Regulation of intrinsic activation thresholds of T cells

Item Type dissertation

Authors Matson, Courtney

Publication Date 2020

Abstract T cells are activated when their T cell receptor (TCR) senses peptide-MHC (pMHC) molecules from pathogens and tumors. A network of signaling molecules downstream of the TCR drives the extent and nature of subsequent cellular responses. The activation...

Keywords activation threshold; T cell; T-Lymphocytes--immunology

Download date 04/10/2021 10:52:05

Link to Item http://hdl.handle.net/10713/14476 Courtney Alexandra Matson Email: [email protected] 660 West Redwood Street, Howard Hall Room 332D, Baltimore, MD, 21201

EDUCATION PhD, University of Maryland, Baltimore, MD December 2020 Graduate Program: Molecular Microbiology and Immunology BS, James Madison University, Harrisonburg, VA May 2015 Major: Biotechnology, Cum Laude with Distinction

RESEARCH EXPERIENCE UMB, Dept. of Microbiology and Immunology Baltimore, MD Summer 2016 - Present PhD Candidate Principle Investigator: Nevil Singh, PhD Regulation of intrinsic activation thresholds of T cells.

UMB, Dept. of Microbiology and Immunology Baltimore, MD Spring 2016 Rotation Student Principle Investigator: Jonathan Bromberg, MD, PhD Optimizing surface lymphotoxin α1β2 and lymphotoxin β receptor regulation assays for ligand stimulation and peptide blockade.

UMB, Dept. of Microbiology and Immunology Baltimore, MD Winter 2015 Rotation Student Principle Investigator: Nevil Singh, PhD Evaluate the role of antigen rate of change in T cell activation.

UMB, Dept. of Microbiology and Immunology Baltimore, MD Summer 2015 Rotation Student Principle Investigator: Sharon Tennant, PhD Characterizing pathogenesis of Salmonella enterica Typhimurium ST19 and ST313 clinical isolates.

JMU, Dept. of Biology Harrisonburg, VA Fall 2012 – Spring 2015 Undergraduate Research Internship Principle Investigator: Timothy Bloss, PhD Evaluate the effects on unfolded protein response, differentiation and co-localization of HSP-4 in fluorescently marked cell lineages in C. elegans resulting from depletion of the icd-1/icd-2 complex.

JMU, Dept. of Biology Harrisonburg, VA Fall 2011 – Fall 2013 Undergraduate Research Experience Principle Investigators: Steven Cresawn, PhD and Louise Temple, PhD Isolation, discovery, and genomic investigation of novel Bacillus thuringiensis bacteriophages

TEACHING EXPERIENCE UMB, Dept. of Microbiology and Immunology Baltimore, MD Summer 2020 Course Organizer and Lecturer – Overview of Immunology Summer Course Organized syllabi, course materials, lecturers, and group projects for three-week virtual immunology course for summer internship students as well as incoming and current graduate students.

UMB, School of Medicine Baltimore, MD Fall 2019 Small Group Leader – Host Defense and Infectious Disease Presented course material and associated journal articles to facilitate critical discussion of seminal journal articles.

UMB, Dept. of Microbiology and Immunology Baltimore, MD Spring 2019 – Spring 2020 Workshop and Paper Discussion Leader – Basic Immunology Led workshop and paper discussion to develop hypothesis building and experimental design skills.

UMB, Dept. of Microbiology and Immunology Baltimore, MD Fall 2017 – Fall 2019 Lecturer – Advanced Immunology Lectured and led paper discussions on Antigen Receptor Signaling and Trained Immunity.

JMU, Dept. of Biology Harrisonburg, VA Spring 2015 Undergraduate Teaching Assistant – Genetics and Development Provided support during classroom component of coursework with 70 - 80 students. Responsible for grading quizzes and recording attendance and grades. Held weekly review session to discuss both classroom and laboratory topics from course.

JMU, Dept. of Biology Harrisonburg, VA Spring 2015 Undergraduate Teaching Assistant – Immunology Provided support during laboratory component of coursework with 15 students. Aided in proctoring and grading of exams and other class assignments. Responsible for recording attendance and grades.

JMU, Dept. of Biology Harrisonburg, VA Fall 2012 - Summer 2014 Undergraduate Teaching Assistant - Organisms Provided support during laboratory component of coursework with 20 - 30 students. Aided in proctoring and grading of exams and other class assignments.

UNIVERSITY SERVICE Student Focus Session Discussion Groups Spring 2020 Student Organizer Molecular Microbiology and Immunology Admissions Committee Winter 2019 Student Representative Immunology Journal Club Fall 2017-Spring 2018 Student Organizer

Student Led Qualifying Exam Practice Sessions Spring 2017-Fall 2018 Student Organizer

ABSTRACTS & PRESETNATIONS Publications Matson CA, Choi S, Livak F, Zhao B, Mirta A, Love PE, Singh NJ (2020). "CD5 dynamically calibrates basal NF-κB signaling in T cells during thymic development and peripheral activation." Proc Natl Acad Sci U S A.

Matson CA and Singh NJ (2020). "Manipulating the TCR signaling network for cellular immunotherapy: Challenges & opportunities." Mol Immunol 123: 64-73.

Ramachandran G, Panda A, Higginson EE, Ateh E, Lipsky MM, Sen S, Matson CA, Permala-Booth J, DeTolla LJ, Tennant SM (2017). "Virulence of invasive Salmonella Typhimurium ST313 in animal models of infection." PLoS Negl Trop Dis 11(8): e0005697.

A diverse repertoire of memory CD4 T cells is maintained in the absence of the cytokine common gamma chain. Johnson L, Matson CA, Martins AJ, Tsang JS, Singh NJ. [Manuscript in prep]

The specificity of autoantibodies in a mouse model of arthritis driven by monoclonal T cells reactive to a systemic self-antigen. Matson CA, Sultan-Reisler Y, Fioretti S, Singh NJ. [Manuscript in prep]

Dynamic downregulation of CD19 by B cells underlies the escapte phenotype after anti-CD19 directed immunotherapy. Fioretti S, Matson CA, Rosenberg KM, Singh NJ. [Manuscript in prep]

Oral Presentations The heterogenous depot of NFκB available for each T cell’s activation is regulated by CD5. Matson CA, Singh NJ. Tumor Immunology and Immunotherapy Seminar Series, University of Maryland Baltimore, Baltimore, MD, June 2020.

The basal transcriptional potential of individual peripheral T cells is actively imprinted during development. Matson CA, Singh NJ. Graduate Student Symposium, University of Maryland Baltimore, Baltimore, MD, June 2019

The self-reactivity intrinsic to all T cells involves the CD5-dependent modulation of NFκB. Matson CA, Singh NJ. Graduate Research Conference, University of Maryland Baltimore, Baltimore, MD, March 2019.

T cell receptor sensitivity tunes in response to chronic antigen stimulation. Matson CA, Choi S, Singh NJ. Graduate Student Symposium, University of Maryland Baltimore, Baltimore, MD, June 2018.

Tuning of TCR signaling pathways in chronically stimulated CD4+ T cells in vivo. Matson CA, Choi S, Singh NJ. American Association of Immunology: Immunology 2018, Austin, TX, May 2018.

Reliance on γc cytokines distinguishes mechanisms maintaining CD4 and CD8 memory T cells. Johnson L, Matson CA, Warner R, Singh NJ. Graduate Student Symposium, University of Maryland Baltimore, Baltimore, MD, June 2017.

Dose or Rate: What Really Activates a T Cell? Matson CA and Singh NJ. Graduate Student Symposium, University of Maryland Baltimore, Baltimore, MD, June 2016.

Characterization of cell-specific responses in C. elegans experiencing misfolded protein stress: how do some cell types save themselves while others die? Matson CA, Wallace EN. Biology Symposium, James Madison University, Harrisonburg, VA, April 2015.

Bacillus Bacteriophages. Balsamo J, Matson CA, Oehler M, Rudloff M, Sauder B. Viral Discovery Symposium, James Madison University, Harrisonburg, VA, April 2012.

Bacillus thuringiensis Bacteriophages. Matson CA, Langouette C, Cain R. Phage Phaire, James Madison University, Harrisonburg, VA, October 2012.

Posters Tuning of TCR signaling pathways in chronically stimulated CD4+ T cells in vivo. Matson CA, Choi S, Singh NJ. American Association of Immunology: Immunology 2018, Austin, TX, May 2018.

Tuning of TCR signaling pathways in chronically stimulated CD4+ T cells in vivo. Matson CA, Choi S, Singh NJ. Graduate Research Conference, University of Maryland Baltimore, Baltimore, MD, March 2018.

A diverse repertoire of memory CD4 T cells is maintained in the absence of the common gamma chain. Johnson L, Matson CA, Warner R, Singh NJ. American Association of Immunology: Immunology 2017, Washington, DC, May 2017.

The Effects of Misfolded Protein Stress on Different Cell Types in C. elegans. Kareem S, Matson CA, Wallace EN, Bloss TA, Slekar K. National Conference of Undergraduate Research, University of Kentucky, Lexington, KY, April 2014.

The Effects of Misfolded Protein Stress on Different Cell Types in C. elegans. Kareem, S., Matson CA, Wallace EN, Bloss TA, Slekar K. Biology Symposium, James Madison University, Harrisonburg, VA, April 2014.

Bacillus Bacteriophages. Balsamo J, Matson CA, Oehler M, Rudloff M, Sauder B. Viral Discovery Symposium, James Madison University, Harrisonburg, VA, April 2012.

HONORS & AWAR DS American Association of Immunology, AAI Trainee Abstract Award May 2018 Graduate Research Conference, Outstanding Presentation Award March 2018

PROFESSIONAL MEMBERSHIPS American Association of Immunology December 2015 – Present American Association for the Advancement of Science May 2020 - Present Association for Women in Science May 2020 - Present

Abstract

Title of Dissertation:

Regulation of intrinsic activation thresholds of T cells

Courtney A. Matson, Doctor of Philosophy, 2020

Dissertation Directed by:

Nevil Singh, PhD, Assistant Professor Department of Microbiology and Immunology University of Maryland School of Medicine

T cells are activated when their T cell receptor (TCR) senses peptide-MHC

(pMHC) molecules from pathogens and tumors. A network of signaling molecules downstream of the TCR drives the extent and nature of subsequent cellular responses. The activation threshold (AT) of these signaling pathways is a critical checkpoint for T cell responses. Here, we examined mechanisms by which the AT of a T cell is first determined and how it changes during the course of responding to antigen. The initial AT of a T cell is set during development by calibrating to how strongly it senses pMHC in the thymus.

This calibration affects the surface levels of a receptor CD5, whose subsequent role is poorly defined. We found that CD5, independent of the TCR, sets basal levels of IκBα in

T cells. Since IκBα critically modulates the transcription factor NFκB, which regulates multiple T cell functions including cell-survival, we hypothesized that variations in basal

AT of T cells stem from varying NFκB depots maintained by CD5. Indeed, blocking NFκB abolished differences in cell-survival of thymocytes with different CD5 levels. The initial

heterogeneities are further modified when peripheral T cells encounter antigen. Resulting memory T cells acquired higher CD5 levels and continuously required CD5 expression to maintain higher IκBα expression. If the stimulating antigen was not cleared efficiently, peripheral T cells further ‘tuned’ their AT in the opposite direction and resulted in loss of sensitivity to antigen (as seen in exhausted T cells). Importantly, this AT tuning involved additional regulation of TCR-proximal kinases such as Zap70 and was reversible in vivo, but not in vitro. We also found that rather than just the duration of antigen exposure, AT- tuning was potentially influenced by the rate at which antigen changes in vivo. This can help us understand how different persistent pathogens or tumors affect T cell responses, potentially based on the rates at which they replicate in the host. Finally, we characterized compounds that can target a T cell’s AT, identified in a high-throughput pharmacological screen, to potentially isolate drugs for altering T cell function during these physiological contexts.

Regulation of intrinsic activation thresholds of T cells

by Courtney A. Matson

Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, Baltimore in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2020

© Copyright 2020 by Courtney Matson

All rights Reserved

Acknowledgements

To my mentor, Nevil J. Singh, thank you for your unwavering patience, support, and guidance these last five years. I truly could not have imagined a better environment to learn and grow during my graduate training.

To Allison Gerber, thank you for your never-ending cheerleading and encouragement for all of my endeavors in and out of the lab, especially during the preparation for this defense.

To Kenneth Rosenberg, thank you for always helping to improve my writing and presentation skills as well as providing critical analysis about my research.

To Gideon Wolf and Zach Fasana, thank you for your critical input on this thesis in the early stages and the enjoyable company in lab.

To my Thesis Readers, Dr. Martin Flajnik and Dr. Gregory Szeto, thank you for your insight and suggestions to improve this thesis. Dr. Flajnik, thank you for always providing me with papers to read as well as pushing the boundaries of my knowledge outside of my comfort zone.

To my Thesis Committee, Dr. Arnob Banerjee, Dr. Amit Golding, Dr. Scheherazade Sadegh-Nasseri, and Dr. Eduardo Sontag, thank you for your input and guidance while developing and executing the work in this dissertation.

To the MMI Program Administration, Dr. Heather Ezelle and Dr. Bret Hassel, thank you for making MMI feel like more of a community than a graduate program (and for all of the free food).

To my family and friends, especially my mother, Mindy Matson, father, Eric Matson, sister, Rileigh Matson, and husband, Phil Fleischer, I truly would not have made it here without you. From the beginning, you have all supported my dream to become a scientist and have continued to do whatever you can to help me reach my goals. Thank you for everything that you have done for me. I hope I make you proud.

iii

Table of Contents List of Tables ...... viii

List of Figures ...... ix

List of Abbreviations ...... xv

Chapter 1: The activation threshold of peripheral T cells ...... 1 1.1. Overview of the immune system ...... 1 1.1.1 Prelude...... 1 1.1.2. The internal circuitry of a cell ...... 4 1.1.3. Decision making in the immune system starts with a single cell ...... 8 1.2. Receptor signaling: Triggering, propagation, and interpretation ...... 9 1.3. The concept of a receptor activation threshold ...... 11 1.4. Levels of the TCR signaling cascade and its essential components ...... 30 1.4.1. Proximal TCR signaling components ...... 33 1.4.2. Adapters and intermediates ...... 38 1.4.3. Distal components ...... 41 1.4.4. Nuclear components and transcription factors ...... 45 1.5. TCR triggering and generating the first intracellular ‘output’ ...... 47 1.6. Control of the T cell’s activation threshold ...... 52 1.6.1. Sensitivity ...... 52 1.6.2. Discrimination ...... 53 1.7 Lay Summary of Thesis Findings ...... 55

Chapter 2: Materials and Methods ...... 57

Chapter 3: The initial setting of the T cell activation threshold ...... 77 3.1. Introduction ...... 77 3.1.1. Steps in T cell development ...... 77 3.1.2. The tuning of activation thresholds during development ...... 87 3.1.3. CD5 ...... 91 3.2. Chapter Summary ...... 97 3.3. Results ...... 99 3.3.1. Analysis strategy for identifying thymocyte subsets...... 99

iv

3.3.2. Signaling variations during development in thymocytes reflects the heterogeneity of T cell responses ...... 101 3.3.3. CD5 expression is required for the upregulation of IκBα and is independent of canonical regulators ...... 103 3.3.4. CD5 expression provides an NFκB-dependent survival advantage in thymocytes ...... 111 3.3. Discussion ...... 119

Chapter 4: Origins of signaling heterogeneity in steady-state peripheral T cells ... 123 4.1. Introduction ...... 123 4.2. Chapter Summary ...... 126 4.3. Results ...... 127 4.3.1. Gating strategy for analyzing peripheral T cell populations...... 127 4.3.2. Peripheral T cells have altered signaling molecule expression compared to their thymic counterparts...... 129 4.3.3. CD5, not TCR signals, is necessary for the continued upregulation of IκBα in the periphery and is independent of canonical regulators ...... 136 4.3.4. CD5 upregulation of IκBα acts at a post-transcriptional level of regulation and involves a saturable mechanism ...... 144 4.3.5. Antigen exposure further increases the expression of CD5 and IκBα ...... 158 4.3.6. CD5 alone is sufficient to upregulate IκBα but requires a lymphocytic signaling mediator and a functional cytoplasmic tail ...... 160 4.3.7. CD5 increases the NFκB depot in peripheral T cells ...... 166 4.4. Discussion ...... 172

Chapter 5: Reversible activation threshold modulation from chronic antigen stimulation ...... 176 5.1. Introduction ...... 176 5.1.1. Tolerance ...... 177 5.1.2. The view past Regulatory T cells ...... 180 5.1.3. Recessive forms of tolerance: Anergy, Exhaustion and Deletion ...... 183 5.1.4. The molecular-genetic signatures of peripheral recessive tolerance ...... 193 5.1.5. Reversibility of hypo responsiveness in T cells ...... 195 5.1.6. The 5C.C.7 → mPCC model system for tolerance ...... 197 5.2. Chapter Summary ...... 201 5.3. Results ...... 201

v

5.3.1. Adoptive transfer of TCR-Tg CD4+ T cells into cognate-self-antigen murine model system ...... 201 5.3.2. Chronic antigen exposure dampens T cell responsiveness ...... 203 5.3.3. In vivo antigen deprivation restores some functional capacity and signaling machinery of chronically stimulated T cells ...... 208 5.3.4. Short term in vitro antigen deprivation does not restore functional capacity in chronically stimulated T cells and reveals distinct signaling alterations ...... 212 5.3.5. High dose antigen stimulation in vitro reveals two distinct functional populations of chronically stimulated T cells ...... 214 5.3.6. Pharmacological treatment does not alter the functional capacity or signaling molecule expression in chronically stimulated T cells ...... 216 5.3.7. Bypassing the proximal TCR signaling machinery restores functional capacity of chronically stimulated T cells ...... 222 5.3.8. Signaling molecules are altered in a chronic self-antigen model system ...... 225 5.4. Discussion ...... 234

Chapter 6: Tuning of activation thresholds and the role of rate of change of antigen in a quantitative framework for immune activation ...... 237 6.1. Introduction ...... 237 6.1.1. The discrimination problem for T cell activation ...... 237 6.1.2. The Tunable activation threshold (TAT) model ...... 239 6.2. Chapter Summary ...... 249 6.3. Results ...... 249 6.3.1. Defining a TCR-Tg murine model system to investigate the role of antigen rate of change...... 249 6.3.2. Sequential exposure to low dose antigen does not elicit T cell activation .... 254 6.3.3. Escalating antigen dose has an altered activation phenotype compared to bolus antigen injection...... 256 6.3.4. Osmotic pump delivery of escalating doses of antigen primes T cells for activation earlier than a single high antigen dose ...... 265 6.4. Discussion ...... 274

Chapter 7: Significant Findings and Future Directions ...... 276 7.1. The initial setting of the T cell activation threshold...... 276 7.1.1. Significant Findings ...... 276 7.1.2. Future Directions ...... 276 7.2. Origins of signaling heterogeneity in steady-state peripheral T cells ...... 276

vi

7.2.1. Significant Findings ...... 276 7.2.2. Future Directions ...... 278 7.3. Reversible activation threshold modulation from chronic antigen stimulation ... 278 7.3.1. Significant Findings ...... 278 7.3.2. Future Directions ...... 279 7.4. Tuning of activation thresholds and the role of rate of change of antigen in a quantitative framework for immune activation ...... 280 7.4.1. Significant Findings ...... 280 7.4.2. Future Directions ...... 281

References ...... 283

vii

List of Tables

Table 1.1. Biochemical properties that contribute to receptor activation thresholds...... 28

Table 2.1. Mouse model genetic designation and abbreviations...... 60

Table 2.2. Antibodies used for flow cytometric analysis...... 63

Table 4.1. Upregulated genes in CD5lo and CD5hi T cells...... 151

Table 4.2. Downregulated genes in CD5lo and CD5hi T cells...... 152

Table 4.3. Differential expression of NFκB pathway genes in CD5lo and CD5hi T cells.

...... 155

viii

List of Figures

Figure 1.1. Overview of the cellular components of the immune system...... 3

Figure 1.2. Ligand receptor interactions initiate complex intracellular signaling networks.

...... 7

Figure 1.3. Biochemical properties that contribute to receptor activation thresholds...... 15

Figure 1.4. Heterogeneity in signaling components can influence the threshold measurement...... 16

Figure 1.5. Positive and negative feedback circuits in signaling...... 23

Figure 1.6. T cell receptor signaling modules...... 32

Figure 1.7. Immune-receptor-tyrosine-based-activation motif sequence and structure within the T cell receptor...... 34

Figure 1.8. Proximal and intermediate modules of TCR signaling...... 37

Figure 1.9. Distal and nuclear components of TCR signaling...... 40

Figure 1.10. TCR signaling to the NFκB pathway...... 44

Figure 1.11. Models of TCR triggering...... 51

Figure 3.1. Overview of thymic development...... 79

Figure 3.2. Steps of selection in the thymus...... 81

Figure 3.3. Geography of thymic T cell development...... 85

Figure 3.4. Thymic development processes and peripheral activation poise the TCR repertoire...... 86

Figure 3.5. Signaling molecules expressed in the thymocytes augment the TCR signaling cascade ...... 88

Figure 3.6. CD5 gene and protein structure and potential interacting partners...... 92

ix

Figure 3.7. CD5 is developmentally regulated and proportional to TCR signaling strength in the thymus...... 96

Figure 3.8. Overview of the components of the NFκB pathway...... 98

Figure 3.9. Flow cytometric staging of thymic development in a C57BL/6 mouse...... 100

Figure 3.10. CD5 and IκBα are developmentally regulated in thymocytes...... 102

Figure 3.11. Loss of CD5 decreases IκBα expression in thymocytes...... 104

Figure 3.12. Construct design and characterization of CD5 inducible knock-out mouse model...... 106

Figure 3.13. Inducible loss of CD5 decreases IκBα in post-selection thymocytes...... 108

Figure 3.14. SHP-1 and THEMIS ablation does not impact CD5-IκBα axis in thymocytes.

...... 110

Figure 3.15. CD5hi thymocytes have a survival advantage over CD5lo cells...... 112

Figure 3.16. Survival advantage of CD5hi thymocytes is dependent on NFκB function. 114

Figure 3.17. Casein kinase II inhibition does not impair survival of CD5hi thymocytes.117

Figure 4.1. Peripheral T cell subset gating...... 128

Figure 4.2. Expression of TCR signaling molecules in peripheral T cells...... 130

Figure 4.3. Zap70, LAT, and PLCγ expression are not correlated with CD5 in peripheral

T cells...... 132

Figure 4.4. CD5 and IκBα expression are tightly correlated in peripheral T cell populations...... 134

Figure 4.5. IκBα expression does not correlate with other T cell surface molecules. .... 135

Figure 4.6. Genomic ablation of CD5 decreases expression of IκBα in peripheral T cells.

...... 137

x

Figure 4.7. Peripheral ablation of CD5 decreases expression of IκBα in T cells...... 139

Figure 4.8. The CD5-IκBα regulation is maintained independently of SHP-1 and

THEMIS...... 141

Figure 4.9. Casein kinase II activity is not required to maintain CD5-dependent IκBα expression...... 143

Figure 4.10. The CD5-IκBα axis is maintained post-transcriptionally...... 145

Figure 4.11. RNA-Seq analysis of CD5lo and CD5hi T cells...... 147

Figure 4.12. RNA-Seq analysis of CD5lo and CD5hi T cells...... 149

Figure 4.13. CD5-driven IκBα expression has a saturation level...... 157

Figure 4.14. Antigen exposure further upregulates CD5 and IκBα expression in peripheral

T cells...... 159

Figure 4.15. CD5 upregulation of IκBα is independent of TCR expression and requires lymphocytic-specific molecules...... 161

Figure 4.16. CD5 upregulation of IκBα requires the intracellular tail of CD5...... 163

Figure 4.17. CD5 expression alters the localization of NFκB in peripheral T cells...... 167

Figure 4.18. CD5 expression alters the localization of NFκB in peripheral T cells...... 171

Figure 5.1. Overview of central and peripheral T cell tolerance mechanisms...... 179

Figure 5.2. Two-signal model of T cell activation...... 185

Figure 5.3. IL-2 promoter requires multiple transcription factors for effective transcription...... 187

Figure 5.4. Outstanding questions related to T cell functional alterations following chronic stimulation...... 191

xi

Figure 5.5. T cell receptor signaling perturbations after chronic stimulation in a self- antigen murine model...... 200

Figure 5.6. Adoptive transfer model system for chronic self-antigen stimulation...... 202

Figure 5.7. Chronic antigen stimulation decreases proliferative capacity and cytokine production of T cells...... 204

Figure 5.8. Adoptive transfer model system for chronic self-antigen stimulation in lympho-replete hosts...... 206

Figure 5.9. Chronic antigen stimulation in replete hosts has loss of proliferative capacity as early as day five...... 207

Figure 5.10. In vivo deprivation of antigen restores some functional capacity of chronically stimulated T cells...... 210

Figure 5.11. In vitro antigen deprivation does not improve proliferative capacity of chronically stimulated T cells...... 213

Figure 5.12. In vitro stimulation with high dose antigen restores proliferative capacity of chronically stimulated T cells...... 215

Figure 5.13. High-throughput drug screening to identify tolerance reversing compounds.

...... 218

Figure 5.14. Pharmacological inhibitors do not improve proliferative capacity of chronically stimulated T cells...... 220

Figure 5.15. Bypassing TCR proximal signaling molecules improves cytokine production but not proliferative capacity in chronically stimulated T cells...... 224

Figure 5.16. Expression of total and phosphorylated Zap70 is altered in chronically stimulated T cells...... 226

xii

Figure 5.17. CD5 and IκBα are upregulated during chronic antigen stimulation...... 228

Figure 5.18. CD5 and IκBα are upregulated during chronic antigen stimulation...... 229

Figure 5.19. CD5 and IκBα are upregulated in tumor infiltrating lymphocytes...... 231

Figure 5.20. Transcription factors associated with exhaustion and checkpoint molecules are upregulated after chronic stimulation ...... 233

Figure 6.1. Delayed negative feedback adjusts activation thresholds...... 242

Figure 6.2. Chronic antigen stimulation elicits continual negative feedback to increase activation thresholds of T cells...... 244

Figure 6.3. Tunable Activation Threshold model kinetic representation...... 247

Figure 6.4. Murine model system to investigate antigen rate of change...... 251

Figure 6.5. Gating strategy for identifying adoptively transferred populations ...... 252

Figure 6.6. Multiple low dose antigen does not activate T cells...... 255

Figure 6.7. Escalating antigen dose over eight-day experimental design...... 257

Figure 6.8. Escalating dose over eight days activates T cells in vivo...... 258

Figure 6.9. Escalating dose over seventeen-day experimental design...... 260

Figure 6.10. Escalating dose over seventeen days activates T cells in vivo...... 261

Figure 6.11. Daily injections results in antigen decay quickly and has fluctuating rate of change of antigen ...... 264

Figure 6.12. Implantable mini osmotic pumps provide continuous antigen delivery so there is steady delivery of antigen ...... 266

Figure 6.13. Overnight culture assay to optimize antigen delivery with mini osmotic pumps...... 268

xiii

Figure 6.14. Adoptive transfer of naïve 5C.C7 T cells to optimize antigen delivery with mini osmotic pumps ...... 269

Figure 6.15. Escalating antigen dose delivered by osmotic pump activates T cells...... 271

Figure 6.16. Escalating antigen dose delivered by osmotic pump activates T cells ...... 272

Figure 7.1. Overview of thesis findings...... 282

xiv

List of Abbreviations

7AAD 7-aminoactinomycin D

AIB1 Amplified in breast cancer 1 protein

AP-1 Activator protein 1

APC Antigen presenting cell

AT Activation threshold

BCR B cell receptor

CABIT Cysteine-containing, all- in Themis

CaMKII Calcium/calmodulin dependent protein kinase type II

CARMA1 Caspase recruitment domain-containing membrane-associated

guanylate kinase protein 1

CBM CARMA1-Bcl10-MALT1 complex

CBP CREB-binding protein

Cdc2 Cell division control protein 2

CFSE Carboxyfluorescein succinimidyl ester

CH Challenge

CKII Casein kinase II

CLP Common lymphoid progenitor

CSK C-terminal Src kinase cTEC Cortical epithelial cells

CTL Cytotoxic T lymphocyte

CTLA-4 Cytotoxic T-lymphocyte-associated protein 4

CTV Cell trace violet

xv

CXCR C-X-C chemokine receptor

DAG Diacylglycerol

DAMP Danger associated molecular patterns

DAPI 4′,6-diamidino-2-phenylindole

DC Dendritic cell

DN Double negative thymocyte

DNA Deoxyribonucleic acid

DP Double positive thymocyte

DT Diptheria toxin

DTR Diptheria toxin receptor

DUSP Dual specificity phosphatase

EC50 Half maximal effective concentration

EGFR Epidermal growth factor receptor

EGR1 Early growth response protein 1

ELISA Enzyme-linked immunosorbent assay

ERK Extracellular signal-regulated kinase

FACS Fluorescence assisted cell sorting

FCS Fetal calf serum

FR4 Folate receptor 4

Gads GRB2-related adapter downstream of Shc gMFI Geometric mean fluorescence intensity

GPI Glucose-6-phosphate isomerase

Grb2 Growth factor receptor bound protein 2

xvi

GRIP-1 GR-interacting protein 1

HA Hemagglutinin

HEL Hen egg lysozyme

IFNγ Interferon γ

IKK Inhibitor of NFκB kinase

IL-10 Interleukin-10

IL-12 Interleukin 12

IL-15 Interlekin-15

IL-2 Interleukin 2

IL-6 Interleukin-6

IL-7 Interleukin-7

IP3 Inositol 1,4,5-triphosphate

IP3R Inositol 1,4,5-triphosphate receptor

ITAM Immunoreceptor tyrosine-based activation motif

ITIM Immunoreceptor tyrosine-based inhibitory motif

ITK Interleukin-2-incudible T cell kinase

IκBα Inhibitor of NFκB

JAK Janus kinase

JNK c-Jun N-terminal kinase

KO Knock-out

LAG-3 Lymphocyte-activation gene 3

LAT Linker for activation

Lck Lymphocyte specific protein tyrosine kinase

xvii

LCMV Lymphocytic choriomeningitis virus

LN Lymph node

LPS Lipopolysaccharide

MALT1 Mucosa associated lymphoid tissue lymphoma translocation

protein 2

MAP2K MAP kinase kinase

MAP3K MAP kinase kinase kinase

MAPK Mitogen activated protein kinase

MCC Moth cytochrome C

M-CSF Macrophage-colony-stimulating-factor

MFI Mean fluorescence intensity

MHC Major histocompatibility complex

MKP MAP kinase phosphatase

ML Multiple low dose mTEC Medullary epithelial cells

MV Multiple varied dose

NFAT Nuclear factor of activated T cells

NFκB Nuclear factor-κB

NIK NFκB-inducing kinase

NK cell Natural killer cell

NKT cell Natural killer T cell

OVA Ovalbumin

PAK2 P21 activated kinase 2

xviii

PAMP Pathogen associated molecular pattern

PARP Poly (ADP-ribose) polymerase

PBS Phosphate-buffered saline

PCC Pigeon cytochrome C

PD-1 Programmed cell death protein 1

PDK1 Phosphoinositide dependent kinase 1

PI3K Phosphoinositide 3-kinase

PIP2 Phosphatidylinositol 4,5-bisphosphate

PKC Protein kinase C

PLA2 Phospholipase A2

PLCγ Phospholipase Cγ

PMA Phorbol myristate acetate pMHC peptide-MHC

PRR Pattern recognition receptors

PTPN22 Protein tyrosine phosphatase, non-receptor type 22

RasGRP-1 Ras guanyl nucleotide releasing protein

RIPK1 Receptor-interacting serine/threonine-protein kinase 1

RNA Ribonucleic acid

SHP-1 Src homology region 2 domain-containing phosphatase-1

SLP76 SH2 domain-containing leukocyte protein of 76 kDa

SOS Son of sevenless

SP Single positive thymocyte

SP1 Specificity protein 1

xix

SRC-1 Steroid receptor cofactor 1

SRCR Scavenger receptor cysteine rich

STAT Signal transducer and activator of transcription

STIM Stromal interaction molecule

TAT Tunable activation threshold

T-bet T-box expressed in T cells

TCF-1 Transcription factor T cell factor 1

TCR T cell receptor

TCR-tg T cell receptor transgenic

Tespa1 Thymocyte Expressed, Positive Selection Associated 1

TF Transcription factor

Tfh T follicular helper cell

THEMIS Thymocyte-expressed-molecule involved in selection

TIF2 Transcriptional intermediary factor 2

TLR Toll-like receptors

TNF Tumor necrosis factor

Tox Thymocyte selection-associated high mobility group box factor

TRAIL TNF-related apoptosis-inducing ligand

Treg T regulatory cell

TSP Thymus seeding progenitor

WT Wild-type

Zap70 Zeta chain associated protein kinase 70

xx

Chapter 1: The activation threshold of peripheral T cells

1.1. Overview of the immune system

1.1.1 Prelude

The immune system is our body’s defense against the world of threats that we encounter throughout our lives. There are two main arms of the immune system that protect us: innate and adaptive (Figure 1.1) [1]. The innate system is responsible for providing fast-acting protection, but it is less nuanced in that it can only discriminate between broad swaths of microbial categories [2]. If this ‘hammer’ approach to protective immunity does not successfully clear out the threatening microbe or tumor, then the precise, scalpel-like adaptive system is activated. The response of most cells in the adaptive immune system is therefore quite delayed but, it is highly specific to match the particular threat to our body

[3]. The functionality of both systems is absolutely critical; where the fast-acting innate system is the first line of defense against threats as they arrive and can potentially destroy them without any assistance from the adaptive response. Conversely, cells of the innate system are also important as they must communicate with the adaptive system to recruit them to the fight. Upon activation, the cells of the adaptive system are deployed in two ways. B cells help make specific proteins called antibodies which can bind the surface of pathogens and mark them for destruction or removal. T cells on the other hand are more versatile; some of them traffic to the site of the threat and ultimately clear it while others are involved in making sure every aspect of the immune response is coordinated and controlled. Together, the army of cells effectively protect the body from any insults over our lifetime. While overtly simplified in the above description, these processes are critical pillars of a successful immune response. This thesis focuses on the operation of T cells

1 during these responses. We will discuss how individual T cells differ in their ability to sense and respond to a threat outside of the genetically encoded differences imbued through their unique T cell receptors. This is controlled by the internal ‘wiring’ of each T cell – whose components and circuitry are still incompletely understood. Being able to manipulate the flow of information within these circuits is the key to future development of therapeutics that can effectively harness T cells during vaccination, tumor immunotherapy, as well as altering responses during autoimmunity and transplant rejection. Indeed, the value of unravelling basic mechanisms of T cell operational logic is that the eventual practical applications to real world problems are quite broad and far reaching.

2

Figure 1.1. Overview of the cellular components of the immune system.

Our immune system is comprised of two main arms; the innate and adaptive systems. The innate system provides fast, relatively non-specific defenses from a variety of phagocytic and cytotoxic cells. The adaptive system conversely, is a slower responding system but has extremely specific targeting mechanisms through either humoral or cellular responses. The humoral adaptive response is coordinated through antibodies secreted from B cells. The cellular adaptive response is coordinated through T cells which come in a variety of flavors. There are two unique T cells, γδ T cells and natural killer T (NKT) that have overlapping functions and kinetics with the adaptive system. αβ T cells are the conventional T cells and they are further subdivided into CD4 and CD8 cells which have helper and cytotoxic functions, respectively. Figures adapted from [© 2020 Oxford Immunotec, http://www.oxfordimmunotec.com/north-america/science/technology/]

3

1.1.2. The internal circuitry of a cell

As we dive deeper into the mechanics of these two systems, it is imperative to appreciate that immune responses are managed by the coordination of actions from single cells. All cells of the immune system develop from precursors in the bone marrow and become specialized artisans of their craft (Figure 1.1). Innate cells such as neutrophils, macrophages, dendritic cells (DC), and natural killer (NK) cells detect any and all pathogen intrusion quickly to limit the damage to the host. This detection is accomplished by having receptors on their surface that bind the pathogen or pathogen product, much in the same way a key fits a lock; where the ‘key’ is the pathogen and the ‘lock’ is the cell surface receptor (Figure 1.2A). The fitting of this ‘lock-and-key’ at any time is an indication to the cell that its receptor protein has now found something outside the cell that is potentially threatening to the body. The cell now goes through a process of acting upon this information. In general, the actions that a cell takes first involves changes inside a cell

(referred to as ‘activation’), and then converts to an altered behavior (referred to as

‘response’) of a cell. The steps involved in activation require the cell to identify what exactly the receptor is engaging with. Understanding the type, duration, persistence, and damage caused by the threat are all integral pieces to identifying and reacting appropriately upon ligand binding. Addressing these questions involves a series of ‘signals’ that must be transmitted from the surface receptor into the cell – which then ‘interprets’ these signals to act upon it. The bulk of this thesis focuses on questions related to how these signals are transmitted (albeit in the context of the T cell) and how they are interpreted by the cell.

Much in the same way a computer chip uses flow of currents through its circuits to calculate and respond, the internal circuitry of the cell depends on biomolecules. Instead of currents,

4 information is transferred in a sequential fashion, by one biomolecule making a change to the next one (similar to flipping a switch), which turns on the next and so on in relay. These biomolecules are called enzymes – and in the context of receptor signaling, the ones that turn on relays of signals are usually kinases. The switch is flipped in these cases through the addition of a chemical (phosphate) group – by a process called phosphorylation. Of course, this would mean that there is an ongoing signal in the circuit when the kinase is active (a ‘positive’ signal). In turn, there are also resistors (or things that stop the flow of information down a circuit). Some of these are enzymes called phosphatases, which remove the chemical that the kinase added – to effectively terminate the signal. Other modifiers of the circuit simply cut away (degrade) pieces of the circuit to prevent flow of information. These enzymes include proteases and ubiquitin ligases that either directly degrade proteins or add other chemical modifications that target proteins for degradation.

Taken together, the action of these resistors prevent the ongoing flow of information and are hence often referred to as ‘negative’ signals. So, it is easy to appreciate how these combinations of parts can be used to build a complex computational device in every single cell (Figure 1.2B, Creative Commons Attribution 4.0). Over evolutionary time, every intricate part of this circuitry has been optimized to perform its biological function at a high efficiency. Such evolutionary optimization has potentially involved patchworks and modifications that we are only now beginning to unravel. And unlike a simple electrical circuit, in many cases, we are unsure of what the parts are ‘supposed’ to do – and often they are capable of performing more than a single function. This analogy of a cellular signaling machine therefore has limits; and has been previously discussed as a thought

5 exercise, where the ways in which a biologist approaches a complex, multi-component system does not necessarily provide adequate contribution to each of the pieces [4].

6

Figure 1.2. Ligand receptor interactions initiate complex intracellular signaling networks.

Pathogen associated molecular patterns (PAMPs) are conserved binding elements that pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs) have evolved specificity for. (A) The binding of the pathogen to the PRR transduces an intracellular signal that is then integrated and generates a biological response appropriate to the initial signal from the pathogen to PRR binding. (B) If we zoom into the signal transduction between the surface receptor to the activation of transcription factors that carry out the generation of a biological response, we find that there is a complex network of signaling molecules that function in coordination to carry the information from the plasma membrane to the nucleus of the cell. Figures adapted from [5, 6].

7

1.1.3. Decision making in the immune system starts with a single cell

As described above, the two arms of the immune system are coordinated by a network of highly specialized cells (Figure 1.1). Once they have differentiated, they populate the various tissues and lymphatic organs to patrol the body for any threats as discussed above. The individual niches that each of these cell subsets occupy provides distinct environmental cues, sensed and integrated by different receptors on these cells, to poise them for whatever they may encounter. These cells interact with their environment through ligand-receptor interactions that transmit information into the cell to modify its behavior. These environmental cues can include nutrients, metabolites, or molecules such as homeostatic cytokines that maintain the proliferation and turnover of cells [7]. Examples of these in T cells include glucose signals through GLUT1 [8] or trophic signals through the common gamma chain cytokines [9.]. Even other T cells, such as T regulatory cells

(Tregs), can provide cues that modify the milieu and drive localized functional changes

[10]. The combination of such signals activate and differentiate T cells to combat the threats that we face – for instance by modulating the biasing of a response towards different

T helper subsets by cytokines from DCs [11, 12]. When distilled to the core principle, these complex biological responses are driven by individual ligand-receptor interactions specific to the environmental niche. Multiple such interactions are integrated to affect an eventual cell fate decision. This means that the same cell type, occupying different niches, may be prone to make different responses when triggered; whether it be an innate cell directly responding to a pathogen or an adaptive cell responding to presented antigen [13]. Cell- type-specific responses are superimposed on the maintenance and homeostatic signals that each cell is already processing from the environmental milieu. Secondly, each of the

8 independently-responding immune cells must often also act in a coordinated fashion to generate an effective immune response. This requires the intercellular communication between different cells – whether it be from one arm of the immune system (the innate system) to the next (the adaptive system) or even within cells of the same arm. Unravelling the principles by which signals from separate ligand-receptor interactions in each T cell are superimposed and then interpreted by the cell to elicit a response that is appropriate for the antigen, the milieu and the needs of other collaborating cells, is challenging. Indeed this is an extensive area of research within ‘systems immunology’ that was overviewed by multiple expert articles in several special issues, including a 2009 Nature Immunology

Focus Issue [14]. These reviews and articles discuss moving to a network level of thinking with signaling in immune cells given the complexities and overlap that they have [14].

Therefore, rather than thinking about TCR signaling in purely biochemical terms, as a series of enzymatic reactions, it is important to think about the underlying logic [15]. This will help relate the details within intricate signaling interactions to the bigger picture – the goal of generating and maintaining a coordinated, multi-cellular, and protective immune response against a wide variety of threats without damaging autoimmunity, all from the predictable actions of individual cells.

1.2. Receptor signaling: Triggering, propagation, and interpretation

As discussed in basic terms above, intracellular signaling functions as a computing machinery to turn environmental stimuli into biological responses. This cascade of events is typically initiated when a receptor capable of receiving the environmental cue as an

‘input’, perceives a critical amount or duration of this to then drive the propagation of downstream biochemical intermediates that progressively amplify the initial input to

9 produce some form of ‘output.’ The ‘output’ can range from metabolic, structural, or gene expression changes that affect cell function and fate. Understanding the precise rules by which these signaling systems interact and lead to cellular decision making is key to developing a complete and therapeutically manipulatable understanding of biological systems, including the working of T cells.

Many principles linking cell signaling to fate decisions in immune cells, are shared with other cell types where they have also been extensively studied [16-18]. Considering these general concepts in the operation of mammalian signaling networks is therefore helpful, before layering on precepts unique to the TCR. Indeed, formulating a general and logical framework that aids our understanding of how parameters in signaling translate to functional consequences within signaling networks (along the lines of the central dogma or the histone code for molecular biology) is a lodestar that is far in the future. We can still draw on biochemical properties of protein signaling to unravel individual components to begin to formulate a framework for how these networks drive cellular decisions. I will first briefly lay out our current understanding of this framework in general signaling parlance and then extend it to the context of the regulation of activation and differentiation programs driven by the TCR itself.

When considering the process of cellular decision-making driven by most cellular receptors, we can interrogate the underlying signaling logic using two broad categories of questions: first, how does a receptor ‘know’ to transmit the signals it is receiving from its ligand? Secondly, how are the downstream molecular networks engaged to turn receptor

‘outputs’ into biological responses? The initial structural ‘lock-and-key’ receptor-ligand interaction for most receptors can be understood by pure biochemical properties – although

10 the T cell receptor (as discussed later) may be different in this regard. These two main questions are linked, but operate at different timescales and are subject to regulation by different parameters. The first is typically discussed as the problem of receptor triggering and the second as that of signal integration. Triggering initiates the flow of a biochemical flux down the circuitry of molecules associated with a receptor, while integration computes multiple parameters leading to functional alterations or gene expression. While these are intertwined in many ways, considering them separately allows us to appreciate the unique regulatory constraints that cells apply to each phase as well as the contributions of each to the eventual functional output of the cell. In many cases, including in the context of T cell activation, both are considered together under the umbrella of an “activation threshold”.

As we will discuss further, the modulation of activation thresholds during different stages of a T cell’s life is a central part of its biology. A better understanding of the underlying circuitry and how it relates to T cell decision-making at each step is therefore critical.

1.3. The concept of a receptor activation threshold

Defining ‘activation threshold’ for any cellular system is not trivial – especially since it is often defined by the downstream consequence (‘output’ or response) which may be immediate or far removed. The quantitative metric typically used to define an activation threshold is the amount of ligand required to elicit a response from the cell. The higher the amount of ligand required, the less sensitive the cell is considered to be for responding to it and therefore assigned a higher activation threshold. Thus, the activation threshold is usually implied to be a barrier to activation – a collection of biochemical ‘resistors’ that limit the receptor from powering downstream signaling networks which has to be overcome for successful responses. As we learn more about the specific role of TCR

11 signaling and the contributions of each individual branch downstream of the receptor, it is helpful to paint a clearer picture of this critical parameter to better understand the biological consequence of this signaling.

The definition of activation threshold is perhaps easiest to define when considering a simplified receptor signaling network like those exemplified by steroid receptors. Here, the simplified ‘input’ to ‘output’ process allows a clearer discussion of the role of thresholding in the interim. Since steroid hormones can diffuse through the membrane, they bind their receptor directly in the cytoplasm. The binding at an allosteric site causes a conformational change that releases it from chaperones that hold the receptor in an inactive state [19]. This complex is then translocated to the nucleus where it dimerizes with another ligand bound receptor to form an active transcription factor that can influence gene expression (Figure 1.3A). Unlike surface receptors, multiple other signaling intermediates are not typically required for the ligand-bound receptor to elicit an output – i.e. exert cellular transcriptional activity to alter gene expression. Therefore, in its simplest form the response is determined by few parameters: the structural properties of the receptor to ligand, translocation to the nucleus, and binding of cognate DNA elements. This results in a traditional sigmoidal response curve as shown in Figure 1.3A, where the response from signaling is proportional to the input of the ligand for the receptor. The shape of the curve here is influenced by biophysical properties linked to the receptor. For instance, a lower affinity ligand-receptor interaction would shift the curve to the right (grey line Figure 1.3A, right panel) and vice versa (dotted grey line Figure 1.3A, right panel). Both of these effectively change the EC50 (see red line in Figure 1.3A right panel), with a higher EC50 indicating an output that is less sensitive to the ligand but, this measurement is essentially

12 at a population level. At the receptor level, we would still expect that the changes for a single molecule to trigger gene expression also relates to this EC50. This is supported by the whole blue curve in Figure 1.3A, right panel, shifting either to the right (grey line) or left (grey dotted line). Therefore, whether you consider the initial phase (e.g. the inflection point at which the response “triggers” or comes off the baseline) or the maximal output

(i.e. the top of the curve representing the saturation response the system is able to make) is unlikely to influence the quantitative interpretation of the response profile. And simplistically, whether we collapse this to a single molecule’s behavior or a single cell’s behavior the derived values for the activation threshold in this system are likely to remain relevant. But, an important point to make here is that each segment of this curve (e.g. trigger, EC50, maxima) can also differ due to two factors - cellular heterogeneity and time.

For instance, if the population of cells to which the steroid is added differ in amounts of steroid receptor expression, then you will expect the differences in responses among individual cells in a population (Figure 1.4). Depending on when you measure the response, not all cells would have responded here – and the greater the heterogeneity, the bigger the impact on the dose response curve. Interestingly, when you plot the cumulative response

(over a period of time from the entirety of the population) from the population (Figure 1.4), you could now have an interesting disconnect between the three parts of the curve. The trigger points can potentially remain the same (since a few cells responding would mark the curve leaving the basal levels), and given time perhaps the maxima as well (since that reflects the eventual maximum possible response). But the dose response can potentially shift around. Additionally, in this situation there is also the possibility that cells that do not respond in the same manner. For example, if there is a population of cells in which cell A

13 makes response X and cell B makes response Y then the average response (summation of all responses at a particular time) will not accurate reflect what each cell was doing. One molecular basis of such heterogeneity in responses among T cells is the focus of Chapters

3 and 4 of this thesis.

14

Figure 1.3. Biochemical properties that contribute to receptor activation thresholds.

(A) The affinity of binding of a ligand to a receptor causes displacement of an inhibitory factor and proceeds to propagate signals. The shifts in curves are reflective of an increase (grey dotted line) or a decrease (solid grey line) of the ligand-receptor interaction. The trigger point is defined as the concentration where the response is generated above baseline. EC50 is the concentration at half maximal response. Maxima is the maximal response generated despite any increase in ligand concentration. (B) Representative response demonstrating the influence of co-factor requirements for response generation where the ligand-receptor interaction alone is shown in blue and the addition of co-factors is shown in the grey dotted line. The addition of cofactors does not significantly alter the trigger or EC50 of this response but rather increases the maxima. (C) Representative response demonstrating a cooperative system where the binding of one ligand-receptor interaction increases the likelihood of a secondary ligand-receptor interaction and facilitates a switch like function from that. (D) Representative response demonstrating the change of feedback that can shift the trigger and EC50 of a response curve. Image generated using Biorender.

15

Figure 1.4. Heterogeneity in signaling components can influence the threshold measurement.

The response of a population versus a single cell can convolve the response curves. (A) Synchronous activation at the population level would have all cells activating at the same time in response to a stimulus to generate a curve as designated in panel D (blue line, A). This though is an unlikely occurrence since there is variability in cellular expression of receptor and the interaction that it has with the ligand within the system. (B) Asynchronous activation takes into account the heterogeneity among cells and the time during which it would take these to activate based on the differences in expression of receptor and the interaction with the ligand as shown in panel D (red line, B). (C) Incomplete activation at a population level is a failure of some of the cells to activate despite any amount of time with that concentration of ligand as shown in panel D (black line, C).

16

A further wrinkle, even in this simplistic paradigm, is the requirement of additional co-factors to affect transcription. In the case of steroid receptors such as the estrogen receptor, the required co-factors include SRC-1 [20, 21], TIF2 [22], GRIP-1 [23], and AIB1

[24] which are p160 family members of steroid receptor coactivators. Additionally, the estrogen receptor also recruits histone acetylases such as CBP and p300 [25, 26]. This assembly of proteins is essential to open the chromatin as well as efficiently recruit the

RNA polymerase complex to initiate gene expression [27]. The limiting factor for eliciting a cellular response in this case is still the accumulation of transcription factor activity, but now it considers not just the activated steroid receptor complex but also the required cofactors. This modifies our simple concept of an activation threshold. The likelihood of activating a response by the receptor relies on two more parameters beyond the ligand availability itself – the expression levels of the steroid receptor and the cofactors necessary to initiate transcription. In this example, it has been shown that overexpression of the estrogen receptor or one of the cofactor proteins (SRC-1) increases transcription of estrogen responsive genes resulting in increased cell survival [28, 29]. Moreover, there is an ordered recruitment of the receptor and cofactors to produce gene expression changes; where the estrogen receptor binds first, then histone modifying cofactors, and finally additional cofactors for assistance in recruitment of the RNA polymerase complex to initial gene transcription [27]. This suggests that outside of the receptor itself, the cofactors and the way in which they are recruited to the DNA are absolutely critical for efficient transcription. As such, heterogeneity in responses to the ligand in this system, can result from the regulation of the required cofactors. As described above, the p160 cofactors are essential for estrogen receptor transcription since they are responsible for aiding in

17 recruitment of the RNA polymerase complex. This suggests that the regulation of these cofactor molecules would be closely tied with the biological outcome of the ligand-receptor interaction. As such, if these molecules are unable to be recruited (or not recruited) efficiently, the ability for transcription to proceed efficiently would significantly decrease.

This has been demonstrated using mutant studies of these p160 cofactors where the mutants are unable to bind the activation estrogen receptor transcription factor and there is diminished gene transcription in the absence of a functional p160 cofactor [27]. While there is diminished activity the reporter is not completely abrogated so the role of these cofactors is instead to alter the maximal gene transcription from an activated ligand-receptor complex rather than altering the activation threshold of the ligand-receptor interaction

(Figure 1.3B, right panel).

As with any biological system, this simple input-output paradigm has additional layers of control (Figure 1.3C and 1.3D). One such layer of control is seen through cooperativity between signaling intermediates. In this case, recruitment of an initial binding partner happens relatively slowly but once the first partners are assembled, the ability to associate with another partner molecule is greatly accelerated. A classic example of cooperativity is the protein hemoglobin, where the first molecule of oxygen to bind one subunit increases the likelihood of another to bind to the additional subunits resulting in a binding curve that is sigmoidal rather than linear (Figure 1.3C). Therefore, this change in interaction between the protein network can change how the response curve looks.

Cooperative interactions in signaling networks at an extreme can lead to a “switch-like” function where there is no evidence of the response being triggered for a while, until the response suddenly transitions to its maximum possible output. These are typically also

18 called “digital” responses, where there are only two states – on or off. A similar framework can be applied to transcription factor complexes where the binding of the transcription factor and its respective cofactors increases the likelihood of the RNA polymerase complex binding and therefore the gene transcription. To further use the example of estrogen receptors, it has been found that ligand-bound estrogen receptor dimers bind cooperatively when there are multiple estrogen-responsive elements within the promoter region, albeit dependent upon spacing and sequence, but this in part explains the synergistic effects seen in vivo [30]. Additionally, feedback mechanisms add an additional layer of control where the output can influence the original input signal [31]. We are most familiar with both positive and negative-feedback mechanisms in cellular signaling systems where a combination of these mechanisms are used to generate a variety of output amplitudes and duration. Again using the estrogen receptor as an example, it has been recently appreciated that CSK and PAK2 provide negative feedback on estrogen-regulated gene expression and have begun to explore modulating pathways for treatment of breast cancer [32]. The feedback ultimately shifts the response curve to the left or right from the initial parameter depending on whether the feedback is positive or negative, respectively (Figure 1.3D, right panel grey lines). Considering these factors together, a picture already emerges of the threshold for triggering responses as a malleable parameter, uniquely responsive to the lineage, experience and state of each cell which would modify the protein networks available to participate in signaling.

The reason we used a simple ‘input - output’ module to discuss these core concepts is because the measurement of output is relatively straightforward here. Concentrations of ligand required to elicit the biological response (initiation of transcription in this simplified

19 system) is considered a direct measurement of the signaling thresholds that need to be overcome in order to induce a response. As we consider signaling pathways with additional steps, the definition of a ‘response’ itself is more layered. Overall, we are still interested in understanding how the ligand for a receptor elicits function – typically in terms of gene expression or (in more complex terms) cellular proliferation, migration, and differentiation to name a few. Yet, at each step of the signaling cascade, there are potentially independently operating input-output modules. So, there are other ways in which to measure output including phosphorylation of signaling intermediates, localization of signaling complexes, alteration of metabolism, etc. but, measuring these other parameters within the signaling cascade are not necessarily a reliable predictor for transcriptional activity since the generation of a phosphorylated intermediate does not guarantee transcription. Similarly, for a specific cellular function like proliferation, measurement of phosphorylated intermediates does not guarantee an accurate readout of these complex functions. Importantly, for each input-output module in the cascade, the amount of ligand elicited may vary. For instance, phosphorylation of ERK by EGFR happens at a concentration of 1ng/mL [33] of ligand, but expression of gene EGR1 coding for Egr-1 requires 50ng/mL concentration of ligand [34]. In these models then the threshold for eliciting ERK phosphorylation is significantly different from the threshold for eliciting gene transcription. Equally important, if the signaling results in ERK phosphorylation, but no subsequent gene expression for a specific reporter, it is also not clear if recruited nuclear transcription factors were able to affect expression of other genes. An additional example is provided by responses to IL-12 in T cells. IL-12 is a critical cytokine that is responsible for skewing T cells towards particular effector fates. This is achieved through JAK-STAT

20 signaling and activation of gene expression programs that synthesizes the cytokine IFNγ.

In human T cells, the stimulation with 50ng/mL IL-12 is sufficient to increase IFNγ production by two-fold [35] whereas the phosphorylation of pSTAT4, a signaling molecule downstream of the IL-12 receptor, can occur at concentrations as low as 2ng/mL [36]. This highlights the discrepancies between measuring the signaling intermediates and the biological outcome of the ligand-receptor interactions. These biochemical principles then afford contextual weight to the quantitation of activation threshold. Aspects of cooperativity in the signaling network can affect the shape of the dose response curve at each step of the cascade. If a very high ligand concentration is required to trigger the initial phosphorylation event, but the subsequent biochemical steps are very cooperative, this would yield a curve as in Figure 1.3C, essentially mimicking a switch-like behavior.

Linear signaling pathways also generate feedback – mechanisms that alter the way in which the initial input is translated into an output [31] (Figure 1.5). There is both positive and negative feedback within signaling pathways that can either enhance or dampen the signaling capacity, respectively. Positive feedback operates with the help of cooperative interactions to feed the output signal back to the input amplifying the response of the original input signal (Figure 1.5A). This type of signaling mechanism is seen with the inositol 1,4,5-triphosphate receptor (IP3R) which is a calcium-gated ion channel in the endoplasmic reticulum. This receptor can be partially activated by the binding of four IP3 molecules which allows an initial release of calcium from the endoplasmic reticulum. This calcium release then fully activates the IP3R causing further calcium release completing a positive feedback loop. This loop provides an amplification of the initial calcium release through the released calcium feeding back on the IP3R [37]. This amplification is critical

21 because it takes the smaller initial output signal (calcium release) and ultimately increases the strength of that signal by feeding back to increase the activation of the initial receptor output. An additional example of a positive feedback loop is the ERK activation of phospholipase 2 (PLA2) which increases production of arachidonic acid to activate protein kinase C (PKC). PKC ultimately feeds back into ERK activation prolonging the signal potentiation [38]. Both of these examples do not require new protein synthesis and are fast acting ways to increase the impact of the initial receptor output signal.

22

Figure 1.5. Positive and negative feedback circuits in signaling.

Simplistic representations of positive and negative feedback loops. (A) Positive feedback is generated from the initial input to output relay and that signal then is integrated back into the input to further perpetuate additional output. (B) Negative feedback is generated from the initial input to output relay and that signal is then integrated back into the input to shut down additional output generation. Image generated using Biorender.

23

In contrast, negative feedback typically serves to turn off the incoming signaling flux by feeding negative signals back to the initial input mechanism (Figure 1.5B). For example, in the case of signal transduction involving kinases there are counterpart molecules that remove the activating phosphate group to terminate signaling. For example, one of the immediate early genes downstream of ERK1/2 activation are a class of phosphatases called DUSPs. These proteins are responsible for dephosphorylating the

ERK1/2 complexes which precludes full activation and terminates their activity [39]. This serves to terminate the signal upon gene expression upregulation and prevents the continued positive signaling if the ligand is still present after these products have been made. Beyond new gene products, there are also immediate examples of phosphatase- kinase regulation with the binding partners on the intracellular tail of CD28. CD28 signaling proceeds through a variety of adapter molecules include Grb2 and MAPK but, it has also been appreciated that among this assembly is a MAP kinase phosphatase, MKP6.

This phosphatase acts to attenuate the CD28 signaling in an immediate manner, in contrast to the previous example that required new protein synthesis [40]. Other forms of negative feedback involve the recruitment of degradative processes. Typically, these include ubiquitin ligases which are responsible for tagging proteins with a post-translational modification to target them to the proteasome. For example, osteoclast differentiation is a tightly controlled process that depends upon the activation of the transcription factor

NFATc1. It has been shown that stimulation through macrophage-colony-stimulating- factor (M-CSF) results in the degradation of NFATc1. This process is mediated by the E3 ubiquitin protein ligase, Cbl, and is critical for downregulation of NFATc1 transcriptional activity [41]. These negative-feedback mechanisms all result in dampening the signals

24 from the initial input typically through post translational modifications to alter the activity of the ‘input to output’ computation. The generation of negative feedback and its consequence on T cell activation is the topic of Chapter 5 of this thesis.

Importantly, in most signaling pathways, negative feedback is generated after a certain delay. Some of the examples given here require the synthesis of new gene products to provide feedback which of course will take a longer period of time to actually feedback on the process compared with the other examples where the feedback molecules are already only have to be recruited or activated in some way rather than synthesized. These differences in timing are critical pieces when understanding how signaling proceeds; does it involve a delay in the recruitment of the feedback or are those molecules made and ready to function? Typically, when feedback is on a delay from the initial output signal in order to allow signals to propagate, this generates a property known as hysteresis. It is a common property among signaling and particularly with regards to feedback mechanisms. This property can significantly alter the way in which a signaling pathway progresses since it provides feedback on the initial input mechanism. As this relates to the TCR, this delay in response is an integral part of the Tunable Activation Threshold (TAT) model proposed by

Dr. William Paul and Dr. Zvi Grossman to explain the ways in which lymphocytes adapt their signaling machinery to adjust their sensitivity to antigen [42, 43]. This framework is the basis of concepts and experiments in Chapter 6.

In the case of signaling from cell surface receptors such as the TCR, the activation of transcription factors is typically the ‘last’ stage leading to gene expression. Often multiple signaling cascades converge to generate a biological response (such as initiation of transcription). The integration of multiple branches adds the final layers of complexity

25 when discussing receptor signaling since each aspect of the ‘input to output’ chain (Figure

1.3B – 1.3D) can be applied to each of the individual branches of the signal flow. This is typically discussed under the umbrella of emergent properties and how they alter the mechanics of the signaling systems as whole [38, 44]. One of these emergent properties is ultrasensitivity where a system of proteins functions as a switchover a narrow change in ligand concentration (Table 1.1). This arises in situations where protein systems work cooperatively, have multiple regulation steps, or by changes in inhibitor or activator concentrations. For example, one of the first demonstrations of this behavior was using

Xenopus oocyte extracts to demonstrate the cooperativity of the MAPK system [45]. It was also demonstrated with the GTPase, Rap in response to cannabinoid receptor-1 signals

[46]. These systems produce a switch-like behavior where the change in the system output is either at baseline or maxima over a narrow change in ligand which is reflected by the ultrasensitive naming. Bistability is another emergent property where positive-feedback loops (or two negative feedback loops) result in switch like behavior as well but in this situation, there are two stable states within the signaling system where above a set amount of ligand the system will achieve a relatively high response steady state of below that amount will be set at a baseline (Table 1.1). These systems use a form of adaptation to convert graded input signals into a singular output behavior which can then persist after the activating stimulus is removed [47]. This type of behavior has been observed in the

JNK cascade of Xenopus where upon stimulation JNK is either on or off. This then provides a positive feedback loop to increase JNK activation promoting factor causing other JNK molecules to be activated. A key piece of this system is that the removal of the initial stimulus does not immediately turn off the signaling due to the positive feedback loop

26 described [48]. Another example of an emergent property is the observation of oscillatory responses. This results from negative feedback loops that gate the maxima of a response as well as the duration of the response (Table 1.1). Interestingly, with this situation, an increase in ligand does not result in increased amplitude and duration but rather increases the number of oscillations of response that are generated. An example of this system is seen within the cell cycle where Cdc2 activates anaphase promoting complex which then feeds back to inactivate Cdc2. This system also has a positive feedback loop with Cdc2-

Wee1/Myt1 and Cdc2-Cdc25 [49]. Together these signaling mechanisms create a distinct pattern of output signal which has a controlled duration and maxima resulting in oscillations of outputs, compared to the other emergent properties. A final characteristic of emergent properties is the use of redundancy and robustness (Table 1.1) where multiple pathways will feed into the same biological response ultimately aiding to withstand perturbations within the system without losing the biological outcome, therefore increasing the robustness of the response from the pathway activation. One example of this is the downstream pathways of EGFRs which there is activation of PKC and Ras to both feed into the ERK pathway [17]. This provides another point of adaptation where there are multiple input signals to result in a singular output by ERK. Taken together, these properties demonstrate the powerful and complex regulatory mechanisms that occur upon the interaction of multiple signaling branches all of which contribute to the ability of the ligand-receptor to generate a biological response and therefore the activation threshold of that particular interaction.

27

Table 1.1. Biochemical properties that contribute to receptor activation thresholds.

Term Definition Citation(s) Cooperativity Protein subunits that have similar binding sites for their ligand [50] and upon binding of one site, there is an increased binding potential for all additional sites. Similar behavior can occur between different proteins as well. Feedback The generation of a signaling output that then influences the [31] initial input signal either positively or negatively Digital signaling Switch-like outputs from ligand-receptor interactions [51] Analog signaling Graded outputs from graded input signals [52] Emergent properties Protein network level phenomena that are not observed within [44] single ligand-receptor systems. This includes: ultrasensitivity, bistability, and oscillatory behavior. Ultrasensitivity Protein systems that have switch-like responses over very narrow [45, 53] changes in ligand concentration typically through protein cooperativity or feedback loop interactions. Bistability This describes a protein system where there are two distinct [47] equilibrium states that the proteins can operate at. These typically arise from feedback loop interactions. Oscillations In certain systems, rather than increasing the duration or [44] amplitude of a response, there is an increase in the number discrete output signals. Redundancy Protein systems will typically have multiple players that can [54] perform similar signaling mechanisms such that if one is rendered inactive it does not shut down the entire system. Robustness The ability of a protein signal system to withstand perturbations [55] from other components. This is closely tied with redundancy because that is the mechanisms that aids to withstand perturbations.

28

Summarizing this section, I have reviewed some of the parameters that contribute to the response of a ligand-receptor interaction including the levels of the ligand, the affinity of the ligand-receptor interaction, cofactor requirements for the biological response, cooperativity and feedback loops within single pathways, and the interactions between pathways activated simultaneously by a single ligand-receptor interaction. From these, we can appreciate that the concept of an activation threshold takes into account the biochemical properties within a ligand-receptor signaling cascade including all of the downstream networks that translate those interactions into biological responses. The necessary partners that aid in producing a final response and the way in which those molecules are regulated, the interactions between those proteins (whether positive or negative mechanisms of action), as well as properties that exist through network interactions are all parameters that build into the framework of an activation threshold.

Therefore, it is perhaps too simplistic to think of activation threshold as a system of negative ‘resistors’ that are working against the positive signal transduction. Rather, this is a network of interactions where each individual piece has its own regulatory processes at work and combines positive and negative components. Each process must be considered individually when defining mechanisms relevant the activation threshold. Each piece of the network contributes to the ability of the transduction cascade to culminate into a biological response and that first response is the point at which we define the activation threshold. Breaking down the cellular components that are responsible for the decision to generate a complex biological response (such as those seen downstream of the TCR) is imperative to understand how there is heterogeneity leading to the variety of survival, activation, and differentiation states that are associated with TCR signals. This is a focus

29 of two Chapters in this thesis, where one of the components (NFκB) of the signaling network downstream of the TCR is calibrated by another receptor (CD5) by modifying the stability of a negative regulator of NFκB. Importantly, for many of the regulatory processes to act in cells and determine cellular fates and behaviors, no concomitant changes in gene expression may be evident. In other words, modern techniques including high resolution single-cell measurements of RNA or the epigenome, may not inform us about unique cellular states that are sustained by purely systems-level biochemical alterations in the cellular signaling network. We discussed these levels of regulation without relying heavily on the literature on TCR signaling, primarily because very little is currently known about the logical principles regulating the signaling networks forming the activation threshold of peripheral T cells. We will now review the basic elements of TCR signaling and how some aspects of systems behavior may be evident in these pathways.

1.4. Levels of the TCR signaling cascade and its essential components

The T cell receptor (TCR) is composed of multiple subunits where the extracellular elements which recognize the ligand and the intracellular components which transduce signals are separated out (Figure 1.6A, Copyright Clearance Center Marketplace

#1053561). The former is contained in the  and  chains, generated uniquely in each T cell precursor as a result of VDJ recombination. The latter complex is made up of the CD3 molecules which engage downstream signals using unique motifs known as ITAMs (Figure

1.6A). The intricacies of signaling via the TCR has been heavily researched and extensively reviewed [56-63]. Here, we will discuss briefly each piece of this signaling pathway to emphasize the key components and regulatory nodes relevant to experiments in this thesis. We can divide the TCR signaling cascade into four distinct regions: proximal,

30 adapters and intermediates, distal, and nuclear (Figure 1.6B, Copyright Clearance Center

#4882631444332). Each section below will describe the main components and the idealized flow through each of these levels leading to the response from pMHC engagement.

31

Figure 1.6. T cell receptor signaling modules

(A) The T cell receptor complex is comprised of the T cell receptor itself (α chain and β chain) and multiple CD3 molecules (one γ, one δ, two ε, and two ζ chains). The CD3 complex is essential for signal transduction since there is no intrinsic signaling capabilities of the TCR. (B) Once the signal is transduced inside the cell from the TCR complex, the signaling cascade can be broken into distinct modules; proximal, intermediates, distal, and nuclear. The proximal module comprises the early kinases Lck and Zap70 that are responsible for transduction of the initial signals. These kinases then activate molecules in the intermediate module that also serve as scaffolds to diversify and amplify the signaling from the TCR. The intermediates branch into multiple distinct distal nodes including calcium, MAPK, and NFκB pathways. These distal components ultimately lead to activation of transcription factors that translocate to the nucleus to initiate gene programs to generate T cell responses as a part of the nuclear module. Figures adapted from [64, 65].

32

1.4.1. Proximal TCR signaling components

The TCR binds small peptides displayed in the context of MHC to transduce signals intracellularly and the resulting signaling cascade has been reviewed [56]. The TCR is a complex of the alpha and beta chains (in the case of conventional T cells) and the association with multiple CD3 chains: one gamma, one delta, two epsilon, and two zeta chains. Since the TCR itself does not have any capability to signal intracellularly alone, the receptors associate with the CD3 complex to enable intracellular signaling through their multiple ITAMs and has been reviewed [66]. ITAMs have a defined amino acid structure that separate the tyrosines that are phosphorylated to create docking sites for downstream signaling molecules (Figure 1.7A). A similar consensus sequence is present in immune- receptor tyrosine inhibitory receptor motifs (ITIM) that recruit inhibitory molecules to dampen signals rather than perpetuate them in other receptors [67]. The ITAMs of the TCR complex are distributed along the CD3 chains where CD3δ, CD3γ, and CD3ε each have 1

ITAM and the ζ chains each have 3 ITAMs (Figure 1.7B, Creative Commons Attribution

4.0). Evidence suggests that the CD3ε and ζ chains are crucial for effective TCR signaling but, the redundancy of ITAMs increases signaling from the complex [68-71]. Interestingly, there does seem to be a preferential binding of downstream molecules to particular ITAMs so, the redundancy may serve to recruit all necessary components in an optimal fashion as well as facilitate development of certain subsets of T cells like T follicular helper (Tfh) cells [72, 73]. Therefore, this particular motif is essential for the ability of the TCR to transduce signals downstream.

33

Figure 1.7. Immune-receptor-tyrosine-based-activation motif sequence and structure within the T cell receptor.

(A) Amino acid sequences of the ITAM on CD3ε, CD3δ, CD3γ, and ζ chains with the consensus ITAM sequence. Conserved residues are highlighted in blue. (B) TCR complex with the ITAM labeled in grey on each molecule. CD3ε, CD3δ, and CD3γ each contain one ITAM and the ζ chains contain three for a total of ten ITAMs per TCR complex. Figures adapted from [66, 73] © Cold Spring Harbor Laboratory Press (panel A).

34

The propagation of signaling downstream of ITAM phosphorylation has been reviewed [56]. Briefly, ITAMs on the zeta chains are phosphorylated by Lck upon TCR binding to pMHC [74]. Coreceptor recruitment to the TCR complex brings in Lck since it is constitutively associated with CD4 and CD8 molecules [75, 76]. Active Lck is present in resting T cells and it is these molecules that initially potentiate TCR signals but, as the signaling continues, additional Lck molecules become activated and assist in initiating signaling [77, 78]. Lck activity is regulated by the phosphatase CD45 and the kinase CSK.

ITAM phosphorylation on the zeta chains provide a docking site for another downstream kinase, Zap70, which then sets off a chain reaction of signaling to ultimately elicit downstream responses from the receptor interaction. The activation of Zap70 is two-fold; first the binding of Zap70 to the SH2 domains within the ITAMs disrupts the autoinhibited conformation and then Zap70 is further activated by phosphorylation via Lck. This proximal module directly translates the extracellular interaction of TCR:pMHC into an intracellular output that can then be fed into a network of signaling molecules towards the goal of T cell activation or differentiation (Figure 1.8, Copyright Clearance Center

Marketplace #1053576).

As discussed above, CD45 is another molecule that could be considered proximal but is not immediately downstream of the TCR itself and the structure and function of this molecule have been reviewed [79, 80]. CD45 is a type I transmembrane glycoprotein that has intracellular phosphatase activity. This intracellular tail interacts with Lck to dephosphorylate an inhibitory tyrosine (initially maintained by CSK) while also, dephosphorylating any targets of Lck such as the ζ chain ITAMs. An important note of

CD45 is that it is both necessary to fully activate Lck but also the segregation of this

35 molecule from the TCR signalosome is necessary to prevent this inhibitory molecule from coming in too close of contact with the phosphorylation cascade that propagates TCR signals (Figure 1.8). Yet, this molecule is absolutely critical for T cell function since the genetic ablation of this molecule prevents T cell development in the thymus [81] and it has recently been appreciated that it functions as a ‘gatekeeper’ to allow graded signaling outputs while maintaining discrimination between weak signaling through the balance of activity with CSK [82].

36

Figure 1.8. Proximal and intermediate modules of TCR signaling.

The TCR complex, upon engagement with pMHC, Lck phosphorylates the tyrosine residues within ITAMs on the zeta chains. This provides a docking site for Zap70 to be recruited which can then begin to phosphorylate its target. This is what comprises the proximal module where there are a few kinase molecules that interact with the pMHC engaged TCR complex. Upon Zap70 activation, it can phosphorylate targets such as LAT (a scaffold molecule) that coordinates the assembly and activation of many of the components of the intermediate module. This includes molecules such as PLCγ and Vav to name a few. Figure adapted from [83].

37

1.4.2. Adapters and intermediates

The activation of Zap70 is a tipping point in the TCR signaling cascade since it is responsible for activating many of the downstream adapters that diversify the signals from the proximal module and its function has been reviewed [84]. Zap70 activates LAT which serves as a scaffold for many other signaling intermediates to bind and localize to the TCR signaling complex. Moreover, the multiple phosphorylation sites on LAT serve as docking sites for downstream mediators such as PLCγ1, Grb2, and Gads. Grb2 and Gads serve as further adapters to recruit SLP76, a target of Zap70, and SOS which potentiates additional signaling pathways (Figure 1.8). Interestingly, SOS and Grb2 can span multiple LAT molecules to serve as a super-scaffold of adapter molecules creating a large platform for amplifying TCR signals [85]. Adapters are a critical component to TCR signaling where the loss of LAT precludes both signaling and development past the double positive stage

[86-89]. While there is redundancy within the system, these adapters each have a distinct role in recruiting individual signaling components such as MAPK, NFκB, and calcium signaling which all contribute to the branching of the TCR signal [90-92].

Taken together these adapters are critical for recruiting many downstream signaling mediators such as PI3K, AKT, ITK, and PKC which will continue to propagate the signals down individualized branches of signaling (Figure 1.9, Copyright Clearance Center

Marketplace #1053576) as reviewed [63]. PI3K activates PKCθ which ultimately feeds into the NFκB pathway. AKT is another target of PI3K which will activate mTOR signaling which is responsible for many alterations including changes in cellular metabolism.

Additionally, ITK phosphorylates PLCγ1which hydrolyzes PIP2 to make IP3 and DAG.

These second messengers, IP3 and DAG, are critical for activation of calcium channels

38 such as IP3R (IP3) and other proteins such as PKCθ and RasGRP-1 resulting in activation of downstream MAPK pathways. Ultimately, the activation of these molecules is what initiates the diversification of the TCR signaling into individual pathways and brings us to the distal components of the pathway.

39

Figure 1.9. Distal and nuclear components of TCR signaling.

After the assembly of intermediates on the scaffolding molecules, the signals begin to diversify into independent pathways including calcium signaling, MAPK, and NFκB. These pathways are maintained relatively independently from each other considering the cascades do not have overlapping players and the endpoint transcription factors are NFAT, AP-1, and NFκB respectively for each distal pathway. These transcription factors are the nuclear components of the TCR signaling pathway that ultimately are responsible for integrating the signals and coordinating the biological response that is generated. Figure adapted from [83].

40

1.4.3. Distal components

The main purpose of the above described adapters is to provide a jumping-off point for the individual branches to amplify the output signals from these adapters. Here we will discuss each of these downstream branches that ultimately lead to activation of a multitude of transcription factors that are essential for T cell activation processes and has been reviewed for the TCR signaling cascade [63].

MAPK

The role of MAPK pathways in T cell signaling has been extensively reviewed [93-

95]. Briefly, these pathways act as one of the main amplification processes of the TCR signaling. There are three main groups: ERK, p38, and JNK kinase cascades. These pathways are mediated by a serial phosphorylation of MAP3K, then MAP2K, and finally the MAPKs. These cascades are not unique to TCR and can be employed by many other receptors such as GPCRs and growth receptors. Specifically, the ERK pathway is activated by Ras via Raf MAP3K. p38 and JNK are activated by Rho GTPases such as Rac and

Cdc42. As mentioned above, the ERK pathway is engaged through the recruitment of adapter molecules SLP76 and Grb2 activating the SOS-Ras-MEK-ERK pathway. ERK is one of the most important MAPK pathways since it plays a distinct role in thymocyte development, activation, and differentiation [96, 97]. The p38 MAPK pathway seems to play a larger role after T cells have differentiated into particular T helper clones such as

Th1, which are critical for controlling viral infections and tumors [98]. Similar roles have been shown with JNK but it seems to be expressed at much lower levels than the other

MAPK groups [99].

41

Calcium signaling

Calcium signaling is one of the main read outs for productive TCR signaling and the role it plays in the cascade has been reviewed [100-102]. T cells have a variety of calcium channels that regulate the cytosolic concentrations of this ion since it is a critical determinant of the activation states of some of the downstream TCR signaling molecules.

Calcium signaling is initiated downstream of the TCR through PLCγ1 production of IP3 molecules that can then act on the intracellular calcium channels to release a large flux of ions. This then sets off a cascade to release additional calcium stores from STIM activation and release of other intracellular stores of calcium. This positive feedback loop further activates calmodulin and calcineurin ultimately leading to the activation of the transcription factor NFAT. Moreover, calcium signaling is one of the earliest outputs for TCR signaling since this operates on a short time scale compared to gene transcription [103].

NFκB pathway

A key pathway that regulates multiple arms of the T cell activation cascade is that of NFκB (Figure 1.10, Copyright Clearance Center #4882680015549). This is a ubiquitous pathway in many different cell types and has important relevance to cancer transformation and, as such, has been reviewed extensively [104-106]. TCR-specific activation of the canonical NFκB pathway is a critical component for effective activation and since this pathway is of specific relevance to Chapters 3 and 4, we discuss more detail here.

The NFκB pathway can be activated through two independent signals from the

TCR; activation of ITK as well as PI3K signaling from the costimulatory molecule CD28.

These pathways converge on the activation of PKCθ which in turn activates the CBM complex composed of CARMA1, Bcl10, and MALT1. This complex adds post-

42 translational modifications to the IKK complex that allows their kinase activity to be initiated. This in turn will phosphorylate the inhibitor of NFκB, IκBα, to target it for degradation. This degradation releases the NFκB subunits to translocate to the nucleus.

The activation downstream of the TCR proceeds through the LAT scaffolding as

SLP76 binds Vav. These two molecules complex with ITK which activate PLCγ1. This phospholipase generates two intermediates, one of which is DAG that activates PKCθ which then continues the cascade as outlined above.

CD28 co-stimulation proceeds to activate PI3K via PDK1. PI3K in turn activates

PKCθ and feeds into the same signaling pathway as the TCR from here forward. Although this is an example of the redundancy in activation of the NFκB pathway, it seems that the role of CD28 in this context is to aid in amplification of these signals rather than drive them alone. Without TCR activation of this pathway as well CD28 is not sufficient to recover full functionality of TCR signaling [107].

Figure 1.10 also shows the wide variety of molecules that can influence the signal transduction of the NFκB pathway. Particularly there is a multitude of kinases that can phosphorylate molecules such as PKCθ and CARMA1. This emphasizes the importance of understanding the signaling molecules in the cell and how those can impinge on signaling pathways that they are not typically associated with. For example, CaMKII which is typically associated with the calcium signaling pathways downstream of the TCR has been shown to phosphorylate CARMA1 in vitro but, the exact role of this interaction in vivo is not clear [108]. Yet, this provides a distinct example of the interplay between these distal components that are typically thought to be relatively independent of one another.

43

Figure 1.10. TCR signaling to the NFκB pathway.

Both the TCR complex and CD28 feed into the activation of the NFκB pathway. From the TCR, SLP76 is phosphorylated and activated by Zap70. This phosphorylation creates a binding site for Vav on SLP76 and, along with ITK, form a complex that activates PLCγ1. Activation of this phospholipase generates two products, IP3 and DAG; of which DAG continues the signaling cascade to ultimately activate PKCθ. This molecule is where the signaling from CD28 also converges but, the upstream activation mechanisms function through PI3K. Upon PKCθ activation, the assembly of CARMA1-Bcl10-MALT1 (CBM) will facilitate poly- ubiquitylation of IKKγ and phosphorylation of IKKβ which then facilitates the degradation of IκBα to release NFκB transcription factors for translocation to the nucleus. Figure adapted from [106].

44

1.4.4. Nuclear components and transcription factors

As discussed in section 1.4.3., the distal components of the TCR end with the activation of multiple transcription factors: NFAT, AP-1 (Fos/Jun heterodimers), and

NFκB. Here we will discuss the role that each of these play in TCR signaling and the gene programs that they help to initiate towards the goal of an effective T cell response.

NFAT

The utility of calcium signaling in a variety of cells culminates in the activation of

NFAT transcription factors as reviewed [109-111]. In T cells, there are a few isoforms of

NFAT: NFAT1 (NFATc2), NFAT2 (NFATc1), and NFAT4 (NFATc3). These transcription factors are aptly named since they were first characterized in a T cell system

[112, 113]. Since their initial identification as Nuclear Factor of Activated T cells, these transcription factors have emerged as important players in many different T cell states. As discussed in Chapter 5, NFAT plays a particularly important role in defining certain subtypes of T cell hyporesponsiveness. This was particularly important to underscore the importance of costimulatory signals since activation of NFAT in the absence of those signals tends to lead T cells down a path towards hyporesponsiveness [114].

These transcription factors bind to DNA regions such as the IL-2 promoter which is a critical cytokine for T cells particularly during their clonal expansion after activation.

Typically, these transcription factors work in concert with other factors activated downstream of the TCR particularly AP-1 [115]. NFAT has also been shown to be important in the upregulation of many cytokines and surface receptors such as CD25 after

TCR stimulation [110].

45

AP-1

AP-1 transcription factors are heterodimeric proteins that are composed of Fos, Jun,

Maf, and ATF partners [116, 117]. In T cells, these are typical heterodimers of Fos and Jun subunits. As mentioned above, the MAPK pathway activates these transcription factors to allow them to translocate to the nucleus. Given the wide breadth of MAPK, there are a variety of stimuli that can interact with these pathways. Additionally, since these pathways are used by a vast number of receptors, the signaling state of the MAPK pathway may be reflective of other environmental situations more so than just TCR signaling.

AP-1 plays a critical function in T cells where it aids to prevent a default anergic state induced by NFAT activation alone [118]. Moreover, the activation of these transcription factors promotes expression of cytokines but, given the redundancy in the system with the variety of heterodimeric partners, it is difficult to discern direct roles with single knock out systems.

NFκB

As alluded to above, Chapter 3 and 4 will focus on the role of NFκB in T cells so here we discuss the role of this transcription factor in effective T cell activation and responses as well as the role in maintaining homeostasis. NFκB plays a critical role in cell survival in T cells and many other cells in the body as reviewed [104, 119]. The gene programs that are initiated by these transcription factors provide insight into how those phenotypes are elicited.

NFκB has a wide variety of target genes including cytokines, growth factors, pro- survival molecules but importantly, it also upregulates expression of its own inhibitors such as IκBα [120]. While the functional consequences of the cytokines and growth factors are

46 easy to interpret for T cells, the upregulation of inhibitors provides a more interesting signaling regulation discussion. The upregulation of an inhibitor is direct negative feedback on the NFκB pathway but, this is upregulation at a transcriptional level. Therefore, the delay before actual inhibition of the transcription factor would allow for the productive transcription of other functional genes providing a uniquely timed regulation to allow signaling for a period of time while shutting down transcription before it gets out of control.

Interestingly, this transcription factor exhibits oscillatory localization between the cytoplasm and nucleus as discussed as an emergent property in section 1.2 [121]. This behavior is likely enforced by the upstream regulators as discussed in sections 1.4.2 and

1.4.3. The translocation of transcription factors is critical for their ability to regulate gene expression and as such understanding the localization patterns may provide insight into the function of NFκB.

1.5. TCR triggering and generating the first intracellular ‘output’

A key characteristic of the TCR that makes the study of nuances in TCR signaling even more relevant, relates to its extremely low affinity for ligand and lack of obvious conformational changes during its activation [122, 123]. As we discuss below, control of

TCR triggering itself relies on special signal-coupling features that are not well understood.

First, the TCR is extremely sensitive to its target ligand (agonist). It has been shown that a single peptide MHC complex is sufficient to trigger naïve T cell activation in the form of cytokine secretion [124]. This suggests that the activation threshold for a T cell is less than the positive signal transduced by a single peptide MHC complex. Since there is no way to have less than a single pMHC, the question become does a T cell really have an activation threshold? In this study, the authors utilized a quantum dot labeling system to

47 visualize the pMHC-TCR contacts. Interestingly, although there was a single pMHC recruited between the T cell and APC contact, the authors did not label the TCR; so multiple TCRs could be recruited to the single visualized pMHC. This suggests that rather than a 1-1 relationship between the pMHC-TCR, there may be a larger number of TCRs present, supported by evidence of ‘nanoclusters’ of TCRs rearranging upon stimulation

[125]. The role of multiple TCRs may explain how a single pMHC is able to lead to cytokine secretion.

To further elaborate on multiple TCRs recruited to a single pMHC, it is important to understand the structures that form upon TCR-pMHC engagement. A distinct molecular structure called an ‘immunological synapse’ forms during T cell-APC contacts as reviewed

[126]. Briefly, this structure includes a variety of receptors and ligands that are recruited into tight clusters to help perpetuate signaling from the TCR. Besides the TCR, coreceptors such as CD4 or CD8, integrins, and others are arranged into rings around the central TCR- pMHC contacts. Importantly, as mentioned above, multiple TCRs are likely present within these clusters [125]. This could provide insight into how a single pMHC was able to cause cytokine secretion from naïve T cells [124]; where the recruited pMHC was not engaging a single TCR but rather sequential TCRs over the period of time that the analysis was conducted (the authors were measuring cytokine secretion over a period of hours). This unique structure provides insight into how a T cell alters the spatial arrangement of receptors to better conduct signal transduction and may play a role in the activation threshold; by physically clustering multiple TCRs the positive signals are able to accumulate from multiple TCR-pMHC engagements increasing the transduction of signal into the cell.

48

Quantity of antigen has always thought to contribute to T cell activation whereby the more antigen that is around, the more T cell activation there is. This is seen in antigen dose response curves where the number of activated T cells increases with increasing dose until it hits a plateau that reflects the saturation of the system. Yet, considering that a single pMHC is able to activate a T cell, quantity is not necessarily the defining factor of activation. The quality of the antigenic signal is also important; meaning, the way in which a TCR experiences antigen presentation plays a role in how that signal is interpreted by the

T cell. For example, TCRs are low affinity receptors by comparison to other immune receptors (such as BCRs) but even within the low affinity, a single ligand is able to activate the cell. This may be due to the fact that when these TCRs are engaged, they are typically engaged within the clusters of the immunological synapse. In this case, the low affinity of the receptor does not have as much effect since they are able to engage and upon disengagement another TCR may be able to bind to the pMHC since they have fast on- rates with the pMHC [127]. Beyond the spatial arrangement of the structure between the T cell-APC contact, the duration of this interaction plays a role in T cell signaling as well.

Prolonged interaction of TCR with pMHC typically results in better activation compared to brief contacts. Both the affinity and duration of interaction are components of the quality of the antigen signal received and these become important concepts when considering dwell times of the receptors. Dwell time refers to the cell-to-cell contact made between the

APC and the T cell and is a crucial component of the kinetic segregation model of TCR triggering. This model suggests that the exclusion of molecules such as CD45 from the immunological synapse will increase the likelihood that the TCR will bind and signal productively from pMHC [128, 129]. Moreover, through the physical clustering of TCRs

49

(to increase contacts of low affinity receptors with their ligand) as well as prolonging those interactions (by increasing the dwell time) to allow the accumulation of signal transduction within a defined space would ultimately aid in downstream signaling perpetuation.

T cell activation models have taken into account the importance of the defined structures and duration of the TCR-pMHC interactions. For example, the kinetic segregation model describes the physical separation of the negative inhibitors like CD45 from the immunological synapse due to their size; where the close interaction of the TCR- pMHC shifts CD45 to the edge of the interaction area as discussed in section 1.4.1 [130].

This provides a spatial separation from the signaling molecules propagating the positive signals from TCR-pMHC and the negative inhibitors that could potentially block those signals. The topology of these receptors and their ligand interactions therefore physically segregate those that are positive and negative regulators of TCR signaling. Beyond this, multiple TCRs can be ‘trapped’ within these structures and continue to interact with the pMHC within the immune synapse. This further perpetuates positive signals all within a defined space which may amplify the overall signal from the pMHC within the synapse.

Taken together, these elements of antigen quality, membrane remodeling, and segregation of inhibitory molecules are all points to consider when building a framework for activation thresholds of T cells (Figure 1.11, Copyright Clearance Center Marketplace

#1053561). The way in which a TCR senses its antigen and the changes that it makes in order to prolong these low affinity interactions to accumulate signaling impacts the way in which a T cell ultimately responds to these signals.

50

Figure 1.11. Models of TCR triggering.

(A) Mechanical force is applied to the TCR upon binding to pMHC which induces a conformational change in the CD3ε complex allowing phosphorylation of the ITAMs by Lck. (B) TCR clustering induces a local environment conducive for productive signaling by clustering multiple TCRs. (C) Segregation of phosphatases from pMHC engaged TCR prevents the dampening of the propagated TCR signal flow typically called kinetic segregation. Figure adapted from [65].

51

1.6. Control of the T cell’s activation threshold

In sections 1.4 and 1.5, we discussed some of the major components of the TCR signaling machinery and how the initial signal transduction down these pathways are initiated. As we discussed in section 1.3, however, it is important to integrate how information flow through this machinery, logically drives the cellular fates and functions typically associated with T cells. Unlike the ‘simpler’ circuitry of a steroid receptor or even a more interwoven one of the EGFR [131], the TCR (as discussed in section 1.4) engages and modulates almost every signaling pathway available in the cell. The evolved logic integrating these are poorly understood. But there are already intriguing suggestions that, even before the first gene expression event is initiated, the signaling networks rewire the T cell and exert key influences on fate decisions. To emphasize this here, we relate how some of the control mechanisms discussed in section 1.3 (Table 1.1) apply to pathways of TCR signaling and to interpreting the data generated by the work described in this thesis.

1.6.1. Sensitivity

As discussed in section 1.5, the biochemical mechanisms responsible for triggering the TCR initially are not yet fully understood. Yet, it is clear that in spite of the relatively poor affinity of the receptor-ligand interaction involved, very few pMHC complexes can trigger full T cell activation, whether it be in terms of cytokine secretion or actual cell killing [124]. Thus, unlike many other receptor systems, a T cell’s ability to respond to antigen is dependent on its internal activation threshold, set by interplay of various molecules in its signaling network, rather than by purely biophysical metrics of the receptor-ligand interaction [132]. This is best illustrated by the mechanisms which determine how sensitive a T cell is to antigen (and the focus of Chapters 4 - 6). As a T cell

52 progresses through development, its activation threshold undergoes multiple changes, even though it carries the same TCR. Many of these changes, discussed in Chapter 4, are related to altered gene expression changes; but some occur without any evident transcriptional alterations. One example is the control of a T cell’s activation threshold by maintenance of contact with endogenous MHC. In one experiment, Stefanova et al demonstrated that T cells freshly isolated from mice rapidly showed an increased activation threshold (i.e. became less sensitive to antigen) just by ‘starving’ them of MHC contact for as little as 30 minutes [133]. These changes were not dependent on gene expression, but reflected an altered signaling complex, anchored by partial phosphorylation of the TCR-zeta chain (in a p21 form instead of the p23 form). This rapid readjustment of sensitivity shows that signaling networks in T cells are constantly in flux, adapting to changes beyond the appearance of cognate antigen alone. While there is yet to be a consensus on the impact of

MHC on a TCR’s functional sensitivity [134], the broader view is that a T cell almost functions like a cellular processing unit or sensory apparatus, with the TCR only providing some of the information required [15].

1.6.2. Discrimination

As discussed in Table 1.1, several properties of the TCR signaling network mimic the higher-level emergent behavior that relies on interactions between multiple pathways.

For instance, how T cells actually distinguish qualitatively different ligands (ligand discrimination) also does not seem to rely heavily on physical TCR-pMHC characteristics alone [132, 135, 136]. Rather, even ligands with similar biophysical constants can trigger different outcomes in T cells, which suggests that they use a different algorithm to discriminate between pMHC complexes. One such mechanism that has been uncovered is

53 the sequential phosphorylation of the ITAMs on the CD3 zeta chains that reveals a switch- like response of Zap70 binding to the phosphorylated ITAMs within mathematical models

[137]. Yet, when this was performed in an in vitro system, the same switch-like responses did not seem to be orchestrated by the sequential phosphorylation [138]. This suggests that there are emergent properties evident within the TCR signaling system that have yet to be completely understood.

The molecular interactions between the TCR signaling components driving properties like cooperativity are becoming more apparent [139]. Part of the problem of building these network of interactions from our current understanding of TCR signaling alone is that we are still identifying new mechanisms of the flow of signaling information downstream of the TCR. For example, it was recently appreciated that Lck functions to bridge the kinase activity of Zap70 to LAT [140], which was previously considered an independent role of Zap70. Understanding all of the potential interactions and the ways in which they influence the flow of information is imperative to further defining the activation threshold of the TCR.

While these vignettes suggest that deciphering biochemical paradigms similar to those discussed in section 1.3 are critical for developing a complete understanding of T cell activation, these data have been limiting. New techniques have recently allowed us to begin evaluating these principles in primary cells. Much of what we know about the building blocks of TCR signaling are derived from using in vitro and cell culture models. As we begin to build a picture of the complex network of interactions downstream of the TCR as it operates in vivo, it is important to keep track of functional significance of important biochemical nuances. While challenging, studying signaling using in vivo models is likely

54 the most promising strategy for dissecting the machinery that is directly relevant to dynamic immune responses.

1.7 Lay Summary of Thesis Findings

The ability of T cells to be activated efficiently against pathogens and malignancies is largely dependent on each T cell’s ability to sense their targets, using the TCR and the signaling machinery associated with it. The composition and status of this signaling machinery determines a T cell’s activation threshold. While this activation threshold is modified during the life of a T cell in order to control immune responses, the underlying molecular mechanisms are unclear. In this thesis, we showed that the activation threshold for a key TCR signaling pathway (NFκB) was set during T cell development by the activity of a surface receptor CD5. We discovered a novel link between CD5 and a known inhibitor of NFκB, IκBα. This meant that T cells with more CD5 also built up a higher intracellular reservoir of NFκB. Accordingly, thymocytes with higher CD5 had an NFκB-dependent cell survival advantage compared to cells with lower expression. We then examined how this signaling circuitry changes when T cells encounter different kinds of targets – one that is transient versus one that is always present (chronic) in the body. Since the latter is typically associated with the body’s own proteins, we find that T cells rapidly adjust various signaling molecules in order to stop the flow of information that could potentially lead to deleterious T cell activation. While this is desirable to prevent autoimmunity, it is a problem in combating tumors and chronic infections; where T cells enter a state that turns off their sensory circuits. Finally, to begin to understand the molecular logic which modifies the signaling circuits in T cells, we turned to the TAT model which suggests that

T cells sense the rates at which their targets appear. If a target increases quickly, T cells

55 equate that with a fast dividing pathogen or tumor, whereas if it increases slowly, it is interpreted to be part of the body. In order to test this logical framework, we used different approaches to change antigen levels at different rates. Interestingly, we empirically determined a rate at which even chronic antigen exposure allows T cells to continue responding rather than shut themselves down. Taken together, the findings in this thesis emphasize a continuum of activation threshold adjustment in T cells, initially driven in part by CD5 levels and subsequently by the rate of change of peripheral antigen, which determines a T cell’s ability to function.

56

Chapter 2: Materials and Methods Cell Culture

BW-5147 TCR- (α−β−) cells originally generated by White et al. [141] were a gift from Philippa Marrack, National Jewish Health, Denver, CO, to B. J. Fowlkes, National

Institute of Allergy and Infectious Diseases (NIAID), NIH, Bethesda, MD, and maintained as described by White et al. NIH 3T3 cells were obtained from ATCC. Cells were maintained in RPMI 1640 (Gibco) supplemented with 10% FCS (Gemini), 2% glutamine

(Gibco), 1% sodium pyruvate (Gibco), and 1% antibiotic/antimycotic (Gibco) referred to as completed RPMI (cRPMI). Phoenix-GP cells were obtained from ATCC and maintained in DMEM (Gibco) supplemented with 10% FCS (Gemini), 2% glutamine (Gibco), 1% sodium pyruvate (Gibco), and 1% antibiotic/antimycotic (Gibco) referred to as complete

DMEM (cDMEM). Primary murine thymocytes and T cells were maintained in RPMI-

1640 (Gibco) supplemented with 10% FCS (Gemini), 1% glutamine (Gibco), 1% antibiotic-antimycotic (Gibco), and 0.000014% β-mercaptoethanol referred to as T cell media (TCM). E.G7-OVA cells were obtained from ATCC and maintained in cRPMI.

Mice

Wild-type mice were obtained from both Jackson Laboratories (C57BL/6J and

B6.SJL-PtprcaPepcb/BoyJ) and Taconic Biosciences (B6NTac and B6.SJL-

Ptprca/BoyAiTac). B6 mice were bred for dual congenic expression using a dam or sire from opposing place of purchase to eliminate any strain drift between the two wild-type strains. Wild-type B10.Q/Ai mice were originally obtained from Ron Schwartz, Institute of Allergy and Infection Diseases, NIH, Bethesda, MD. Wild-type B10.A-H2a H2-

T18a/SgSnJ mice were obtained from Jackson laboratories. B10.A mice were bred for

57 dual-congenic expression using a dam or sire from opposing genotypes. 5C.C7 TCR transgenic mice were obtained from Taconic Biosciences (B10.A-Rag2tm1FwaH2-

T18aTg (Tcra5CC7,Tcrb5CC7)lwep). SMARTA TCR transgenic mice were obtained from

Jackson Laboratories (B6.Cg-Ptprca Pepcb Tg(TcrLCMV)1Aox/PpmJ). TCRαβ knockout mice were obtained from Jackson Laboratories (B6.129S2-Tcratm1Mom/J). THEMIS knockout animals were generated as described [142]. SHP-1 knockout (C57BL/6J Ptpn6me- v/me-v) animals were originally obtained from Samuel Lunenfeld Research Institute at the

Mount Sinai Hospital and subsequently obtained from Paul Love, National Institute of

Child Health and Human Development, NIH, Bethesda, MD for the experiments conducted here. CD5 knockout animals were generated as described [143] and obtained from Ron

Schwartz Institute of Allergy and Infection Diseases, NIH, Bethesda, MD. CD5 transgenic mice were generated as described [144] and obtained from Paul Love, National Institute of

Child Health and Human Development, NIH, Bethesda, MD. CD5 inducible knockout mice were generated by microinjection of constructs shown in Figure 3.12A into C57BL/6

ES cells and implantation into super ovulated females. CD5 inducible knockout mice were obtained from Paul Love, National Institute of Child Health and Human Development,

NIH, Bethesda, MD. CD3ε knockout mice were obtained from NIAID/Taconic mouse contract (B10.A-Cd45a(Ly5a)/NAi N5). PCC-CD3ε knockout mice were generated as described [145-147] and obtained from Ron Schwartz, Institute of Allergy and Infection

Diseases, NIH, Bethesda, MD. Briefly, these mice have expression of a membrane targeted form of pigeon cytochrome C (PCC) under the control of the MHC class I promoter and an Ig enhancer and carry a deletion of CD3ε gene such that they do not generate a functional

T cell repertoire. PCC+ mice were generated by crossing PCC-CD3ε knock out mice with

58

B10.A wild-type mouse to generate heterozygotes that carry the PCC transgene and generate a full compartment of T cells.

All animals were bred and maintained in modified specific-pathogen free (SPF) facilities at the University of Maryland Baltimore or at the National Cancer Institute at the

National Institute of Health. Experiments were performed with animals at least six weeks of age and approved by University of Maryland Baltimore Institutional Animal Care and

Use Committee or by the Animal Care and Use Committee of the National Institute of

Child Health and Human Development. Table 2.1 contains the genetic designation, description of phenotype, and the abbreviation used within this thesis.

59

Table 2.1. Mouse model genetic designation and abbreviations.

Mouse Genetic Strain Description Abbreviation Used C57BL/6J Wild-type B6 animal B6 B6.SJL-PtprcaPepcb/BoyJ Wild-type B6 animal B6 B6NTac Wild-type B6 animal B6 B6.SJL-Ptprca/BoyAiTac Wild-type B6 animal B6 B10.Q/Ai Wild-type B10.A animal B10.A B10.A-H2a H2-T18a/SgSnJ Wild-type B10.A animal B10.A B10.A-Rag2tm1FwaH2-T18aTg CD4+ TCR-tg specific for pigeon 5C.C7 (Tcra5CC7,Tcrb5CC7)lwep or moth cytochrome C peptide B6.Cg-Ptprca Pepcb CD4+ TCR-tg specific for SMARTA Tg(TcrLCMV)1Aox/PpmJ LCMV peptide B6.129S2-Tcratm1Mom/J Animals lack expression of TCR TCRαβ-KO alpha and beta chain which prevents T cell development B6.129S-Themistm1Gasc/J Animals lack expression of THEMIS-KO THEMIS C57BL/6J Ptpn6me-v/me-v Animals lack expression of SHP- SHP-1-KO 1 B6.129P2-Cd5tm1Cgn Animals lack expression of CD5 CD5-KO C57BL/6J. CD5Tg Animals have transgenic CD5-Tg expression of CD5 under the control of human CD2 promoter C57BL/6J.CD5iKO Animals have floxed CD5 gene CD5-iKO and tamoxifen inducible expression of Cre recombinase (B10.A-Cd45a(Ly5a)/NAi N5 Animals lack expression of CD3ε CD3ε-KO which prevents T cell development B10.A, mPCC (tg), CD3e-/- Animals lack expression of CD3ε PCC-CD3ε-KO which prevents T cell development. Animals also have transgenic expression of membrane localized pigeon cytochrome C under the MHC class I promoter and Ig enhancer. B10.A, mPCC (tg), Animals have single allele of PCC+ CD3ε and transgenic expression of membrane localized pigeon cytochrome C under the MHC class I promoter and Ig enhancer

60

T cell purification

Lymph node and spleen (unless otherwise indicated) were isolated from mice and mechanically processed into single cell suspensions by mashing tissues through 100μM nylon mesh. The cell suspensions were then processed through a Ficoll-HiPaque Density

Gradient to isolate lymphocytes. Magnetic depletion was performed with MyOne

Streptavidin T1 DynaBeads with biotinylated anti-CD45R/B220 (RA3-6B2, Biolegend

#103204), biotinylated anti-NK1.1 (PK136, Biolegend #108704), biotinylated anti-I-A/I-

E (M5/114, Biolegend #107604), and biotinylated anti-CD11b (M1/70, Biolegend

#101204) for isolation of CD4 and CD8 T cells according to manufacturer instructions. For isolation of CD4 T cells only, the addition of biotinylated anti-CD8 (53-6.7, Biolegend,

#100703).

Proliferation dye labeling

T cells were labeled with 2μM carboxyfluorescein succinmidyl ester (CFSE,

Molecular Probes) or Cell Trace Violet (CTV, Thermo) in PBS supplemented with 0.5%

FCS in 37°C water bath for 12 minutes. The cells were washed with neat FCS to quench labeling and then resuspended in crushing buffer or TCM dependent upon assay.

Flow cytometry staining

Fc-receptor blocking was performed for 15 minutes at 4°C in PBS supplemented with 2% FCS, 0.01% azide, 1% mouse serum, 1% hamster serum, and 1% rat serum

(FcBlock). Surface staining was performed for 30 minutes at 4°C in PBS supplemented with 2% FCS, 0.01% azide, and 0.02% EDTA (FACS Buffer) using the antibodies provided in Table 2.2. Cells were washed once with FACS Buffer. Intracellular staining

61 was performed with eBioscience Fixation/Permeabilization Kit. Cells were resuspended in

1X eBioscience Fixation/Permeabilization buffer and incubated overnight at 4°C.

Intracellular staining was performed overnight at 4°C in 1X eBioscience Permeabilization

Buffer using the antibodies listed in Table 2.2. Stained cells were washed twice with FACS buffer. Cells were analyzed on an LSR-Fortessa flow cytometer (BD). All data were analyzed using FlowJo (Tree Star).

62

Table 2.2. Antibodies used for flow cytometric analysis.

Antibody Fluorochrome Manufacturer Catalog # Clone 7AAD Biolegend 420403 B220 APC-Cy7 Biolegend 103224 RA3.6B2 B220 PE BD Pharmingen 553090 RA3-6B2 CD11b APC eBioscience 17-0112-83 M1/70 CD11b PE eBioscience 12-0112-81 M1/70 CD11c APC Biolegend 117309 N418 CD11c PE eBioscience 12-0114-83 N418 CD127 APC eBioscience 17-1271-82 A7R34 CD19 APC eBioscience 17-0193-82 eBio1D3 CD19 APC-Cy7 BD Pharmingen 557655 ID3 CD25 BV650 Biolegend 102037 PC61 CD25 APC-Cy7 BD Pharmingen 557658 PC61 CD25 APC-Cy7 BD Pharmingen 557658 PC61.5 CD4 BV605 Biolegend 100451 GK1.5 CD4 APC-Cy7 BD Pharmingen 552051 GK1.5 CD4 PE BD Pharmingen 553730 GK1.5 CD4 PE eBioscience 12-0042-83 RM4-5 CD4 Pac Orange Thermo MCD0430 RM4-5 CD4 BV605 Biolegend 100451 GKJ1.5 CD4 eFluor 450 eBioscience 48-0041-82 GK1.5 CD4 FITC BD Pharmingen 553055 RM4-4 CD4 BV605 Biolegend 100451 GK1.5 CD4 APC Biolegend 100515 RM4-5 CD44 FITC Biolegend 103006 IM7 CD44 PerCPCy5.5 Biolegend 103032 IM7 CD44 FITC BD Pharmingen 553133 IM7 CD44 FITC eBioscience 11-0441-82 IM7 CD45.1 FITC Biolegend 110706 A20 CD45.1 APC-Cy7 Biolegend 110716 A20 CD45.1 APC Biolegend 110714 A20 CD45.1 PE-Cy7 eBioscience 25-0453-82 A20 CD45.1 AF700 Biolegend 110724 A20 CD45.2 PE-Cy5.5 Invitrogen 35-0454-82 104 CD45.2 APC Biolegend 109814 104 CD45.2 APC Biolegend 109814 104 CD5 eFluro450 eBioscience 48-0051-82 53-7.3 CD5 PE-Cy7 eBioscience 25-0051-81 53-7.3 CD62L PE-Cy7 Biolegend 104418 MEL-14

63

Table 2.2. Antibodies used for flow cytometric analysis (cont.).

CD62L AF700 Biolegend 104426 MEL-14 CD69 PerCP-Cy5.5 BD Pharmingen 551113 H1.2F3 CD69 FITC BD Pharmingen 553236 H1.2F3 CD69 eFluor450 eBioscience 48-0691-82 H1.2F3 CD69 APC eBioscience 17-0691-82 H1.2F3 CD69 PE-Cy7 eBioscience 25-0691-82 H1.2F3 CD8α PerCP-Cy5.5 BD Pharmingen 551162 53-6.7 CD8α APC-Cy7 BD Pharmingen 557654 53-6.7 CD8α FITC Biolegend 100706 53-6.7 CD8α PerCP-Cy5.5 BD Pharmingen 551162 53-6.7 CD8α FITC eBioscience 11-0081-85 56-6.7 CD8α APC-eFluro780 eBioscience 47-0081-82 53-6.7 CD8α BV650 BD OptiBuild 740552 H35-17.2 CD8α BB515 BD Pharmingen 564422 53-6.7 CD8α APC eBioscience 17-0081-81 53-6.7 CD8α PE-Cy7 eBioscience 25-0081-81 53-6.7 CTLA-4 APC eBioscience 17-1522-82 UC10-4B9 Ghost Dye Red780 Tonbo 13-0865-T100 Ghost Dye Violet450 Tonbo 13-0863-T100 Ghost Dye Violet 510 Tonbo 13-0870-T100 IκBα PE Invitrogen 12-9036-41 MFRDTRK LAT APC eBioscience 17-9967-42 LAT.10-17 LiveDead Violet Invitrogen L34963 LiveDead Aqua Invitrogen L34965 NFAT1c PE Biolegend 649605 7A6 NFκB p65 AF647 CST 8801S D14E12 NK1.1 APC eBioscience 17-5941-82 PK136 NK1.1 PE Biolegend 108708 PK136 PD-1 FITC Biolegend 135213 29F.A12 PLCγ PE BD Phosflow 558575 10/PLCg pZap70 (Y292) AF647 BD Biosciences 558515 J34-602 pZap70 (Y319) PE-Cy7 BD Biosciences 561458 17A/P-ZAP70 TCRβ PE-Cy7 BD 560720 H57-597 TCRβ PE-Cy7 Biolegend 109222 H57-597 TCRβ FITC BD Pharmingen 553171 H57-597 Tox PE Invitrogen 12-6502-80 TXRX10 Vβ3 BV650 BD OptiBuild 743415 KJ25 Vβ3 PE BD Pharmingen 553209 KJ25 Zap70 FITC eBioscience 11-6695-82 1E7.2

64

Flow assisted cell sorting

Purified T cells were labeled with the antibodies listed in Table 2.2 and 7AAD in

PBS supplemented with 2% FCS and 1% antibiotic-antimycotic (Gibco). Cells were sorted on a BD FACSAria II cytometer into the following groups: (1) 7AAD-CD44loCD4+CD5hi,

(2) 7AAD-CD44loCD4+CD5lo, (3) 7AAD-CD44loCD8+CD5hi, or (4) 7AAD-

CD44loCD8+CD5lo.

Adoptive transfer and antigen challenge

Lymph nodes were isolated from 5C.C7 mice and mechanically dissociated by mashing tissues through 100μM nylon mesh. 1 x 105 cells were transferred into B10.A mice and 24 hours later were challenged with MCC (30μG) and LPS (0.5μg) or PBS as a control.

Spleen and lymph node tissue were isolated at varying time points (12, 60, and 84 hours post-challenge) and analyzed by flow cytometry.

Imaging flow cytometry analysis

B6 lymph nodes were stained with antibodies to CD4, CD8, CD5, and NFκB listed in Table 2.2. Nucleus was stained with DAPI. Cells were acquired on an ImageStream

MarkII imaging flow analyzer (Luminex) equipped with two CCD cameras, at 60x magnification and low speed (7mm core stream). Analysis was performed using the IDEAS software package (Luminex). After spectral compensation, in focus, single cells were identified on the first camera bright field images and were further subdivided using two dimensional plots of CD4 and CD8 followed by CD5 histogram plots. The upper and lower

20% of CD5-expressing subsets of CD4+ and CD8+ cells were analyzed for NFκB nuclear translocation and for nuclear NFκB fluorescent intensity measurements. Nuclear

65 translocation was determined by masking the nucleus based on DAPI staining and expressed as the log-transformed Pearson’s correlation coefficient between the nuclear

(DAPI)- and NFκB (AF647)-masked images. The amount of nuclear NFκB was expressed as the geometric mean fluorescent intensity of AF647 staining signal within the DAPI- masked nuclear image.

Western blot analysis

Sorted cells as described above were lysed in 1% Nonidet P-40, 10 mM Tris-HCl

(pH 7.2), 140 mM NaCl, 2 mM EDTA, 5 mM iodoacetamide, 1mM Na3VO4, and complete protease inhibitor mixture (Roche) and incubated for 30 minutes on ice. Lysates were spun at 14,000rpm at 4°C and clarified lysate was removed. Samples were run on an SDS-PAGE gel and transferred onto nitrocellulose membranes. Membranes were blocked with 5% skim milk at room temperature for 1 hour. Membranes were incubated with primary antibodies of anti-CD5 (SC6986, SantaCruz Biotechnology), anti-IκBα (9242, Cell

Signaling Technology), anti-phosphoSHP1(Y564)(8849, Cell Signaling Technology), anti-actin (A5441, Sigma Aldrich), anti-PLCγ1 (SC81, SantaCruz Biotechnology), anti- p65 NFκB (SC372, SantaCruz Biotechnology), or anti-SP1 (SC59, SantaCruz

Biotechnology), for 4 hours followed by HRP-conjugated anti-mouse IgG, anti-rabbit IgG or anti-goat IgG. Membranes were developed and visualized using the ECL technique

(Thermo Scientific). Band densitometry was analyzed with MultiGauge software

(Fujifilm).

Cellular fractionation

66

Nuclear and cytoplasmic extracts were prepared by using Subcellular Protein Fractionation

Kit (78840, Thermo Scientific).

Thymocyte ex vivo culture

Thymus tissue was isolated and mechanically dissociated by mashing tissue through 100μM nylon mesh. 2 x 105 thymocytes were plated in T cell media and incubated for 24 hours at 37°C. Cells were harvested and surface stained as described above. Cells were washed twice with 1X Annexin Binding buffer (BD). Cells were resuspended in 1X

Annexin Binding Buffer with Annexin V PE (BD) and 7AAD (BD) and incubated at 27°C for 10 minutes and then analyzed by flow cytometry.

CX4945 and Bay 11-7082 ex vivo culture

Thymocytes or peripheral T cells (lymph node and spleen) were plated at a density of 1 x 105 cells per well in T cell media. CX4945 (Enzo) or Bay 11-7082 (Sigma-Aldrich) was resuspended in DMSO and diluted in T cell media and then plated with thymocytes.

Samples were incubated at 37°C harvested for flow cytometric analysis at varying time points (12, 15, 24, and 48 hours).

Real-time quantitative PCR analysis

Sorted cells were resuspended in Trizol and frozen at -80°C. Trizol suspension was filtered through a QIAShredder Mini Column (Qiagen). Chloroform was applied to filtrate and RNA layer was removed. RNA was further isolated using RNEasy MinElute Kit

(Qiagen) according to manufacturer instructions. Total RNA was transcribed using

SuperScript II reverse transcriptase and oligo(dT)20 primers (Invitrogen). The expression of the mRNA for Nfkbia and Actb were determined by real-time PCR using the following

67 primers. Gene-specific primer sets used in the real-time PCR assays were: Nfkbia 5’-

GGAGACTCGTTCCTGCACTT -3’ and 5’-TCTCGGAGCTCAGGATCACA- 3’; Actb

5’- GGAGCACCCTGTGCTGCTCACCGAGG 3’ and 5’-

ATCTACGAGGGCTATGCTCTCCC -3’. The expression of genes was normalized to

Actb.

In Vitro Overexpression of CD5

CD5 cDNA (amplified from Origene, Catalog #MR207937) was cloned into

MIGR1 retroviral expression vector (submitted to Addgene, Catalog #27490) or pQ2aB retroviral vector (Addgene, Catalog #124887) after PCR amplification (primers 5′-CGC

CGGAATTAGATCCCATTATCCATGGACTCCCACGAAGT-3′ and 5′-GGGGGG

GGCGGAATTTTACAGTCTCTGAGCCACTTGCAG-3′ for MIGR1 and primers 5’-

AATTGATCCGCGGCCGGCCATTATCCATGGACTCCCAC-3’ and 5’-

GCGAGGCCTCCTAGGAGTCTCTGAGCCACTTGCAGG-3’ for pQ2aB) and in-fusion reaction. CD5 retroviral constructs and empty vectors were used to transfect Phoenix-GP cells, retroviral supernatants collected, concentrated, and used to transduce BW-5147

TCR− and NIH3T3 cells by spinoculation.

RNA-Seq and Data Analysis

Total lymph node cells from B6 mice were stained with fluorochrome-conjugated antibodies against CD4, CD8, CD5, CD25, CD62L, and CD44. Sorted populations consisted of the one-third of CD25−CD62Lhi CD44lo CD4+ or CD8+ T cells expressing the highest surface levels of CD5 (CD5hi) or the one-third of CD25−CD62Lhi CD44lo CD4+ or

CD8+ T cells expressing the lowest surface levels of CD5(CD5lo). RNA was extracted from

68

FACS-sorted T cells using TRIzol reagent (Invitrogen) followed by purification and DNase treatment with RNeasy columns (Qiagen) and submitted for library construction by the

TruSeq RNA sample preparation kit v2 or TruSeq stranded mRNA sample preparation kit

(Illumina). Sequencing was done on an Illumina HiSEq. 2500 sequencer, yielding 30 to 70 million reads of 101 bp. The raw reads were aligned using STAR v2.5.2a [148] and reads quantified with featureCounts v1.4.6-p3 [149]. Differential expression analysis was performed using DESeq2 [150]. We used a false-discovery rate cutoff of <1% to define statistical significance. Data was analyzed using R Studio and Bioconductor. Heatmaps were generated using pHeatmap and Volcano plots with EnhancedVolcano. RNA-seq data are available in the Gene Expression Omnibus (accession GSE151395).

Adoptive transfer for chronic self-antigen stimulation

Lymph nodes were isolated from 5C.C7 and mechanically dissociated by mashing tissues through 100μM nylon mesh. 0.5 – 1 x 106 cells were injected intravenously (i.v.) into CD3ε-KO and PCC-CD3ε-KO hosts. 24 hours later, the CD3ε-KO hosts were challenged with MCC (30μg) and LPS (0.5μg) or PBS as a control. Cells were left resident for 28 days unless otherwise specified before isolation and analysis.

Lymph nodes were isolated from 5C.C7 and mechanically dissociated by mashing tissues through 100μM nylon mesh. 0.5 – 1 x 106 cells were injected intravenously (i.v.) into B10.A and PCC+ hosts. 24 hours later, the B10.A hosts were challenged with MCC

(30μg) and LPS (0.5μg) or PBS as a control. Cells were left resident for 5-18 days before isolation and analysis.

Ex vivo proliferation assay

69

Lymph nodes and spleens were isolated from 5C.C7 mice or the CD3ε-KO or PCC-CD3ε-

KO hosts bearing adoptively transferred 5C.C7 cells. Tissue was mechanically dissociated through 100μM nylon mesh. Cells were purified as described above. Purified cells were then stained for expression of CD4 and Vβ3 to normalize number of 5C.C7 cells from each host. 1 x 104 CFSE or CTV labeled 5C.C7 cells were plated with 2.5 – 5 x 105 splenocytes from CD3ε-KO mice with pigeon cytochrome C (PCC) peptide at indicated concentrations in TCM. Cells were cultured at 37°C for 72 hours prior to flow cytometric analysis.

Cytokine production quantification

Supernatants were collected from the ex vivo proliferation assays and IL-2 was measured using Quantikine ELISA systems (R&D Systems).

Adoptive transfer of acute and chronic cells into antigen free hosts

Lymph nodes and spleens were isolated from 5C.C7 mice or the CD3ε-KO or PCC-

CD3ε-KO hosts bearing adoptively transferred 5C.C7 cells at indicated time points. Tissue was mechanically dissociated through 100μM nylon mesh. Cells were then stained for expression of CD4 and Vβ3 to normalize number of 5C.C7 cells from each host. 1 x 106 cells were injected i.v. into CD3ε-KO and left resident for times indicated in each experiment (10 – 32 days). Cells were re-isolated and stained for expression of CD4 and

Vβ3 to normalize number of 5C.C7 cells from each culture before CFSE labeling and plating for proliferation assay.

In vitro culture system for naïve, acute, and chronic cells

Lymph nodes and spleens were isolated from 5C.C7 mice or the CD3ε-KO or PCC-

CD3ε-KO hosts bearing adoptively transferred 5C.C7 cells. Tissue was mechanically

70 dissociated through 100μM nylon mesh. Aliquot of cells were stained for expression of

CD4 and Vβ3 to normalize number of 5C.C7 cells from each culture before CFSE labeling and plated for proliferation assay (designated as initial). Remaining cells were cultured in

TCM in 24 well plates with 2.5 x 106 splenocytes from CD3ε-KO mice at 37°C for 48 hours. Cells were then isolated and stained for expression of CD4 and Vβ3 to normalize number of 5C.C7 cells from each culture before CFSE labeling and plating for proliferation assay.

Peptide and drug in vitro cultures and proliferation assays for naïve, acute, and chronic cells

Lymph nodes and spleens were isolated from 5C.C7 mice or the CD3ε-KO or PCC-

CD3ε-KO hosts bearing adoptively transferred 5C.C7 cells. Tissue was mechanically dissociated through 100μM nylon mesh. Cells were stained for expression of CD4 and Vβ3 to normalize number of 5C.C7 cells and then 1 x 105 cells were plated in 24 well plates with 2.5 x 106 splenocytes from CD3ε-KO mice and 10μM PCC peptide. Cells were cultured for 48 hours before re-isolation for CFSE labeling and plating for proliferation assay. A second experiment was performed where the cells were labeled with CTV prior to peptide culture for 48 hours and then subsequently labeled with CFSE prior to plating for proliferation assay.

Lymph nodes and spleens were isolated from 5C.C7 mice or the CD3ε-KO or PCC-

CD3ε-KO hosts bearing adoptively transferred 5C.C7 cells. Tissue was mechanically dissociated through 100μM nylon mesh. Cells were stained for expression of CD4 and Vβ3 to normalize number of 5C.C7 cells and then 1 x 105 CFSE labeled cells were plated in 24 well plates with 2.5 x 106 splenocytes from CD3ε-KO mice. Phosphatase inhibitor cocktail

71 set IV (1:100 dilution, Millipore Sigma #524628), okadaic acid (10nM, Millipore Sigma

#O9381), and Trichostatin A (500nM, Millipore Sigma #T8552) were added to the cultures. Cells were cultured at 37°C for 48 hours before re-isolation for plating for proliferation assay.

Lymph nodes and spleens were isolated from 5C.C7 mice or PCC-CD3ε-KO hosts bearing adoptively transferred 5C.C7 cells. Tissue was mechanically dissociated through

100μM nylon mesh. Cells were stained for expression of CD4 and Vβ3 to normalize number of 5C.C7 cells and CFSE labeled and then plated at a density of 2.5 x 104 in TCM in 96 well plates coated with varying doses of anti-CD3ε (145-2C11, Biolegend, #100301) as indicated. SU6656 (Cayman Chemical #13338) and SB216763 (Cayman Chemical

#10010246) were added to cultures at indicated concentrations. Cells were cultured for 72 hours before flow cytometric analysis

Tumor inoculation and isolation of tumor infiltrating lymphocytes

B6 mice were inoculated with 5 x 105 E.G7-OVA cells subcutaneously (s.c.) and allowed to grow for 21 days. At day 21 tumors, draining lymph node (popliteal), and non- draining lymph node (contralateral popliteal) were isolated. Tumors were mechanically dissociated with a razor and placed in 1mL enzyme solution (1g collagenase, 100mg, hyaluronidase, 20,000U DNase IV in HBSS) on shaker at 37°C for 1 hour. After incubation, all tissues were mechanically dissociated through 100μM nylon mesh and stained for flow cytometric analysis.

Microarray analysis

72

Naïve T cells from 5C.C7 mice or 5C.C7 cells that were acutely or chronically primed in vivo (described above) were enriched by negative selection and FACS sorted to greater than 99% purity gated as 7AAD-CD44loCD4+. Sorted T cells were suspended in

Trizol reagent (Invitrogen/Life Technologies, CA) followed by homogenization by spinning through a Qiashredder column (Qiagen, MD). 0.2 volumes of chloroform was added to the lysate and the aqueous phase separated out after centrifugation. An equal volume of 70% ethanol was added to this phase and suspension was passed through a

RNeasy mini column (Qiagen). Bound RNA was eluted and quantitated. 50ng of total RNA was then amplified using the Ovation RNA amplification system v2 (Nugen, CA), purified with DNA Clean & Concentrator -25 (Zymo Research), labeled with Cy3 using the Agilent

ULS protocol. A standard reference RNA used in the laboratory derived from pooled mouse T cells was similarly labeled using the Cy5 dye. The two probes were mixed and hybridized to Agilent Mouse V2 chips at the NIAID Microarray core facility. Data extracted from the Genepix software was subjected to quality control on the mAdb platform

(NIAID). Subsequent analyses were performed using BRB-ArrayTools developed by Dr.

Richard Simon and BRB-ArrayTools Development Team.

Adoptive transfer model system to investigate rate of change of antigen

Lymph nodes were isolated from 5C.C7 and mechanically dissociated by mashing tissues through 100μM nylon mesh. Cells were purified, CFSE labeled, and 1 – 2.5 x 105 cells were i.v. transferred into B10.A mice. Cells were allowed to seed for 24 hours prior to any antigen exposure.

Daily intraperitoneal injection for antigen dosing

73

B10.A mice bearing 5C.C7 cells were injected intraperitoneally (i.p.) with varying concentrations of PCC peptide suspended in 1X PBS. iPrecio mini osmotic pump preparation, programming, and refilling

iPrecio pumps were turned on and primed with 1X PBS. Using iPrecio management system, pumps were programmed according to dosing regimen as indicated in each experimental system. Within 2 hours of programming, pumps were implanted into animals.

Reservoir of antigen solution was exchanged through dorsal port by aspiration using insulin syringe and refilled through dorsal port with an insulin syringe. This was done every 2 days to maintain sufficient fluid in the reservoir for delivery.

Surgical implantation of mini osmotic pumps

Administration of Tribromoethanol is used for anesthesia. Once animal loses righting reflex and this confirmed by lack of tail flex reflex, protective ophthalmic ointment is applied to both eyes (lacri-lube puralube) to reduce chance of corneal desiccation. Pre- emptive analgesic of Buprenorphine is administered subcutaneous (s.c., 3 injections of 50 ul each spread over the back area) and the post injection site is prepared with 70% alcohol on clean cotton gauze or clean Q tip. Buprenex is given at 0.05-0.1 mg/ kg using a 25 G or smaller bore needle every 8-12 hours for 2 days following surgical procedure. The surgery site hair is removed with electric clippers using a #40 blade in area adjacent to surgery table. Fur is removed to provide 1 cm margins minimum for planned incision. After the removal of any loose hairs, the animal is placed on the surgery table warming pad. The clipped site will be scrubbed with an 8-12% final concentration of Betadine Surgical scrub

74 using clean Q tips or sterilized cotton gauze. Scrub is removed using 70 % alcohol on Q tips or gauze and repeated twice for a total of three applications.

A 10 mm incision is made through the prepared skin site on the back of the mouse

– dorsal opposition to the xiphoid cartilage just after the true ribs, above the rib cage. Blunt dissection creates a pocket for the iPrecio pump. The iPrecio pump is introduced into the pocket and slid in with frequent moistening of the pocket (using sterile saline drops). A 18 gauge needle is used to make a trocar for delivering the iPrecio pump catheter into the peritoneal cavity and the catheter introduced through the portal. Pharmaceutical grade

Bupivacaine is used (dripped) on incisions (0.25 % dilute 1: 4, dose not to exceed 2 mg/ kg bupivacaine) prior to closure to provide topical analgesia. Sutures, monofilament 4.0, or wound clips are used to close the skin and supplementation with tissue adhesive if necessary.

After the surgery, animals are wrapped in sterile “blankets” of autoclaved tissue wads (i.e. paper towel) and placed in a cage atop a heating pad. The cage is placed with ½ of the cage bottom on a thick towel or piece of cardboard and the other ½ of the cage bottom on a thermal supportive device at 37ºC. Subsequently, animals are placed in sterile cages and monitored approximately every 2-3 hours for a 6-hour period to ensure complete recovery from procedure. Post-operative analgesia includes Pharmaceutical grade butorphanol (i.e. 1 – 5 mg/kg s.c. q 4 hrs.), or Pharmaceutical grade buprenex (i.e. 0.05 -

0.1 mg/kg s.c. q 6 – 12 hrs.) as needed after surgical recovery. Animals may receive post- operative analgesics for an additional 3 days if required. Skin staples or clips are removed

7-14 days post-surgery. Animals are monitored twice a day for 72 hours and then every day for the remainder of the week following the surgical procedure

75

Serum collection and culture assay

Mandibular bleeds were performed and collected in BD Microtainer Gold Tubes.

Samples were spun at 14,000rpm for 2 min to collect serum. Serum was plated at varying concentrations with 1 x 104 naive 5C.C7 cells (collected as described above) with 2.5 x 106

CD3ε-KO splenocytes (collected as described above) in a 96 well plate. Cells were incubated at 37ºC for 15 hours and subsequently stained for flow cytometric analysis.

76

Chapter 3: The initial setting of the T cell activation threshold

3.1. Introduction

3.1.1. Steps in T cell development

All peripheral T cells develop from precursors in a specialized organ called the thymus. They acquire T cell receptors (TCR) for the first time as immature thymocytes.

They subsequently go through multiple selection processes, ultimately developing into single positive mature thymocytes. The details of this process have been reviewed extensively [151-157]. In this chapter, we discuss the current understanding of how the signaling machinery downstream of the TCR is first assembled and how an initial activation threshold for the T cell is also set at this time. We will then outline how our studies have improved our understanding of this critical process.

After T cell precursors leave the bone marrow committed to a lymphoid lineage

(Common Lymphoid Progenitors or CLP), these cells subsequently mature in a specialized organ called the thymus (Figure 3.1). This developmental progression can be traced by the expression of the co-receptors, CD4 and CD8. While mature cytotoxic T cells express CD8 and mature helper T cells express CD4, the initial precursors that home to the thymus do not express either molecule, so these are deemed double negative (DN) cells. The DN cells can be further fractionated into 4 stages, based on the expression of CD25 and CD44. The initial DN1 stage stemming from the Thymus Seeding Precursor (TSP) that homed in from the CLP is marked by CD44 alone, but no CD25. Subsequently, expression of CD25 comes on, leading to the CD25+CD44+ DN2 cell. This cell then loses CD44, giving rise to the

DN3 cell and eventually also loses CD25 becoming the DN4 cell. The exact roles of these molecules are not necessarily clear but, based on mixed-chimera studies, it seems that

77

CD44 provides a survival advantage to the thymocytes since the loss of this surface receptor decreases thymocyte survival in a competitive environment [158]. Yet, loss of either CD44 or CD25 does not completely abrogate thymocyte development so it is unclear what precise roles they may play during this early developmental stage [159, 160].

78

Figure 3.1. Overview of thymic development.

Common lymphoid progenitor cells first enter the thymus and then become the precursor cell. This stage is designated as double-negative (DN) since these cells lack expression of both CD4 and CD8 surface co- receptors. The DN stages can be further subsetted by CD44 and CD25 expression where the cells first are CD44+ and CD25- (DN1), then gain expression of CD25 (DN2), then lose expression of CD44 (DN3) and finally CD25 expression is lost (DN4). During the DN3 stage, VDJ recombination rearranges the TCRβ chain and tests to see if it is functional during a process called β-selection where a surrogate α-chain is partnered with the newly rearranged TCRβ. The T cells then gain expression of both CD4 and CD8 co-receptors and begin rearrangement of the TCRα chain. The newly formed TCRαβ pairing is then tested again to ensure proper signaling; both that it can signal effectively but that it does not signal too strongly. Finally, these cells commit to either the CD4 or CD8 lineage to become single-positive thymocytes. Figure adapted from [© April 2013 British Society for Immunology, https://www.immunology.org/public-information/bitesized- immunology/immune-development/t-cell-development-in-thymus]

79

These stages marked by CD44 and CD25 are helpful, since the first time a developing thymocyte acquires a part of the TCR is at the DN3 stage (Figure 3.1). At this step, T cells undergo VDJ recombination to rearrange the endogenous TCR-β chain locus.

The resulting TCR-β chain must pair with a surrogate for the alpha chain, called pre-Tα, and signal successfully into the T cell in order to further develop (Figure 3.2A). The key point is that while the TCR is assembled to form a stable structure it is ultimately the signaling of the TCR itself that is being tested. So, a T cell which has defects in molecules necessary for the signaling machinery (discussed in Chapter 1) will not develop past this stage. This concept is true for all subsequent selection steps discussed below. Efficient and appropriate signaling is a key quality control benchmark in all stages of a T cell’s development and life.

80

Figure 3.2. Steps of selection in the thymus.

(A) Rearrangement of the β chain begins during the DN3 stage. Termination of β chain rearrangement occurs after successful β selection. These cells then proliferate and become double positive where the α chain is rearranged and tested by both positive and negative selection. This shows the phases in which either D-J or V-DJ or V-J arrangement occurs within each of these phases. (B) Fates of double-positive thymocytes as they proceed through positive selection are to either survive through both selection processes through successful but not overtly strong signaling. In other cases, T cells may die through lack of proper signals or signaling too strongly. This ultimately populates a repertoire with T cells that are within a particular window of affinity. Figures adapted from Janeway Immunobiology and [161].

81

After β selection, thymocytes go through rounds of proliferation to ensure that a successful β chain is preserved in the population. These dividing cells then proceed to develop to the double-positive stage, marked by the expression of both the CD4 and CD8 molecules. In the double-positive stage, rearrangement of the α chain completes the TCR and sets the stage for at least three potential outcomes (Figure 3.2B, Creative Commons

Attribution 3.0). All of these outcomes are driven by a sequence of antigen presentation events by specialized cells in the thymus. In the thymus, cortical epithelial cells present a milieu of self-peptides to first test if the newly generated TCR can engage peptide on self-

MHC. This ensures MHC restriction and overall utility of the TCR in that host; but also drives the lineage commitments to CD4 versus CD8. The lineage decision between CD4 and CD8 begins when the cells downregulate CD8 and if they lose signal, from the lack of their interaction with class I pMHC, then they are instructed to upregulate CD8 and become

CD8 T cells Conversely, if the downregulation of CD8 does not decrease signaling, since the TCR would still be able to interact appropriately with class II pMHC, then the cells are instructed to become CD4 T cells [162]. During this process the T cells must actively signal their ability to sense self-peptides in order to survive; failure to do so results in death via apoptosis – also known as death by neglect. This is a fundamentally important step, since the critical components of the TCR signaling machinery are not only checked here, but also adjusted for optimal sensitivity. On completing positive selection in the thymus, the maturing single-positive T cell tunes its TCR signaling using multiple processes, in such a way that the endogenous positive selecting ligand is no longer stimulatory. In addition to the intracellular signaling alterations, there is also an upregulation of the TCR itself after positive selection [163]. Moreover, there are changes in co-receptor expression

82

(particularly CD8) that modifies signaling downstream of the TCR, termed ‘co-receptor tuning’ [164]. As I discuss further below in section 3.1.2, this adjustment of reactivity alters the TCR activation threshold through intrinsic mechanisms and has a consequence on all subsequent steps of T cell activation.

The third selection event is negative selection (Figure 3.2B) where TCRs that are overtly autoreactive are eliminated from the repertoire. The two major selection stages after a full TCR is generated (positive and negative) operate with the help of different antigen presenting cells (Fig 3.3, Copyright Clearance Center #4882680753890). Positive selection requires interactions with epithelial cells in the cortical regions of the thymus (cortical

Epithelial cells or cTECs). Negative selection is more diffuse, requiring either dendritic cells (DC) or medullary epithelial cells (mTECs). This sets up a geographical sequence of events (Figure 3.3) where each stage of development is closely tied to interactions within a specific microenvironment. This is important to consider since signals other than TCR- pMHC interactions can significantly impact T cell development. While these are not major focal points for this thesis, some such signals include cytokines [165, 166], chemokines

[167], cell adhesion receptors [168, 169] and key regulators such as Notch [166, 170]. Once each of these selection steps are completed, the T cell finally becomes a mature single positive cell (SP) that is either CD4 (CD4-SP) or CD8 (CD8-SP) and emigrates to the periphery. Yet even at these early stages within a T cell’s life, there is a wide variety in signaling molecule expression and functionality in the remaining thymocytes. As a result of these molecular changes, the initial activation threshold of a T cell is determined by selection events in the thymus. In effect, while we are used to thinking about specific TCRs being selected for peripheral function from the vast diversity originally generated in the

83 thymus (Figure 3.4, Creative Commons Attribution 3.0), this is quite dependent on the actual signaling criteria as well. Independent of which TCR a T cell leaves the thymus with, intricate regulatory changes in the intrinsic signaling network are more deterministic, i.e. only T cells with a certain signaling framework successfully seed the periphery.

84

Figure 3.3. Geography of thymic T cell development.

T cells develop in discrete stages within particular microenvironments of the thymus. β selection occurs within the cortex of the thymus where the thymocytes interact with cortical epithelial cells as the first quality control step of the newly generated TCRβ. These cells then develop into double-positive thymocytes and migrate to the medulla where they interact with either medullary epithelial cells or dendritic cells/mTECs and go through positive and negative selection, respectively. The location of each of these processes is critical because specialized antigen presenting cells are critical for the quality control of developing thymocytes. Figure adapted from [162].

85

Figure 3.4. Thymic development processes and peripheral activation poise the TCR repertoire.

T cells go through thymic selection where a distinct window of affinity is selected. This leaves the remainder of generated thymocytes to die either through neglect or negative selection from either too low or too high affinities, respectively. This selection process results is a great deal of death since the random generation of TCR chains does not guarantee that they are functionally active. Beyond the thymus, there are additional pressures within the periphery, particularly upon antigen encounter, that further dwindle the repertoire of T cells. Figure adapted from [171].

86

3.1.2. The tuning of activation thresholds during development

The molecular changes that accompany development serve a variety of purposes.

The first series of signals (at the β selection step) is thought to be ligand independent; but all subsequent steps require a pMHC interaction. The key difference between positive and negative selection is in the strength of signal perceived by the TCR. If it can detect a pMHC complex, albeit very weakly, then the T cell survives positive selection. But negative selection is much more stringent. Based on the strength of TCR signaling, a decision is made that removes the T cell from the system if it signals too strongly [172]. Between these two extremes of selection, there is a wide variety of signaling states and concomitant T cell fates. The adjustments in the signaling machinery in the thymus cater to these choices.

First, the TCR is kept maximally sensitive to any ligand, which would allow any weak interaction to positively select the TCR. Next, the TCR undergoes a second tuning phase where the activation threshold is adjusted to prevent being functionally triggered by that positively selecting self-ligand again. It should be noted that some of the changes after positive selection may also relate to lineage commitment [162]. While there are multiple key modifications of TCR signaling that are not completely understood now (Figure 3.5,

Copyright Clearance Center #4882631444332), I will focus on some key regulators relevant to my thesis work including some of the best characterized regulators; mir181a,

THEMIS, Tespa1, and finally, CD5.

87

Figure 3.5. Signaling molecules expressed in the thymocytes augment the TCR signaling cascade

There are a multitude of signaling molecules that regulate the strength of TCR signals within the thymus. Some of these molecules, like miR181a function to inhibit negative regulators which allows TCR signaling to be heightened during the selection phases. Conversely, the upregulation of negative inhibitors such as PTPN22 dampens TCR signaling after the selection process. Together the checks and balances of signaling molecule expression aids to select thymocytes that bear TCRs that can signal upon encounter with peptide- MHC but do not signal too strongly. Figure adapted from [64].

88

The first, miR181a is expressed highly in pre-positive selection stages of thymic development. As a microRNA, miR181a potentially regulates multiple genes in T cells.

Relevant to the TCR signaling cascade, it has been shown to negatively regulate some known negative regulators of TCR signaling. This double negative impact means that thymocytes with high levels of miR181a are more easily activated by weak ligands; those that are typically involved in positive selection. Targets of miR181a include ubiquitin ligases and phosphatases such as, DUSP5, DUSP6, SHP-2 and PTPN22 [173] (Figure 3.5).

The loss of these negative regulators increases the signal transduction of the TCR and ultimately increases the antigen sensitivity of the T cell. miR181a expression seems to reflect TCR signaling. After positive selection, the expression of miR181a is decreased and allows the expression of the negative regulators it was keeping at bay to accumulate again.

This results in an alteration of the overall activation threshold of the T cell preventing it from signaling in response to the original positively selecting self-peptides. The use of an antagomir to miR181a, prevents ERK activation during positive selection [173] to illustrate the importance of this regulator for activating key distal components of TCR signaling. It is important to point out that altering miR181a expression in T cells does not only affect positive selection, but also negative selection and peripheral autoreactivity [174]. Taken together, these data demonstrate the importance of balancing the expression of signaling molecules to modulate the activation threshold of a T cell.

Two more recently appreciated molecules that affect TCR signal strength in the thymus are THEMIS and Tespa1. Genetic ablation of THEMIS reduces the number of single-positive thymocytes [142]. Later studies demonstrated that the activity of this adapter is through SHP-1 mediated repression of TCR signaling [175, 176]. Therefore, it

89 is not entirely clear if the loss of thymocytes in THEMIS knockout models is due to reduced positive selection or enhanced negative selection. The genetic ablation of THEMIS increased calcium flux and ERK phosphorylation in thymocytes compared to wild-type controls [175] further indicating that this molecule restrains TCR signaling. Recent studies have indicated that THEMIS inhibits the phosphatase activity of SHP-1 through CABIT domain interactions [177] suggesting that the loss of thymocytes in the THEMIS knock- out models may be due to loss of positively selecting TCRs of lower affinities. Beyond thymic selection, the use of peripheral knock-out models of THEMIS from Brozstek et al found that proliferation under lymphopenic conditions requires the presence of THEMIS

[178] where there is a competitive advantage to the wild-type cells compared to the knock- outs.

Tespa1 was first reported as a molecule that transduces positive selection signals since the genetic knock-out reduces the number of single positive thymocytes [179]. This has been further studied to show that Tespa1 recruits IP3R1 to the TCR proximal region to potentiate efficient calcium signaling [180]. More recently, there have been efforts to dissect the thymic versus peripheral roles of these two novel thymic regulators. Lyu et al. used conditional knock-out of Tespa1 to demonstrate that Tespa1, unlike THEMIS, does not play a role in maintaining TCR signaling in the periphery since there were no defects in TCR signaling or proliferation [181]. Taken together, this suggests that THEMIS augments TCR signaling and therefore the activation threshold through inhibiting negative regulators while Tespa1 aids in propagating signaling.

90

3.1.3. CD5

One of the first mechanisms implicated in the tuning of signaling during T cell maturation is the surface glycoprotein CD5. CD5 is a Scavenger Receptor Cysteine Rich

(SRCR) family protein expressed on all T cells and on subsets of B cells [182]. CD5 was initially identified and named T1/Leu-1 as a marker of CD4 T cells [183]. CD5 expression is tightly regulated by a variety of transcription factors including Ets-1 and AP-1 as well as other lymphocytic specific transcription factors which may explain the cell type restricted expression of the gene [184, 185]. The gene and protein structure of CD5 is shown in Figure 3.6A, Copyright Clearance Center #4882681061622. Briefly, there are 11 exon regions in the gene that code for the following protein domains: a signal peptide, three

SRCR domains, a hydrophobic transmembrane region, and a cytoplasmic tail that contain

ITAM-like and ITIM-like sequences [186]. Unlike the consensus ITAM discussed in section 1.4.1, this contains tyrosines that are phosphorylated but, they are not spaced in the same manner as a canonical ITAM [186]. Moreover, ITIMs have similar structures but instead recruit molecules that typically dampen signaling rather than aid to propagate it. It has recently been appreciated that there is not strict positive and negative roles in signaling of these two amino acid sequences so the exact mechanism that CD5 recruits molecules to these sequences has yet to be fully appreciated [67]. To begin to investigate CD5 function,

CD5 knockout (CD5-KO) animals were generated by homologous recombination to disrupt exon 7 and 8 in embryonic stem cells so that no functional protein can be formed

[143]. CD5-KO thymocytes have higher calcium fluxes suggesting that CD5 restrains these signals through a negative regulatory mechanism [187].

91

Figure 3.6. CD5 gene and protein structure and potential interacting partners.

(A) The CD5 gene is composed of 11 exons that codes for 6 different protein coding regions. These include a signal peptide, three scavenger receptor cysteine region domains, a transmembrane domain, and a cytoplasmic tail that contains ITAM-like and ITIM like domains with multiple target sites of other TCR signaling molecules as shown in the diagram. (B) CD5 has many proposed intracellular binding partners that mediate effects on the proximal TCR machinery such as Zap70 and Lck. Many of these attenuate the activity of TCR signaling molecules through post-translational modifications. Figures adapted from [186, 188].

92

CD5 does not have a single identified ligand but rather, has been shown to interact with a variety of molecules including CD6, IL-6, CD72, and CD5 itself through its extracellular domain [189, 190]. Yet, any potential ligand interaction or even the extracellular domain itself does not seem to be critical for CD5’s ability to negatively regulate TCR-signaling [191]. The intracellular domain of CD5 has multiple tyrosines within ITAM-like and ITIM-like motifs which are phosphorylated at various stages of T cell activation [188]. These phosphorylated tyrosines create SH2 binding sites for a variety of intracellular molecules to dock on the intracellular tail of CD5. The activation-induced recruitment of molecules such as the phosphatase SHP-1, the ubiquitin ligases Cbl-b and

Cbl, CSK, and casein kinase II to these residues and thereby to the TCR signalosome is largely thought to mediate the negative regulatory impact of CD5 in lymphocytes [192-

202] (Figure 3.6B, Creative Commons Attribution 4.0). SHP-1 is one of the most well characterized binding partners of CD5. It has been shown that SHP-1 is recruited and mediates the dephosphorylation of proximal TCR signaling molecules to attenuate TCR signals [194]. However, SHP-1 is not essential to the inhibitory function of CD5 since in

SHP-1 knock-out (SHP-1-KO) T cells, there is still an inhibitory effect of CD5 on TCR signaling [201]. Therefore, while SHP-1 associates with CD5 it is not essential to its function. Additionally, the recruitment of Cbl molecules to the intracellular tail of CD5 are able to target TCR signaling molecules for degradation such as the CD3 chains, Lck,

Zap70, and PI3K [202-207]. Ultimately, downregulation of these molecules results in attenuation of TCR signaling, further contributing to the negative inhibition from CD5.

Another contribution of these ubiquitin ligases is that they seem to form a complex on the intracellular tail of CD5 with CSK [202]. CSK impacts TCR signaling by regulating Src

93 family kinases such as Lck [208] and is thought to have an impact on actin remodeling which is critical for effective TCR signaling [209]. Finally, another binding partner of CD5 is casein kinase II (CKII) [193]. CKII is a constitutively active serine/threonine kinase, composed of alpha and beta subunits, that phosphorylates the intracellular tail of CD5

[210]. CKII has shown to influence thymocyte survival by modulating ERK signaling

[211]. Recently, additional CD5 binding partners were identified through immunoprecipitation and mass-spectrometry analysis which may provide novel insight into the signaling modalities of CD5 beyond the above mentioned molecules [212, 213].

Taken together, there are a variety of mechanisms by which CD5 can mediate its inhibitory action, but the exact mechanisms are not well understood.

Work from Paul Love and others demonstrated that CD5 is also upregulated in proportion to each TCR’s self-reactivity during positive selection [214]. As a negative regulator of TCR signaling, this is thought to dampen the signaling potential of the TCR such that it becomes virtually ignorant of self-peptide presentation. The authors find that alterations in CD5 expression can either decrease (through overexpression) or increase

TCR signaling (through genetic ablation) to change the signaling predisposition of a particular TCR. Importantly, these effects are only evident within TCR-Tg models [144].

In a polyclonal repertoire, modulation of CD5 expression is masked by a shift in the repertoire itself rather than a loss or gain of cells in general [144]. Therefore, CD5 acts to control the selection boundaries of both positive and negative selection. CD5 restrains signaling of TCRs that are too highly reactive through upregulation of expression but, for those TCR already of low signaling, there is little upregulation of CD5 to prevent any dampening of the signals. Therefore, through the proportional upregulation of a negative

94 inhibitor, a variety of affinities of TCR during thymic development can be selected (Figure

3.7, Copyright Clearance Center #4882631444332).

95

Figure 3.7. CD5 is developmentally regulated and proportional to TCR signaling strength in the thymus.

During positive selection thymocytes must effectively signaling against self-peptide MHC to survive. During this process, CD5 is proportionally upregulated to the strength of that signal. As such, the negative inhibitory effects of CD5 exerted on the TCR signaling cascade is dictated by the original capacity to signal. This regulatory axis tunes the TCR to prevent it from begin able to react to self-peptide MHC presented in the thymus after these selection steps are completed. Figured adapted from [64].

96

Taken together, while the regulation of signaling networks in developing T cells is complex, CD5 remains one of the best indicators of these dynamic changes. Yet, it is not clear how exactly CD5 influences the developing thymocyte at a biochemical level.

3.2. Chapter Summary

Given that CD5 is widely used as a marker for broad alteration of T cell’s activation thresholds starting within the thymus, we set out to examine the signaling network in developing T cells that best correlates with CD5. We find that the negative regulator of

NFκB, namely IκBα (see Figure 3.8 for a brief overview of the NFκB pathway), is expressed in very strict correlation with the thymocyte surface levels of CD5. IκBα, similar to CD5, is increased in expression after positive selection. Ablation of CD5 (in the germline or using a new inducible knockout mouse model) led to decreased IκBα in thymocytes.

Interestingly, this regulatory axis is not maintained by canonical binding partners of CD5.

Finally, we find that there is a significant survival advantage of CD5hi cells that is dependent on the activity of NFκB. Taken together, these data demonstrate the variability in T cell responses may reflect the alterations that are made early at the selection phase. In turn, this variability is a reflection of the activation threshold of each cell which is modulated, in this particular example, through changes in the NFκB pathway.

97

Figure 3.8. Overview of the components of the NFκB pathway.

There are two arms of the NFκB pathway, the canonical and non-canonical pathways. The canonical pathway lies downstream of the TCR signaling pathway where proximal kinases activate PKCθ. This kinase then activates the CBM complex which is composed of CARMA1, Bcl10, and MALT1. This complex activates the IKK complex which then causes the degradation of IκB to release NFκB transcription factors, p50 and p65. These then translocate to the nucleus to exert their transcription factor function. In parallel, there is also a non-canonical NFκB that is induced by activation of NFκB-inducing kinase (NIK) downstream of TNF receptors and other costimulatory molecules. NIK activation leads to activation of IKKα that then processes p100 into p52. The heterodimer of processes p52 and RelB are then translocated to the nucleus to regulate gene expression. Figure adapted from [© 2020 Novus BiologicalsTM a Bio-Techne® Company, https://www.novusbio.com/nfkbpathway]

98

3.3. Results

3.3.1. Analysis strategy for identifying thymocyte subsets.

To first begin to analyze thymocytes, we used a flow cytometry approach to subset the thymocyte populations by their developmental stages as discussed in section 3.1.1. We used a combination of CD4, CD8, and CD69 staining to subset thymocyte populations in

B6 mice. CD69 was used to distinguish pre-selection and post-selection thymocytes as previously described [215]. The DN populations were gated as CD4-CD8-, DP pre- selection were gated as CD4+CD8+ CD69lo, DP post-selection were gated as

CD4+CD8+CD69hi, single-positive (SP) CD4 were gated as CD4+CD8-, and SP CD8 were gated as CD4-CD8+ (Figure 3.9). This was the gating strategy that is used for the remainder of the thymocyte analysis in all subsequent sections of Chapter 3.

99

Figure 3.9. Flow cytometric staging of thymic development in a C57BL/6 mouse.

Thymocytes were isolated from wild-type mice and analyzed by flow cytometry. First setting a lymphocyte gate, singlet gate, then gating live cells as designated by 7AAD negativity. To separate thymocyte populations, the expression of CD4 and CD8 were examined to bin these into either double-negative, double- positive, or single-positive thymocytes. The double-positive population was further subsetted by CD69 expression to designate those thymocytes that had undergone positive selection.

100

3.3.2. Signaling variations during development in thymocytes reflects the heterogeneity of

T cell responses

The increasing levels of CD5 on T cells typically reflects a developmental continuum in the thymus as discussed above. As previously reported, early thymic precursors to T cells (CD4-CD8- double-negative (DN) thymocytes) express very little surface CD5 [214]. During the positive-selection step (transition of the CD4+CD8+ double- positive (DP) cells to express CD69), there is a proportional upregulation of CD5 to the strength of the positively-selecting TCR signal [214]. We therefore examined if the levels of IκBα are also dynamically regulated with CD5 during T cell development (Figure 3.10).

As previously described, CD4-CD8- thymocytes had lower expression of CD5

(gMFI 1917.55 ± 729.76) compared to developing T cells that are DP CD69hi cells (gMFI

4647.27 ± 1644.39, Figure 3.10). In concert with CD5 upregulation, IκBα expression also increased through these developmental stages in a tight relationship, such that it resembles two epitopes on the same molecule based on the proportional relationship (Figure 3.10).

CD5 levels remained low prior to positive selection at the double-positive CD69lo stage but there was concurrent upregulation of both CD5 and IκBα at the positive-selection stage where double-positive CD69lo cells transition to double-positive CD69hi cells (Figure

3.10B) This suggested that not only does IκBα expression correlate with CD5 but, it may be a novel quantitative marker for positive selection and TCR self-reactivity as previously reported for CD5 and Nur77 [216].

101

Figure 3.10. CD5 and IκBα are developmentally regulated in thymocytes.

(A) Representative plots showing the levels of CD5 versus IκBα during progressive stages of T cell development in the thymus. Thymocytes were stained for CD4, CD8, and CD69 and gated on markers listed on top of each panel (gating strategy shown in Figure 3.9). DP = CD4+CD8+ and SP = single positive for either CD4 or CD8 expression as designated. (B) The CD5 and IκBα gMFI on thymocyte subsets as identified in panel A. (n = 11 from 3 independent experiments. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

102

3.3.3. CD5 expression is required for the upregulation of IκBα and is independent of canonical regulators

Given our current understanding of how TCR affinity to self-ligands, and by extension TCR-signal intensity, influences CD5 expression, it was formerly possible that the concordance we observed between CD5 and IκBα expression were both independently regulated by TCR; where higher affinity interactions with self-peptides resulted in upregulation of IκBα as well as CD5. If this were true, we would expect the IκBα levels in a CD5 knockout (CD5-KO) T cell to be either unchanged relative to a wild-type (WT) T cell or potentially slightly higher due to stronger TCR signals from a loss of CD5 inhibition.

In the absence of CD5, to correlate with the levels of IκBα, we examined the gradation of

IκBα levels as the CD5-KO T cells progress from DN through positive selection and transition into SP thymocytes (Figure 3.11).

We found that CD5-KO thymocytes had significantly lower expression of IκBα in

DP CD69hi, SP CD4, and SP CD8 thymocytes compared to wild-type (CD5+/+) animals

(Figure 3.11) but not within the CD4-CD8- and DP CD69lo. This suggests a failure to upregulate IκBα in CD5-KO (CD5-/-) animals as thymocyte development progresses through TCR-dependent selection stages.

103

Figure 3.11. Loss of CD5 decreases IκBα expression in thymocytes.

(A) A similar gating strategy as in Figure 3.9, was used to examine different thymocyte subsets in wild-type (WT, top row) or CD5 knock out (CD5-KO, bottom row) mice and representative flow plots are shown. (B) The gMFI of CD5 and IκBα on thymocyte populations as gated in panel A from WT or CD5-KO mice (n = 3 - 5 from 1 experiment. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

104

While these data revealed an unexpected role for CD5 in directly modulating IκBα levels, the possibility still existed that our observations resulted from the selection of different TCRs in the WT versus CD5-KO mice [194, 217]. Since CD5 regulates TCR signals in developing thymocytes, the absence of CD5 could lead to the selection of a lower self-affinity TCR repertoire resulting in basal differences of IκBα expression. To evaluate this, we used an inducible knockout mouse model in which CD5 was ablated only after administration of Tamoxifen (referred to as CD5-iKO) as shown in Figure 3.12.

Administration of tamoxifen induced expression of Cre recombinase. Exons 3-5 of the CD5 gene were flanked by loxP sites so, upon expression of Cre, this gene segment is excised and removed (Figure 3.12A). This leaves the gene without critical domains and renders the expression of CD5 nonfunctional. For validation, tamoxifen was administered for 5 days and then thymocyte populations were measured (Figure 3.12B and 3.12C). There were no changes in the frequency or cell number of the populations within the tamoxifen administered ERT2-Cre CD5 floxed animals compared to their floxed only littermate controls (Figure 3.12B). Additionally, there was complete ablation of CD5 expression in

SP CD4 and SP CD8 thymocytes (Figure 3.12C).

105

Figure 3.12. Construct design and characterization of CD5 inducible knock-out mouse model.

(A) Schematic of CD5-floxed mice with the CAG-ERT2 transgene showing the excision of CD5 gene before and after tamoxifen administration. (B) Representative flow plots of thymocytes isolated from CD5-floxed mice with CAG-CreERT2 transgene (CD5 inducible knock-outs; CD5-iKO) after treatment with tamoxifen for 5 days. (C) Histograms of CD5 expression are shown in single-positive CD4 and CD8 T cells. (D) Frequency of each thymocyte subsets and absolute cell counts from WT and CD5-iKO animals.

106

Upon observation of the thymocyte populations in the CD5-iKO mice, we found that there was a significant decrease in the expression of IκBα, in post selection thymocytes: DP CD69hi, SP CD4, and SP CD8 (Figure 3.13). This was consistent with the previous data in the genomic deletion of CD5, further supporting an active role of CD5 in maintaining IκBα expression.

107

Figure 3.13. Inducible loss of CD5 decreases IκBα in post-selection thymocytes.

(A) A similar gating strategy as in Figure 3.9, was used to examine different thymocyte subsets in wild-type (WT, top row) or CD5 inducible knock out (CD5-iKO, bottom row) mice and representative flow plots are shown. (B) The gMFI of CD5 and IκBα on thymocyte populations as gated in Figure 3.9 from WT or CD5- iKO mice (n = 2 - 5 from 1 experiment. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

108

Given that there was a direct role of CD5 expression to increase IκBα expression, we first investigated a known binding partners of CD5 that could influence the expression of IκBα, SHP-1 as well as THEMIS, an additional thymic regulator, as discussed in section

3.1.2. We found that genomic ablation of either SHP-1 or THEMIS did not impair expression of CD5 nor IκBα in bulk thymocytes via western blot analysis (Figure 3.14).

109

Figure 3.14. SHP-1 and THEMIS ablation does not impact CD5-IκBα axis in thymocytes.

Western blot analysis of CD5, IκBα, pSHP-1, and actin in thymocytes from wild-type (WT), THEMIS knock- out (THEMIS-KO), and SHP-1 knock-out (SHP-1-KO) animals (n = 1 from 1 independent experiment). Western blot, courtesy Seeyoung Choi.

110

3.3.4. CD5 expression provides an NFκB-dependent survival advantage in thymocytes

NFκB transcription factors are involved in multiple cellular functions including the survival of thymocytes and T cells both under homeostatic conditions or after selection

[218]. We examined the survival of CD5hi and CD5lo thymocytes using a simple ex vivo culture assay (mimicking trophic factor withdrawal induced apoptosis as described [219]).

After 24 hours of culture without additional cytokines, we found that the expression of

CD5 had a significant impact on cell death as measured by 7AAD and Annexin V staining

(Figure 3.15). CD5hi populations, as gated by the top 20% expressors within each thymocyte population (and conversely CD5lo populations are the lowest 20% expressors), had lower proportions of Annexin V+ cells in the developmental stages after positive selection including DP CD69hi, SP CD4, and SP CD8 (Figure 3.15B). This suggests that the upregulation of CD5 after TCR stimulation, and by extension IκBα, during positive selection provides a survival advantage, potentially through NFκB signaling.

111

Figure 3.15. CD5hi thymocytes have a survival advantage over CD5lo cells.

Thymocytes from B6 mice were cultured ex vivo for 24 hours. Cells were stained and gated as in Figure 3.9, together with Annexin V and 7AAD to assess cell death. The populations were grouped based on top and bottom 20% of CD5 expression within each population. (A) Representative flow plots are shown. (B) The frequencies of Annexin V+ cells within the top and bottom 20% of CD5 expressers are shown (n= 4 from 1 experiment is shown and n = 3 from second independent experiment). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

112

We then investigated if NFκB signaling was required for this survival effect. We cultured thymocytes ex vivo in the presence or absence of an NFκB inhibitor, Bay 11-7082, at varying concentrations. After 15 hours, cell death was analyzed by staining with 7AAD and Annexin V. We found that with increasing doses of Bay 11-7082, the significant differences in death between the CD5hi and CD5lo in post-selection DP CD69hi, SP CD4, and SP CD8 populations was abrogated (Figure 3.16). This suggests that the survival advantage of CD5hi cells as seen in Figure 3.15, is mediated through NFκB signaling.

113

Figure 3.16. Survival advantage of CD5hi thymocytes is dependent on NFκB function.

Thymocytes from B6 mice were cultured ex vivo for 15 hours with or without varying Bay11-7082 doses. Cells were stained and gated as in Figure 3.9, together with Annexin V and 7AAD to assess cell death. (A) Frequency of Annexin V cells in each thymocyte subsets. Representative flow plots are shown for the, double-negative (B), double-positive CD69lo (D), double-positive CD69hi population (F), single-positive CD4 (H), single-positive CD8 (J). The frequencies of Annexin V+ cells within the top and bottom 20% of CD5 expressers of the double-negative (C), double-positive CD69lo (E), double-positive CD69hi population (G), single-positive CD4 (I), single-positive CD8 (K). (n = 3 from 1 independent experiment). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

114

Figure 3.16. Survival advantage of CD5hi thymocytes is dependent on NFκB function (cont.). Thymocytes from B6 mice were cultured ex vivo for 15 hours with or without varying Bay11-7082 doses. Cells were stained and gated as in Figure 3.9, together with Annexin V and 7AAD to assess cell death. (A) Frequency of Annexin V cells in each thymocyte subsets. Representative flow plots are shown for the, double-negative (B), double-positive CD69lo (D), double-positive CD69hi population (F), single-positive CD4 (H), single-positive CD8 (J). The frequencies of Annexin V+ cells within the top and bottom 20% of CD5 expressers of the double-negative (C), double-positive CD69lo (E), double-positive CD69hi population (G), single-positive CD4 (I), single-positive CD8 (K). (n = 3 from 1 independent experiment). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

115

A second molecule involved in CD5 signaling is casein kinase II (CKII). CKII interacts with the cytoplasmic tail of CD5 [193] to influence T cell subset survival in the thymus [220]. We were interested in understanding the potential role of CKII activity in survival of thymocytes since CKII can influence ERK and NFκB signaling [211, 221]. To address this, we cultured thymocytes ex vivo in the presence or absence of a well- characterized CKII inhibitor CX4945 [222, 223]. This is an ATP-competitive inhibitor of

CKII that abrogates kinase activity [224, 225]. After 15 hours of culture, there were no significant differences in survival compared to DMSO controls in total thymocytes within each subset (Figure 3.17A). When comparing CD5hi and CD5lo thymocytes of each stage we found that CX4945 treatment did not impair the survival advantage of CD5hi thymocytes in any subsets (Figure 3.17B – 3.17K), unlike what was seen with the NFκB inhibitor treatment (Figure 3.16). As seen with the initial thymocyte culture (Figure 3.15), the survival advantage was not observed in pre-selection thymocytes (Figure 3.17B –

3.17E). Upon positive selection and the upregulation of both CD5 and IκBα (Figure 3.10), the survival advantage seen in Figure 3.15 was maintained in the presence of CKII inhibition suggesting that it does not play a role in maintaining the survival phenotype of

CD5hi cells (Figure 3.17F – 3.17K).

116

Figure 3.17. Casein kinase II inhibition does not impair survival of CD5hi thymocytes.

Thymocytes were cultured ex vivo for 15 hours with varying concentrations of the casein kinase II inhibitor, CX4945. (A) The frequency of Annexin V+ cells within each population gated as in Figure 3.9 is shown (n = 3 from 1 independent experiment). Representative flow plots are shown for the, double-negative (B), double-positive CD69lo (D), double-positive CD69hi population (F), single-positive CD4 (H), single-positive CD8 (J). The frequencies of Annexin V+ cells within the top and bottom 20% of CD5 expressers of the double-negative (C), double-positive CD69lo (E), double-positive CD69hi population (G), single-positive CD4 (I), single-positive CD8 (K). (n = 3 from 1 independent experiment). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

117

Figure 3.17. Casein kinase II inhibition does not impair survival of CD5hi thymocytes (cont.). Thymocytes were cultured ex vivo for 15 hours with varying concentrations of the casein kinase II inhibitor, CX4945. (A) The frequency of Annexin V+ cells within each population gated as in Figure 3.9 is shown (n = 3 from 1 independent experiment). Representative flow plots are shown for the, double-negative (B), double-positive CD69lo (D), double-positive CD69hi population (F), single-positive CD4 (H), single-positive CD8 (J). The frequencies of Annexin V+ cells within the top and bottom 20% of CD5 expressers of the double-negative (C), double-positive CD69lo (E), double-positive CD69hi population (G), single-positive CD4 (I), single-positive CD8 (K). (n = 3 from 1 independent experiment). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

118

3.3. Discussion

CD5 has typically been used as a surrogate marker of signal strength received during positive selection and is a well characterized negative inhibitor of TCR signaling.

Yet, the exact mechanisms of how these roles are exerted is not well understood. CD5 recruits a variety of molecules through its ITAM-like and ITIM-like motifs such as SHP-

1, Cbl molecules, and CKII but these molecules do not explain the roles of CD5 entirely.

Here we find that there is a distinct correlation between the expression of CD5 and an inhibitor of NFκB, IκBα, suggesting a novel link to the NFκB pathway.

It was first appreciated that SHP-1 was recruited to the cytoplasmic tail of CD5 and recruited into the TCR signaling complex to dephosphorylate kinases within the TCR signaling cascade [194]. Yet, it was subsequently shown that the inhibitory function of

CD5 is maintained in the absence of SHP-1 [201]. Despite the non-essential role of SHP-

1 we were interested in understanding if this molecule plays a role in the CD5- IκBα. We found that SHP-1-KO mice do not have altered relationship between CD5 and IκBα which suggests that it does not play an essential role in maintaining this signaling axis either

(Figure 3.14). These data further highlight that the molecular mechanisms of CD5 have yet to be completely explained.

As such, there is continuing research to identify binding partners of CD5. CD5 has also been shown to bind Cbl proteins [202]. These proteins are ubiquitin ligases that target proteins for degradation and can negatively regulate TCR signaling [203]. Therefore, it can be interpreted that CD5 acts as a scaffold to allow the recruitment of Cbl molecules into proximity of the TCR complex to negatively regulate the signaling cascade. Moreover, recent proteomics analysis of these molecules has highlighted novel interactions between

119

CD5, Cbl, and other regulatory molecules that may provide further insight into the potential regulatory mechanism [213]. We have yet to investigate the role of Cbl molecules in the

CD5- IκBα axis.

The intracellular tail of CD5 also contains a CK2 binding/activation domain which facilitates interactions with CKII. CKII is a broad acting serine threonine kinase that has a multitude of cellular targets [226]. Interestingly, disrupting the CK2 domain on the intracellular tail of CD5 modulates ERK activation and thymocyte survival, indicating that

CKII via CD5 may influence TCR signaling [211]. Moreover, there is evidence that CKII interacts with the NFκB pathway by increasing IκBα expression via NFκB mediated synthesis [221] as well as through a post-translational stabilization via phosphorylation in the C-terminal PEST domain of IκBα [227]. Given these data we were interested in understanding the potential role of CKII activity in the novel CD5- IκBα axis but, as discussed above there was no change in thymocyte survival with CKII kinase inhibition

(Figure 3.17) suggesting that the kinase activity of this molecule does not affect the functionality of the CD5- IκBα axis at least in short term culture conditions.

The NFκB pathway is essential for thymocyte development [228]. First, it is a critical pathway engaged downstream of the TCR where the loss of NFκB activity prevents the development of SP CD8 thymocytes [229]. For example, the loss of NEMO or

IKK1/IKK2 results in arrest at the SP stage [230, 231]. Interestingly, over-active NFκB signals preclude development of SP CD4 T cells emphasizing distinct control of this pathway to balance the dichotomy between CD4 and CD8 T cells. These checkpoints all seem to operate at the level of selection where NFκB signals are critical for the cell to have functional TCR signaling.

120

Secondly, NFκB signals provide pro-survival cues in response to cytokine stimulation within the thymus where cells with NFκB impairments had increased cell death in response to cytokine stimulation such as with TNFα [231]. Unconventionally, this is through the inhibition of RIPK1 in SP thymocytes [219]. Briefly, the authors were interested in understanding the role of NFκB signaling downstream of TNF signals in thymocytes. What they found was that the loss of canonical NFκB subunits, Nfkb1 and

Rel, did not alter development as was expected given the essential roles of NFκB signaling described above. The authors found that there was a developmentally regulated kinase,

RIPK1 that has dual functionality in thymocyte survival; where RIPK1 in the absence of proper NFκB signaling sensitized thymocytes to TNF induced death but, in the absence of

RIPK1 there was also cell death. The IKK complex was responsible for inhibiting the formation of RIPK1 complex IIb for induction of cell death. This demonstrates a novel and complex molecular interaction in thymocytes involving NFκB signaling for survival.

We found a distinct role of NFκB activity in maintaining the phenotype observed with higher expression of CD5 and IκBα. We found that CD5hi T cells have a survival advantage that when cultured with an NFκB inhibitor, abrogated the survival advantage. In this assay we used Bay11-7082 which was initially characterized as an inhibitor of IκB kinase activity [232]. This compound prevents the phosphorylation and degradation of

IκBα therefore inhibiting the translocation of NFκB to the nucleus. Yet, it has recently been demonstrated that this compound has off target effects including irreversible inhibition of protein tyrosine phosphatases [233]. So, while we interpret the loss of survival effect in the

CD5hi cells as due to the loss of NFκB activity based on the previous characterization, it does not eliminate the potential for an off-target mediator to be involved in this pathway.

121

Our data provides a novel mechanistic link between the CD5 and NFκB pathways.

Where an increase in CD5 is concurrent with the upregulation of a negative inhibitor of the

NFκB pathway, IκBα. Interestingly, an increase in expression of both these negative regulators provides a survival advantage that is NFκB dependent. This points to a positive regulatory role of CD5 and by extension, IκBα, in thymocyte survival which is counter- intuitive to their canonical inhibitory roles. Importantly, this reveals a new pathway that sets differential thresholds for T cells prior to their emigration from the thymus into the periphery.

122

Chapter 4: Origins of signaling heterogeneity in steady-state peripheral T cells

4.1. Introduction

Once the T cells complete development in the thymus, they migrate to the peripheral lymphoid organs as naïve CD4 or CD8 T cells and await cognate antigen stimulation. At this stage, the reactivity of individual T cells reflects a continuum of their thymic experience as discussed in Chapter 3. As a result, each T cell’s activation threshold, set by the affinity of the TCR for thymic-selecting ligands, continues to be the baseline in the periphery.

There is accumulating evidence suggesting that T cells are aware of self-pMHC complexes within the periphery as well [234]. While these interactions were sufficient to complete selection in the thymus, within the periphery weak signals through the TCR may have functional relevance to T cell responses generated against foreign agonists. As of date, there are varying reports of what exactly the functional relevance of self-pMHC in the periphery may be [133, 164, 235-240] and the current consensus has been reviewed [241-

243]. For example, one functional consequence of self-pMHC signaling to naïve peripheral

T cells is the maintenance of phosphorylated zeta chains as evidenced by the loss of these post-translational modifications upon transfer into MHC class II deficient hosts [235].

Many of the reports seem to demonstrate a positive role of self pMHC in the peripheral compartment, particularly through poising the TCR signaling pathway to respond better upon engagement with foreign agonists. For example, transfer of T cells into MHC-KO hosts has been shown to decrease the activity of small GTPases, Rap1 and Rac1 but with strong stimulation through CD3 crosslinking, the cells are still able to respond [240]. When considering these observed changes and their effects on peripheral activation thresholds, it

123 is important to note the signaling changes that are made with or without pMHC. The presence of pMHC seems to continually provide low level stimulation which keeps the

TCR signaling machinery primed for responsiveness, as evidenced by the phosphorylation of the zeta chains in the presence of pMHC interactions.

Woven into these studies, is the use of CD5 as a surrogate for measuring TCR self- reactivity [244]. As discussed in Chapter 3, CD5 is upregulated proportional to the TCR signal strength in the thymus and that expression is maintained into the periphery [214].

As such, this has used as a marker to measure T cell self pMHC reactivity via surface receptor expression of CD5 which has made possible many adoptive transfer experiments through flow assisted cell sorting (FACS). Additionally, Nur77 is used as another marker for self-reactivity [245] but, unlike CD5, is an intracellular molecule so it does not provide the same advantages for subsetting T cells for FACS unless used as a reporter system [246,

247]. As such, it has been accepted that CD5 is a surrogate marker for self pMHC signal strength during thymic selection.

Multiple labs have shown that high CD5 expression on T cells, and by extension the population of more self-reactive T cells, respond better to a variety of pathogens compared to T cells with lower CD5 expression in both CD8 and CD4 T cells [248-251]. Specifically, one study finds CD8 CD5hi cells accumulate in higher numbers after Listeria monocytogenes infection by day 7 and persist at higher numbers through day 30 [248]. This suggests that high self-reactivity of T cells, as denoted by higher CD5 expression, during thymic selection improves the peripheral responses of T cells to invading pathogens. The continued expression of higher CD5 (and the functional benefit of that) is thought to accrue

124 from the ability of some TCRs to continue sensing their positively selecting ligands in the periphery but, this has not been explicitly tested.

Beyond its associated with self-reactivity, CD5 has been associated with other functional alterations in lymphocytes and was recently reviewed [252]. In human T cells isolated from cancer patients, it was shown that higher expression of CD5 protects clones from activation induced cell death [253]. Similarly, it was found that higher CD5 expression also protects T cells from Fas-mediated cell death programs [254]. Within murine disease models, similar results were obtained since there was decreased severity of induced experimental autoimmune encephalitis due to increased cell death in CD5-KO animals or through CD5 blockade [255]. Taken together, these data demonstrate a stiking role of CD5 in maintaining T cell survival although the exact molecular mechanisms are not well characterized.

Moreover, Paul Allen’s lab has generated two listeria specific T cell clones that vary only in their expression of CD5 [251]. Interestingly, these cells have different responses to bacterial infection, where the CD5lo clone initial responds better but the persistence and secondary responses are dominated by the CD5hi clone [251, 256]. The follow up studies from these findings are reviewed [257] but there are distinct changes in the calcium responses as well as cytokine secretion kinetics [258, 259]. Yet, the exact molecular mechanisms that mediate these cellular responses are not well characterized.

The consequences of CD5 signaling has typically been measured by its impact on TCR proximal adapters, cytokine secretion, and proliferation of T cells [190, 200, 217, 260-

263]. As discussed in Chapter 3, SHP-1 is one of the most well characterized binding partners and signaling mediators downstream of CD5 [194]. There are also links to the

125 mTOR pathway that could alter T cell metabolism and impair the proliferative burst upon antigen encounter [262]. Interestingly, higher expression of CD5 seems to allow faster secretion of IL-2 upon stimulation [246]. Moreover, there are differentiation effects seen between expression of CD5 particularly with the balance between Th17 and T regulatory cells with a role for casein kinase II in this process as well [223, 262, 264]. There is also recent evidence demonstrating the binding of IL-6 to CD5 which then triggers STAT3 signaling within the cell [190]. Taken together, this suggests that there is varied functional consequences of CD5 in peripheral T cells.

Taken together, the field has coalesced around the idea that self-reactivity of individual

TCRs dictates their ability to respond to antigen. Moreover, the current consensus is that heterogeneity between individual T cells is directly a reflection on TCR affinity for tonic self-antigens as well as the amounts of those antigens in the body. Our findings in Chapter

3, where CD5 itself directly controls the expression of IκBα as well as the survival advantage of CD5hi thymocytes, suggested an alternate hypothesis. Contrary to the prevailing idea we suggest that CD5 dictates the signaling thresholds – at least of the NFκB pathway – without additional help from the TCR. In this chapter we evaluated this hypothesis using peripheral T cells.

4.2. Chapter Summary

We found that, similar to the data in the thymus, CD5 and IκBα are tightly correlated in peripheral T cell populations. CD5 expression was continually required for maintaining IκBα in the periphery rather than only a reflection of thymic commitment.

Mechanistically, CD5 regulated IκBα levels post-transcriptionally and independently of canonical regulators. Antigen stimulation upregulated CD5 and IκBα suggesting a two-

126 step model wherein TCR signals regulate transcript levels of IκBα whereas CD5 controls the post-transcriptional regulation. We found that CD5 is sufficient to upregulate IκBα in a TCR deficient lymphoma line but not in a fibroblast line which suggests the role of a lymphocytic signaling molecule to mediate this correlation. Finally, to address the potential activity of NFκB, we found enhanced NFκB accumulation in cells with higher CD5 and

IκBα expression. Taken together, these data create a picture where the expression of two canonical regulators, CD5 and IκBα, create a depot of a critical transcription factor, NFκB, that may imbue high expressor cells with advantageous responses upon antigenic stimulation.

4.3. Results

4.3.1. Gating strategy for analyzing peripheral T cell populations.

In order to analyze peripheral T cells, we used a flow cytometric approach and staining with TCRβ, CD4, CD8α, CD44, and CD25. CD4 and CD8 T cells were defined as

TCRβ+ and positivity for their respective co-receptor (Figure 4.1). Furthermore, from these subsets, we gated by CD44 expression to designate antigen experienced cells as represented by CD44hi cells within naïve animals (Figure 4.2). Within CD4 subsets, T regulatory cells can be defined by a variety of transcription factor but, here we used CD25hi as a surrogate marker (Figure 4.1) acknowledging that it was not a pure T regulatory population for these analyses since CD25 is also an activation marker on T cells. These gating strategies were used throughout the remainder of this Chapter for all subsequent experiments.

127

Figure 4.1. Peripheral T cell subset gating.

Lymph nodes and/or spleens were isolated from C57BL/6 mice and analyzed by flow cytometry. Cells were stained with 7AAD, TCRβ, CD4, CD8α, CD25, and CD44 to subset populations. CD4 and CD8 populations were gated as 7AAD-, TCRβ+, and either CD4+ or CD8+ respectively. Further subsets were generated based on CD44hi expression as ‘memory phenotype’ T cells. Within the CD4 population, CD25hi was used as a surrogate marker to enrich for T regulatory cells.

128

4.3.2. Peripheral T cells have altered signaling molecule expression compared to their thymic counterparts

We first investigated the expression of TCR signaling molecules to determine if there were any basal variation among a polyclonal population of T cells that may reflect the tonic signaling from self-peptide MHC engagement (Figure 4.2). As discussed in

Chapter 1, TCR signal transduction is initiated by the activity of the CD4/CD8 co-receptor associated kinase Lck, which phosphorylates immunoreceptor tyrosine-based activating motifs (ITAMS) on the CD3ζ chain, resulting in recruitment of the kinase Zap70 and subsequent signaling through LAT, ERK, PLCγ, p38 etc. Typically, the TCR proximal signaling machinery converges on key transcription factor complexes which activate gene expression. Of these, the NFκB pathway involves a tightly regulated cytoplasm to nucleus transit of the transcription factor (TF) complex. Typically, the dimeric TFs of NFκB p50 or p65 are complexed with the inhibitor, IκB – which masks their nuclear localization signals and retains them in the cytoplasm. Upon TCR signaling, proximal kinases phosphorylate IκB and mark it for ubiquitin tagging and degradation, allowing the cytoplasmic pool of NFκB TF to translocate to the nucleus and drive gene expression. We found that there was tight expression of Zap70 and LAT but variable expression of PLCγ and IκBα (Figure 4.2).

129

Figure 4.2. Expression of TCR signaling molecules in peripheral T cells.

Lymph nodes were isolated from wild-type mice and analyzed via flow cytometry for the expression of Zap70, LAT, PLCγ, and IκBα in CD4 and CD8 T cells using the gating strategy in Figure 4.1.

130

The functional differences in peripheral T cells with varying levels of CD5 from thymic development signals or from tonic signals in the periphery have been described

[249, 265-268], but the underlying molecular mechanisms are not fully understood. This is especially true in the context of reports that CD5hi T cells respond better to foreign antigens

[248, 249, 268], although CD5 is reported to be a negative regulator of TCR signaling.

Moreover, Previous studies have examined gene expression changes between CD5hi and

CD5lo cells as well as direct consequences on TCR signaling [268]. We reasoned that any basal differences in the signaling capacity of these cells might also be discerned by analyzing the protein expression of key mediators of signaling in T cells, in relation to the levels of CD5 (Figure 4.2 and 4.3).

131

Figure 4.3. Zap70, LAT, and PLCγ expression are not correlated with CD5 in peripheral T cells.

Lymph nodes were isolated and stained for flow cytometric analysis for expression of CD5, Zap70, LAT and PLCγ in CD4 and CD8 T cell populations as gated in Figure 4.1. Representative flow plots of Zap70 (A), LAT (B), and PLCγ (C) in either CD4 (left panel) or CD8 (right panel) T cells are shown.

132

Since peripheral T cells express a continuum of CD5 ([265], also see Y-axis in

Figure 4.3) we first examined whether higher CD5 expression correlated with higher or lower expression of any TCR signaling intermediates. There was no apparent correlation between CD5 levels to the total intracellular levels of Zap70, PLCγ or LAT (Figure 4.3).

We found that the expression levels of IκBα in T cells consistently showed a broad distribution similar to the distribution of CD5 in CD4 and CD8 T cells (Figure 4.4A and

4.4B) similar to the relationship that we observed in the thymus as discussed in Chapter 3.

Furthermore, as shown in Figure 4.4B, peripheral CD4 T cells typically expressed a higher level of CD5 (mean fluorescence intensity (MFI) of 10631 ± 3025) than CD8s (gMFI of

5953 ± 2252) and the IκBα levels showed a similar trend (gMFI of 4709 ± 1041 in CD4s and gMFI of 2155 ± 364 in CD8s) as seen in Figure 4.4B. This striking correlation between

CD5 and IκBα expression was not shared with other surface receptors such as CD127,

CD44, CD62L, or TCRβ (Figure 4.5) or signaling molecules as discussed above (Figure

4.3).

133

Figure 4.4. CD5 and IκBα expression are tightly correlated in peripheral T cell populations.

Lymph nodes and spleens were isolated and stained for flow cytometric analysis for expression of CD5 and IκBα in peripheral T cell populations as gated in Figure 4.1. (A) Representative flow plot of CD5 and IκBα expression in CD4 (left) and CD8 (right) T cells. (B) The geometric mean fluorescence intensity (gMFI) of CD5 and IκBα in CD4 and CD8 T cells (n= 11 from 3 independent experiments. Statistical significance was calculated using a paired t-test.) (C) Representative flow plot of CD5 and IκBα expression in CD25lo and CD25hi CD4 T cells (n= 7 from 3 independent experiments. Statistical significance was calculated using paired t tests). (D) The gMFI of CD5 and IκBα in CD25lo and CD25hi CD4 T cells. (E) Representative flow plots of CD5 and IκBα in CD44lo (left column) or CD44hi (right column) CD4 (top row) and CD8 (bottom row). (F) The gMFI of CD5 and IκBα in CD44 subsets of CD4 and CD8 T cells (n = 11 from 3 independent experiments. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

134

Figure 4.5. IκBα expression does not correlate with other T cell surface molecules.

(A) Lymph nodes were isolated and stained for flow cytometric analysis for expression of CD44, CD127, CD62L, and TCRβ in either CD4 or CD8 T cells as gated in Figure 4.1. (B) Lymph nodes were isolated and stained for flow cytometric analysis for expression of CD5 and CD44 in either CD4 or CD8 T cells as gated in Figure 4.1.

135

Typically, T cells that complete thymic development and emigrate to the periphery maintain the levels of CD5 they acquired in the thymus [214]. Tregs – which also have highly self-reactive TCRs - have been shown to express high levels of CD5 [269, 270]. We also found higher IκBα expression in the peripheral CD25hi population of CD4+ T cells compared to the CD25lo population (Figure 4.4C and 4.4D).

The overall levels of CD5 were higher in the discrete memory T cell population identified as CD44hi T cells (Figure 4.4E and 4.4F); but the distribution of CD44 itself did not follow a correlation with either CD5 or IκBα (Figure 4.5). Peripheral “memory” phenotype T cells in a polyclonal C57BL/6 repertoire (identified as CD44hi) also maintained higher CD5 expression (Figure 4.4E and 4.4F). Similarly, the basal IκBα expression in peripheral CD44hi CD4 and CD8 T cells increased correspondingly (Figure

4.4E and 4.4F).

4.3.3. CD5, not TCR signals, is necessary for the continued upregulation of IκBα in the periphery and is independent of canonical regulators

As discussed in Chapter 3, CD5 was essential for the developmental upregulation of IκBα expression. We were interested in understanding how this relationship is set either at the developmental phase in the thymus or if the continuous expression of CD5 in the periphery was necessary for IκBα expression. As expected, IκBα expression was decreased in both the peripheral CD4 and CD8 T cells of CD5-KO mice, regardless of the expression of CD44 (Figure 4.6). This suggested that IκBα upregulation during antigen encounter required the presence of CD5 but, could alternatively reflect the loss of this upregulation step within the thymus.

136

Figure 4.6. Genomic ablation of CD5 decreases expression of IκBα in peripheral T cells.

(A) CD4 or CD8 T cells as gated in Figure 4.1 from the pooled lymph nodes of WT or CD5 knock-out (CD5- KO) mice were analyzed for CD5 and IκBα expression and a representative set of flow plots are shown. (B) The gMFI of CD5 and IκBα gMFI in peripheral CD4 or CD8 T cells gated as in Figure 4.1 from lymph nodes (n= 3 - 5 from 1 experiment. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

137

To dissect the differential roles of CD5 in the thymus and periphery, we next examined the peripheral population of T cells in the CD5-iKO mice. Although the ablation of CD5 will impact the recent thymic emigrants within the periphery, this composes less than 25% of the total peripheral population within the window of analysis and therefore should not significantly impact selection and the analysis of IκBα expression [271]. First, we found that within 5 days of tamoxifen administration the entire peripheral compartment of both CD4 and CD8 T cells lacked CD5 expression (Figure 4.7A). Interestingly, we found that the loss of IκBα expression was significant in the CD44hi CD4 T cells, but not within the CD44lo CD4 T cells (Figure 4.7B and 4.7C). Although CD8 T cells also trended similarly, the levels of IκBα in the WT versus CD5-iKO in these cells were not statistically significant at the time frame of CD5 ablation or with the number of mice used here.

138

Figure 4.7. Peripheral ablation of CD5 decreases expression of IκBα in T cells.

(A) Representative flow plots of peripheral T cells isolated from lymph nodes of CD5-floxed mice with CAG-CreERT2 transgene (CD5 inducible knock-outs; CD5-iKO) after treatment with tamoxifen for 5 days. Histograms of CD5 expression are shown. (B) Frequency and absolute number of peripheral cell subsets. (C) The expression of CD5 and IκBα in peripheral T cells in either WT or CD5-iKO mice after treatment with tamoxifen for 5 days. Representative flow plots are shown. (D) CD5 and IκBα of peripheral CD4 or CD8 T cells gated as in Figure 4.1 above from lymph nodes of WT or CD5-iKO mice (n = 2 - 5 from 1 experiment. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

139

Since we have unveiled a novel connection between the surface molecule CD5 and an intracellular component of the NFκB pathway, we next investigated the molecular mechanisms that maintain the regulatory relationship between CD5 and IκBα in peripheral

T cell compartments. Although recent reports hint at “non-canonical” roles for CD5 [190], it is typically thought to act by negatively regulating the TCR-proximal signaling pathway as discussed in section 3.1.3. The intracellular tail of CD5 recruits negative regulators, most prominently the phosphatase SHP-1 to TCR signaling complexes [194]. We therefore examined the levels of CD5 and IκBα in SHP-1 knockout (SHP-1-KO) mice as well.

Similarly, the adapter protein THEMIS is a critical integrator of TCR signaling, especially in the thymus. More as a matter of opportunity, since THEMIS-KO cells were available from our collaborator Dr. Paul Love, we examined CD5 and IκBα in THEMIS-KO mice as well (Figure 4.8). Intriguingly, the proportional relationship between CD5 and IκBα expression was maintained despite increased expression of both CD5 and IκBα in the SHP-

1-KO T cells and conversely, decreased in expression in the THEMIS-KO T cells (Figure

4.8). The two populations in these plots were likely from contaminating cells since this particular flow panel did not have a TCRβ gating strategy due to technical constraints.

140

Figure 4.8. The CD5-IκBα regulation is maintained independently of SHP-1 and THEMIS.

(A) Flow cytometry analysis of wild-type (WT), SHP-1-KO, and THEMIS-KO splenocytes (n = 1 from 1 independent experiment). (B) CD5 and IκBα gMFI in WT, SHP-1-KO, and THEMIS-KO splenocytes (n = 1 from 1 independent experiment).

141

A second molecule involved in CD5 signaling is casein kinase II (CKII). CKII binds to the cytoplasmic tail of CD5 influences T cell function and differentiation and has also been shown to modulate IκBα activation [200, 211, 272]. Treatment with a well- characterized CKII inhibitor CX4945 [222, 223] also did not affect the levels of CD5 or

IκBα in peripheral T cells in a dose-dependent manner (Figure 4.9). Taken together this suggests that the regulation between CD5 and IκBα is maintained independently of SHP-

1, THEMIS, and CKII kinase activity.

142

Figure 4.9. Casein kinase II activity is not required to maintain CD5-dependent IκBα expression.

Pooled lymph node and spleens were homogenized into a single cell suspension and cultured ex vivo with varying concentrations of CX4945 and harvested at the indicated time points. Cells were analyzed for expression of CD5 and IκBα within CD4 (A) and CD8 (B) T cell populations as gated in Figure 4.1 via flow cytometry (n = 3 from 1 independent experiment).

143

4.3.4. CD5 upregulation of IκBα acts at a post-transcriptional level of regulation and involves a saturable mechanism

Since we were unable to find any direct links from canonical binding partners of

CD5, we examined if CD5hi cells expressed more Nfkbia mRNA, the gene that encodes the

IκBα protein (Figure 4.10). Cells were sorted as gated in Figure 4.10A and the expression of CD5 did not seem to reflect higher expression of Nfkbia transcript by qRT-PCR (Figure

4.10B, left panel). Furthermore, additional analysis of transcripts for NFκB subunits by qRT-PCR did not show significant differences between CD5hi and CD5lo T cells (Figure

4.10B middle and right panels). To validate our flow cytometry findings, western blot analysis was performed on FACS sorted CD5hi and CD5lo cells and immunoblotted for expression of CD5 and IκBα. We find that similar to our flow cytometric analysis, there was increased expression of IκBα in the CD5hi sorted populations (Figure 4.10C). Taken together, these findings suggest that the relationship between CD5 and IκBα is likely maintained at a post-transcriptional level.

144

Figure 4.10. The CD5-IκBα axis is maintained post-transcriptionally.

(A) Peripheral T cells from wild-type animals were sorted on 20% CD5hi and CD5lo expression as indicated. Representative flow plots are shown. (B) Sorted cells were lysed and RNA was harvested for reverse transcription. cDNA was analyzed for expression of Actb and Nfkbia between groups. Fold change (2^ΔΔCT) of CD5hi cells compared to CD5lo cells of either CD4 or CD8 populations normalized to expression of Actb (n = 6 from 2 independent experiments). (C) Western blot analysis from B6 lymphocytes sorted on 20% CD5hi and CD5lo within CD4+ and CD8+ T cell (TCRβ+) populations. Representative western blot is shown (n = 3 from 3 independent experiments). Western blot, courtesy Seeyoung Choi.

145

To further interrogate transcriptional differences in CD5lo and CD5hi cells to potentially identify factors that may mediate the CD5- IκBα axis, we analyzed an RNA- sequencing (RNA-seq) dataset of FACS-sorted CD5lo and CD5hi T cells (both CD4 and

CD8). We first generated heat maps and performed a principal component analysis between

CD5lo and CD5hi T cells and found there is distinct expression differences between these populations (Figure 4.11). Moreover, we constructed volcano plots of the differentially expressed genes in CD5lo and CD5hi T cells (Figure 4.12). We found identical changes in the genes that were upregulated between CD4 and CD8 T cells (Table 4.1) but, the downregulated genes vary between the two T cell subsets (Table 4.2). Similar analysis has been performed by others in CD8 T cells sorted by CD5 expression and shares some overlapping gene profiles from our analysis [248]. To specifically investigate any expression alterations in the NFκB pathway, we queried the expression of genes in this pathway. We found that none of the genes in the pathway, including IκBα, had differential expression based on CD5 surface expression (Table 4.3). Taken together, these data suggest that the regulation of IκBα protein levels in correspondence to CD5 levels involves posttranscriptional mechanisms that identifies a novel regulatory axis between CD5 and

IκBα.

146

Figure 4.11. RNA-Seq analysis of CD5lo and CD5hi T cells.

CD4 and CD8 T cells were sorted based on CD5 expression, RNA was collected, and sequenced for analysis. (A) Unsupervised hierarchical clustering of genes significantly different between each group.

147

Figure 4.11. RNA-Seq analysis of CD5lo and CD5hi T cells (cont.). CD4 and CD8 T cells were sorted based on CD5 expression, RNA was collected, and sequenced for analysis. (B) Principal component analysis of RNAseq.

148

Figure 4.12. RNA-Seq analysis of CD5lo and CD5hi T cells.

CD4 and CD8 T cells were sorted on based on CD5 expression, RNA was collected, and sequenced for analysis. Volcano plots of the top 100 differentially expressed genes (p<0.001) for CD4 (A) and CD8 (B) samples.

149

Figure 4.12. RNA-Seq analysis of CD5lo and CD5hi T cells (cont.). CD4 and CD8 T cells were sorted on based on CD5 expression, RNA was collected, and sequenced for analysis. Volcano plots of the top 100 differentially expressed genes (p<0.001) for CD4 (A) and CD8 (B) samples.

150

Table 4.1. Upregulated genes in CD5lo and CD5hi T cells.

Upregulated CD4 Upregulated CD8 A530032D15Rik A530032D15Rik Glrp1 Glrp1 Hmcn1 Hmcn1 Arfgef3 Arfgef3 Adamts14 Adamts14 Lyz2 Lyz2 Grm6 Grm6 Trim80 Trim80 Ly86 Ly86 Pcsk1 Pcsk1 Gzma Gzma Lrrc3b Lrrc3b Ly6c1 Ly6c1 Prss30 Prss30 Nrg2 Nrg2 Dntt Dntt Lamc3 Lamc3 S100a8 S100a8 Plac8 Plac8 Gprin3 Gprin3 Cd163 Cd163 Nkg7 Nkg7 1-Mar.1 1-Mar.1 Cilp Cilp

151

Table 4.2. Downregulated genes in CD5lo and CD5hi T cells.

Downregulated CD8 Downregulated CD4 Gm4956 Gm4956+G46D1F2:F31F2:F63 Kcnq5 Nrp2 Steap3 Map2 Kif14 Ikzf2 Aspm Fam124b Dpt Lad1 Xcl1 Rgs16 Nuf2 Gm20743 Armc2 Atp1a2 Cdk1 Phactr2 Ntn4 Ntn4 E2f7 9530020I12Rik Fignl1 Msgn1 Hmmr Kcnf1 Aurkb 9430078K24Rik Fam64a Mctp1 Spag5 Gm10406 Tbx21 Gm3696 Tbkbp1 4930431P03Rik Cdc6 Gjb2 BC030867 Cldn10 Kif18b Prlr Tcam1 Apol9b Kcnf1 Scn8a Lrr1 Litaf Hist1h3c Slc22a2 F13a1 Pstpip2 Mxd3 Cd5 Crhbp A430093F15Rik 4930431P03Rik Plce1 Gjb2 Itih5 Pbk C1ql3 Pdgfb Nebl Cenpm Fam163b Espl1 Dbhos Gpr15 Fibcd1 Cdc25c Acvr1c Ska1 Ccdc148 A430093F15Rik Prr5l Kif11 D430041D05Rik Entpd1 Pla2g4f

152

Table 4.2. Downregulated genes in CD5lo and CD5hi T cells (cont.).

Mcm10 Il2 Itih5 Fgf2 Nebl Gucy1b3 Adamtsl2 Gucy1a3 Fam163b Slc9b2 1700026L06Rik Ccdc180 Ak8 Coro2a Fibcd1 D630039A03Rik Ccdc148 Clcnkb Bub1b Dhrs3 Rad51 Tnfrsf9 Nusap1 Kit Bub1 Syna Ckap2l Slc29a4 Tpx2 Chn2 Ube2c Chl1 Bmp7 Cecr6 Ccna2 Lag3 Mnd1 Bcat1 Nes Trpm1 Iqgap3 Sh3gl3 Slc9b2 Map6 Depdc1a Tubb3 Rad54b Izumo1r Melk Nrgn Coro2a Arhgap20 Stil Kdelc2 Rad54l Gsta4 Kif2c Nt5e Stmn1 Itga9 Ncapg Cdcp1

Kit

Cit

Kntc1

Slc29a4

Wdr95

Osbpl3

Chn2

Chl1

Srgap3

Cecr6

Lag3

153

Figure 4.2. Downregulated genes in CD5lo and CD5hi T cells (cont.).

Rad51ap1

Klra3

Klra1

Bcat1

Ccdc155

E2f8

Mki67

Shcbp1

Neil3

Gins2

Chek1

Nrgn

Arhgap20

Kdelc2

2810417H13Rik

Ttk

Kif15

Kif4

154

Table 4.3. Differential expression of NFκB pathway genes in CD5lo and CD5hi T cells.

155

To begin to unravel this novel regulatory pathway, we wanted to address whether further synthetic increases in CD5 levels would also increase IκBα expression by using a

CD5 transgenic mouse model [144]. In these mice, CD5 levels in peripheral T cells increased by less than two-fold (1.52-fold increase in CD4 T cells; 1.39-fold increase in

CD8 T cells) (Figure 4.13A and 4.13B). Accordingly, a subtle shift in IκBα (gMFI of 4734

± 867 in CD5-Tg and 3412 ± 452 in WT) in CD4 T cells, was also observed but did not approach significance (Figure 4.13A and 4.13B). Importantly, although CD5 overexpression in the transgenic mice did increase, this was relatively small compared to the normal expression range. For instance, the gMFI of CD5 in peripheral T cells range from 1060 to 32150 in CD4 T cells and 1211 to 14873 in CD8 T cells, while in CD5-Tg mice this was 4193 to 37951 and 3202 to 19213, respectively. Intriguingly, on further examination of the data, we observed that the maximum value of IκBα expression between

WT and CD5-Tg mice was similar (Figure 4.13A, grey dashed line) but, the average of

IκBα expression trended higher in CD5-Tg cells compared to WT cells (Figure 4.13B). To further elaborate this, we examined the frequency of the population of T cells in 7 bins based on equivalent CD5 levels in both WT and CD5-Tg mice (Figure 4.13C). Plotting of the frequency of T cells in each bin clearly demonstrated slight skewing, with significantly less of the population of CD5-Tg T cells in the lower bins (Figure 4.13D, Bin 4) and increased populations in the higher bins (Figure 4.13D, Bin 6). This suggests that within resting (steady state, non-antigen-activated) T cells, the levels of IκBα are limited to a point of saturation despite the enforced, marginal increase in CD5 expression through transgenic overexpression.

156

Figure 4.13. CD5-driven IκBα expression has a saturation level.

(A) Wild-type (WT) and CD5 transgenic (CD5-Tg) pooled lymph nodes were harvested and analyzed for expression of CD5 and IκBα. Representative flow plots are shown. (B) CD5 and IκBα gMFI on peripheral CD4 and CD8 T cells from WT and CD5-Tg animals (n = 3 from 1 independent experiment. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test.) (C) WT and CD5-Tg lymphocytes were harvested from pooled lymph nodes and grouped into bins by CD5 expression as shown. (D) Frequency of population in each bin is displayed for CD4 and CD8 peripheral T cells. (n = 3 from 1 independent experiment. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test.) Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

157

4.3.5. Antigen exposure further increases the expression of CD5 and IκBα

Given that the transgenic overexpression had a very subtle and non-significant increase in IκBα expression and knowing that CD5 is upregulated when peripheral T cells undergo antigen-dependent activation ([258] and Figure 4.14A), we wanted to understand if TCR signaling could upregulate IκBα expression. We first validated the CD5 upregulation after antigenic stimulation with an in vivo activation time course. We adoptively transferred 5C.C7 TCR-tg CD4+ T cells specific for the model antigen, pigeon cytochrome C (PCC), with a subsequent PCC peptide and adjuvant (LPS) challenge and recovered the cells at various time points for analysis of surface CD5 expression (Figure

4.14A). Within 12 hours after challenge, there was a significant upregulation of CD5 on the adoptively transferred 5C.C7 population that continued to increase up to 60 hours after challenge (Figure 4.14A). Importantly, the saturation of IκBα expression as seen in the overexpression mouse model (Figure 4.13) was overcome during T cell activation by antigen – i.e. the physiological increase of CD5 in such cases leads to a corresponding and further increase in IκBα (Figure 4.14B) beyond the saturating limit found in resting T cells.

This reveals another level of regulation for CD5-mediated control of IκBα in peripheral T cells.

158

Figure 4.14. Antigen exposure further upregulates CD5 and IκBα expression in peripheral T cells.

(A) 5C.C7 TCR-tg CD4+ T cells were adoptively transferred to congenically distinct B10.A mice and challenged with MCC peptide and LPS challenge in B10.A hosts. Expression of CD5 was analyzed at varying time points after challenge on adoptively transferred TCR-tg CD4+ T cells. (n = 2-3 from 1 independent experiment. Statistical significance was calculated using Tukey’s multiple comparisons). (B) Lymphocytes from SMARTA CD4+ TCR-Tg mice were isolated and stimulated ex vivo with TCRαβ-KO APCs and cognate peptide for the indicated time points. Cells were harvested and analyzed by flow cytometry. CD5 and IκBα gMFI of SMARTA CD4+ TCR-Tg T cells (n = 1 from 1 independent experiment). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

159

4.3.6. CD5 alone is sufficient to upregulate IκBα but requires a lymphocytic signaling mediator and a functional cytoplasmic tail

Given the difficulty in effectively separating the roles of tonic TCR signals and

CD5 toward the regulation of IκBα levels (Figure 4.13 and Figure 4.14), we used a cell culture model with the T cell (lymphoma) line BW-5147 that lacks TCR expression and expresses relatively low levels of CD5 natively. Retroviral transduction for overexpression of CD5 in these cells also resulted in a proportional upregulation of IκBα (Fig 4.15).

Interestingly, in a control group using the fibroblastic cell line NIH 3T3, we did not find any consequence to IκBα levels even though high levels of CD5 is expressed (Fig 4.15).

Apart from demonstrating a necessary and sufficient role for CD5 in the control of lymphocytic IκBα, this also suggests the involvement of lineage-specific interactors to mediate the proportional relationship between CD5 and IκBα.

160

Figure 4.15. CD5 upregulation of IκBα is independent of TCR expression and requires lymphocytic- specific molecules.

(A) BW-5147 TCR deficient lymphoma cells or NIH3T3 cells were transduced with empty or CD5 retroviral vector. Cells were rested for 5 d (BW5147) or (48 hours) after transduction and then analyzed by flow cytometry. Representative flow plots are shown. (B) Fold change of CD5 and IκBα gMFI on retrovirally transduced BW-5147 TCR deficient lymphoma cells (n = 5 from 3 independent experiments; statistical significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test for BW5147 experiments and n = 1 from 1 independent experiment for NIH3T3 experiments).

161

A fortuitous finding while generating CD5 overexpression constructs was that the construct in which the CD5 gene is overexpressed matters. We used two different retroviral overexpression plasmids, MIGR1 and pQ2aB (Figure 4.16A and 4.16B). MIGR1 does not have any additional sequence added to the cDNA of CD5 whereas the pQ2aB plasmid adds a small peptide sequence. In transduced BW5147 TCR deficient cells, we found that the pQ2aB plasmid was unable to upregulate the expression of IκBα in contrast to the MIGR1 construct (Figure 4.16C and 4.16D). This suggests that disruption of the cytoplasmic tail, even by a small peptide abrogates the CD5-IκBα axis. Furthermore, this disruption suggests that a binding partner is recruited to the cytoplasmic tail of CD5 to mediate this relationship with IκBα.

162

Figure 4.16. CD5 upregulation of IκBα requires the intracellular tail of CD5.

(A) MIGR1 retroviral vector with CD5 insertion.

163

Figure 4.16. CD5 upregulation of IκBα requires the intracellular tail of CD5 (cont.). (B) pQ2aB retroviral vector with CD5 insertion.

164

Figure 4.16. CD5 upregulation of IκBα requires the intracellular tail of CD5 (cont.). (C) BW-5147 TCR deficient lymphoma cells were transduced with CD5 containing MIGR1 and pQ2aB vectors. Cells were rested for 5 d after transduction and then analyzed by flow cytometry. Representative flow plots are shown. (D) Fold change of CD5 and IκBα gMFI on retrovirally transduced BW-5147 TCR deficient lymphoma cells (n = 5 from 3 independent experiments).

165

4.3.7. CD5 increases the NFκB depot in peripheral T cells

Although IκBα is well characterized as a cytoplasmic protein that binds components of the NFκB transcription factor and prevents them from triggering gene expression in the nucleus, the expression of IκBα is also regulated by NFκB itself [273,

274]. Therefore, it is not trivial to discern if the enhanced IκBα expression would lead to lower expression of NFκB target genes or conversely it simply reflects a greater amount of

NFκB activity within the cell. Given the results from the RNA-Seq data analysis (Figure

4.11 and 4.12 and Table 4.1 – 4.3), the increased expression of IκBα does not seem to be exclusively driven by increased NFκB signaling. To begin to understand the impact of increased IκBα on the NFκB pathway, we first examined the expression patterns of NFκB p65. We found that CD5hi cells had an increase in expression of p65 compared to CD5lo cells in both CD4 and CD8 T cells and, as discussed previously, CD8 T cells have higher expression of NFκB p65 (Figure 4.17A and 4.17B). To resolve whether this implies higher

NFκB activity we used localization of NFκB p65 as a surrogate marker – since active NFκB is translocated to the nucleus.

166

Figure 4.17. CD5 expression alters the localization of NFκB in peripheral T cells.

(A) The surface levels of CD5 and intracellular levels of NFκB for either CD4 or CD8 T cells from pooled lymph nodes as gated in Figure 4.1. Representative flow plots are shown. (B) The gMFI of NFκB expressed by CD4 or CD8 T cells from lymph node or spleen gated in Figure 4.1 (n= 8 from 2 independent experiments. Statistical significance was calculated using a paired t-test. Statistical significance was calculated using two- way ANOVA with Sidak’s multiple comparisons test). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

167

Figure 4.17. CD5 expression alters the localization of NFκB in peripheral T cells (cont.). (C) Screenshot of dashboard from IDEAS analysis.

168

Figure 4.17. CD5 expression alters the localization of NFκB in peripheral T cells (cont.). (D) Quantitation of nuclear and cytoplasmic localization of NFκB from imaging flow cytometric analysis. (n = 8 from 3 independent experiments; statistical significance was calculated using two-way ANOVA with Sidak’s multiple comparisons test). Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

169

Intriguingly, using imaging flow cytometry to determine the localization of the

NFκB p65, we found that the proportion of nuclear-localized NFκB p65 was not significantly different between the CD5lo and CD5hi cells in either CD4 or CD8 T cells despite the higher expression of NFκB p65 in the CD8 CD5hi T cell population (Figure

4.17C and 4.17D). To further validate this, we separated the nuclear and cytoplasmic fractions from sorted CD5hi and CD5lo populations of both CD4 and CD8 T cells and the relative location of p65 was quantified using western blot (Figure 4.18). We found that there was no significant increase in the nuclear fraction of NFκB p65 in CD5hi cells compared to CD5lo populations in both CD4 and CD8 T cells (Figure 4.18). Interestingly, despite the lack of difference in the proportion of nuclear NFκB p65, there was a significant increase in the gMFI of nuclear NFκB in CD5hi CD8 but not, CD5hi CD4 T cells (Figure

4.18). This suggests that the overall increase in NFκB p65 in CD5hi CD8 cells reflects an increase in both the nuclear and cytoplasmic compartments within the cell. Since this was not evident in CD4 T cells, it suggests that there is a specific accumulation of cytoplasmic

NFκB p65 in CD5hi CD4 T cells that accounts for the significant difference in total NFκB p65 in CD5hi CD4 (Figure 4.18). This was further supported by the cellular fractionation and western blot analysis of NFκB p65 in sorted T cells that shows a higher proportion of

NFκB p65 in the cytoplasmic fraction in CD5hi CD4 compared to CD5lo CD4 T cells

(Figure 4.18).

170

Figure 4.18. CD5 expression alters the localization of NFκB in peripheral T cells.

CD4 and CD8 T cells were sorted based on CD5 expression. Cell fractionation was performed to separate cytoplasmic and nuclear components for western blot analysis. Three independent cell sorts were performed and western blot analysis for NFκB p65 was performed (panels A, B, and C). NFκBp65 was quantitated to SP1 or PARP as indicated. (D) Relative values of nuclear localized NFκBp65. Western blots, courtesy of Seeyoung Choi.

171

Taken together these data suggest that increased expression of IκBα in CD5hi T cells may allow more NFκB proteins to be present in CD5hi T cells (and potentially retained in the cytoplasm) providing an advantageous depot of NFκB transcription factors, ready to be translocated to the nucleus upon TCR-driven activation.

4.4. Discussion

CD5 plays a critical role as a rheostat for TCR signaling in T cells, but the molecular mechanisms by which it regulates multiple aspects of T cell function in the periphery are not completely understood. This is especially true in the context of responses to pathogens, where unexpectedly, T cells with higher levels of CD5 (a negative regulator) survive better and form a larger memory compartment than their CD5lo companions [248, 249, 251].

Here, we identify a molecular mechanism that can help clarify such a duality of functions.

We find that similar to the expression in the thymus as discussed in Chapter 3, CD5 regulates the expression of IκBα within the peripheral compartment. The increased expression of IκBα allows for an accumulation of NFκB p65, potentially helping to

‘jumpstart’ the NFκB pathways during T cell activation by antigen. The ease of triggering robust NFκB driven gene expression in these cells can also help explain the enhanced survival of the cells previously observed during peripheral immune responses [248, 249].

Overall, these data suggest an intriguing model for the duality of CD5 function in peripheral T cells – i.e. the role as a negative regulator which also imbues T cells with better reactivity to agonist stimulation. The enhanced expression of IκBα confers a basal inhibitory component to the CD5hi T cells. However, these T cells also express a higher level of NFκB p65, which is potentially restrained in the cytoplasm. This “poised” state of the NFκB pathway in CD5hi T cells may also confer a faster activation of NFκB-dependent

172 gene transcription when it receives strong agonistic TCR signals to facilitate IκBα degradation releasing the depot of NFκB to translocate to the nucleus and initiate gene transcription. Thus, CD5 dependent differences in depots of NFκB may account for heterogeneity and intrinsic biases in subsequent antigen-responses by peripheral T cells with distinct phenotypes in CD4 and CD8 populations.

The identification of CD5 as a rheostat for NFκB function can also help resolve many biological paradoxes in T cell development and activation. CD5 has been implicated in multiple different signaling activities in peripheral T cells. CD5hi cells have been shown to have reduced calcium flux after stimulation via CD3 crosslinking – consistent with its role as an inhibitory molecule - yet generate faster IL-2 secretion after anti-CD3/CD28 stimulation [246]. Interestingly, there is less IFNγ secretion in CD5hi cells after 5 days of

TCR stimulation but, the early kinetic differences as seen with IL-2 have yet to be explored

[275]. While biochemical data support a TCR proximal effect of CD5 (e.g. on TCRζ phosphorylation) it has also been shown to affect phosphorylation of ERK after PMA stimulation [250] – which does not typically require TCRζ phosphorylation. As discussed above, the enhanced survival of CD5hi T cells in the context of an infectious challenge is also not intuitive, given its negative regulatory role. Our data suggest that many of these disparate phenotypes stem from the differential signaling potential of the NFκB pathway in CD5hi compared to CD5lo cells. Since NFκB is involved in regulating the expression of multiple genes in T cells, some of the other altered functions in CD5hi T cells discussed above could stem from the gene products of NFκB signaling that are triggered either at basal state or more rapidly following T cell activation. Indeed, the NFκB pathway can directly account for some of the responses – such as the increased early IL-2 production

173 and increased survival in CD5hi T cells [276]. Interestingly, the expression of CD5 is also maintained on some population of B cells [277] so it is not hard to speculate that similar mechanisms governing NFκB function in B cells may exist.

The duality of the effect of CD5 on IκBα levels is also underscored by the saturation effect we observe (Figure 4.13). In resting or steady state T cells, although IκBα follows a striking correlation with CD5 (Figure 4.4A) in wild type mice, artificially increasing CD5 through transgenic overexpression in the CD5-Tg mice, cannot increase IκBα beyond the maximum value that wildtype T cells normally have (Figure 4.13). The simplest explanation for this is that CD5 only regulates the post-translational levels of IκBα – whereas antigen activation by TCR signals allows a transcriptional upregulation of IκBα

[278]. Indeed, this is validated by the further increase in IκBα in antigen-activated T cells

(Figure 4.14). While the increase was attributed to the increase in antigen, the challenge in these studies did include an LPS adjuvant so it is not possible to dissect the effects from co-stimulation on the increase in CD5 and IκBα with the present studies. It is not yet clear if this additional increase would serve a positive or negative role in T cell activation. While the basal levels of IκBα can help maintain an NFκB depot, the further increase (driven by

TCR rather than CD5) may in fact serve an inhibitory role as it has been commonly studied.

Further, antigen activated T cells, which may rest down from the acute TCR signaling, set

IκBα levels higher than naïve T cells (Figure 4.4) in a CD5-dependent fashion (Figure 4.6 and 4.8). This is critical because the protein levels of NFκB members are known to require post-translational control of their stoichiometry [273, 274]. CD5 could be the mechanism that fixes the IκBα expression gain (initially driven the TCR signals) by stabilizing a larger

NFκB pool in resting memory-phenotype T cells. Whether this increased NFκB depot in

174 memory T cells could drive some of the rapid responses associated with memory function is a significant question that remains to be studied. It is also important to note that NFκB does not necessarily mean positive or negative responses, but rather relies on the unique chromatin landscape in a cell. So, in cells like Tregs the effect of an NFκB depot could be quite contrary to that in memory T cells depending on the genes transcribed given the varied accessibility of genes between these two populations. The precise dynamics of TCR signaling in peripheral T cells and its relationship to heterogeneity in cell fates is still being resolved. A variable and CD5 dependent NFκB depot in such cells can be a key part of this signaling network.

Finally, it is intriguing to find that the cytoplasmic tail is critical for maintenance of the

CD5- IκBα axis as evidence by the loss of the relationship in a cell line overexpression system if protein tags are added to the cytoplasmic portion of the tail (Figure 4.16). This suggests that a protein-binding partner to the intracellular tail is necessary to maintain this relationship. Yet, this binding partner does not seem to be SHP-1, THEMIS, or CKII given that the loss or inhibition of the kinase activity for CKII disrupts the CD5- IκBα relationship (Figure 4.8 and 4.9). Recent immunoprecipitation studies identify interacting partners of the TCR signaling complex diversification including CD5 but, none of the identified partners of CD5 were IκBα [213]. This suggests that there is likely an intermediate between CD5 and IκBα through a novel interaction.

175

Chapter 5: Reversible activation threshold modulation from chronic antigen

stimulation

5.1. Introduction

In the previous chapters we have discussed how the basal NFκB signaling machinery, a critical component of the T cell’s activation threshold, is first set in the thymus and then actively maintained and regulated by CD5 within the periphery. Here, we take a broader vantage on multiple signaling pathways and how they change when a T cell faces a unique stimulation context: persistent antigen stimulation. It is increasingly clear that constant stimulation by either self-antigens, tumors or persistent pathogens, triggers pathways of T cell tolerance in peripheral T cells. While both intrinsic mechanisms

(anergy, exhaustion, etc.) as well as regulatory interactions mediate tolerance, we focus on the former as they are directly contingent on modification of a T cell’s ability to activate.

Indeed, a key readout of a T cell’s intrinsic tolerance is the loss of responsiveness to antigen. Many transcriptional and epigenetic changes associated with such a form of tolerance have been characterized, but key signaling changes that unify multiple different contexts of tolerance are not well understood. In Chapters 5 and 6, we will discuss how the regulation of a T cell’s activation threshold is a key regulator of tolerance to chronic antigen stimulation. Given that self-antigens or persistent foreign antigens result in the loss of T cell responses, similar molecular mechanisms may underly both situations; where T cells alter their activation threshold to reflect the persistence of the antigen and shut down their ability to generate productive responses.

Understanding the ways in which activation thresholds are altered during chronic stimulation may provide insight into how to reverse the hyporesponsive phenotype. If we

176 are able to understand the molecular logic that the T cell uses to decide between responding and not-responding, we can harness that logic to design better therapeutics through modulation of those molecules that mediate the activation threshold of these cells. This is a critical focus of research in the area of immunotherapy for improving T cell function during chronic infection and cancer where the re-invigoration of T cell function is critical for clearance of these threats.

5.1.1. Tolerance

Tolerance is a broad term that encompasses the regulation of T cell function both within the thymus and the periphery to control overt reactions to self-antigens. The concept emerges from studies in tissue transplantation by Medawar, Billingham, and Brent where neonatal exposure to antigen prevented lifelong responses to tissues bearing that antigen

(the ‘tolerogen’) while preserving responses to unrelated targets (third party) [279, 280].

Over the last 70 years, research into the cellular and molecular mechanisms of antigen- specific tolerance have unraveled multiple cellular processes that mesh together to result in the tolerant state. We now know that this happens in two distinct phases - central and peripheral tolerance (Figure 5.1). Central tolerance is enforced during thymic selection by a combination of different processes. The most well understood of which is that T cells that are too highly reactive undergo apoptotic death (deletion) during negative selection [172].

In addition to deletional tolerance, T cells that are autoreactive are also converted to Tregs in the thymus. These cells are marked by the expression of the transcription factor Foxp3,

Tregs are the major subset of T cells that enforce a ‘dominant’ form of tolerance – i.e. they suppress other effector T cell responses to prevent autoreactive T cells from causing pathology. In contrast, cell-intrinsic silencing of T cell functions only limits that particular

177 autoreactive T cell from causing damage and is deemed a ‘recessive’ form of tolerance. It is important to note that Treg suppression can be done on a larger scale; where a single

Treg is able to suppress multiple effector T cells simultaneously. In contrast, intrinsic changes to autoreactive T cells happen on a single cell basis. As T cells complete development in the thymus, dominant (conversion to thymically derived natural Tregs) as well as recessive (deletion and the process of intrinsically tuning T cell activation thresholds, discussed in Chapter 3) forms of controlling autoreactive responses are invoked

[281]. Since this is still an imperfect process and self-reactive T cells are able to emigrate to the periphery, there are mechanisms to continue to limit the function of these potentially deleterious T cells as well. These include, the activity of thymically-derived regulatory T cells (Fig 5.1), new induction of peripheral T regs, peripheral deletion, as well as new changes to the T cell that allow for cell intrinsic alteration of the activation threshold of the

TCR for its self-pMHC such that it renders the cell non-functional. The last of these, involving intrinsic changes to T cell function is the focus of this Chapter. In the literature, these phenomena have been discussed under the umbrella of three terms – anergy, exhaustion and co-inhibition – yet, each of these shares considerable overlap at the cellular level since they all involve a failure of T cells to respond maximally to a new antigenic stimulus. This hyporesponsiveness is acquired and may have shared molecular characteristics in all three cell states as we discuss further below. Nevertheless, it is easy to get the impression from reading the current literature that these recessive mechanisms are secondary to the activity of dominant regulation by Tregs during peripheral tolerance

[282]. This notion, however, as we discuss below, is mostly fueled by an incomplete understanding of peripheral tolerance.

178

Figure 5.1. Overview of central and peripheral T cell tolerance mechanisms.

T cells during development undergo mechanisms that prevent self-reactive T cell clones from entering the periphery. This is achieved through negative selection where if T cells bind too strongly to self-pMHC they undergo apoptosis. This is called central tolerance. Yet, this is an imperfect process and T cells that are self- reactive have been found in the periphery. As such, there are also peripheral tolerance mechanisms such as regulatory T cells (either thymically derived or peripherally induced) that suppress the function of auto- reactive clones, peripheral deletion, or the induction of the anergic state. Figure adapted from [Robbins and Cotran Pathologic Basis of Disease, 2020]

179

5.1.2. The view past Regulatory T cells

The identification of a continuous role Treg activity in the maintenance of peripheral tolerance has perhaps been one of the major breakthroughs in the field over the past two decades. Studies by multiple groups identified subsets of CD4 T cells which constitutively expressed the high affinity IL-2 receptor CD25, as key suppressive cells in the peripheral immune system, whose development and function was controlled by the master transcription factor Foxp3 [283-287]. Soon afterwards, a key experiment was influential to our understanding of peripheral tolerance as we know it today. The Rudensky and Sparwasser groups demonstrated that removing Tregs only, from an otherwise healthy adult mouse, without any other overt triggering event, was sufficient for the development of autoimmunity in mice [288, 289]. Rightfully, this established Tregs as an identifiable and primary cellular mechanism regulating peripheral tolerance. Studies other than those directly evaluating fundamental mechanisms of tolerance have since tended to focus on these cells and the mechanisms by which they act as a surrogate of peripheral tolerance as a whole. Indeed, these days when a publication comes from labs that are further removed from directly investigating mechanisms of peripheral tolerance, oftentimes the enumeration and frequency of regulatory T cells or even the expression of Foxp3 itself, are used as the primary readouts (or even sole indicators) of tolerance. Yet, there are several reasons why dominant immunosuppression by Tregs alone does not offer a complete picture of the tolerance machinery operating in our peripheral immune system.

Critically, the tremendous advances in basic immunology of regulatory T cells come from two singular advances – a fairly reliable cell surface marker, CD25 [283] that allows researchers to identify, isolate, and ablate these cells using antibodies against this

180 receptor as well as the critical molecular mapping of Foxp3 as a fundamental regulator of this group of T cells. After decades of controversy over the nature of ‘suppressor cells’

[282], these molecular tools allowed the Rudensky, Sakaguchi, and other labs to develop multiple approaches to critically eliminate these cells (without many off-target effects) from the system or even introduce them to new animals by adoptive transfer through isolation from fluorescence-activated cell sorting (FACS) using CD25 and other markers for isolating pure populations of these cells. As discussed above, this led them, (for e.g. the

Rudensky group,) to conclude that, (as quoted from [288]) “… Foxp3-independent recessive and dominant tolerance mechanisms established in adult mice are not sufficient to protect mice from fatal autoimmunity after Treg cell elimination..”. This is an accurate statement but, the missing corollary is whether elimination of other (recessive) mechanisms of tolerance would have the same outcomes. So, for instance, if a master transcription factor driving a recessive form of peripheral tolerance through mechanism X was available – if we disable the master transcription factor through a genetic ablation, would the deficiency in mechanism X also behave the same way and allow us to state that in the absence of mechanism X, “… Foxp3-dependent dominant tolerance mechanisms established in adult mice are not sufficient to protect mice from fatal autoimmunity ....” ? Indeed, this is exactly what would be expected, if all forms of peripheral tolerance, whether dominant or recessive, play co-equal roles in cooperatively maintaining peripheral tolerance. The roadblock, however, is that there is currently no known ‘master regulator’ for any other form of tolerance other than Tregs so, this remains to be investigated. A close approximation is data from elimination of other molecules, which also disrupt peripheral tolerance, without directly targeting Tregs. For instance, in the absence of Fas/FasL

181 interactions that mediate deletion of autoreactive clones, autoimmunity develops despite the presence of Tregs [290]. This suggests that other peripheral tolerance mechanisms are also critical for maintenance of a healthy state.

Additionally, hints for the requirement of additional mechanisms towards stable peripheral tolerance are already present in the existing literature analyzing CD25+ and

Foxp3 cells. In early experiments depleting Tregs of adult mice using the anti-CD25 antibody PC61, spontaneous autoimmunity did not result [291]. This was superseded by the diphtheria toxin mediated deletion of Tregs in mice expressing the diphtheria toxin

(DT) receptor (DTR) under the control of the Foxp3 promoter (Foxp3-DTR) [288, 289]. In the absence of a clear understanding of the recessive mechanisms, it has not been possible to examine if DT treatment also disrupts other forms of tolerance at the same time (since even the attenuated toxin can activate DCs and drive up some inflammatory markers).

Further, in these elegant studies, the elimination of Tregs by neonatal DT treatment allowed the mice to survive for 24 days before developing immunopathology, while adults developed it more quickly [288]. Importantly, in a parallel study using a different Foxp3-

DTR construct, elimination of Tregs only led to spontaneous immunopathology in neonates, but not in adults [289]. While these wrinkles do not reduce the significance of

Tregs, it certainly justifies a search for parallel and cooperative mechanisms that contribute to peripheral tolerance mechanisms.

A precise biochemical and molecular blueprint of other forms of tolerance are also likely to clarify the activity of Tregs themselves. For instance, it has long been known that

Tregs display an anergic phenotype [292] but, it is not clear if that is central to their ability to be functionally suppressive as well. Similarly, eliminating co-inhibitory molecules such

182 as CTLA-4 in Tregs also affects suppressive activity [293-295]. For example, ablation of

CTLA-4 expression on Tregs did not prevent their in vitro suppression activity but rather caused an increased immunoregulatory phenotype in the conventional T cell compartment where there was upregulation of IL-10, LAG-3, and PD-1 suggesting that the loss of suppression activity from CTLA-4 on Tregs causes compensatory tolerance mechanisms to be activated [294]. Given these, it is not hard to envision how recessive mechanisms may facilitate the dominant ones and how these are more interconnected than is currently appreciated.

5.1.3. Recessive forms of tolerance: Anergy, Exhaustion and Deletion

Anergy was first coined by Dr. Gustav Nossal to describe the loss of response in B cells and this eventually became a universally accepted term for the loss of response in any lymphocyte including T cells [296]. Although R.H. Schwartz, the co-discoverer of anergy in T cells with M.K. Jenkins, advocated the use of the term clonal anergy mostly for in vitro derived anergic cells and a new term ‘adaptive tolerance’ for in vivo anergized T cells

[297], T cell anergy is currently used to refer to both. Clonal anergy was first discovered when Jenkins and Schwartz showed that in the absence of active co-stimulation, APCs that only present antigen to a T cell, would induce biochemical changes in the T cell, such that any subsequent stimulation failed to trigger robust proliferation or cytokine production by the same T cell [298]. Several subsequent studies elaborated on this process and led to the appreciation that T cells require two separate signals (signal 1 via the TCR and signal 2 via the CD28 pathway) for full activation (Figure 5.2, Copyright Clearance Center

#4882690225672). Intriguingly, as the molecular paradigms intertwining the two signals were better described, it was also appreciated that the nature of the TCR signals received

183 by a T cell also plays a key role in the induction of anergy [299-301]. Overall, the a consensus emerged that a partial/incomplete activation signal due to absence of co- stimulation or weak signals from altered peptides or low levels of antigen, could trigger anergy in T cells [302].

184

Figure 5.2. Two-signal model of T cell activation.

T cell activation is typically thought of in the context of two main signals. The first of which is the interaction between the TCR and the pMHC complex on the APC. The second signal is co-stimulation, typically through B7 and CD28 interactions but there are also other molecules such as CD40 and CD40L that are used in other contexts. A final and third signal that helps to differentiate CD4 T cells in particular is cytokine secretion from APCs. Figure adapted from [303].

185

But, even in its early iterations, anergy was not associated with a complete loss of functional responsiveness. Instead, T cells which were anergized (i.e. stimulated once without co-stimulation) would subsequently show a gradation in their ability to make different cytokines – e.g. they could make some amount of IFN but not much IL-2 at all

[304, 305] as such, the loss of IL-2 production became a key signature of the anergic state in T cells [306]. Studies also established that while anergy was very stable in culture, even in the absence of continued stimulation, these cells could be “reversed” by culture with exogenous IL-2 [304, 307]. Finally, biochemical analysis of TCR signaling in anergic cells led to the appreciation that the Ras-MAPK pathway and AP-1 activation (necessary for transcription of the IL-2 gene, Figure 5.3) were impaired in anergic T cells [308-312].

186

Figure 5.3. IL-2 promoter requires multiple transcription factors for effective transcription.

IL-2 is a critical cytokine for T cell activation. Interestingly, the IL-2 promoter has binding sites for many different transcription factors that are required for successful transcription illustrating the importance of proper activation along multiple signaling pathways for induction of IL-2 during T cell activation. Figure adapted from [313].

187

Experiments to extend studies on anergy to in vivo models used a variety of experimental setups – immunization with altered peptide ligands, superantigens, constitutive expression of self-proteins as well as other model systems [314-320]. In many of these cases, these resulted in T cells that entered states of anergy marked by poor proliferation and IL-2 production but, they also revealed differences from the strict parameters of the in vitro Jenkins-Schwartz paradigm of clonal anergy. For instance, the

Rocha lab showed that anergy was reversible if the T cells were removed from chronic stimulation and ‘parked’ in a new animal without antigen – leading them to suggest that exhaustion (induced by persistent low level antigen) and anergy (high antigen stimulation) are distinct processes [321, 322]. A critical difference in the reversal of T cells from this in vivo form of tolerance was that it was independent of IL-2, but rather relied on the removal of continuous antigen stimulation. Using a similar approach (which we discuss further below) the Schwartz group showed that CD4 T cells could fluidly shift between tolerant, non-tolerant and more tolerant states – depending on the presence, absence or increased presence of peripheral antigen [323]. This state was coined (primarily by the Schwartz group) as adaptive tolerance. Overall, these multiple phenotypes associated with anergy in vivo, may also reflect some of the heterogeneity observed in the in vitro models – related to differences in the TCR ligand quality and quantity.

In studies from the Rocha lab, where they discussed anergy, they also considered a distinct form of tolerance called exhaustion [322]. This term owes its origins to work from the Zinkernagel and Hengartner groups where they studied how persistent infection with

LCMV led to the loss of effector functions against the virus [324]. Both the Rocha and

Zinkernagel groups referred to ‘exhaustion’ to imply a gradual death of effector T cells,

188 such that CTLs capable of mediating immune functions were no longer present in the animal (i.e. literally, the immune system was exhausted out of the pool of available CTLs).

Both groups also discussed that there can be gradations of exhaustion based on the number of CTLs that remained in vivo. Therefore, the effort by the Rocha group to define anergy and exhaustion as two different physiological processes of tolerance was in effect a separation of anergy (mechanisms of recessive tolerance, that were non-deletional) from slow elimination of cells (mechanisms of recessive tolerance, that were ‘eventually’ deletional in nature). Further studies, once tetramers were available to track LCMV- specific T cells directly these distinctions were not so easily discernable.

Work from the Rafi Ahmed and Zinkernagel groups showed that during the T cell response to chronic LCMV infection a gradual loss of response in the LCMV-specific

CTL’s effector functions preceded any loss of CTLs themselves [325, 326]. In addition, these poorly functional T cells could persist in vivo for weeks, without necessarily undergoing clonal deletion unlike the previous models discussed above. The loss of function was progressive - with a relatively rapid loss of IL-2 production, while loss of cytokine production such as TNFα or IFNγ from T cells took longer to extinguish and proceeded in a step-wise fashion [327]. It is also noteworthy that even though CTLs became progressively poor functioning, they were still capable of clearing virus in different tissues [324]. In essence, this metamorphosis of the definition of exhaustion was not fundamentally different from that of anergy. In fact the disconnect between growth promoting cytokines (IL-2) and effector function (e.g. secretion of IFN or B cell helper functions) was also appreciated early in the history of anergy – although sometimes called

‘split-anergy’ for CTLs [328]. Over time, the broad swath of tolerant (hyporesponsive) T

189 cell states have now been collectively referred to as exhaustion. While a rudimentary distinction from anergy is sometimes made, it is no longer clear if all cases of exhaustion or anergy are truly distinct at the molecular level. Perhaps, a simple discriminator is the need for persistent antigenic stimulation to induce exhaustion, but perhaps not anergy. Yet, as we discuss in 5.1.4, this is also not a clear distinction now either.

In its currently defined state, exhaustion is characterized by an accumulation of inhibitory receptors (e.g. PD-1, CTLA-4, LAG-3, etc) and persistent antigen stimulation as reviewed [329-333]. Additionally, lack of CD4 T cell help [334] or a variety of other extrinsic mechanisms can influence the development of CD8 T cell exhaustion (reviewed in [330]). Ultimately, this results in functional and molecular changes to these cells that separate them from classical memory differentiation. CD8 memory cells can survive in vivo in the absence of antigen using IL-7 and IL-15, but chronic stimulation causes an

‘addiction’ to antigen so that exhausted T cells cannot survive without the persistence of cognate antigen [335]. In tumor-specific T cells, constant exposure to the tumor antigen is also thought to drive an exhaustion/anergy phenotype; but in this case, an additional term

‘dysfunction’ has also been coined [336] and there is disagreement within the field what this terms exactly refers to [337]. In light of the above discussion, perhaps the use of the term ‘tolerance’ (or more aptly non-deletional tolerance) is an appropriate umbrella to consider all these overlapping mechanisms, until we can parse them out based on discrete molecular operators. At this time, however, major outstanding questions related to the physiological, molecular, phenotypic, and functional consequences for T cells exposed to chronic and acute antigenic stimulation can be seamlessly transferred to each of these terms

(Figure 5.4, Copyright Clearance Center #4882690521166).

190

Figure 5.4. Outstanding questions related to T cell functional alterations following chronic stimulation.

The differentiation of T cells from naïve to memory or exhausted T cells has many unanswered questions. First, defining the differences between memory and exhausted precursors cells remains to be completely understood but, unraveling the differences between these two programs may provide insight into potential therapeutic avenues. Secondly, it is not well defined if there is conversion from exhausted cells to memory cells and along the same thread the reversal of exhausted cells back to effector cells. Reversal will be further discussed in this Chapter. Finally, defining the molecular networks that govern the signaling and transduction potential of exhausted T cells remains to be completely understood. Significant progress has been made towards this goal particularly in the area of transcription factor expression and epigenetic modifications. Figure adapted from [337].

191

In addition to anergy and exhaustion, the other forms of recessive tolerance that have been described including clonal deletion, senescence and ignorance. Clonal deletion plays a significant role in the control of effector T cell populations after every immune response. In addition, interaction of molecules such as Fas/FasL and TRAIL are also involved in controlling the viability of T cells around specific organs [338]. It is not clear how relevant these processes are to antigen-specific peripheral tolerance versus broad homeostasis. Similarly, both ignorance and senescence are not equated with actively regulated tolerance as both are part of the continuum of a T cell’s life – even outside of tolerogenic antigen encounters. While some organs may shield antigens from T cells to offer ‘immunological privilege’ and lead to the T cell being ‘ignorant’ of the antigen, it is increasingly clear that this is not absolute [339]. Organs like the brain have their own lymphatic drainage and may elicit T cell responses distinct from those that were previously measured. Along these lines, senescence is typically seen in aging T cells where they have lost some ability to proliferative in response to TCR signaling [340]. General cell biology processes such as telomere shortening may play a role here. Yet, the proliferation loss results as a consequence of persistent antigen or self-tissues and it is not clear if senescence is sufficiently distinct from anergy or exhaustion. In a polyclonal repertoire, the selective retention of some TCRs versus others are also a confounding factor in interpreting models of senescence [341]. Although we don’t discuss ignorance or senescence further here, it is possible that once we have a better molecular understanding of these processes, overlapping molecular patterns can eventually be defined.

192

5.1.4. The molecular-genetic signatures of peripheral recessive tolerance

As discussed above, the functional readouts for the fate of T cells that have encountered a chronic self or viral antigen in vivo, are similar in the case of anergy and exhaustion (or for that matter dysfunction). These typically include reduced proliferation to antigen challenge and poorer production of cytokines. Since T cells use a definitive set of signaling and gene expression networks to express cytokines, proliferate, and differentiate, it is natural that some of the molecular changes associated with anergy and exhaustion considerably overlap since they both result in the loss of these programs.

Although exhausted CD8 T cells were initially thought of as a relatively homogenous albeit progressively acquired state, recent research has begun to parse these cells into distinct subsets, and importantly has identified a progenitor population responsible for maintaining the terminally differentiated cells [342]. There are markers that have been used to subset these populations (some shown in Figure 5.4) that include the overlap in transcription factors like T-bet, Eomes, TCF-1, and Tox as well as surface markers like PD1, KLRG1, IL7R, and others that are typical of T cell effector and memory differentiation [332]. These stages of exhaustion differentiation provide tools to determine where along the hierarchical and progressive path of exhaustion these cells are. With the recent advanced in genomic analysis, there have been studies that examines these populations of T cells to parse out the different molecular patterns as they differentiate from naïve to terminally differentiated states [343].

Compared to CD8 T cells, the hyporesponsive states in CD4 T cells are more complex. As they lose the ability to proliferate they seem to still have the capacity to make some cytokines, albeit different ones than the canonical IL-2 and IFNγ [329]. Recently,

193 exhausted CD4 T cells have more T follicular helper (Tfh) like phenotypes with increased production of IL-10 and IL-21 as well as similar gene signatures such as the expression of

Bcl-6 [331, 344-349]. This suggests that similar to the altered differentiation fate that is seen with CD8 T cells, CD4 T cells may progress toward a Tfh like state in the context of chronic stimulation. Interestingly, a similar Tfh like gene expression pattern is seen in CD8

T cells but has yet to be completely understood but the expression of PD-1 and CXCR5 are common surface markers of both Tfh and exhausted cells [350-352]. These heterogeneities have made it more difficult to stage the onset of exhaustion in CD4s, without parallel measurement of functional correlates.

In other studies of T cells specific for a self-antigen (glucose-6-phosphate isomerase, GPI) induce an arthritic phenotype. In the presence of Tregs in a lymphopenic system, these cells become anergic and do not cause pathology. Interestingly, there is a significant upregulation of two molecules, folate receptor 4 (FR4) and CD73, an ecto-5’- nucleotidase, on anergic cells [353]. These molecules were previously observed on Tregs

[354, 355] and seem to be concurrently upregulated as cells become anergic but, a functional role of these has yet to be completely understood. Yet, one can speculate that these markers are likely part of a non-responsive program.

Beyond changes at the level of surface receptors, it has recently been appreciated that there are epigenetic changes within exhausted T cells such that they have heritable alterations to their genetic programs [330, 343, 349, 356]. Most of the studies thus far have focused on comparing activated or memory cells with exhausted T cells such that they eliminate any changes that occur from antigen interaction alone. The changes occur at genes for effector responses to silence them while opening the chromatin around genes like

194 checkpoint molecules. Alterations at this level provide a stability to the phenotype that is not possible with other mechanisms like the upregulation of checkpoint molecules or altering other surface receptors.

Taken together, while there are shared factors between both CD4 and CD8 T cells during exhaustion, the programs that are initiated within these cells during chronic viral infection seem to have distinct molecular consequences between these two T cell lineages.

5.1.5. Reversibility of hypo responsiveness in T cells

A key feature described during peripheral tolerance in vivo, driven by the encounter of T cells with their cognate antigen either as a self-antigen or as a chronic viral infection was reversibility. Unlike clonally anergic T cells in vitro, whose functioning could be reversed upon treatment with non-TCR mitogens such as IL-2, tolerant T cells in vivo, could recover functions upon removal of antigen [321, 323]. The significance of this finding was that it offered a therapeutic recourse for instances where T cell tolerance was not desirable such as in the case of chronic infections and tumors. It is still debated whether or not reversal is possible in all chronically stimulated T cells [337]. Some suggest that there is a terminal differentiation state where upon entry, there is no way to recover the functionality of these cells due the epigenetic modifications that silence their effector molecule expression [357]. Here we will discuss the evidence of reversibility and the specifics of what has been achieved thus far.

Given the importance of antigen continuity in all of these situations, as proof of principle one of the easiest ways to determine the reversibility of this state is through removal of antigen. In the Rocha and von Boehmer experiments, this was done through a secondary adoptive transfer of the tolerant T cells into antigen free hosts as discussed in

195 section 5.1.3 where the cells recover the ability to proliferate in vitro [358]. Similarly, cells from a different model system after removal from chronic stimulation, there is a recovery of proliferation and production of cytokines (IL-2 and IFNγ) [359]. Additionally, another model system had recovery of proliferation and cytokine production but interestingly if the tolerant cells were re-transferred into an antigen bearing host they further lost functionality

[146]. Taken together, these multiple model systems provide strong evidence that the removal of antigen can reverse the hyporesponsive state and allow cells to recover both their proliferative and cytokine production capacity.

One regulatory pathway of T cells that has been exploited for reversal purposes has been checkpoint molecules. T cells upregulate these surface receptors as the molecular breaks to prevent overly responding to pathogenic insults. Unfortunately, in cases of chronic infection or cancer, these pathways shut down T cell responses to the insult and prevents effective clearance. As such, the use of checkpoint inhibitor therapies are extremely successful in re-invigorating the responses of some populations T cells in the context of cancer [360]. Yet, these do not result in complete restoration of responses so there are other mechanisms left to be understood and targeted to restore T cell responses, potentially ones that could re-invigorate the entire hyporesponsive population. Checkpoint inhibitors are monoclonal antibodies that block coinhibitory pathways such as PD-1 or

CTLA-4 but, as mentioned above, it is variable on the T cell populations that are responsive to this therapy [360]. It seems as though the progenitor cells are those that are able to be recovered rather than the terminally differentiated subsets [361]. This presents a problem where this only selectively expands and may not provide ample effector responses to combat the threat at hand. One explanation for this selective response may be due to the

196 epigenetic modifications in these T cells that preclude recovery of function [362, 363]. So, finding an alternative to these options is a current pressing question within the field. While some are approaching these therapies by combining multiple checkpoint molecules, we believe it may be a better strategy to find an underlying molecular program that governs all hyporesponsive T cells to target that pathway rather than the downstream inhibitory pathways that are engaged from chronic stimulation like the checkpoint molecules.

There is a strong push to find ways to re-invigorate T cells from these hypo responsive states particularly in the context of chronic infection and cancer. If we are able to understand the molecular paradigms that drive these responses, can we design ways to reverse their course? Recently, an unbiased screening of T cells in LCMV clone 13 infected hosts identified 19 pharmacological compounds that recover IFNγ production of which, a

PKC activator was the top hit [364]. This suggests that targeting the TCR signaling machinery itself may restore the functionality of T cells. Moreover, this strategy could provide a unifying theme to re-invigorating cell responses in multiple contexts of chronic stimulation. Therefore, understanding the ways in which the TCR signaling molecules themselves establish the hypo responsive state and how we can augments those programs is of great interest for therapeutic application.

5.1.6. The 5C.C.7 → mPCC model system for tolerance

The advent of TCR-Transgenic (TCR-tg) systems allowed systematic investigation of peripheral tolerance mechanisms using adaptive transfer strategies, popularized by

Jenkins and others using Ovalbumin as a model antigen [365]. The ability to track the fate of a monoclonal population from the naïve state all the way to tolerance (and reversal) allowed for rigorous experimental design that was not possible with polyclonal T cells and

197 their heterogenous TCRs. In the decade since, multiple model systems were developed to examine these questions – including the adoptive transfer of Hen Egg Lysozyme (HEL)- specific T cells (3A9 TCR transgenic) to mice expressing different forms of HEL by the

Goodnow lab [366, 367], HA-specific T cells to HA-expressing mice (or those challenged with a HA-expressing virus) from the von Boehmer, Pardoll, and Caton labs [368-373],

Ova-specific DO11.10 T cells transferred to variable Ova-transgenic mice by the Abbas

Lab [374], transfer of male-specific T cells to male or female mice by the Cobbold, Rocha, and von Boehmer labs [322] to name a few examples.

In this chapter, we use a model system that was originally developed by Dr. Corinne

Tanchot and Dr. Ron Schwartz [323], using TCR-tg mice made by Dr. Mark Davis. The

5C.C7 TCR transgenic mice express a monoclonal TCR containing a V11 and V3 pair specific for cytochrome c peptides derived from either moth or pigeon (MCC or PCC) presented on the class II MHC molecule I-E of the k haplotype (specific for I-Ek-M/PCC).

In this model, 5C.C7 T cells were adoptively transferred to PCC-transgenic mice, where the target antigen (PCC) was expressed as a transgene driven by the ubiquitous MHC-I promoter coupled to an Ig enhancer. The recipient mice were also deficient in CD3ε such that there is no development of endogenous T cells. Upon transfer into hosts that lack T cells and have cognate antigen expression, there is a rapid proliferative burst of the TCR- tg CD4+ T cells followed by a plateau phase around days 7-10 after transfer. Concomitant with this cessation of in vivo proliferation, T cells recovered from these hosts show a striking decrease in proliferative capacity and production of IL-2. There is also a wealth of information regarding bulk changes in the TCR signaling cascade within these cells as summarized in Figure 5.5, Copyright Clearance Center #4882690664729 [146, 147, 375-

198

377]. Briefly, there are both proximal (Zap70, LAT, etc) as well as distal alterations

(calcium signaling, ERK, IκBα, etc) that could contribute to the loss of functionality from these cells. This suggests that the chronic stimulation within these cells augments the signaling molecules themselves which in turn may cause changes to the activation threshold reflected by the loss of functionality. Given the ability of this established model system for peripheral T cell responses to chronic versus acute antigen, we set out to examine the molecular mechanisms underlying activation threshold modulation using this platform.

199

Figure 5.5. T cell receptor signaling perturbations after chronic stimulation in a self-antigen murine model.

5C.C7 TCR-tg CD4+ T cells are transferred into hosts that bear a transgene for constitutive expression of their cognate antigen PCC. Within this system, there is progressive loss of function (loss of proliferative capacity and IL-2) as well as distinct biochemical changes along all modules of the TCR signaling pathway. Within the proximal module, there is decreased Lck and Zap70 activity. Additionally, there is an accumulation of Zap70 but, a defect in its translocation to the TCR complex. Within the intermediate module, there are additional defects in translocation of LAT, PLCγ, and Vav. Phosphorylation status of these molecules are also altered. At the distal pathways there is impaired calcium mobilization, PKCθ translocation, and decreased expression of phosphorylated ERK. Taken together, this model system has systemic defects in TCR signaling. Figure adapted from [378].

200

5.2. Chapter Summary

We used an adoptive transfer model system to characterize changes in naïve antigen-specific T cells as they experienced acute versus chronic antigenic stimulation in vivo. We found that compared to acutely stimulated or naïve T cells, chronically stimulated cells were about 100-fold less sensitive to antigen. Although previous data shows that this unresponsiveness can be recovered when transferred to mice without constant antigen stimulation, we find that short term deprivation of antigen in vitro is not sufficient to restore

T cell functionality. Chronically stimulated cells show multiple alterations in their signaling machinery. Finally, we used pharmacological reagents to disrupt the molecular targets identified from a drug screen hit to see if those will recover any functional capacity of chronically stimulated T cells.

5.3. Results

5.3.1. Adoptive transfer of TCR-Tg CD4+ T cells into cognate-self-antigen murine model system

To investigate the progression into a hyporesponsive state without any confounding variables from infection or the tumor microenvironment, we employed an adoptive transfer mouse model where 5C.C7 TCR-Tg CD4+ T cells (B10.A,5C.C7, RAG2-/-) are transferred into host mice of two different genotype – one bearing the PCC transgene for chronic stimulation and the other without antigen expression. The latter is then primed with an intraperitoneal injection of PCC peptide and LPS, to provide a single stimulation. (Figure

5.6A). After various days of stimulation, T cells are isolated by negative selection using magnetic beads (Figure 5.6B) for further analysis as outlined in subsequent sections in this

Chapter.

201

Figure 5.6. Adoptive transfer model system for chronic self-antigen stimulation.

Naïve 5C.C7 TCR-Tg CD4+ T cells were isolated and adoptively transferred into CD3εKO or PCC-CD3εKO animals. Cells were left resident in hosts for 14-28 days before ex vivo analysis was performed. (A) Experimental diagram of adoptive transfer system. (B) Flow cytometric analysis of naïve, acute, and chronic 5C.C7 T cells before and after purification protocol.

202

5.3.2. Chronic antigen exposure dampens T cell responsiveness

T cells recovered from acute or chronic hosts were re-stimulated in vitro to evaluate the proliferative capacity of these cells. We found that there was a significant loss of proliferative capacity such that the EC50 of the dose response curve shifts from 0.0015μM or 0.007μM for naïve and acute stimulation respectively to 1μM for chronic stimulation

(Figure 5.7A – 5.7B) Additionally, there was a significant loss in IL-2 cytokine production upon re-stimulation (Figure 5.7C). This was critical in establishing the baseline for all further assays and analysis that aim to recover functional responses of these T cells.

Moreover, given that the microbiome can influence the hyporesponsive state of T cells in this model system [379], it was imperative to determine these functional parameters with our animal housing conditions.

203

Figure 5.7. Chronic antigen stimulation decreases proliferative capacity and cytokine production of T cells.

5C.C7 TCR-Tg CD4+ T cells were isolated from Chronic or Acute hosts as depicted in Figure 5.6. or from naïve 5C.C7 TCR-Tg animals. The cells were loaded with CFSE and then normalized numbers of 5C.C7 cells from each host (as determined based on flow cytometric analysis) were plated with CD3ε-KO splenocytes with varying doses of peptide and cultured for 72 hours. Cells were stained and analyzed via flow cytometry. (A) Representative histogram of 5C.C7 T cells from naïve (green), acute (blue), and chronic (red) hosts were re-stimulated with 0.1μM peptide. (B) Dose response curve of naïve, acutely, and chronically stimulated 5C.C7 T cells (n = 6 from 3 independent experiments). Significance was calculated using one- way ANOVA with Tukey’s multiple comparisons test, # indicate comparisons of acute versus chronic and * indicate comparisons of naïve versus chronic. (C) Naïve (green), acute (blue), and chronic (red) 5C.C7 T cells were isolated and re-stimulated ex vivo with 1μM peptide and supernatants were collected to measure IL-2 by ELISA. Significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

204

The data presented in Figure 5.7 was using the lymphopenic model system shown in Figure 5.6 but, there are conflicting phenotypes when TCR-Tg T cells are transferred into lympho-replete hosts (those that have a full compartment of T cells, Figure 5.8). For example, an arthritic phenotype developed within the lymphopenic model shown in Figure

5.6 but, this was not present in the lympho-replete model [380] suggesting that there are distinct differences in T cell function between these two states. Within the lympho-replete model, there was a shorter period of residency before these cells become hyporesponsive so the timing of the experiments and analysis is different from Figure 5.6. We found that there was an increase in the EC50 of chronically stimulated cells compared to acutely stimulated counterparts (Figure 5.9). We harvested these cells at two separate timepoints,

5 days and 14 days and found at both timepoints there was a loss in proliferative capacity compared to acutely stimulated cells (Figure 5.9B and 5.9D). Interestingly, the EC50 of these cells was shifted to approximately 0.03μM at 5 days and approximately the same at day 14 showing no additional loss of functionality over time (Figure 5.9B and 5.9D). Taken together these data provide a baseline to use for our other analysis of T cells isolated from a lympho-replete model.

205

Figure 5.8. Adoptive transfer model system for chronic self-antigen stimulation in lympho-replete hosts.

Naïve 5C.C7 TCR-Tg CD4+ T cells were isolated and adoptively transferred into B10.A or PCC+ animals. Cells were left resident in new hosts for 4-18 days as indicated in each experiment before ex vivo analysis was performed.

206

Figure 5.9. Chronic antigen stimulation in replete hosts has loss of proliferative capacity as early as day five.

Acute or Chronic 5C.C7 TCR-Tg CD4+ T cells were isolated from hosts as indicated in Figure 5.8. Cells were CFSE labeled and plated with APCs and peptide and cultured for 72hrs before flow cytometric analysis. (A) Representative flow plots of acute (blue) and chronic (red) T cells isolated at day 5 and cultured with APC and peptide at indicated concentrations for 72hrs and then analyzed via flow cytometry. (B) Dose response curve from day five-isolated acute and chronic T cells. (n=2 from 1 independent experiment) (C) Representative flow plots of acute (blue) and chronic (red) T cells isolated at day 14 and cultured with APC and peptide at indicated concentrations for 72hrs and then analyzed via flow cytometry. (D) Dose response from day fourteen-isolated acute and chronic T cells (n=1 from 1 independent experiment).

207

5.3.3. In vivo antigen deprivation restores some functional capacity and signaling machinery of chronically stimulated T cells

To investigate how to functionally reverse the T cell, we used an in vivo model system where we re-transferred cells that were either acutely or chronically stimulated into antigen free hosts to allow time to readjust their activation threshold, similar to previous models [146]. We set up two different transfers, one where the cells were isolated at 14 days after initial transfer and then transferred into antigen free hosts and left resident for

10 days before re-isolated to re-stimulate ex vivo. The second time point was after 21 days of the initial transfer, cells were isolated and transferred into antigen free hosts and left resident for 32 days before isolation and re-stimulation (Figure 5.10A). We found that there was no significant increase in the proliferative capacity of chronically stimulated T cells

10 days after re-transfer into antigen free hosts (Figure 5.10B and 5.10C). The proliferative capacity of the chronically stimulated T cells was still significantly impaired compared to the naïve controls. Additionally, we found that the acutely stimulated T cells had decreased proliferative capacity compared to naïve cells as well. Since this was unexpected and inconsistent with previous data as presented in section 5.1.6, additional experiments are needed to examine these differences. One known factor to influence T cell responses in this system is the microbiome [379] as mentioned above and that has yet to be examined at this institution so, it remains a potential explanation for our inconsistent results.

We also performed this analysis at a second timepoint that had a longer residency period in the antigen free hosts. We found there was heterogeneity among the recovery of proliferative capacity in chronically stimulated T cells (Figure 5.10D and 5.10E). The chronically stimulated T cells from this experiment were all isolated 21 days after transfer

208 of naïve cells into the antigen-bearing hosts. The cells were all isolated from the same animal and adoptively transferred into three separate antigen-free hosts. Of these three hosts, the recovered chronically stimulated cells had variable proliferative capacity; where one biological replicate had an EC50 of approximately 0.001μM whereas the two other biological replicates had higher values (Figure 5.10E) more consistent with the previous results in Figure 5.10C. This suggests that there is heterogeneity among the T cells that may reflect their potential recovery of function.

209

Figure 5.10. In vivo deprivation of antigen restores some functional capacity of chronically stimulated T cells.

(A) Acute and Chronic T cells were isolated from hosts as depicted in Figure 5.6 at the timepoints indicated. and transferred into CD3ε-KO animals for the timepoints indicated. Previously acute or chronic T cells were then re-isolated and plated with APC and peptide at indicated concentrations and cultured for 72hrs prior to flow analysis.

210

Figure 5.10. In vivo deprivation of antigen restores some functional capacity of chronically stimulated T cells (cont.). (B) Representative flow plots from naïve (green), acute (blue), and chronic (red) T cells isolated 14 days after first transfer and 10 days after second transfer. (C) Dose response curve of naïve, acute, and chronic T cells from panel B. (D) Representative flow plots from naïve (green), acute (blue), and chronic (red) T cells isolated 21 days after first transfer and 32 days after second transfer. (E) Dose response curve of naïve, acute, and chronic T cells from panel D.

211

5.3.4. Short term in vitro antigen deprivation does not restore functional capacity in chronically stimulated T cells and reveals distinct signaling alterations

One way to potentially readjust the activation threshold is to remove the chronic stimulus which has been shown in vivo [146]. While this is not physiologically or therapeutically an option for most patients, as a proof of principle we used an in vitro model system of antigen deprivation where the naïve, acute, or chronic T cells were cultured without antigen in the presence of APCs isolated from CD3ε-KO animals. We used a 48- hour rest period and then re-stimulated the cells with antigen. As expected, there was no difference in proliferative capacity of naïve T cells whether they are analyzed immediately ex vivo or rested within culture for 48 hours (Figure 5.11B) Additionally, there was not a significant recovery of proliferative capacity in chronically stimulated T cells after an in vitro rest period (Figure 5.11D). Interestingly, we found that there was a significant increase in the proliferative capacity of the acutely stimulated T cells (Figure 5.11). This suggests that there is an enhancement of proliferative activity of acutely-stimulated cells upon antigen deprivation. Moreover, this suggests that upon removal from an acutely- challenged host, 48 hours of culture can lower the activation threshold of these cells to enhance their antigen sensitivity. Taken together, these data suggest that the increased activation threshold of T cells that have experienced chronic stimulation is stable for at least 48 hours of antigen deprivation ex vivo.

212

Figure 5.11. In vitro antigen deprivation does not improve proliferative capacity of chronically stimulated T cells.

(A) Naïve, acute, and chronic T cells as generated in Figure 5.6. were isolated and cultured ex vivo with APCs for 48hrs. T cells were isolated and re-stimulated with peptide concentrations as indicated and cultured ex vivo for 72hrs prior to flow analysis. Dose response curve of naïve (B), acute (C), or chronic (D) T cells immediately ex vivo (initial) or after 48hrs with APC and subsequent re-stimulation (rested). (n = 3 from 3 independent experiments). Significance was calculated with two-way ANOVA with Sidak’s multiple comparisons. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant. 213

5.3.5. High dose antigen stimulation in vitro reveals two distinct functional populations of chronically stimulated T cells

We cultured chronically stimulated T cells in the presence of antigen for 48 hours and then re-stimulated these cells to see if the pre-treatment with high dose antigen would augment their activation threshold (Figure 5.12A). We found that high-dose antigen increased the proliferative capacity of chronically stimulated T cells compared to those that did not receive pre-stimulation (Figure 5.12B and 5.12C). This suggests that the introduction of high-peptide concentration is sufficient to induce proliferation in a population of chronically stimulated 5C.C7 T cells.

We investigated this further by using two different proliferation dyes to separate the proliferation of the high-dose antigen and the second stimulation with the lower dose antigen. Prior to the 48-hour high-dose peptide culture, cells were labeled with CFSE so that we were able to distinguish the populations that divided from the initial high dose peptide. The cells were then re-isolated after 48 hours and re-labeled with CTV such that we were able to distinguish cells that divided from the re-stimulation during the proliferation assay. Interestingly, we found that there was a distinct population of chronically stimulated T cells that do not respond to either high-dose peptide or re- stimulation after high-dose peptide stimulus (Figure 5.12D and 5.12E). Similar to the results in Figure 5.10D and 5.10E, this further supports the heterogeneity among these cells despite the use of a TCR-Tg model system, emphasizing the role that intrinsic signaling modulation may play in the ability of a T cell to respond to antigen to generate two distinct populations; one that is able to proliferate in response to high dose peptide and one that is not.

214

Figure 5.12. In vitro stimulation with high dose antigen restores proliferative capacity of chronically stimulated T cells.

(A) Naïve, acute, and chronic T cells as generated in Figure 5.6. were isolated and cultured ex vivo with APCs ±10μM concentration of peptide for 48hrs. T cells were isolated and re-stimulated with peptide concentrations as indicated and cultured ex vivo for 72hrs prior to flow analysis. (B) Representative flow plots of naïve (green), acute (blue), and chronic (red) after 48hr ex vivo culture and subsequent re-stimulation with APC and peptide concentration as indicated. (C) Frequency of undivided cells from experiment in panel B (D) Representative flow plots of naïve (green), acute (blue), and chronic (red) T cells after 48hr ex vivo culture with 10μM PCC peptide and subsequent re-stimulation with APC and peptide concentration as indicated. (E) Frequency of undivided cells from experiment in panel D.

215

5.3.6. Pharmacological treatment does not alter the functional capacity or signaling molecule expression in chronically stimulated T cells

Previously, a drug screen was conducted using the NCATS catalog and identified multiple potential pharmacological agents that would restore IL-2 production from chronically stimulated T cells. This was screened with a high-throughput cell-based assay that measures the bypass of T cell tolerance through the expression of GFP and DsRed, under the control of two distinct IL-2 promoters (Figure 5.13A). One promoter is the endogenous IL-2 promoter driving GFP expression and the second is a transgenic IL-2 promoter in which specific sequences known to be subject to epigenetic regulation have been removed driving DsRed expression (Figure 5.13A). This provides a multiplexed measurement of IL-2 regulation with and without opportunities for epigenetic influence, collected simultaneously within the primary assay itself. Compounds that show differential levels of GFP and DsRed expression will be further investigated to determine whether any disparities are truly due to epigenetic regulation (marking them as members of an additionally interesting class) or due to some off-target effect. Transgenic 5C.C7 T cells bearing both reporter constructs were used to generate acute and chronically stimulated cells as indicated in Figure 5.6.

The IL-2 production was validated with flow cytometric analysis to verify that the reporter systems were sufficient in the transgenic cells (Figure 5.13B). Moreover, a high- throughput culture system was optimized for number of cells and stimulation conditions for quantitation of positive signals from both reporter constructs (Figure 5.13C). Screening was performed on the LOPAC drug library and found that there were very few compounds that were able to restore IL-2 production (Figure 5.13D and 5.13E). Of the identified

216 compounds, interestingly PMA, a PKC activator was the top hit (Figure 5.13E), similar to a recent study [364].

To validate these findings, we set up proliferation assays using anti-CD3ε stimulus to prevent any off-target effects from the APC in the presence two candidates from the high-throughput screening, SU6656 and SB216763. SU6656 is a Src family kinase inhibitor and SB216763 is a highly selective GSK-3 inhibitor [381, 382]. The cells were cultured for 72 hours in the presence of SU6656 or SB216763 at indicated concentrations and harvested for flow cytometric analysis. From this single pilot assay, we did not find any recovery of proliferative capacity from any concentration of SU6656 or SB216763

(Figure 5.14).

217

Figure 5.13. High-throughput drug screening to identify tolerance reversing compounds.

(A) Reporter constructs with endogenous IL-2 promoter driving expression of GFP and transgenic IL-2 promoter lacking epigenetic modification sites driving expression of DsRed. (B) 5C.C7 T cells bearing transgenic IL-2 promoter were stimulated with peptide and IL-2 production was quantitated via flow cytometric analysis through GFP and intracellular staining. Drug screening and analysis, courtesy of Eleanore Chuang and Nevil Singh, unpublished.

218

Figure 5.13. High-throughput drug screening to identify tolerance reversing compounds (cont.). (C) Experimental set up for high-throughput screening. (D) Quantitation of compounds identified as positive from screening in panel C. (E) List of compounds identified through screening. Drug screening and analysis, courtesy of Eleanore Chuang and Nevil Singh, unpublished.

219

Figure 5.14. Pharmacological inhibitors do not improve proliferative capacity of chronically stimulated T cells.

(A) Naïve (green) or chronic (red) T cells were isolated from hosts as indicated in Figure 5.6. Cells were cultured on anti-CD3ε coated plates at the indicated concentrations with no treatment (No Tx), DMSO (DMSO), SU6656 at indicated concentrations, or SB216763 at indicated concentrations. (B) Frequency of undivided populations of naïve and chronic T cells from experiment in panel A (n=1 from 1 independent experiment).

220

Figure 5.14. Pharmacological inhibitors do not improve proliferative capacity of chronically stimulated T cells (cont.). Naïve, acute, and chronic T cells were isolated from hosts as indicated in Figure 5.4. (C) Cells were cultured ex vivo with APC and drugs at indicated concentrations for 48hrs. Cells were then isolated and plated for proliferation assay with peptide concentrations as indicated for 72hrs and subsequently analyzed via flow cytometry. Representative flow plots of T cells from hosts as indicated cultured with no treatment (No Tx), phosphatase inhibitor cocktail (PI, 1:100), okadaic acid (OKA, 10nM), or trichostatin A (TSA, 500nM). Are shown. (D) Frequency of undivided populations of naïve, acute, and chronic T cells from experiment in panel C (n=1 from 1 independent experiment).

221

Given that the balance of signaling molecules is crucial to the productive TCR signaling cascade, we also tested other pharmacological inhibitors such as phosphatase inhibitor cocktails, okadaic acid (protein phosphatase inhibitor), and trichostatin A (HDAC inhibitor) in a proliferation assay as well. We pre-treated the naïve, acute, or chronic T cells with the drugs for 48 hours and then subsequently washed to remove the drug from the media. The cells were then harvested and prepped for a proliferation assay as described in

Chapter 2. After 72 hours, cells were harvested and stained for flow cytometric analysis.

We did not find any recovery of proliferation within the chronic T cells in these assays

(Figure 5.14). Again, while this was a pilot assay, this suggests that either the drug concentrations were insufficient, the timing of the culture period did not provide adequate time to augment the molecular machinery of the T cells, or that they simply do not change the proliferation capacity of the chronic T cells.

5.3.7. Bypassing the proximal TCR signaling machinery restores functional capacity of chronically stimulated T cells

Previous evidence suggested that there is a proximal blockade of the TCR signaling cascade that prevents the production of cytokines [375]. We found that stimulation with

PMA and ionomycin recovered some, but not all, cytokine production suggesting that while there is a proximal blockade (i.e. above calcium and PKC activation) there are also distal mechanisms at work to prevent productive TCR signaling (Figure 5.15). Moreover, we found that the proliferative capacity was not recovered showing distinct functional recovery in some aspects, like IL-2 production, but not others, such as proliferation (Figure

5.15). Taken together these data suggest that there is multiple signaling blockades in chronic cells; both at the proximal and distal modules. Additionally, this highlights the

222 differential requirements for functional recovery of two distinct responses where the proximal bypass can increase IL-2 production but not proliferative capacity.

223

Figure 5.15. Bypassing TCR proximal signaling molecules improves cytokine production but not proliferative capacity in chronically stimulated T cells.

Naïve, acute, and chronic T cells were isolated from hosts as indicated in Figure 5.6. Cells were plated for proliferation assay with APCs and either peptide or PMA (0.3μg/mL) and varying concentrations of ionomycin (μg/mL) as indicated and cultured for 72 hours before collecting supernatants and staining for flow cytometric analysis. (A) Representative flow plots of CFSE from naïve (green), acute (blue), and chronic (red) cells treated as indicated. (B) Frequency of undivided cells from naïve, acute, and chronic cells treated as indicated (n = 3 from 3 independent experiments). (C) IL-2 production as measured by ELISA (n = 1 from 1 independent experiment).

224

5.3.8. Signaling molecules are altered in a chronic self-antigen model system

As discussed above, there are multiple TCR signaling alterations in this murine model system. The accumulation of data summarized in Figure 5.5 was performed with

Western blot analysis. This bulk analysis does not capture the heterogeneity of the populations being analyzed as evidenced by the differences seen in recovery of function in

Figure 5.10. We set out to use a flow cytometric approach to have single cell resolution of the molecular changes within this model system.

Given that there are likely both proximal and distal blockades in signaling, we first investigated if there were differences in the early signaling components of chronically stimulated T cells. We found, similar to previous findings, that there was increased Zap70 expression in chronically stimulated T cells compared to both naïve and acutely stimulated

T cells (Figure 5.16A and 5.16B, left panels). There was also increased expression of phosphorylated forms of Zap70 at the Y292 and Y319 tyrosines (Figure 5.16A and 5.16B, middle and right panels). The Y292 and Y319 phosphosites are thought to serve as docking sites for c-Cbl and Lck, respectively, and are located within the interdomain B region of the molecule [84]. Interestingly, when normalizing the expression of the phosphoforms of

Zap70 to the increase in total Zap70, there was a decrease in the Y292 phosphoform of

Zap70 but not the Y319 phosphoform (Figure 5.16C). This analysis was performed in resting cells from each antigen exposure condition (Figure 5.6) which suggests that there are basal differences in the phosphorylation of Zap70 dependent upon previous antigen exposure. Taken together, this suggests that one of the earliest kinases within the TCR signaling cascade has a functional defect which would cause global dysregulation of the entire signaling cascade.

225

Figure 5.16. Expression of total and phosphorylated Zap70 is altered in chronically stimulated T cells.

Naïve, acutely, and chronically stimulated 5C.C7 T cells were isolated from hosts as indicated in Figure 5.6 and stained for flow cytometric analysis. (A) Representative flow plots of total Zap70, pZap70 Y292, and pZap70 Y319 in naïve, acute, and chronic T cells. (B) gMFI of total Zap70, pZap70 Y292, and pZap70 Y319 in naïve, acute, and chronic T cells. (C) Ratio of pZap70 Y292 to total Zap70 (left panel) and ratio of pZap70 Y319 to total Zap70 (right panel) (n=1 from 1 independent experiment).

226

Given the data from Chapter 4 that showed that acute antigen stimulation increased

CD5 and IκBα expression, we were interested in using this chronic stimulation model to observe the expression of these molecules. We found that there was an increase in both

CD5 and IκBα in chronically stimulated T cells (Figure 5.17A and 5.17B). Interestingly, this did not reach statistical significance with the number of animals analyzed here.

Furthermore, the lack of difference between acutely and chronically stimulated cells suggests that continued antigen stimulation did not further increase expression of either

CD5 or IκBα (Figure 5.17C). We performed similar analysis in the lympho-replete model system as shown in Figure 5.8 to observe expression of CD5 and IκBα and found increased

CD5 and IκBα expression on varying days after transfer (Figure 5.18A). Interestingly, the kinetics of expression were different in this model system where there was first a transient increase in CD5 and IκBα in both the acute and chronic cells but, this was not maintained at later time points (Figure 5.18B). This suggests that there is transient upregulation of CD5 and IκBα that is decreased after the initial antigen encounter similar to the kinetics we observed in our in vivo challenge experiments in Figure 4.14.

227

Figure 5.17. CD5 and IκBα are upregulated during chronic antigen stimulation.

Naïve, acutely, and chronically stimulated 5C.C7 T cells were isolated from hosts as indicated in Figure 5.6 and stained for flow cytometric analysis. (A) Representative flow plot of CD5 and IκBα expression in naïve (green), acute (blue), and chronic (red) T cells. (B) Representative histograms of CD5 and IκBα expression in naïve (green), acute (blue), and chronic (red) T cells. (C) gMFI of CD5 and IκBα in naïve, acute, and chronic T cells (n = 3 from 3 independent experiments). Statistical significance was calculated by one-way ANOVA with Tukey’s multiple comparisons. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

228

Figure 5.18. CD5 and IκBα are upregulated during chronic antigen stimulation.

Naïve, acute, and chronic 5C.C7 T cells were isolated from hosts as indicated in Figure 5.6 at varying timepoints indicated. (A) Representative flow plot of CD5 and IκBα expression in naïve (green), acute (blue), and chronic (red) T cells. (B) CD5 and IκBα gMFI normalized to naïve expression at varying time points (n=1-2 from 1 independent experiment).

229

Finally, to address whether this is a specific phenomenon of the self-antigen model system we used an E.G7-OVA tumor model to measure expression of CD5 and IκBα in a polyclonal population to evaluate tumor-infiltrating T cells as well as draining and non- draining T cells (Figure 5.19). We found that there was an upregulation of IκBα in the tumor-infiltrating CD4 and CD8 T cells compared to both the non-draining and draining lymph nodes yet there was only significant upregulation of CD5 in the tumor infiltrating

CD4 T cells (Figure 5.19B). This further suggests that the upregulation of these molecules occurs in a polyclonal population under physiological chronic stimulation.

230

Figure 5.19. CD5 and IκBα are upregulated in tumor infiltrating lymphocytes.

Tumors, draining popliteal, and non-draining popliteal lymph nodes were isolated from E.G7-OVA tumor bearing B6 mice. Cells were stained for flow cytometric analysis. (A) Representative flow plot of CD5 and IκBα expression in tumor infiltrating (red), draining LN (grey), and non-draining LN (black) CD4 and CD8 T cells. (B) CD5 and IκBα gMFI in tumor infiltrating, draining LN, and non-draining LN CD4 and CD8 T cells (n = 5 from 1 independent experiment). Statistical significance was calculated by one-way ANOVA with Holm-Sidak’s multiple comparisons. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

231

Given the wealth of markers that have been characterized in exhausted T cells we wanted to investigate the expression of checkpoint molecules as well as exhaustion associated transcription factors NFAT1c and Tox [383-388]. We found an increase in expression of the checkpoint molecules, PD-1 and CTLA-4 on chronically stimulated T cells (Figure 5.20). Additionally, we found an increase in both Tox and NFATc1 in chronically stimulated T cells (Figure 5.20). Interestingly, there was an intermediate phenotype of the acutely stimulated T cells for both Tox and NFATc1 (Figure 5.20). While these results were preliminary, these data help to provide us with a distinct signaling signature that results from chronic stimulation through the TCR and may provide insight into molecules that are involved in altering the activation threshold of these cells to render them hyporesponsive. Moreover, similar marker expression provides evidence that using this model may provide insight to other systems of chronic simulation such as tumors or chronic infections.

232

Figure 5.20. Transcription factors associated with exhaustion and checkpoint molecules are upregulated after chronic stimulation

Naïve (green), acute (blue), and chronic (red) 5C.C7 T cells were isolated from hosts as indicated in Figure 5.6 and stained for flow cytometric analysis. gMFI of each population is listed on the histogram (n=1 from 1 independent experiment).

233

5.4. Discussion

Here we validate a self-antigen adaptive tolerance model where there is a significant loss of both proliferative capacity and cytokine production to evaluate reversibility and signaling alterations in both chronic T cells isolated from lymphopenic and lympho-replete models (Figure 5.6 and 5.8). We find that in vivo-deprivation of antigen through a subsequent adoptive transfer of chronic T cells into antigen free hosts recovers some proliferative capacity but results in heterogeneity of these recovered responses (Figure 5.9). Additionally, an in vitro antigen deprivation culture system does not restore proliferative capacity of chronic T cells (Figure 5.10) suggesting that the hyporesponsive state is stable for at least 48 hours without antigen. Using this culture system, we find that stimulation with high concentration of peptide restores the proliferative capacity of some, but not all, chronic T cells further suggesting the heterogeneity among the population despite the use of a TCR-Tg system (Figure 5.12). To further investigate the reversibility of chronic T cells we used a culture system model to test a variety of pharmacological agents but found that none of those tested provided any recovery of proliferative capacity (Figure 5.14).

As discussed in section 5.1.6, the self-antigen model system we use has known defects in Zap70 [376]. The previous analysis had all been performed on bulk samples and we were interested in understanding this on a single cell level to understand the heterogeneity within this population. Similarly, we find that the total amount of Zap70 is increased in chronic cells but, the ratio of phosphorylated forms of Zap70 is decreased

(Figure 5.16) [375]. This defect is interesting to considering since Zap70 is within the proximal modulate of the TCR signaling cascade and controls the entirety of the

234 downstream response. As such, a defect within this proximal module would be expected to have a global effect on the remainder of the signaling cascade. Others have found distal component dysregulation (Figure 5.5) which may be a result of the problems at the proximal module. Interestingly, we find that bypassing the proximal module through PMA and ionomycin stimulation recovers some IL-2 production but not proliferative capacity

(Figure 5.15). This suggests that while there is a proximal blockade that is able to be overcome through these stimulation conditions, there exists a distal blockade that prevents full recovery. Moreover, there is differential requirement for recovery of IL-2 production versus proliferative capacity that has not yet been interrogated. It is not surprising that there are differential requirements for recovery of these two functions since they operate with very different signaling machinery. Where the cell cycle machinery must be functional for proliferation, this is not needed for cytokine production. Similarly, despite the fact that cytokines require multiple transcription factors for the efficient transcription, this may be less impaired compared to the blockades that exist within the cell cycle pathways [313].

Given the novel identification of the CD5-IκBα axis in Chapter 3 and 4, and the upregulation of these molecules during activation we measured the expression of these in both acute and chronic cells. We find that there is an increase in both of these molecules in the acute and chronic cells but, there is no further upregulation within chronic cells suggesting that there is not a summation effect of continued antigen stimulation and the upregulation of these molecules. Additionally, given the proximal blockade of the TCR signaling, it is not surprising that the upregulation of these molecules does not provide any functional advantage since the activation of the distal NFκB pathway is downstream of

Zap70.

235

The state of the exhaustion field is still working to identify reliable markers to delineate populations of T cells. Last year, there was a wealth of papers describing the importance of the transcription factor Tox in setting the stage for exhaustion in many different murine model systems and human contexts of chronic stimulation [383, 385-388].

Along with other markers of exhaustion such as checkpoint molecules and NFAT we found that this self-antigen system maintains similar phenotypes. The direct roles of these transcription factors in particular have yet to be fully elucidated.

Ultimately, the goal to understanding the molecular processes that mitigate the exhaustion phenotype is to harness that information to improve T cell function. To that end we have employed an antigen-deprivation model both in vitro and in vivo and found that there is heterogeneity among the recovery of the cells. The hyporesponsive phenotype is relatively stable since a short in vitro culture deprivation does not improve proliferative capacity (Figure 5.11). Yet, in vivo there seems to be recovery after 30 days but this is highly variable within the pilot study conducted here (Figure 5.10). While these provide proof of principle studies that the state is reversible, this is not a therapeutic option. To harness information from these systems, the molecular signatures of the cells needs to be studied to provide clues of what pathways are altered upon regain of functionality.

236

Chapter 6: Tuning of activation thresholds and the role of rate of change of antigen

in a quantitative framework for immune activation

6.1. Introduction

A fundamental question that is relevant to almost every aspect of immunology – whether studies on basic immunology or the design of clinical therapeutics and vaccines – is how an immune response is initiated. In this context, activation of T cells has a profound consequence, since it modifies and regulates multiple other cells types involved in an immune response. Optimal activation of T cells can lead to an effective orchestration for successful anti-pathogen response, while inadvertent activation of self-reactive T cells can trigger a cascade of autoimmune events. Given the critical nature of this step, the immune system must have evolved a robust set of principles to make sure that immune activation is mostly directed against threats the body faces and subsequently controlled to avoid collateral damage to the host.

6.1.1. The discrimination problem for T cell activation

The first question that the biochemical machinery in T cells has to answer in this regard is which targets the T cell should be activated against. Since the TCR is generated randomly in the thymus, as discussed previously in Chapter 3, there is the likelihood that that the TCR binds a target pMHC derived from a harmless self-antigen. Despite thymic negative selection, a sizeable fraction of the peripheral repertoire continues to be self- reactive [389]. Understanding how these T cells do not routinely and continually get activated has been a conceptual challenge that has occupied theoretical immunologists for a long time. The current model, as learned in textbooks is based on the two-signal model for T cell activation [390] (Figure 5.2). In this framework, the key decision of whether to

237 respond to a pathogen or not is essentially made by the dendritic cells. As discussed in

Chapter 1, DCs which sense pathogens or danger in the tissues are activated by PAMPs or

DAMPs to upregulate co-stimulatory ligands. The T cell which sees pMHC on an activated

DC, receives co-stimulation (Signal 2) in addition to the TCR signal (Signal 1). The presence of both signals together is ‘interpreted’ by the T cell as a ‘GO’ message and leads to full T cell activation, proliferation, and differentiation. In contrast, only having Signal

1, which would be expected when a resting DC presents self-antigen in the absence of infection or danger signals, would lead to a ‘STOP’ message, which limits the T cells activation – and triggering instead, pathways of anergy and tolerance (as discussed in

Chapter 5). This is largely a qualitative model, since the presence or absence of distinct signals from a second receptor (not the TCR) essentially drives the decision-making process in T cells. Yet, it is increasingly clear that the critical costimulatory molecules involved in early T cell activation do not provide a distinct signal to the T cell, instead mostly amplifying pathways that the TCR already triggers [391-395]. Another conceptual challenge to a purely two-signal model-based logic for T cell activation is that DCs, which are activated by PAMPs or DAMPs, will also present self-antigens to T cells under infectious or danger situations. An argument to get around this was the involvement of

Tregs which might suppress any autoreactive T cells that is activated inadvertently. Yet, it is again not clear how these cells would spare the pathogen-reactive T cell while suppressing the autoreactive cell. This is especially poignant since inflammatory cytokines such as IL-6 made during pathogen activation, also inactivate Tregs [396, 397]– potentially making them ineffective against halting autoreactive T cell activation in a bystander fashion during a pathogen response. Alternate mechanisms could be that PAMP

238 stimulation is linked very strictly to compartments from which antigen in DC is collected for presentation [398, 399]. The key issue with these models is that the ‘GO’ versus ‘STOP’ decision in T cells is made in an antigen-non-specific manner. If instead, at least part of the logic for T cell discrimination (self versus pathogen) is deciphered by properties that are antigen-specific, this would offer T cells a level of control that is more focused. In this

Chapter, I will summarize my efforts to explore the predictions of one such theoretical model which offers a quantitative framework for decision making in T cell activation [400,

401].

6.1.2. The Tunable activation threshold (TAT) model

The notion that T cells are capable of adapting to their environment was hypothesized by Dr. Zvi Grossman and Dr. William E. Paul [42] as part of the Tunable

Activation Threshold (TAT) model proposed in 1992 (further reviewed [43, 402, 403]). As we discussed in previous Chapters, the signaling machinery in T cells sets an activation threshold that is initially programmed in the thymus, and maintained (with contributions from molecules such as CD5) in the periphery. In order to activate the T cell, antigenic signals must overcome this threshold. But as we saw in Chapter 5, peripheral T cells can dynamically alter their activation thresholds by changing parameters of their signaling network. So the question becomes, what determines whether a T cell will be activated (by antigenic stimulus being strong enough to overcome the threshold) or left alone (by

‘silently’ changing its activation threshold)? The relevance of the TAT model for our work is that it predicted the dynamic behavior of T cell activation thresholds in 1992, before all of the modern advances in signal integration discussed in Chapter 1.

239

The TAT model suggests that T cells alter their activation threshold based on recent antigen encounter by altering signaling molecules within the T cell [42]. Therefore, the most recent experiences of antigen, will have an impact on the signaling capacity of the receptor and ultimately the response of the cell. Although it was evident before this model that T cells incorporate intracellular features such as negative feedback into their response upon antigen encounter, T cell activation was (and still largely is), thought of as a binary process. In other words, this school of thought predicts that a T cell has only two options when responding to antigen. Upon encounter, a response is elicited if the dose of antigen

(plus co-stimulation) can overcome the T cell’s activation threshold, and if the dose is below that cutoff value, then it does not respond. Of course, if there is no cognate antigen present or if the dose/affinity of antigen is too low to overcome the activation threshold, then the T cell will not respond as discussed extensively in Chapter 5. Overall, what this predicts is a binary on/off state of T cells based solely on presence or absence of (proper) antigenic signal. Unlike this binary model, the TAT model suggested that T cell activation reflects information processing along an analog scale where the integration of antigenic signals over time is a key feature of the logic used to drive T cell activation.

First, as we discussed in Chapters 3 and 4, all resting (naïve) T cells have a certain activation threshold associated with their TCR. In any model for T cell activation, including the TAT model, the critical requirement is for antigenic signals to be strong enough to overcome this threshold. It is what happens next that the TAT model offers a unique perspective, that is quite relevant to our data on chronic versus acute T cell activation in

Chapter 5.

240

The relevant and central tenet of the TAT model is that T cells adjust their activation threshold every time they experience antigenic stimulus. A strong antigenic signal, in addition to eliciting activation, will also trigger a proportional feedback, that acts to ‘shut off’ the signal. The key is that this feedback (as discussed in Chapter 1) is delayed. It is during this window of delay (the absolute value of which is not defined in the TAT model) that positive signals which activate a T cell can propagate and trigger full T cell activation.

So, for any T cell to continue responding, the levels of antigenic signal they sense has to be on an upward trend. In other words, the positive signal from the antigen via the TCR has to continue to increase before the delayed negative signal from the last antigen signal catches up and shuts it off (Figure 6.1).

241

Figure 6.1. Delayed negative feedback adjusts activation thresholds.

T cells receive signals from pMHC complexes that produces some ‘amount’ of positive signal. That ‘amount’ of signal must be greater than the intrinsic balance of regulators that mediate the signal propagation. As a T cell continues to see that same ‘amount’ of pMHC, there is an accumulation of negative feedback that respond, albeit at a delay, to balance the pMHC signals.

242

This image of an adaptable lymphocyte is consistent with the data of reversibility and signaling adjustment that we find in Chapter 5. Indeed, as observed in our model system, in order to have an adaptable lymphocyte, Grossman and Paul suggested that the

TCR is capable of tuning its sensitivity to antigen by dynamically altering its activation threshold [43]. In our model of chronic antigen stimulation, we can envision the continual antigenic signal allowing such TCR-proximal negative feedback to accumulate and form a stable counteraction to the positive signals that are transduced from the pMHC:TCR interaction and therefore renders the T cells hyporesponsive (Figure 6.2). Overall, this process results in an increased activation threshold within the original T cell where a larger amount of ligand interaction is needed for the generation of the same biological output than what was previously needed. As with relevance to the model system we use, these aspects of the TAT model are quite relevant to phenomena such as exhaustion in T cells responding to tumors and chronic infections.

243

Figure 6.2. Chronic antigen stimulation elicits continual negative feedback to increase activation thresholds of T cells.

Acute stimulation of pMHC proceeds through a series of biochemical signaling pathways that ultimately generates a T cell response through transcriptional alterations. Slow changes in antigen or the continual presence of antigen allows the accumulation of negative feedback along the biochemical pathways which then shut down those pathways to prevent induction of transcriptional changes and T cell responses.

244

The value of a theoretical paradigm is in being able to offer new hypotheses – especially ones that may not have been obvious without the framework offered by the new theory. In that light, the data we have discussed so far, while consistent with the TAT model, does not necessarily need this framework to be studied. In other words, looking at altered TCR signaling after chronic T cell activation is a natural next step after the observation that we and others have when studying T cell responses in vivo. Yet the next tenet that is offered by the TAT model is not necessarily so linearly linked to previous data.

Since the TAT model envisions tuning of the T cell as a result of competing positive and negative signals in each T cell, it makes some intriguing predictions regarding the fundamental logic of T cell activation itself. Although we have discussed the positive and negative signals so far in very conventional terms, in the TAT model these are along a continuum (Figure 6.3, Copyright Clearance Center Marketplace #1053580). T cells are constantly signaling via their TCRs using low affinity self-peptides and cross-reactive ligands as briefly discussed in Chapter 4, many of which may be too weak to trigger complete T cell activation. Yet, in the TAT model, these weak signals (called ‘sub- threshold interactions’ in the model) also elicit proportional negative feedback [43].

Therefore, a T cell’s activation threshold is constantly adapting to its environmental pMHC milieu whether they are strong or weak signals. The critical parameter is the delay involved in the recruitment of negative feedback for T cell activation. For weak signals that do not activate a T cell fully, the activation threshold is adjusted after a while, even though the T cell never really responded to the antigen. If the antigen in the environment increases slightly thereafter, the T cell is unlikely to be activated still since the initial signal elicited generation of a negative feedback response. Remember that the T cell has already adjusted

245 its activation threshold to be higher, so the new and ‘slightly’ increased antigen levels will still be subthreshold. This cycle can continue indefinitely if there is a continual and gradual increase in environmental antigen with the caveats that the increase in antigen is (a) small enough to be still within the new activation threshold and (b) slow enough to only change pMHC levels after the threshold has been adjusted by the delayed onset of negative feedback from the previous signal. This could also be evident between T cells that have been in the periphery receiving pMHC signals tonically compared to recent thymic emigrants. In other words, if the antigen levels change too quickly – before the negative feedback is set in place from the earlier change, then the new change will elicit an above- threshold response from the T cell resulting in activation. The sum total of these arguments is that T cell activation is critically reliant on the rate of change of antigen presented by the

APC, in addition to the absolute amounts itself alone [43]. The elegance of this simple precept is that it can even serve as an alternate model for T cell activation – changing the way we currently interpret the two-signal based model.

246

Figure 6.3. Tunable Activation Threshold model kinetic representation.

The TAT model proposes the element of time as a key determinant of T cell activation. Each panel represents a different amount of signal from antigenic stimulus (blue line). The black line and grey line are the activation threshold and the intrinsic negative feedback response (termed de-excitation index) within the cell, respectively. Each situation has a basal starting activation threshold yet, with changes to the signal, this threshold is altered. Over a period of time, the signal that a T cell comes into contact with varies. The signal will elicit delayed negative feedback responses and those negative feedback responses will increase the activation threshold of the cell. Such that in situations where there are either weak antigenic signals or continual strong signals there is no apparent activation (A-C). In contrast, the only situation that elicits activation is one in which there is a dramatic increase of signal over a short duration of time. This does not provide enough time for proportional negative feedback to dampen these signals and as such the T cell is activated (D). Figure adapted from [43].

247

In this model, the discrimination between self and ‘foreign’ is not simply made on the basis of PAMPs or other contextual indicators. Instead, the critical discriminator would be rate of change of the antigenic stimulus (Figure 6.3). If an antigen changes too quickly in the local tissue environment (or for that matter, all over the body) then the T cell is able to sense those changes and respond. T cells would get activated and begin an adaptive response cascade; this would ensure that the adaptive response will effectively respond to sources of antigen including fast replicating viruses, bacteria, tumors etc. On the other hand, if the antigen levels do not change at all (rate of change = 0) whether at high or low absolute amounts, or if the antigen levels change too slowly, then the T cells ‘assume’ that these are part of healthy changes in your tissue (growth, puberty, pregnancy, etc.) and avoid responding. Indeed, this would also predict that T cells have a harder time controlling chronic versus acute infections – which of course is a relevant issue in T cell responses as discussed in Chapter 5. The teleological incentive for developing such a paradigm can be discussed further; but it does resolve some limitations of the two-signal model. On the other hand, it is possible that co-stimulation itself is a ‘rate-changer’ for TCR signaling, since it contributes a sudden burst of positive signals towards the core TCR signaling pathway itself. Since it amplifies many of the signaling cascades downstream of the TCR.

Twenty-eight years after the tuning model was proposed, while the concept has been widely acknowledged, studies that directly examine this central question – i.e. rate of change of antigen as a driver of T cell activation itself have been limited. The fact that

T cell activation, particularly in vivo, is usually measured using readouts that are influenced by multiple signals and inputs has made a clear dissection of TCR-level tuning challenging. Clearly negative feedback in pathways of co-stimulation [404] , cytokine

248 signaling [12, 405], cell adhesion pathways [406], and chromatin remodeling [407] can all influence T cell responses. However, the teleological identity and operational logic of a T cell is tightly linked with the TCR itself. It is therefore important to clearly delineate how tuning of this central receptor operates and impacts of the T cell response as a whole.

6.2. Chapter Summary

The key question we set out to examine in this chapter is whether changing the rate at which antigen increases in vivo can determine ‘GO’ versus ‘STOP’ signals for T cells.

We focused on developing a system in which antigen can be increased at very slow rates

(so that a T cell would not be activated, even when antigen levels reached high absolute concentrations). Since the TAT model (or any available literature) does not specify what this rate could be, we focused on empirically determining rates that are either activating or non-activating. Preliminary data optimizing two different experimental approaches towards this goal are discussed in this Chapter.

6.3. Results

6.3.1. Defining a TCR-Tg murine model system to investigate the role of antigen rate of change

In order to isolate antigen rate of change as the sole variable within an experimental system, we used an adoptive transfer mouse model where TCR-tg CD4+ T cells were transferred into a wild-type host and then subsequently injected with peptide antigen without any adjuvant or other stimulatory reagents (Figure 6.4). We used 5C.C7 TCR-tg

CD4+ T cells adoptively transferred into wild type B10.A animals with dual congenic expression. This allowed us to track the antigen specific T cells within an endogenous

249 population of T cells. Moreover, we labeled the cells with CFSE to track the proliferation in vivo. We transferred these cells at a low density (under 1 – 2.5 x 105 cells) to prevent any supraphysiological population of these cells. The seeding density of adoptively transferred cells is approximately 10% so, the expected recovery from these transfers was approximately 1 – 2.5 x 104 cells. Within this system, we were able to identify cells based on the congenic expression to evaluate the response through CFSE dilution as a marker of proliferation and with upregulation of activation markers within the transferred population

(Figure 6.5).

250

Figure 6.4. Murine model system to investigate antigen rate of change.

5C.C7 TCR-tg CD4+ T cells were isolated and CFSE labeled before adoptive transfer into B10.A hosts. These hosts are then subjected to daily intraperitoneal injections based on their experimental grouping. Three days after the final injection, T cells are recovered and analyzed via flow cytometry.

251

Figure 6.5. Gating strategy for identifying adoptively transferred populations

Spleen and lymph nodes from hosts bearing adoptively transferred 5C.C7 T cells were isolated, purified, and stained for cytometric analysis. Cells were gated to separate congenic populations to identify donor versus endogenous T cell populations.

252

To address whether the antigen rate of change alters activation thresholds, we set up four experimental groups receiving daily i.p. injections. One group with PBS only injections (PBS, green), a second with injections of low dose antigen with no change in the rate (ML, blue), a third with PBS injections and a final challenge injection with a bolus of antigen (CH, red), and the final group will receive escalating doses of antigen to simulate a particular rate of change (MV, purple) (Figure 6.4). The PBS group served as a vehicle control for the peptide injections as well as to control for any variables introduced by daily intraperitoneal injections (inflammation from injection site, stress from injection, etc). The second experimental group received daily injections of low dose peptide. This group served to provide a baseline of what the T cell activation state is in the presence of antigen that was below an activating dose with no rate of change. This discerned if there was any accumulation of antigen within the system that changes the activation state of the T cells.

The third experimental group was one which received PBS daily and the final injection was a high dose bolus of antigen. This provided a comparison to what a challenge antigen stimulation was the cells were seeded within the host for a number of days. The final experimental group was one that has an escalating antigen dose of a particular rate of change. The rates of changes are defined below within each experiment described.

Since there were experimental groups with cells that would not be activated, we optimized a purification protocol to enrich for cells of interest and a flow cytometric approach to identify the cells. As discussed in Chapter 2, T cells were purified by both density gradient separation and magnetic depletion. The remaining enriched CD4 T cells were then stained for flow cytometric analysis by using congenic markers to separate the adoptively transferred population from the endogenous cohort of T cells. Moreover, since

253 we were tracking a transgenic population, we were able to use the specific Vβ chain to identify these cells. Representative flow plots are shown of the gating strategy used for subsequent experiments in Figure 6.5.

6.3.2. Sequential exposure to low dose antigen does not elicit T cell activation

To determine an in vivo activating dose of antigen, we used previous in vivo activation models with 5C.C7 TCR-tg CD4+ T cells to guide our initial experiments [408].

Given that we had a starting in vivo activating dose, we first wanted to determine the role of continued low dose antigen exposure. We injected low dose antigen (0.01μg/100μL) for

5 days and analyzed the activation of the adoptively transferred population (Figure 6.6A).

We found that there was no increase in size, activation marker expression (CD25 or CD44), or dilution of CFSE in these populations (Figure 6.6B and Figure 6.6C). This provided a control group for our overarching experiments where any depot effect does not cause activation of these cells so, we were able to separate that variable from the antigen rate of change in our experimental model system.

254

Figure 6.6. Multiple low dose antigen does not activate T cells.

5C.C7 TCR-tg CD4+ T cells were isolated and CFSE labeled before adoptive transfer into B10.A hosts. These hosts are then subjected to daily intraperitoneal injections based on their experimental grouping as indicated. Three days after the final injection, T cells are recovered and analyzed via flow cytometry. (A) Experimental set up of five-day daily injections of low does PCC peptide and protein. (B) Histograms of FSC, CD25, CD44, and CFSE expression in adoptively transferred populations. All animals from experiment are shown in histograms. (C) MFI and gMFI of FSC, CD25, and CD44 as well as frequency of undivided cells (as indicated by CFSE expression). (n = 2-5 per experimental group from 1 independent experiment) Significance was calculated using one way ANOVA with Tukey’s multiple comparisons test. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

255

6.3.3. Escalating antigen dose has an altered activation phenotype compared to bolus antigen injection

We then moved to our experimental model system with daily injections of escalating antigen dose over an eight-day period with a rate of change as shown in Figure

6.7. With this pilot injection dosing regimen, there was activation in all groups that received antigen (Figure 6.8), including the group that received low dose antigen which was not seen in our previous experiments. So, we decided to re-evaluate our starting doses and change the escalation of antigen.

We therefore augmented the initial dosing by lowering the starting point and increasing the duration of injections to a seventeen-day period (Figure 6.9). With this injection-dosing regimen, we found that there was an increase in size, expression of CD44, as well as a decrease in the population of undivided cells all indicating that the T cells that received the escalating antigen dose were activated. Interestingly, these cells had increased expression of CD44 and a lower average frequency of undivided cells compared to the challenge group (Figure 6.10). This suggests that the escalating dose of antigen may have activated these cells prior to the last day of injections since they have divided more than cells that had single bolus challenge.

256

Figure 6.7. Escalating antigen dose over eight-day experimental design.

5C.C7 TCR-tg CD4+ T cells were isolated and CFSE labeled before adoptive transfer into B10.A hosts. These hosts are then subjected to daily intraperitoneal injections for eight days based on their experimental grouping. Three days after the final injection, T cells are recovered, purified, and analyzed via flow cytometry. (A) Experimental design. (B) Injection dose for each day. (C) Graphed dose escalation.

257

Figure 6.8. Escalating dose over eight days activates T cells in vivo.

5C.C7 TCR-tg CD4+ T cells were isolated from experimental hosts as depicted in Figure 6.7. (A) Flow plots from each individual animal with expression of CD44 and CFSE from PBS injected (PBS, green), bolus challenged (CH, red), low dose injected (ML, blue), and escalating dose (MV, purple).

258

Figure 6.8. Escalating dose over eight days activates T cells in vivo (cont.). 5C.C7 TCR-tg CD4+ T cells were isolated from experimental hosts as depicted in Figure 6.7. (B) MFI and gMFI of FSC and CD44 as well as frequency and number of undivided cells (as indicated by CFSE expression) from PBS only (PBS, green), multiple low dose injections (ML, blue), bolus antigen challenge (CH, red), and escalating doses (MV, purple). (n = 2-3 per experimental group from 1 independent experiment) Significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

259

Figure 6.9. Escalating dose over seventeen-day experimental design.

5C.C7 TCR-tg CD4+ T cells were isolated and CFSE labeled before adoptive transfer into B10.A hosts. These hosts are then subjected to daily intraperitoneal injections for seventeen days based on their experimental grouping. Three days after the final injection, T cells are recovered and analyzed via flow cytometry. (A) Experimental design. (B) Injection dose for each day. (C) Graphed dose escalation.

260

Figure 6.10. Escalating dose over seventeen days activates T cells in vivo.

5C.C7 TCR-tg CD4+ T cells were isolated from experimental hosts as depicted in Figure 6.9. (A) Flow plots from each individual animal with expression of CD44 and CFSE from PBS injected (PBS, green), bolus challenged (CH, red), low dose injected (ML, blue), and escalating dose (MV, purple).

261

Figure 6.10. Escalating dose over seventeen days activates T cells in vivo (cont.). 5C.C7 TCR-tg CD4+ T cells were isolated from experimental hosts as depicted in Figure 6.9. (B) MFI and gMFI of FSC and CD44 as well as frequency and number of undivided cells (as indicated by CFSE expression) from PBS only (PBS, green), multiple low dose injections (ML, blue), bolus antigen challenge (CH, red), and escalating doses (MV, purple). (n = 2-3 per experimental group from 1 independent experiment) Significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

262

We decided to take a step back and evaluated the longevity of the peptide within the model to understand the kinetics of the peptide in vivo. We i.p. injected a high-dose bolus of peptide antigen into B10.A animals and then bled those animals at varying time points to evaluate the serum concentration of the peptide. The serum was plated with naïve

5C.C7 T cells overnight and subsequently measured expression of the early activation marker CD69 (Figure 6.11). We found that serum taken 1 hour after i.p. injection had a significant loss of expression of CD69 on naïve T cells compared to earlier time points indicating that there was a loss in circulating peptide as compared to the earlier time points.

This suggested that the daily injection experimental set up that we had originally designed may be inherently flawed. Given the quick clearance of peptide from serum, daily injections would ultimately result in an extremely varied rate of change of the antigen within the system with continual peaks and valleys upon peptide turnover.

263

1 0 0

)

% (

8 0

y

c n

e 6 0

u

q

e r

4 0

F

+

9 6

2 0

D C

0

r M m m h u 0 0 1 0 1 3

3 0 0 u g I . P . P C C

Figure 6.11. Daily injections results in antigen decay quickly and has fluctuating rate of change of antigen

B10.A mice intraperitoneally challenged with indicated dose of PCC. At varying time points after injection, mandibular bleeds were performed to isolate serum. Serum was plated with naïve 5C.C7 TCR-tg CD4+ T cells and splenocytes from CD3ε-KO mice and incubated for 15 hours. Cells were then harvested and stained for flow cytometric analysis of CD69 frequency.

264

6.3.4. Osmotic pump delivery of escalating doses of antigen primes T cells for activation earlier than a single high antigen dose

In order to address the issue of peptide clearance within the system in vivo we turned to an osmotic pump system. These pumps were programmable such that we are able to dictate the antigen rate of change and provide continual delivery of antigen so that there was no bolus of delivery via i.p. injection that dissipated over the next 24 hours. We adapted our previous experimental model system to implant the pre-programmed osmotic pump 24 hours prior to the adoptive transfer of the 5C.C7 TCR-tg T cells into the hosts.

Upon transfer, the animals were monitored and the pump fluid was exchanged as necessary to maintain the antigen dosing as shown in Figure 6.12.

265

Figure 6.12. Implantable mini osmotic pumps provide continuous antigen delivery so there is steady delivery of antigen

B10.A animals were implanted with iPrecio mini programmable osmotic pumps and allowed to recover for 24 hours. 5C.C7 TCR-tg CD4+ T cells were isolated and CFSE labeled before adoptive transfer into pump bearing B10.A hosts. Three days after final antigen delivery, T cells are recovered and analyzed via flow cytometry.

266

To first validate the antigen delivery with this pump system, animals were bled to collect serum to validate the peptide concentration using the same overnight assay to measure CD69 expression from naïve 5C.C7 T cells. The pumps were programmed to release 24μg of peptide over a 24-hour period to examine high antigen dissemination since we had not previously validated this system. The animal was bled, serum was collected, and plated with naïve 5C.C7 T cells overnight and subsequently analyzed via flow cytometry. With serum at varying concentrations, there was no detectable CD69 expression

(Figure 6.13). This experiment was repeated with increased serum concentrations in the overnight assay as well as higher peptide delivery without any success potentially due to technical limitations with the assay if the serum concentration of the peptide was not sufficient for in vitro activation of naïve cells. To functionally validate if the pumps were delivering any antigen in vivo, we increased the pump antigen delivery to 240μg per day and adoptively transferred naïve CFSE labeled 5C.C7 TCR-Tg cells into the pump-bearing host. The cells were recovered from the animal three days after for flow cytometric analysis. We found that despite the negative data from the ex vivo assay, there was robust proliferation of the adoptively transferred cells in vivo which suggests that there was peptide delivery from the pump (Figure 6.14).

267

Figure 6.13. Overnight culture assay to optimize antigen delivery with mini osmotic pumps.

(A) Experimental design for analyzing pump antigen delivery. B10.A animals were implanted with iPrecio mini programmable osmotic pumps and allowed to recover for 24 hours. Mandibular bleeds were performed to isolate serum. Serum was plated with naïve 5C.C7 TCR-tg CD4+ T cells and splenocytes from CD3ε-KO mice and incubated for 16 hours. Cells were then harvested and stained for flow cytometric analysis of CD69 frequency. (B) Histogram of CD69 expression from control (grey), peptide stimulated (red), or serum stimulated (blue) 5C.C7 cells. (C) gMFI of CD69 expression from cells treated as indicated.

268

Figure 6.14. Adoptive transfer of naïve 5C.C7 T cells to optimize antigen delivery with mini osmotic pumps

(A) Experimental design for analyzing pump antigen delivery. B10.A animals were implanted with iPrecio mini programmable osmotic pumps and allowed to recover for 24 hours. After 10 days of pump antigen delivery, 5C.C7 TCR-tg CD4+ T cells were adoptively transferred into the hosts. Three days later cells were recovered and analyzed via flow cytometry (B) Histogram of CFSE expression from adoptively transferred 5C.C7 T cells.

269

Since we had a functional validation of the pump delivery system, we proceeded with a pilot experiment as outlined in Figure 6.15. We found that there was activation as evidenced by increased expression of activation markers such as CD44, CD69, and CD25

(Figure 6.16A – 6.16F). Additionally, there was increased proliferation of the adoptively transferred population as indicated by the decreased frequency and number of undivided cells (Figure 6.16H). This suggests that the rate at which the pumps were programmed was activating for the adoptively transferred population. Interestingly, based on the expression of the challenge group, there was significant downregulation of CD69 within the rate of change group compared to the challenge group (Figure 6.16C and 6.16D). This is expected since CD69 is an early activation marker and is downregulated as T cells begin to proliferate [409]. Yet, this suggests that the escalating dose experimental group was potentially activated at an earlier time point than the challenge group. Taken together, this suggest that rather than desensitizing the T cells, the increasing antigen may have actually sensitized T cells to be activated at a lower dose of antigen than would be expected as a single bolus challenge.

270

Figure 6.15. Escalating antigen dose delivered by osmotic pump activates T cells.

B10.A animals were implanted with iPrecio mini programmable osmotic pumps and allowed to recover for 24 hours. 5C.C7 TCR-tg CD4+ T cells were isolated and CFSE labeled before adoptive transfer into pump bearing B10.A hosts. Three days after the final antigen delivery, T cells are recovered, purified, and analyzed via flow cytometry. (A) Experimental design. (B) Antigen delivery for each day. (C) Graphed dose escalation.

271

Figure 6.16. Escalating antigen dose delivered by osmotic pump activates T cells

5C.C7 TCR-Tg CD4+ T cells were isolated from hosts as depicted in Figure 6.16 for flow cytometric analysis. Flow plots from each individual animal with expression of CD44 and CFSE (A), CD69 and CFSE (C), and CD25 and CFSE (E) from PBS injected (PBS, green), bolus challenged (Challenged, red), or escalating dose (Varying Dose (pump), purple). gMFI of CD44 (B), CD69 (D), CD25 (F), and MFI of FSC (G) are shown for each experimental group. (H) Frequency and number of undivided cells (as indicated by CFSE expression) from experimental groups. (n = 2 per experimental group from 1 independent experiment) Significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

272

Figure 6.16. Escalating antigen dose delivered by osmotic pump activates T cells (cont.). 5C.C7 TCR- Tg CD4+ T cells were isolated from hosts as depicted in Figure 6.16 for flow cytometric analysis. Flow plots from each individual animal with expression of CD44 and CFSE (A), CD69 and CFSE (C), and CD25 and CFSE (E) from PBS injected (PBS, green), bolus challenged (Challenged, red), or escalating dose (Varying Dose (pump), purple). gMFI of CD44 (B), CD69 (D), CD25 (F), and MFI of FSC (G) are shown for each experimental group. (H) Frequency and number of undivided cells (as indicated by CFSE expression) from experimental groups. (n = 2 per experimental group from 1 independent experiment) Significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test. Significance is indicated as follows: p<0.05 = * ; p<0.01 = ** ; p<0.001 = *** ; p<0.0001 = ****. Anything not marked is not statistically significant.

273

6.4. Discussion

With these experiments we have designed an experimental system to isolate the antigen rate of change as a variable for T cell activation. Using an adoptive transfer model system, we are able to track an antigen-specific population of T cells within an endogenous repertoire. We first used an injection-based strategy but found that this had drawbacks due to antigen dissipation within the serum (Figure 6.11). We then moved to using an implantable programmable osmotic pump that provides continual delivery of antigen over the duration of the experiment. Since the optimization of these systems was technically challenging, we have only been able to perform a single pilot experiment with a single dose escalation regimen.

Within this pilot study we find that the cells that were exposed to an escalating dose of antigen were activated compared to their non stimulated controls (Figure 6.16).

Moreover, we find that these cells have decreased expression of CD69 compared to the challenge group (Figure 6.16C and 6.16D). This indicates that the cells exposed to an escalating dose were likely activated earlier than the cells that received the challenge dose.

There are two ways to interpret this data; one is that the activating dose of antigen is different from what was previously optimized in the injection system such that a lower dose of antigen is activating to naïve cells. A second interpretation is that the presence of antigen actually decreased the activation threshold of these cells to make them more sensitive to antigen. Since this is a single pilot experiment with a low number of mice used, it is only speculation to draw anything from these experiments. Moreover, there are many questions to be answered about the effective of the osmotic pump delivery system since it

274 is a surgical implantation that could initiate damage signals along with the peptide delivery which could impact how the T cells sense the peptide.

Given that we now have a system for investigating the rate of change of antigen, the interesting question becomes testing many different rates of antigen delivery. This is extremely clinically relevant when discussing tolerance induction as is the goal for allergy therapies [410]. The question really becomes where to start when deciding which rates to test? Since this is a novel paradigm of understanding T cell activation, even defining the activating and non-activating rate becomes informative. If a certain rate augments the activation threshold, whether activating or non-activating, then that suggests that the rate of change does have an influence on the overall activation of T cells.

275

Chapter 7: Significant Findings and Future Directions

7.1. The initial setting of the T cell activation threshold

7.1.1. Significant Findings

We identified a novel regulatory pathway between the surface receptor CD5 and an inhibitor of NFκB, IκBα. We found that CD5 was necessary for setting the basal levels of

IκBα and NFκB signaling in thymocytes. Additionally, IκBα levels were dynamically adjusted as T cells develop in the thymus – correlating with each TCR-dependent step – and mirrored the dynamics of CD5 expression. Intriguingly, in the absence of CD5, developing T cells only retained a basal level of IκBα, even though all TCR-dependent- selection steps were completed. This suggests that even in the presence of intact TCR signaling, CD5 is critical for setting the NFκB signaling threshold in thymocytes.

Additionally, we found that enhanced IκBα had an NFκB dependent pro-survival function in thymocytes with higher CD5 expression.

7.1.2. Future Directions

Identifying the molecular mechanisms that maintains the relationship between CD5 and IκBα may reveal a previously underappreciated mechanism for the phenotypes observed with high CD5 expression. Moreover, these data suggest that the heterogeneity of T cells begins within the thymus and impacts the ways in which these cells are poised to react after they are in the periphery.

7.2. Origins of signaling heterogeneity in steady-state peripheral T cells

7.2.1. Significant Findings

We evaluated the novel regulatory pathway identified in Chapter 3 in peripheral T cells. Similar to our findings in thymocytes, the relationship between CD5 and IκBα was

276 maintained by continuous expression of CD5. Examining signaling molecule expression in peripheral T cells, we found that most critical signaling molecules were relatively homogenously expressed by peripheral T cells, although there were differences between naïve, memory and regulatory T cells. Intriguingly, even within naïve and memory T cells, two signaling molecules showed an extended range of expression – PLCγ and IκBα. Given the data from Chapter 3 and the previously reported heterogeneity in CD5 expression, we evaluated any correlation between CD5 and these molecules. While CD5 levels did not correlate with PLCγ, they showed a linear relationship to IκBα. To further explore the axis between CD5 and IκBα, we used both in vivo and in vitro models of overexpression of

CD5 and subsequently measured IκBα. In vivo, we found there is a two-step regulation of

IκBα expression. First, CD5 expression set a basal level of IκBα but this was a saturable adjustment, as CD5 transgenic overexpression in mice did not result in a proportional upregulation of IκBα. Secondly, IκBα expression was robustly upregulated during activation beyond the basal saturation point that was seen within the overexpression animal models. This suggests that there is a post-transcriptional regulatory axis between CD5 and

IκBα, in contrast to a transcriptional upregulation of IκBα upon TCR stimulation. Further mechanistic insight into CD5-depedent regulation of IκBα was obtained from the overexpression of CD5 in a TCR deficient lymphoma cell line suggesting that CD5 directly stabilizes IκBα protein levels, even in the absence of TCR-signals. Interestingly, the upregulation of IκBα from CD5 overexpression was not seen in a fibroblast cell line or if the expression vector disrupts the cytoplasmic tail of CD5. Taken together these data suggest that the expression of CD5, a canonical negative regulator, post-transcriptionally regulates the NFκB pathway in peripheral T cells potentially explaining some of the

277 heterogeneity within peripheral T cell responses to antigen. Finally, to address the role of increased IκBα on NFκB function, we used localization studies and found that there is increased NFκB expression in CD5hi cells in both CD4 and CD8 T cells. The increased expression of NFκB may poise CD5hi cells to respond more quickly upon stimulation since there is an accumulation of the transcription factor.

7.2.2. Future Directions

Similar to the thymocyte data, unraveling how the relationship between CD5 and

IκBα are maintained in the periphery may provide some insight into the data previously attributed to self-pMHC interactions. In a similar vein, dissecting the roles of TCR and

CD5 from these functional consequences is of interest as well since the field currently uses

CD5 as a marker for self-reactivity and parsing the exact contribution of each of these molecules may provide insight into better vaccine design since it has been shown that these correlate with better memory responses after pathogenic challenge [248, 249, 251].

7.3. Reversible activation threshold modulation from chronic antigen stimulation

7.3.1. Significant Findings

To more broadly examine activation threshold dynamics in the periphery we used a model of self-antigen stimulation side by side with an acute-antigen stimulation model to understand the molecular changes that occur between these two states. We first evaluated the loss of proliferative capacity and cytokine production in both a lymphopenic and lympho-replete model system. We also performed initial experiments for both in vivo and in vitro model systems of antigen deprivation and found that in vitro deprivation does not recover any function while in vivo deprivation has variability in proliferative capacity recovery. Interestingly, we also find that there is a division within the population of

278 chronically stimulated cells; one in which high-dose antigen allows proliferation to be restored and a second in which it is not. Furthermore, we investigated pharmacological agents that could restore hyporesponsive T cell function but from those tested, did not find any promising recovery of function. Finally, we began to characterize some of the molecular changes within these distinct states and find both proximal and distal dysregulation. We find that there is an increase in total Zap70 in chronically stimulated T cells compared to acutely stimulated or naïve cells, yet, the phosphorylated forms of Zap70 are decreased. This suggests that while there is ample Zap70 available to propagate TCR signals, it is a faulty circuit. Beyond this proximal T cell signaling regulator, we also find an accumulation of CD5 and IκBα as well as a key transcription factor, NFAT, in chronically stimulated T cells. Interestingly, Tox, another transcription factor, is increased in chronically stimulated cells; consistent with findings from last year in multiple cancer and chronic viral infection model systems [383, 385-388]. Furthermore, we find an increase in the expression of checkpoint molecules, especially PD-1, which suggests that these cell likely upregulate similar molecular programs after chronic stimulation to those in cancer and chronic viral infection.

7.3.2. Future Directions

Since we have now set up systems for understanding reversal, performing signaling molecule analysis at different time points to potentially identify those that are altered during reversal may provide insight into what pathways are involved in maintaining the hyporesponsive phenotype in T cells. Additionally, while the self-antigen model provides a starting point for these analyses, examining similar signaling molecule signatures in other contexts of chronic stimulation would potentially reveal a unifying molecular program for

279

T cell hyporesponsiveness. Moreover, further interrogating the pharmacological screen may provide other interesting candidates such as, benzamidine which is a serine protease inhibitor, that could alter the balance of signaling molecules in chronic-antigen stimulated

T cells.

7.4. Tuning of activation thresholds and the role of rate of change of antigen in a quantitative framework for immune activation

7.4.1. Significant Findings

One unifying principle to understand the alteration of activation thresholds is through the lens of antigen rate of change. In the context of chronic-antigen exposure, the rate of change is low or zero and as such the T cell would experience this antigen as it would experience most self-antigens ultimately perceiving it is a non-threat and therefore the T cell does not respond. In contrast, most pathogenic insults provide a bolus of new antigen resulting in a high rate of change which the T cell perceives as a threat and responds robustly. We set up to understand the role of antigen rate of change on naïve T cell activation in two in vivo model systems. First, using daily peptide injections we found preliminary evidence that low-dose antigen and escalating dose of antigen did not elicit the same activation phenotype as a bolus challenge alone. Although this was promising data, it was not a perfect model system so we moved to using an implantable osmotic pump to ameliorate some of the shortcomings of daily injections. In a preliminary experiment, we found that animals that received an escalating dose of antigen were more activated than those animals that received a challenge dose alone. These data suggest that the even chronic stimuli, which is not a constant rate of change, can be activating. This is relevant to

280 comparisons between slowly replicating pathogens versus self-antigens, where the increase rates may differ.

7.4.2. Future Directions

Since much of the data presented in Chapter 6 was optimization and pilot experiments, there is now a wide array of questions that can be addressed using this model system but, most pressingly; what do different antigen rates of change do to T cells? With this model system we are able to program an almost infinite number of rates of change so, to first hone in on what rates may be activating or non-activating is the first step towards understanding the nuanced role rate of change may play in T cell activation.

281

Figure 7.1. Overview of thesis findings.

(A) In the thymus, T cells are subjected to a selection process that tests the ability of the TCR to signal. We found a novel link between the surface receptor CD5 and an inhibitor of NFκB, IκBα. CD5 (as indicated in orange) is upregulated during positive selection via TCR signals. This step proportionally upregulated IκBα. The upregulation of these molecules correlates with an increase in the activation threshold (AT) of thymocytes (as indicated in black) (B) The CD5- IκBα axis was maintained into the periphery and we found that NFκB expression also correlates with these two molecules (as indicated in grey). Upon acute antigen stimulation, there was an upregulation of both CD5 and IκBα. Interestingly, after antigen clearance from an acute stimulus, higher expression of both CD5 and IκBα was maintained. Acute antigen stimulus decreases the activation threshold of a T cells (as indicated in blue) (C) Chronic stimulation, similar to acute stimulation, upregulated CD5 and IκBα. Moreover, we found that molecules such as checkpoints as well as transcription factors associated with exhaustion, were also upregulated after chronic stimulation from a self-antigen. In this state there was a loss of proximal signal transduction likely through a decrease in Zap70 function. The activation threshold is increased during chronic stimulation (as indicated in red). We explored conditions that would restore function of cells and found that high dose antigen as well as in vivo antigen deprivation returned some function of T cells but only in select populations. In order to restore function of these cells the activation threshold would need to be decreased (as indicated in red). Generated using Biorender.

282

References

1. Chaplin, D.D., Overview of the immune response. J Allergy Clin Immunol, 2010.

125(2 Suppl 2): p. S3-23.

2. Brubaker, S.W., et al., Innate immune pattern recognition: a cell biological

perspective. Annu Rev Immunol, 2015. 33: p. 257-90.

3. Bonilla, F.A. and H.C. Oettgen, Adaptive immunity. J Allergy Clin Immunol,

2010. 125(2 Suppl 2): p. S33-40.

4. Lazebnik, Y., Can a biologist fix a radio?--Or, what I learned while studying

apoptosis. Cancer Cell, 2002. 2(3): p. 179-82.

5. Iwasaki, A. and R. Medzhitov, Regulation of adaptive immunity by the innate

immune system. Science, 2010. 327(5963): p. 291-5.

6. Oda, K. and H. Kitano, A comprehensive map of the toll-like receptor signaling

network. Mol Syst Biol, 2006. 2: p. 2006.0015.

7. Freitas, A.A. and B. Rocha, Population biology of lymphocytes: the flight for

survival. Annu Rev Immunol, 2000. 18: p. 83-111.

8. Rathmell, J.C., et al., Akt-directed glucose metabolism can prevent Bax

conformation change and promote growth factor-independent survival. Mol Cell

Biol, 2003. 23(20): p. 7315-28.

9. Di Santo, J.P., et al., The common cytokine receptor gamma chain and the pre-T

cell receptor provide independent but critically overlapping signals in early

alpha/beta T cell development. J Exp Med, 1999. 189(3): p. 563-74.

10. Shevach, E.M., Mechanisms of foxp3+ T regulatory cell-mediated suppression.

Immunity, 2009. 30(5): p. 636-45.

283

11. Boyman, O., et al., Homeostatic maintenance of T cells and natural killer cells.

Cell Mol Life Sci, 2012. 69(10): p. 1597-608.

12. Zhu, J. and W.E. Paul, Peripheral CD4+ T-cell differentiation regulated by

networks of cytokines and transcription factors. Immunol Rev, 2010. 238(1): p.

247-62.

13. Hu, W. and C. Pasare, Location, location, location: tissue-specific regulation of

immune responses. J Leukoc Biol, 2013. 94(3): p. 409-21.

14. Embrace the complexity. Nature Immunology, 2009. 10(4): p. 325-325.

15. Davis, M.M., et al., T cells as a self-referential, sensory organ. Annu Rev

Immunol, 2007. 25: p. 681-95.

16. Yamane, H. and W.E. Paul, Early signaling events that underlie fate decisions of

naive CD4(+) T cells toward distinct T-helper cell subsets. Immunol Rev, 2013.

252(1): p. 12-23.

17. Morrison, D.K., MAP kinase pathways. Cold Spring Harb Perspect Biol, 2012.

4(11).

18. Bray, S.J., Notch signalling in context. Nature Reviews Molecular Cell Biology,

2016. 17(11): p. 722-735.

19. Wilkenfeld, S.R., C. Lin, and D.E. Frigo, Communication between genomic and

non-genomic signaling events coordinate steroid hormone actions. Steroids,

2018. 133: p. 2-7.

20. Chen, H., et al., Nuclear receptor coactivator ACTR is a novel histone

acetyltransferase and forms a multimeric activation complex with P/CAF and

CBP/p300. Cell, 1997. 90(3): p. 569-80.

284

21. Oñate, S.A., et al., Sequence and characterization of a coactivator for the steroid

hormone receptor superfamily. Science, 1995. 270(5240): p. 1354-7.

22. Li, H., P.J. Gomes, and J.D. Chen, RAC3, a steroid/nuclear receptor-associated

coactivator that is related to SRC-1 and TIF2. Proc Natl Acad Sci U S A, 1997.

94(16): p. 8479-84.

23. Hong, H., et al., GRIP1, a transcriptional coactivator for the AF-2 transactivation

domain of steroid, thyroid, retinoid, and vitamin D receptors. Mol Cell Biol,

1997. 17(5): p. 2735-44.

24. Anzick, S.L., et al., AIB1, a steroid receptor coactivator amplified in breast and

ovarian cancer. Science, 1997. 277(5328): p. 965-8.

25. Chakravarti, D., et al., Role of CBP/P300 in nuclear receptor signalling. Nature,

1996. 383(6595): p. 99-103.

26. Hanstein, B., et al., p300 is a component of an estrogen receptor coactivator

complex. Proc Natl Acad Sci U S A, 1996. 93(21): p. 11540-5.

27. Shang, Y., et al., Cofactor dynamics and sufficiency in estrogen receptor-

regulated transcription. Cell, 2000. 103(6): p. 843-52.

28. Tolhurst, R.S., et al., Transient over-expression of estrogen receptor-α in breast

cancer cells promotes cell survival and estrogen-independent growth. Breast

Cancer Res Treat, 2011. 128(2): p. 357-68.

29. Tai, H., N. Kubota, and S. Kato, Involvement of nuclear receptor coactivator

SRC-1 in estrogen-dependent cell growth of MCF-7 cells. Biochem Biophys Res

Commun, 2000. 267(1): p. 311-6.

285

30. Anolik, J.H., et al., Cooperative binding of estrogen receptor to DNA depends on

spacing of binding sites, flanking sequence, and ligand. Biochemistry, 1995.

34(8): p. 2511-20.

31. Brandman, O. and T. Meyer, Feedback loops shape cellular signals in space and

time. Science, 2008. 322(5900): p. 390-5.

32. Xiao, T., et al., Estrogen-regulated feedback loop limits the efficacy of estrogen

receptor-targeted breast cancer therapy. Proc Natl Acad Sci U S A, 2018.

115(31): p. 7869-7878.

33. Capuani, F., et al., Quantitative analysis reveals how EGFR activation and

downregulation are coupled in normal but not in cancer cells. Nat Commun,

2015. 6: p. 7999.

34. Ryu, J.W., et al., Paradoxical induction of growth arrest and apoptosis by EGF

via the up-regulation of PTEN by activating Redox factor-1/Egr-1 in human lung

cancer cells. Oncotarget, 2017. 8(3): p. 4181-4195.

35. Vacaflores, A., et al., Exposure of Human CD4 T Cells to IL-12 Results in

Enhanced TCR-Induced Cytokine Production, Altered TCR Signaling, and

Increased Oxidative Metabolism. PLoS One, 2016. 11(6): p. e0157175.

36. Li, Q., et al., IL-12-programmed long-term CD8+ T cell responses require

STAT4. J Immunol, 2006. 177(11): p. 7618-25.

37. Bezprozvanny, I., J. Watras, and B.E. Ehrlich, Bell-shaped calcium-response

curves of Ins(1,4,5)P3- and calcium-gated channels from endoplasmic reticulum

of cerebellum. Nature, 1991. 351(6329): p. 751-4.

286

38. Azeloglu, E.U. and R. Iyengar, Signaling networks: information flow,

computation, and decision making. Cold Spring Harb Perspect Biol, 2015. 7(4): p.

a005934.

39. Huang, C.Y. and T.H. Tan, DUSPs, to MAP kinases and beyond. Cell Biosci,

2012. 2(1): p. 24.

40. Marti, F., et al., Negative-feedback regulation of CD28 costimulation by a novel

mitogen-activated protein kinase phosphatase, MKP6. J Immunol, 2001. 166(1):

p. 197-206.

41. Kim, J.H., et al., Negative feedback control of osteoclast formation through

ubiquitin-mediated down-regulation of NFATc1. J Biol Chem, 2010. 285(8): p.

5224-31.

42. Grossman, Z. and W.E. Paul, Adaptive cellular interactions in the immune

system: the tunable activation threshold and the significance of subthreshold

responses. Proc Natl Acad Sci U S A, 1992. 89(21): p. 10365-9.

43. Grossman, Z. and W.E. Paul, Dynamic Tuning of Lymphocytes: Physiological

Basis, Mechanisms, and Function*. Annual Review of Immunology, 2015. 33: p.

677-713.

44. Bhalla, U.S. and R. Iyengar, Emergent properties of networks of biological

signaling pathways. Science, 1999. 283(5400): p. 381-7.

45. Huang, C.Y. and J.E. Ferrell, Jr., Ultrasensitivity in the mitogen-activated protein

kinase cascade. Proc Natl Acad Sci U S A, 1996. 93(19): p. 10078-83.

287

46. Lipshtat, A., et al., Design of versatile biochemical switches that respond to

amplitude, duration, and spatial cues. Proc Natl Acad Sci U S A, 2010. 107(3): p.

1247-52.

47. Ferrell, J.E., Jr., Self-perpetuating states in signal transduction: positive feedback,

double-negative feedback and bistability. Curr Opin Cell Biol, 2002. 14(2): p.

140-8.

48. Bagowski, C.P. and J.E. Ferrell, Jr., Bistability in the JNK cascade. Curr Biol,

2001. 11(15): p. 1176-82.

49. Pomerening, J.R., S.Y. Kim, and J.E. Ferrell, Jr., Systems-level dissection of the

cell-cycle oscillator: bypassing positive feedback produces damped oscillations.

Cell, 2005. 122(4): p. 565-78.

50. Koshland, D.E., Jr. and K. Hamadani, Proteomics and models for enzyme

cooperativity. J Biol Chem, 2002. 277(49): p. 46841-4.

51. Das, J., et al., Digital signaling and hysteresis characterize ras activation in

lymphoid cells. Cell, 2009. 136(2): p. 337-51.

52. Munsky, B. and G. Neuert, From analog to digital models of gene regulation.

Phys Biol, 2015. 12(4): p. 045004.

53. Ferrell, J.E., Jr. and S.H. Ha, Ultrasensitivity part I: Michaelian responses and

zero-order ultrasensitivity. Trends Biochem Sci, 2014. 39(10): p. 496-503.

54. Saba-El-Leil, M.K., C. Frémin, and S. Meloche, Redundancy in the World of

MAP Kinases: All for One. Front Cell Dev Biol, 2016. 4: p. 67.

55. Masel, J. and M.L. Siegal, Robustness: mechanisms and consequences. Trends

Genet, 2009. 25(9): p. 395-403.

288

56. Courtney, A.H., W.L. Lo, and A. Weiss, TCR Signaling: Mechanisms of Initiation

and Propagation. Trends Biochem Sci, 2018. 43(2): p. 108-123.

57. Navarro, M.N. and D.A. Cantrell, Serine-threonine kinases in TCR signaling. Nat

Immunol, 2014. 15(9): p. 808-14.

58. Acuto, O. and D. Cantrell, T cell activation and the cytoskeleton. Annu Rev

Immunol, 2000. 18: p. 165-84.

59. Sherman, E., V. Barr, and L.E. Samelson, Super-resolution characterization of

TCR-dependent signaling clusters. Immunol Rev, 2013. 251(1): p. 21-35.

60. Bhattacharyya, N.D. and C.G. Feng, Regulation of T Helper Cell Fate by TCR

Signal Strength. Front Immunol, 2020. 11: p. 624.

61. Wu, L., et al., Signaling from T cell receptors (TCRs) and chimeric antigen

receptors (CARs) on T cells. Cell Mol Immunol, 2020. 17(6): p. 600-612.

62. Hwang, J.R., et al., Recent insights of T cell receptor-mediated signaling

pathways for T cell activation and development. Exp Mol Med, 2020. 52(5): p.

750-761.

63. Smith-Garvin, J.E., G.A. Koretzky, and M.S. Jordan, T cell activation. Annu Rev

Immunol, 2009. 27: p. 591-619.

64. Gaud, G., R. Lesourne, and P.E. Love, Regulatory mechanisms in T cell receptor

signalling. Nat Rev Immunol, 2018. 18(8): p. 485-497.

65. Goyette, J., et al., How does T cell receptor clustering impact on signal

transduction? J Cell Sci, 2019. 132(4).

66. Love, P.E. and S.M. Hayes, ITAM-mediated signaling by the T-cell antigen

receptor. Cold Spring Harb Perspect Biol, 2010. 2(6): p. a002485.

289

67. Underhill, D.M. and H.S. Goodridge, The many faces of ITAMs. Trends Immunol,

2007. 28(2): p. 66-73.

68. Irving, B.A., A.C. Chan, and A. Weiss, Functional characterization of a signal

transducing motif present in the T cell antigen receptor zeta chain. J Exp Med,

1993. 177(4): p. 1093-103.

69. Irving, B.A. and A. Weiss, The cytoplasmic domain of the T cell receptor zeta

chain is sufficient to couple to receptor-associated signal transduction pathways.

Cell, 1991. 64(5): p. 891-901.

70. Letourneur, F. and R.D. Klausner, Activation of T cells by a tyrosine kinase

activation domain in the cytoplasmic tail of CD3 epsilon. Science, 1992.

255(5040): p. 79-82.

71. Wegener, A.M., et al., The T cell receptor/CD3 complex is composed of at least

two autonomous transduction modules. Cell, 1992. 68(1): p. 83-95.

72. Osman, N., et al., The protein interactions of the immunoglobulin receptor family

tyrosine-based activation motifs present in the T cell receptor zeta subunits and

the CD3 gamma, delta and epsilon chains. Eur J Immunol, 1996. 26(5): p. 1063-

8.

73. Hwang, S., et al., TCR ITAM multiplicity is required for the generation of

follicular helper T-cells. Nature Communications, 2015. 6(1): p. 6982.

74. Duplay, P., et al., p56lck interacts via its src homology 2 domain with the ZAP-70

kinase. J Exp Med, 1994. 179(4): p. 1163-72.

290

75. Thome, M., et al., Syk and ZAP-70 mediate recruitment of p56lck/CD4 to the

activated T cell receptor/CD3/zeta complex. J Exp Med, 1995. 181(6): p. 1997-

2006.

76. Yamasaki, S., M. Takamatsu, and M. Iwashima, The kinase, SH3, and SH2

domains of Lck play critical roles in T-cell activation after ZAP-70 membrane

localization. Mol Cell Biol, 1996. 16(12): p. 7151-60.

77. Wong, J., D. Straus, and A.C. Chan, Genetic evidence of a role for Lck in T-cell

receptor function independent or downstream of ZAP-70/Syk protein tyrosine

kinases. Mol.Cell Biol., 1998. 18(5): p. 2855-2866.

78. Pelosi, M., et al., Tyrosine 319 in the interdomain B of ZAP-70 is a binding site

for the Src homology 2 domain of Lck. J Biol Chem, 1999. 274(20): p. 14229-37.

79. Holmes, N., CD45: all is not yet crystal clear. Immunology, 2006. 117(2): p. 145-

55.

80. Hermiston, M.L., Z. Xu, and A. Weiss, CD45: a critical regulator of signaling

thresholds in immune cells. Annu Rev Immunol, 2003. 21: p. 107-37.

81. Byth, K.F., et al., CD45-null transgenic mice reveal a positive regulatory role for

CD45 in early thymocyte development, in the selection of CD4+CD8+

thymocytes, and B cell maturation. J Exp Med, 1996. 183(4): p. 1707-18.

82. Courtney, A.H., et al., CD45 functions as a signaling gatekeeper in T cells. Sci

Signal, 2019. 12(604).

83. Huse, M., The T-cell-receptor signaling network. J Cell Sci, 2009. 122(Pt 9): p.

1269-73.

291

84. Wang, H., et al., ZAP-70: an essential kinase in T-cell signaling. Cold Spring

Harb Perspect Biol, 2010. 2(5): p. a002279.

85. Balagopalan, L., et al., The linker for activation of T cells (LAT) signaling hub:

from signaling complexes to microclusters. J Biol Chem, 2015. 290(44): p.

26422-9.

86. Finco, T.S., et al., LAT is required for TCR-mediated activation of PLCgamma1

and the Ras pathway. Immunity, 1998. 9(5): p. 617-26.

87. Zhang, W., et al., Functional analysis of LAT in TCR-mediated signaling

pathways using a LAT-deficient Jurkat cell line. Int Immunol, 1999. 11(6): p. 943-

50.

88. Zhang, W., et al., Essential role of LAT in T cell development. Immunity, 1999.

10(3): p. 323-32.

89. Shen, S., et al., The essential role of LAT in thymocyte development during

transition from the double-positive to single-positive stage. J Immunol, 2009.

182(9): p. 5596-604.

90. Sommers, C.L., L.E. Samelson, and P.E. Love, LAT: a T lymphocyte adapter

protein that couples the antigen receptor to downstream signaling pathways.

Bioessays, 2004. 26(1): p. 61-7.

91. Jang, I.K., J. Zhang, and H. Gu, Grb2, a simple adapter with complex roles in

lymphocyte development, function, and signaling. Immunol Rev, 2009. 232(1): p.

150-9.

292

92. Koretzky, G.A., F. Abtahian, and M.A. Silverman, SLP76 and SLP65: complex

regulation of signalling in lymphocytes and beyond. Nat Rev Immunol, 2006.

6(1): p. 67-78.

93. Rincón, M., R.A. Flavell, and R.J. Davis, Signal transduction by MAP kinases in

T lymphocytes. Oncogene, 2001. 20(19): p. 2490-7.

94. Dong, C., R.J. Davis, and R.A. Flavell, MAP kinases in the immune response.

Annu Rev Immunol, 2002. 20: p. 55-72.

95. Davis, R.J., Signal transduction by the JNK group of MAP kinases. Cell, 2000.

103(2): p. 239-52.

96. O'Shea, C.C., et al., Raf regulates positive selection. Eur J Immunol, 1996.

26(10): p. 2350-5.

97. Alberola-Ila, J., et al., Selective requirement for MAP kinase activation in

thymocyte differentiation. Nature, 1995. 373(6515): p. 620-3.

98. Rincón, M., et al., Interferon-gamma expression by Th1 effector T cells mediated

by the p38 MAP kinase signaling pathway. Embo j, 1998. 17(10): p. 2817-29.

99. Dong, C., et al., JNK is required for effector T-cell function but not for T-cell

activation. Nature, 2000. 405(6782): p. 91-4.

100. Trebak, M. and J.P. Kinet, Calcium signalling in T cells. Nat Rev Immunol, 2019.

19(3): p. 154-169.

101. Joseph, N., B. Reicher, and M. Barda-Saad, The calcium feedback loop and T cell

activation: how cytoskeleton networks control intracellular calcium flux. Biochim

Biophys Acta, 2014. 1838(2): p. 557-68.

293

102. Xia, F., et al., TCR and CD28 Concomitant Stimulation Elicits a Distinctive

Calcium Response in Naive T Cells. Front Immunol, 2018. 9: p. 2864.

103. Lewis, R.S., Calcium signaling mechanisms in T lymphocytes. Annu Rev

Immunol, 2001. 19: p. 497-521.

104. Hayden, M.S. and S. Ghosh, NF-kB in Immunobiology. Cell Research, 2011: p.

223-244.

105. Oh, H. and S. Ghosh, NF-κB: roles and regulation in different CD4(+) T-cell

subsets. Immunol Rev, 2013. 252(1): p. 41-51.

106. Paul, S. and B.C. Schaefer, A new look at T cell receptor signaling to nuclear

factor-κB. Trends Immunol, 2013. 34(6): p. 269-81.

107. Diehn, M., et al., Genomic expression programs and the integration of the CD28

costimulatory signal in T cell activation. Proc Natl Acad Sci U S A, 2002. 99(18):

p. 11796-801.

108. Ishiguro, K., et al., Ca2+/calmodulin-dependent protein kinase II is a modulator

of CARMA1-mediated NF-kappaB activation. Mol Cell Biol, 2006. 26(14): p.

5497-508.

109. Hogan, P.G., Calcium-NFAT transcriptional signalling in T cell activation and T

cell exhaustion. Cell Calcium, 2017. 63: p. 66-69.

110. Alam, M.S., et al., Unique properties of TCR-activated p38 are necessary for

NFAT-dependent T-cell activation. PLoS Biol, 2018. 16(1): p. e2004111.

111. Crabtree, G.R. and E.N. Olson, NFAT signaling: choreographing the social lives

of cells. Cell, 2002. 109 Suppl: p. S67-79.

294

112. Shaw, J.P., et al., Identification of a putative regulator of early T cell activation

genes. Science, 1988. 241(4862): p. 202-5.

113. Serfling, E., et al., Ubiquitous and lymphocyte-specific factors are involved in the

induction of the mouse interleukin 2 gene in T lymphocytes. Embo j, 1989. 8(2): p.

465-73.

114. Macian, F., et al., Transcriptional Mechanisms Underlying Lymphocyte

Tolerance. Cell, 2002. 109: p. 719-731.

115. Macián, F., C. López-Rodríguez, and A. Rao, Partners in transcription: NFAT

and AP-1. Oncogene, 2001. 20(19): p. 2476-89.

116. Karin, M., The regulation of AP-1 activity by mitogen-activated protein kinases. J

Biol Chem, 1995. 270(28): p. 16483-6.

117. Shaulian, E. and M. Karin, AP-1 as a regulator of cell life and death. Nat Cell

Biol, 2002. 4(5): p. E131-6.

118. Rincon, M. and R.A. Flavell, AP-1 transcriptional activity requires both T-cell

receptor-mediated and co-stimulatory signals in primary T lymphocytes. Embo j,

1994. 13(18): p. 4370-81.

119. Ghosh, S. and M. Karin, Missing pieces in the NF-kappaB puzzle. Cell, 2002. 109

Suppl: p. S81-96.

120. Oeckinghaus, A. and S. Ghosh, The NF-kappaB family of transcription factors

and its regulation. Cold Spring Harb Perspect Biol, 2009. 1(4): p. a000034.

121. Zambrano, S., et al., NF-κB oscillations translate into functionally related

patterns of gene expression. Elife, 2016. 5: p. e09100.

295

122. Wang, J.H. and E.L. Reinherz, The structural basis of αβ T-lineage immune

recognition: TCR docking topologies, mechanotransduction, and co-receptor

function. Immunol Rev, 2012. 250(1): p. 102-19.

123. Birnbaum, M.E., et al., Molecular architecture of the αβ T cell receptor-CD3

complex. Proc Natl Acad Sci U S A, 2014. 111(49): p. 17576-81.

124. Huang, J., et al., A single peptide-major histocompatibility complex ligand

triggers digital cytokine secretion in CD4(+) T cells. Immunity, 2013. 39(5): p.

846-57.

125. Pageon, S.V., et al., Functional role of T-cell receptor nanoclusters in signal

initiation and antigen discrimination. Proc Natl Acad Sci U S A, 2016. 113(37):

p. E5454-63.

126. Dustin, M.L., The immunological synapse. Cancer Immunol Res, 2014. 2(11): p.

1023-33.

127. Govern, C.C., et al., Fast on-rates allow short dwell time ligands to activate T

cells. Proc Natl Acad Sci U S A, 2010. 107(19): p. 8724-9.

128. Chang, V.T., et al., Initiation of T cell signaling by CD45 segregation at 'close

contacts'. Nat Immunol, 2016. 17(5): p. 574-582.

129. Fernandes, R.A., et al., A cell topography-based mechanism for ligand

discrimination by the T cell receptor. Proc Natl Acad Sci U S A, 2019. 116(28):

p. 14002-14010.

130. Davis, S.J. and P.A. van der Merwe, The kinetic-segregation model: TCR

triggering and beyond. Nat Immunol, 2006. 7(8): p. 803-9.

296

131. Oda, K., et al., A comprehensive pathway map of epidermal growth factor

receptor signaling. Mol Syst Biol, 2005. 1: p. 2005.0010.

132. Kersh, G.J., et al., High- and low-potency ligands with similar affinities for the

TCR: the importance of kinetics in TCR signaling. Immunity., 1998. 9(6): p. 817-

826.

133. Stefanova, I., J.R. Dorfman, and R.N. Germain, Self-recognition promotes the

foreign antigen sensitivity of naive T lymphocytes. Nature, 2002. 420(6914): p.

429-34.

134. Hogquist, K.A., T.K. Starr, and S.C. Jameson, Receptor sensitivity: when T cells

lose their sense of self. Curr Biol, 2003. 13(6): p. R239-41.

135. Beeson, C., et al., Early biochemical signals arise from low affinity TCR-ligand

reactions at the cell-cell interface. J Exp.Med., 1996. 184(2): p. 777-782.

136. Singh, N.K., et al., Emerging Concepts in TCR Specificity: Rationalizing and

(Maybe) Predicting Outcomes. The Journal of Immunology, 2017. 199(7): p.

2203-2213.

137. Mukhopadhyay, H., et al., Systems model of T cell receptor proximal signaling

reveals emergent ultrasensitivity. PLoS Comput Biol, 2013. 9(3): p. e1003004.

138. Mukhopadhyay, H., et al., Multisite Phosphorylation Modulates the T Cell

Receptor ζ-Chain Potency but not the Switchlike Response. Biophys J, 2016.

110(8): p. 1896-1906.

139. Isakov, N., et al., ZAP-70 binding specificity to T cell receptor tyrosine-based

activation motifs: the tandem SH2 domains of ZAP-70 bind distinct tyrosine-

based activation motifs with varying affinity. J Exp Med, 1995. 181(1): p. 375-80.

297

140. Lo, W.L., et al., Lck promotes Zap70-dependent LAT phosphorylation by bridging

Zap70 to LAT. Nat Immunol, 2018. 19(7): p. 733-741.

141. White, J., et al., Two better cell lines for making hybridomas expressing specific T

cell receptors. J Immunol, 1989. 143(6): p. 1822-5.

142. Lesourne, R., et al., Themis, a T cell-specific protein important for late thymocyte

development. Nat Immunol, 2009. 10(8): p. 840-7.

143. Tarakhovsky, A., W. Muller, and K. Rajewsky, Lymphocyte populations and

immune responses in CD5-deficient mice. Eur J Immunol, 1994. 24(7): p. 1678-

84.

144. Azzam, H.S., et al., Fine tuning of TCR signaling by CD5. J.Immunol., 2001.

166(9): p. 5464-5472.

145. Oehen, S., et al., Antigen compartmentation and T helper cell tolerance induction.

J Exp Med, 1996. 183(6): p. 2617-26.

146. Tanchot, C., et al., Adaptive tolerance of CD4+ T cells in vivo: multiple

thresholds in response to a constant level of antigen presentation. J Immunol,

2001. 167(4): p. 2030-9.

147. Singh, N.J. and R.H. Schwartz, The strength of persistent antigenic stimulation

modulates adaptive tolerance in peripheral CD4+ T cells. J Exp Med, 2003.

198(7): p. 1107-17.

148. Dobin, A., et al., STAR: ultrafast universal RNA-seq aligner. Bioinformatics,

2013. 29(1): p. 15-21.

298

149. Liao, Y., G.K. Smyth, and W. Shi, featureCounts: an efficient general purpose

program for assigning sequence reads to genomic features. Bioinformatics, 2014.

30(7): p. 923-30.

150. Love, M.I., W. Huber, and S. Anders, Moderated estimation of fold change and

dispersion for RNA-seq data with DESeq2. Genome Biol, 2014. 15(12): p. 550.

151. Kisielow, P. and H. von Boehmer, Development and selection of T cells: facts and

puzzles. Adv Immunol, 1995. 58: p. 87-209.

152. Zúñiga-Pflücker, J.C. and M.J. Lenardo, Regulation of thymocyte development

from immature progenitors. Curr Opin Immunol, 1996. 8(2): p. 215-24.

153. Marrack, P. and J. Kappler, Positive selection of thymocytes bearing alpha beta T

cell receptors. Curr Opin Immunol, 1997. 9(2): p. 250-5.

154. Killeen, N., et al., Signaling checkpoints during the development of T

lymphocytes. Curr Opin Immunol, 1998. 10(3): p. 360-7.

155. Gascoigne, N.R., et al., TCR Signal Strength and T Cell Development. Annu Rev

Cell Dev Biol, 2016. 32: p. 327-348.

156. Rothenberg, E.V. and T. Taghon, Molecular genetics of T cell development. Annu

Rev Immunol, 2005. 23: p. 601-49.

157. Ellmeier, W., S. Sawada, and D.R. Littman, The regulation of CD4 and CD8

coreceptor gene expression during T cell development. Annu Rev Immunol, 1999.

17: p. 523-54.

158. Graham, V.A., A.L. Marzo, and D.F. Tough, A role for CD44 in T cell

development and function during direct competition between CD44+ and CD44-

cells. Eur J Immunol, 2007. 37(4): p. 925-34.

299

159. Protin, U., et al., CD44-deficient mice develop normally with changes in

subpopulations and recirculation of lymphocyte subsets. J Immunol, 1999.

163(9): p. 4917-23.

160. Willerford, D.M., et al., Interleukin-2 receptor alpha chain regulates the size and

content of the peripheral lymphoid compartment. Immunity, 1995. 3(4): p. 521-

30.

161. Hu, Q.N. and T.A. Baldwin, Programmed cell death in T cell development.

Apoptosis, Justine Rudner, IntechOpen, 2013.

162. Germain, R.N., T-cell development and the CD4-CD8 lineage decision. Nat Rev

Immunol, 2002. 2(5): p. 309-22.

163. Bendelac, A., et al., Activation events during thymic selection. J Exp Med, 1992.

175(3): p. 731-42.

164. Takada, K. and S.C. Jameson, Self-class I MHC molecules support survival of

naive CD8 T cells, but depress their functional sensitivity through regulation of

CD8 expression levels. J Exp Med, 2009. 206(10): p. 2253-69.

165. Carding, S.R., A.C. Hayday, and K. Bottomly, Cytokines in T-cell development.

Immunol Today, 1991. 12(7): p. 239-45.

166. Hosokawa, H. and E.V. Rothenberg, Cytokines, Transcription Factors, and the

Initiation of T-Cell Development. Cold Spring Harb Perspect Biol, 2018. 10(5).

167. Lancaster, J.N., Y. Li, and L.I.R. Ehrlich, Chemokine-Mediated Choreography of

Thymocyte Development and Selection. Trends Immunol, 2018. 39(2): p. 86-98.

168. Bertoni, A., et al., Integrins in T Cell Physiology. Int J Mol Sci, 2018. 19(2).

300

169. Crisa, L., et al., Cell adhesion and migration are regulated at distinct stages of

thymic T cell development: the roles of fibronectin, VLA4, and VLA5. J Exp Med,

1996. 184(1): p. 215-28.

170. Tsaouli, G., et al., Molecular Mechanisms of Notch Signaling in Lymphoid Cell

Lineages Development: NF-κB and Beyond. Adv Exp Med Biol, 2020. 1227: p.

145-164.

171. Attaf, M., E. Huseby, and A.K. Sewell, αβ T cell receptors as predictors of health

and disease. Cell Mol Immunol, 2015. 12(4): p. 391-9.

172. Singh, N.J. and R.H. Schwartz, Primer: mechanisms of immunologic tolerance.

Nature Clinical Practice Rheumatology, 2006. 2(1): p. 44-52.

173. Li, Q.J., et al., miR-181a is an intrinsic modulator of T cell sensitivity and

selection. Cell, 2007. 129(1): p. 147-61.

174. Ebert, P.J., et al., An endogenous positively selecting peptide enhances mature T

cell responses and becomes an autoantigen in the absence of microRNA miR-

181a. Nat Immunol, 2009. 10(11): p. 1162-9.

175. Fu, G., et al., Themis sets the signal threshold for positive and negative selection

in T-cell development. Nature, 2013. 504(7480): p. 441-5.

176. Mehta, M., et al., Themis-associated phosphatase activity controls signaling in T

cell development. Proc Natl Acad Sci U S A, 2018. 115(48): p. E11331-e11340.

177. Choi, S., et al., THEMIS enhances TCR signaling and enables positive selection

by selective inhibition of the phosphatase SHP-1. Nat Immunol, 2017. 18(4): p.

433-441.

301

178. Brzostek, J., et al., T cell receptor and cytokine signal integration in CD8(+) T

cells is mediated by the protein Themis. Nat Immunol, 2020.

179. Wang, D., et al., Tespa1 is involved in late thymocyte development through the

regulation of TCR-mediated signaling. Nat Immunol, 2012. 13(6): p. 560-8.

180. Liang, J., et al., Tespa1 regulates T cell receptor-induced calcium signals by

recruiting inositol 1,4,5-trisphosphate receptors. Nat Commun, 2017. 8: p. 15732.

181. Lyu, J., et al., Thymic-specific regulation of TCR signaling by Tespa1. Cell Mol

Immunol, 2019. 16(12): p. 897-907.

182. Raman, C., CD5, an important regulator of lymphocyte selection and immune

tolerance. Immunol.Res., 2002. 26(1-3): p. 255-263.

183. Jones, N.H., et al., Isolation of complementary DNA clones encoding the human

lymphocyte glycoprotein T1/Leu-1. Nature, 1986. 323(6086): p. 346-9.

184. Weichert, T.R. and R.C. Schwartz, Cloning of the murine CD5 promoter and its

tissue-specific regulation. J Immunol, 1995. 154(9): p. 4603-12.

185. Tung, J.W., et al., The regulation of CD5 expression in murine T cells. BMC Mol

Biol, 2001. 2: p. 5.

186. Carnero-Montoro, E., et al., Evolutionary and functional evidence for positive

selection at the human CD5 immune receptor gene. Mol Biol Evol, 2012. 29(2):

p. 811-23.

187. Tarakhovsky, A., et al., A role for CD5 in TCR-mediated signal transduction and

thymocyte selection. Science, 1995. 269(5223): p. 535-7.

188. Voisinne, G., A. Gonzalez de Peredo, and R. Roncagalli, CD5, an Undercover

Regulator of TCR Signaling. Front Immunol, 2018. 9: p. 2900.

302

189. Brown, M.H. and E. Lacey, A ligand for CD5 is CD5. J Immunol, 2010. 185(10):

p. 6068-74.

190. Zhang, C., et al., CD5 Binds to Interleukin-6 and Induces a Feed-Forward Loop

with the Transcription Factor STAT3 in B Cells to Promote Cancer. Immunity,

2016. 44(4): p. 913-923.

191. Bhandoola, A., et al., CD5-mediated inhibition of TCR signaling during

intrathymic selection and development does not require the CD5 extracellular

domain. Eur J Immunol, 2002. 32(6): p. 1811-7.

192. Raman, C. and R.P. Kimberly, Differential CD5-dependent regulation of CD5-

associated CK2 activity in mature and immature T cells: implication on

TCR/CD3-mediated activation. J Immunol, 1998. 161(11): p. 5817-20.

193. Raman, C., et al., Regulation of casein kinase 2 by direct interaction with cell

surface receptor CD5. J Biol Chem, 1998. 273(30): p. 19183-9.

194. Perez-Villar, J.J., et al., CD5 negatively regulates the T-cell antigen receptor

signal transduction pathway: involvement of SH2-containing phosphotyrosine

phosphatase SHP-1. Mol Cell Biol, 1999. 19(4): p. 2903-12.

195. Gary-Gouy, H., et al., The pseudo-immunoreceptor tyrosine-based activation

motif of CD5 mediates its inhibitory action on B-cell receptor signaling. J Biol

Chem, 2000. 275(1): p. 548-56.

196. Ochi, H. and T. Watanabe, Negative regulation of B cell receptor-mediated

signaling in B-1 cells through CD5 and Ly49 co-receptors via Lyn kinase activity.

Int Immunol, 2000. 12(10): p. 1417-23.

303

197. Thien, C.B., et al., Loss of c-Cbl RING finger function results in high-intensity

TCR signaling and thymic deletion. Embo j, 2005. 24(21): p. 3807-19.

198. Demydenko, D., c-Cbl mediated ubiquitylation and regulation of cell surface

exposure of CD5. Biochem Biophys Res Commun, 2010. 392(4): p. 500-4.

199. Bamberger, M., et al., A new pathway of CD5 glycoprotein-mediated T cell

inhibition dependent on inhibitory phosphorylation of Fyn kinase. J Biol Chem,

2011. 286(35): p. 30324-36.

200. Sestero, C.M., et al., CD5-dependent CK2 activation pathway regulates threshold

for T cell anergy. J Immunol, 2012. 189(6): p. 2918-30.

201. Dong, B., et al., CD5-mediated inhibition of TCR signaling proceeds normally in

the absence of SHP-1. Int J Mol Med, 2016. 38(1): p. 45-56.

202. Voisinne, G., et al., Co-recruitment analysis of the CBL and CBLB signalosomes

in primary T cells identifies CD5 as a key regulator of TCR-induced

ubiquitylation. Mol Syst Biol, 2016. 12(7): p. 876.

203. Huang, F. and H. Gu, Negative regulation of lymphocyte development and

function by the Cbl family of proteins. Immunol Rev, 2008. 224: p. 229-38.

204. Wang, H.Y., et al., Cbl promotes ubiquitination of the T cell receptor zeta

through an adaptor function of Zap-70. J Biol Chem, 2001. 276(28): p. 26004-11.

205. Thien, C.B., D.D. Bowtell, and W.Y. Langdon, Perturbed regulation of ZAP-70

and sustained tyrosine phosphorylation of LAT and SLP-76 in c-Cbl-deficient

thymocytes. J Immunol, 1999. 162(12): p. 7133-9.

304

206. Thien, C.B. and W.Y. Langdon, c-Cbl and Cbl-b ubiquitin ligases: substrate

diversity and the negative regulation of signalling responses. Biochem J, 2005.

391(Pt 2): p. 153-66.

207. Chiang, Y.J., et al., Cbl-b regulates the CD28 dependence of T-cell activation.

Nature, 2000. 403(6766): p. 216-20.

208. Okada, M., et al., CSK: a protein-tyrosine kinase involved in regulation of src

family kinases. J Biol Chem, 1991. 266(36): p. 24249-52.

209. Tan, Y.X., et al., Inhibition of the kinase Csk in thymocytes reveals a requirement

for actin remodeling in the initiation of full TCR signaling. Nat Immunol, 2014.

15(2): p. 186-94.

210. Calvo, J., et al., Human CD5 signaling and constitutive phosphorylation of C-

terminal serine residues by casein kinase II. J Immunol, 1998. 161(11): p. 6022-9.

211. Mier-Aguilar, C.A., et al., CD5-CK2 Signaling Modulates Erk Activation and

Thymocyte Survival. PLoS One, 2016. 11(12): p. e0168155.

212. Blaize, G., et al., CD5 signalosome coordinates antagonist TCR signals to control

the generation of Treg cells induced by foreign antigens. Proc Natl Acad Sci U S

A, 2020. 117(23): p. 12969-12979.

213. Voisinne, G., et al., Quantitative interactomics in primary T cells unveils TCR

signal diversification extent and dynamics. Nat Immunol, 2019. 20(11): p. 1530-

1541.

214. Azzam, H.S., et al., CD5 expression is developmentally regulated by T cell

receptor (TCR) signals and TCR avidity. J Exp Med, 1998. 188(12): p. 2301-11.

305

215. Yamashita, I., et al., CD69 cell surface expression identifies developing

thymocytes which audition for T cell antigen receptor-mediated positive selection.

Int Immunol, 1993. 5(9): p. 1139-50.

216. Ashouri, J.F. and A. Weiss, Endogenous Nur77 Is a Specific Indicator of Antigen

Receptor Signaling in Human T and B Cells. J Immunol, 2016.

217. Pena-Rossi, C., et al., Negative regulation of CD4 lineage development and

responses by CD5. J Immunol, 1999. 163(12): p. 6494-501.

218. Sun, S.C., The non-canonical NF-κB pathway in immunity and inflammation. Nat

Rev Immunol, 2017. 17(9): p. 545-558.

219. Webb, L.V., et al., Survival of Single Positive Thymocytes Depends upon

Developmental Control of RIPK1 Kinase Signaling by the IKK Complex

Independent of NF-kappaB. Immunity, 2019. 50(2): p. 348-361.e4.

220. Ribeiro, S.T., et al., Casein kinase 2 controls the survival of normal thymic and

leukemic gammadelta T cells via promotion of AKT signaling. Leukemia, 2017.

31(7): p. 1603-1610.

221. Jaksch, C. and P. Thams, A critical role for CK2 in cytokine-induced activation of

NFkappaB in pancreatic beta cell death. Endocrine, 2014. 47(1): p. 117-28.

222. Gibson, S.A., et al., Protein Kinase CK2 Controls the Fate between Th17 Cell

and Regulatory T Cell Differentiation. J Immunol, 2017. 198(11): p. 4244-4254.

223. Jang, S.W., et al., Casein kinase 2 is a critical determinant of the balance of Th17

and Treg cell differentiation. Exp Mol Med, 2017. 49(9): p. e375.

306

224. Pierre, F., et al., Pre-clinical characterization of CX-4945, a potent and selective

small molecule inhibitor of CK2 for the treatment of cancer. Mol Cell Biochem,

2011. 356(1-2): p. 37-43.

225. Siddiqui-Jain, A., et al., CX-4945, an orally bioavailable selective inhibitor of

protein kinase CK2, inhibits prosurvival and angiogenic signaling and exhibits

antitumor efficacy. Cancer Res, 2010. 70(24): p. 10288-98.

226. Gibson, S.A. and E.N. Benveniste, Protein Kinase CK2: An Emerging Regulator

of Immunity. Trends Immunol, 2018. 39(2): p. 82-85.

227. Lin, R., et al., Phosphorylation of IkappaBalpha in the C-terminal PEST domain

by casein kinase II affects intrinsic protein stability. Mol Cell Biol, 1996. 16(4):

p. 1401-9.

228. Gerondakis, S., et al., NF-κB control of T cell development. Nat Immunol, 2014.

15(1): p. 15-25.

229. Jimi, E., et al., Differential role of the transcription factor NF-kappaB in selection

and survival of CD4+ and CD8+ thymocytes. Immunity, 2008. 29(4): p. 523-37.

230. Schmidt-Supprian, M., et al., Mature T cells depend on signaling through the IKK

complex. Immunity, 2003. 19(3): p. 377-89.

231. Webb, L.V., S.C. Ley, and B. Seddon, TNF activation of NF-κB is essential for

development of single-positive thymocytes. J Exp Med, 2016. 213(8): p. 1399-407.

232. Pierce, J.W., et al., Novel inhibitors of cytokine-induced IkappaBalpha

phosphorylation and endothelial cell adhesion molecule expression show anti-

inflammatory effects in vivo. J Biol Chem, 1997. 272(34): p. 21096-103.

307

233. Krishnan, N., et al., The anti-inflammatory compound BAY-11-7082 is a potent

inhibitor of protein tyrosine phosphatases. Febs j, 2013. 280(12): p. 2830-41.

234. Takada, K. and S.C. Jameson, Naive T cell homeostasis: from awareness of space

to a sense of place. Nat Rev Immunol, 2009. 9(12): p. 823-32.

235. Dorfman, J.R., et al., CD4+ T cell survival is not directly linked to self-MHC-

induced TCR signaling. Nat Immunol, 2000. 1(4): p. 329-35.

236. Bhandoola, A., et al., Peripheral expression of self-MHC-II influences the

reactivity and self-tolerance of mature CD4(+) T cells: evidence from a

lymphopenic T cell model. Immunity, 2002. 17(4): p. 425-36.

237. Hochweller, K., et al., Dendritic cells control T cell tonic signaling required for

responsiveness to foreign antigen. Proc Natl Acad Sci U S A, 2010. 107(13): p.

5931-6.

238. De Riva, A., et al., Noncognate interaction with MHC class II molecules is

essential for maintenance of T cell metabolism to establish optimal memory CD4

T cell function. J Immunol, 2007. 178(9): p. 5488-95.

239. Khanom, U.S., et al., TCR Affinity for In Vivo Peptide-Induced Thymic Positive

Selection Fine-Tunes TCR Responsiveness of Peripheral CD8(+) T Cells. J

Immunol, 2019. 203(4): p. 881-887.

240. Fischer, U.B., et al., MHC class II deprivation impairs CD4 T cell motility and

responsiveness to antigen-bearing dendritic cells in vivo. Proc Natl Acad Sci U S

A, 2007. 104(17): p. 7181-6.

241. Stefanova, I., et al., On the role of self-recognition in T cell responses to foreign

antigen. Immunol Rev, 2003. 191: p. 97-106.

308

242. Hogquist, K.A. and S.C. Jameson, The self-obsession of T cells: how TCR

signaling thresholds affect fate 'decisions' and effector function. Nat Immunol,

2014. 15(9): p. 815-23.

243. Garbi, N., et al., Tonic T cell signalling and T cell tolerance as opposite effects of

self-recognition on dendritic cells. Curr Opin Immunol, 2010. 22(5): p. 601-8.

244. Stritesky, G.L., S.C. Jameson, and K.A. Hogquist, Selection of self-reactive T

cells in the thymus. Annu Rev Immunol, 2012. 30: p. 95-114.

245. Ashouri, J.F. and A. Weiss, Endogenous Nur77 Is a Specific Indicator of Antigen

Receptor Signaling in Human T and B Cells. J Immunol, 2017. 198(2): p. 657-

668.

246. Zinzow-Kramer, W.M., A. Weiss, and B.B. Au-Yeung, Adaptation by naive

CD4(+) T cells to self-antigen-dependent TCR signaling induces functional

heterogeneity and tolerance. Proc Natl Acad Sci U S A, 2019. 116(30): p. 15160-

15169.

247. Moran, A.E., et al., T cell receptor signal strength in Treg and iNKT cell

development demonstrated by a novel fluorescent reporter mouse. J Exp Med,

2011. 208(6): p. 1279-89.

248. Fulton, R.B., et al., The TCR's sensitivity to self peptide-MHC dictates the ability

of naive CD8(+) T cells to respond to foreign antigens. Nat Immunol, 2015.

16(1): p. 107-17.

249. Mandl, J.N., et al., T cell-positive selection uses self-ligand binding strength to

optimize repertoire recognition of foreign antigens. Immunity, 2013. 38(2): p.

263-274.

309

250. Persaud, S.P., et al., Intrinsic CD4+ T cell sensitivity and response to a pathogen

are set and sustained by avidity for thymic and peripheral complexes of self

peptide and MHC. Nat Immunol, 2014. 15(3): p. 266-74.

251. Weber, K.S., et al., Distinct CD4+ helper T cells involved in primary and

secondary responses to infection. Proc Natl Acad Sci U S A, 2012. 109(24): p.

9511-6.

252. Burgueno-Bucio, E., C.A. Mier-Aguilar, and G. Soldevila, The multiple faces of

CD5. J Leukoc Biol, 2019. 105(5): p. 891-904.

253. Friedlein, G., et al., Human CD5 protects circulating tumor antigen-specific CTL

from tumor-mediated activation-induced cell death. J Immunol, 2007. 178(11): p.

6821-7.

254. Ryan, K.R., D. McCue, and S.M. Anderton, Fas-mediated death and sensory

adaptation limit the pathogenic potential of autoreactive T cells after strong

antigenic stimulation. J Leukoc Biol, 2005. 78(1): p. 43-50.

255. Axtell, R.C., et al., Cutting edge: critical role for CD5 in experimental

autoimmune encephalomyelitis: inhibition of engagement reverses disease in

mice. J Immunol, 2004. 173(5): p. 2928-32.

256. Graw, F., et al., Dynamics of CD4(+) T cell responses against Listeria

monocytogenes. J Immunol, 2012. 189(11): p. 5250-6.

257. Milam, A.V. and P.M. Allen, Functional Heterogeneity in CD4(+) T Cell

Responses Against a Bacterial Pathogen. Front Immunol, 2015. 6: p. 621.

258. Freitas, C.M.T., et al., Naive helper T cells with high CD5 expression have

increased calcium signaling. PLoS One, 2017. 12(5): p. e0178799.

310

259. Milam, A.A.V., et al., Tuning T Cell Signaling Sensitivity Alters the Behavior of

CD4(+) T Cells during an Immune Response. J Immunol, 2018. 200(10): p. 3429-

3437.

260. Perez-Villar, J.J., et al., CD5 negatively regulates the T-cell antigen receptor

signal transduction pathway: Involvement of SH2-containing phosphotyrosine

phosphatase SHP-1. Molecular and Cellular Biology, 1999. 19(4): p. 2903-2912.

261. Chan, S., et al., An influence of CD5 on the selection of CD4-lineage T cells.

Eur.J.Immunol., 1999. 29(9): p. 2916-2922.

262. Henderson, J.G., et al., CD5 instructs extrathymic regulatory T cell development

in response to self and tolerizing antigens. Immunity, 2015. 42(3): p. 471-83.

263. McGuire, D.J., et al., CD5 enhances Th17-cell differentiation by regulating IFN-

gamma response and RORgammat localization. Eur J Immunol, 2014. 44(4): p.

1137-42.

264. McGuire, D.J., et al., CD5 enhances Th17-cell differentiation by regulating IFN-γ

response and RORγt localization. Eur J Immunol, 2014. 44(4): p. 1137-42.

265. Azzam, H.S., et al., CD5 expression is developmentally regulated by T cell

receptor (TCR) signals and TCR avidity. J.Exp.Med., 1998. 188(12): p. 2301-

2311.

266. Stamou, P., et al., Chronic exposure to low levels of antigen in the periphery

causes reversible functional impairment correlating with changes in CD5 levels

in monoclonal CD8 T cells. J Immunol, 2003. 171(3): p. 1278-1284.

267. Marquez, M.E., et al., CD8 T cell sensory adaptation dependent on TCR avidity

for self-antigens. J.Immunol., 2005. 175(11): p. 7388-7397.

311

268. Fulton, R.B., et al., The TCR's sensitivity to self peptide-MHC dictates the ability

of naive CD8 T cells to respond to foreign antigens. Nat Immunol, 2014.

269. Delpoux, A., et al., TCR signaling events are required for maintaining CD4

regulatory T cell numbers and suppressive capacities in the periphery. J

Immunol, 2014. 193(12): p. 5914-23.

270. Sprouse, M.L., et al., High self-reactivity drives T-bet and potentiates Treg

function in tissue-specific autoimmunity. JCI Insight, 2018. 3(2).

271. Houston, E.G., Jr., L.E. Higdon, and P.J. Fink, Recent thymic emigrants are

preferentially incorporated only into the depleted T-cell pool. Proc Natl Acad Sci

U S A, 2011. 108(13): p. 5366-71.

272. Chu, Z.L., et al., Basal phosphorylation of the PEST domain in the

I(kappa)B(beta) regulates its functional interaction with the c-rel proto-oncogene

product. Mol Cell Biol, 1996. 16(11): p. 5974-84.

273. Scott, M.L., et al., The p65 subunit of NF-kappa B regulates I kappa B by two

distinct mechanisms. Genes Dev, 1993. 7(7a): p. 1266-76.

274. Sun, S.C., et al., NF-kappa B controls expression of inhibitor I kappa B alpha:

evidence for an inducible autoregulatory pathway. Science, 1993. 259(5103): p.

1912-5.

275. Sood, A., et al., Differential interferon-gamma production potential among naive

CD4(+) T cells exists prior to antigen encounter. Immunol Cell Biol, 2019.

97(10): p. 931-940.

312

276. Knudson, K.M., et al., NFkappaB-Pim-1-Eomesodermin axis is critical for

maintaining CD8 T-cell memory quality. Proc Natl Acad Sci U S A, 2017. 114(9):

p. E1659-e1667.

277. Martin, F. and J.F. Kearney, B1 cells: similarities and differences with other B

cell subsets. Curr Opin Immunol, 2001. 13(2): p. 195-201.

278. Oh, H. and S. Ghosh, NF-kappaB: roles and regulation in different CD4(+) T-cell

subsets. Immunol Rev, 2013. 252(1): p. 41-51.

279. Billingham, R.E., L. Brent, and P.B. Medawar, Acquired tolerance of skin

homografts. Ann N Y Acad Sci, 1955. 59(3): p. 409-16.

280. Billingham, R.E., L. Brent, and P.B. Medawar, Actively acquired tolerance of

foreign cells. Nature, 1953. 172(4379): p. 603-6.

281. Klein, L., E.A. Robey, and C.S. Hsieh, Central CD4(+) T cell tolerance: deletion

versus regulatory T cell differentiation. Nat Rev Immunol, 2019. 19(1): p. 7-18.

282. Germain, R.N., Special regulatory T-cell review: A rose by any other name: from

suppressor T cells to Tregs, approbation to unbridled enthusiasm. Immunology,

2008. 123(1): p. 20-27.

283. Sakaguchi, S., et al., Immunologic self-tolerance maintained by activated T cells

expressing IL-2 receptor alpha-chains (CD25). Breakdown of a single mechanism

of self-tolerance causes various autoimmune diseases. J.Immunol., 1995. 155(3):

p. 1151-1164.

284. Hori, S., T. Nomura, and S. Sakaguchi, Control of regulatory T cell development

by the transcription factor Foxp3. Science, 2003. 299(5609): p. 1057-61.

313

285. Khattri, R., et al., An essential role for Scurfin in CD4(+)CD25(+) T regulatory

cells. Nature Immunology, 2003. 4(4): p. 337-342.

286. Hsieh, C.S., et al., Recognition of the peripheral self by naturally arising CD25+

CD4+ T cell receptors. Immunity., 2004. 21(2): p. 267-277.

287. Fontenot, J.D., M.A. Gavin, and A.Y. Rudensky, Foxp3 programs the

development and function of CD4+CD25+ regulatory T cells. Nat.Immunol.,

2003. 4(4): p. 330-336.

288. Kim, J.M., J.P. Rasmussen, and A.Y. Rudensky, Regulatory T cells prevent

catastrophic autoimmunity throughout the lifespan of mice. Nat.Immunol, 2007.

8(2): p. 191-197.

289. Lahl, K., et al., Selective depletion of Foxp3+ regulatory T cells induces a scurfy-

like disease. J Exp.Med., 2007. 204(1): p. 57-63.

290. Stranges, P.B., et al., Elimination of antigen-presenting cells and autoreactive T

cells by Fas contributes to prevention of autoimmunity. Immunity, 2007. 26(5): p.

629-641.

291. McHugh, R.S. and E.M. Shevach, Cutting edge: depletion of CD4+CD25+

regulatory T cells is necessary, but not sufficient, for induction of organ-specific

autoimmune disease. J Immunol, 2002. 168(12): p. 5979-5983.

292. Schwartz, R.H., T Cell Anergy. Annu.Rev.Immunol, 2003. 21: p. 305-334.

293. Ise, W., et al., CTLA-4 suppresses the pathogenicity of self antigen-specific T cells

by cell-intrinsic and cell-extrinsic mechanisms. Nat Immunol, 2010. 11(2): p.

129-35.

314

294. Paterson, A.M., et al., Deletion of CTLA-4 on regulatory T cells during adulthood

leads to resistance to autoimmunity. J Exp Med, 2015. 212(10): p. 1603-21.

295. Paterson, A.M. and A.H. Sharpe, Taming tissue-specific T cells: CTLA-4 reins in

self-reactive T cells. Nat Immunol, 2010. 11(2): p. 109-11.

296. Nossal, G.J. and B.L. Pike, Clonal anergy: persistence in tolerant mice of

antigen-binding B lymphocytes incapable of responding to antigen or mitogen.

Proc.Natl.Acad.Sci.U.S.A, 1980. 77(3): p. 1602-1606.

297. Schwartz, R.H., T cell anergy. Annu Rev Immunol, 2003. 21: p. 305-34.

298. Jenkins, M.K. and R.H. Schwartz, Pillars article: antigen presentation by

chemically modified splenocytes induces antigen-specific T cell unresponsiveness

in vitro and in vivo. J. Exp. Med. 1987. 165: 302-319. J Immunol, 2009. 183(7):

p. 4150-67.

299. Sloan-Lancaster, J., B.D. Evavold, and P.M. Allen, Induction of T-cell anergy by

altered T-cell-receptor ligand on live antigen-presenting cells. Nature, 1993.

363(6425): p. 156-159.

300. Korb, L.C., et al., Induction of T cell anergy by low numbers of agonist ligands. J

Immunol, 1999. 162(11): p. 6401-9.

301. Sloan-Lancaster, J. and P.M. Allen, Altered peptide ligand-induced partial T cell

activation: molecular mechanisms and role in T cell biology.

Annu.Rev.Immunol., 1996. 14: p. 1-27.

302. Kersh, E.N., G.J. Kersh, and P.M. Allen, Partially phosphorylated T cell receptor

zeta molecules can inhibit T cell activation. J Exp.Med., 1999. 190(11): p. 1627-

1636.

315

303. Kambayashi, T. and T.M. Laufer, Atypical MHC class II-expressing antigen-

presenting cells: can anything replace a dendritic cell? Nat Rev Immunol, 2014.

14(11): p. 719-30.

304. Beverly, B., et al., Reversal of in vitro T cell clonal anergy by IL-2 stimulation.

Int.Immunol., 1992. 4(6): p. 661-671.

305. Jenkins, M.K., et al., Molecular events in the induction of a nonresponsive state in

interleukin 2-producing helper T-lymphocyte clones. Proc.Natl.Acad.Sci.U.S.A,

1987. 84(15): p. 5409-5413.

306. Mueller, D.L., M.K. Jenkins, and R.H. Schwartz, Clonal expansion versus

functional clonal inactivation: a costimulatory signalling pathway determines the

outcome of T cell antigen receptor occupancy. Annu.Rev.Immunol., 1989. 7: p.

445-480.

307. Essery, G., M. Feldmann, and J.R. Lamb, Interleukin-2 can prevent and reverse

antigen-induced unresponsiveness in cloned human T lymphocytes. Immunology,

1988. 64(3): p. 413-417.

308. Fields, P.E., T.F. Gajewski, and F.W. Fitch, Blocked Ras activation in anergic

CD4+ T cells. Science, 1996. 271(5253): p. 1276-1278.

309. Li, W., et al., Blocked signal transduction to the ERK and JNK protein kinases in

anergic CD4+ T cells. Science, 1996. 271(5253): p. 1272-1276.

310. Mondino, A., et al., Defective Transcription of the IL-2 Gene Is Associated with

Impaired Expression of c-Fos, FosB, and JunB in Anergic T Helper 1 Cells.

Journal of Immunology, 1996. 157(5): p. 2048-2057.

316

311. Desilva, D.R., et al., The p38 mitogen-activated protein kinase pathway in

activated and anergic Th1 cells. Cellular Immunology, 1997. 180(2): p. 116-123.

312. Kang, S.M., et al., Transactivation by AP-1 is a molecular target of T cell clonal

anergy. Science, 1992. 257(5073): p. 1134-1138.

313. Rothenberg, E.V. and S.B. Ward, A dynamic assembly of diverse transcription

factors integrates activation and cell-type information for interleukin 2 gene

regulation. Proc Natl Acad Sci U S A, 1996. 93(18): p. 9358-65.

314. Rammensee, H.G., R. Kroschewski, and B. Frangoulis, Clonal anergy induced in

mature Vβ6+ T lymphocytes on immunizing Mls-1b mice with Mls-1a expressing

cells. Nature, 1989. 339(6225): p. 541-544.

315. Kawabe, Y. and A. Ochi, Selective anergy of vβ8+, CD4 + T cells in

staphylococcus enterotoxin B-primed mice. Journal of Experimental Medicine,

1990. 172(4): p. 1065-1070.

316. Rellahan, B.L., et al., In vivo induction of anergy in peripheral vβ8+ T cells by

staphylococcal enterotoxin B. Journal of Experimental Medicine, 1990. 172(4): p.

1091-1100.

317. Ramsdell, F. and B.J. Fowlkes, Maintenance of in vivo tolerance by persistence of

antigen. Science, 1992. 257(5073): p. 1130-1134.

318. Rocha, B., C. Tanchot, and H. Von Boehmer, Clonal anergy blocks in vivo

growth of mature T cells and can be reversed in the absence of antigen. Journal of

Experimental Medicine, 1993. 177(5): p. 1517-1521.

319. Ferber, I., et al., Levels of peripheral T cell tolerance induced by different doses of

tolerogen. Science, 1994. 263(5147): p. 674-676.

317

320. Kearney, E.R., et al., Visualization of peptide-specific T cell immunity and

peripheral tolerance induction in vivo. Immunity, 1994. 1(4): p. 327-339.

321. Rocha, B., C. Tanchot, and H. Von Boehmer, Clonal anergy blocks in vivo

growth of mature T cells and can be reversed in the absence of antigen. J

Exp.Med., 1993. 177(5): p. 1517-1521.

322. Rocha, B., A. Grandien, and A.A. Freitas, Anergy and exhaustion are independent

mechanisms of peripheral T cell tolerance. J Exp.Med., 1995. 181(3): p. 993-

1003.

323. Tanchot, C., et al., Adaptive tolerance of CD4+ T cells in vivo: multiple

thresholds in response to a constant level of antigen presentation. J.Immunol.,

2001. 167(4): p. 2030-2039.

324. Moskophidis, D., et al., Virus persistence in acutely infected immunocompetent

mice by exhaustion of antiviral cytotoxic effector T cells. Nature, 1993.

362(6422): p. 758-61.

325. Zajac, A.J., et al., Viral immune evasion due to persistence of activated T cells

without effector function. Journal of Experimental Medicine, 1998. 188(12): p.

2205-2213.

326. Gallimore, A., et al., Induction and exhaustion of lymphocytic choriomeningitis

virus-specific cytotoxic T lymphocytes visualized using soluble tetrameric major

histocompatibility complex class I-peptide complexes. J Exp.Med., 1998. 187(9):

p. 1383-1393.

318

327. Wherry, E.J., et al., Viral persistence alters CD8 T-cell immunodominance and

tissue distribution and results in distinct stages of functional impairment. J Virol.,

2003. 77(8): p. 4911-4927.

328. Otten, G.R. and R.N. Germain, Split anergy in a CD8+ T cell: receptor-

dependent cytolysis in the absence of interleukin-2 production. Science, 1991.

251(4998): p. 1228-1231.

329. Wherry, E.J., T cell exhaustion. Nat Immunol, 2011. 12(6): p. 492-9.

330. Wherry, E.J. and M. Kurachi, Molecular and cellular insights into T cell

exhaustion. Nat Rev Immunol, 2015. 15(8): p. 486-99.

331. Vella, L.A., R.S. Herati, and E.J. Wherry, CD4(+) T Cell Differentiation in

Chronic Viral Infections: The Tfh Perspective. Trends Mol Med, 2017. 23(12): p.

1072-1087.

332. McLane, L.M., M.S. Abdel-Hakeem, and E.J. Wherry, CD8 T Cell Exhaustion

During Chronic Viral Infection and Cancer. Annu Rev Immunol, 2019.

333. Blackburn, S.D., et al., Coregulation of CD8+ T cell exhaustion by multiple

inhibitory receptors during chronic viral infection. Nat Immunol, 2009. 10(1): p.

29-37.

334. Matloubian, M., R.J. Concepcion, and R. Ahmed, CD4+ T cells are required to

sustain CD8+ cytotoxic T-cell responses during chronic viral infection. J Virol,

1994. 68(12): p. 8056-63.

335. Wherry, E.J., et al., Antigen-independent memory CD8 T cells do not develop

during chronic viral infection. Proc Natl Acad Sci U S A, 2004. 101(45): p.

16004-9.

319

336. Schietinger, A. and P.D. Greenberg, Tolerance and exhaustion: defining

mechanisms of T cell dysfunction. Trends Immunol, 2014. 35(2): p. 51-60.

337. Blank, C.U., et al., Defining 'T cell exhaustion'. Nat Rev Immunol, 2019.

338. Rossin, A., G. Miloro, and A.O. Hueber, TRAIL and FasL Functions in Cancer

and Autoimmune Diseases: Towards an Increasing Complexity. Cancers (Basel),

2019. 11(5).

339. Forrester, J.V., et al., Immune privilege or privileged immunity? Mucosal

Immunol, 2008. 1(5): p. 372-81.

340. Vallejo, A.N., C.M. Weyand, and J.J. Goronzy, T-cell senescence: a culprit of

immune abnormalities in chronic inflammation and persistent infection. Trends

Mol Med, 2004. 10(3): p. 119-24.

341. Tu, W. and S. Rao, Mechanisms Underlying T Cell Immunosenescence: Aging

and Cytomegalovirus Infection. Front Microbiol, 2016. 7: p. 2111.

342. Paley, M.A., et al., Progenitor and terminal subsets of CD8+ T cells cooperate to

contain chronic viral infection. Science, 2012. 338(6111): p. 1220-5.

343. Wherry, E.J., et al., Molecular signature of CD8+ T cell exhaustion during

chronic viral infection. Immunity, 2007. 27(4): p. 670-84.

344. Brooks, D.G., et al., Interleukin-10 determines viral clearance or persistence in

vivo. Nat Med, 2006. 12(11): p. 1301-9.

345. Ejrnaes, M., et al., Resolution of a chronic viral infection after interleukin-10

receptor blockade. J Exp Med, 2006. 203(11): p. 2461-72.

346. Elsaesser, H., K. Sauer, and D.G. Brooks, IL-21 is required to control chronic

viral infection. Science, 2009. 324(5934): p. 1569-72.

320

347. Frohlich, A., et al., IL-21R on T cells is critical for sustained functionality and

control of chronic viral infection. Science, 2009. 324(5934): p. 1576-80.

348. Fahey, L.M., et al., Viral persistence redirects CD4 T cell differentiation toward T

follicular helper cells. J Exp Med, 2011. 208(5): p. 987-99.

349. Crawford, A., et al., Molecular and transcriptional basis of CD4⁺ T cell

dysfunction during chronic infection. Immunity, 2014. 40(2): p. 289-302.

350. Im, S.J., et al., Defining CD8+ T cells that provide the proliferative burst after

PD-1 therapy. Nature, 2016. 537(7620): p. 417-421.

351. Leong, Y.A., et al., CXCR5(+) follicular cytotoxic T cells control viral infection

in B cell follicles. Nat Immunol, 2016. 17(10): p. 1187-96.

352. He, R., et al., Follicular CXCR5- expressing CD8(+) T cells curtail chronic viral

infection. Nature, 2016. 537(7620): p. 412-428.

353. Martinez, R.J., et al., Arthritogenic self-reactive CD4+ T cells acquire an

FR4hiCD73hi anergic state in the presence of Foxp3+ regulatory T cells. J

Immunol, 2012. 188(1): p. 170-81.

354. Kobie, J.J., et al., T regulatory and primed uncommitted CD4 T cells express

CD73, which suppresses effector CD4 T cells by converting 5'-adenosine

monophosphate to adenosine. J Immunol, 2006. 177(10): p. 6780-6.

355. Yamaguchi, T., et al., Control of immune responses by antigen-specific regulatory

T cells expressing the folate receptor. Immunity, 2007. 27(1): p. 145-59.

356. Sen, D.R., et al., The epigenetic landscape of T cell exhaustion. Science, 2016.

354(6316): p. 1165-1169.

321

357. Schietinger, A., et al., Tumor-Specific T Cell Dysfunction Is a Dynamic Antigen-

Driven Differentiation Program Initiated Early during Tumorigenesis. Immunity,

2016. 45(2): p. 389-401.

358. Rocha, B., C. Tanchot, and H. Von Boehmer, Clonal anergy blocks in vivo

growth of mature T cells and can be reversed in the absence of antigen. J Exp

Med, 1993. 177(5): p. 1517-21.

359. Stamou, P., et al., Chronic Exposure to Low Levels of Antigen in the Periphery

Causes Reversible Functional Impairment Correlating with Changes in CD5

Levels in Monoclonal CD8 T Cells. The Journal of Immunology, 2003. 171(3): p.

1278-1284.

360. Miller, B.C., et al., Subsets of exhausted CD8(+) T cells differentially mediate

tumor control and respond to checkpoint blockade. Nat Immunol, 2019. 20(3): p.

326-336.

361. Blackburn, S.D., et al., Selective expansion of a subset of exhausted CD8 T cells

by alphaPD-L1 blockade. Proc Natl Acad Sci U S A, 2008. 105(39): p. 15016-21.

362. Pauken, K.E., et al., Epigenetic stability of exhausted T cells limits durability of

reinvigoration by PD-1 blockade. Science, 2016. 354(6316): p. 1160-1165.

363. Zebley, C.C., S. Gottschalk, and B. Youngblood, Rewriting History: Epigenetic

Reprogramming of CD8(+) T Cell Differentiation to Enhance Immunotherapy.

Trends Immunol, 2020.

364. Marro, B.S., et al., Discovery of Small Molecules for the Reversal of T Cell

Exhaustion. Cell Rep, 2019. 29(10): p. 3293-3302.e3.

322

365. Kearney, E.R., et al., Visualization of peptide-specific T cell immunity and

peripheral tolerance induction in vivo. Immunity., 1994. 1(4): p. 327-339.

366. Akkaraju, S., et al., A range of CD4 T cell tolerance: partial inactivation to

organ-specific antigen allows nondestructive thyroiditis or insulitis. Immunity.,

1997. 7(2): p. 255-271.

367. Ausubel, L.J., et al., Functional tolerance is maintained despite proliferation of

CD4 T cells after encounter with tissue-derived antigen. Dev.Immunol., 2002.

9(3): p. 173-176.

368. Lanoue, A., et al., Conditions that induce tolerance in mature CD4+ T cells.

J.Exp.Med., 1997. 185(3): p. 405-414.

369. Adler, A.J., et al., CD4+ T cell tolerance to parenchymal self-antigens requires

presentation by bone marrow-derived antigen-presenting cells. J.Exp.Med., 1998.

187(10): p. 1555-1564.

370. Adler, A.J., et al., In vivo CD4+ T cell tolerance induction versus priming is

independent of the rate and number of cell divisions. J.Immunol., 2000. 164(2): p.

649-655.

371. Jordan, M.S., et al., Anergy and suppression regulate CD4(+) T cell responses to

a self peptide. Eur.J.Immunol., 2000. 30(1): p. 136-144.

372. Riley, M.P., et al., Graded deletion and virus-induced activation of autoreactive

CD4+ T cells. J.Immunol., 2000. 165(9): p. 4870-4876.

373. Riley, M.P., et al., CD4+ T cells that evade deletion by a self peptide display Th1-

biased differentiation. Eur.J.Immunol., 2001. 31(1): p. 311-319.

323

374. Knoechel, B., et al., Sequential development of interleukin 2-dependent effector

and regulatory T cells in response to endogenous systemic antigen. J.Exp.Med.,

2005. 202(10): p. 1375-1386.

375. Chiodetti, L., et al., Adaptive Tolerance and Clonal Anergy Are Distinct

Biochemical States. The Journal of Immunology, 2006. 176(4): p. 2279-2291.

376. Choi, S. and R.H. Schwartz, Molecular mechanisms for adaptive tolerance and

other T cell anergy models. Semin Immunol, 2007. 19(3): p. 140-52.

377. Choi, S. and R.H. Schwartz, Impairment of immunological synapse formation in

adaptively tolerant T cells. J Immunol, 2011. 187(2): p. 805-16.

378. Brownlie, R.J. and R. Zamoyska, T cell receptor signalling networks: branched,

diversified and bounded. Nat Rev Immunol, 2013. 13(4): p. 257-69.

379. Chappert, P., et al., Specific gut commensal flora locally alters T cell tuning to

endogenous ligands. Immunity, 2013. 38(6): p. 1198-210.

380. Singh, N.J., C. Chen, and R.H. Schwartz, The impact of T cell intrinsic antigen

adaptation on peripheral immune tolerance. PLoS Biol, 2006. 4(11): p. e340.

381. Coghlan, M.P., et al., Selective small molecule inhibitors of glycogen synthase

kinase-3 modulate glycogen metabolism and gene transcription. Chem Biol,

2000. 7(10): p. 793-803.

382. Blake, R.A., et al., SU6656, a selective src family kinase inhibitor, used to probe

growth factor signaling. Mol Cell Biol, 2000. 20(23): p. 9018-27.

383. Alfei, F., et al., TOX reinforces the phenotype and longevity of exhausted T cells

in chronic viral infection. Nature, 2019. 571(7764): p. 265-269.

324

384. Chen, J., et al., NR4A transcription factors limit CAR T cell function in solid

tumours. Nature, 2019. 567(7749): p. 530-534.

385. Khan, O., et al., TOX transcriptionally and epigenetically programs CD8(+) T

cell exhaustion. Nature, 2019. 571(7764): p. 211-218.

386. Scott, A.C., et al., TOX is a critical regulator of tumour-specific T cell

differentiation. Nature, 2019. 571(7764): p. 270-274.

387. Yao, C., et al., Single-cell RNA-seq reveals TOX as a key regulator of CD8(+) T

cell persistence in chronic infection. Nat Immunol, 2019. 20(7): p. 890-901.

388. Seo, H., et al., TOX and TOX2 transcription factors cooperate with NR4A

transcription factors to impose CD8(+) T cell exhaustion. Proc Natl Acad Sci U S

A, 2019. 116(25): p. 12410-12415.

389. Yu, W., et al., Clonal Deletion Prunes but Does Not Eliminate Self-Specific

alphabeta CD8(+) T Lymphocytes. Immunity, 2015. 42(5): p. 929-41.

390. Baxter, A.G. and P.D. Hodgkin, Activation rules: the two-signal theories of

immune activation. Nat.Rev.Immunol., 2002. 2(6): p. 439-446.

391. Cai, Z. and J. Sprent, Influence of antigen dose and costimulation on the primary

response of CD8+ T cells in vitro. J.Exp.Med., 1996. 183(5): p. 2247-2257.

392. Teh, H.S. and S.J. Teh, High concentrations of antigenic ligand activate and do

not tolerize naive CD4 T cells in the absence of CD28/B7 costimulation. Cell

Immunol, 1997. 179(1): p. 74-83.

393. Wang, B., et al., Naive CD8+ T cells do not require costimulation for

proliferation and differentiation into cytotoxic effector cells. J.Immunol., 2000.

164(3): p. 1216-1222.

325

394. Kane, L.P., J. Lin, and A. Weiss, It's all Rel-ative: NF-kappaB and CD28

costimulation of T-cell activation. Trends Immunol, 2002. 23(8): p. 413-420.

395. Fujii, S., et al., The linkage of innate to adaptive immunity via maturing dendritic

cells in vivo requires CD40 ligation in addition to antigen presentation and

CD80/86 costimulation. J.Exp.Med., 2004. 199(12): p. 1607-1618.

396. Pasare, C. and R. Medzhitov, Toll pathway-dependent blockade of CD4+CD25+

T cell-mediated suppression by dendritic cells. Science, 2003. 299(5609): p.

1033-6.

397. Wan, S., C. Xia, and L. Morel, IL-6 produced by dendritic cells from lupus-prone

mice inhibits CD4+CD25+ T cell regulatory functions. J Immunol, 2007. 178(1):

p. 271-9.

398. Blander, J.M. and R. Medzhitov, Toll-dependent selection of microbial antigens

for presentation by dendritic cells. Nature, 2006. 440(7085): p. 808-812.

399. Delamarre, L., et al., Differential lysosomal proteolysis in antigen-presenting cells

determines antigen fate. Science, 2005. 307(5715): p. 1630-1634.

400. Grossman, Z. and W.E. Paul, Adaptive cellular interactions in the immune

system: the tunable activation threshold and the significance of subthreshold

responses. Proc.Natl.Acad.Sci.U.S.A, 1992. 89(21): p. 10365-10369.

401. Grossman, Z. and W.E. Paul, Dynamic tuning of lymphocytes: physiological

basis, mechanisms, and function. Annu Rev Immunol, 2015. 33: p. 677-713.

402. Grossman, Z. and W.E. Paul, Self-tolerance: context dependent tuning of T cell

antigen recognition. Semin Immunol, 2000. 12(3): p. 197-203; discussion 257-

344.

326

403. Grossman, Z. and W.E. Paul, Autoreactivity, dynamic tuning and selectivity. Curr

Opin Immunol, 2001. 13(6): p. 687-98.

404. Sharpe, A.H., Mechanisms of costimulation. Immunol Rev, 2009. 229(1): p. 5-11.

405. Cox, M.A., L.E. Harrington, and A.J. Zajac, Cytokines and the inception of CD8

T cell responses. Trends Immunol, 2011. 32(4): p. 180-6.

406. Bierer, B.E. and S.J. Burakoff, T cell adhesion molecules. Faseb j, 1988. 2(10): p.

2584-90.

407. Schmidl, C., et al., Epigenetic mechanisms regulating T-cell responses. J Allergy

Clin Immunol, 2018. 142(3): p. 728-743.

408. Singh, N.J., M. Cox, and R.H. Schwartz, TLR ligands differentially modulate T

cell responses to acute and chronic antigen presentation. J Immunol, 2007.

179(12): p. 7999-8008.

409. Cibrián, D. and F. Sánchez-Madrid, CD69: from activation marker to metabolic

gatekeeper. Eur J Immunol, 2017. 47(6): p. 946-953.

410. Fujita, H., et al., Mechanisms of allergen-specific immunotherapy. Clin Transl

Allergy, 2012. 2(1): p. 2.

327