Thus, TRAF3 associates with the LAT complex and requires CD28 stimulation for formation of the SMAC, bringing TRAF3 into the TCR/CD28 complex. 22

40 Figure 3. The LAT signalosome. CD3 and CD28 stimulation initiates formation of the LAT complex. Phosphorylation of LAT by Zap70 recruits the Gab family members Grb2 and Gads. Grb2 associates with Y226 and Y171 on LAT and recruits SOS1 to the LAT complex. Gads association with phosphorylated LAT at Y191 also aids in the production of the LAT signalosome by recruiting SLP-76 and PLC 1. Together, Grb2 and Gads are required for downstream transcriptional activation, cytokine production and increased metabolism. 23

41 Signal 3: cytokine stimulation While signals 1 and 2 are widely regarded to occur through the association of a T cell with a single APC, signal 3 can occur through a second, independent cytokinesecreting cell [22]. In T cell activation, signal three occurs upon binding of the cytokine to its respective receptor on the T cell. Signal three, also called polarization, directs the differentiation of the T cell into one of the various T helper (Th) subsets, such as Th1 and Th2 [121]. The distinct T cell subsets are characterized by the cytokines they secrete, which determine their particular effector functions. For example, Th1 cells secrete TNF- and IFN to combat intracellular bacterial and viral infections. Th2 cells secrete IL-4, IL-5 and IL-13 to protect against extracellular parasites and to regulate humoral immunity [22]. While there are several cytokine receptors that qualify as signal 3 receptors on T cells, this dissertation will focus upon the regulatory role of TRAF3 in IFN / -induced IFN R signaling. IFN R1 signaling pathway Type 1 interferons (IFN) are polypeptides secreted almost ubiquitously from infected cells to limit the spread of infections, modulate innate immune responses and activate the adaptive immune system [122]. Of the 7 type 1 IFNs, IFN and IFN are the most well-characterized and widely expressed family members [123]. IFN is expressed predominately by hematopoietic cells; it is encoded by 13 different genes in humans and 14 in mice [122, 124]. In contrast, IFN is expressed by most cell types and is encoded by one gene [122]. The ubiquitously expressed IFN receptor (IFN R) is a 24

42 heterodimeric transmembrane receptor that is composed of two subunits: IFN R1 and IFN R2. Binding of type 1 IFN to IFN R results in the induction of signaling pathways referred to as canonical and non-canonical. The canonical pathway is Janus kinase (JAK)/ signal transducer and activator of transcription (STAT) dependent and is the more commonly researched IFN R signaling pathway. The non-canonical pathways signal through PI3K or MAPK, and less is known about these pathways in response to type 1 IFN [125]. Both canonical and non-canonical pathways originate from JAK1/ Tyrosine kinase 2 (Tyk2) activation. Previous research identified an association between TRAF3 and JAK1 [126], thus TRAF3 can potentially regulate both canonical and non-canonical pathways of IFN R signaling. Canonical type 1 IFN signaling As depicted in Figure 4, upon binding of type 1 IFN to IFN R in the canonical pathway, the kinases JAK1 and Tyk2 are phosphorylated and activated. Together these two kinases phosphorylate the cytoplasmic transcription factors STAT1 and the constitutively IFN R-bound STAT2 [123]. Phosphorylated STAT1 and STAT2 dimerize to form a complex that translocates away from IFN R to the nucleus. Prior to entry into the nucleus, the STAT1/2 complex recruits the transcription factor termed Interferon regulatory factor 9 (IRF9) to form a tri-molecular complex designated IFNstimulated gene factor 3 (ISGF3). The ISGF3 complex then migrates to the nucleus where it binds to the IFN-stimulated response element (ISRE) DNA sequence to initiate transcription of IFN-induced response genes including OAS, MX1 and IRFs [122]. There are hundreds of interferon response genes that harbor and are regulated by an ISRE 25

43 sequence which ultimately regulate the antiviral response. Alternatively, STAT1 can also homodimerize and translocate to the nucleus, where it interacts with the Gammaactivated sequence (GAS) to induce production of the chemokine CXCL9 and the transcriptional regulator IRF1 [122]. In 2010 Terawaki et al identified the sequence between the and bp region of the PD-1 promoter as the region required for transcription of the PD-1 gene, Pdcd1 [127]. This gene plays an important inhibitory role in T cell activation, and is identified as a key checkpoint inhibitor; it is currently of great interest as a target for cancer immunotherapy [128]. Contained within this newly identified region of Pdcd1 is an ISRE binding site. The identification of the ISRE led investigators to classify PD-1 as an additional IFN-responsive gene. In the 2B4.11 T cell line, PD-1 is upregulated on T cells upon TCR and IFNα stimulation, but not with IFNα stimulation alone [127]. This suggests cross-talk between TCR signaling (signal 1) and IFN R (signal 3) for the induction of IFN response genes, an interaction we explore further in Chapter 5 of this dissertation. Non-canonical type 1 IFN signaling As stated above, there are two identified non-canonical IFN R signaling pathways. The first pathway signals through MAPK and is illustrated in Figure 4. Upon stimulation of IFN R by type 1 IFN, Tyk2 associates with and phosphorylates Vav, a guanine nucleotide exchange factor [129]. Tyrosine phosphorylation of Vav allows for the Vav-induced GDP to GTP exchange on Ras-Related C3 Botulinum Toxin Substrate 1 RAC1, which allows RAC1 to activate downstream MAPKs, including Erk [130]. 26

44 Type 1 IFN-induced Erk phosphorylation leads to an increase in cellular growth, differentiation and STAT1 phosphorylation at S727 [131, 132]. In T cells, type 1 IFN alone is not sufficient for IFN R-induced Vav activation. For this event, T cells also require signaling through the TCR. Interestingly, in Jurkat T cells IFN-induced Erk activation will not occur in the absence of either TCR stimulation or following removal of critical LAT complexassociated proteins. In TCR-deficient Jurkat T cells, IFN R-induced MAPK stimulation is hindered, with no alteration occurring in JAK/STAT signaling, showing that MAPK activation requires TCR and IFN R co-stimulation [133]. Members of the LAT complex, including Slp76, are essential for the IFN-induced activation of Vav [ ]. Similarly, Slp76 is also phosphorylated in response to IFN stimulation at three different tyrosine sites, all of which are required for Erk activation. In response to type 1 IFN stimulation, Vav associates with Slp76. Furthermore, deletion of either Vav or Slp76 in Jurkat T cells inhibits IFN-induced Erk phosphorylation [133]. Together these data indicate that the cross talk between TCR and IFN R signaling is essential for IFNinduced MAPK activation. The second non-canonical IFN R signaling pathway utilizes PI3K. Briefly, JAK1 and Tyk2 phosphorylate the insulin receptor substrates 1 (IRS1) and IRS2. The phosphorylation of IRS1/2 provides a docking site for PI3K, with STAT3 providing further support for the association of PI3K with IFN R1 [131]. Induction of PI3K leads to the activation of Akt, and regulates downstream signaling proteins, including NF B, mechanistic target of rapamycin (mtor) and GSK-3. NF- B can also be regulated by the association of TRAFs and NF B-inducing kinase (NIK). Ultimately, the activation of 27

45 NF- B leads to pro-survival signals and increased gene expression of GTP-binding proteins, as well as that of antigen processing and/or presentation proteins [131]. Similar to Erk, activation of PI3K can also drive the Protein Kinase C (PKC) δ-induced phosphorylation of STAT1 on S727 [136]. Tonic IFN R signaling In addition to an acute response to type 1 IFN, immune cells can also undergo homeostatic activation of IFN R from exposure to low levels of IFN [137]. IFN is continuously produced at levels in the serum just above the limit of detection. The low concentration of IFN allows for tonic signaling to occur in the absence of an acute infection. Constitutive low levels of type 1 IFN secretion was first proposed in the early 1980s by Velio Bocci, who suggested that IFN production occurs at the mucosal lining of the gut and bronchial-associated lymphoid tissue as a normal low-grade adaptive immune response to foreign microbes, although some IFN will disseminate throughout the whole body, resulting in systemic mild exposure [138]. While researchers believed low level IFN exposure occurred, the inability to detect such small quantities of IFN hindered further research into this phenomenon. With the advent of more sensitive IFN detection technology, researchers were able to identify low levels of IFN in the tissue environment [139, 140]. Both IFN mrna and protein have been detected in healthy individuals prior to physical activity or virally-induced IFN production [ ]. These data together identified a new potential role for IFN in chronic low-grade stimulation of cells. Tonic IFN signaling has been proposed as a priming mechanism for cells to produce a faster, more intense response to a higher grade stimulation. The faster 28

46 response to high-dose type 1 IFN stimulation is due to the priming-induced maintenance of increased protein levels of STAT1, STAT2 and IRFs [137]. Cells lacking this priming capacity due to deletion of IFN R1 or the prevention of IFN-β stimulation display decreased basal STAT1 expression levels [145, 146]. IFN-induced cellular priming is not restricted to enhancing IFN R signaling only. Instead, IFN priming allows for faster response times of all JAK1/STAT1 pathways, in response to a variety of ligand-receptor interactions. The reduced STAT1 protein levels in IFN R1 -/- cells also hinder the activation of other JAK1/STAT1 pathways, such as responses to IFN and IL-6 [145, 147]. Thus, regulating the protein levels of STAT1/2 can determine the signaling intensity not only of IFN R1, but also other receptors utilizing the JAK1/STAT1 pathway. TRAF3 in type 1 IFN production and IFN R signaling Research investigating the mechanisms for type 1 IFN secretion identified TRAF3 as an enhancer of Toll-like receptor (TLR)-induced type 1 IFN production. There are three different TLR-induced pathways that regulate the production of type 1 IFN: 1) TRIF-dependent, 2) MYD88 dependent (both TLR-mediated) and 3) a TLRindependent pathway, all of which involve roles for TRAF3 [148]. Despite research into the regulatory role of TRAF3 in IFN production in myeloid cells and fibroblasts, the role of TRAF3 in IFN R1 signaling pathways has gone unexplored. The research supporting the hypothesis that TRAF3 plays a regulatory role in IFN R signaling is at times contradictory. In two anaplastic large cell lymphoma (ALCL) cell lines, called Karpas and Michel, loss of TRAF3 decreases normal cellular 29

47 proliferation and basal Akt phosphorylation, with no change in total Akt protein levels. Additionally, the loss of TRAF3 results in an increase in plasma membrane expression levels of the Phosphatase and tensin homolog (PTEN), an inhibitor of the PI3K/Akt pathway [149]. While these data suggest an enhancing role for TRAF3 in the PI3K/Akt pathway, one caveat to this conclusion is that the cells studied were unstimulated, indicating a need to explore the function of TRAF3 upon type 1 IFN stimulation. The regulatory role of TRAF3 in T cell JAK/STAT pathways has not been explored. Recent data from B lymphocytes support the concept that TRAF3 can play important regulatory roles in these cytokine receptor pathways. Lin et el identified TRAF3 as an inhibitor of the B cell IL-6 receptor signaling pathway. Upon IL-6 stimulation, TRAF3 associates with JAK1 to recruit PTPN22. The close proximity of PTPN22 to JAK1 allows PTPN22-induced JAK1 dephosphorylation. In the absence of TRAF3, inhibition of the IL-6 receptor signaling pathway cannot occur, because there is no longer effective recruitment of PTPN22 [126]. Additionally, in T cells, TRAF3 also inhibit JAK1 and JAK3 in IL-2R signaling through the recruitment and dephosphorylation of the JAK proteins by TCPTP [20]. These findings suggest that there may also be an inhibitory role for TRAF3 in type 1 IFN-induced activation of the JAK/STAT pathway in T lymphocytes. While TRAF3 regulation of JAK/STAT signaling pathways in T vs B lymphocytes are unlikely to be precisely the same, these early results support a new role for TRAF3 in both canonical and non-canonical IFN R signaling pathways. 30

48 PTPN22 in IFN R signaling PTPN22, a known inhibitor of TCR signaling [88, 150], also negatively regulates IFN R signaling in hematopoietic progenitor cells. In the absence of PTPN22, proliferation and activation of progenitors increases with IFN-α stimulation. Furthermore, PTPN22-deficient hematopoietic cells stimulated with recombinant mouse IFN-α4 double their phosphorylation of STAT1 Y701 compared to LMC cells. Moreover, the increase in STAT1 activation in the PTPN22-deficient progenitors is consistent with an increase in levels of the IFN response genes Mx1, Irf4, Osmr and Isg15. In IFN-α-responsive 3T3 fibroblasts, PTPN22 associates with the IFN R signaling complex, and dephosphorylates STAT1 and STAT2 [151]. Consistent with these data and as stated previously, PTPN22 inhibits IL-6 signaling in B cells, through the TRAF3-induced recruitment of PTPN22 to JAK1, allowing PTPN22 dephosphorylation/inhibition of JAK1 [126]. These data suggest a potential role for PTPN22 as a TRAF3-recruited inhibitor of JAK1 and IFN R signaling in T cells. PTPN22 can have varying roles depending on its associating proteins. While PTPN22 inhibits the JAK1/STAT1 pathway in B cell IL-6R signaling, this phosphatase positively regulates the PI3K/AKT pathway downstream of the B cell receptor. In B cells, PTPN22 expression levels directly correlate with Akt phosphorylation on S473. Conversely, levels of the Akt inhibitor SHIP decreased with increased PTPN22 levels, resulting in increased Akt activation. These data suggest PTPN22 enhances Akt activation by inhibiting SHIP in B cells [152]. Similar to B cells, T cell PTPN22 can also positively enhance Akt phosphorylation at S473. In the absence of PTPN22, T cells display decreased levels of phosphorylated Akt at S473 and as a result, increased 31

49 apoptosis [153]. Together these data indicate that PTPN22 has varying contextdependent roles, determined by the associated proteins. Because it is now known that TRAF3 is one such protein, it is essential to determine which regulatory roles of PTPN22 involve its association with TRAF3. Dissertation objective for signal 3 Previous research, including that from our lab, strongly indicates a role for TRAF3 in signal 3 of T cell activation, by regulating both canonical and non-canonical IFN R signaling pathways. The goal of Chapter 5 is to define the role of TRAF3 in IFN R signaling, which may occur through the regulation of PTPN22, or other unknown protein associations. Our working hypothesis is that TRAF3 recruits PTPN22 to the IFN R complex after IFN stimulation, where PTPN22 dephosphorylates/inhibits JAK1. Thus alterations in JAK1 activation could also affect both the canonical and noncanonical signaling pathways. 32

50 Figure 4. IFN R canonical and non-canonical signaling. Upon ligation of IFN R with type 1 IFN, both the canonical and non-canonical pathways are induced. In the canonical pathway, phosphorylation and activation of JAK1 and Tyk2 allows for these kinases to in turn phosphorylate both STAT1 and STAT2. Phosphorylated STAT1 and STAT2 dimerize to form a complex that disassociates from IFN R and recruits the transcription factor IRF9, to form the tri-molecular complex termed ISGF3. The ISGF3 complex then translocates to the nucleus where it binds to the DNA-binding sequences called ISRE and GAS to initiate IFN-induced response genes. The first non-canonical IFN R signaling pathway primarily utilizes MAPK. Following stimulation of IFN R by type 1 IFN, JAK1 and Tyk2 activation leads to the phosphorylation of Vav. Vav activation in turn induces the downstream activation of MAPK/Erk by phosphorylation. Induction of Erk activation ultimately promotes phosphorylation of STAT1 at S727 along with cellular growth and differentiation. In T cells, type 1 IFN-induced Vav activation requires simultaneous stimulation via the TCR. Vav activation occurs through signaling from the TCR-induced LAT complex and IFN R. The second noncanonical IFN R signaling pathway utilizes PI3K. JAK1 and Tyk2 phosphorylate the IRS1 to provide a docking site for PI3K, and activation of STAT3 further stabilizes PI3K binding to IFN R. Induction of PI3K activation leads to the activation of Akt and its downstream signaling pathway, including pathways regulated by NF B. NF- B can also be regulated by TRAF3/TRAF2/cIAP-induced degradation of NIK. Ultimately, the activation of NF- B leads to pro-survival signals, increased expression of GTP-binding proteins, antigen processing and/or presentation proteins, and phosphorylation of STAT1 S

51 34

52 Overall dissertation objective It is clear that TRAF3 has the potential to regulate various receptors with which it associates, playing both inhibiting and enhancing roles for signaling pathways. This context dependency of TRAF3 function is largely determined by its binding partners in signaling complexes. Little is known about how TRAF3 regulates important receptors expressed by T lymphocytes. In this dissertation, I address this knowledge gap by exploring the specific roles of TRAF3 in TCR, CD28 and IFN R signaling, the three types of signals required for complete activation of T cells. 35

53 CHAPTER II: MATERIALS AND METHODS Mice Traf3 flox/flox mice, backcrossed >10 generations to C57BL/6 mice [9], and crossed to Cd4-Cre mice [20], created the T-traf3 -/- strain used here, together with their littermate controls (LMC). Adult mice (2-4 months old) were used for all experiments. Mice were housed in specific pathogen-free conditions and used in accordance with NIH guidelines, under a protocol approved by the Animal Care and Use Committee, University of Iowa. Similar numbers of male and female mice were used interchangeably. Cell lines A subclone of the human CD4 + T cell line HuT78, transfected to stably express CD28 (HuT28.11), was the gift of Dr. Arthur Weiss, University of California, San Francisco [154]. HuT28.11 cells were cultured in RPMI 1640 medium (Invitrogen, Grand Island, NY) supplemented with 100 U/mL penicillin, 100 U/mL streptomycin, 2 mm L-glutamine, 10μM β-mercaptoethanol (all from Invitrogen), and 10% fetal calf serum (FCS; Atlanta Biologicals, Atlanta, GA) (this medium is referred to as BCM10). Subclones of HuT28.11, including shluc, shtraf3, crtraf3 -/- and crptpn22 -/-, and subclones of HuT78, including shgads, shgrb2 and shluc, are described below; cultures of the HuT28.11 clones included 2 mg/ml G418 (Research Products International, Mount Prospect, IL). Viral RNA constructs, cell culture and transduction 36

54 The shgads, shgrb2 and shluc HuT78 clones were a gift from Dr. Jon C.D. Houtman (University of Iowa, Iowa City, IA) and produced as described [116, 117]. Production of HuT28.11 subclones stably expressing the small hairpin (sh) RNA constructs shluc (vector control) and shtraf3 was performed as indicated [116, 155], with minor modifications as follows. Using the htraf3 sequence as a template, shrnas targeting TRAF3 were obtained from the algorithm of Dr. Ravi Sachidanandam ( The following sequences were used for production of shtraf3 (TRAF3-8 sense 5 GAACCTACC GGTCCGTGTGTCCCTGCTCATAAAGTAGTGAAGCCACAG 3 TRAF3-8 antisense 5 GTTCCGAATTCAAAAAATCGTGTGTCCCTGCTCATAAAGTACAT CTGTGGCTTC3 ; TRAF3-14 sense 5 GAACCTACCGGTAACTGGTTATCACTT GTGATAGTAGTGAAGCCACAG 3 TRAF3-14 anti-sense 5 GTTCCGAATTCAAA AAACACTGGTTATCACTTGTGATAGTACATCTGTGGCTTC 3 ) (Integrated DNA Technologies (IDT), Coralville, IA). Both shtraf3-8 and shtraf3-14 were used together to produce the most effective inhibition of TRAF3 expression. To make shrna-containing virus, HEK 293T cells were transfected using lipofectamine 2000 (Invitrogen), according to the manufacturer s instructions. Each transfection included 5μg of each shrna plasmid (plko.1 shtraf3-8 and -14), with viral packaging vectors VSV-G (4μg), and Pax2 (10μg) (all these plasmids were generous gifts from Dr. Jon Houtman, University of IA, Iowa City, IA). This mixture was incubated with the cells at 37 C for 6-8h, after which cells were washed and cultured with 25mL fresh DMEM10 (Invitrogen) supplemented with 100 U/mL penicillin, 100 U/mL streptomycin, 2 mm L-glutamine, 10mM HEPES, 1 x MEM NEAA (all Invitrogen), and 37

55 10% FCS. Culture supernatant containing recombinant virus was collected at 24 and 48h and virus was isolated as described [117]. Virus was resuspended in 1.5mL BCM10. HuT28.11 T cells (3-5 x 10 5 ) were resuspended in 1.5mL of virus-containing supernatant, with 8μg/mL hexadimethrine bromide (Polybrene). Cells were cultured for 1 week, after which shrna-expressing cells were selected with 1μg/mL puromycin (Sigma, St. Louis, MO). Production of CRISPR subclones Guide RNA/Cas9 vector constructs (px330 Addgene, Cambridge, MA) for disruption of the TRAF3 gene were prepared as described [156], using the CRISPR design tool (crispr.mit.edu) maintained by Dr. Feng Zhang (MIT, Cambridge, MA). Two constructs were prepared for both TRAF3 and PTPN22 deletion. For TRAF3, one construct targeted intron 1 upstream of the ATG, and a second targeted exon 5. The double-stranded synthetic oligonucleotides for intron 1 were: 5 CACCG CCATCATATCCTCTCATGCA 3, and 5 AAACTGCATGAGAGGATATGATGGC 3. The exon 5 oligonucleotide pairs were: 5 CACCGGTTCCGATGATCGCGCTGC 3 and 5 AAACGCAGCGCGATCATCGGAACC 3. For PTPN22, one construct targeted intron 1 upstream of the ATG, and a second the intron between exon 1 and 2. The double-stranded synthetic oligonucleotides for intron 1 were: 5 CACCGTAGTACGTAA CCACCTTCTG 3, and 5 AAACCAGAAGGTGGTTACGTACTAC 3. The intron between exon 1 and 2 oligonucleotide pairs were: 5 CACCGATTATTGATGCAAT 38

56 GGCAGC 3 and 5 AAACGCTGCCATTGCATCAATAATC 3 (all IDT). Pairs were annealed and phosphorylated as described [156]. px330 (ID 42230) was cut with BbsI and treated with calf intestinal phosphatase, then purified (QIAquick PCR purification column, Qiagen, Hilden, Germany). Phosphorylated double-stranded oligonucleotides were ligated into the cut vector and ligated DNA used to transform competent E. coli (DH5 Invitrogen). Plasmid DNA was sequenced to verify proper insertion. 2.5 x 10 6 HuT28.11 cells were resuspended in 400μl Optimem (Invitrogen) with 2.5μg of each of the two guide RNA/Cas9 vectors, 0.5μg pegfp-c1 (Clontech, Mountain View, CA), and 5μg double-stranded filler DNA oligonucleotides (random sequence[157]). The cell suspension was electroporated in 4mm cuvettes, 225V for 30 ms (BTX square wave electroporator, Holliston, MA). After a 10 minute rest at 37 C, cells were resuspended in 10mL BCM10 and cultured for 5d. GFP-expressing cells were sorted using a Becton Dickinson FACS Fusion (Franklin Lakes, NJ) at 1 cell/well into 96-well plates. Clones were screened by PCR of genomic DNA using the following primers: TRAF3 targeting primers include 5 CTGAAAGACAGCAGGTCTCAGGCAC 3, and 5 GAATGTATCATATAGGAATTGAGTGG 3 (Int-5R3) and PTPN22 targeting primers are 5' TGG AGT GAG ATG ATG GCT GTG 3' (PTPN22 Fwd) and 5' TGC TAT CCG GTA CCA GAG TG 3' (all IDT). A PCR product of ~100bp (TRAF3) or ~500bp (PTPN22) indicated the desired deletion. DNA samples exhibiting this product were retested with primers specific for sequences within the deleted region (TRAF3: 5 GGTTTCATTGCATAGAGATTAGAATC 3, and Int-5R3 (above) or PTPN22: 5' TGA GAG GGT CAC ATA CAG GAC 3', and PTPN22 Fwd (above)). Clones testing 39

57 negative for the 300bp (TRAF3) or 200bp (PTPN22) intact gene product were screened by Western blot to cofirm disruption of TRAF3 or PTPN22 protein expression. Western blot analysis Western blot analysis was performed as described [126]. Resolution of proteins was performed on SDS-PAGE gels and proteins were transferred to PVDF membranes (Millipore, Billerica, MA). Using 5% milk in TBST (120mM NaCl, 0.08% Tween 20, and 40mM TRIS ph 8.0), blots were blocked for 1 hour at room temperature and then washed. Primary antibodies (Abs) were added and incubated at 4 C overnight. Blots were then washed with TBST and incubated with secondary Ab for 1 hour at room temperature. Using the developer Supersignal West PICO or FEMTO reagents (Thermo Scientific), chemiluminescence was detected on the LAS-4000 low-light camera and analyzed with Multigauge software (Fujifilm Life Sciences, Edison, NJ). Primary Abs used for Western blotting analysis were: mouse anti-csk (BD Biosciences, San Jose, CA), rabbit anti-csk (C-20), rabbit anti-traf3 (H122,), mouse anti-cd3ζ (6B10.2, all from Santa Cruz, Dallas, TX), rabbit anti-ifn R1 (EP899Y), rabbit anti-pcsk S364 (Abcam, Cambridge, MA), rabbit anti-ptpn22 (D6D1H), rabbit anti-psrc 416, rabbit anti-lck, rabbit anti-plck 505, rabbit anti-lat, rabbit anti-stat1, rabbit antipstat1 Y701 (58D6), rabbit anti-pstat1 S727, rabbit anti-akt, rabbit anti-pakt S473, rabbit anti-p44/42 MAPK (Erk1/2), rabbit anti-p44 Y202 /42 Y204 MAPK (Erk1/2), rabbit anti-stat2, rabbit anti-stat2 Y690, rabbit anti-stat3 (all from Cell Signaling, Danvers, MA), mouse anti-flag (M2), mouse anti-ha (HA-7, all from Sigma), mouse anti-py (4G10), and mouse anti-actin (Clone 4, Millipore, Billerica, MA). Secondary 40

58 Abs are as follow: HRP-conjugated goat anti-mouse IgG, ant-rabbit Ig (Jackson ImmunoResearch Laboratories) and light chain specific HRP-conjugated mouse antirabbit Abs (Southern Biotech, Birmingham, AL). Immunoprecipitation Primary mouse splenic T cells were isolated using a Pan T cell negative purification kit (StemCell Technologies, Vancouver, Canada). 30 x10 6 primary T or HuT28.11 T cells or subclones were used/time point. Cells were washed and resuspended in serum-free RPMI 1640 medium and stimulated at 37 C with anti-cd3 (mouse: 145-2C11, human: OKT3) and anti-cd28 stimulatory mabs (mouse: 37.51, human: CD28.2) (ebiosciences (ThermoFisher), Waltham, MA) at 10μg/mL for TCR/CD28 stimulation. For IFN R stimulation, 500U/mL of universal type 1 IFN (Pbl Assay Science, Township, NJ) was added. Cell lysis was performed as described [116] with slight modifications as follows. Cells were lysed using 500μl of Brij 97 buffer (25 mm Tris ph 8.0, 150 mm NaCl, 1% Brij-97, 0.5% n-octyl-β-d-glucopyranoside, 2 mm Na3VO4, EDTA-free mini-complete protease inhibitor tablets (Roche, Basal, Switzerland), and 5μg/mL DNase 1. For TCR/CD28 stimulation, AffiniPure F(ab ) 2 Ab fragments (for mouse T cells,rabbit anti-syrian hamster IgG; for human T cells, goat antimouse IgG, Jackson ImmunoResearch Laboratories), were added to each lysate (20μg/mL) to inhibit stimulatory Ab association with protein G Dynabeads (Life Technologies, Carlsbad, CA). Lysates were pre-cleared for 1 hour with protein G beads to remove nonspecific binding. Lysates were incubated + immunoprecipitation Abs for 1 hour or overnight on a rotator at 4 C (the lane label C indicates the sample was 41

59 stimulated but no immunoprecipitation Ab was added). The following Abs were used for immunoprecipitation: 2μg of rabbit anti-traf3 Ab, 5μl of mouse anti-csk mab, 5μl anti-ifn R1 mab, or 6μl of rabbit anti-ptpn22 Ab. An hour later or the next day protein G beads were added for 1h, rotating at 4 C. Using a magnet, beads were washed three times with 25 mm Tris ph 8.0, 150 mm NaCl, 0.5% SDS. To elute immunoprecipitated proteins, 2X sample buffer was added to the beads and subsequently boiled at 95 C for 5. Samples were separated by SDS-PAGE and analyzed via Western blotting. For TCR complex (CD3+CD28) immunoprecipitation, the above protocol was used with the following alterations. Rabbit anti-mouse IgG (3.5μL/IP, Jackson ImmunoResearch Laboratories) was added instead of the F(ab ) 2 Ab fragments to crosslink the CD3/CD28 antibodies. Whole cell lysates Primary mouse splenic T cells were isolated using a Pan T, CD4 + or CD4 + Treg - negative purification kit (StemCell Technologies). For isolation of CD4 + Treg - T cells, a negative purification was first performed to isolate CD4 + T cells. To obtain CD4 + Treg - T cells, Tregs were positively selected using a CD25-specific Ab and removed by magnetic beads according to manufacturer s instructions. 1x10 6 human immortalized T cells or 5x10 6 primary mouse T cells were used per time point. Cells were washed and resuspended in serum-free RPMI medium and placed on ice. Both TCR/CD28 and IFN R stimulation was performed as stated previously (10μg/mL anti-cd3 and anti- CD28 Abs, 500U/mL of universal type 1 IFN) in a 37 C water bath for the specified times. After stimulation, samples were centrifuged at 1280 x g for 2 minutes. The 42

60 medium was aspirated and 80μL of pre-warmed (95 C) 2x sample buffer was added. The samples were then sonicated using a Branson Sonifier 250 (VWR International, Radnor, PA) for 12 pulses at 90% duty cycle, output 1.5 and subsequently boiled at 95 C for 5. Western blot analysis was then performed on SDS-PAGE-separated samples. Cell fractionation T cells were washed and stimulated as above. After activation, cells were resuspended in a solution of 10mM HEPES [ph 7.4], 5mM MgCl2, 2mM Na3VO4, EDTA-free mini-complete protease inhibitor tablets, and placed on ice for 15. Cells were sheared by passaging 5 times through a l ml syringe fitted with a 27g needle. The lysate was centrifuged at 200 x g for 10 mins. The supernatant was recovered and subjected to ultra-centrifugation (Beckman Coulter Optima MAX-XP, Bea, CA) at 90,000 x g for 3h. The supernatant, representing the cytoplasmic fraction, was saved and the pellet resuspended in membrane solution, representing the plasma membrane fraction (20mM Tris [ph7.4], 150mM NaCl, 1% NP-40, 2mM Na3VO4, EDTA free mini-complete protease inhibitor tablets). Immunoprecipitation from the cytoplasmic and membrane fractions was performed as described above using 5μl anti-ptpn22 mab or 2μg rabbit anti-traf3 Ab. HEK 293T cell transfection HEK 293T cells (1 x 10 6 /well in a 6-well plate) were grown overnight in DMEM. Prior to transfection, DMEM was replaced with Optimem-1 medium for 30. Plasmids used were: pcdna-ptpn22-flag, pcdna-ptpn22r620w (a gift from Dr. Eric 43

61 Peterson, University of MN, Minneapolis, MN [158]), pslx-csk (a gift from Dr. David Schlaepfer, Scripps Research Institute, La Jolla, CA [159]), IFN R1 (a gift from Dr. Lawrence M. Pfeffer, University of TN, Memphis TN [160]), plk4-lat constructs, including full length WT and mutants (a gift from Dr. Jon C.D. Houtman, University of Iowa, Iowa City, IA), pcdna-traf3-ha, and pcdna-traf3δtrafc-ha [161]. For each transfection, 8μg of plasmid and 12.5μl of lipofectamine 2000 were suspended in Optimem-1 medium, as specified in the company s protocol, and added to the HEK 293T cells. 6-8h later medium was removed and fresh DMEM10 was added. Two days later cells were washed with PBS and lysed with 1 x RIPA buffer (150mM NaCl, 5mM EDTA [ph 8.0], 50mM TRIS [ph 8.0], 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 2mM Na3VO4, EDTA-free mini-complete protease inhibitor tablets, 5μg/mL DNase 1) for 30 on ice. Cell lysate was centrifuged for 10 at 21,000 x g. The supernatant was saved and an immunoprecipitation was performed as described above using 5μl of mouse anti-csk or 5μl rabbit anti-ifn R1 mabs. Cell numbers, cytokine and protein production with prolonged type 1 IFN stimulation From the HuT28.11 parent cells and subclones, 0.5x10 6 cells were isolated per time point and resuspended in fresh BCM10 (containing 500U/mL type 1 IFN) at a concentration of 0.25x10 6 cells/ml. For the 0 time point, cells were not exposed to type 1 IFN and medium was fresh BCM10. In a 12-well plate, 2mL of resuspended cells in BCM10 + type 1 IFN were added to each well, with individual wells representing distinct time points. Cells were stimulated in culture for 0.5, 1, 2 and 3 days. Every day at a 44

62 specified time, 500U/mL of type 1 IFN was added to each well or the cells/medium were harvested. After 1 day of culture, 2mL of fresh BCM10 + type 1 IFN (500U/mL) were added to each of the remaining wells. Upon harvest, cell numbers were determined prior to the production of a whole cell lysate as previously described. Using cell counts, calculations were performed to determine the volume of whole cell lysates required to equal 0.5x10 6 cells. STAT1 and actin protein quantification was performed via Western blot analysis. Medium was collected at the time of cellular harvest to assay for cytokine production via ELISA. Flow Cytometry Single cell suspensions of spleens were prepared by mashing between frosted ends of glass slides and filtering through a 100μM nylon mesh cells/well in 96-well plates were incubated with FACS buffer (1x PBS + 2% heat-inactivated FBS) containing 5μg/mL anti-cd16/32 (clone 2.4G2, Tonbo biosciences, San Diego, CA) and 2% rat serum (Jackson Immunoresearch) on ice for 15 mins to block non-specific antibody binding. Cells in FACS buffer were then stained with fluorescently labeled antibodies on ice in the dark for 30 mins. Cells were washed twice with FACS buffer, fixed for 10 mins at room temperature with BD FACS lysing solution (BD Biosciences, product # ), then resuspended in PBS. Intracellular staining was carried out using the ebiosciences transcription factor staining kit according to the manufacturer s protocol. The following fluorescently labeled antibodies were used: anti-cd4 (clone GK1.5, 1:100 dilution, ebioscience); anti-cd8α (clone , 1:250 dilution, ebioscience, BD Biosciences, and BioLegend); anti-cd44 (clone IM7, 1:500 dilution, ebioscience and Biolegend); anti- 45

63 CD90.2/Thy1.2 (clone , 1:500 dilution, BD Biosciences); anti-cd279/pd-1 (BioLegend, clone RMP1-30, 1:100 dilution); anti-foxp3 (clone FJK-16s, 1:250 dilution, ebioscience). Cells were analyzed on a BD LSR II flow cytometer and data were analyzed using FlowJo software (FlowJo LLC, Ashland, Oregon). Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number P30CA Total Internal Reflection Fluorescence (TIRF) microscopy Imaging was accomplished using the Leica AM TIRF MC imaging system (Meyer Instruments, Houston, TX) as in [116] with modifications below. 5x10 5 primary mouse splenic T cells from LMC or T-traf3 -/- mice were placed into glass chamber slides (LabTek II, Scotts Valley, CA) pre-coated with 10μg/mL anti-cd3 and anti-cd28 Abs. T cells were stimulated for the indicated times. Cells were blocked with SEA blocking buffer (ThermoFisher Scientific) for 1h and stained with 5μl rabbit anti-ptpn22 Ab (Proteintech, Chicago, IL) overnight at 4 C. Cells were washed and incubated at room temperature with Alexa 488-conjugated goat anti-rabbit IgG (Invitrogen) secondary Ab for 2h. Cells were washed and fresh PBS was added to each well. Images were obtained using a 100x oil submersion lens and Leica AF software. In individual experiments, all images were obtained using the same laser intensity and exposure parameters. Images were processed using PhotoShop (Adobe, San Jose, CA) and ImageJ (NIH) software. Analysis via ImageJ was performed by circling individual cells and maximum intensity was obtained using the colocalization-coloc2 plugin. Data from approximately 50 cells analyzed/time point were compiled and graphed using scattered dot plots with mean 95% 46

64 confidence interval (GraphPad Prism Software, San Diego, CA). Data were obtained from two independent experiments. Quantitative PCR CD4 + splenic T cells were isolated from T-traf3 -/- or LMC mice using a CD4 + negative selection isolation kit (StemCell Technologies). For all samples, mrna extraction was performed on unstimulated T cells using the RNeasy Mini Kit (Qiagen). Equal amounts of RNA were DNAse (Invitrogen) treated and cdna was made using the Superscript III First Strand kit (Invitrogen) as indicated in the manufacturer s protocol. cdna was then used in the qpcr reaction containing 1x Absolute Blue qpcr low ROX master mix (ThermoFisher Scientific) with 1μL of the specified Taqman probes: Csk (Mm _g1), Ppia (Mm _g1), Ifnar1 (Mm _m1), Cxcl9 (Mm _m1), Actin b (Mm _g1), and Gapdh (Mm _g1) (ThermoFisher Scientific). Analysis was performed using the Applied Biosystems 7900HT machine. The Pdcd1 qpcr analysis was performed using PerfeCta SYBR Green FastMix with ROX (Quanta Biosciences, Beverly, MA) with the following probes: Pdcd1 qstar qpcr primer pair (MP Origene, Rockville, MD) and Hprt RT2 primer assay (PPM03559F, Qiagen). The SYBR Green qpcr analysis was performed using the Mastercycler ep realplex (Eppendorf, Hamburg, Germany). For all samples, assembly of the reactions and cycling conditions were performed according to the manufacturer s protocols. Microsoft Excel software was used to calculate the ΔCT values (averaged gene value/averaged control values), ΔΔCT ((T-traf3-/- ΔCT value)- (LMC ΔCT value)) and fold change (2 -(ΔΔCT) ). The pooled data were then graphed using 47

65 Prism Software (GraphPad). Data presented herein were obtained at the Genomics Division of the Iowa Institute of Human Genetics which is supported, in part, by the University of Iowa Carver College of Medicine and the Holden Comprehensive Cancer Center (National Cancer Institute of the National Institutes of Health under Award Number P30CA086862). Statistical analyses Statistical significance was performed on the average of individual experimental groups using either two-way ANOVA (for multiple groups), T-test or Wilcoxon matchedpairs signed rank test (two groups). Statistical significance was calculated using Prism software with significance set at * = P<0.05, **=P<0.01, ****=P<

66 CHAPTER III: TRAF3 ENHANCES TCR SIGNALING BY REGULATING THE INHIBITORS CSK AND PTPN22 Introduction TRAF3 associates with the TCR complex following co-ligation of CD3 and CD28; ligation of either alone is not sufficient for effective TRAF3 recruitment [7]. T cell-specific TRAF3-deficient mice revealed that loss of TRAF3 in conventional T cells leads to reduced CD3/CD28-stimulated activation of the TCR signaling proteins LAT, ZAP70, ERK and PLC-γ1 [7]. Activation events upstream of ZAP70 phosphorylation were previously unexplored. Induction of the TCR signaling pathway involves equilibrium between phosphorylation and de-phosphorylation of Lck at activating (Y394) and inhibitory (Y505) sites [56]. Lck activity is negatively regulated by both Csk and PTPN22, with Csk being the most widely recognized Lck inhibitor [53]. Csk phosphorylates Y505 of Lck [53], which results in a conformational change in Lck that reduces the phosphorylation of the activating Y394 site [51]. The phosphatase PTPN22 inhibits Lck activation after TCR induction by dephosphorylating Y394 on Lck [88, 89]. PTPN22-mediated dephosphorylation of Lck prevents abnormally prolonged TCR activation [77]. Csk and PTPN22 associate in the cytoplasmic fraction of T cells, and this association dissipates shortly after TCR activation. After 15 mins of TCR stimulation, Csk and PTPN22 begin to re-associate [89]. Previous studies identified an inherited human coding variant (PTPN22-R620W), associated with an increased likelihood of autoimmunity and susceptibility to pathogens such as Mycobacterium tuberculosis [77]. The association between Csk and PTPN22 is defective in PTPN22-R620W-expressing cells [81, 89, 162]. 49

67 These observations led to the hypothesis that TCR-induced Csk and PTPN22 dissociation regulates T cell activation by regulating the release of PTPN22 to the TCR complex after stimulation to inhibit TCR-associated proteins such as Lck and the CD3 complex [77]. Recent studies demonstrate that TRAF3 associates with PTPN22 in PBMC [163] and B cells [126], and that the autoimmunity-associated R620W variant diminishes the association with TRAF3 [163]. We thus hypothesized that TRAF3 promotes Lck activation and downstream events of early TCR signaling by regulating two key Lck inhibitors, Csk and PTPN22. We addressed this hypothesis using primary splenic T cells from T-traf3 -/- mice and their LMC, as well as a complementary model of human T cell lines and TRAF3-deficient subclones that we produced. In TRAF3-deficient T cells, activation of Lck, as determined by Y394 phosphorylation, was consistently reduced. This suggested that the overall reduction in TCR signaling and TCR-dependent downstream functions in these T cells is primarily a consequence of this initial decrease in Lck activation. In WT or LMC T cells, TRAF3 association with Csk was dynamic, with the TRAF3-Csk complex re-locating from the membrane to the cytoplasm upon CD3/CD28 stimulation. In TRAF3-deficient T cells, decreased Lck activation levels coincided with an increase in Csk at the plasma membrane. TRAF3 also associated with PTPN22 in T cells, and regulated PTPN22 localization to the membrane after TCR stimulation. Interestingly, the absence of TRAF3 reduced, but did not eliminate, the Csk-PTPN22 association, suggesting that TRAF3 enhances the formation of a potential tri-partite complex. The present findings reveal a new mechanism by which TCR signaling is regulated by the multi-functional adapter protein TRAF3. 50

68 Results The impact of TRAF3 upon early TCR signaling events TRAF3-deficient T cells display a ~50% decrease in TCR-mediated activation of signaling proteins, evident as early as phosphorylation of the tyrosine kinase ZAP70 [7]. To address the role of TRAF3 in the earliest steps of TCR signaling, we utilized shrnas targeting TRAF3 (shtraf3), with Luciferase shrna (shluc) as a control. The shtraf3 cell line exhibited a ~60% reduction in TRAF3 protein compared to shluc cells (Figure 5a). To remove TRAF3 more completely from the human T cell line, CRISPR/Cas9 technology was used to produce a crtraf3 -/- cell line subclone of the HuT28 cells (Figure 5c). Neither the shrna-developed lines, shluc or shtraf3, nor the crtraf3 -/- cell line showed significantly altered total cellular levels of Lck, Csk or PTPN22 (Figure 5b and d). With the newly established TRAF3-deficient human T cell lines and our T-traf3 -/- mouse model, we examined one of the earliest measurable changes seen after engagement of the TCR complex, the activation of Src family kinases, particularly Lck. There was a basal decrease in plcky394 in the human shtraf3 cells (Figure 6a, b) and the crtraf3 -/- cells (Figure 6c, d) that continued through 15 minutes of stimulation via CD3/CD28. To confirm that the crtraf3 -/- results were not subclone-specific, we measured the reduction in plcky394 in a second crtraf3 -/- clone (Figure 7). The most profound difference in relative levels of plcky394 occurred within 5 minutes of stimulation, so we focused upon early times post-stimulation in primary mouse T cells, to avoid unnecessary animal use. In the mouse model, TRAF3 -/- T cells displayed a similar 51

69 decrease in activated plcky394 (Figure 6e, f). Furthermore, both total and plcky394 association with the TCR/CD28 complex decreased in TRAF3-deficient human T cells (Figure 8a). It is important to note that these data specifically evaluate Lck associated with the TCR/CD28 complex, rather than total membrane associated Lck, as a significant pool of membrane Lck will not be associated with the TCR/CD28 complex. Thus, these results do not suggest a reduction in levels of total membrane associated Lck. Additionally, in crtraf3 -/- cells, phosphorylation of the CD3ζ subunit of the TCR complex decreased in the absence of TRAF3 (Figure 8b). Consistent with this decrease in plcky394 and pcd3, loss of TRAF3 in the crtraf3 -/- T cell lines resulted in a reduction of tyrosine phosphorylation levels of TCR signaling proteins after TCR/CD28 stimulation (Figure 8c). The reduction in both levels of plcky394 and the recruitment of Lck to the TCR/CD28 complex seen here extends previous work showing reduced phosphorylation of downstream TCR signaling proteins in the absence of TRAF3 [7], and reveals that the defect in TCR signaling in TRAF3-deficient T cells occurs very early in the TCR signaling cascade. TRAF3 association with Csk in T cells The Lck inhibitor Csk contains both SH2 and SH3 domains. As TRAF3 contains several potential SH2 and SH3 binding sites, we predicted that TRAF3 associates with Csk in T cells. This prediction was confirmed by immunoprecipitating TRAF3 from mouse splenic T cells, following CD3/CD28 stimulation. The TRAF3-Csk interaction remained robust for minutes post-stimulation, returning to near unstimulated levels 52

70 after 15 minutes (Figure 9a). TRAF3 also associated with Csk in HuT28.11 human T cells (Figure 9b). Consistent with the nature of transformed cells, some association was seen even in unstimulated HuT28.11 cells. However, as in the primary mouse T cells, this association increased upon CD3/CD28 activation. Reciprocally, immunoprecipitation of Csk in HuT28.11 T cells confirmed the inducible association of Csk with TRAF3 (Figure 9c). To further investigate the TRAF3:Csk association, extended CD3/CD28 activation was performed in human T cells, revealing a decline in association by 30 minutes (Figure 9d). Due to the disassociation of Csk from the membrane after TCR stimulation, and the different cellular compartments in which Csk can reside [53], we performed cell fractionation to determine where the association between TRAF3 and Csk occurred. Results indicated a dynamic association between TRAF3 and Csk in distinct cell sub-compartments. Upon CD3/CD28 stimulation, levels of associated Csk:TRAF3 decreased in the insoluble fraction containing the plasma membrane fraction, while increasing in the cytoplasmic fraction (Figure 5e). Association with Csk prior to engagement of the TCR complex involves the membrane pool of TRAF3. Thus, the Csk- TRAF3 association seen in resting T cells likely involves the membrane pool, while the increased association following stimulation can involve both the membrane pool of TRAF3, as well as newly-recruited TRAF3 from the cytoplasm pool. To determine if the interaction of Csk and TRAF3 requires the TRAFC domain, which typically participates in protein-protein interactions involving TRAF3 [3], HEK 293T cells were transfected with plasmids containing full length Csk and WT TRAF3, or TRAF3 lacking this domain (ΔTRAFC) (Figure 9f). Co-immunoprecipitation confirmed 53

71 the TRAF3:Csk association, and indicated that the TRAFC domain of TRAF3 was required for this association (Figure 9f). Together, these results are consistent with the hypothesis that T cell TRAF3 associates with Csk at the plasma membrane and facilitates translocation of Csk to the cytoplasm upon TCR/CD28 stimulation. Csk regulation by TRAF3 Results above suggest that TRAF3 recruited to the TCR complex affects relative Csk levels in the plasma membrane of T cells. Following CD3/CD28 stimulation, shtraf3 T cells displayed an increase in Csk in the membrane fraction, compared to control T cells (Figure 10a). TRAF3 -/- mouse primary T cells also had an increase in membrane-associated Csk at the peak association time at 7.5 minutes, whereas membrane Csk levels in LMC T cells showed a decrease at this time (Figure 10b). The increase in Csk in the membrane of T-traf3 -/- T cells is not due to significantly increased levels of total Csk protein (Figure 11a, b) or mrna levels (Figure 11c). Phosphorylation of Csk at S364 inhibits its catalytic activity [75, 164], so we next examined whether T cell TRAF3 influences this mechanism of Csk regulation. In both shtraf3 and TRAF3 -/- T cells, pcsks364 levels were decreased compared to control T cells (Figure 10c, d). These findings suggest that more Csk is in the activated state in TRAF3-deficient T cells. The increase in Csk at the membrane and the reduction in levels of pcsks364 predicted a correlative increase in plcky505 phosphorylation in TRAF3-deficient T cells. However, there was no significant increase in the level of plcky505 in either shtraf3 or TRAF3 -/- primary T cells. The modest apparent decrease in plcky505 is not 54

72 statistically significant (Figure 10e, f). Several possibilities could account for these results. Technical limitations of Western blotting may fail to reveal transient changes. Alternatively, the reduction in TCR signaling seen in the absence of TRAF3 is not solely dependent on Csk. This increased our interest in further investigating the interactions between TRAF3 and Csk s binding partner, the phosphatase PTPN22. TRAF3 regulation of PTPN22-Csk association Csk associates with PTPN22 in the cytoplasm of resting human T cells. Upon TCR stimulation, PTPN22 is released from Csk to translocate to the membrane, where it dephosphorylates LckY394 [89]. Interestingly, the variant PTPN22-R620W has markedly decreased association with Csk [81], and the same variant displays decreased TRAF3- PTPN22 association [163]. These reports, together with the findings above demonstrating T cell Csk-TRAF3 association, led us to ask whether TRAF3 competes with Csk for association with PTPN22 at amino acid 620. We also wished to know whether all three proteins could form a complex together, and whether each individual protein is required for this complex to form. HEK 293T cells were transfected with constructs encoding full length or mutant TRAF3 and PTPN22, together with full length Csk (Figure 12a). In cells expressing WT TRAF3, PTPN22 and Csk, both TRAF3 and PTPN22 associated with Csk (Figure 12b). The association between Csk and PTPN22 was reduced in the presence of the ΔTRAFC mutant, which does not associate with Csk (Figure 12b). There was also a modest reduction in the total level of PTPN22 in the ΔTRAFC mutant- transfected cells. These data indicate that optimal interactions between Csk and PTPN22 requires WT TRAF3 (Figure 12b). Association between Csk and TRAF3 was also reduced in cells co- 55

73 transfected with PTPN22-R620W, compared to WT PTPN22 (Figure 12c). Furthermore, utilizing the crtraf3 -/- T cell line, we found association between Csk and PTPN22 decreased in the absence of TRAF3 in unstimulated cells (Figure 12d and 13). However, upon stimulation of T cells with CD3/CD28-specific Abs, Csk and PTPN22 association displayed no difference in association when comparing parent to crtraf3 -/- cells (Figure 13). Together these data indicate that Csk, TRAF3 and PTPN22 can potentially form a complex together, and both TRAF3 and PTPN22 are required for maximum association of the reciprocal protein with Csk. That all three proteins are not absolutely required for the association supports the model that there are both TRAF3-dependent effects upon Csk or PTPN22, but there are also TRAF3-independent roles for Csk and PTPN22 complexed together. TRAF3 regulation of PTPN22 localization Because the reduction in plcky394 in TRAF3-deficient T cells was not demonstrably correlated with an increase in plcky505 (Figure 6 and 10e, f), and Figure 12 revealed that PTPN22 can associate with both TRAF3 and Csk, we hypothesized that TRAF3 regulates PTPN22-induced dephosphorylation of LckY394 [88, 89]. First, we queried whether the relative amount of PTPN22 at the T cell plasma membrane is regulated by TRAF3. In CD3/CD28-stimulated TRAF3-deficient human T cells, more PTPN22 was observed in the insoluble membrane fraction of cell lysates, compared to control T cells (Figure 14a). 56

74 To further address the potential role of TRAF3 in PTPN22 trafficking within intact T cells, we examined PTPN22 localization in mouse splenic T cells stimulated with plate bound αcd3/αcd28 Abs, using Total Internal Reflection Fluorescence (TIRF) microscopy (Figure 14b). TIRF microscopy permits visualization of proteins within nm of a coverslip upon which cells are stimulated. This allowed us to examine PTPN22 localized to the plasma membrane, without the distraction of background cytoplasmic fluorescence. We observed an increase in PTPN22 clustering at the membrane in TRAF3 -/- compared to LMC T cells (Figure 14c). Reciprocally, the loss of PTPN22 in T cells did not alter plasma membrane levels of TRAF3 detected by TIRF microscopy (Figure 15a, b). Taken together, these data indicate that TRAF3 restrains the amount of PTPN22 associated with the membrane after CD3/CD28 stimulation. Results thus indicate that TRAF3 regulates relative levels of plcky394 and TCR activation by controlling the amount of PTPN22 migrating to the membrane. Conclusions Upon the discovery of TRAF3 association with the TCR/CD28 complex, and a positive role for TRAF3 as a regulator of TCR signaling, it has been a high priority to understand the mechanism by which TRAF3 exerts this regulation. Loss of TRAF3 in our T-traf3 -/- mouse model results in reduced activation of several TCR signaling proteins, including ZAP70, LAT, ERK and PLCγ1 [7]. We demonstrate here that TRAF3 enhanced the earliest detectable TCR signaling event, the activation of Src kinases. In the absence of TRAF3, less Lck was associated with the TCR/CD28 complex. This suggests that upon CD3/CD28 stimulation, less membrane-associated Lck in TRAF3-deficient T 57

75 cells is recruited to the TCR/CD28 complex, preventing Lck from phosphorylating the CD3 subunits of the TCR. This does not suggest that the reduction in Lck association with the TCR/CD28 complex is a result of decreased total membrane levels of Lck; we found no data to support this. TRAF3 enhanced TCR signaling by regulating the localization of two inhibitors of TCR and Lck signaling, the kinase Csk and the phosphatase PTPN22. The first mechanism by which TRAF3 regulates TCR/CD28 signaling occurs upon CD3/CD28 activation, when TRAF3 sequesters Csk away from the plasma membrane (Figure 16a). Removal of Csk from the membrane allows more Lck to become activated and subsequently phosphorylate the TCR complex. Because the current study indicates that TRAF3 regulates Csk movement, an area of interest for further research is to determine the mechanism that leads to TRAF3:Csk association. There are two pools of TRAF3 that have the potential to associate with and subsequently regulate Csk: membrane-localized TRAF3 and TRAF3 recruited to the TCR complex from the cytoplasm in activated T cells. It is of particular interest to determine if TRAF3 in different cellular compartments associates with and regulates Csk differently. Several minutes after the activation of the TCR/CD28 complex, TRAF3 also regulates the translocation of PTPN22 to the membrane (Figure 16b). Following this translocation, signaling proteins such as Lck and the CD3 subunits of the TCR complex are dephosphorylated and inhibited. This allows TRAF3 to regulate the duration and intensity of TCR/CD28 signaling. In the absence of TRAF3, there are increased levels of both Csk and PTPN22 at the membrane. In TRAF3-deficient T cells, Csk is not removed from the plasma membrane upon stimulation, while PTPN22 protein levels not only 58

76 significantly increases at the membrane, but the increase is also seen earlier after stimulation (Figure 16c). 59

77 Figure 5. Protein levels of TRAF3. TCR signaling protein expression levels were determined by Western blot analysis in shluc and shtraf3 (a and b), HuT28.11 and crtraf3 -/- (Clone 45) (c and d) whole cell lysates. Western blots were cropped to focus upon specific proteins indicated. Quantification was performed by normalizing relative amounts of indicated proteins to actin and subsequently calculating the fold change of normalized shtraf3 or crtraf3 -/- values to the normalized control shluc or HuT28.11 values, respectively. Data from at least 3 independent experiments were pooled and the mean values ± SEM are shown. Statistical analysis was performed using the Wilcoxon matched-pairs signed rank test, which indicated no statistical differences between TRAF3-deficient vs. sufficient T cells in b and d. 60

78 Figure 6. Requirement for TRAF3 in Src kinase activation by the TCR complex. T cells were stimulated via CD3/CD28 for indicated times. Whole cell lysates were prepared from HuT28.11 subclones shluc or shtraf3 (a), HuT28.11 or the HuT28.11 subclone crtraf3 -/- (c), or primary mouse splenic T cells from LMC or T- traf3 -/- mice (e), or as described in Methods. Western blot analysis was performed to detect the indicated proteins. Expression levels of pfyn Y417 /plck Y394, detected by antipsrc Y416 Ab (b, d and f) were calculated as: (plck/actin)/(lck/actin) = plck Relative expression level (REL). Fold change was then determined based upon the obtained expression levels, as follows. (plck REL)/(pLck REL of control time point 0) = plck Fold change. Data from at least 3 independent experiments were pooled and the mean values ± SEM are shown. A 2-way ANAOVA was performed and was not corrected for multiple comparisons using the Fisher s LSD test to determine statistical significance; * = P<0.05 comparing the TRAF3 deficient cells to the respective control time point. T-traf3 -/- time point 7.5 is trending towards significance with p= Western blot and blots were cropped to focus upon the specific proteins indicated. Data are representative of 3 individual experiments. 61

79 Figure 7. TCR signaling in crtraf3 -/- T cells. TCR signaling proteins were determined by Western blot analysis in HuT28.11 and crtraf3 -/- (Clones 28 and 45) whole cell lysates. Western blots were cropped to focus upon specific proteins indicated. Quantification was performed by normalizing relative amounts of indicated proteins to actin and subsequently calculating the fold change of normalized crtraf3 -/- values to the normalized control HuT Data from at least 3 independent experiments were pooled and the mean values + SEM are shown. Statistical analysis was performed using the Wilcoxon matched-pairs signed rank test, * = P<0.05 comparing the crtraf3 -/- cells to the respective control time point. 62

80 Figure 8. Loss of TRAF3 decreases TCR signaling and CD3ζ activation. T cells were stimulated via CD3/CD28 for indicated times. (a) Immunoprecipitation of the TCR (CD3)/CD28 complex was performed on shluc and shtraf3 whole cell lysates. The relative amount of plck Y394 and total Lck association with the TCR/CD28 complex was analyzed by Western blot. (b) Whole cell lysates were prepared from HuT28.11 and subclone crtraf3 -/- (clone 45) cells. Immunoprecipitation of the CD3/CD28 complex was performed and the relative amounts of activated (p CD3ζ) and total CD3ζ were analyzed, respectively, by Western blot. (c) Western blot analysis was performed on whole cell lysates to detect tyrosine phosphorylated proteins (antipy) and actin. Data are representative of 2-3 individual experiments. Blots were cropped to focus upon the specific proteins indicated. 63

81 Figure 9. TRAF3 association with Csk in T cells. (a-d) T cells were stimulated via CD3/CD28 for the indicated times. Fab fragments targeting the stimulatory Abs were added to prevent their association with the magnetic protein G beads, as in Methods. Samples were also precleared to remove any unbound stimulatory Ab. TRAF3 immunoprecipitation was then performed as described in Methods, from whole cell lysates of WT primary mouse splenic T cells (a) or Hut28.11 T cells (b and d). (c) Immunoprecipitation of Csk from Hut28.11 whole cell lysates. (e) Cellular fractionation was performed as in Methods on HuT T cells. TRAF3 was immunoprecipitated from soluble and insoluble cell lysate fractions, isolated as in Methods. (a-e) Analysis was performed by Western blot to detect the indicated proteins. C=Control samples, in which cells were stimulated for 5 and no immunoprecipitation Ab was added, to detect any residual stimulatory Ab association with the protein G beads which would result in non-specific immunoprecipitation. (f) HEK 293T cells transfected with TRAF3 and Csk constructs (depicted in Fig. 8a) were lysed, and Csk was immunoprecipitated. Total input prior to immunoprecipitation is shown at right. Data shown are representative of 3-6 independent experiments. 64

82 Figure 10. Role of TRAF3 in the regulation and localization of Csk. T cells were stimulated with αcd3/αcd28 Abs for the indicated times and a cellular fractionation was performed, as described in Methods, on lysates from shluc and shtraf3 cells (a) or isolated primary mouse splenic T cells from LMC and T-traf3 -/- mice (b). Western blot analysis of the insoluble (membrane) fractions was performed to detect the designated proteins, with actin or LAT serving as loading controls (Top panels). Expression levels of Csk were first normalized to LAT/actin and the subsequent fold change was calculated from the control 0 time point (bottom panels). Data from T- traf3 -/- cells stimulated for 7.5 minutes is trending towards significance with p= Western blot analysis was performed on whole cell lysates from human T cell lines (c) or primary mouse splenic T cells (d) and blotted for pcsk S364 (Top panels). Expression levels of pcsk S364 were first normalized to actin and subsequent fold change was calculated from the control 0 time point (bottom panels). Western blot analysis was performed on whole cell lysates to detect the expression levels of plck Y505 (e and f, top panels). Western blots were cropped to focus upon specific proteins indicated. Expression levels of plck Y505 were first normalized to actin and subsequent fold change was calculated from the control 0 time point. Differences between TRAF3- deficient and TRAF3-sufficient cells did not reach statistical significance. Data from at least 3 independent experiments were pooled and the mean values ± SEM are shown. A 2-way ANOVA was performed and was not corrected for multiple comparisons using the Fisher s LSD test to determine statistical significance in all 6 panels; * = P<0.05, **=P<0.01 indicates statistical significance of values compared to those at the respective control time point. NS=Nonspecific band. 65

83 66

84 Figure 11. Csk expression in T-traf3 -/- T cells. (a) Protein levels of Csk were determined by Western blot analysis in unstimulated LMC and T-traf3 -/- T cell whole cell lysates. (b) Western blots were cropped to focus upon specific proteins indicated. Quantification was performed by normalizing relative amounts of indicated proteins to actin and subsequently calculating the fold change of normalized T-traf3 -/- values to the normalized control LMC values. (c) mrna levels of Csk were determined by qpcr of samples from unstimulated LMC and T-traf3 -/- T cells. Data from at least 3 independent experiments were pooled and the mean values + SEM are shown. Statistical analysis was performed using the Wilcoxon matched-pairs signed rank test, which indicated no statistical differences between TRAF3 deficient vs. sufficient T cells in b and c. 67

85 Figure 12. TRAF3, PTPN22 and Csk association. (a) TRAF3, PTPN22 and Csk constructs used in HEK 293T cell transfections. (b and c) Immunoprecipitation of Csk from transfected HEK 293T cell lysates is shown at left, with total cell lysate input prior to immunoprecipitation shown at right. Western blot analysis was performed for the specified proteins. Quantification of bands, shown in right hand panels, was performed by dividing the relative expression of PTPN22 by Csk (c) or TRAF3 by Csk. The obtained values for the cells transfected with the mutant plasmids were then normalized to the values obtained for cells transfected with the control wild type plasmids. (d) An immunoprecipitation for Csk was performed as described previously, from unstimulated HuT28.11 or crtraf3 -/- (clone 45) T cell whole cell lysates. Western blots were cropped to focus upon specific proteins indicated. C=Control samples, cells were unstimulated and no immunoprecipitation Ab was added, to detect any nonspecific binding to the protein G beads. Error bars indicate SEM of three experiments. A Wilcoxon matched-pairs signed rank test was performed to determine statistical significance in b and c; * = P<0.05, **=P<0.01. Data are representative of 3-8 independent experiments. 68

86 Figure 13. Csk and PTPN22 association in TRAF3-deficient activated T cells. An immunoprecipitation for Csk was performed as described previously, from CD3/CD28-stimulated HuT28.11 or crtraf3 -/- (clones 28 and 45) T cell whole cell lysates. Western blots were cropped to focus upon specific proteins indicated. C=Control samples, cells were unstimulated and no immunoprecipitation Ab was added, to detect any nonspecific binding to the protein G beads. Error bars indicate SEM of three experiments. 69

87 Figure 14. TRAF3 regulation of PTPN22 localization and association with membrane Csk. T cells were stimulated for indicated times via CD3/CD28. (a) Cellular fractionation of whole-cell lysates was performed on shluc and shtraf3 T cells to obtain the insoluble (membrane) fractions, which were assessed for relative PTPN22 protein levels by Western blot analysis, using LAT and actin expression as controls. Western blots were cropped to focus upon the designated proteins. (b and c) Primary mouse splenic T cells isolated from LMC and T-traf3 -/- mice were permeabilized and stained with an anti-ptpn22 Ab followed by a secondary Ab for Alexa-488, then analyzed by total Internal Reflection Fluorescence (TIRF) microscopy. (b) Representative TIRF images and (c) quantification of each cell for maximum intensity of PTPN22 fluorescence was determined as in Methods (n=50 cells per time point). A 1-way ANOVA was performed to establish statistical significance, **** = P< Data are representative of 3 (a) or 2 (b and c) independent experiments. 70

88 Figure 15. Membrane TRAF3 clustering is unaffected by the loss of PTPN22. Primary mouse splenic T cells isolated from LMC and Ptpn22 -/- mice were permeabilized and stained with an anti-traf3 Ab followed by a secondary Ab conjugated to Alexa-488, then analyzed by total Internal Reflection Fluorescence (TIRF) microscopy. (a) Representative TIRF images and (b) quantification of each cell for maximum intensity of PTPN22 fluorescence was determined as in Methods (n=50 cells per time point). A 1-way ANOVA was performed indicating no statistical significance. Data were obtained from 1 experiment. 71

89 Figure 16. TRAF3 regulation of Csk and PTPN22 in TCR/CD28 signaling. TRAF3 regulates TCR signaling via at least 2 distinct mechanisms. During early TCR activation, TRAF3 associates with Csk to translocate Csk away from the plasma membrane to the cytoplasm (a). This allows the recruitment of activated Lck to the TCR complex where Lck phosphorylates the CD3 subunits. A second mechanism of TRAF3 regulation of TCR/CD28 signaling is by controlling the localization of cytosolic PTPN22 to the plasma membrane. To inhibit prolonged TCR/CD28 activation, TRAF3 regulates the amount and time at which PTPN22 is released to the membrane from the cytoplasm (b). Once at the membrane, PTPN22 dephosphorylates Lck and the CD3 subunits of the TCR complex to inhibit TCR/CD28 signaling. In the absence of TRAF3 both mechanisms of TCR/CD28 inhibition are enhanced. The loss of TRAF3 results in more Csk at the membrane and more PTPN22 recruited to the membrane earlier in TCR/CD28 activation (c). 72

90 73

91 CHAPTER IV: TRAF3 ASSOCIATION WITH THE LAT COMPLEX Introduction Signal 2 of T cell activation, delivered by co-stimulatory receptors, is required for the enhancement of TCR signaling and formation of SMAC protein clustering. Engagement of the co-stimulatory receptor CD28 enhances TCR signaling by converging on the TCR signaling pathway at the step of PLC- 1 activation [91], while also inducing TCR-independent pathways as described previously [30]. CD28 signaling enhances cellular proliferation and survival, cytokine production and increased metabolism [30]. In a previous study in our lab, TRAF3 was found to associate with the TCR/CD28 complex upon treatment of T cells with both anti-cd3 and anti-cd28 stimulatory Abs, but not with stimulation of either receptor alone [7]. Thus, TRAF3 recruitment to the TCR/CD28 complex requires CD28 stimulation, via several possible mechanisms: 1) TRAF3 associates with CD28 directly after TCR/CD28 stimulation, and/or 2) TRAF3 interacts with LAT complex-associated proteins that cluster with the TCR/CD28 complex after stimulation. The objective of this chapter is to explore the potential association between TRAF3 and the LAT complex. T cells stimulated through the TCR/CD28 receptors by the antigenic peptide- MHC and CD80/CD86 on the APC induce SMAC formation on the T cell. SMAC formation includes the establishment of the c-smac that surrounds the TCR/CD28 receptors. The c-smac consists of TCR signaling proteins, including the LAT complex [95, 99, 100]. LAT complex formation occurs through the recruitment of several associated proteins, including Grb2, Gads, Grap, PLC- 1 and PI3K [ ]. In this chapter, I will focus upon LAT and the Grb2 family proteins, Grb2 and Gads. Our 74

92 working hypothesis is that TRAF3 associates with either LAT or LAT complexassociating proteins in T cells (Figure 17). This chapter focuses on three essential proteins for LAT complex formation and activation: LAT, Grb2 and Gads. Results TRAF3 association with the LAT complex Understanding how TRAF3 associates with the TCR/CD28 complex after anti- CD3/anti-CD28 Ab stimulation is a current knowledge gap. In the absence of CD28 stimulation there is very little detectable TRAF3 association with the TCR/CD28 complex [7], suggesting that signaling through CD28 is essential for the recruitment of TRAF3. There are two working hypotheses for the requirement of CD28 signaling, described previously. The focus of this chapter is to explore an association between TRAF3 and the TCR signaling proteins that make up the LAT complex. In the resting T cell, LAT is distributed throughout the cell membrane and is not observed in direct association with the TCR or CD28 receptors [165]. Our hypothesis predicts that TRAF3 associates with LAT in the resting T cell, as seen in Figure 17. The distribution of LAT throughout the resting T cell membrane prevents TRAF3 association with the TCR/CD28 complex. Following anti-cd3/anti-cd28 Ab stimulation, formation of the T cell SMAC sequesters TCR signaling proteins in close proximity to the TCR/CD28 complex. TCR/CD28-induced formation of the LAT signalosome surrounds the TCR/CD28 complex (Figure 17). Removal of the TCR/CD28 complex via immunoprecipitation would also remove the LAT signalosome. The Grb2 family members, Grb2 and Gads, are required for formation of the LAT signalosome and/or 75

93 downstream signaling. Both Grb2 and Gads proteins contain SH2 and SH3 domains which harbor a potential TRAF binding site. Our hypothesis is that TRAF3 associates with LAT and/or the Grb2 family members, which results in TRAF3 interaction with the TCR/CD28 complex after stimulation. TRAF3 association with wild-type LAT We explored the association between LAT and TRAF3, using LAT mutants. The LAT mutant constructs include deletion of amino acids (LAT 23 ) or rearrangement of the LAT protein sequence by placing amino acids to the end of LAT (LAT 132 ) (Figure 18a). Following HEK293T cell transfections with the indicated constructs, TRAF3 associated with full length wild-type LAT, with diminished association seen between TRAF3 and both LAT mutant constructs (Figure 18b, c). To explore whether TRAF3 requires a specific phosphorylation site on LAT, the phosphorylation sites Y171, Y191 and Y226 and the surrounding 9-10 amino acids was mutated so that all three phosphorylation sites expressed the same sequence (LAT 171x3, LAT 191x3 and LAT 226x3 ) (Figure 18d). Additionally, point mutations in all 4 tyrosine phosphorylation sites (amino acids 132, 171, 191, and 226) were changed from a tyrosine to a phenylalanine (LAT Y F ) (Figure 18d). Again, TRAF3 associated with full length, wild-type LAT, but this association decreased in all the LAT phosphorylation mutants (Figure 18e, f). These data indicate that TRAF3 associated with LAT, and this association required the wild-type LAT protein sequence, including the 4 known tyrosine phosphorylation sites. 76

94 Association of Grb2 family members with TRAF3 As shown in the previous chapter, TRAF3 associates with Csk, a SH2 and SH3 domain containing protein, which contains similar domains to the Grb2 family of proteins [150]. This suggests a possible interaction between TRAF3 and Grb2 family members such as Grb2 and Gads. To address this possibility, full length and TRAFC-deleted TRAF3 (ΔTRAFC) constructs were co-transfected with wild-type Gads or Grb2 into HEK293T cells. Immunoprecipitation of TRAF3 identified an association between full length, wild-type TRAF3 and Gads but not Grb2 (Figure 19a, b), despite the similarity in protein structure between Grb2 and Gads. Furthermore, the association between TRAF3 and Gads was dependent on the TRAFC domain of TRAF3 (Figure 19a, b). Membrane localization of TRAF3 and Grb2 family proteins Upon observing a potential association between TRAF3 and Gads but not Grb2, we wanted to determine if membrane clustering of TRAF3 or the Grb2 family proteins were altered in the absence of TRAF3, Grb2 or Gads. Primary mouse T cells were stimulated with plate-bound anti-cd3 and anti-cd28 Abs, then permeabilized and stained with anti-gads Ab to detect Gads membrane clustering via TIRF microscopy. Primary T cells deficient in TRAF3 did not alter Gads clustering after TCR/CD28 stimulation (Figure 20a). To address the level of membrane clustering of TRAF3 in the absence of Gads, we utilized previously established T cell lines transduced with shrnas targeting Gads (shgads) and Luciferase (shluc) as a control [117]. A caveat of these cells is that they do not endogenously express the CD28 receptor. In contrast to the TRAF3-deficient primary T cells, shgads T cells stimulated with anti-cd3 Ab increased 77

95 TRAF3 membrane clustering compared to shluc, despite the absence of CD28 stimulation (Figure 20b). TRAF3 and Gads associates when exogenously overexpressed in HEK293T cells, but clustering of TRAF3 after TCR stimulation of HuT transformed human T cells is independent from the association with Gads. We next explored the membrane clustering of Grb2 and TRAF3 in TRAF3- deficient (T-traf3 -/- ) or Grb2-deficient (shgrb2) T cells. Cells of the Grb2-deficient HuT78 T cell line were stimulated with plate bound anti-cd3 Ab, fixed and subsequently stained with an anti-traf3 Ab for visualization by TIRF microscopy. Loss of Grb2 decreased TRAF3 clustering after TCR stimulation (Figure 21a). A similar result is seen in TRAF3-deficient primary T cells stained for Grb2, which displayed reduced Grb2 clustering after anti-cd3/anti-cd28 Ab stimulation (Figure 21b). The reduced clustering of TRAF3 and Grb2 in T cells deficient in the partner protein can be explained by the decreased TCR signaling displayed in both altered T cell lines [7, 116]. Using a TRAF3- deficient HuT28.11 cell line (shtraf3), that expresses the CD28 receptor, T cells were stimulated with anti-cd3/anti-cd28 Abs and a cellular fractionation was performed for the isolation of the plasma membrane. Total protein expression levels for Grb2 in the membrane fraction increased in the absence of TRAF3 (Figure 21c). We hypothesize that the increase in Grb2 at the membrane may occur as a result of the T cell attempting to compensate for the decrease in TCR signaling caused by TRAF3 deficiency. Alternatively, TRAF3 and Grb2 may compete for association with LAT, and the loss of either TRAF3 or Grb2 allows more of the reciprocal protein to associate with LAT. Despite the increase in membrane levels of Grb2, LAT complex formation will not occur in TRAF3 deficient cells due to decreased upstream TCR activation (Chapter 3). 78

96 TRAF3 and Grb2 association with the TCR/CD28 complex The decrease in clustering of TRAF3 and Grb2 indicates reduced association of either protein with the TCR/CD28 complex in the absence of the reciprocal protein. To determine if Grb2 association with the TCR/CD28 complex is altered in the absence of TRAF3, TRAF3-deficient T cells were stimulated with anti-cd3/anti-cd28 Abs. Subsequently, immunoprecipitation of the TCR/CD28 complex was performed, and Grb2 association was detected in Western blots of proteins subjected to SDS-PAGE. Loss of TRAF3 increased early Grb2 association with the TCR/CD28 complex (Figure 22a). Similarly, TRAF3 associated with the TCR increased after anti-cd3 Ab stimulation in the Grb2-deficient T cells (Figure 22b). Together, these data suggest that TRAF3 and Grb2 compete for association with the TCR or TCR/CD28 complex. Conclusions Upon anti-cd3/anti-cd28 Ab stimulation, TRAF3 associates with the TCR/CD28 complex, though association does not occur with stimulation of either receptor alone [7]. TRAF3-mediated enhancement of TCR signaling requires CD28 stimulation via several possible mechanisms: 1) TRAF3 associates with CD28 directly after TCR/CD28 stimulation, and/or 2) TRAF3 interacts with the LAT complex, and TCR/CD28 stimulation allows LAT complex association with TCR/CD28. The objective of this chapter was to explore the second hypothesis of a potential association between TRAF3 and the LAT complex. 79

97 Collectively, the data from this chapter identify an association between TRAF3 and the LAT complex in T cells. Deletion of the membrane-associated region of LAT, rearrangement of its protein sequence, or mutation of LAT phosphorylation sites decreased TRAF3 association. TRAF3 also associates with Csk, a SH2 and SH3 domain containing protein [150], indicating a possible association between TRAF3 and the SH2 or SH3 domains of the Grb2 family of proteins. Despite the structural similarities between Grb2 and Gads, TRAF3 associated with Gads, but not Grb2 in HEK293T cells, following exogenous expression. A future area of research in our lab is to further examine this association in T cells. While we could not demonstrate a direct TRAF3-Grb2 association, loss of TRAF3 increased Grb2 expression levels at the membrane prior to and following TCR/CD28 stimulation. Despite this increase in Grb2 at the membrane, loss of TRAF3 inhibited TCR/CD28-induced Grb2 clustering in T cells. Additionally, TCR/CD28 stimulation of TRAF3-deficient T cells increased early Grb2 association with the TCR/CD28 complex. Although there were increased membrane levels of Grb2 and TCR/CD28-associated Grb2, Grb2 clustering in the absence of TRAF3 was not sufficient for detection by TIRF microscopy. This is likely a result of decreased early TCR signaling in the absence of TRAF3. Proximal TCR signaling is required for LAT activation and subsequent formation of the LAT complex. In the absence of Grb2, TRAF3 association with the TCR increased, with a decrease in TRAF3 membrane clustering after anti-cd3 Ab stimulation. These data suggest TRAF3 and Grb2 possibly compete for association with the TCR/CD28 complex. 80

98 Unlike the apparent absence of TRAF3-Grb2 association, HEK293T cell transfection experiments identified an association between TRAF3 and Gads that required the TRAFC domain of TRAF3. Clustering of membrane-associated Gads was unaltered in the absence of TRAF3. Conversely, Gads-deficient T cells displayed increased TRAF3 clustering after TCR stimulation, in contrast to what was observed in the Grb2-deficient T cells. The difference in TRAF3 membrane clustering between Grb2 and Gads-deficient T cells is likely a result of their individual roles in the LAT complex. First, Gads-deficient T cells do not alter LAT induced cytoskeletal rearrangement upon TCR stimulation [116], suggesting LAT complex formation still occurs in the absence of Gads, although LAT-driven downstream effects which require PLC- 1, such as T cell adhesion capabilities, are decreased in these cells [117]. Together these data suggest that the increase in TRAF3 clustering in the Gads-deficient T cells is due to increased LAT complex formation. Secondly, Grb2-deficient T cells lack LAT complex formation upon TCR activation [116], which would inhibit TRAF3 clustering. This chapter identified a new role for TRAF3, in association with LAT and Gads. Although there was not a demonstrable association between TRAF3 and Grb2 when overexpressed in an epithelial cell line, competition between TRAF3 and Grb2 for LAT association may occur. Examination of these associations in T lymphocytes is needed. More research is also needed to dissect the exact mechanisms for TRAF3 regulation of the LAT complex, and how TRAF3 impacts LAT-mediated downstream signaling pathways. Additionally, further clarification for the requirement of CD28 is needed to determine if TRAF3 association with the TCR/CD28 complex is a result of CD28- enhanced SMAC formation. 81

99 Figure 17. Predicted model of TRAF3 in the LAT signalosome. (a) In a resting T cell, TRAF3 associates with LAT. In the absence of TCR/CD28 stimulation, LAT can freely distribute throughout the plasma membrane and is unassociated with the TCR or CD28 receptors. (b) Following CD3 and CD28 stimulation, development of the T cell SMAC occurs, with the TCR/CD28 receptors clustering in the center. Surrounding the SMAC TCR/CD28 core are the TCR signaling proteins, including LAT, which is depicted here. Phosphorylation of LAT recruits the Gab family members Grb2 and Gads. Grb2 associates with SOS1 to bring SOS1 to the LAT complex. Ultimately this leads to the creation of the LAT signalosome. Gads association with LAT also aids in the production of the LAT signalosome by recruiting SLP-76 and PLC 1. Together, Grb2 and Gads are required for downstream transcriptional activation, cytokine production and increased metabolism. We hypothesize that LAT brings associated TRAF3 into the SMAC. Therefore, immunoprecipitates of the TCR/CD28 complex would contain the LAT signalosome, harboring TRAF3. 82

100 Figure 18. LAT association with TRAF3 in HEK293T cell transfections. (a, d) TRAF3 and LAT constructs used in HEK293T cell transfections. (b and e) Immunoprecipitation of HA-tagged TRAF3 from transfected HEK293T cell lysates is shown at left, with total cell lysate input prior to immunoprecipitation shown at right. Western blot analysis was performed for the specified proteins. (b, e) Quantification of bands was performed by dividing the relative expression of LAT over TRAF3 (c, f). The fold change was calculated as cells transfected with the mutant plasmids, normalized to the values obtained for cells transfected with the control wild type LAT plasmid. Western blots were cropped to focus upon specific proteins indicated. Error bars indicate SEM of mean values of two experiments. 83

101 Figure 19. TRAF3 association with Grb2 family members in HEK293T cell transfections. HEK293T cells were transfected with TRAF3 or ΔTRAF3 (TRAF3 lacking the TRAFC domains), and Gads (a) or Grb2 (b). Immunoprecipitation of HAtagged TRAF3 from transfected HEK293T cell lysates is shown in upper panels, with total cell lysate input prior to immunoprecipitation shown in lower panels. Western blots were cropped to focus upon specific proteins indicated. Data are representative of 3 independent experiments. 84

102 Figure 20. TRAF3 and Gads membrane clustering upon TCR/CD28 stimulation. Primary mouse splenic T cells isolated from LMC and T-traf3 -/- mice were stimulated with αcd3/αcd28 Abs for 2 minutes. T cells were permeabilized and stained with an anti-gads Ab (a). The human immortalized T cell line knock-downs (shluc and shgads) were permeabilized and stained with an anti-traf3 Ab (b). A secondary Ab for Alexa-488 was used for fluorescence detection by Total Internal Reflection Fluorescence (TIRF) microscopy. Representative TIRF images are on the left and quantification of each cell for maximum intensity of fluorescence was determined as in Methods (n=50 cells per time point). A T-test was performed to establish statistical significance, ** = P<

103 Figure 21. TCR/CD28-induced membrane clustering of Grb2 and TRAF3. (a and b) T cells were stimulated with αcd3/αcd28 Abs for 2 minutes. The human immortalized T cell lines (shluc and shgrb2) were permeabilized and stained with an anti-traf3 Ab (a). Primary mouse splenic T cells isolated from LMC and T- traf3 -/- mice were permeabilized and stained with an anti-grb2 Ab (b). A secondary Ab for Alexa-488 was used for fluorescence detection by Total Internal Reflection Fluorescence (TIRF) microscopy. Representative TIRF images are on the left and quantification of each cell for maximum intensity of fluorescence was determined as in Methods (n=50 cells per time point). A T-test was performed to establish statistical significance, * = P<0.05. (c) shluc and shtraf3 T cells were stimulated with αcd3/αcd28 Abs for the indicated times and a cellular fractionation was performed, as described in Methods, on whole cell lysates. Western blot analysis of the insoluble (membrane) fractions was performed to detect the designated proteins, with actin serving as loading controls. Data are representative of 2 independent experiments. 86

104 Figure 22. Grb2 and TRAF3 competition for TCR association. shluc and shtraf3 T cells were stimulated via CD3/CD28 for the indicated times. (a) Immunoprecipitation of the TCR (CD3)/CD28 complex was performed on whole cell lysates. The relative amount of Grb2 association with the TCR/CD28 complex was analyzed by Western blot. (b) Whole cell lysates were prepared from HuT28.11 and shgrb2 cells. Immunoprecipitation of CD3 was performed and the relative amounts of TRAF3 were analyzed by Western blot. Data are representative of 2 individual experiments. Blots were cropped to focus upon the specific proteins indicated. 87

105 CHAPTER V: TRAF3 INHIBITION OF BOTH CANONICAL AND NON- CANONICAL IFN R1 SIGNALING PATHWAYS IN T LYMPHOCYTES. Introduction In B cells, IL-6 receptor signaling is inhibited by the recruitment of PTPN22 to JAK1, where PTPN22 dephosphorylates and inhibits JAK1 kinase activity. The recruitment and association of PTPN22 with JAK1 is facilitated by TRAF3, which associates with the IL-6R following its engagement by IL-6. In the absence of TRAF3, PTPN22 does not associate with JAK1, resulting in an increase in activation of the IL-6 signaling pathway [126]. This led us to hypothesize that TRAF3 also inhibits JAK1 signaling pathways, in other hematopoietic cell types. The focus of this dissertation is the regulatory roles of TRAF3 in T lymphocyte activation. In T cells, the IFN receptor, another JAK1-utilizing cytokine receptor, plays important biological roles. We thus investigated the potential role(s) of TRAF3 in regulating IFN R signaling by T cells. IFNs are polypeptides that are secreted from infected cells to limit the spread of viral infections, modulate innate immune responses and activate the adaptive immune system [122]. IFN R is a heterodimeric transmembrane receptor that is composed of two subunits, IFN R1 and IFN R2. Ligation of type 1 IFN to IFN R results in the induction of both canonical and non-canonical pathways. The canonical pathway is JAK/STATdependent while the non-canonical pathways signal through PI3K or MAPK. Both canonical and non-canonical pathways originate from JAK1 activation. With previous research identifying an association between TRAF3 and JAK1 [126], there is reasonable evidence to suggest that TRAF3 regulates both canonical and non-canonical pathways downstream of IFN R-induced JAK1 signaling. 88

106 Upon ligation of type 1 IFN with IFN R in the canonical pathway, the kinases JAK1 and Tyk2 are phosphorylated and activated. Together these two kinases phosphorylate the cytoplasmic transcription factors STAT1 and STAT2. Phosphorylated STAT1 and STAT2 dimerize to form a complex that moves away from IFN R. While in the cytoplasm, the STAT1/2 complex recruits the transcription factor IRF9 to form the tri-molecular complex termed ISGF3. The ISGF3 complex then translocates to the nucleus where it binds to the DNA-binding sequences called ISRE and GAS to initiate interferon induced response genes [122]. One of the two non-canonical IFN R signaling pathways primarily utilizes MAPK. Following stimulation of IFN R by type 1 IFN, Tyk2 activation leads to the phosphorylation of Vav, a guanine nucleotide exchange factor. Vav activation in turn induces the downstream activation of MAPKs, including Erk phosphorylation, which is also altered in TRAF3-deficient T cells after TCR/CD28 activation [7, 19]. Induction of Erk activation promotes STAT1 phosphorylation at S727, as well as pathways leading downstream to cellular growth and differentiation [131]. In T cells, type 1 IFN-induced Vav activation requires simultaneous stimulation via the TCR. Vav activation occurs through signaling from the TCR-induced LAT complex and IFN R-induced Tyk2 phosphorylation. TCR-deficient Jurkat T cells highlight this requirement for joint receptor engagement, as IFN R stimulation is not sufficient for MAPK activation; TCR and IFN R co-stimulation together activate the MAPK pathway [133, 134]. The other non-canonical IFN R signaling pathway utilizes PI3K. Briefly, JAK1 and Tyk2 phosphorylate IRS1 to provide a docking site for PI3K, and activation of STAT3 further stabilizes PI3K binding to IFN R. Induction of PI3K activation leads to 89

107 the activation of Akt and its downstream signaling pathway, including pathways regulated by NF- B. NF- B can also be regulated by TRAF3/TRAF2/cIAP-induced degradation of NIK. Ultimately, the activation of NF- B leads to pro-survival signals, increased expression of GTP-binding proteins, antigen processing and/or presentation proteins, and phosphorylation of STAT1 S727 [131, 136]. There are thus numerous possible pathways TRAF3 can regulate in IFN R signaling, either through JAK1 and/or via additional mechanisms. Thus, we wished to explore the role(s) of TRAF3 in regulating different aspects of these pathways to understand how TRAF3 impacts IFN R signaling in T cells. Results TRAF3 inhibition of canonical IFN R1 signaling Previous research identified an inhibitory role for TRAF3 in JAK/STAT signaling [20, 126], so we initially focused on the canonical IFN R pathway. We hypothesize that TRAF3 negatively regulates JAK1 activation by recruiting the phosphatase PTPN22 to JAK1 (Figure 23). In the absence of TRAF3, we predict that PTPN22 is not recruited to JAK1, allowing constitutive phosphorylation of JAK1 upon IFN R activation. We addressed the experimental questions using two complementary models: a mouse with T cell-specific deletion of TRAF3 in primary T cells (T-traf3 -/- ) and CRISPR-induced TRAF3-deleted human T cell lines (crtraf3 -/- ). We first focused upon the activation of STAT1, as a key early event in IFN R signaling. In primary CD4 + T cells lacking TRAF3, stimulation with type 1 IFN increased in both the phosphorylation of Y701 and S727 of STAT1 (Figure 24a-c). Normalization of relative amounts of 90

108 pstat1 to total STAT1 indicated that the increase in phosphorylation of STAT1 in the TRAF3 deficient cells is due to an increase in total STAT1 levels (Figure 24b, c). Normalization of activated STAT1 levels to total STAT1 levels identified no difference in the ratio of STAT1 phosphorylation at S727, but did show a decrease in the ratio of Y701-phosphorylated/total STAT1 at at 30 minutes in the T-traf3 -/- compared to LMC T cells (Figure 24b, c). T-traf3 -/- mice have an increased number of thymus-derived Treg [20]. To determine if the increase in total STAT1 levels was specific to CD4 + T cells or due to the increase in Tregs, we isolated CD4 + Treg - splenic T cells to assess STAT1 activation in the absence of Tregs. The increased total and phosphorylated STAT1 following INF treatment was recapitulated in CD4 + Treg - T cells (Figure 24d). Together these data identify a new inhibitory role for TRAF3 in the regulation of total STAT1 protein and its Y701 phosphorylation levels, independent of Treg. To determine if TRAF3 inhibition of IFN R signaling also occurred in human T cells, we performed the same experiment using the crtraf3 -/- cell line. Surprisingly, deletion of TRAF3 in human immortalized T cells did not significantly increase phosphorylation of Y701 or total levels of STAT1 as seen in the primary T cells, but these T cells did display an increase in phosphorylation of STAT1 on the S727 residue (Figure 24e). Furthermore, type 1 IFN-induced phosphorylation of STAT2 at Y690 was increased in the crtraf3 -/- cells compared to the control cells (Figure 24f). Due to the species specificity of the STAT2 Y690 antibody, we were unable to explore this phenotype in primary mouse T cells. Deletion of TRAF3 in primary mouse but not the T cells significantly altered STAT1 levels in primary T cells but did not alter other IFN Rassociated signaling total protein levels (Figure 25a, b). Together these data support an 91

109 inhibitory role for TRAF3 in the canonical IFN R signaling pathway, with overlapping but possibly distinct mechanisms in mouse vs. human T cells or alternatively, in primary vs. transformed T cells. The impact of TRAF3 and PTPN22 on T cell IFN R activation To determine if PTPN22 is playing an important role in the IFN R signaling pathway in T cells, we developed a PTPN22-deficient subclone of the HuT28.11 human T cell line, using CRISPR/Cas9 technology (crptpn22 -/- )(Figure 26a). Neither deletion of PTPN22 nor the production of the cell line altered the total protein levels of IFN R signaling proteins (Figure 26b). Upon type 1 IFN stimulation, crptpn22 -/- cells displayed enhanced phosphorylation of both Y701 and S727 sites on STAT1, confirming findings above, and suggesting that PTPN22 plays an important role in TRAF3-mediated restraint of STAT1 activation (Figure 27a). The enhanced activation of STAT1 was confirmed in 2 different subclones of crptpn22 -/- (Figure 26c). Deletion of PTPN22 enhanced STAT1 activation on S727 to a similar level as that seen in the crtraf3 -/- cell line at 30 mins (Figure 27a). Interestingly, the crptpn22 -/- cells also displayed enhanced IFN-stimulated Y701 phosphorylation in comparison to the HuT28.11 and crtraf3 -/- cells (Figure 27a). The similar increase in S727 activation observed in the crtraf3 -/- and crptpn22 -/- cell lines suggests that PTPN22 and TRAF3 work together to regulate this site. Because PTPN22 is a tyrosine phosphatase, the increase in S727 phosphorylation of STAT1 would not be a direct effect; rather, PTPN22 must be regulating a serine-threonine kinase that phosphorylates S727. Because PTPN22 appears 92

110 to have a greater impact than TRAF3 in determining levels of STAT1 Y701 phosphorylation, PTPN22 may only partially require TRAF3 for JAK1 association. To determine if TRAF3 and PTPN22 regulated STAT1 production, immortalized human T cells were stimulated with Type 1 IFN for 3 days. Initially, TRAF3 deficient T cells displayed an increase in total STAT1 levels, though STAT1 production after prolonged exposure increased more in the HuT28.11 than the TRAF3 and PTPN22 deficient T cells (Figure 27b, c). These data suggest TRAF3 regulates the basal levels of STAT1 in T cells and T cells require TRAF3 and PTPN22 for full production of prolonged IFN-induced STAT1 production. TRAF3 associates with and inhibits IFN R activation Upon finding an increase in the amount of phosphorylated STAT1 protein in IFN -stimulated T cells, we explored more proximal IFN R signaling events by determining the effects of TRAF3 on the IFN R itself. Similar to previous research [126], PTPN22 association with JAK1 increased upon type 1 IFN stimulation in the HuT28.11 T cell line (Figure 28a). These results further support a role for PTPN22 in regulating IFN R signaling. To determine if other JAK1 targets were altered in the absence of TRAF3, we examined IFN R1 phosphorylation. Upon type 1 IFN stimulation, JAK1 phosphorylates IFN R, providing the docking sites for the STAT family of proteins [122]. In the absence of TRAF3, IFN R1 phosphorylation increased in comparison to the control HuT28.11 T cell line upon type 1 IFN stimulation (Figure 28b). Previous research identified the association of TRAF3 and JAK1 via immunoprecipitation [20, 126], which does not determine direct vs indirect association 93

111 between two proteins. We thus considered the possibility that TRAF3 may associate with other IFN R proteins directly, as it does with the IL-6R and IL-2R. To address this question, we transfected HEK293T cells with full length TRAF3 or a TRAF3 mutant lacking the TRAFC domain (TRAF3 TRAFC), together with full length IFN R1 constructs. Results identified an association between IFN R1 and TRAF3, requiring the TRAFC domain (Figure 28c). While the association between JAK1 and PTPN22 upon type 1 IFN stimulation is potentially driven by TRAF3, this finding demonstrates that TRAF3 associates with and regulates IFN R1. The role of TRAF3 in canonical IFN R-mediated transcriptional regulation in T cells Inasmuch as TRAF3 inhibits signaling mediated by IFN R in T cells, we predicted that both ISRE and GAS response genes would also be upregulated in resting T cells. To test this prediction, mrna was isolated from resting CD4 + T-traf3 -/- and control T cells. Expression of IFN R1, which is itself down-regulated after type 1 IFN stimulation, was decreased in TRAF3-deficient T cells (Figure 29a). Despite the decrease in IFN R1 expression, CxCl9, a GAS induced gene, and Pdcd-1, an ISRE induced gene, were both upregulated at a basal level in the absence of TRAF3 (Figure 29b,c). This is consistent with continuous exposure to a low level of type 1 IFN in T- traf3 -/- mice in vivo, which is likely, and could also be partially the result of a low level of IFN R signaling in the absence of TRAF3 that does not require type 1 IFN. To further confirm the mrna data, we examined the level of PD-1, the protein encoded by Pdcd-1, on unstimulated primary splenocytes from the LMC or T-traf3 -/- mice. There was an increased percentage of T-traf3 -/- CD4 + T cells compared to the 94

112 LMC mouse (Figure 30b). This result is explained by the increase in Treg numbers in the T-traf3 -/- mouse described in our previous publication [20]. Despite this increase, there was no difference in the number of antigenexperienced CD4 + T cells (CD4 + CD44 hi ) (Figure 30c). Consistent with gene expression data, PD-1 + cells represented a larger proportion of CD4 + TRAF3-deficient T cells compared to LMC (Figure d). Furthermore, this increase occurred in both CD4 + Treg - antigen-experienced (CD44 + ) and inexperienced (CD44 - ) T cells (Figure 30e). Thus, neither Treg numbers nor antigen exposure accounts for the increase in PD-1 expression on TRAF3-deficient T cells. PD-1 upregulation in the TRAF3-deficient T cell occurs in the absence of antigen activation. Similar to CD4 + T cells, an increased proportion of CD8 + T cells also displayed PD-1 expression (Figure 30f) although there is no alteration in CD8 + T cell numbers in the T-traf3 -/- mouse [7]. IFN R response gene mrna levels are thus globally upregulated in primary TRAF3-deficient CD4 + T cells with protein levels of PD-1 upregulated on both CD4 + and CD8 + T cells. TRAF3 regulation of the non-canonical pathways of IFN R signaling. The increase in total STAT1 and Y701 phosphorylation observed in primary T cells was not recapitulated in the crtraf3 -/- cell line. Instead, the crtraf3 -/- T cells displayed an increase in S727 phosphorylation of STAT1 (Figure 24a). The increased S727 was not observed in the T-traf3 -/- T cells when S727 levels were normalized to the increased STAT1 levels (Figure 24b). These findings suggested that the two types of T cells use distinct mechanisms to regulate IFN R signaling. To explore this possibility 95

113 further, we focused on the two non-canonical IFN R signaling pathways that upregulate STAT1 S727 phosphorylation, PI3K and MAPK. We first explored the role of TRAF3 in IFN R induced MAPK kinase activation as depicted in Figure 31. Isolated and type 1 IFN-stimulated CD4 + primary T cells from the T-traf3 -/- mice displayed increased Erk phosphorylation 5 and 15 minutes after stimulation (Figure 32a) compared to the LMC cells. A similar phenotype was also seen in the CD4 + Treg - T-traf3 -/- cells at 30 minutes post-stimulation (Figure 32b). To prevent unnecessary mouse harvest, the 30 minute time point was chosen to show IFN R signaling disparities in both canonical and non-canonical pathways. The crtraf3 -/- cell line also displayed an increase in Erk phosphorylation upon 5 minutes of type 1 IFN stimulation (Figure 32c). This increase was not a result of increased total Erk protein. To explore a role for PTPN22 in this pathway, the crptpn22 -/- cell line was stimulated with type 1 IFN. While the absence of PTPN22 slightly increased Erk phosphorylation at 0 and 5 minutes, Erk activation did not increase to the levels seen in crtraf3 -/- T cells (Figure 33a). Furthermore, prolonged IFN-induced proliferation increased in the crtraf3-/- compared to the HuT28.11 T cells, while loss of PTPN22 did not increase proliferation to that of the TRAF3 deficient T cells (Figure 33b). This suggests that the TRAF3-induced regulation of Erk is not PTPN22-dependent. Because Erk activation leads to STAT1 S727 phosphorylation, the increase in Erk activation seen in the TRAF3-deficient T cells can explain the increase in S727 phosphorylation of STAT1. We also examined the non-canonical IFN R signaling pathway that involves activation of PI3K as depicted in Figure 34. The PI3K pathway can induce 96

114 phosphorylation of STAT1 at S727 upon IFN stimulation [125], suggesting a potential role in type 1 IFN signaling. Similar to the observed Erk activation, Akt activation increased after 5 minutes of type 1 IFN stimulation in the CD4 + T-traf3 -/- cells and 30 minutes in the CD4 + Treg - T-traf3 -/- cells compared to the LMC T cells (Figure 35a,b). This finding was further confirmed in the crtraf3 -/- T cells, which displayed increased Akt activation at all times post-type 1 IFN stimulation (Figure 35c). Because PI3K association to IFN R is stabilized by STAT3, we looked for TRAF3-induced alteration in STAT3 protein levels. Unlike STAT1, STAT3 total protein levels were unaffected by the absence of TRAF3 (Figure 35d). Furthermore, type 1 IFN stimulation of the crptpn22 -/- cell line at 5 minutes caused a similar increase in Akt activation to that seen in the crtraf3 -/- cell line (Figure 36). Thus, TRAF3 inhibits IFN R-induced PI3K signaling indirectly, potentially through its recruitment of PTPN22. Together these results identified a new inhibitory role for TRAF3 in the noncanonical IFN R signaling pathways. The absence of TRAF3 enhanced S727 phosphorylation of STAT1 in both the IFN R-induced MAPK and PI3K pathways. Additionally, our data suggest a cooperation between TRAF3 and PTPN22 to regulate the PI3K pathway. As with the canonical IFN R signaling pathway, our findings point towards an unidentified serine-threonine kinase impacted by IFN R-associated TRAF3 and PTPN22. Conclusions Recently, a negative regulatory role for TRAF3 in IL-6 receptor signaling to B cells was discovered, in which TRAF3-induced inhibition of IL-6R occurs through the 97

115 recruitment of PTPN22 to JAK1. PTPN22 association with JAK1 allows dephosphorylation and subsequently inhibition of JAK1 and the downstream IL-6R signaling pathway. In T cells, TRAF3 also inhibits IL-2R signaling, via recruitment of the phosphatase TCPTP [21]. Further, Chapter 3 demonstrated that TRAF3 associates with PTPN22 to regulate inhibition of the TCR signaling pathway [150]. A previous microarray analysis of differences in gene expression of TRAF3-deficient vs. sufficient T cells identified an overall increase in IFN response genes. The present study thus addressed the hypothesis that TRAF3 inhibits IFN R signaling to T cells. TRAF3 inhibited both canonical and non-canonical IFN R signaling. In the canonical pathway, TRAF3 deficiency in primary CD4 + T cells resulted in increased STAT1 total protein levels, which consequently altered levels of pstat1 as well. Primary naïve TRAF3 -/- T cells displayed these increased STAT1 levels, suggesting previous exposure to IFN in the mouse prior to isolation of these cells, and thus elevated constitutive IFN R signaling. Constant exposure to low levels of type 1 IFN, coupled with the inability to inhibit IFN R signaling in TRAF3-deficient T cells could result in increased STAT1 levels. A second possibility is that the IFN R pathway is constitutively active in the absence of both TRAF3 and IFN stimulation, thus leading to increased STAT1. Current studies are aimed at addressing both these alternatives. Furthermore, JAK1-induced phosphorylation of IFN R1 increased in the crtraf3 -/- immortalized human T cells. These data identify a new inhibitory role for TRAF3 in canonical IFN R signaling, potentially through the recruitment of PTPN22 to JAK1. In the absence of PTPN22, STAT1 Y701 phosphorylation increased to levels above the crtraf3 -/- and HuT28.11 cell levels. This suggests two possibilities. TRAF3 may not 98

116 be the sole protein able to recruit PTPN22 to JAK1. Additionally, there may be another inhibitory function for PTPN22, alternative to dephosphorylation of JAK1. Further studies are needed to clarify how TRAF3 and PTPN22, separately and through their interactions, inhibit the canonical IFN R signaling pathway. While the inhibition of tyrosine kinase JAK1 by TRAF3 and PTPN22 can explain the increase in STAT1 Y701 phosphorylation levels, it cannot directly explain the observed increase in phosphorylation of STAT1 at S727. As a result, we explored the role of TRAF3 in non-canonical IFN R signaling pathways, both of which have the potential to regulate STAT1 S727 phosphorylation. Our results show TRAF3 also plays an inhibitory role in both non-canonical pathways. In IFN R-induced activation of the MAPK pathway, TRAF3 deficiency enhanced Erk activation in both primary mouse and immortalized human T cells. Interestingly, loss of PTPN22 in the crptpn22 -/- cell line did not recapitulate the crtraf3 -/- phenotype. Thus, TRAF3 inhibits MAPK signaling, but this does not require PTPN22. The second non-canonical pathway signals through PI3K, which leads to Akt activation. In the absence of TRAF3, both primary and immortalized T cells displayed increased Akt activation at 5 minutes, confirming an inhibitory role for TRAF3 in PI3K signaling. Unlike Erk activation, loss of PTPN22 leads to an increase in Akt activation, similar to the crtraf3 -/- phenotype. This suggests the inhibitory role for TRAF3 in PI3K signaling can occur at least in part through the inhibitory action of PTPN22. Interestingly, PTPN22 can potentially inhibit two different kinases that regulate Akt activation, JAK1 and PI3K. Further research is needed to determine the exact role for TRAF3 in IFN R induced PI3K signaling. 99

117 Figure 23. Predicted model of TRAF3 regulation of canonical IFN R1 signaling via JAK1. Upon type 1 IFN engagement by IFN R1/2, JAK1 and Tyk2 are phosphorylated and activated. Activated JAK1 and Tyk2 phosphorylate STAT1 and STAT2, causing the dimerization of STAT1:STAT2 and their subsequent dissociation from IFN R. In the cytoplasm, the STAT1:STAT2 complex recruits IRF9 to form a complex termed ISGF3. The ISGF3 complex then translocates to the nucleus where it interacts with the DNA-binding sequence termed interferon-stimulated response element (ISRE), thus inducing interferon response genes. Our hypothesis is that TRAF3 negatively regulates IFN R1/2 signaling by recruiting PTPN22 to JAK1. The recruitment of PTPN22 to JAK1 results in the dephosphorylation and inhibition of JAK1. Inhibition of JAK1 prevents STAT1 phosphorylation and activation of downstream signaling. In the absence of TRAF3, the IFN R1/2 signaling pathway is enhanced and inhibition cannot occur without the recruitment of PTPN

118 Figure 24. Requirement for TRAF3 in type 1 IFN-induced STAT1 activation. T cells were stimulated with type 1 IFN for the indicated minutes. Whole cell lysates were prepared from primary mouse splenic CD4 + (a-c) or CD4 + Treg - T cells (d) from LMC or T-traf3 -/- mice or from the immortalized human T cell lines HuT28.11 or its subclone crtraf3 -/- (e, f), as described in Methods. Lysate proteins were separated by SDS-PAGE and Western blot analysis was performed to detect the indicated proteins. Western blots were cropped to focus upon the specific proteins indicated. Expression levels of pstat1 normalized to total STAT1 (b, c) were calculated as: (pstat1/actin)/(stat1/actin) = pstat1 Relative expression level (REL). Fold change was then determined based upon the obtained expression levels, as follows. (pstat1 REL)/(pSTAT1 REL of control time point 0) = pstat1 Fold change. PSTAT1 not normalized to total STAT1 was normalized to actin alone but used the same fold change calculations (b, c). Data from at least 3 independent experiments were pooled and the mean values ± SEM are shown. A 2-way ANOVA was performed and was not corrected for multiple comparisons using the Fisher s LSD test to establish statistical significance; * = P<0.05, ** =P 0.01 comparing the T-traf3 -/- values to the respective LMC time point. Data are representative of 3 individual experiments. 101

119 Figure 25. Levels of IFN R signaling proteins in TRAF3 deficient T cells. Total protein levels from unstimulated HuT28.11 and the subclone crtraf3 -/- (clone 45) (a) or splenic primary LMC and T-traf3 -/- (b) T cells were determined by Western blot analysis. Western blots were cropped to focus upon specific proteins indicated. Quantification was performed by normalizing relative amounts of the indicated protein to actin. The relative values were then used to determine the fold change by normalizing the crtraf3 -/- or T-traf3 -/- relative values to the control values. Data from at least 4 experiments were pooled and the mean values ± SEM are shown. 102

120 Figure 26. Production and selection of crptpn22 -/- clones. (a) Western blot analysis for PTPN22 expression in unstimulated crptpn22 -/- clones derived from the HuT28.11 T cell line. (b) Total levels of the IFN R1 signaling proteins indicated were determined by Western blot analysis (top) and relative amounts quantified (bottom). (c) Whole cell lysates were obtained from type 1 IFN-stimulated HuT28.11 and crptpn22 -/- (clones 43 and 120) T cells for the indicated minutes. Line indicates a croppingt of the image to remove gel empty lanes. Western blot analysis was performed upon proteins separated by SDS-PAGE to detect the activation levels for IFN R1 signaling induction. Western blots were cropped to focus upon the indicated proteins. Data are representative of at least 3 individual experiments per panel. 103

121 Figure 27. STAT1 activation in the absence of TRAF3 or PTPN22. STAT1 activation was determined by Western blot analysis on SDS-PAGE gels of proteins separated from whole cell lysates. Lysates were derived from type 1 IFN-stimulated HuT28.11, crtraf3 -/- and crptpn22 -/- T cells after short stimulation (a) and prolonged stimulation (b). Western blots were cropped to focus upon the specific proteins indicated. (c) Compiled data from prolonged type 1 IFN stimulation in (b) with error bars indicating ± SEM of mean values from pooled experiments. A 2-way ANOVA was performed and corrected for multiple comparisons using Turkey analysis to determine statistical significance in b; * = P<0.05 and ** = P<0.01 comparing the CRISPR knockout cell line to the respective control time point. Data are representative of 3-4 individual experiments. 104

122 Figure 28. TRAF3 association with and regulation of the IFN R1 complex. (a) Whole cell lysates were obtained from type 1 IFN-stimulated HuT28.11 T cells. Using a JAK1-specific Ab, an immunoprecipitation was performed on the whole cell lysates and PTPN22 association was determined by Western blot analysis. (b) IFN R1 immunoprecipitation from type 1 IFN-stimulated HuT28.11 and crtraf3 -/- whole cell lysates. Following SDS-PAGE gel separation of proteins, Western blot analysis (left) was performed to determine IFN R1 phosphorylation levels (py) and total levels with relative quantification of two similar experiments presented in the bar graph to the right of the representative blot. (c) HEK293T cells were transfected with TRAF3 or the TRAFC-deleted TRAF3 mutant (TRAF3ΔTRAFC) and the IFN R1 construct. Cells were lysed and IFN R1 was immunoprecipitated. Error bars in bar graph to the right of (b) indicate SEM of values from pooled experiments (b Western blots were cropped to focus upon the specified proteins. Control samples, identified with a c, were from cells stimulated for 15 minutes in the absence of immunoprecipitation Ab. Data shown are representative of 2 (a,b) or 3 (c) independent experiments. 105

123 Figure 29. mrna expression of IFN R1 and IFN response genes in TRAF3- deficient T cells. RNA was extracted from unstimulated CD4 + LMC or T-traf3 -/- T cells. mrna levels were determined by qpcr for Ifnar1 (a), Cxcl9 (b) or Pdcd1 (c). Both Ifnar1 and Cxcl9 levels were individually normalized to Gapdh and Actin b (a,b), while Pdcd1 was normalized to Hprt (c). All samples were normalized to the control LMC sample values. Data shown are representative of 2-3 independent experiments. (c, Data obtained by Emma Hornick) 106

124 Figure 30. PD-1 surface expression by CD4 + and CD8 + T cells. (a) Gating strategy for detection of PD-1 expression on unstimulated splenic T cells isolated from LMC and T-traf3 -/- mice. Pooled data of percent CD4 + (b), CD44 + CD4 + (c), PD-1 + of CD4 + (d), PD-1 + of Foxp3 - CD4 + (e) and PD-1 + of CD8a + (f) T cells. Data shown are representative of 3 (a-d) independent experiments or 1 (e,f) experiment with 1 mouse per experiment. T-test was performed to determine statistical significance in a-d as identified. (Data obtained by Emma Hornick) 107

125 Figure 31. Predicted model of TRAF3 regulation of non-canonical IFN R1 signaling via JAK1. IFN R1/2 activates both canonical and non-canonical signaling pathways. In one of the latter, JAK1 and Tyk2 are phosphorylated and activated. This leads to the activation of Vav and the Ras, RAF, Mek pathway, ultimately leading to Erk activation. Erk activation results in cellular growth, differentiation and serine phosphorylation of STAT1. Another mechanism that positively regulates VAV operates through TCR activation. Because TRAF3 plays a role in TCR signaling, we focused this study upon IFN R1/2 stimulation only. Our hypothesis is that TRAF3 negatively regulates IFN R1/2 signaling by recruiting PTPN22 to JAK1. The recruitment of PTPN222 to JAK1 allows PTPN22 to dephosphorylate JAK1 and subsequently inhibit JAK1. Inhibition of JAK1 prevents the activation of IFN R1 and thus the downstream induction of the Ras, RAF, Mek pathway. In the absence of TRAF3, induction of the IFN R1/2 signaling pathway cannot be properly regulated without recruitment of PTPN22 to JAK1. 108

126 Figure 32. Inhibition by TRAF3 of early type 1 IFN-induced Erk activation. T cells were stimulated with type 1 IFN for the indicated times. Whole cell lysates were prepared from primary mouse splenic CD4 + (a) or CD4 + Treg - (b) T cells from LMC or T-traf3 -/- mice or from the human immortalized T cell lines HuT28.11 and the subclone crtraf3 -/- (c), as described in Methods. Following SDS-PAGE separation of proteins, Western blot analysis was performed to detect the activation of Erk p44/p42. Blots were cropped to focus upon the specific proteins indicated. Data are representative of > 3 individual experiments. 109

127 Figure 33. Erk activation and proliferation in the absence of TRAF3 and PTPN22. (a) Erk activation of p44/p42 was determined by Western blot analysis of SDS-PAGE-separated whole cell lysates derived from type 1 IFN-stimulated HuT28.11, crtraf3 -/- and crptpn22 -/- T cells. Western blots were cropped to the specific protein indicated. Quantification of pooled data is depicted below. (b) Prolonged type 1 IFN stimulation of HuT28.11, crtraf3 -/- and crptpn22 -/- T cells for the indicated number of days. Data were compiled and the average values are shown with error bars indicating SEM of pooled experiments. A 2-way ANOVA was performed to determine statistical significance in b; ** = P<0.01 comparing the CRISPR knockout cell line to the respective control time point. Data are representative of > 3 individual experiments. 110

128 Figure 34. Predicted model of TRAF3 regulation of the non-canonical IFN R1 signaling pathway via JAK1 and PI3K. Binding of type 1 IFN to IFN R1/2 induces a non-canonical NF- B pathway. Upon stimulation of IFN R1/2, JAK1and Tyk2 are activated, which leads to the recruitment of PI3K to the receptor. PI3K is further stabilized to IFN R by STAT3. Activation of PI3K leads to the activation and recruitment of Akt to the receptor. Akt activation positively regulates the activation of NF- B through the degradation of the NF- B inhibitor, I B. Upon activation, NF- B translocation to the nucleus allows for regulation of GTP binding protein expression and Ag processing/presentation proteins, as well as an increase transcriptional regulation of survival signals. Our first hypothesis is that TRAF3 negatively regulates IFNaR1 signaling by recruiting PTPN22 to JAK1. PTPN22 associates with and dephosphorylates JAK1, subsequently inhibiting JAK1. Inhibition of JAK1 prevents JAK1-induced activation that leads to Akt phosphorylation and TRAF translocation. Our second hypothesis is that TRAF3 positively enhances IFN R1/2 signaling by PTPN22 association with PI3K. PTPN22 dephosphorylates the inhibitory site on PI3K allowing for increased phosphorylation of Akt. 111

129 Figure 35. Regulatory role of TRAF3 in type 1 IFN induced Akt activation. T cells were stimulated with type 1 IFN for the indicated times. Whole cell lysates were prepared from primary mouse splenic CD4 + (a, d) or CD4 + Treg - T cells (b) from LMC or T-traf3 -/- mice or the human immortalized T cell lines HuT28.11 and its subclone crtraf3 -/- (c), as described in Methods. Western blot analysis was performed on proteins separated by SDS-PAGE to detect S473 phosphorylation of Akt (a-c) and STAT3 (d), and cropped to focus upon the specific proteins indicated. Data from at least 2 individual experiments were compiled and the average values are shown below with error bars indicating ± SEM of pooled experiments. 112

130 Figure 36. Akt activation in the absence of TRAF3 and PTPN22. S473 phosphorylation of Akt was determined by Western blot analysis on SDS-PAGEseparated whole cell lysates derived from type 1 IFN-stimulated HuT28.11, crtraf3 - /- and crptpn22 -/- T cells. Western blots were cropped to the specific protein indicated. Data from 3 individual experiments were compiled and the average values are shown below with error bars indicating ± SEM of pooled experiments. 113