THE ALTERNATIVE NF-B PATHWAY IN MATURE B CELL DEVELOPMENT

Nilushi De Silva

Submitted in partial fulfillment of the

requirements for the degree of

Doctor of Philosophy

in the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY

2015

© 2015 Nilushi De Silva All rights reserved ABSTRACT

The alternative NF-B pathway in mature B cell development

Nilushi De Silva

The nuclear factor-B (NF-B) signaling cascade is comprised of two branches, the canonical and alternative NF-B pathways. Signaling through the alternative NF-B pathway culminates in the activation of the downstream transcription factor subunits, RELB and NF-B2. The biological roles of RELB and NF-B2 within the B cell lineage have been obscured in constitutional knockout mice by the diverse functions of these subunits in non-B cell types. To overcome these limitations, conditional alleles were generated to investigate the roles of RELB and NF-

B2 in B cell development. These alleles allowed the identification of complex functional requirements for RELB and/or NF-B2 in naïve B cells, germinal center (GC) B cells and plasma cells (PCs). These functional requirements may have implications for B cell malignancies that display mutations that constitutively activate the alternative NF-B pathway.

A large body of work has demonstrated that B cell activating factor (BAFF) signaling is critical for the maintenance of mature B cells. However, the contribution of the alternative NF-B subunits that are activated downstream of BAFF remained unclear, especially in regards to their specific target genes. We have identified critical, B cell-intrinsic roles for RELB and NF-B2 in the maintenance of mature B cells. In response to BAFF, these subunits were found to control the expression of anti-apoptotic genes, genes that ensure correct positioning within the B cell niche, and genes involved in promoting B–T cell interactions that allow effective antigen- mediated activation.

During the GC B cell reaction, light zone (LZ) B cells undergo affinity-based selection mediated by T follicular helper (Tfh) cells. A subset of LZ B cells show activation of the NF-B signaling cascade, suggesting a critical role for NF-B in the selection of high-affinity GC B cells. We here report that GC B cell development occurred normally in mice with conditional deletion of either relb (RELB) or nfkb2 (NF-B2) in GC B cells. In contrast, the simultaneous ablation of both subunits caused rapid involution of established GCs, similar to what has been observed for ablation of the canonical NF-B transcription factor subunit c-REL. Intriguingly,

RNA-sequencing analysis of relb/nfkb2-deleted GC B cells revealed no overlap between the genes controlled by RELB/p52 and c-REL within GC B cells. This suggests that signaling through the separate NF-B pathways in GC B cells results in the expression of different biological programs that are independently required for the maintenance of the GC reaction.

In addition, we observed that human PCs and PC precursors within the LZ showed high protein levels of NF-B2 compared to surrounding lymphocytes, suggesting a biological role for this subunit in PCs. Indeed, ablation of nfkb2 alone in GC B cells led to a dramatic decrease in antigen-specific serum IgG1 and antigen-specific IgG1-secreting cells. Interestingly however, the mice developed normal frequencies of PCs, suggesting a role for NF-B2 in PC physiology rather than differentiation.

TABLE OF CONTENTS

LIST OF FIGURES ...... iii

ACKNOWLEDGMENTS ...... v

COMMONLY USED ABBREVIATIONS ...... vii

CHAPTER 1: Background ...... 1

Section 1.1: Overview of B cells in adaptive immunity ...... 2

Section 1.2: Mature B cell development ...... 3

Section 1.3: The germinal center (GC) reaction ...... 10

Section 1.4: Plasma cell (PC) development ...... 18

Section 1.5: The GC and lymphomagenesis ...... 24

CHAPTER 2: Impairment of mature B cell maintenance upon combined ablation of the alternative NF-B transcription factors RELB and NF-B2 in B cells ...... 49

Section 2.1: Introduction ...... 50

Section 2.2: Materials and methods ...... 54

Section 2.3: Results ...... 56

Section 2.4: Discussion ...... 67

CHAPTER 3: Activation of the alternative NF-Bpathway in human GC B cells ...... 97

Section 3.1: Introduction ...... 98

Section 3.2: Materials and methods ...... 100

Section 3.3: Results ...... 104

Section 3.4: Discussion ...... 111

CHAPTER 4: Maintenance of the GC reaction requires signaling through the alternative NF-B pathway ...... 131

Section 4.1: Introduction ...... 132

Section 4.2: Materials and methods ...... 134

i

Section 4.3: Results ...... 138

Section 4.4: Discussion ...... 146

CHAPTER 5: Conclusions ...... 173

REFERENCES ...... 183

APPENDICES ...... 209

ii

LIST OF FIGURES

Figure 1.1. Stages of mature B cell development ...... 36

Figure 1.2. Structure of the spleen ...... 38

Figure 1.3. Dynamics of the GC reaction and selection of high-affinity antibody mutants ...... 40

Figure 1.4. Mutually repressive genetic programs govern the B cell/GC B cell vs. PC fate ...... 42

Figure 1.5. The NF-B signaling cascade ...... 44

Figure 1.6. Genetic aberrations in multiple components of the NF-B signaling cascade are observed in B cell malignancies ...... 46

Figure 2.1. Generation of a conditional relb allele ...... 72

Figure 2.2. Generation of a conditional nfkb2 allele ...... 74

Figure 2.3. Generation of mice with conditional deletion of relb or nfkb2 in B cells and simultaneous expression of eGFP ...... 76

Figure 2.4. Reduced fractions of splenic B cells in the absence of alternative NF-B subunits ..... 78

Figure 2.5. Reduced numbers of splenic B cells in the absence of alternative NF-B subunits ..... 80

Figure 2.6. relbfl/flnfkb2fl/flCD19-Cre mice display fewer B cell follicles within the white pulp, and exhibit more heterogeneity in the size of B cell follicles, compared to control mice...... 82

Figure 2.7. Reduced fraction of mature B cells in the bone marrow in relbfl/flnfkb2fl/flCD19-Cre mice ...... 84

Figure 2.8. Counter selection against relb/nfkb2-deleted MZ B cells in relbfl/flnfkb2fl/flCD19-Cre mice (flow cytometry) ...... 86

Figure 2.9. Counter selection against relb/nfkb2-deleted MZ B cells in relbfl/flnfkb2fl/flCD19-Cre mice (immunohistochemistry) ...... 88

Figure 2.10. Identification of BAFF-responsive genes controlled by the alternative NF-kB pathway by RNA-seq analysis of BAFF-stimulated nfkb2-deleted and WT B cells ...... 90

Figure 2.11. Control of BAFF-mediated survival by the alternative NF-B pathway ...... 92

Figure 2.12. Differential expression of BAFF-responsive genes controlled by the alternative NF-B pathway ...... 94

Figure 3.1. Activation of the canonical and alternative NF-B pathways in the human Burkitt lymphoma line, P3HR1, upon CD40 stimulation ...... 118

iii

Figure 3.2. Expression and activation of canonical and alternative NF-B subunits in normal and transformed human GC B cells and PCs ...... 120

Figure 3.3. ChIP analyses in human GC-derived lymphoma cell lines and normal GC B cells ...... 122

Figure 3.4. ChIP analyses in human GC-derived lymphoma cell lines ...... 124

Figure 3.5. Western blot analyses of ChIPs of human GC-derived lymphoma lines ...... 126

Figure 3.6. Knockdown of NF-B subunits in P3HR1 cells ...... 128

Figure 4.1. Combined GC B cell-specific deletion of relb and nfkb2 impairs the GC reaction ...... 154

Figure 4.2. Normal fractions of GC B cells are present in relbfl/flnfkb2fl/flCγ1-Cre mice up to day 7 of the GC reaction ...... 156

Figure 4.3. Normal DZ/LZ fractions and proliferation capacity of GC B cells in relbfl/flnfkb2fl/flCγ1- Cre mice ...... 158

Figure 4.4. Identification of genes controlled by the alternative NF-B subunits, RELB and NF- B2, in GC B cells at day 7 of the GC reaction ...... 160

Figure 4.5. Reduced expression of genes with metabolic and immune trafficking functions in RELB/NF-B2-deficient GC B cells at day 7 of the GC reaction ...... 162

Figure 4.6. Normal PC frequencies following GC B cell-specific deletion of alternative NF-B subunits ...... 164

Figure 4.7. Normal PC frequencies in nfkb2fl/flCγ1-Cre mice ...... 166

Figure 4.8. GC B cell-specific deletion of nfkb2 results in reduced antigen-specific serum IgG1 and antigen-specific IgG1 ASCs ...... 168

Figure 4.9. In vitro plasmablastic differentiation of NF-B2-deficient B cells ...... 170

Figure 5.1. Requirement of NF-B subunits in GC B cell and PC development ...... 180

iv

ACKNOWLEDGMENTS

My time at Columbia has been a challenging yet ultimately very rewarding experience and I would not have been able to complete this journey without the support of many people. First and foremost, I must thank Ulf Klein for being an exceptional mentor over the last 6 years. I thank him for taking a chance on me and for being a constant source of encouragement and advice, and a reassuring voice of reason when things were not going well. Ulf set high expectations and entrusted me with difficult projects. His confidence that I could tackle any challenge inspired me to try many new things and I have grown immeasurably as a scientist under his mentorship. On a personal level I also want to thank Ulf for being a caring, funny and engaged mentor. I have never once regretted joining Ulf’s lab and it has been my honor to be his first graduate student.

I must also thank all the members of the Klein lab for creating a warm and collaborative lab environment. To Amanda Carette, who has become a dear friend, I want to thank her for her exuberant personality and being so full of life, these traits made it fun to come to work every day. I want to thank Nicole Heise, for being a calm, patient presence within the lab. Nicole and I collaborated extensively and I have learned countless things from her. I also thank Giorgia for being an example of pure determination and Maja for approaching all challenges with a great attitude. To our technicians, Katie and Mick, I thank them for taking care of my mice, this has been an invaluable contribution to my work and allowed me to focus on experiments.

I would also like to thank my committee members, Sankar Ghosh, Adolfo Ferrando and

Steve Reiner for their guidance, time and valuable suggestions over the course of many meetings. I have always left these meetings feeling encouraged, with a clear concept of what the next steps in my project should be. I also thank Lei Ding and Sergei Koralov for serving on my thesis committee.

On a professional level I thank all the members of the Dalla-Favera lab for advice and reagents, in particular Claudio Scuoppo for helping me establish knockdown protocols, David

v

Dominguez-sola for teaching me how to use GSEA and Anthony Holmes for bioinformatic assistance. I also thank Govind Bhagat for histological evaluations and Joseph Haddad for providing us with human tonsils. I want to acknowledge Victor Lin and the HICCC Transgenic

Mouse Facility, Columbia Genome Center, New York Genome Center and the HICCC Flow

Cytometry Facility for critical support. I am grateful to the Cancer Biology Training Program

(T32) grant for financial support. I also want to express my appreciation to the Department of

Microbiology and Immunology for being a supportive, tight-knit department and in particular

David Fidock, for recruiting me from Australia.

On a personal note, I must thank Matt Borok, for almost 6 years of love and companionship, for being my greatest supporter and for forcing me to take a break when I needed it. I also thank him professionally, as by chance we had oddly parallel projects, and spent many days troubleshooting common experimental issues. I have shared my experience as a Columbia graduate student with fantastic classmates – Lucja Grajkowska, Sarah Deng,

Angelo Pefanis, Tulsi Patel and Olya Yarychkivska. None of us expected to make such great friends when we started, and I am especially grateful for this as it has eased my transition into starting a new life in the USA. These friendships have also made it easier to survive the many ups and downs of graduate school, and I want to thank Lucja is particular for being an almost daily life-line.

I want to end by thanking my amazing parents, Ruvani and Sarath De Silva, for all that they have done for me. It’s not possible for me to articulate all the ways in which I am grateful to them. To name just a few, I thank them for the courage, hard work and resilience it took to give me the beautiful childhood I had in Australia. I thank them for providing me with opportunities that they never had and for pushing me to take advantage of them. I thank them for instilling me with a strong sense of discipline, which has made it easier to navigate all stages of my education. Finally, I thank them for all their years of love and for their unshakeable belief and pride in me, and all the things I do.

vi

COMMONLY USED ABBREVIATIONS

ABC-DLBCL: activated B cell-like diffuse large B cell lymphoma

ASC: antibody secreting cell

BCR: B cell receptor

B-NHL: B cell non-Hodgkin lymphoma

ChIP: chromatin immunoprecipitation

CSR: class switch recombination

DZ: dark zone

FDC: follicular dendritic cell

GC: germinal center

GCB-DLBCL: germinal center B cell-like diffuse large B cell lymphoma

Ig: immunoglobulin

IgV: immunoglobulin variable region

IP immunoprecipitation

LZ: light zone

MHC: major histocompatibility complex

MM: multiple myeloma

MZ: marginal zone

PC: plasma cell

SHM: somatic hypermutation

SRBCs: sheep red blood cells

TAD: transactivation domain

TD: T-dependent

TI: T-independent

T1, T2: transitional stage 1, 2

vii

This page has been intentionally left blank.

viii

CHAPTER 1: Background

1

Section 1.1: OVERVIEW OF B CELLS IN ADAPTIVE IMMUNITY

Following exposure to a foreign pathogen, the innate branch of the immune system is engaged to mount an early response. This response is coordinated by a range of cell types, which may include macrophages, neutrophils and/or natural killer cells, depending on the nature of the pathogen. Early responses are rapid and characterized by lack of antigen specificity to the invading pathogen. Following this first wave of innate immunity, an antigen-specific immune response is mounted. This adaptive immune response enables both effective pathogen clearance and protection against future encounters with the invading pathogen via the generation of immunological memory.

The mediators of adaptive immunity are B and T lymphocytes. T lymphocytes (T cells) lyse infected cells and provide activation signals to B lymphocytes (B cells). B cells differentiate into plasmablasts or plasma cells (PCs) that secrete antibodies, also known as immunoglobulin

(Ig), which are crucial for protection against microorganisms. This is best exemplified by the immunodeficiency syndromes that occur when B cells are absent or function abnormally.

Antibodies allow binding to antigenic structures on the surface of microbes enabling neutralization, opsonization and clearance. They are comprised of two identical heavy chains and two identical light chains which consist of variable and constant regions. The variable regions of the heavy and light chain mediate binding to antigens whereas the constant regions of the heavy chain impart various effector functions on the antibody that tailor it to optimally combat particular microbes.

Antibodies are the secreted form of the B cell-receptor (BCR) that is assembled and expressed on the surface of B cells undergoing B cell development in the bone marrow.

Developing B cells progress through the pro, pre and immature B cell stages during which BCR assembly occurs through a process of somatic DNA recombination known as VDJ recombination. VDJ recombination assembles the variable region of an antibody via the juxtaposition of distinct gene segments encoded within the heavy and light chain Ig loci.

2

Variable (V), diversity (D) and joining (J) segments are combined in the case of the heavy chain, whereas V and J gene segments are combined in the case of the light chain. As the heavy and light chain Ig loci within germ-line DNA encode many different V, D and J gene segments, a multitude of combinations are possible during VDJ recombination. The addition of non-germline encoded nucleotides (N-nucleotides) by terminal nucleotidyl transferase (TdT) inbetween the

VD, DJ or VJ gene segments further contribute to the diversity of the antibody repertoire. As early B cell development proceeds within the bone marrow, processes of positive and negative selection ensure the survival of only those immature B cells expressing a functional BCR that is not strongly reactive to self. An enormous repertoire of B cells, each expressing a BCR with a unique antigenic specificity, is produced during B cell development. This repertoire is theoretically capable of recognizing any foreign antigen.

Section 1.2: MATURE B CELL DEVELOPMENT

B cell developmental processes lead to the generation of three subsets of mature B cells, namely follicular B cells, marginal zone (MZ) B cells and B1 cells1,2. B1 cells develop during fetal life3. In contrast, follicular and MZ B cells, also known as B2 cells or conventional B cells, are derived from immature B cells completing B cell development within the bone marrow (Fig. 1.1).

B1 cells and MZ B cells survey sensitive sites that experience continual exposure to microbial antigens and provide a first line of defense. In the case of B1 cells, these are the peritoneal and pleural cavities, whereas MZ B cells reside within a specialized compartment of the spleen where they monitor the blood. Follicular B cells are named for their primary localization within the B cell follicles of lymphoid tissues; however, these cells recirculate between the follicles of the spleen, lymph nodes and Peyer’s patches via the blood and lymph. Follicular B cells are also known as naïve B cells, as they have not yet encountered antigen. Upon antigen encounter, follicular B cells become activated, ultimately leading to their differentiation into PCs

3 that secrete high affinity antibodies (Fig. 1.1). The development, localization and function of B2 cells is discussed in this section.

Transitional B cell stages prior to mature B cell development

Immature B cells developing within the bone marrow progress through two transitional stages

(T1 and T2) prior to becoming a mature B cell4 (Fig. 1.1). B cells expressing BCRs with intermediate reactivity to self can progress through these transitional stages; however, these cells are removed prior to final maturation5-7. T1 and T2 B cells retain the B cell lineage marker of immaturity, CD93/AA4.1, and as T1 cells transition into T2 B cells, they acquire the ability to recirculate1. The final maturation of T2 B cells into mature B cells occurs within the bone marrow environment for approximately one quarter of immature B cells, whereas the majority of developing B cells traffic to the spleen to complete maturation8,9. T2 B cells are able to follow two distinct mature B cell developmental pathways resulting in the generation of either follicular

B cells or MZ B cells (Fig. 1.1). Follicular B cells are characterized by high expression of the antibody isotypes IgM and IgD and intermediate expression of the complement receptor CD21, whereas MZ B cells have high expression of IgM and CD21 and low expression of IgD. Most T2

B cells develop into follicular B cells and only a small proportion of cells follow the MZ B cell developmental path.

Follicular vs. MZ B cell fate commitment

The integration of several signals dictates the decision to follow a follicular vs. MZ B cell fate, and these signals often represent a balance between opposing regulatory factors. NOTCH2 signaling is required for commitment to the MZ B cell fate as MZ B cells do not develop in the absence of NOTCH2 or a DNA binding molecule critical for NOTCH signaling, recombination signal binding protein for immunoglobulin kappa J region (RBP-J)10,11. The physiologically relevant ligand of NOTCH2 that activates signaling in B cells is delta-like 1 (DL1), which is absent from the bone marrow and is highly expressed in the splenic venules12. MSX2-

4 interacting protein (MINT) is a protein capable of competing with the intracellular domain of

NOTCH for binding to RBP-J and is therefore an inhibitor of NOTCH signaling. While low expression of MINT is observed in MZ B cells, it is highly expressed in follicular B cells and inactivation leads to expansion of MZ B cells13.

The follicular vs MZ B cell fate is also governed by a similar dichotomy between the transcription factor E2A and its antagonist, inhibitor of DNA binding 3 (ID3). Low levels of E2A and high levels of ID3 promote MZ B cell development, whereas high levels of E2A and low levels of ID3 are associated with follicular B cell development14. The transcription factor friend leukemia integration 1 (FLI1) further regulates this balance15 and NOTCH signaling may also have a regulatory role via the degradation of E2A16.

Persistent, weak BCR signaling is required throughout B cell development (discussed later in this section). The strength of the signal through the BCR at the T2 stage of development is thought to influence the decision to differentiate into either a follicular or MZ B cell. In this signal-strength model, weak BCR signaling is believed to promote MZ B cell differentiation, whereas strong signaling favors follicular B cell differentiation17-19. This signaling strength is linked to the balance between Bruton’s tyrosine kinase (BTK), which is activated by BCR signaling, and the transcription factor Aiolos, which may negatively regulate BTK. Therefore, in the absence of Aiolos, strong BCR signaling is present and follicular B cell development is favored, whereas MZ B cell development is promoted in the absence of BTK. However,

NOTCH2 signaling may be sufficient to drive the MZ fate as constitutive NOTCH2 signaling can promote the differentiation of MZ B cells in a BCR-independent manner20.

Follicular vs. MZ B cell localization

Splenic structure

The spleen is comprised of red pulp and white pulp. Whereas the red pulp functions to filter the blood and remove aged erythrocytes, the white pulp is lymphoid tissue comprised of B cell

5 follicles that are bordered by T cell areas21 (Fig. 1.2). Blood is supplied to the spleen via the splenic artery which branches into central arterioles that further divide into smaller arterioles.

These branching arterial vessels are sheathed by white pulp. Branching arterioles penetrate the marginal sinus, a fenestrated vessel with low resistance21. Circulating blood enters the marginal sinus and filters slowly into an adjacent region, the MZ, before entering the red pulp and the general circulation21 (Fig. 1.2). While the spleen and lymph nodes differ in a number of structural ways, the basic organization of the splenic white pulp into T cell areas and B cells follicles is very similar to that observed within lymph nodes22.

Follicular B cell localization

B cells travelling in the circulation to the spleen migrate into the B cell follicles of the white pulp through the marginal sinus following a gradient of the chemokine, CXCL13. This movement is directed by the expression of the chemokine receptor, CXCR5, which is expressed on mature B cells and is essentially absent on immature B cells23,24,25. CXCR5 is required for homing of mature B cells into splenic follicles as mice lacking CXCR5 do not form splenic follicles and B cells are instead present as a ring at the boundary of the T cell area25,26. Together with adjacent stromal cells27, CXCL13 is also produced by follicular dendritic cells (FDCs) that form a network within the B cell follicles25. FDCs sequester antigen within the follicle and present it in antigen- antibody complexes using Fc receptors and complement receptors28. FDCs are non- haematopoietic in origin and are named for their dendritic morphology rather than relation to dendritic cells (DCs)29. B cells and FDCs reinforce their respective localizations, as signaling through CXCR5 induces the expression of lymphotoxin-α1β2 (LT-α1β2) on B cells, which in turn promotes both FDC differentiation and CXCL13 expression25.

The T cell areas of the white pulp are comprised of T cells and DCs, and these cells are attracted to the T cell area via the chemokines CCL19 and CCL21 that bind the chemokine receptor CCR730,31. T cells are activated upon the entry of activated DCs bearing antigen into

6 the white pulp. This activation leads to upregulation of CXCR5 expression together with a downregulation of CCR7 expression; these events result in the migration of T cells towards the border of the B cell follicles32. Within B cell follicles, binding of antigen to BCRs induces the upregulation of CCR7 on the surface of B cells, promoting their migration to the boundary of the

B cell follicles33-35. At this B/T interface B cells receive signals from activated T cells that initiate their activation33-35; this T cell-dependent activation leads to the initiation of a germinal center

(GC) response (see Section 1.3).

MZ B cell localization

Murine MZ B cells are found only in the splenic MZ, a region that separates the outer rim of the white pulp from the red pulp (Fig. 1.2). The architecture of the spleen ensures that the MZ is exposed to blood from the circulation21. It consists of a reticular fibroblast framework that is continuous with the reticular network of the red pulp, and the MZ is therefore considered an open compartment21. MZ macrophages and MZ B cells reside within this reticular network, while

DCs, lymphocytes and granulocytes transit through this area21.

Sphingosine-1-phosphate (S1P) is a lysophospholipid present in blood that binds to the

S1P receptors, S1P1 and S1P3. The MZ has a high concentration of S1P as it is a blood-rich environment whereas low concentrations are found in the follicles and T cell areas of lymphoid

36,37 tissues . The S1P receptors S1P1 and S1P3 are expressed on MZ B cells at higher levels than those observed on follicular B cells38. In addition to this S1P ligand/receptor distribution, although follicular B cells and MZ B cells express similar levels of CXCR5, the ligand for this receptor, CXCL13, is found in abundance within the follicles, whereas CXCL13 transcripts are absent in the MZ27,38,39. As such, while follicular B cells home to the follicles following a CXCL13 gradient, the S1P ligand/receptor interaction (principally through S1P1) is able to overwhelm this attraction resulting in the migration of MZ B cells into the MZ38,40. Once MZ B cells enter the

MZ, they are retained within this environment due to their high expression of several integrins

7 that enable them to resist the strong sheer forces created by the flow of blood. These include the integrin, lymphocyte function-associated antigen 1 (LFA1, comprised of αL and β2 chains), which binds to intracellular adhesion molecule 1 (ICAM1), and the integrin, very late antigen 4

(VLA4, comprised of α4 and β1 chains chains), which binds to vascular cell adhesion molecule 1

(VCAM1)41. ICAM1 and VCAM1 are expressed on stromal cells in the MZ. Finally, in addition to directing MZ B cell development, continued expression of NOTCH2 is also important for the retention of MZ B cells within the MZ42.

MZ B cell function

MZ B cells are situated at the interface between the circulatory system and the lymphoid tissue of the white pulp and are thereby ideally located to survey the blood and respond rapidly to blood-borne pathogens. MZ B cells are activated upon interaction with antigen captured by metallophilic macrophages that line the marginal sinus, resident MZ macrophages or circulating

DCs and granulocytes43,44. Activation can also occur by direct interaction with antigen opsonized with complement45,46. MZ B cells are particularly well equipped to respond rapidly to antigen in a

T-independent (TI) fashion. They have a lower antigen-activation threshold than follicular B cells and high expression of Toll-like receptors (TLRs)47. TLR signaling can integrate with BCR signaling to promote optimal MZ B cell activation48-50 and TLR stimulation alone is also capable of driving the differentiation of MZ B cells into plasmablasts49. MZ B cells also have high expression of the complement receptors CD21 and CD35, allowing efficient binding of antigens opsonized by complement45,51,52.

Upon antigen activation, MZ B cells rapidly differentiate into plasmablasts primarily via a

TI response to carbohydrate antigens44,45,53-55. Mice that lack MZ B cells10,45 and asplenic individuals56-58 are therefore particularly susceptible to infection with encapsulated bacteria.

Antibodies secreted by these plasmablasts generally have low affinity for antigen; however, as

8 the generation of a high-affinity response by follicular B cells requires one week to develop (see

Section 1.3), MZ B cells contribute to bridging this temporal gap in immunity.

MZ B cells are also capable of differentiating into plasmablasts under homeostatic conditions, giving rise to “natural” antibodies. These natural antibodies are capable of recognizing conserved molecular patterns present on self and foreign antigen59. MZ B cells are therefore involved in the recognition of bacteria, including commensal bacteria, and host cell debris59. B1 cells provide an additional source of natural antibodies59.

Signals required for mature B cell survival

Only those B cells that have assembled a functional BCR survive to maturity60. Mature B cells are then dependent on BCR signaling for survival, as conditional or inducible ablation of the

BCR or BCR associated components leads to the rapid death of mature B cells in vivo61,62.

Persistent BCR signaling is therefore a requirement for mature B cell maintenance and is termed “tonic” signaling. The nature of this BCR stimulation, however, remains unclear and may occur due to low-affinity interactions with self antigens or proceed in a ligand-independent manner63.

An additional survival signal is required from the T2 stage of B cell development onwards as B cells enter the follicular environment. This signal is derived from binding of the B cell activating factor-receptor (BAFF-R) to the ligand BAFF. The importance of this signal is exemplified by the severe reduction in mature B cells observed in mice deficient for either the

BAFF-R or BAFF and the rapid death of mature B cells following BAFF neutralization64-73. B cell expansion is also observed in mice with transgenic BAFF expression or following BAFF administration74-77. The BAFF-R is first expressed in T2 B cells and sensitivity to BAFF-mediated survival signals begins at this stage64,67,78-81. BAFF is a soluble factor produced by FDCs and other cell types within the B cell follicle82 and the level of available BAFF serves as a limiting factor for B cell survival and determines the size of the B cell pool. BAFF signaling activates

9 many pathways that promote survival83-87, including the NF-B signaling cascade which will be explored further in Section 1.5.

Section 1.3: THE GERMINAL CENTER (GC) REACTION

GCs are transient structures that form within peripheral lymphoid organs in response to T- dependent (TD) antigens88. Within the GC, mutational processes lead to the generation of mutant clones possessing a broad spectrum of affinity for the immunizing antigen. Highly effective selection processes within the GC then ensure that inferior antibody-mutants or those with autoreactive specificities are outcompeted by higher affinity competitors89. The GC reaction is remarkably efficient as antigen-specific memory and PCs appear within one week after antigen-encounter90. Over time, several iterative rounds of Darwinian-like mutation and selection occur, ultimately resulting in the generation of memory B cells and PCs that exit the GC reaction expressing an antibody repertoire whose affinity increases in a stepwise manner. This phenomenon is known as affinity maturation and drives the effective clearance of invading pathogens, which is best illustrated by the immunodeficiency syndromes observed in patients who are unable to form GCs91. Discussed in this section are the events leading to the formation of the GC, processes that drive mutation and selection within the GC and molecular regulators that coordinate the GC reaction.

Initiation of the GC B cell reaction

GCs form within the B cell follicles of lymphoid organs (Fig. 1.2). The first step in GC formation is the activation of naïve B cells by exogenous antigen within the follicle92 (Fig. 1.3). Intravital imaging93,94 has demonstrated that at day 1 post-immunization, antigen-activated B cells migrate to the border of the T cell/B cell zone or the interfollicular region in the lymph nodes where they proliferate and form long-lived interactions with antigen-specific T cells to become fully activated34,95. At this time, T cells, which have committed to T follicular helper (Tfh) cell development through priming by DCs96, begin to upregulate B cell lymphoma 6 (BCL6)93,97, a

10 critical regulator of both the Tfh98-100 and GC B cell program (discussed later in this section). In contrast, low expression of BCL6 is only detectable in interacting B cells at day 2. At day 3, Tfh cells migrate from the interfollicular area into the follicle, and pre-GC B cells follow at day 4 and form a primordial GC at the center of the follicle. This movement is dependent on downregulation of the chemotactic receptor Epstein-Barr virus induced molecule-2 (EBI2), which drives B cell localization to the outer follicle by responding to oxysterol ligand101-105. The S1P receptor, S1P2, is also required for GC-confinement of B cells by inhibiting migration to the outer follicle106.

The formation of the dark and light zones of the GC

Primordial GCs can be identified histologically at day 4 after immunization within follicles. At this time, B cells grow and differentiate into blasts that rapidly divide and begin to fill the network of

FDCs in center of the follicle88. During this process, the blasts displace the IgM+IgD+ B cells resulting in the formation of the mantle zone around the GC. From day 5-6 post-immunization, the GC increases in size due to the rapid proliferation of the blasts, which continues until day 7, the time-point at which the GC has been fully established and is polarized into two microenvironments known as the dark and light zones88,107 (Fig. 1.3). The dark zone (DZ) is named for its histological appearance and consists of densely packed GC B cells that proliferate within a recently identified interconnected network of reticular cells that express the chemokine

CXCL12108. These reticular cells are morphologically similar to the FDCs of the light zone (LZ).

The LZ is more sparsely populated by B cells than the DZ and is comprised of several cell types, including Tfh cells, FDCs and macrophages107.

Generation of mutant GC B clones

Antibody-dependent immune responses are characterized by affinity maturation of the immune response. The mechanism of affinity maturation is centered on somatic hypermutation (SHM) of the variable regions of the Ig loci within proliferating GC B cells in the DZ89,109 (Fig. 1.3). The GC

11 is the primary site of SHM; however, extrafollicular SHM has been reported110. The process of

SHM is driven by activation-induced cytidine deaminase (AID)111,112 and involves the induction of

DNA strand breaks resulting primarily in the introduction of single nucleotide exchanges113,114.

While undergoing SHM, GC B cells proliferate at a rate unparalleled in mammalian tissues115.

The estimated mutational rate within the variable regions of the Ig loci is 10-3 per base pair per generation which is 10-6 fold higher than the rate of spontaneous mutation at other loci115. SHM is therefore a second point of diversification of the Ig loci, following VDJ recombination within the bone marrow, and is generally limited to the proliferative GC B cells within the DZ116. Robust division and SHM produces a pool of mutant GC B cell clones possessing a large spectrum of affinities to the immunizing antigen (Fig. 1.3).

Movements of B cells between the DZ and LZ of GCs

The DZs and LZs of the GC are organized by the expression of the chemokine receptors

CXCR4 and CXCR5117. CXCR4 is expressed on DZ B cells and is responsive to CXCL12 produced in the DZ108, and CXCR5 is expressed on LZ B cells (and follicular B cells) and is responsive to CXCL13 within the LZ (and throughout the follicle)117. An important advance in the field has been the ability to identify DZ vs. LZ B cells phenotypically using CXCR4 and the activation markers CD83 and CD86117-119. DZ B cells can be identified as CXCR4hiCD83loCD86lo and LZ B cells as CXCR4loCD83hiCD86hi. Classical models of the GC define the existence of two distinct GC B cell differentiation states88. Centroblasts were thought to reside within the DZ and differentiate into smaller centrocytes upon entry into the LZ. However, this classical model has been challenged by findings derived from studies that incorporated the use of intravital microscopy and photoactivation to visualize the movements of GC B cells for the first time119-122.

Rapid flux of GC B cells was observed between the DZ and LZ, and GC B cells in the DZ and

LZ were found to have similar morphology and a great overlap in gene expression. These cells

12 are therefore perhaps more accurately described as transient states that are associated with positioning in the DZ and LZ107,123.

Recirculation between the DZ and LZ is known as cyclic reentry. Mutant GC B cell clones move to the LZ where high affinity mutants are selected and several iterative rounds of cyclic reentry drive affinity maturation (Fig. 1.3). In the absence of CXCR4, GC B cells are restricted to the LZ and affinity maturation is inefficient108. GC polarization into DZs and LZs is therefore critical for optimal GC function and it is likely that the spatial separation of SHM and the selection of high-affinity GC B cells is required for efficient rounds of affinity maturation. In the absence of this polarization, inappropriate or premature selection may occur. The importance of the unique zonal structure of GCs is perhaps most strongly evidenced by its conservation in humans118,124.

Selection of high-affinity GC B cells

Tfh cells drive selection of high-affinity GC B cells

Following SHM within the DZ, high-affinity GC B cell mutants must be selected in the LZ before undergoing further rounds of hypermutation in the DZ. Traditionally, it was believed that the selection of high-affinity GC B cells was largely driven by a process of direct competition for antigen acquired from FDCs88. High-affinity GC B cells would sequester all available antigen via their BCRs and thereby deprive lower-affinity competitors of survival signals triggered by BCR ligation. However, using an experimental system in which antigen is delivered directly to GC B cells independent of the BCR, it was recently demonstrated that Tfh cells promote selection within the LZ119. GC B cells capture antigen via the BCR and present the processed antigen on major histocompatibility complexes (MHCs) to Tfh cells (Fig. 1.3). Higher BCR affinity is directly associated with greater antigen-capture and leads to a higher density of MHC-peptide presentation on the surface of the B cell. This results in the greatest share of T cell help, which in turn drives selection. Active BCR signaling is also largely quiescent within B cells125.

13

Therefore, within the context of the GC, it appears that the role of the BCR is primarily to capture and internalize antigen rather than induce BCR signaling. Visualization of these selection processes in vivo demonstrated that Tfh cells patrol GC B cells and form the largest and longest contacts with those cells presenting the highest density of antigen on MHC complexes126. These contacts are transient, thereby allowing Tfh cells to continuously survey many different LZ B cells, resulting in the selection of the antibody mutants with the highest affinity on a population level.

Mechanism of selection of high-affinity GC B cells

Tracking of cell division within the GC demonstrated that B cells that receive the greatest magnitude of T cell help are programmed to divide the greatest number of times upon reentry into the DZ127. As SHM and proliferation are intimately linked128, GC B cells that undergo the greatest number of divisions harbor significantly more affinity-enhancing mutations in their Ig variable region (IgV) genes compared to less proliferative competitors. Therefore, in a “feed- forward” loop, the GC B cells with the highest affinity are repeatedly instructed to recirculate to the DZ to undergo further rounds of division and SHM (Fig. 1.3). As a result of this iterative process, GC B cells with the highest affinity dominate the population of B cells participating in the GC reaction over time.

Signals derived from T cells during selection within the LZ

As outlined above, Tfh cells provide crucial signals to high-affinity GC B cells to allow their selection, transit back to, and division in the DZ. The exact nature of these signals and their downstream effects remains unclear and is being actively explored in the field. Murine and human LZ B cells are enriched for several early activation signatures, including CD40, BCR, nuclear factor-B (NF-B) and c-MYC signatures, suggesting that these pathways are activated during selection119,124,129. CD40 stimulation is achieved via ligation of the CD40 receptor, which is expressed on B cells, by CD40 ligand (CD40L), expressed on the surface of Tfh cells. An

14 additional T cell/B cell interaction is mediated by inducible T cell co-stimulator (ICOS) ligand

(ICOSL) which is expressed on GC B cells and promotes transient yet expansive ‘entanglement’ with Tfh cells that express the surface receptor ICOS. This interaction stimulates the increased display of CD40L on the surface of Tfh cells, thereby providing GC B cells with CD40- stimulation130. CD40-stimulation also leads to the upregulation of ICOSL on GC B cells, resulting in a feed-forward loop that enables high-affinity B cells to reiteratively acquire a greater share of

T cell help than lower-affinity competitors130. T cell-mediated CD40 stimulation and BCR signaling lead to the activation of several downstream pathways, including NF-B activation, which will be discussed in Section 1.5. Contact between Tfh and GC B cells has also been shown to induce increased calcium signaling and co-expression of IL-4 and IL-21 in Tfh cells, both of which are cytokines that augment GC B development126.

An additional DNA-modifying event occuring within the GC is Ig class switch recombination (CSR). Like SHM, this process is driven by AID111,112. CSR allows B cells to switch the constant region of the heavy chain Ig locus, enabling Ig isotype switching from IgM and IgD to other classes such as IgG, IgE or IgA, while retaining the same BCR specificity. This switch occurs via DNA recombination of switch regions that intercede the constant region gene segments of the Ig heavy chain locus. Different isotypes are associated with specific effector functions. While CSR can occur both in a TI or TD manner131, it is believed that within the GC, signals that induce CSR are derived from Tfh cells located in LZ132 (Fig. 1.3). CSR to various isotypes is promoted by co-stimulatory signals, such as CD40 and ICOSL stimulation from T cells133-135, and exposure to cytokines, such as IL4 which drives switching to IgG1136. It is believed that SHM and CSR are temporally separated. However, the precise developmental stage at which CSR occurs is unclear, as class-switched cells can be observed among DZ cells and post-GC memory B cells and PCs, suggesting that CSR can occur prior to cyclic reentry and/or post-GC differentiation137.

15

Molecular control of the GC reaction

The GC-reaction is a highly complex process involving several distinct, timed events that are segregated topographically within the GC microenvironment. The coordination of this complex cascade of events is underpinned by precise molecular control achieved by a series of transcription factors, epigenetic mediators and post-transcriptional mechanisms138,139. The function of the major regulators of these processes is discussed below.

BCL6

Within the B cell lineage, BCL6 is specifically expressed in GC B cells and is rapidly upregulated upon antigen activation140,141. BCL6 is essential during GC responses, since BCL6-deficient mice fail to form GCs142,143. During GC initiation, BCL6 controls a transcriptional program within pre-GC B cells that facilitates the migration of these cells to the center of the follicle via the

94,144 94 downregulation of both EBI2 and S1P1 , and the upregulation of CXCR4 . BCL6- upregulation in pre-GC B cells also appears to enhance integrin-dependent conjugation of B and T cells94.

BCL6 is a transcriptional repressor that is responsible for the specialized physiology of

GC B cells, including all DZ B cells and most LZ B cells. BCL6 represses the expression of genes encoding a range of targets that are involved in B cell activation and post-GC activation including CD80145, which promotes T cell interactions via the binding of CD28. BCL6 also represses genes encoding factors that drive PC development such B lymphocyte-induced maturation protein 1 (BLIMP1) and interferon regulatory factor 4 (IRF4)146-149 (see Section 1.4) and genes encoding several factors that modulate chemokine and cytokine responses146,150.

GC B cells are tolerant of SHM-induced DNA damage as BCL6 inhibits DNA damage-induced apoptosis by inhibiting the expression of the tumor suppressor p53151 and other DNA damage- response factors146,152-154. Repression of p53 also promotes rapid proliferation, which is further facilitated by BCL6 via the transcriptional repression of the cell cycle regulator p21152. BCL6 also

16 modulates a range of apoptotic genes146,147 and has been shown to directly target the gene encoding the anti-apoptotic factor B cell lymphoma 2 (BCL2)147,155. GC B cells express a pro- apoptotic transcriptional profile156 and are particularly prone to apoptosis ex vivo157-159; this susceptibility is thought to enable the rapid elimination of mutant GC B clones that are not rescued by survival signals provided during LZ selection. BCL6 therefore controls a broad genetic program that maintains and promotes the DZ phenotype by collectively ensuring that

GC B cells do not undergo premature activation or differentiation, are permissive to genotoxic stress induced by SHM and can be rapidly eliminated if they are not selected by T cells in the

LZ. c-MYC c-MYC exerts a critical role in virtually all proliferating cells. Through both transcriptional and non-transcriptional mechanisms, c-MYC supports robust proliferation by regulating diverse processes including cell-cycle progression, metabolism and telomere maintenance160. Despite the highly proliferative nature of DZ B cells, it was believed that GC B cells do not express c-

MYC156,161 as BCL6 represses transcription of the MYC locus162. Recently however, it was observed that the primordial GC arising at day 4 post-immunization is characterized by the appearance of a sizable fraction of c-MYC+BCL6+ blasts, and ablation of c-MYC in activated B cells shortly after immunization (day 2) inhibits GC formation163. These findings demonstrate an essential function for c-MYC in GC initiation. At day 4, the c-MYC+BCL6+ GC B cells, in contrast to their c-MYC–BCL6+ counterparts, transcribed the known direct c-MYC target gene ccnd2

(encoding Cyclin D2) in addition to ccnd3 (encoding Cyclin D3)163, the DZ-specific D-type cyclin164,165. As c-MYC+BCL6+ cells appear to be the precursors of c-MYC–BCL6+ GC B cells163, this finding suggests that c-MYC establishes a Ccnd2-dependent proliferation program in GC- precursor B cells that is replaced by the typical Ccnd3-dependent proliferation program of DZ B

17 cells164,165 upon c-MYC downregulation. From this point onwards, c-MYC is virtually absent in

DZ B cells until it is reexpressed at day 7162,163.

c-MYC expression occurs within a small subset of LZ B cells within mature GCs that have segregated into DZs and LZs at day 7 post-immunization162,163. This pattern of expression coincides with the time-point at which selection of high-affinity mutants is thought to begin90,166.

Moreover, c-MYC+ GC B cells demonstrated evidence of T cell-mediated selection as they possessed an activated phenotype similar to that described by gene expression profiling of LZ B cells119,162,163, were actively cycling, expressed high-affinity BCRs, and T cell help was found to induce c-MYC expression in LZ B cells prior to cyclic reentry162,163. These findings collectively suggest that c-MYC+ B cells within the LZ are undergoing affinity-based positive selection.

MCL1

Despite the general pro-apoptotic nature of GC B cells, one antiapoptotic BCL2 family member, myeloid cell leukemia 1 (MCL1), has recently been identified as the principal regulator of cell survival during GC-formation167. The conditional deletion of mcl1 early after immunization (d2) prevented the formation of GCs, suggesting a continuous requirement of MCL1 throughout GC development. In contrast, the pro-survival factor BCL2-like 1 (BCL-XL) is dispensable within this

167 context and antagonizing BCL-XL and BCL2 leads to a reduction in naïve B cells, but has no effect on GC B cells168.

Section 1.4: PLASMA CELL (PC) DEVELOPMENT

Upon completion of GC development, GC B cells exit the GC as either memory B cells or PCs

(Fig. 1.3). Memory B cells have the capacity to rapidly differentiate into PCs that secrete high affinity antibody upon reencounter with the immunizing antigen. The molecular events that drive memory B cell development remain largely unknown; however, it is clear that extensive heterogeneity exists within this cell population169. In contrast, PC development is understood in greater depth. All B cell subpopulations are capable of differentiating into ASCs upon antigen

18 encounter. ASCs include plasmablasts and PCs. Plasmablasts are short-lived and actively cycling and can further mature into PCs, which are long-lived and non-proliferative. The kinetics of this differentiation and the type of antibodies secreted can differ depending on the nature and location of the antigen, the type of responding B cell and the co-stimuli present. During a TD response, an initial low affinity wave of antibody production provides immediate protection.

Following a GC response, PCs are generated that produce a high affinity, persistent antibody response. The B cell—PC transition can be considered a lineage switch, and PCs are terminally differentiated as they have permanently extinguished the genetic program of the B cell.

Described in this section are early antibody responses, the mutually repressive genetic programs that govern B cell/GC B cell vs. PC fate and PC survival within the bone marrow.

Early antibody responses

As mentioned in Section 1.2, antibody responses can occur even in the absence of infection via the secretion of natural antibodies by PCs derived from MZ B cells or B1 cells170. Following infection, MZ B cells respond rapidly to blood-borne pathogens by differentiating into plasmablasts that secrete large amounts of low-affinity antibody. These plasmablasts develop as early as 1-3 days following microbial exposure, reside within the red pulp and are short- lived53,59. The antibody class these cells secrete is generally of the IgM subtype; however, switching via CSR to IgA and IgG can occur45,171. Movement to the red pulp is achieved via downregulation of S1P receptors, CXCR5 and CCR7 and upregulation of CXCR4, allowing plasmablasts to enter the red pulp following a CXCL12 gradient that is produced by stromal and innate immune cell types within the red pulp38,171-173.

Follicular B cells can also differentiate into plasmablasts and contribute to an early antibody response, as not all antigen-activated B cells enter the GC reaction170. Following interaction with T cells, a subset of B cells differentiates into short-lived plasmablasts that secrete low-affinity antibodies directed against the invading pathogen174. Among the pool of

19 responding B cells, it appears to be those with high-affinity antigen specificities that predominantly differentiate into plasmablasts175,176. Plasmablastic differentiation occurs in extrafollicular sites and movement to these sites is dependent on the expression of EBI2101,102.

In the lymph nodes, these extrafollicular sites are the medullary cords; in the spleen, foci of plasmablasts form in the red pulp170. Two days following immunization with a TD antigen, foci of plasmablasts appear and expand until day 8, when they recede and die via apoptosis177,178.

Following this initial wave of low affinity antibody, PCs derived from the GC reaction begin to secrete high affinity antibody.

Antagonism of the PC fate in B cells

In the bone marrow, B cells develop from haematopoietic stem cells. The master regulator of commitment to the B cell fate is the transcription factor paired box 5 (PAX5)179. Expression of

PAX5 is consistent throughout all stages of B cell development and preservation of B cell identity is dependent on continued expression, as removal of PAX5 from mature B cells leads to dedifferentiation back to a progenitor stage180,181. PAX5 binds thousands of genomic sites within

B cells, and by acting as either a transcriptional activator or repressor, directs a broad transcriptional program182-184. PAX5 activates genes involved in BCR signaling (such as the IgH locus and Igα), adhesion and migration (such as CD44), immune receptors (such as CD19,

CD21, CD40) and important B cell transcriptional regulators (such as IRF4, interferon regulatory factor 8 (IRF8), BTB and CNC homology 2 (BACH2), Aiolos and SPIB). In addition, PAX5 represses a range of lineage inappropriate genes that are associated with driving a T cell or myeloid fate185. PAX5 also represses the expression of several PC regulators including X-box binding protein 1 (XBP1), BLIMP1 and joining (J) antibody chain185-188 (Fig. 1.4), and PC differentiation is associated with a loss of B cell identity via the inhibition of PAX5185,187,189. It was believed that repression of PAX5 in PCs was induced by BLIMP1, a critical regulator of PC development190. However, an early “pre-plasmablast” stage, characterized by PAX5 inhibition,

20 can develop in the absence of BLIMP1191. These pre-plasmablasts secrete low amounts of antibody likely due to the derepression of the genes encoding XBP1 (a factor that drives the PC secretory phenotype) and J chain191. Therefore, it appears that the repression of PAX5 is the earliest event in PC differentiation. Subsequent expression of essential regulators of PC development, BLIMP1 and IRF4, then drive the development of mature PCs, that secrete large amounts of antibody, and PAX5 repression is reinforced by BLIMP1 (Fig. 1.4). The molecular events that initiate PAX5 repression remain to be determined.

In addition to PAX5, several other B cell factors also repress drivers of PC development.

For example, in addition to establishing the DZ phenotype, BCL6 also guards against premature

PC differentiation by repressing the transcription of PRDM1 (the gene encoding BLIMP1)148,149

(Fig. 1.4). In addition, BACH2 cooperates with BCL6 to promote the repression of PRDM1192,193 and in naïve B cells, microphthalmia-associated transcription factor (MITF) represses irf4194.

Molecular drivers of the PC fate

Regulators of B cell and GC B cell identity repress drivers of PC development, and in turn regulators of PC development repress both B cell and GC B cell genetic programs to reinforce the PC fate195 (Fig. 1.4). The function of the major regulators of these processes is discussed below.

IRF4

IRF4 has diverse, context-dependent functions in GC B cell and PC development196,197.

Together with various co-factors, IRF4 has roles in promoting very early GC initiation198,199, CSR and PC differentiation200,201, and in the T cell lineage IRF4 regulates the differentiation of Tfh cells202. Its function appears to be dose-dependent as low levels are associated with GC formation and CSR, whereas high, sustained levels are associated PC development201. The role of IRF4 in GC formation appears to be confined to a very early initiation stage when pre-GC

B cells still reside in the outer follicle, as GCs develop normally in the absence of IRF4 from day

21

2 of the GC reaction onwards200. At this pre-GC stage, IRF4 induces the expression of BCL6 and AID198. In GC B cell and PC development, IRF4 has a biphasic activation pattern during which it is required for early GC initiation, before it is silenced throughout the majority of GC B cell development, and is then reactivated in late GC differentiation. High expression of IRF4 is essential for PC differentiation201 and may direct PC differentiation either in parallel with200 or upstream201 of another essential regulator, BLIMP1. High expression of IRF4 also results in inhibition of BCL6, thus extinguishing the GC B program, while also indirectly inducing BLIMP1 expression by alleviating BCL6-mediated repression of PRDM1201,203 (Fig. 1.4).

BLIMP1

BLIMP1 is a transcriptional repressor, and within the B cell lineage BLIMP1 is uniquely expressed in PCs, with PCs expressing higher levels than plasmablasts204-206. While BLIMP1 is not required for initiation of PC development191, it is essential for the generation of mature

PCs207, and the expression of exogenous BLIMP1 can induce plasmablast differentiation208.

Continuous expression of BLIMP1 is also required for the maintenance of long-lived PCs209.

BLIMP1 acts upstream of another important PC factor, XBP1210 (see below), and represses the genes encoding B cell and GC B cell factors PAX5, c-MYC and BCL6190,211-213 (Fig. 1.4). In response to activation, B1 and MZ B cells upregulate BLIMP1 more rapidly than follicular B cells; this kinetic difference may underlie the swift differentiation of these cell types into plasmablasts upon activation49,214.

XBP1

PCs are essentially factories of antibody production and secretion, creating conditions that induce endoplasmic reticulum (ER) stress. When a cell experiences ER stress caused by unfolded proteins, it induces the unfolded protein response (UPR) in order to tolerate these conditions. The key mediator of this response is the transcription factor XBP1215. While PCs form in the absence of XBP1, these cells secrete reduced amounts of antibody, and fail to

22 undergo the ER remodeling that creates the characteristic morphology of PCs210,216-219. The expression of XBP1 may be activated by IRF4200 and BLIMP1210.

Long term survival of PCs

Long-lived PCs primarily reside within the bone marrow. Within this niche, their survival is maintained by several trophic signals220 and PCs are dependent on these non-intrinsic trophic factors for survival. Only a subset of plasmablasts generated in secondary lymphoid organs will travel via the blood to the bone marrow to become long-lived PCs221. The expression of the S1P

221 receptor S1P1 is required for the migration into the blood , after which expression of CXCR4 together with other chemokine receptors promotes homing to and retention within the bone marrow environment172,222-225. This movement is also associated with reduced expression of

CXCR5 and CCR7, allowing PCs to leave the B cell follicles and T cell areas172. Retention within the bone marrow environment involves the interactions of several integrins and adhesion molecules226,227, including VLA4-mediated interactions with VCAM1-expressing epithelial cells228. BLIMP1 at least in part drives these events as it has been shown to repress CXCR5211, while inducing CXCR4 and VLA4 expression229. The cell types that make up the bone marrow

PC niche remain somewhat unclear226, but include a CXCL12-producing stromal cell type222 and eosinophils that colocalize with bone marrow PCs230. PCs express high levels of the receptor B cell maturation protein (BCMA)231 in a BLIMP1-independent manner232, and low levels of the

BAFF-R relative to follicular B cells81,231. The bone marrow niche produces the B cell survival factors a proliferation-inducing ligand (APRIL) and BAFF which are critical for the survival of bone marrow PCs231,233,234. BAFF and APRIL are both able to bind and signal through BCMA, and the survival of long-lived PCs is impaired in the absence of BCMA or following sequestration of APRIL or BAFF231,232,235. Signaling through BCMA induces high levels of the anti-apoptotic factor MCL1, and PCs are depleted in vivo upon removal of MCL1232. Several other important factors are produced within the bone marrow niche that support PC survival227,

23 including IL-6236,237. Signals within the bone marrow niche also promote survival via the activation of the NF-B signaling cascade, which will be discussed in Section 1.5.

Section 1.5: THE GC AND LYMPHOMAGENESIS

The mutational processes occurring within the GC drive the generation of a high affinity immune response that is critical for effective pathogen clearance. However, the advantage of these Ig remodeling processes for immunity is counterbalanced by the threat of malignant transformation when these events, that induce DNA strand breaks, are dysregulated and create genomic aberrations. Indeed, the vast majority of mature B cell lymphomas are derived from GC B cells at various stages of GC development238,239. Genomic sequencing of human lymphomas has revealed that most cases develop as a result of genetic lesions involving loci encoding key regulators of normal GC development240-245. Therefore, genetic programs important for normal

GC differentiation are frequently hijacked during lymphomagenesis. Understanding the molecular control of the GC reaction therefore provides the basis for the development of therapies aimed at the treatment of lymphoma. In turn, recurrent mutations identified in lymphoma may provide clues that identify novel pathways that function within the GC, and this has been the case for many factors now recognized as critical for normal GC development.

Lymphoma classification and recurrent genetic lesions are discussed in this section, with particular focus on mutations in NF-B signaling components.

Dysregulation of Ig remodeling mechanisms

B cell non-Hodgkin lymphomas (B-NHLs) comprise several different lymphoma subtypes with distinct genetic and phenotypic composition and heterogenous prognosis. The majority of these lymphomas are of GC origin, as evidenced by the presence of somatically mutated Ig genes238,246. A major type of genetic alteration observed in B-NHLs are chromosomal translocations and these translocations frequently involve the Ig loci. The analysis of translocation breakpoints has revealed that chromosomal translocations can be caused by

24 errors in SHM or CSR. These processes normally induce DNA lesions in Ig switch regions or in regions targeted by SHM and can aberrantly induce these lesions in non-Ig loci247-249. These errors can result in the juxtaposition of the coding region of a proto-oncogene with Ig regulatory elements, leading to dysregulated expression of the corresponding gene. A second major type of genetic lesion observed in B-NHLs is aberrant SHM (ASHM). Under normal conditions the process of SHM acts almost exclusively on the variable regions of the Ig loci; during ASHM, however, several non-Ig loci are targeted250,251. These mutations can target 5’ regulatory regions and coding regions leading to nucleotide and amino acid exchanges, respectively250. A direct link between the Ig DNA remodeling events operating within the GC and lymphomagenesis has been provided via the demonstration that AID, the enzyme that directs SHM and CSR, is required for the development of GC-derived B-NHL in mouse models of B cell lymphoma252,253.

B-NHL classification

Comparison of the transcriptional profiles of normal GC B cell types and lymphoma subtypes has revealed that different B-NHLs subtypes can often be traced back to specific cells of origin within the GC, suggesting that these lymphomas are stalled at particular phases of GC development124,254,255. In addition, while there are several commonalities, different subtypes of

B-NHL are often characterized by the recurrence of characteristic genetic lesions that frequently involve known regulators of the GC reaction. A selection of representative examples are discussed below.

Burkitt lymphoma

The cell of origin in Burkitt lymphoma is believed to be a DZ B cell124,254, and virtually all cases are associated with translocations between the Ig and MYC loci256. MYC was the first gene associated with GC-derived lymphomagenesis257,258 and encodes the transcription factor c-

MYC. c-MYC is an important example of a critical regulator of the GC whose role in lymphomagenesis was appreciated long before a physiological function within the GC could be

25 identified. Only recently, as outlined in Section 1.3, have important roles for c-MYC been defined specifically in early GC initiation and during T cell-mediated selection; in all other phases of the GC reaction, the expression of c-MYC is tightly controlled by BCL6-mediated repression162,163. Oncogenic activation of c-MYC expression allows escape from BCL6-mediated repression and is likely to promote lymphomagenesis by driving inappropriate cell growth and also via the induction of genomic instability, which may facilitate the acquisition of additional genetic aberrations160,259,260.

Diffuse large B cell lymphoma (DLBCL)

DLBCL is the most common B-NHL, comprising ~40% of adult B-NHLs256. Transcriptional analysis of DLBCL has further segregated this type of lymphoma into two distinct forms known as GC B cell-like DLBCL (GCB-DLBCL) and activated B cell-like DLBCL (ABC-DLBCL). GCB-

DLBCL is characterized by the expression of GC B cell genes, whereas genes associated with the in vitro activation of B cells are expressed in ABC-DLBCL255. Recently, additional work further defined this distinction and it is now believed that the cell of origin in GCB-DLBCL is a LZ

B cell, whereas ABC-DLBCL appears to derive from a cellular stage at which a LZ B cell has initiated post-GC plasmablastic differentiation124,255. These distinctions have important clinical implications, as GCB-DLBCL is associated with a more favorable prognosis than ABC-

DLBCL255,261. One of the most common oncogenic events observed in DLBCLs is dysregulation of BCL6 expression which affects ~30% of cases256,262 and can occur via several mechanisms.

Translocations involving BCL6 frequently occur263,264, and mice engineered to express an Ig-

BCL6 translocation develop a lymphoproliferative disease that closely mirrors human DLBCL265.

In addition, binding sites within the BCL6 promoter that normally mediate repression of BCL6 transcription are frequently rendered non-functional by mutations in DLBCL203,266. These genetic aberrations collectively lead to constitutive BCL6 expression and thus prevent the silencing of

BCL6 that is necessary for post-GC differentiation, as described in Section 1.4. These types of

26 aberrations are therefore thought to cause a block in development and consequently lock GC B cells into a DZ B cell phenotype characterized by rapid proliferation and tolerance of DNA damage, which in turn create opportunities for further, potentially transforming, genetic hits267.

Distinctive genetic lesions are also observed in GCB-DLBCL vs. ABC-DLBCL256,262.

GCB-DLBCL is associated with translocations involving the MYC and BCL2 loci and mutations in the genes encoding polycomb methyltranferase enhancer of zeste homolog 2 (EZH2)241,268-270

271 106,272 and regulators of GC confinement , including the chemokine receptor S1P2 . Genetic lesions resulting in the inhibition of the PC regulator BLIMP1 are observed in ~30% of ABC-

DLBCL cases256,262. Bi-allelic inactivation can occur via a variety of mechanisms including deletions, mutations and epigenetic silencing273-275. These lesions are likely to contribute to lymphomagenesis by creating a block in PC differentiation, thereby trapping LZ B cells in a state of early plasmablastic differentiation. Indeed, conditional removal of BLIMP1 in B cells promotes the generation of a lymphoproliferative disorder with several phenotypic similarities to human

ABC-DLBCL275. Finally, in the last several years, a range of mutations that converge on activation of the NF-B signaling cascade were identified in ~95% of ABC-DLBCL cases256,262, providing an explanation for the predominant expression of NF-B target genes observed in

ABC-DLBCL vs. GCB-DLBCL that had been noted in early studies276. Constitutive activation of

NF-B signaling is therefore a hallmark of ABC-DLBCL. The normal function of NF-B and the consequences of lymphoma-associated mutations are discussed below.

The NF-B signaling cascade

The NF-B signaling transduction pathway can be divided into two major branches, a canonical and an alternative pathway that are activated through specific cell surface receptors277,278 (Fig.

1.6). The canonical NF-B pathway has been intensively studied over the past 29 years, with critical roles defined in both innate and adaptive immune cell types. The range of functions ascribed to the canonical NF-B pathway can be inferred from the diverse range of signals

27 known to activate this pathway. These include exposure to pathogen-associated molecular patterns (PAMPs) that bind TLRs, pro-inflammatory cytokines such as tumor necrosis factor

(TNF) that binds the TNF-receptor (TNF-R) and IL-1 which binds the IL-1-receptor (IL1-R).

Canonical NF-B activation also occurs following T cell-receptor (TCR) and BCR antigen stimulation. In response to signaling through this array of immune receptors, canonical NF-B activation induces the expression of a range of host defense genes including cytokines, growth factors and other effector molecules that augment the immune response. In addition, canonical

NF-B signaling also induces the expression of genes with pro-proliferative and anti-apoptotic functions, promotes effector cell activation and regulates the development of various lymphocyte subpopulations279. In contrast, the alternative NF-B pathway is associated with a specific set of functions280. This pathway was discovered more recently than the canonical pathway with the first reports defining the alternative pathway appearing in the early 2000s. The alternative NF-B pathway is activated by signals through a limited number of receptors including the LT β-receptor (LTβR)281,282, CD40283,284, receptor activator of NF-B (RANK)285 and the BAFF-R286-288, all of which to varying degrees also have some capacity to activate the canonical pathway. CD40 stimulation in particular is capable of strongly activating both NF-B pathways283,284,289,290. Alternative NF-B pathway activation in response to signaling through each of these receptors leads to heterogeneous outcomes280. LTβR signaling promotes the organization of lymphoid architecture by stromal cells, while CD40 signaling induces T cell- mediated activation of B and T cells, DCs and monocytes. Signaling through RANK has important roles in osteoclast development, and BAFF-R signaling is required for the survival of mature B cells. The role of alternative NF-B signaling in mature B cell survival will be further discussed in the introduction to Chapter 2 (Section 2.1). As outlined below, under physiological conditions, activation of the NF-B cascade occurs only in response to inducing signals and is subject to tight regulation. Therefore, under normal circumstances, NF-B activation is transient,

28 and the constitutive activation that occurs as a result of genetic alternations in lymphoma is an aberrant, pathogenic event.

The canonical NF-B pathway

The downstream mediators of the canonical NF-B pathway are the transcription factors reticuloendotheliosis oncogene (c-REL), v-rel reticuloendotheliosis viral oncogene homolog A

(RELA) and p50 (which is derived from constitutive, co-translational processing of the precursor, p105, also known as NF-B1) that mainly occur as heterodimers277,278 (Fig. 1.6). In B cells, translocation of canonical NF-B heterodimers from the cytoplasm to the nucleus occurs in response to BCR, TLR and CD40 stimulation291,292. These signals are relayed through a series of kinases and adaptor proteins that differ depending on the initiating signal but culminate in the activation of the inhibitory BIB) kinase β subunit (IKKβ)277,278. IKKβ is present in a complex with another catalytic subunit, IKKα, and the regulatory subunit IKKγ, together these three subunits from the IKK complex. Activated IKKβ phosphorylates inhibitory IBmolecules (IBα,

IBβ or IBε), an event which marks these subunits for polyubiquitylation and subsequent proteasomal degradation (Fig. 1.6). IB degradation represents the rate limiting step in the activation of the canonical NF-B pathway277,278. Under steady state, unstimulated conditions,

IBs hold the canonical heterodimers in an inactive form in the cytoplasm by masking nuclear localization sequences. Following proteasomal degradation of IBs, homo- or heterodimers of c-

REL, RELA and p50 (usually c-REL-p50 and RELA-p50) are free to enter the nucleus and bind to B binding sites in the regulatory regions of target genes. c-REL and RELA contain transactivation domains (TADs) that enable these subunits to drive the transcription of target genes277,278. IBα can also serve to inhibit DNA binding and promote the export of NF-B dimers from the nucleus. Transcription of the gene encoding IBα is a frequent event following

29 canonical NF-B activation, providing a mechanism by which NF-B signaling can be rapidly extinguished in order to restore steady-state conditions293,294.

The alternative NF-B pathway

The corresponding downstream mediators of the alternative NF-B pathway are avian reticuloendotheliosis viral (v-rel) oncogene related B (RELB) and p52 (which is generated by proteolytic cleavage of the precursor p100, also known as NF-B2) and these subunits mainly occur as heterodimers280 (Fig. 1.6). In B cells, alternative NF-B signaling is activated by soluble

BAFF or CD40 stimulation283,286. These signals activate the alternative NF-B pathway via the release of NF-B-inducing kinase (NIK) from continuous turnover mediated by a complex comprised of TNF receptor-associated factor 2 and 3 (TRAF2 and TRAF3) and baculoviral IAP repeat-containing 2 and 3 (cIAP1/2), allowing its stabilization295-302. NIK activates IKKα homodimers, which phosphorylate and subsequently induce proteasomal processing of p100 that is bound to RELB303,304. This results in the removal of an inhibitory C-terminal ankyrin domain of p100, resulting in both the generation of p52 and the release of RELB from p100- mediated inhibition and retention within the cytoplasm303,304 (Fig. 1.6). The processing of p100 to p52 therefore represents the rate limiting step in the activation of the alternative NF-B pathway and allows the nuclear translocation of RELB/p52 heterodimers280. RELB, but not p52, contains a TAD and is thus capable of activating transcription280.

NF-B mutations in ABC-DLBCL

Activating mutations in canonical NF-B signaling components

Mutations affecting NF-B signaling components observed in ABC-DLBCL predominantly converge on constitutive activation of the canonical NF-B pathway (Fig. 1.7). BCR signaling activates the canonical NF-B pathway, and chronic BCR signaling is observed in ABC-

DLBCL305. Mutations in the genes encoding the intracellular components of the BCR signaling

30 complex, CD79A and CD79B, prevent receptor internalization and inhibit Yamaguchi sarcoma viral (v-yes-1) oncogene homolog (LYN), a negative regular of BCR signaling305. Mutations in the gene encoding coiled-coil domain of the scaffolding protein caspase recruitment domain family, member 11 (CARD11), which is a downstream signaling component of the BCR present in a complex with mucosa-associated lymphoid tissue lymphoma translocation gene 1 (MALT1) and B cell lymphoma 10 (BCL10), lead to constitutive oligomerization and canonical NF-B activation independent of antigen stimulation306-308. Mutations in the gene encoding myeloid differentiation primary response gene 88 (MYD88), an adaptor protein that mediates signaling downstream of TLRs and the IL-1 receptor, were also observed309. Mutant MYD88 complexes with the adaptors proteins interleukin-1 receptor-associated kinase 1 and 4 (IRAK1 and IRAK4) in a spontaneous manner, thereby leading to constitutive downstream pathway activation309,310.

Mutant MYD88 promotes survival via the constitutive activation of several signaling pathways in addition to the canonical NF-B pathway, including the Janus kinase (JAK), signal transducer and activator of transcription 3 (STAT3) and type I interferon (IFN) pathways309. Finally, tumor necrosis factor, alpha-induced protein 3 (TNFAIP3, also known as A20), a negative regulator of the canonical NF-B signaling cascade, is frequently subjected to bi-allelic inactivation311,312.

Functional validation of addiction to constitutive canonical NF-B activation

The addiction of ABC-DLBCL to constitutive canonical NF-B signaling has been directly demonstrated using both in vitro and in vivo approaches. Suppression of canonical NF-B activation via the introduction of a super-repressor form of IBα or a dominant negative form of

IKKβ leads to cell death and growth arrest in ABC-DLBCL but not GCB-DLBCL cell lines276. The oncogenic function of mutant CARD11 was demonstrated using complementation assays in which ABC-DLBCL lines expressing CARD11 mutants could be rescued from cell death following CARD11 knockdown via the introduction of mutant CARD11, but not wild-type

CARD11308. In addition, the reintroduction of A20 into ABC-DLBCL cell lines with bi-allelic

31 inactivation of TNFAIP3 induced apoptosis and cell growth arrest. In vivo, tumorigenicity was also observed, demonstrating a tumor suppressor role for A20311,312. Finally, constitutive activation of NF-B signaling cooperated with BLIMP1 inactivation in mouse GC B cells to drive the development of a lymphoma closely resembling human ABC-DLBCL313.

Activating mutations in alternative NF-B signaling components

Mutations that activate the NF-B signaling cascade in ABC-DLBCL do so predominantly via the canonical pathway. However, mutations leading to activation of the alternative NF-B pathway were also observed in DLBCL with a frequency of ~15%242,314,315. Nuclear p52 was however observed in ~30% of cases, suggesting a more expansive role for this pathway311. These genetic lesions occurred in DLBCL irrespective of the GCB or ABC subtype, and were either mutations or mono- or bi-allelic deletions of TRAF3 or TRAF2242,314 that abrogated the ability of these factors to degrade NIK, resulting in constitutive alternative NF-B pathway activation (Fig.

1.7). Finally, NIK overexpression in GC B cells cooperated with deregulated BCL6 to promote the development of DLBCL in mice, thereby demonstrating an oncogenic role for the alternative

NF-B pathway in this disease315.

NF-B mutations in multiple myeloma (MM)

MM is a post-GC malignancy of PCs during which a monoclonal PC population accumulates in the bone marrow316. The disease has several subtypes defined by particular genetic and cytogenetic alterations; however, it is believed that dysregulation of Cyclin D (Cyclin D1, D2 or

D3) may be a unifying event317. Multiple mutations in NF-B signaling components have also been identified in this disease with a frequency of ~17%318-320. However, in contrast to ABC-

DLBCL, most of these mutations affected signaling components of the alternative NF-B pathway318-320 (Fig. 1.7).

Activating mutations in alternative NF-B signaling components

32

Several mutations that lead to the overexpression of NIK were observed in MM318,320 (Fig. 1.7).

Overexpression of NIK was achieved both directly via amplification or translocations and indirectly via mechanisms that lead to protein stabilization318,320. Indirect mechanisms involved mutations in the region of NIK targeted by TRAF3 for NIK destabilization. Inactivation of several negative regulators of NIK protein stability were also observed, including bi-allelic deletion, truncating mutations and epigenetic silencing of TRAF3 and TRAF2 and BIRC2/BIRC3 (genes encoding the ubiquitin ligases c-IAP2/3, which promote NIK degradation)318,320. Copy number amplifications of CD40, LTBR and TACI (another receptor which can activate NF-B signaling) and homozygous deletion of cylindromatosis (CYLD), a negative regulator of NF-B signaling, were also observed318,320. Similar to the canonical pathway mutations observed in B-NHL, most of these aberrations therefore affected upstream components of the alternative NF-B pathway, with a few exceptions such as truncating mutations of NFKB2 (the gene encoding p100)318,320.

These NFKB2 mutations disrupt the C-terminal ankryin repeats that serve as an inhibitory domain that maintains cytoplasmic localization. Similar mutations have previously been reported for MM and other lymphoid malignancies321,322. Finally, amplification of NFKB1 (the gene encoding p105, a canonical pathway component) was also observed318,320.

Cross-talk between the canonical and alternative NF-B pathways

While the majority of these mutations in MM affect upstream components of the alternative NF-

B pathway, analysis of MM cell lines revealed that these mutations were found to constitutively activate both the alternative and canonical pathways in the majority of cases318,319,323. In addition, toxicity was induced in many of these cell lines following inhibition of the canonical NF-

B pathway via treatment with a specific inhibitor of IKKβ318. Knockdown of NIK in NIK- overexpressing cell lines induced toxicity and reduced both canonical and alternative pathway activation and induction of TRAF3 in TRAF3-deficient lines had similar effects318,319,323. Indeed, the toxicity induced by NIK inhibition may be dependent on the canonical pathway, since it was

33 observed that knockdown of IKKα, downstream of NIK in the alternative pathway, in NIK- overexpressing lines did not affect viability318.

This dual activation is likely a result of cross-talk between the two pathways. For example, in certain cellular contexts, NIK has been shown to activate the canonical pathway in response to particular stimuli, including CD40L and BAFF290 (Fig. 1.7). In addition, overexpression of NIK is associated with canonical pathway activation via activation of the IKK complex324-327. It is possible therefore that elevated NIK may promote MM survival primarily via activation of the canonical pathway318. However, as nuclear localization of alternative subunits was also observed in these cases, it is likely that both pathways are utilized318,319,323.

Role of the bone marrow niche in MM

The findings described above collectively demonstrate that constitutive NF-B activation has an important role in MM pathogenesis. A unique aspect of MM, in comparison to B-NHLs, is dependence on the bone marrow microenvironment. As outlined in Section 1.4, PCs rely on signals from bone marrow niches for long term survival. These signals include BAFF and

APRIL, related molecules that can activate the NF-B signaling cascade by binding to the

BCMA receptor expressed on PCs231,232,234,235. Normal PCs are therefore characterized by activated NF-B signaling and the magnitude of this signaling may be the highest observed among B cell subpopulations318. MM cells also reside within the bone marrow environment and rely on bone marrow trophic signals for survival328-332. It has been suggested that mutations causing constitutive NF-B activity may allow MM cells to escape the survival limitations imposed by the bone marrow environment, thus leading to growth that is independent of bone marrow factors333,334.

NF-B mutations in other B cell malignancies

34

In addition to the mutations in NF-B signaling components observed in DLBCL and MM, other types of B cell malignancies also display similar aberrations. These B cell malignancies include

Hodgkin lymphoma, splenic marginal zone lymphoma, mantle cell lymphoma, mediastinal large

B cell lymphoma and chronic lymphocytic leukemia (reviewed in ref.335). Mutations activating the canonical pathway and/or the alternative NF-B pathway were identified in most cases. The acquisition of activating mutations in NF-B signaling components is therefore common among

B cell malignancies. These activating mutations appear to be a rare feature as constitutive NF-

B signaling is commonly achieved via infection and inflammatory environmental stimuli in other contexts of cancer336,337.

Function of NF-B activation in GC and B cell development?

The selection of mutations converging on NF-B activation in various GC-derived lymphomas suggests that the NF-B signaling cascade may have an important biological role during normal

GC development. Understanding the biological function of NF-B may therefore provide insights into the oncogenic function of this signaling cascade. Cross-talk between the canonical and alternative NF-B pathways necessitates analysis of the downstream transcription factors of the

NF-B cascade in order to assign specific functions to either the canonical or alternative NF-B pathway. As such, the role of the canonical subunits c-REL and RELA in the GC reaction was the subject of a project within the lab that was recently published338 (discussed in Section 4.1).

The importance of the subunits of the alternative NF-B pathway, RELB and p52, during GC development however remained an open question and is the focus of this thesis in Chapters 3 and 4. In contrast, the alternative NF-B pathway has been extensively studied within the context of mature B cell survival. However, specific questions remained, regarding B cell intrinsic functions and the events that occur downstream of BAFF-mediated activation of the alternative NF-B pathway, and these issues are addressed in Chapter 2.

35

Figure 1.1. Stages of mature B cell development. Immature B cells expressing a functional B cell receptor (BCR) complete development within the bone marrow and progress through two transitional stages (T1 and T2) prior to becoming a mature B cell. For the majority of immature B cells, these transitional stages occur in the spleen. T2 B cells mature into either follicular B cells

(also known as mature B cells) or marginal zone (MZ) B cells. Mature B cells activated via antigen (red) in the presence of T cell help enter the germinal center (GC) reaction, GC B cells then differentiate into either plasma cells (PCs) or memory B cells.

36

37

Figure 1.2. Structure of the spleen. Blood is supplied to the spleen by the splenic artery which branches into arterioles (not depicted). These arterioles penetrate a vessel called the marginal sinus. From the marginal sinus, blood filters into the marginal zone (MZ) prior to reentering the general circulation. The white pulp of the spleen is lymphoid tissue comprised of B cell follicles that are bordered by T cell areas. B cells travelling in the blood migrate into the B cell follicles from the MZ. This splenic structure is shown schematically (top) and via immunohistochemical analysis of mouse spleen for the expression of IgM (bottom, image provided by Nicole Heise).

Germinal centers (GCs) form within B cell follicles and are marked by BCL6 expression

(bottom).

38

39

Figure 1.3. Dynamics of the GC reaction and selection of high-affinity antibody mutants.

Antigen-activated germinal center (GC) precursor B cells form the early GC at day 4, during which they differentiate into blasts that undergo clonal expansion over the following days until the mature GC, that is characterized by a dark zone (DZ) and light zone (LZ), forms at day 7.

During proliferation, the process of somatic hypermutation (SHM) introduces base-pair changes into the V(D)J region of the rearranged immunoglobulin variable region (IgV) genes (red lines); some of these base-pair mutations lead to a change in the amino acid sequence. DZ B cells then move to the LZ, where the modified B cell receptor (BCR) of the LZ B cell, with help from immune cells, including T follicular helper (Tfh) cells and follicular dendritic cells (FDCs), is selected for improved binding to the immunizing antigen. Among the newly generated LZ B cells that express BCR mutants resulting from SHM in the DZ, higher BCR affinity is directly associated with greater antigen capture and leads to a higher density of peptide–major histocompatibility (MHC) complexes presented on the surface of the B cell. This results in the greatest share of T cell help, in the form of CD40 stimulation and other signals, which in turn drives positive selection. Newly generated LZ B cells that produce an unfavourable antibody are rendered unable to capture sufficient antigen and undergo apoptosis. Following positive selection, a subset of LZ B cells is instructed to recirculate to the DZ. LZ B cells may undergo Ig class switch recombination (CSR) before LZ–DZ recirculation, whereas other cells switch and directly differentiate. Back in the DZ, these cells undergo further proliferation and SHM, thus potentially generating antibody mutants with an improved affinity. Recirculation between the DZ and the LZ facilitates several iterative rounds of mutation and selection, and within a short time, leads to the generation of high-affinity memory B cells and plasma cells (PCs). Antigen-selected

LZ B cells eventually differentiate into memory B cell precursor cells and plasmablasts, which are the precursors of memory B cells and PCs. TCR, T cell receptor. Figure adapted from ref.139.

40

41

Figure 1.4. Mutually repressive genetic programs govern the B cell/GC B cell vs. PC fate.

In B cells and/or germinal center (GC) B cells, the transcription factors PAX5, BACH2 and BCL6 inhibit plasma cell (PC) development by repressing the transcription of genes encoding the PC regulator BLIMP1. PC development is further inhibited by BCL6-mediated repression of the gene encoding IRF4, PAX5-mediated repression of the genes encoding J chain and XBP1 and repression of the gene encoding IRF4 by MITF. In PCs, BLIMP1 represses the genes encoding the B cell/GC B cell factors PAX5, c-MYC and BCL6. In addition, IRF4 represses the gene encoding BCL6. Repression of the PC genetic program in B cells and GC B cells prevents premature PC differentiation and repression of the B cell and GC B cell genetic program in PCs reinforces the PC fate.

42

43

Figure 1.5. The NF-B signaling cascade. The NF-B signaling cascade is comprised of two pathways, the canonical and alternative NF-B pathways. Signaling through the T cell receptor

(TCR), B cell receptor (BCR), Toll-like receptors (TLRs), interleukin 1-receptor (IL1-R) and tumor necrosis factor-receptors (TNF-Rs) lead to activation of the canonical NF-B pathway.

Signaling through these receptors culminates in the activation of IKKβ. IKKβ is present in a complex with another catalytic subunit, IKKα, and the regulatory subunit IKKγ. Activated IKKβ phosphorylates inhibitory IBmolecules, marking these subunits for proteasomal degradation.

Under unstimulated conditions, these IBs hold the canonical heterodimers in an inactive form in the cytoplasm. Following proteasomal degradation of IBs, homo- or heterodimers of the transcription factor subunits c-REL, RELA and p50 (usually c-REL-50 and RELA-p50) are free to enter the nucleus and bind to B binding sites in the regulatory regions of target genes to drive transcription. Activation of the alternative NF-B pathway is induced by a more limited set of signals, including the lymphotoxin β-receptor (LTβR) and B cell activating factor-receptor

(BAFF-R). CD40 stimulation is capable of activating both the canonical and alternative NF-B pathways. Signaling through these receptors releases NIK from continuous turnover mediated by a complex comprised of TRAF2, TRAF3 and cIAP1/2. NIK activates IKKα homodimers, which phosphorylate and subsequently induce proteasomal processing of p100 which is bound to the transcription factor subunit, RELB. This results in the generation of the transcription factor subunit p52 and the release of RELB from p100-mediated inhibition and retention within the cytoplasm. RELB-p52 heterodimers then enter the nucleus and activate transcription of target genes. In the text p100/p52 is also referred to as NF-B2.

44

45

Figure 1.6. Genetic aberrations in multiple components of the NF-B signaling cascade are observed in B cell malignancies. Mutations in genes encoding multiple components of the

NF-B signaling cascade are observed in various B cell malignancies. These mutations lead to constitutive activation of the canonical and/or alternative NF-B pathway. Activating mutations are marked with a red asterisk. Inactivating mutations, including bi-allelic mutations and/or deletions, are marked with a blue asterisk. Figure adapted from ref.335. The publications originally describing the genetic alterations are cited in the text.

46

47

This page has been intentionally left blank.

48

CHAPTER 2:

Impairment of mature B cell maintenance upon combined ablation of the

alternative NF-B transcription factors RELB and NF-B2 in B cells

49

Section 2.1: INTRODUCTION

As mentioned in Section 1.2, it is clear that the maintenance of mature B cells is dependent on tonic BCR signaling61,62 and signaling induced by binding of BAFF to the BAFF-R64-73. The downstream events that orchestrate this dependence are however convoluted and have been the subject of intense investigation.

As BCR stimulation via cognate antigen during B cell activation is known to activate canonical NF-B signaling, it was believed that the canonical NF-B pathway was also activated within the context of tonic BCR signaling. A role for the canonical pathway in tonic signaling was further supported by the observation that mature B cells were greatly reduced upon B cell- specific ablation of the canonical pathway components IKKγ and IKKβ339,340. However, removal of downstream components of the BCR pathway that are necessary for canonical NF-B activation, CARD11, MALT1 and BCL10, had minimal impact on the resting peripheral B cell pool306,341-345. Furthermore, BCR signaling is capable of activating many pathways346, including the phosphoinositide 3-kinase (PI3K) pathway, and it was demonstrated that BCR-deficient mature B cells could be completely rescued by PI3K activation in vivo347. In contrast, minimal rescue was observed following activation of canonical NF-B signaling using a constitutively active IKKβ347. These results suggest that B cell survival mediated by tonic BCR signaling does not critically involve signaling through the canonical NF-B pathway. Tonic BCR signaling is therefore functionally distinct from BCR signaling induced by cognate antigen interactions. This is further supported by the inability of mitogen-activated protein kinase (MAPK) activation, another pathway involved in antigen-dependent B cell activation, to rescue BCR deficiency347.

Genetic ablation of BAFF or the BAFF-R results in a dramatic reduction in peripheral B cells due to maturation arrest at the T1 developmental stage348. In terms of NF-B activation,

BAFF-BAFF-R signaling is predominantly associated with activation of the alternative NF-B pathway286-288. However, the phenotypes of mice deficient for the alternative NF-B signaling

50 components IKKα303,349 or NIK350 did not fully recapitulate the severe B cell defects observed in

BAFF-R deficient mice. In these mice, B cell numbers were reduced, although not to the dramatic extent observed in the absence of BAFF survival signals. This indicates that BAFF also mediates B cell survival via pathways other than the alternative NF-B pathway. BAFF has been reported to activate several downstream signaling routes including protein kinase B (PKB or AKT), PI3K, extracellular signal-regulated kinase (ERK) and protein kinase Cβ (PKCβ) signaling axes which sustain B cell survival and metabolic fitness and may be a means of crosstalk with tonic BCR signaling83-87. Indeed, sustained activation of the PI3K pathway is able to partially restore B cell maturation in the absence of BAFF83. In regards to NF-B activation, it has been reported that activation of the canonical pathway may also contribute to BAFF- mediated pro-survival functions351,352. Most strikingly, BAFF-R deficiency can be fully rescued by constitutive canonical NF-B activation353. It appears therefore that the canonical pathway can replace signals derived from BAFF-mediated activation of the alternative NF-B pathway as well as pathways. These findings suggest that BAFF-mediated activation of the canonical NF-B pathway may explain in part the importance of canonical NF-B signaling to mature B cell maintenance339,340.

The body of literature described above highlights the involvement of several signaling routes in maintaining the survival of mature B cells in the periphery. These signaling routes appear to direct overlapping functions, suggesting that mechanisms driving mature B cell survival are highly reinforced. In order to determine the function of alternative NF-B signaling in

BAFF-mediated B cell survival, many studies have focused on analyzing upstream components of the alternative pathway and have confirmed that this pathway has a critical role in maintaining

B cell homeostasis. Constitutional deletion of these upstream components produces dramatic phenotypes often associated with severe lymphoid architectural defects. These defects are likely due to disrupted activation of the alternative NF-B pathway by LTβ in non-hematopoietic

51 cell types354. Due to these severe phenotypes, described below are studies in which radiation chimeras were utilized to investigate the function of upstream components of the alternative

NF-B pathway in hematopoietic cells, or studies in which these components were conditionally deleted in B cells. First, adoptive bone marrow transfers involving alymphoplasia

(aly) mice, which harbor a defective NIK, resulted in a marked reduction in mature B cells350.

Second, conditional deletion of TRAF2 or TRAF3 and double deletion of cIAP1 and cIAP2, which are negative regulators of NIK, led to peripheral B cell hyperplasia and conferred independence from BAFF survival signals355-357. Third, fetal liver cell transfer studies reported that inactivation of IKKα was associated with a marked decrease in the number of peripheral B cells303,349. However, this issue requires further investigation as more recent work using conditional alleles found that while deletion of ikkα in early B cell development led to a reduction in the number of peripheral B cells due to a partial block in B cell maturation at the T2 stage, inducible deletion of ikkα in mature B cells had little effect, and p100 processing was not impaired83. Finally, it is important to note that, while NIK and IKKα have critical roles in inducing the processing of p100 and thereby facilitating nuclear translocation of RELB/p52 heterodimers, these factors can have additional functions. In certain cell systems, NIK has been found to also activate the canonical pathway290,324-327, and IKKα has NF-B-independent roles358. These alternative NF-B pathway-independent functions of NIK and IKKα make it difficult to conclusively identify the biological role of RELB/p52-mediated target gene transcription via the study of NIK or IKKα.

Similar to constitutional deletion of the upstream alternative NF-B components described above, ablation of relb or nfkb2 produces severe phenotypes. In the absence of

RELB or NF-B2 (p100/p52), disrupted lymphoid organization was observed, and RELB- deficient mice in addition were characterized by multiorgan inflammation359-362. Due to these limitations, the roles of RELB and NF-B2 in mature B cell development have been investigated

52 with bone marrow chimeras360,362. Relb–/–→wild-type (WT) chimeras displayed a strong reduction in MZ B cells, suggesting that RELB is required in hematopoietic cells for the generation of this B cell subset362. Reduced peripheral B cells and a distorted MZ were observed in nfkb2–/– mice360,361; however, the study of nfkb2–/–→RAG-1 (recombination activating gene 1) chimeras did not specifically report reduced B cell numbers or impaired MZ B cell development360. Finally, combined constitutive ablation of RELB and NF-B2 has not been undertaken. As such, similar experiments with hematopoietic cells from relb–/–nfkb2–/– mice to determine the biological consequences of the combined deletion of RELB and NF-B2 in B cells have not been conducted.

The studies outlined above have revealed much information about the role of the alternative NF-B pathway in BAFF-mediated B cell homeostasis; however, several issues remain to be clarified. First, the cell-intrinsic requirement of the NF-B2 or RELB subunits for mature B cell maintenance has not been determined using conditional alleles that can be deleted specifically in B cells. Second, knockout of either relb or nfkb2 alone does not allow complete ablation of the alternative NF-B pathway. Indeed, in the absence of either RELB or

NF-B2, the remaining alternative NF-B subunit may be able to drive transcription by binding to canonical NF-B subunits. Furthermore, in nfkb2 knockout mice, lack of NF-B2 not only prevents the generation of p52, but also eliminates the p100 inhibitor. NF-B2 retains RELB in the cytoplasm, and in its absence, inappropriate translocation of RELB into the nucleus may occur, potentially resulting in target gene transcription via the formation of heterodimers with other NF-B subunits. Therefore, identification of the biological consequences of complete inactivation of the alternative NF-B pathway requires ablation of both RELB and NF-B2 subunits. Finally, the transcriptional program activated by the alternative NF-B pathway in B cells following stimulation with BAFF has not been defined. To address these issues, we have generated conditional relb and nfkb2 alleles. We found that combined deletion of relb and nfkb2

53 in B cells had a markedly stronger effect on the survival of mature B cells compared to the single gene deletions. These results reveal the extent to which the activity of the alternative NF-

B pathway controls B cell homeostasis in vivo, and provide new insights into the biological program controlled by the alternative NF-B pathway in response to BAFF.

Section 2.2: MATERIALS AND METHODS

Generation of mice carrying a floxed relb or nfkb2 allele

The targeting vector used to target relb and nfkb2 has been described previously in the context of targeting the irf4 gene200. The relb and nfkb2 vectors were constructed such that upon Cre- mediated deletion, the promoter regions and the exons comprising the first ATG of relb or nfkb2 were deleted with simultaneous activation of eGFP expression (Fig. 2.1-2.3). The conditional relb and nfkb2 alleles were backcrossed to C57BL/6 mice (n ≥ 7). CD19-Cre mice have been described363. Mice were housed and treated in compliance with the US Department of Health and Human Services Guide for the Care and Use of Laboratory Animals and in accordance with the guidelines of the Institute of Comparative Medicine at Columbia University (New York). The animal protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of

Columbia University.

B cell isolation and culture

Single cell suspensions of mouse spleen were subjected to hypotonic lysis and B cells were purified by depletion of magnetically labeled non-B cells using the MACS B cell isolation kit

(Miltenyi Biotec). Purified B cells from the indicated genotypes were cultured in the presence of:

1 g/ml anti-mouse CD40 (clone HM40-3; BD Pharmingen) or 25 ng/mL BAFF (R&D Systems).

Cell density in CD40 and BAFF stimulation experiments was 1.5 x 106 cells/mL using RPMI with

15% FBS. For the RNA-seq analysis, cell pellets were lysed with TRIzol reagent (Life

Technologies) for RNA isolation. For Western blot analysis, purified B cells were washed twice with PBS and subjected to NP40-based lysis.

54

Flow cytometry

Spleen cell suspensions or cultured B cells were stained on ice in PBS/0.5% BSA with the following antibodies. From BD Pharmingen: APC-conjugated anti-IgM (clone II/41), PE- conjugated anti-IgD (clone 11-26c.2a) and PE-conjugated anti-CD23 (clone B3B4). From

Biolegend: PerCP-conjugated anti-B220 (clone RA3-6B2), APC-conjugated anti-CD21 (clone

7E9), PE-conjugated anti-β2 (clone M18/2), PE-conjugated anti-CD93 (clone AA4.1), biotin- conjugated anti-ICOSL (clone HK5.3) followed by streptavidin-PerCP (BD Pharmingen) and PE- conjugated anti-BCL2 (clone BCL/10C4). Intracellular BCL2 was detected using the BD

Cytofix/Cytoperm Fixation Permeabilization Kit. Annexin V/7-AAD stainings were conducted using the APC Annexin V Apoptosis Detection Kit with 7-AAD (Biolegend). For DNA content analysis, cells were lysed and stained with propidium iodide (PI). The cells or nuclei were washed and analyzed on a FACSCalibur or LSRII (BD). Data were analyzed using FlowJo software.

Immunohistochemistry

Sections of 2-3 m in thickness were cut from spleen tissue that was fixed overnight in 10% formalin and embedded in paraffin. Sections were stained with either H&E, unlabeled anti-CD3 rabbit antibody (Thermo Scientific, clone SP7), or anti-GFP rabbit antibody (Molecular Probes,

Invitrogen). Antibodies were incubated overnight at 4°C, stained with anti-rabbit HRP-labeled polymer (Dako) and developed in aminoethylcarbazole (AEC; Sigma). These slides were then counterstained for IgM by overnight incubation at 4°C with alkaline peroxidase (AP)-conjugated anti-IgM antibody (Southern Biotech). IgM stainings were developed in nitro blue tetrazolium chloride–5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP; Roche, New York, NY). The images were acquired by means of a Digital Sight camera (Nikon) mounted on a Nikon Eclipse E600 microscope (Nikon).

Immunoblot analysis

55

Purified B cells were subjected to NP40-based lysis, separated by SDS-PAGE and blotted on nitrocellulose membranes (GE Healthcare). Samples were incubated with the following primary antibodies overnight at 4°C: rabbit anti-RELB (Santa Cruz, clone C-19), rabbit anti-p100/p52

(Cell Signaling Technologies, CST), mouse anti-β-actin (Sigma, clon: AC-15), rabbit anti-pAKT

(Cell Signaling Technologies, CST, clone D9E) and rabbit anti-AKT (Cell Signaling

Technologies, CST). Horseradish peroxidase-conjugated secondary antibodies (anti-rabbit IgG,

GE Healthcare and anti-mouse IgG, Thermo Scientific) and ECL Western Blotting Substrate or

SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) were used for detection.

RNA-seq

RNA was harvested from 3 culture replicates/genotype of BAFF-stimulated B cells and submitted for RNA-sequencing at the Columbia University Genome Center. 30 million single- end, 100bp reads were sequenced per sample on the Illumina HiSeq2000/2500 V3 instrument.

DESeq analysis was used to identify differentially expressed genes and determine fold expression changes between genotypes. Padj is the p value for the statistical significance of the corresponding fold changes adjusted for multiple testing correction to control the false discovery rate (FDR).

Statistical analysis

P values were obtained using unpaired Student’s t test. P values < 0.05 were considered significant.

Section 2.3: RESULTS

Conditional deletion of relb and nfkb2 in B cells

To determine the B cell-intrinsic functions of RELB and NF-B2 in the maintenance of B cells in vivo, we generated transgenic mouse strains carrying loxP-flanked alleles of relb and nfkb2,

56 respectively. The region of the relb or nfkb2 genes flanked by loxP sites comprised the promoter region and the exon encoding the translational start site, thus allowing the creation of relb or nfkb2 null alleles upon Cre-mediated recombination of loxP sites. A previously described targeting vector was utilized to generate these alleles200. To generate a relb targeting vector, 3

DNA fragments were successively inserted into the cloning sites of this targeting vector – (1)

2.2kb of the region upstream of the relb promoter region; (2) the relb promoter region and exon

1 which contains the translation start site and exon 2 (overall 2.8kb); and (3) 3.8 kb of the region downstream of exon 2 (Fig. 2.1). Likewise, to generate a nfkb2 targeting vector, the following

DNA fragments were successively inserted into the targeting vector – (1) 2.0 kb of the region upstream of the nfkb2 promoter region; (2) the nfkb2 promoter region and exon 1 and exon 2, which contains the translational start site (overall 2.4kB); and (3) 4.6 kb of the region downstream of exon 2 (Fig. 2.2). The linearized vectors were electroporated into KV1 embryonic stem (ES) cells (a 129:B6 hybrid ES cell line), and correctly targeted ES cell colonies were identified after selection with gancyclovir and G418 via a PCR-based screening strategy followed by confirmation via Southern blot analysis (Fig. 2.1, Fig. 2.2). 288 ES cell clones transfected with the relb targeting vector were screened via PCR for integration of the 5’ and 3’ regions of the targeting vector. 20 (6.9%) of the screened ES cell clones were positive for 5’ integration, and of these, 5 (25.0%) were positive for 3’ co-integration. The frequency of correctly targeted ES cell clones was therefore 1.7% with a total of five clones identified, which were expanded and correct integration was confirmed via Southern blot analysis (Fig. 2.1). 2 ES cell clones were subsequently injected into blastocysts derived from C57BL/6 mice and one of the 2 injections yielded chimeras. 188 ES cell clones transfected with the nfkb2 targeting vector were screened via PCR for integration of the 5’ and 3’ region of the targeting vector. 18 (9.6%) of the screened ES cell clones were positive for 5’ integration, and of these, 8 (44.4%) were positive for 3’ co-integration. The frequency of correctly targeted ES cell clones was therefore

4.3% with a total of 8 clones identified. 6 of these ES cell clones were expanded and correct

57 integration was confirmed via Southern blot analysis (Fig. 2.2). 2 ES cell clones were subsequently injected into blastocysts derived from C57BL/6 mice and chimeras were obtained from 1 of the 2 injections. Following breeding of chimeras to C57BL/6 females, mice with the

‘conditional’ relb or nfkb2 alleles in the germ-line were obtained.

To enable B cell-specific ablation of relb and nfkb2, these mice were crossed to CD19-

Cre mice which express Cre recombinase specifically in B cells from the pre-B stage of bone marrow development363. The targeting vector used to generate the relb and nfkb2 alleles was constructed such that the recombination of loxP sites is accompanied by the expression of an enhanced green fluorescent protein (eGFP), thus enabling tracking of relb or nfkb2-deleted cells in the tissues (Fig. 2.1, 2.2, 2.3). To achieve this, an eGFP minigene was placed upstream of the relb or nfkb2 promoter regions in the opposite orientation; this minigene is silent in the targeted locus. However, following recombination of loxP sites, a phosphoglycerate kinase

(PGK) promoter (placed in intron 2 of relb or nfkb2) is placed proximal to the eGFP gene, allowing transcription. A similar strategy was previously used for the conditional deletion of irf4, rel (the gene encoding c-REL) and rela200,338. In this system, gene deletion is directly linked to induction of eGFP expression, and therefore eGFP positivity is a marker for gene deletion. In addition, eGFP expression is quantitative, as B cells from mice heterozygous or homozygous for the floxed allele can be distinguished by their mean eGFP fluorescence200,338 (also see Fig.

2.4A,B lower panels).

Western blot analysis of B cells isolated from relbfl/flCD19-Cre and nfkb2fl/flCD19-Cre mice demonstrated that the protein levels of RELB and p52 were strongly reduced compared to

B cells from littermate controls (Fig. 2.3B,E), confirming the creation of null alleles upon recombination of the loxP-flanked relb and nfkb2 alleles. The remaining protein is likely to be derived from eGFP– B cells that have escaped Cre-mediated deletion. This possibility was confirmed by the complete absence of RELB protein in flow-sorted eGFP+ B cells from relbfl/flCD19-Cre mice (Fig. 2.3C). The low eGFP expression in B cells of nfkb2fl/flCD19-Cre mice

58 prevented us from unambiguously separating eGFP+ and eGFP– B cells; we were therefore unable to perform a similar experiment for NF-B2. B cells from relbfl/+CD19-Cre and nfkb2fl/+CD19-Cre mice respectively showed reduced protein expression of RELB and p52 relative to the CD19-Cre controls, indicating haploinsufficiency of RELB and p52 expression in B cells (Fig. 2.3B,C,E). Importantly, in the absence of Cre, the unrearranged conditional relb and nfkb2 alleles produced physiological amounts of RELB and p52 protein, respectively (Fig.

2.3B,E).

Reduced fractions of splenic B cells in the absence of alternative NF-B subunits

To assess the B cell-intrinsic functions of the alternative NF-B subunits in B cell maintenance, we determined the fraction of mature splenic B cells present in mice with B cell-specific ablation of RELB or NF-B2 individually, or combined RELB and NF-B2 inactivation. Upon B cell- specific ablation of RELB in relbfl/flCD19-Cre mice, we observed a slight but significantly reduced fraction of splenic B cells compared to relbfl/+CD19-Cre and CD19-Cre control mice (Fig. 2.4A; for cell numbers, see Fig. 2.5A). B cell-specific ablation of NF-B2 was associated with a marked reduction in the fraction of splenic B cells; B220+ cells comprised 28% of total splenocytes in nfkb2fl/flCD19-Cre mice compared to 46% and 50% in nfkb2fl/+CD19-Cre and

CD19-Cre control mice, respectively (Fig. 2.4B; for cell numbers, see Fig. 2.5B). In order to obtain complete ablation of RELB/p52, the major heterodimer of the alternative NF-B pathway, the conditional relb and nfkb2 alleles were interbred to generate relbfl/flnfkb2fl/flCD19-Cre mice to allow simultaneous deletion of relb and nfkb2 in B cells. The fraction of splenic B cells in these mice was similar to that observed following ablation of NF-B2 alone (B220+ cells comprised

30% of total splenocytes in relbfl/flnfkb2fl/flCD19-Cre mice compared to 45% and 50% in relbfl/+nfkb2fl/+CD19-Cre and CD19-Cre control mice, respectively; Fig. 2.4C; for cell numbers, see Fig. 2.5C). However, analysis of eGFP expression among B cells of relbfl/flnfkb2fl/flCD19-Cre mice identified two distinct eGFP+ and eGFP– peaks of approximately equal proportions (Fig.

59

2.4C; bottom panel). The presence of a sizable eGFP– population in mice with combined ablation of RELB and NF-B2 indicates that eGFP+ B cells deficient for both RELB and NF-B2 are outcompeted by eGFP– B cells that escaped Cre-mediated deletion. As a result, relb/nfkb2- deleted eGFP+ B cells were present at a frequency of only 15% of total splenocytes in relbfl/flnfkb2fl/flCD19-Cre mice compared to the B cell fraction of 50% observed in CD19-Cre control mice.

The unique eGFP-expression pattern in relbfl/flnfkb2fl/flCD19-Cre mice contrasts with the single eGFP+ peak observed in B cells from relbfl/flCD19-Cre mice (Fig. 2.4A; bottom panel), suggesting that there is virtually no counter-selection against RELB-deficient B cells in these mice. Due to the generally low eGFP-expression in B cells from nfkb2fl/flCD19-Cre mice, a clear distinction between eGFP+ and eGFP– populations is not possible; however, the eGFP expression of B cells from nfkb2fl/flCD19-Cre mice is generally higher than that of B cells from mice with heterozygous deletion of nfkb2 (nfkb2fl/+CD19-Cre mice) (Fig. 2.4B; bottom panel). In addition, Western blot analysis (Fig. 2.3E) demonstrated that the majority of B cells from nfkb2fl/flCD19-Cre mice do not express p52 protein. Together, these observations argue against extensive counter-selection of NF-B2-deficient B cells in these mice, unlike what is observed in relbfl/flnfkb2fl/flCD19-Cre mice. Nonetheless, deletion of nfkb2 alone led to a strong reduction in the fraction of B cells that was considerably higher than that observed in relbfl/flCD19-Cre mice, indicating that functionally, p52 appears to be the major subunit downstream of the alternative pathway in B cells.

Immunohistochemical staining analysis of spleen sections for IgM and CD3 revealed that in accordance with the reduced B cell fraction observed by flow cytometry, relbfl/flnfkb2fl/flCD19-

Cre mice displayed fewer B cell follicles within the white pulp compared to relbfl/+nfkb2fl/+CD19-

Cre and CD19-Cre control mice (Fig. 2.6). The size of these B cell follicles was also more heterogeneous in relbfl/flnfkb2fl/flCD19-Cre mice compared to the controls. In summary,

60 combined B cell-specific deletion of relb and nfkb2 results in a greatly reduced fraction of B cells that is reflected by abnormalities in splenic white pulp architecture. Moreover, the severity of the

B cell phenotype upon relb/nfkb2 deletion appears to surpass that observed with the corresponding single gene deletions.

Marked counter-selection against relb/nfkb2-deleted MZ B cells

The bone marrow compartment is comprised of immature B cells and mature B cells that have yet to exit into the periphery. Mice with B cell-specific ablation of both RELB and NF-B2 compared to controls displayed reduced frequencies of mature B cells (B220+CD93–) in the bone marrow, whereas immature B cells were present at normal frequencies (Fig. 2.7A). This is in agreement with published observations from relb–/– and nfkb2–/– mice which concluded that B cell development in the bone marrow is not impaired in these mice359,360. In addition, we observed that ~70% of immature CD93+B220int B cells in the bone marrow were positive for eGFP, which is a readout for gene deletion (Fig. 2.7B). This frequency is in the range of the deletion efficiency observed for CD19-Cre in immature B cells (77%±8%)363 which is lower than the frequency observed in splenic B cells (93%±4%)363; therefore, there does not appear to be counter-selection against relb/nfkb2-deleted immature B cells. In contrast, the fraction of mature

B cells in the bone marrow that are positive for eGFP was ~40% (Fig. 2.7B), indicating that during the transition from the immature to the mature B cell stage, the fraction of relb/nfkb2- deleted B cells is strongly reduced. This transition is also the timeframe during which BAFF responsiveness is initiated64,67,78-81, and the counter-selection against relb/nfkb2-deleted B cells observed following this transition is likely due to the requirement of the alternative NF-B pathway to transmit BAFF signals286-288. As such, eGFP– B cells successfully outcompeted eGFP+ B cells (i.e. relb/nfkb2-deleted) and became a significant population comprising approximately half of splenic peripheral B cells (Fig. 2.4C; bottom panel).

61

Among the splenic B cell subsets in relbfl/flnfkb2fl/flCD19-Cre mice, a slightly reduced fraction of follicular B cells (CD23+CD21int) and an increased fraction of MZ B cells

(CD21hiCD23–) was observed compared to CD19-Cre mice (~70% vs. ~83% follicular B cells and ~11% vs. ~6% MZ B cells) (Fig. 2.8A). Similar observations were made when analyzing

IgM+IgD+ and IgMhiIgDlo B cell subsets (data not shown). However, the extent of counter- selection against relb/nfkb2 deleted B cells differed considerably between the follicular and MZ

B cell fractions (Fig. 2.8B), with the strongest counter-selection observed in the MZ B cell compartment. Analysis of eGFP expression within MZ B cells (Fig. 2.8B) demonstrated that only

~25% of MZ B cells were eGFP+ compared to ~57% of follicular B cells. The observed dramatic counter-selection against relb/nfkb2-deleted MZ B cells is consistent with previous publications that invoked roles for RELB or NF-B2 in the development of MZ B cells from the study of constitutional knockout mice352,362. Curiously, morphologic evaluation of hematoxylin & eosin stained splenic sections from relbfl/flnfkb2fl/flCD19-Cre mice demonstrated focal MZ enlargement in these mice compared to the controls (Fig. 2.9; left panels). The prominent MZs presumably reflect the higher fractions of MZ vs. follicular B cells that were observed among splenic B cells by flow cytometric analysis (see Fig. 2.8A). These observations suggest that while there is an overall reduction in both follicular B cells and MZ B cells in relbfl/flnfkb2fl/flCD19-Cre mice, of the

B cells that remain, the fraction of MZ B cells is increased relative to follicular B cells.

Importantly, immunohistochemical staining analysis of splenic sections revealed that the majority of the B cells in the enlarged MZs of relbfl/flnfkb2fl/flCD19-Cre mice did not stain for eGFP, whereas eGFP+ follicular B cells were readily observed (Fig. 2.9; right panels), a finding which is in accordance with the flow cytometric results (Fig. 2.8B). In summary, relbfl/flnfkb2fl/flCD19-Cre mice show strong counter-selection against relb/nfkb2-deleted MZ B cells.

Identification of BAFF-responsive genes controlled by the alternative NF-B pathway

62

The data outlined above demonstrate that the alternative NF-B pathway controls a biological program that promotes B cell maintenance. In order to gain insight into this program, we sought to identify genes that are transcriptionally regulated by the alternative NF-B subunits in B cells stimulated with BAFF in vitro. A time-course analysis of BAFF stimulation on normal B cells revealed that at 6h, the majority of p100 was processed to p52, demonstrating robust activation of the alternative NF-B pathway (Fig. 2.10A). Phosphorylation of AKT, an additional event known to occur downstream of BAFF stimulation85, was observed after 24h of stimulation (Fig.

2.10A). We chose to stimulate B cells for 6h in order to identify early BAFF-responsive genes. B cells purified by magnetic cell separation from nfkb2fl/flCD19-Cre mice were used for this transcriptional profiling experiment as these mice showed a marked reduction in mature B cells without evidence of strong counter-selection against nfkb2-deleted B cells, unlike what is observed in relbfl/flnfkb2fl/flCD19-Cre mice. B cells isolated from three nfkb2fl/flCD19-Cre and

CD19-Cre mice each were stimulated with BAFF for 6 hours (6h) in vitro, and RNA harvested from the cultures was subjected to RNA sequencing.

Differentially expressed sequence analysis (DESeq) of NF-B2-proficient vs. NF-B2- deficient BAFF-stimulated B cells identified 150 genes with reduced (Fig. 2.10B), and 174 genes with increased expression in the absence of NF-B2 at a significance threshold of padj<0.05

(Appendix A). The fold changes were mostly small, but highly significant. Genes were assigned to putative functional categories (Fig. 2.10B and Appendix A), including B cell markers, apoptosis, cellular trafficking, and immune response–T-dependent (Fig. 2.10C). A transcriptional profile analysis of BAFF-responsive genes controlled by the alternative NF-B pathway has not previously been undertaken. However, cd23 (encoding FcεRII) is a known BAFF-controlled gene in B cells364, and the gene encoding inducible costimulator ligand (ICOSL) has been identified as a bona fide target of the alternative subunits in response to BAFF-R stimulation in murine B cells365. In the RNA-seq analysis, we observed reduced expression of mRNA encoding

63 cd23 and icosl in NF-B2-deficient vs. control B cells (Fig. 2.10C). Accordingly, following BAFF or CD40 stimulation, ICOSL protein was expressed at significantly reduced levels on eGFP+ (i.e. relb/nfkb2-deleted) B cells derived from relbfl/flnfkb2fl/flCD19-Cre mice, whereas eGFP– B cells from the same mice displayed surface levels of ICOSL similar to those observed on WT B cells

(Fig. 2.12A). These results validated our strategy for identifying BAFF-responsive targets. The major findings of the RNA-seq analysis are presented below.

Control of BAFF-mediated survival by the alternative NF-B pathway

We observed reduced expression of mRNA encoding the anti-apoptotic gene bcl2 in NF-B2- deficient vs. control B cells (Fig. 2.10C). In accordance, flow cytometric analysis revealed lower

BCL2 protein expression by B220+ B cells in relbfl/flnfkb2fl/flCD19-Cre mice compared to CD19-

Cre control mice (Fig. 2.11A). In addition, the genes encoding the pro-apoptotic factors BCL2 modifying factor (BMF) and BCL2 binding component 3 (BBC3, PUMA) were expressed at higher levels in BAFF-stimulated relb/nfkb2-deleted vs. WT B cells (see Appendix A). Together, these findings suggest that B cells with impaired signaling through the alternative NF-B pathway have an increased propensity to undergo apoptosis. In agreement, following BAFF- stimulation, increased cell death of B cells that express a RELB mutant protein which is unable to bind DNA has previously been demonstrated compared to WT controls366. To assess the survival capacity of RELB/NF-B2-deficient B cells in vitro, we stimulated B cells from relbfl/flnfkb2fl/flCD19-Cre or CD19-Cre control mice with either BAFF or CD40 for three days, as both BAFF and CD40 are strong activators of the alternative NF-B pathway283,286-288. The results showed significantly enhanced cell death in the cultures of RELB/NF-B2-deficient vs. control B cells in response to BAFF stimulation (~63% vs. ~23% by propidium iodide (PI) staining and ~66% vs. ~21% by annexin V/7AAD staining) (Fig. 2.11B,C). In contrast, cell viability was comparable among the genotypes in response to CD40 stimulation (~22% vs.

~14% by PI staining and ~20% vs. ~17% by annexin V/7AAD staining), although the small

64 difference observed in the PI staining reached significance (Fig. 2.11B,C). These findings suggest that CD40 stimulation activates a transcriptional program in RELB/NF-B2-deficient B cells that is able to overcome the viability defect observed in response to BAFF.

Control of B cell trafficking by the alternative NF-B pathway

Both relb–/– and nfkb2–/– mice show altered lymphoid organization359-362, and analysis of total splenic mRNA from RELB-deficient mice revealed reduced expression of the chemokines

CXCL13, CCL19 and CCL21, and mildly reduced expression of the chemokine receptors

CXCR5 and CCR7 compared to controls362. These findings provided the first suggestion that the alternative pathway may be required for proper lymphoid organization by controlling the expression of chemokines or chemokine receptors in stromal and/or other cell types. In support of this notion, RNA-seq analysis revealed reduced mRNA expression of the genes encoding the chemokine receptors CXCR5, CCR7 and EBI2 (Gpr183), as well as the adhesion molecule L- selectin (encoded by sell) in NF-B2-deficient B cells compared to controls following BAFF stimulation (Fig. 2.10C). These findings suggests that in B cells, the alternative NF-B pathway controls a gene expression program that permits homing to and positioning within B cell follicles in response to cues from stromal cells.

Since the absence of alternative NF-B subunits correlated with a reduction in the fraction of MZ B cells (ref.352,362, and Fig. 2.8B, Fig. 2.9), we determined the surface protein expression of the integrin chains αL and β2, that together form the LFA-1 complex, on relb/nfkb2-deleted and WT control B cells. LFA-1 is important for the retention of B cells within the splenic MZ41. Whereas we detected no significant differences in the expression level of αL between eGFP+ (i.e. relb/nfkb2-deleted) B cells from relbfl/flnfkb2fl/flCD19-Cre and CD19-Cre control mice, relb/nfkb2-deleted B cells showed strongly reduced surface expression of β2 compared to B cells from CD19-Cre control mice (Fig. 2.12B). This observation is likely to be relevant to the strong counter-selection against eGFP+ MZ B cells observed in

65 relbfl/flnfkb2fl/flCD19-Cre mice (Fig. 2.8B, Fig. 2.9), and may suggest an impaired retention of these cells in the MZ compartment (see discussion, Section 2.4).

CD21 has been identified as a target of BAFF receptor signaling364. We observed lower protein levels on the surface of eGFP+ relb/nfkb2-deleted B cells in comparison to CD19-Cre control mice (Fig. 2.12C), suggesting that BAFF mediates expression of CD21 via the alternative NF-B pathway. CD21 is a complement receptor and high expression of CD21 is a marker of MZ B cells59. Reduced expression of CD21 may therefore result in less efficient binding to complement opsonized antigens, an important function of MZ B cells45,51,52. It is also possible that reduced expression of CD21 on RELB/NF-B2-deficient MZ B cells impairs the antigen-stimulation properties of those cells as the CD19/CD21 co-receptor complex has been implicated in lowering the threshold of BCR signaling367.

Genes involved in T cell-B cell interactions regulated by the alternative NF-B pathway

As mentioned above, the alternative NF-B pathway has been implicated in the regulation of

ICOSL expression on B cells365. ICOSL is involved in facilitating interactions with T cells368 and its expression on B cells is required for the development of Tfh cells365. In addition to decreased expression of icosl in BAFF-stimulated NF-B2-deficient vs. control B cells, our RNA-seq analysis revealed reduced expression of several genes involved in MHC class-II (MHCII) antigen presentation (Fig. 2.10C). These included the genes ctss, h2-dmb2 and h2-oa, which respectively encode cathepsin S, H2-DMb2 and H2-Oa and are involved in the processing and removal of invariant chain peptide (CLIP) from MHCII molecules to allow loading of antigenic peptides369. In addition, cd83 was expressed at lower levels in NF-B2-deficient B cells compared to WT B cells (Fig. 2.10C). CD83 is a member of the Ig superfamily that is associated with B cell activation, and CD83-deficient splenic B cells also have reduced MHCII antigen presentation370. Thus, it appears that besides regulating ICOSL expression, the alternative NF-

66

B pathway may have a more substantial role in activating the expression of genes required for optimal T cell-dependent B cell responses in response to BAFF.

Section 2.4: DISCUSSION

BAFF-mediated activation of the alternative NF-B pathway has long been implicated in the survival of mature B cells. However, the B cell-intrinsic roles of transcription factors downstream of the pathway, RELB or NF-B2, have not as yet been identified. Perhaps most importantly, it is unknown how the combined deletion of relb and nfkb2 affects the development and function of B cells. We here crossed newly generated conditional relb and nfkb2 alleles, either separately or in combination, to CD19-Cre mice in order ablate these subunits specifically in B cells. We found that the individual gene deletions led to a decrease in the fraction of mature B cells; however, the most dramatic reduction in peripheral B cells was observed upon simultaneously ablating RELB and NF-B2. Thus, by ablating the downstream transcription factors of the alternative NF-B pathway in B cells, our results for the first time define the extent to which this pathway regulates mature B cell homeostasis. In addition, we identified BAFF-responsive target genes of the alternative NF-B pathway that are involved in the survival and trafficking of peripheral B cells and their interaction with T cells.

Similar to what has been described in mice with conditional B cell ablation of major regulators of the canonical pathway, IKKβ or IKKγ339,340, we observed a marked counter- selection of relb/nfkb2-deleted B cells in mice with B cell-specific RELB/NF-B2 deficiency, which was strongest in the MZ B cell compartment. Nevertheless, none of these mouse models showed a complete disappearance of ikkβ, ikkγ or relb/nfkb2-deleted B cells, indicating that the

B cell phenotypes are less severe than those observed in mice deficient in BAFF signaling that show a virtually complete absence of B cells. One possibility to explain these findings is that downstream of BAFF, the canonical and alternative NF-B pathways may complement each other to a certain extent during B cell development and maintenance, as suggested by the

67 observation that adoptive transfer of bone marrow from nfkb1–/–nfkb2–/– mice into RAG-1- deficient animals resulted in the generation of only few mature B cells371. Of note, ablation of the canonical subunits c-REL or RELA in B cells does not appear to impact mature B cell numbers

(Nicole Heise (Klein lab), unpublished findings), which may be suggestive of redundancy between these two subunits. The generation of mice with B cell-specific double ablation of c-

REL and RELA will clarify this issue. Finally, it is known that non-NF-B-mediated effects of

BAFF signaling contribute to the survival of mature B cells83-87 and these effects may also promote the survival of relb/nfkb2-deleted B cells.

In order to make a comparative assessment of the severity of the B cell maintenance defects that manifest in the absence of NF-B pathway components, the experimental systems should be comparable. In regards to the alternative NF-B pathway, CD19-Cre has been used to ablate TRAF2, TRAF3 and cIAP1/2 which all result in B cell expansion355-357. CD19-Cre- mediated ablation of IKKα has also been conducted and the reduction in peripheral B cells, while similar to what we report here for combined ablation of RELB and NF-B2, is slightly less severe83. Also, no counterselection against ikkα-deficient B cells was evident83. Moreover, while deletion of ikkα in early B cell development using CD19-Cre (which deletes prior to the expression of the BAFF-R) led to a reduction in peripheral B cells due to a partial block in B cell maturation at the T2 stage, inducible deletion in mature B cells had little effect on peripheral B cell survival83. This study highlights an important distinction between a defect in the differentiation of T2 B cells into mature B cells and a defect in the maintenance of mature B cells once they have passed this transitional developmental stage. It is therefore possible that IKKα is required for progression through the T2 stage, while it is not required for continued maintenance of mature peripheral B cells83. Most studies aimed at understanding the roles of both canonical and alternative NF-B pathway components in B cell maintenance, including our study, cannot clearly distinguish between these two possibilities due to the use of bone marrow transfers or a

68

Cre expressed in early B cell development. While we cannot make such a distinction without the use of an inducible Cre, this issue can be further investigated by determining the effect of

RELB/NF-B2 deficiency on transitional B cell subsets in the spleen, as to date we have only broadly characterized immature vs. mature B cells in the bone marrow of relb/nfkb2-deleted mice. Of note, the maintenance of mature B cells was recently assessed following inducible ablation of NIK372. However, in contrast to the IKKα report described above83, NIK ablation was induced ubiquitously in all cell types372. Reduced B cells were observed over time in mesenteric lymph nodes and Peyer’s patches upon inducible ablation of NIK, demonstrating that continued

NIK expression is important for the survival of mature B cells in these sites372.

In in vitro-cultures, relb/nfkb2-deleted B cells were sensitive to apoptosis when stimulated with BAFF, suggesting that the alternative NF-B pathway controls an anti-apoptotic program in normal B cells that is activated in response to BAFF stimulation. BAFF signaling has long been implicated in controlling B cell survival78, at least in part through the regulation of

BCL2 expression373. In agreement with this notion, ectopic expression of BCL2 led to the accumulation of B cells in BAFF-R-deficient mice67,374,375. Our observation that peripheral B cells in mice with combined B cell-specific deletion of relb and nfkb2 died rapidly in vitro following

BAFF stimulation and showed reduced protein levels of BCL2 suggests that bcl2 expression is at least partially mediated via the alternative NF-B pathway. Indeed, bcl2 has been described to be a target of nfkb2 in a cell line376. It is interesting to note that activation of relb/nfkb2-deleted

B cells with CD40, also a strong inducer of p100 processing283, did not lead to increased cell death. This result may indicate that CD40 signaling can transmit survival signals by a different route, presumably via the canonical NF-B pathway377. Similarly, while B cells derived from mice with inducible ablation of NIK were found to exhibit greater susceptibility to cell death following BAFF stimulation, proliferation in response to CD40 stimulation was comparable to control B cells372. By means of exploring CD40 signaling in further detail we plan to conduct

69 gene expression profiling on CD40-stimulated nfkb2-deficient B cells, similar to the approach used to identify BAFF-responsive genes.

BCL2 overexpression does not fully rescue the loss of B cells in BAFF-R-deficient mice as only follicular B cells with an immature surface phenotype develop in these mice. These B cells display improper follicular organization67,374,375, indicating that BAFF stimulation has additional functions besides providing pro-survival signals. In accordance, our RNA-seq analysis identified several BAFF responsive genes controlled by the alternative NF-B pathway that are involved in lymphocyte trafficking, including adhesion molecules and chemokine receptors. Two previous studies have implicated the alternative pathway in the regulation of B cell homing to the lymphoid follicles and the MZ, respectively. Based on an RT-PCR analysis of mRNA derived from whole spleen of relb–/– mice, Weih et al. reported slightly reduced mRNA levels of the genes encoding the chemokine receptors CXCR5 and CCR7362. Expression of CXCR5 is required for homing of B cells to the follicle25,26, while CCR7 directs positioning at the B-T cell boundary following antigen stimulation34,35. We found both receptors to be expressed at lower mRNA levels in BAFF-stimulated NF-B2-deficient B cells compared to controls. In addition, we observed reduced expression of the chemokine receptor EBI2, which guides B cell localization to the outer follicle and must be downregulated to enable GC formation101-105, and the adhesion molecule L-selectin (CD62L), which mediates the entry of lymphocytes from the blood into lymphoid tissue378. These results suggest that RELB/NF-B2-deficient B cells that fail to express chemokine receptors and adhesion molecules at physiological levels may be impaired in their homing to and organization within the follicle. In agreement with this notion, mice with B cell- specific deletion of both relb and nfkb2 have a decreased number of splenic B cells and fewer B cell follicles (Fig. 2.4C, Fig. 2.5C, Fig. 2.6). In vitro migration assays that assess the ability of

RELB/NF-B2-deficient B cells to travel towards the CXCR5 ligand CXCL13 may lend further support to a role for the alternative NF-B pathway in B cell migration.

70

A critical role for the alternative NF-B pathway in the development of MZ B cells has been suggested based on two observations: First, relb–/–→WT chimeras show a significant reduction of MZ B cells compared to the controls362. Second, BAFF-transgenic mice, mice with inactivation of upstream components of the alternative pathway (that induce processing of p100) and mice that lack p100 but still express p52, are characterized by MZ hyperplasia75,302,355-357,379.

As reported by Enzler et al., the latter phenotype is likely due to the upregulation of the integrin chains αLβ2 (LFA-1) and α4β1 (VLA-4)352 that facilitate homing to the MZ41. The expression of these genes was found to be at least partially impaired in BAFF-Tg/nfkb2–/– mice that presented with a strongly reduced fraction of MZ B cells relative to BAFF-Tg mice352. In light of these findings, our observation that, compared to the controls, relb/nfkb2-deleted B cells had significantly reduced surface-expression of the β2 integrin chain suggests that these cells might have defects in homing to and retention within the MZ. In turn, this may contribute to the strong reduction in RELB/NF-B2-deficient MZ B cells observed in the spleens of relbfl/flnfkb2fl/flCD19-

Cre mice. It is possible that relb/nfkb2-deleted MZ B cells that fail to be retained by the stroma are flushed out into the blood and eventually disappear due to the lack of stroma-associated survival signals, similar to what has been observed in various other experimental systems41,42,352.

Adding to a previous report365, we observed reduced protein expression of ICOSL on the cell surface of B cells upon BAFF as well as CD40 stimulation in relb/nfkb2-deleted B cells. Our

RNA-seq analysis in addition revealed that several genes involved in MHC class II-antigen presentation failed to be expressed at physiological levels in BAFF-stimulated NF-B2-deficient

B cells. These findings further suggest that the alternative NF-B pathway appears to control the expression of genes involved in facilitating optimal T–B cell interactions.

71

Figure 2.1. Generation of a conditional relb allele. Relb (RELB) targeting strategy. (Top)

Shown from top to bottom i) the relb locus with the promoter region, exons and translational start site in exon 1 (E1); ii) the targeting vector comprising an enhanced green fluorescent protein (eGFP), a triple simian virus 40 (SV40) polyA site (tpA), a phosphoglycerate kinase

(PGK) neomycin (neopA) cassette, a PGK promoter (P) downstream of exon 1, a thymidine kinase (TK) gene, and loxP and frt sites; arrows denote the orientation; iii) the targeted relb locus and iv) the locus after Cre-mediated recombination of loxP sites. Restriction sites, probes used for detection and the expected fragments detected by Southern blot analysis are indicated.

(Bottom) Correctly targeted embryonic stem (ES) cell lines were identified by Southern blot analysis of EcoRV-digested DNA that displayed the 5.9 kb band of the integrated transgene along with the 11.4 kb wild-type (WT) band. Co-integration of the 3’ loxP site was verified by the presence of the expected 4.8 kb transgenic band.

72

73

Figure 2.2. Generation of a conditional nfkb2 allele. Nfkb2 (NF-kB2) targeting. (Top) Shown from top to bottom i) the nfkb2 locus with the promoter region, exons and translational start site in exon 2 (E2); ii) the targeting vector comprising an enhanced green fluorescent protein

(eGFP), a triple simian virus 40 (SV40) polyA site (tpA), a phosphoglycerate kinase (PGK) neomycin (neopA) cassette, a PGK promoter (P) downstream of exon 2, a thymidine kinase

(TK) gene, and loxP and frt sites; arrows denote the orientation; iii) the targeted nfkb2 locus and iv) the locus after Cre-mediated recombination of loxP sites. Restriction sites, probes used for detection and the expected fragments detected by Southern blot analysis are indicated.

(Bottom) Correctly targeted embryonic stem (ES) cell lines were identified by Southern blot analysis of PvuII-digested DNA that displayed the 8.3 kb band of the integrated transgene along with the 5.2 kb wild-type (WT) band. Co-integration of the 3’ loxP site was verified by PCR analysis using primers that hybridize in a unique region spanning the PGK promoter and the 3’ frt site (forward primer A) and in a region upstream of exon 3 of nfkb2 (reverse primer B).

74

75

Figure 2.3. Generation of mice with conditional deletion of relb or nfkb2 in B cells and simultaneous expression of eGFP. (A and D) Targeting strategy showing the status of relb and nfkb2 before (top) and after (bottom) Cre-mediated recombination. Numbers indicate the respective exons. (B and E) Western blot analysis of RELB and p52 protein levels in purified splenic B cells of the indicated genotypes, and (C) of flow-sorted eGFP+ B cells from relbfl/flCD19-Cre mice and relbfl/+CD19-Cre and eGFP– CD19-Cre control mice.

76

77

Figure 2.4. Reduced fractions of splenic B cells in the absence of alternative NF-B subunits. (A-C, top) The fractions of splenic B cells from relbfl/flCD19-Cre, nfkb2fl/flCD19-Cre or relbfl/flnfkb2fl/flCD19-Cre mice and the corresponding heterozygous and CD19-Cre control mice were determined by flow cytometry. Each symbol represents a mouse. Data are shown as mean

± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05; **,

P<0.01; ***, P<0.001). (A-C, bottom) Flow cytometry of eGFP expression in splenic B cells of the indicated genotypes. The number above the gate indicates the percentage of eGFP+ B cells among B220+ B cells of relbfl/flnfkb2fl/flCD19-Cre mice. The “eGFP+” arrow denotes the fraction of eGFP+ B cells present in relbfl/flnfkb2fl/flCD19-Cre mice.

78

79

Figure 2.5. Reduced numbers of splenic B cells in the absence of alternative NF-B subunits. (A-C) Absolute numbers of splenic B cells from relbfl/flCD19-Cre, nfkb2fl/flCD19-Cre or relbfl/flnfkb2fl/flCD19-Cre mice and the corresponding heterozygous and CD19-Cre control mice.

Data are shown as mean ± standard deviation. Statistical significance was determined by

Student’s t test (*, P<0.05; ***, P<0.001). The “eGFP+” arrow denotes the number of eGFP+ B cells present in relbfl/flnfkb2fl/flCD19-Cre mice.

80

81

Figure 2.6. relbfl/flnfkb2fl/flCD19-Cre mice display fewer B cell follicles within the white pulp, and exhibit more heterogeneity in the size of B cell follicles, compared to control mice. Spleen sections from relbfl/flnfkb2fl/flCD19-Cre mice and the corresponding heterozygous and CD19-Cre control mice were analyzed via immunohistochemistry for the expression of CD3 and IgM. One representative mouse out of 3 per group is shown. 40x magnification (left) and

100x magnification (right).

82

83

Figure 2.7. Reduced fraction of mature B cells in the bone marrow in relbfl/flnfkb2fl/flCD19-

Cre mice. (A) B220 and CD93 expression of bone marrow lymphocytes from relbfl/flnfkb2fl/flCD19-Cre mice and the corresponding heterozygous and CD19-Cre control mice were analyzed by flow cytometry. Numbers beside each gate indicate the percentage of immature (CD93+B220int) and mature (B220+CD93–) bone marrow B cells (top). Summary of the frequencies of immature and mature bone marrow B cells (bottom). (B) The fractions of eGFP+ cells among immature (CD93+B220int) and mature (B220+CD93–) bone marrow B cells in relbfl/flnfkb2fl/flCD19-Cre mice was determined by flow cytometry. Numbers above the gates indicate the percentage of eGFP+ B cells among the indicated B cell subsets (top). Summary of the frequency of eGFP+ cells among the corresponding B cell subsets (bottom). Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05; **, P<0.01, ***, P<0.001).

84

85

Figure 2.8. Counter selection against relb/nfkb2-deleted MZ B cells in relbfl/flnfkb2fl/flCD19-

Cre mice (flow cytometry). (A) CD21 and CD23 expression of splenic B cells from relbfl/flnfkb2fl/flCD19-Cre mice and the corresponding heterozygous and CD19-Cre control mice was analyzed by flow cytometry. Numbers beside each gate indicate the percentage of follicular

(CD23+CD21int) and marginal zone (MZ) (CD21hiCD23–) B cells (top). Summary of the frequencies of follicular and MZ B cells (bottom). (B) The fractions of eGFP+ cells among splenic follicular (CD23+CD21int) and MZ (CD21hiCD23–) B cells in relbfl/flnfkb2fl/flCD19-Cre mice was determined by flow cytometry. Numbers above the gates indicate the percentage of eGFP+ B cells among the indicated B cell subsets (top). Summary of the frequency of eGFP+ cells among the corresponding B cell subsets (bottom). Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*,

P<0.05; ***, P<0.001).

86

87

Figure 2.9. Counter selection against relb/nfkb2-deleted MZ B cells in relbfl/flnfkb2fl/flCD19-

Cre mice (immunohistochemistry). (left) Hematoxylin & eosin (H&E) stained splenic sections from relbfl/flnfkb2fl/flCD19-Cre mice and CD19-Cre control mice demonstrating focal MZ enlargement (arrows) in the former mice. 40x. (right) Spleen sections from mice of the indicated genotypes were analyzed for the expression of eGFP and IgM by immunohistochemistry. FO, follicular area; MZ, marginal zone area. One representative mouse out of 3 per group is shown.

400x.

88

89

Figure 2.10. Identification of BAFF-responsive genes controlled by the alternative NF-B pathway by RNA-seq analysis of BAFF-stimulated nfkb2-deleted and WT B cells. (A)

Western blot analysis of purified WT B cells ex vivo (day 0), after 3 hours (h) of culture in medium only, and stimulated for 6h and 24h with BAFF. (B) Differentially expressed sequence analysis (DESeq) of NF-B2-proficient vs. NF-B2-deficient B cells stimulated with BAFF for 6h identified 150 genes with reduced expression in the absence of NF-B2 at a significance threshold of padj<0.05. Genes were assigned to putative functional categories. For the identity of the corresponding genes, fold-change and padj values, see Appendix A. (C) Genes of selected categories discussed in the text. Fold-change and padj values are indicated.

90

91

Figure 2.11. Control of BAFF-mediated survival by the alternative NF-B pathway. (A)

(left) Flow cytometric analysis of ex vivo B cells from relbfl/flnfkb2fl/flCD19-Cre and CD19-Cre mice for BCL2 expression. (right) Summary of the results obtained for relbfl/flnfkb2fl/flCD19-Cre mice represented as percentage median fluorescence intensity (MFI) of CD19-Cre mice. (B)

Flow cytometric analysis of BAFF or CD40-stimulated purified B cells from relbfl/flnfkb2fl/flCD19-

Cre and CD19-Cre mice at day 3 for DNA content by propidium iodide (PI) staining (top) and summary of the corresponding percentage Sub-G1 (bottom). (C) Flow cytometric analysis of

BAFF or CD40-stimulated purified B cells from relbfl/flnfkb2fl/flCD19-Cre and CD19-Cre mice at day 3 for apoptotic/dead cells by annexin V/7AAD staining (top) and summary of the corresponding percentage of annexin V/7AAD+ cells (bottom). Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05; **, P<0.01; ***, P<0.001).

92

93

Figure 2.12. Differential expression of BAFF-responsive genes controlled by the alternative NF-B pathway. (A) Flow cytometric analysis of BAFF or CD40-stimulated B cells from relbfl/flnfkb2fl/flCD19-Cre and CD19-Cre mice at day 3 for the expression of ICOSL. Staining of eGFP+ and eGFP– B cells from relbfl/flnfkb2fl/flCD19-Cre mice and B cells from CD19-Cre mice

(top). Summary of the corresponding median fluorescence intensities (MFI) (bottom). (B) Flow cytometric analysis of ex vivo B cells from relbfl/flnfkb2fl/flCD19-Cre and CD19-Cre mice for β2 expression (left) and summary of the corresponding MFIs (right). (C) Flow cytometric analysis of ex vivo B cells from relbfl/flnfkb2fl/flCD19-Cre and CD19-Cre mice for CD21 expression (left) and summary of the corresponding MFIs (right). eGFP+ and eGFP– identifies relb/nfkb2-deleted and non-deleted B cells from relbfl/flnfkb2fl/flCD19-Cre mice, respectively. Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05; **, P<0.01; ***, P<0.001).

94

95

This page has been intentionally left blank.

96

CHAPTER 3:

Activation of the alternative NF-Bpathway in human GC B cells

97

Section 3.1: INTRODUCTION

Although the NF-B signaling cascade has been extensively studied within the context of mature B cell development, very little is understood about the role of this cascade in GC development. As described in a Section 1.5, recurrent, activating mutations in both canonical and alternative NF-B pathway components are observed in various GC-derived B cell malignances (summarized in ref.335). These mutations are suggestive of a physiological role of the NF-B cascade in GC B cells that is hijacked during lymphomagenesis. However, initial gene expression profiling studies revealed that NF-B signature genes are not expressed in GC

B cells, which was thought to reflect the pro-apoptotic nature of these cells156,161. In addition, a

CD40 signaling signature, which includes several NF-B target genes, as CD40 activates the

NF-B signaling cascade283,284,289,290, was not detectable in gene expression data obtained from purified human GC B cells129. This observation was surprising since the importance of the

CD40/CD40L interaction during the GC reaction was well established: first, GCs do not develop in mice or humans with nonfunctional CD40 or CD40L380; second, established GCs collapse upon CD40L inhibition381; and third, CD40 stimulation can prevent the rapid apoptosis of GC B cells ex vivo157. Immunofluorescence analysis of human lymphoid tissue confirmed that the vast majority of GC B cells are not subject to active NF-B signaling as evidenced by the predominantly cytoplasmic rather than nuclear localization of NF-B subunits within these cells129. Nuclear localization of NF-B subunits was however detectable in a subset of LZ B cells129. Indeed, this finding was supported by recent gene expression profiling analyses of mouse and human LZ B cells that revealed the presence of CD40, BCR and NF-B gene signatures119,124. As LZ B cells are subject to T cell-mediated selection, these findings suggest that while NF-B activation is silent in the bulk of GC B cells, it is induced in a subset of LZ B cells that receive activation signals during T cell-mediated selection. These activation signals are likely to include both CD40 and BCR stimulation, which converge on NF-B activation291,292,

98 resulting in the nuclear translocation of NF-B subunits. In the case of CD40 stimulation, this pattern of activation is consistent with the localization of T cells only within the LZ of GCs. Of note, only the canonical subunits c-REL and RELA were tested in the immunofluorescence analyses described above129. Therefore, the expression and activation pattern of alternative NF-

B subunits within the GC remained unknown.

Following this initial description of NF-B activation in LZ B cells, Saito et al. defined a signaling axis in human Burkitt lymphoma cell lines, in which CD40-mediated activation of NF-

B led to the repression of BCL6 expression203. In a step-wise series of events, activated canonical NF-B subunits were found to directly transactivate IRF4, and, in turn, IRF4 bound to the promoter of BCL6 leading to its transcriptional repression203. These signaling events were also observed upon CD40 stimulation of native human GC B cells and may therefore represent an important pathway by which the GC program, which is directed by BCL6267, can be extinguished. This presumably occurs during the selection of LZ B cells, in order to allow GC exit. Since NF-B activation can be induced by BCR signaling, BCR signals may also activate this signaling axis during selection. BCR signals are also able to extinguish BCL6 expression via

MAPK-mediated phosphorylation of BCL6, which leads to its ubiquitin-mediated proteasomal degradation382. The NF-B-IRF4-BCL6 signaling axis, in addition to extinguishing the GC program, is also likely to promote PC development. As mentioned in Section 1.4, IRF4 drives

PC development200,201 and loss of BCL6 would be predicted to relieve BCL6-mediated transcriptional repression of BLIMP1149, another key regulator of the PC fate207. In turn, BLIMP1 would further reinforce silencing of BCL6 expression and the B cell program190,211. Additional evidence that is supportive of this NF-B-IRF4-BCL6 signaling axis include the coexpression of

IRF4 and BLIMP1 in BCL6-negative LZ B cells205,383. However, GCs are not expanded in mice with GC B cell-specific ablation of IRF4200, which would be a possible consequence of IRF4- deficiency in these cells due to the unrestricted expression of BCL6. It therefore appears that

99 the NF-B-IRF4-BCL6 axis identified by Saito et al. may promote GC exit and PC development in some, but not all, contexts of GC development.

The body of literature described above provides evidence that the canonical NF-B pathway is active in LZ B cells. It appears likely that the alternative NF-B pathway is also activated during the GC reaction, as CD40 stimulation is known to induce both the canonical and alternative NF-B pathways and appears to be a critical signal during B cell selection in the

LZ. In addition, as described in Section 1.5, NF-B mutations that activate the canonical pathway are predominantly associated with ABC-DLBCL, while MM is associated with mutations that activate both the canonical and alternative NF-B pathways318,319,335. Since ABC-DLBCL is believed to originate from the malignant transformation of a LZ B cell that has initiated plasmablastic differentiation256, whereas the cell of origin of MM is a bona fide PC316, the mutational pattern suggests the possibility that the canonical and alternative NF-B pathways may have distinct physiological functions in GC and post-GC cell types. These distinct functions may explain the selection of genetic mutations in either canonical or alternative NF-B pathway components in B cell malignancies of different cellular origin. For these reasons we assessed the expression and activation pattern of both the canonical and alternative NF-B subunits in normal GC B cells. Similar to the canonical NF-B subunits, nuclear translocation of alternative

NF-B subunits was observed following CD40 stimulation of human GC B cells. A unique expression pattern of alternative vs. canonical NF-B subunits was revealed within human PCs.

We next attempted to identify specific gene targets of the alternative NF-B pathway via chromatin immunoprecipitation (ChIP) and gene expression profiling in various human cellular systems. These experiments were however unsuccessful due to technical limitations.

Section 3.2: MATERIALS AND METHODS

Isolation of GC B cells from human tonsil

100

Tonsils were obtained from children undergoing routine tonsillectomies at the Babies and

Children’s Hospital of Columbia-Presbyterian Medical Center. Residual material was obtained after diagnosis and was exempt from informed consent as samples were fully anonymized.

Approval from the Institutional Review Board (IRB) was obtained for all procedures. Biopsies were transferred to ice immediately upon surgical removal and tonsillar mononuclear cells were obtained via mincing of the tissue in RPMI medium followed by Ficoll-Isopaque density centrifugation. Human GC B cells were isolated via magnetic labelling of CD77+ cells followed by magnetic cell separation using the MidiMACS system (Miltenyi Biotec).

Cell culture

Human GC B cells and the CD40L-expressing mouse feeder cell lines157 were cultured in RPMI with 10% FBS. The P3HR1, SUHDL2 and JJN3 cell lines were cultured in Iscove’s Modified

Dulbecco’s Medium (IMDM) with 10% FBS. The U266 line was cultured in IMDM with 20% FBS.

LY10 was cultured in IMDM with 15% human serum obtained from the New York Blood Center.

Immunoblot analysis

Cell lines or human GC B cells were subjected to NP40-based lysis, separated by SDS-PAGE and blotted on nitrocellulose membranes (GE Healthcare). Samples were incubated with the following primary antibodies overnight at 4°C. From Santa Cruz: rabbit anti-RELB (clone C-19), rabbit anti-BCL6 (clone N-3), rabbit anti-RELA (clone C-20), and rabbit anti-c-REL (clone C).

From Cell Signaling Technologies (CST): rabbit anti-p100/p52, mouse anti-Phospho-IB(clone

5A5), rabbit anti-IB, rabbit anti-p105/p52. From Sigma, mouse anti-β-actin (clone AC-15) and from Abcam, rabbit anti-RELA (clone ab7970). Horseradish peroxidase-conjugated secondary antibodies (anti-rabbit IgG, GE Healthcare and anti-mouse IgG, Thermo Scientific) and ECL

Western Blotting Substrate or SuperSignal West Dura Extended Duration Substrate (Thermo

Scientific) were used for detection.

101

Immunofluorescence

For single cell staining, cells were spun onto slides using a cytospin machine and fixed in 10% formalin for 20 minutes followed by an additional 20 minutes of methanol fixation. Nuclear permeabilization was achieved via incubation with 0.2% triton in PBS. Cytospin slides were incubated overnight with each of the antibodies specific for NF-B subunits followed by incubation with a secondary anti-rabbit Cy3 antibody (Cy3-conjugated AffiniPure Donkey anti- rabbit IgG; Jackson ImmunoResearch Laboraties). During slide mounting, slides were counterstained with DAPI (Prolong Gold antifade reagent with DAPI; Molecular Probes by Life

Technologies). The images were acquired by means of an Eclipse E400 microscope (Nikon) with a Spot2 charge-coupled device camera (Diagnostic Instruments) mounted on top. Images were acquired in monochromatic JPEG format prior to merging as two-color images in

Photoshop (Adobe Systems).

Chromatin immunoprecipitation (ChIP)

Cells were subjected to cross-linking via incubation with 1% formaldehyde for 10 minutes at room temperature. The reaction was stopped with glycine. The following steps were completed via the use of buffers and reagents provided by the ChIP-IT Express Kit (Active Motif). Nuclei were released following resuspension in lysis buffer followed by dounce homogenization.

Chromatin was then sonicated in a Bioruptor sonicator (Diagenode) for 15-30 minutes, depending on the cell type, in order to achieve sheared fragments of 200-1000bp. Sheared chromatin was used in ChIP reactions with the anti-RELA, anti-c-REL and anti-RELA antibodies detailed above together with protein G magnetic beads and immunoprecipitated chromatin was isolated using a magnetic stand. Immunoprecipitated chromatin was then subjected to cross-link reversal and isolated DNA was analyzed via PCR or quantitative PCR (qPCR) for various target genes. Primer sets were designed to flank NF-B “B” binding sites in putative target genes identified using a web based transcription factor binding site search tool, available at:

102 http://diyhpl.us/~bryan/irc/protocol-online/protocol-cache/TFSEARCH.html. The forward (F) and reverse (R) primer pairs used were: BCL6 (F: gccacgtagcagtggtaaag, R: tgtccggcctttcctagaaac),

BFL1 (F: agctgggattacaggcacac, R: agcactttgggaggacaagg), CD44 (F: cttgcttgggtgtgtccttc, R: aaacttgtccatggtgtccggag), CCR7 (F: actgagcagagcaagggaag, R: cagaagaaaggtgacagagggtg),

CXCR5 (F: caagtgccctaacttgccag, R: cacgcaaagtcagcaggttg), IBa (F: agaaactccctgcgatgagc,

R: agggacaggattacagggtg), β-ACTIN (F: tgtaatcccagctactcaggagg, R: tgtgttagccaggatggtctcg),

BCL2 (F: gcatcacagaggaagtagactg, R: gctttgcattcttggacgagg), IRF4 (F: agctcaacgggtgaaagctcag, R: tacttgagaggacgagccacgag) and PDX1 (F: agcacttgcaaatgctggctc,

R: tctaactgtgaccaaac). shRNA knockdown

Hairpins were designed against the mRNA sequences of RELA, c-REL and RELB using a shRNA design algorithm384. These hairpins were:

RELA 1: tgctgttgacagtgagcgcccgctgcatccacagtttccatagtgaagccacagatgtatggaaactgtggatgcagcggttgcctact gcctcgga RELA 2: tgctgttgacagtgagcgatcagcgcatccagaccaacaatagtgaagccacagatgtattgttggtctggatgcgctgactgcctact gcctcgga RELA 3: tgctgttgacagtgagcgacaggcgagaggagcacagatatagtgaagccacagatgtatatctgtgctcctctcgcctggtgcctact gcctcgga RELA 4: tgctgttgacagtgagcgcaaggacatatgagaccttcaatagtgaagccacagatgtattgaaggtctcatatgtccttttgcctactgc ctcgga RELA 5: Tgctgttgacagtgagcgaaaggtgcagaaagaggacatttagtgaagccacagatgtaaatgtcctctttctgcaccttgtgcctact gcctcgga c-REL 1: tgctgttgacagtgagcgaccgtatatagagataattgaatagtgaagccacagatgtattcaattatctctatatacgggtgcctactgc ctcgga c-REL 2: tgctgttgacagtgagcgaaagctattattacaagaataatagtgaagccacagatgtattattcttgtaataatagcttctgcctactgcct cgga c-REL 3: tgctgttgacagtgagcgaaaagttcagaaagatgacatatagtgaagccacagatgtatatgtcatctttctgaactttgtgcctactgc ctcgga c-REL 4:

103 tgctgttgacagtgagcgcgagaattacattagtaacaaatagtgaagccacagatgtatttgttactaatgtaattctcatgcctactgcc tcgga c-REL 5: tgctgttgacagtgagcgctacggcaataaagcaaagaaatagtgaagccacagatgtatttctttgctttattgccgtaatgcctactgc ctcgga RELB 1: tgctgttgacagtgagcgacaggagcacagatgaattggatagtgaagccacagatgtatccaattcatctgtgctcctggtgcctact gcctcgga RELB 2: tgctgttgacagtgagcgataggttggttgttcagagtcttagtgaagccacagatgtaagactctgaacaaccaacctactgcctactg cctcgga RELB 3: tgctgttgacagtgagcgccgacgagtacatcaaggagaatagtgaagccacagatgtattctccttgatgtactcgtcgatgcctactg cctcgga RELB 4: tgctgttgacagtgagcgcccatcaggaagtagacatgaatagtgaagccacagatgtattcatgtctacttcctgatggttgcctactgc ctcgga RELB 5: tgctgttgacagtgagcgcagaggacatatcagtggtgtttagtgaagccacagatgtaaacaccactgatatgtcctctttgcctactgc ctcgga

These hairpins were cloned into a viral vector (pMSCV-mir30-Luci.1309-PIG (MLP))385 and transfected using polyethylenimine into an amphotropic retroviral packaging cell line385 to create retroviral particles. These viral particles were used to infect cell lines via three rounds of spin infection in the presence of polybrene. Successfully infected cells were then selected using puromycin.

Section 3.3: RESULTS

Expression and activation of the alternative NF-B subunits in human GC B cells

To assess the activation of the alternative NF-B pathway within the context of human GC B cells we began by stimulating a human cell line with CD40 in vitro. P3HR1 is a Burkitt lymphoma cell line that was used to define the NF-B-IRF4-BCL6 signaling axis described in the introduction to this chapter203 (Section 3.1). Since P3HR1 is believed to originate from a DZ B cell256, NF-B signaling is expected to be quiescent in this cell line. Indeed, in the absence of stimulation, minimal phosphorylation of IBα, a readout of canonical NF-B activation, was observed (Fig. 3.1A). Accordingly, the canonical NF-B subunits RELA, c-REL and p50 were

104 present in a cytoplasmic, inactive state (Fig. 3.1C). In regards to the alternative NF-B pathway, expression of the p100 precursor itself was not detectable via Western blot analysis (Fig. 3.1B).

The alternative NF-B subunits RELB and NF-B2 (p100/p52) were detected in the cytoplasm by immunofluorescence analysis, confirming a quiescent signaling state (Fig. 3.1C). CD40- stimulation of the P3HR1 cell line was achieved via co-culture with a feeder cell line engineered to express CD40L157. 6h and 24h of stimulation resulted in the accumulation of phosphorylated

IBα (Fig. 3.1A), and after 24h of stimulation, nuclear translocation of canonical NF-B subunits was observed (Fig. 3.1C), in accordance with the results presented by Saito et al.203. 6h of

CD40-stimulation resulted in both the expression of the p100 precursor and its processing to p52, demonstrating activation of the alternative NF-B pathway (Fig. 3.1B). This activation was further increased following 24h of stimulation, and nuclear translocation of the alternative NF-B subunits was observed (Fig. 3.1B,C). The activation of the alternative pathway was accompanied by downregulation of BCL6 expression, as predicted by the NF-B-IRF4-BCL6 signaling axis described in Section 3.1203.

We next determined whether alternative NF-B activation was also detectable in CD40- stimulated native GC B cells. Native GC B cells can be isolated from human tonsils obtained during routine tonsillectomies performed on children experiencing recurrent infections. Freshly isolated tonsillar GC B cells were devoid of activation of the alternative NF-B pathway, as evidenced by the presence of unprocessed p100 (Fig. 3.2A; left panel). Following 24h of CD40- stimulation, p100-p52 processing was observed, which was accompanied by BCL6 downregulation (Fig. 3.2A; left panel). Accordingly, the p52 subunit along with the canonical NF-

B subunit p50 translocated into the nucleus of GC B cells (Fig. 3.2A; right panel). These results demonstrate that the alternative NF-B pathway is activated in human GC B cells in response to

CD40-stimulation. CD40-stimulation is a signal thought to be important during the selection of

105

LZ B cells; our data suggest that both the alternative and canonical NF-B pathways may be activated in LZ B cells during selection.

To further investigate the expression pattern of canonical and alternative NF-B subunits in lymphoid tissue we performed immunofluorescence analysis on human tonsil sections (Fig.

3.2B, data generated by Amanda Carette, a technician in the Klein lab). In the tonsillar subepithelium where PCs reside, strong cytoplasmic staining of the alternative NF-B subunit p100/p52 (NF-B2) and weak staining of the canonical NF-B subunit p105/p50 (NF-B1) was observed in PCs, which were identified by BLIMP1 expression (Fig. 3.2B, top panel). In contrast, strong cytoplasmic staining of the canonical NF-B subunit p105/p50 and weak staining of the alternative subunit p100/p52 was observed in surrounding lymphocytes (Fig.

3.2B, top panel). The same trend was also evident in the LZ of GCs, as BLIMP1+ PC precursors demonstrated strong cytoplasmic staining of the alternative subunit p100/p52 and weak staining of the canonical subunit p105/p50 (Fig. 3.2B; bottom panel). Weak staining of the canonical subunit c-REL was also observed in PCs and PC precursors (data not shown). These data suggest that the expression of the alternative NF-B subunit p100/p52 is upregulated in PCs and their precursors, whereas the expression of the canonical NF-B subunit p105/p50 is downregulated.

We were also able to identify differential activity of canonical vs. alternative NF-B subunits when analyzing subunit expression in selected ABC-DLBCL and MM cell lines. The

ABC-DLBCL lines SUDHL2 and LY10 have mutations that constitutively activate the canonical

NF-B pathway (mutations in MYD88, CD79A, amplification of REL and TNFAIP3 deletions242), whereas the MM lines U266 and JJN4 have mutations in alternative NF-B pathway components that result in dual activation of the canonical and alternative pathways (mutations in

NIK and TRAF3)319,318. In the MM cell lines, we observed strong processing of p100 to p52, denoting alternative NF-B activation, and weak activation was observed in the ABC-DLBCL

106 cell lines (Fig. 3.2C), as previously reported311,319. No significant expression differences in the alternative subunit RELB and the canonical subunits RELA and p50 could be detected among the cell lines. Strikingly however, c-REL protein levels were virtually absent in the MM lines, whereas robust expression could be detected in the ABC-DLBCL lines (Fig. 3.2C).

Collectively, the data presented in Figures 3.1 and 3.2 demonstrate that the subunits of the alternative NF-B pathway are activated in response to CD40 stimulation, and there appears to be preferential expression of alternative over canonical subunits in PCs and PC precursors, which could also be identified in MM cell lines. This differential expression NF-B subunits could be suggestive of distinct functional roles for the canonical vs. alternative pathways during the GC reaction and in post-GC development.

Attempts to identify NF-B target genes

The identification of the target genes activated by alternative NF-B subunits in GC B cells is an important step towards understanding the downstream functional events that occur following activation of this pathway during the GC reaction. In light of the apparent preferential expression of an alternative NF-B subunit in PC and PC precursors, we were interested in identifying target genes of both the canonical and alternative NF-B subunits in order to detect potential differential targets. As the ABC-DLBCL cell lines presented in Fig. 3.2C demonstrated preferential expression and activation of the canonical NF-B subunits242, these cell lines may provide a cellular system suitable for the identification of specific canonical NF-B target genes.

Similarly, since the MM lines shown in Fig. 3.2C demonstrate dual activation of the canonical and alternative pathways (Fig. 3.2C and ref.319) and low expression of the canonical subunit c-

REL (Fig. 3.2C), these cell lines may prove amenable to the identification of specific alternative

NF-B targets. ChIP sequencing (ChiP-seq) in combination with gene expression profiling represents a suitable approach to identify transcriptionally active, direct NF-B target genes.

107

We sought to assess the robustness of our ChIP protocol using an anti-BCL6 antibody that had been successfully used in ChIP assays on the Burkitt lymphoma line P3HR1266. BCL6 modulates its expression via binding to its own promoter, thus repressing transcription266. We assayed for this interaction following immunoprecipitation (IP) of cross-linked P3HR1 chromatin with the validated anti-BCL6 antibody and were able to detect binding of BCL6 to the BCL6 promoter at 6-fold enrichment over binding of a control IgG to this region (Fig. 3.3A). When calculated over binding of the anti-BCL6 antibody to the β-ACTIN locus, a fold enrichment of 12- fold was detected (Fig. 3.3A; bottom panel). These experiments demonstrated the functionality of our ChIP protocol.

NF-B antibodies specific for the three NF-B subunits that possess TAD domains and thereby directly drive transcription – RELA, c-REL and REL – were utilized in subsequent ChIP experiments. These antibodies were previously used by Saito et al. in Burkitt lymphoma cell lines to elucidate the NF-B-IRF4-BCL6 signaling axis described in Section 3.1203. ChIP experiments were conducted on CD40-stimulated P3HR1 to replicate the ChIP assays presented by Saito et al. and assayed for binding to an NF-B “B” binding site in the IBα loci, which had been used as a positive control region that was strongly bound by all NF-B subunits tested. RELA was found to bind to the IBα loci at ~4-fold enrichment over binding of a control

IgG to the same region (Fig. 3.3B). In addition, binding of RELA to another predicted NF-B target gene that was shown to be induced by CD40 stimulation in B cells, BFL1386, was also tested by designing primers to flank predicted B sites within the gene that were identified by a web-based transcription factor binding site predictor. Binding of RELA to BFL1 was detected at a fold enrichment of ~2-fold over control IgG (Fig. 3.3B). Binding to β-ACTIN was also assayed as a negative control and a fold enrichment of ~5-fold was detected over IgG (Fig. 3.3B).

Perhaps unexpectedly, this locus contained several putative B binding sites; however, binding by RELA was detected despite designing primers around these sites. Since binding of greater

108 than 10-fold over a negative control is preferred before the submission of ChIP samples for deep sequencing, the IBα and BFL1 enrichments over IgG were not sufficient, and larger enrichments were also not observed for binding of c-REL or RELB to these loci (data not shown). Similar observations were made in ChIP assays on CD40-stimulated native tonsil GC B cells. Via PCR assay, binding of NF-B subunits to the BFL1 and IBα loci and to the defined

B binding site in the IRF4 loci203 was observed, however, binding to β-ACTIN was also detected (Fig. 3.3C).

In order to remove one variable from these assays - the efficiency and strength of CD40- mediated induction of NF-B subunits into the nucleus - RELA ChIP assays were conducted on two ABC-DLBCL lines that harbor mutations which lead to strong constitutive canonical NF-B activation (LY10 and SUDHL2). High levels of RELA protein were detected upon Western blot analysis of chromatin derived from these cell lines following cross-link reversal, confirming the nuclear presence of the RELA subunit (Fig. 3.3D). However, significant binding of RELA to a range of both defined and predicted targets in either the LY10 or SUDH2 cell lines could not be detected (Fig. 3.3E; top panel). In addition, quantitative analysis of these ChIP reactions revealed essentially no enrichment over binding to the “negative control” target gene PDX1, a pancreatic developmental gene not expressed in B cells which was also found to lack any predicted B binding sites (Fig. 3.3F). This negative control was utilized due to problems with using β-ACTIN as a negative control (Fig. 3.3E; bottom panel). To determine if another anti-

RELA antibody may be more efficient in the ChIP assays, another Western blot test was performed in which RELA was immunoprecipitated from cross-linked CD40-stimulated P3HR1 chromatin with the anti-RELA clone c-20 antibody (Santa Cruz) and an additional anti-RELA clone ab7970 antibody (Abcam), that had been used in a published ChIP analysis of macrophages387 (Fig. 3.4A). Following cross-link reversal, immunoprecipitated chromatin was probed via Western blot and the anti-RELA c-20 antibody appeared to be more efficient at

109 immunoprecipitating RELA than the anti-RELA ab7970 antibody. The Western blot analysis also confirmed that the anti-RELA c-20 antibody was capable of recognizing its corresponding RELA epitope following cross-linking, which can be a concern in ChIP assays as epitopes can be altered via exposure to the cross-linking agent formaldehyde. Neither the anti-RELA c-20 or ab7970 antibody was capable of binding to a range of targets with significant enrichment over binding of these targets by an IgG control or over binding of RELA to PDX1 (Fig. 3.4B-C). For these later tests, an expanded list of target gene candidates was utilized by designing primers flanking predicted B binding sites in several genes that were strongly induced in response to

CD40 stimulation in a Burkitt lymphoma line – CD44, BCL2, CCR7 and CXCR5129.

Finally, the ability of anti-c-REL and anti-RELB antibodies to immunoprecipitate c-REL and RELB, respectively, from cross-linked chromatin was tested, as shown for RELA in Fig.

3.4A. The anti-c-REL antibody was able to immunoprecipitate its target from CD40-stimulated

P3HR1 chromatin (Fig. 3.5A; middle panel). The specificity of c-REL IP was confirmed by the inability to detect RELA or RELB protein (Fig. 3.5A; middle and right panel). Furthermore, no c-

REL protein could be detected following c-REL IP of chromatin from unstimulated P3HR1, where c-REL is localized in the cytoplasm (Fig. 3.1C and Fig. 3.5A; left panel). Similar results were observed using an anti-RELB antibody, which was able to immunoprecipitate its target from chromatin derived from the MM cell line U266 (Fig. 3.5B; left panel), which is characterized by strong activation of the alternative NF-B pathway (Fig. 3.2C). RELB, but not RELA, was detected in this IP (Fig. 3.5B; right panel). Collectively, the Western blot analyses (Fig. 3.4A and

3.5) demonstrate that the NF-B antibodies used were capable of binding their targets in cross- linked chromatin, and that the antibody-target interactions survive the washing and cross-link reversal steps of the ChIP protocol. Despite these data that suggest that the employed ChIP protocols and antibodies were suitable for this assay, we observed no consistent enrichment for a range of known and predicted NF-B targets over negative controls. In the absence of this

110 form of validation, we did not have the confidence to submit any samples for deep sequencing and therefore pursued different approaches for the identification of NF-B target genes.

Knockdown of NF-B subunits in human cell lines

To develop an alternative strategy for identifying NF-B target genes within a human system, we adopted a knockdown strategy. Using a shRNA prediction algorithm, five hairpins were designed to target c-REL, RELA and RELB mRNAs. The efficacy of each hairpin was tested in the P3HR1 cell line. In each case, hairpins that produced effective knockdown of their target protein could be identified (Fig. 3.6). As NF-B signaling in P3HR1 is quiescent, this cell line presented an ideal system to test these hairpins without the risk of inhibiting growth or cell survival. We next tested a selection of effective c-REL hairpins in LY10, an ABC-DLBCL cell line with constitutive canonical NF-B signaling, and effective knockdown of c-REL was observed

(Fig. 3.6B). Knockdown of RELB in the MM cell line U266, which displays constitutive alternative

NF-B activation, was also evident (Fig. 3.6C). Following knockdown, however, this cell line grew poorly. This observation is in accordance with a publication reporting the toxicity of RELB knockdown in U266 that was published while these experiments were being conducted388. Gene expression profiling of ABC-DLBCL cell lines following knockdown of canonical NF-B subunits, and MM cell lines following knockdown of alternative NF-B subunits, appears to be a valid strategy for identifying canonical and alternative NF-B target genes, respectively. This approach is being pursued by another member of the Klein lab using inducible targeting vectors.

Section 3.4: DISCUSSION

NF-B-activating mutations identified in various GC-derived lymphomas suggest a functional role for NF-B signaling during the GC reaction335. Early studies, perhaps surprisingly, found no detectable NF-B activity in the vast majority of normal GC B cells156,161. However, it was later revealed that NF-B activation occurs in a subset of LZ B cells129. In addition, CD40 and NF-B

111 gene signatures were found to be enriched in LZ B cells119,124. It is therefore likely that CD40 stimulation occurs during the selection of high affinity LZ B cells. Via CD40 stimulation of both native and transformed human GC B cells we were able to detect activation of both the canonical and alternative NF-B pathway. Therefore, T cell help afforded to LZ B cells in the form of CD40L-CD40 interactions appears to activate both pathways of the NF-B signaling cascade.

While CD40 stimulation clearly mediates activation of the canonical and alternative NF-

B pathways in LZ B cells, it is possible that additional signaling events contribute to NF-B activation in these cells. BCR signaling is known to specifically activate the canonical NF-B pathway291,292; therefore, antigen binding to the BCR during selection may induce canonical NF-

B activation. However, recent evidence suggests that BCR signaling within murine GC B cells is quiescent, suggesting that the sole function of the BCR during the GC reaction is to enable antigen capture for presentation on MHC-II molecules in order to acquire T cell help125.

Nevertheless, the same study did report a defined window in which active BCR signaling could

125 be detected, specifically in the G2/M phase of the cell cycle . This observation suggests that transient BCR signaling can occur within the GC reaction, which is further supported by the enrichment of a BCR gene signature in the gene expression profile of murine LZ B cells compared to DZ B cells119. Activation of BCR signaling may have a role in testing that the signaling capacity of mutant BCRs has “survived” the SHM process and could potentially also function as a selection signal in LZ B cells. An additional signal occurring in LZ B cells that may be involved predominantly in the activation of the alternative NF-B pathway is stimulation by

BAFF, which is produced within the GC environment by FDCs82. Despite the severe reduction in peripheral B cells observed in mice deficient for BAFF or the BAFF-R, the remaining B cells were capable of forming small GCs70,389,390. However, while GC initiation and formation appeared normal, these GCs were found to involute progressively over time from day 7 post-

112 immunization onwards. In the case of mice deficient for BAFF or treated with a BAFF antagonist, this defect is unlikely to be intrinsic to B cells and rather appears to be a result of perturbed FDC development389,390. However, in the case of BAFF-R deficient mice, the impaired

GC maintenance may be a result of restricted GC B cell proliferation70,389. These studies are limited by the severe B cell defect observed at the T2 stage of B cell development and therefore in essence investigate only the function of the few B cells that have survived in the absence of

BAFF. How these surviving B cells differ functionally from WT B cells prior to GC entry remains unknown. Potentially, the mutant cells may be unable to sustain a GC reaction. In addition, as these studies were entirely based on histological evaluation of GCs, there is no information on the quality or function of the GCs that formed in the mutant mice70,389,390. Determining the function of BAFF signaling during the GC reaction would require conditional deletion of the

BAFF-R in GC B cells. Nevertheless, these studies raise the possibility that BAFF signaling in

GC B cells may be functionally relevant, and could lead to activation of the alternative and potentially also the canonical NF-B pathway.

Immunofluorescence analysis of human tonsil tissue for the expression of the separate

NF-B subunits of the canonical and alternative pathways revealed preferential expression of the alternative NF-B subunit p100/p52 (NF-B2) in PCs and their precursors within the LZ of

GCs. As p100/p52 was associated with cytoplasmic localization, this suggests that upon appropriate activation by cell surface receptors, PCs and their precursors may be predisposed to signal through the alternative NF-B pathway. Furthermore, Western blot analysis of ABC-

DLBCL and MM cell lines for NF-B subunits revealed preferential activation of the alternative

NF-B pathway in the MM lines together with a striking downregulation of the canonical NF-B subunit c-REL. The virtual absence of c-REL expression is interesting as the MM lines tested were reported to show active canonical and alternative NF-B signaling. In regards to canonical

NF-B activation, this designation was made based on sensitivity to an IKKβ inhibitor and via

113 the nuclear presence of RELA and p50 in these cell lines; c-REL localization was not analyzed318,319. The functional implications of the absence of the c-REL subunit in MM lines are unclear but suggest that this subunit may be dispensable in PC-derived cell lines. The role of c-

REL during the GC reaction was recently investigated by the Klein lab338 and will be discussed in Chapter 4.

A preferential role for the alternative NF-B pathway in PC biology is supported by additional reports in the literature. Gene expression profiling of human B cell subpopulations revealed that PCs transcriptionally downregulate the expression of the canonical NF-B subunits (c-REL and p105/p50) and upregulate that of the alternative NF-B subunits (RELB and p100/p52)391 in comparison to B cells. In addition, gel shift studies found that nuclear RELB- p52 heterodimers are preferentially found in mouse plasmacytoma cell lines (mouse PC tumor cell lines) in comparison to mouse B cell lines392. In plasmacytoma cell lines, RELB-p52 heterodimers were found in addition to canonical NF-B heterodimers; in contrast, mouse B cell lines were characterized by the presence of only canonical NF-B heterodimers392. Finally, an important role for the alternative NF-B pathway in PCs is supported by the selection of mutations that lead to activation of both the canonical and alternative pathways in MM, whereas

ABC-DLBCL is predominantly associated with mutations that activate the canonical NF-B pathway335. These findings collectively suggest that the alternative NF-B pathway may be preferentially activated in late B cell development and may promote a certain aspect of PC biology.

The emerging differential expression of canonical vs. alternative NF-B subunits is suggestive of distinct functions at distinct developmental stages. This raises the possibility that the separate NF-B subunits activate differential target gene expression. In order to identify target genes that are differentially regulated by canonical vs. alternative NF-B dimers, we attempted to establish a ChIP assay to determine the target genes bound by RELA and c-REL

114 in ABC-DLBCL lines with preferential canonical NF-B activation, and the target genes bound by RELB in MM lines with activation of the canonical and alternative NF-B pathways. We were unable to detect binding of NF-B subunits to specific target genes that were identified in a published report203 at a fold enrichment of more than 2-4-fold over negative controls, despite using the same set of antibodies, primer pairs and cellular system. Of note, the report performed

ChIP enrichments using an agarose gel-based PCR strategy rather than quantitative (qPCR).

The fold enrichments for binding to target genes over a negative control were therefore difficult to estimate and thus it remains unclear how their results compared to ours. Although it is possible that we obtained comparable binding to these target genes, this enrichment was considered insufficient for ChIP-sequencing as fold enrichments of >10-fold were required. In an attempt to improve fold enrichments, we assayed several cell lines with constitutively active rather than induced NF-B signaling for the binding of all three transcriptionally active NF-B subunits to a range of known and predicted target genes. No consistent fold enrichments could be reliably detected. The problems were two-fold: enrichments for predicted target genes or target genes described in the literature were generally low, and binding to supposedly negative targets such as β-ACTIN, GAPDH and regions outside the defined B site in the IRF4 locus was readily detectable on agarose gels. In regards to the additional gene targets utilized in our assays that were not published by Saito et al. it is possible that these genes were unsuitable targets for our particular cellular systems or that we designed primers flanking inappropriate regions of the respective loci. Our primer design was guided by computational identification of the “B” binding consensus site 5’-GGGRNYYYCC-3’. In most cases, candidate target genes possess several of these predicted sites, and it was unclear which were the most functionally relevant. Therefore, while several target genes seemed to represent bone fide NF-B target genes as their expression was induced by CD40 in a Burkitt lymphoma cell line129, the actual B sites within the loci to which we designed flanking primers were essentially determined

115 randomly. Western blot analysis of chromatin immunoprecipitated by each of the NF-B antibodies did however demonstrate specific pull down of the expected target proteins, suggesting that the antibodies and ChIP procedure worked satisfactorily.

Attempts to achieve robust fold enrichments were hampered by the limited number of previous studies in B cell systems from which reliable target genes and binding sites could be obtained for validation of the ChIPs. The results of a single report in the literature that defined target sites in a CD40-stimulated Burkitt lymphoma line203 were used as our standard for validation. Relatively few genome wide NF-B ChIP studies can be found in the literature. In addition, most of these publications specifically analyzed the RELA subunit and employed non-

B cell systems under inflammatory stimuli that activate only the canonical NF-B pathway387,393-

399. Only two recent studies focused on B cell systems396,400. One group performed ChIP-seq for all five NF-B subunits in an EBV-transformed lymphoblastoid B cell line with constitutively active canonical and alternative NF-B signaling400; this study revealed several interesting observations. Nearly a third of the identified NF-B sites did not contain the B binding site; instead, several alternative motifs were enriched. In addition, binding sites bound by each of the five NF-B subunits were not significantly different; in other words, NF-B subunit-specific binding sites were not identified. Furthermore, it was found that the majority of target genes were co-occupied by subunits of the canonical and alternative NF-B pathways, suggesting no distinction in the target genes controlled by these pathways. It is unclear how these results that were derived from a cell line with constitutive activation of both the canonical and alternative pathways (and thus have nuclear localization of all five subunits) can be extrapolated to other cellular contents. The question remains as to how functional specificity of the canonical vs. alternative NF-B pathway occurs in different contexts. Specific dimer composition paired with defined B sites appears to be the simplest explanation, although, as outlined above, evidence for this mechanism is lacking. Additional factors such as the binding of co-factors,

116 posttranslational modifications and chromatin configurations may have important roles in directing the transcription of NF-B target genes401. As an alternative strategy for NF-B target gene identification, we pursued knockdown of NF-B subunits in cell lines with constitutive

NFB activation. We identified suitable hairpins that could effectively target RELA, c-REL and

RELB in a cell line without constitutive NF-B activation. Due to the possibility that knockdown of these subunits may be toxic to cell lines that have selected constitutive NF-B activation, the employment of inducible targeting vectors appears necessary. NF-B subunit-specific target genes could be identified by inducibly knocking down each subunit prior to gene expression profiling. This strategy is limited in that it will not allow direct target genes to be conclusively identified, but is likely to provide a range of useful target gene candidates that could be validated in ChIP experiments. Similarly, BAFF-responsive target genes identified in Chapter 2 could be utilized as candidates for ChIP experiments in in vitro BAFF-stimulated mouse B cells.

Another aim of these knockdown studies is to assess the response of cell lines with constitutive

NF-B activation to single NF-B subunit knockdown. Identifying cell lines that are particularly susceptible to the removal of specific NF-B subunits is of interest biologically but may also have therapeutic implications, which will be further explored in the concluding discussion

(Chapter 5). As part of this thesis, the use of the conditional alleles generated in Chapter 2 was adopted to identify either direct or indirect targets of the alternative NF-B pathway in murine

GC B cells; this strategy will be discussed in Chapter 4.

117

Figure 3.1. Activation of the canonical and alternative NF-B pathways in the human

Burkitt lymphoma line, P3HR1, upon CD40 stimulation. (A) Western blot analysis of phosphorylated IBα in CD40L feeder cells, unstimulated P3HR1 cells and P3HR1 cells following co-culture with CD40L-expressing feeders cells for 6 hours (h) and 24h. (B) Western blot analysis of p100/p52 and BCL6 protein expression in P3HR1 cells following co-culture with

CD40L-expressing feeder cells for 1h, 6h and 24h. (C) Immunofluorescence analysis of the cellular localization of canonical and alternative NF-B subunits (red) in unstimulated P3HR1 cells and following 24h of co-culture on CD40L-expressing feeder cells. Slides were co-stained with DAPI (blue).

118

119

Figure 3.2. Expression and activation of canonical and alternative NF-B subunits in normal and transformed human GC B cells and PCs. (A) Human tonsillar GC B cells ex vivo

(0 hours (h)) and following 24h of co-culture on CD40L-expressing feeders, Western blot analysis of p100/p52 and BCL6 (left) and immunofluorescence analysis for p105/p50 and p100/52 (right, red), slides were co-stained with DAPI (blue). (B) Immunofluorescence analysis of tonsil sections for p105/p50 and p100/p52 (red), BLIMP1 (green) and DAPI (blue) in the subepithelium and GC light zone. Data generated by Amanda Carette. (C) Western blot analysis of activated B cell-type diffuse large B cell lymphoma (ABC-DLBCL) and multiple myeloma

(MM) cell lines for p100/p52, RELB, c-REL, RELA and p105/p50.

120

121

Figure 3.3. ChIP analyses in human GC-derived lymphoma cell lines and normal GC B cells. (A) Analysis of binding of BCL6 or IgG control to the BCL6 locus following chromatin immunoprecipitation (ChIP) of P3HR1 chromatin, by agarose gel (top) and qPCR (bottom). (B)

Analysis of binding of RELA or control IgG to IBa, BFL1 and β-ACTIN following ChIP of chromatin from P3HR1 cells stimulated for 24 hours (h) with CD40, by agarose (top) and qPCR

(bottom). (C) Analysis of binding of RELA, c-REL, RELB or IgG control to BFL1, IBa, IRF4 and

β-ACTIN following ChIP of human tonsil GC B cells that were stimulated with CD40 for 24h via agarose gel electrophoresis. (D) Western blot analysis of RELA protein in chromatin from

P3HR1 cells following CD40-stimulation for 24h and in LY10 and SUDHL2 chromatin (E)

Analysis of binding of RELA to PDX1, IBa, IRF4 and BFL1 following ChIP of SUDHL2 and

LY10 chromatin by agarose gel (top) and qPCR (bottom).

122

123

Figure 3.4. ChIP analyses in human GC-derived lymphoma cell lines. (A) Western blot analysis of RELA protein bound by anti-RELA antibodies (clone c-20, Santa Cruz; clone ab7970, Abcam) following ChIP of P3HR1 cells following CD40-stimulation for 24 hours (h).

Equal amounts of chromatin were used in each immunoprecipitation (IP), percentages denote the fraction of each IP or unbound protein loaded on the Western blot gel. (B) Analysis of binding of c-20 or ab7979 anti-RELA antibodies or control IgG to CD44, CCR7, CXCR5, BCL2 and BFL1 by agarose gel following ChIP of SUDHL2 chromatin. (C) qPCR analysis of binding of c-20 or ab7979 anti-RELA antibodies to CD44.

124

125

Figure 3.5. Western blot analyses of ChIPs of human GC-derived lymphoma lines. (A)

Western blot analysis of c-REL, RELB and RELA protein bound by anti-c-REL antibody following ChIP of P3HR1 (left) and P3HR1 following CD40-stimulation for 24 hours (h) (middle and right). (B) Western blot analysis of RELB, c-REL, and RELA protein bound by anti-RELB antibody and control IgG following ChIP of the multiple myeloma (MM) cell line U266. Equal amounts of chromatin were used in each immunoprecipitation (IP) and equal fractions of each

IP were loaded on the Western blot gel.

126

127

Figure 3.6. Knockdown of NF-B subunits in P3HR1 cells. (A) Western blot analysis of c-

REL (top left), RELA (top right) and RELB (bottom) protein expression in P3HR1 cells following infection with 5 c-REL, RELA or RELB-specific shRNAs and a shRNA specific for Renilla which served as a negative control. (B) Western blot analysis of c-REL protein expression in LY10 in uninfected cells, and following infection with c-REL-specific shRNAs and a Renilla-specific shRNA. (C) Western blot analysis of RELB protein expression in U266, in uninfected cells and following infection with RELB-specific shRNAs and a Renilla-specific shRNA.

128

129

This page has been intentionally left blank.

130

CHAPTER 4:

Maintenance of the GC reaction requires signaling through the alternative NF-B pathway

131

Section 4.1: INTRODUCTION

Activating mutations in GC-derived lymphomas that result in constitutive activation of NF-B are suggestive of a functional role for this signaling cascade during normal GC development335. This putative function is likely to be associated with LZ B cells, as human studies have revealed that

NF-B activation within the GC is only detectable in a subset of LZ B cells129. To specifically dissect the roles of the canonical subunits c-REL and RELA during the GC reaction, conditional alleles were generated338. Using a Cre recombinase that allows deletion in GC B cells from day

2 of the GC reaction onwards, it was found that conditional ablation of c-REL induces the progressive collapse of established GCs past day 7 of the GC reaction338. This collapse could not be rescued by a BCL2 transgene and was therefore not caused by the defective delivery of pro-survival signals in the absence of c-REL. Instead, c-REL appears to be required for the establishment of a metabolic program that enables cell growth prior to cell division338. DZ and

LZ B cell fractions were equally lost upon GC B cell-specific deletion of c-REL, suggesting that c-REL is essential for the maintenance of the GC reaction. In contrast, RELA appears to be dispensable for GC development and instead has a role in PC development, as PCs were greatly reduced in mice following ablation of RELA in GC B cells338. This PC defect may be due at least in part to RELA-mediated regulation of BLIMP1 expression338.

Interestingly, many parallels are evident when comparing the activity of c-MYC and NF-

B in the GC reaction. Both NF-B and c-MYC gene signatures were found to be enriched in LZ

B cells119,124, and NF-B gene signatures were also specifically enriched in c-MYC+ GC B cells162,163. Evidence strongly suggests that c-MYC+ GC B cells are undergoing T cell-mediated selection as they displayed an activated phenotype, were actively cycling and expressed high affinity BCRs162,163. As described in Section 1.3, positive selection of a high-affinity B cell within the LZ programs reentry to and division within the DZ, and T cell help was found to induce c-

MYC expression in LZ B cells prior to cyclic reentry into the DZ162,163. Similar to c-REL, ablation

132 of c-MYC also resulted in the dissolution of established GCs162,163,338. As c-MYC was found to be induced prior to cyclic reentry, GC collapse upon removal of c-MYC suggests that cyclic reentry sustains the population of cells proliferating within the DZ and thereby maintains the GC reaction over time. Therefore, and by analogy, it is possible that impaired GC maintenance in the absence of c-REL may be due to a role for c-REL in licensing antigen-selected LZ B cells to grow and divide upon reentry into the DZ. This is consistent with the impairment of the GC reaction observed in the absence of c-REL manifesting only after day 7 of the GC reaction, the time-point at which it is believed that the GC has reached maturity and is polarized into a DZ and a LZ that can support cyclic reentry88,107.

To date, the role of the alternative NF-B pathway during the GC reaction has been studied using mice with constitutional knockout of either relb or nfkb2. These mice fail to form

GCs following immunization, or do so in a highly impaired fashion360-362. However, experiments with bone marrow chimeras demonstrated that this GC defect was due to the absence of RELB or NF-B2 in non-hematopoietic, stromal accessory cells360,362. Mice with germline inactivation of upstream regulators of the alternative NF-B pathway, NIK or IKKα, also fail to form

GCs303,349,350. In contrast to RELB or NF-B2 however, inactivation of NIK or IKKα in hematopoietic cells was found to inhibit GC development303,349,350. These findings are again suggestive of additional roles for NIK and IKKα independent of activation of the alternative pathway that may be important for GC development290,324-327,358, or may be indicative of redundancy between RELB and NF-B2.

GCs in mice that received RELB- or NF-B2-deficient bone marrow were analyzed at day 9 or 10 following immunization, which implies that these subunits are not required for GC development at least up to 10 days post-immunization360,362. However, the occurrence of mutations identified in GC-derived lymphomas that result in constitutive activation of both the canonical and alternative NF-B pathways335, the activation of the alternative pathway observed

133 in human GC B cells and the preferential expression of alternative subunits observed in PCs

(Fig. 3.2) all provided us with an impetus to investigate the role of the alternative NF-B pathway within the GC using conditional alleles. We reasoned that the initial studies followed

GC development only up to day 10 post-immunization, thus preventing the detection of a later

GC maintenance defect. Finally, it is possible that the RELB and NF-B2 subunits have redundant functions, and an effect on GC development may be uncovered only upon combined ablation, an experiment which has not previously been conducted using constitutional knockout mice. For these reasons, we utilized the relb and nfkb2 conditional alleles generated in Chapter

2 to determine the consequences of the GC B cell-specific ablation of RELB and NF-B2 either alone or in combination. These studies revealed that GC maintenance is dependent on the combined activity of RELB and NF-B2. In contrast, single ablation of either subunit had no effect. In addition, while GC B cell-specific deletion of nfkb2 did not appear to impair PC development, reduced antigen-specific serum IgG1 and IgG1 ASCs was observed.

Section 4.2: MATERIALS AND METHODS

Mice

The conditional relb and nfkb2 alleles are described in Chapter 2 and were backcrossed to

C57BL/6 mice (n≥ 7). Cγ1-Cre mice have been described402. Mice were housed and treated in compliance with the US Department of Health and Human Services Guide for the Care and Use of Laboratory Animals and according to the guidelines of the Institute of Comparative Medicine at Columbia University (New York). The animal protocol was approved by the Institutional

Animal Care and Use Committee (IACUC) of Columbia University.

Immunization

Mice were immunized via intraperitoneal injection with 109 sheep red blood cells (SRBCs,

Cocalico Biologicals) in PBS, or NP-KLH (100 µg per mouse; Biosearch Technologies) in

134 complete Freund’s adjuvant (Sigma-Aldrich). Peripheral blood and spleens were removed at the indicated time points for analysis.

Flow cytometry

Spleen cell suspensions or cultured B cells were stained on ice in PBS/0.5% BSA with the following antibodies, PerCP-conjugated anti-B220 (Biolegend, clone RA3-6B2), Alexa Fluor

700-conjugated anti-CD38 (eBioscience, clone 90), PE-conjugated anti-CD138 (BD

Pharmingen, clone 281-2), APC-conjugated anti-CD86 (eBioscience, clone GL1), PE- conjugated anti-CD95 (BD Pharmingen, clone Jo2), PerCP-eFluor710-conjugated anti-CXCR4

(eBioscience, clone 2B11), biotin-conjugated anti-PNA (Vector Laboratories) followed by streptavidin-APC (BD Pharmingen), APC-conjugated anti-CD36 (Biolegend, clone HM36) and

PE-CF594-conjugated anti-CD19 (BD Horizon, clone 1D3). The cells were washed and analyzed on a FACSCalibur or LSRII (BD). To assess DNA content, cells were fixed using the

BD Cytofix/Cytoperm Fixation Permeabilization Kit and stained with 3 M DAPI (Molecular

Probes, Invitrogen) prior to analysis on a LSRII (BD). Data were analyzed using FlowJo software.

ELISA and ELISPOT

For ELISA, 96-well immune plates (Thermo Fisher Scientific) were coated with NP9-BSA

(Biosearch Technologies). Mouse serum samples were incubated for 2 h at room temperature.

Standard curves were generated using mouse IgG1 (Southern Biotech). Bound antibodies were detected using AP-conjugated anti–mouse IgG1 antibody (Southern Biotech). Plates were developed with p-nitrophenylphosphate (Southern Biotech) dissolved in substrate buffer. For

ELISPOT analysis, the spleens and bone marrow of NP-KLH–immunized mice were removed at the indicated time points, and single cell suspensions were subjected to hypotonic lysis. The samples were plated overnight on 96-well filtration plates (Millipore) coated with NP25-BSA

(Biosearch Technologies). The cells were plated in 1:2 serial dilutions, starting at 8 × 105 cells

135 for bone marrow-derived samples, and at 105 cells for spleen-derived fractions. NP-specific

IgG1 ASCs were detected by AP-conjugated anti–mouse IgG1 antibody (Southern Biotech).

Plates were developed with nitro blue tetrazolium chloride-5-bromo-4-chloro-3- indolyl phosphate (NBT/BCIP; Roche).

Immunohistochemistry

Sections of 2-3 m in thickness were cut from spleen tissue that was fixed overnight in 10% formalin and embedded in paraffin. Sections were stained with unlabeled anti-BCL6 rabbit antibody (Santa Cruz, clone N-3). Antibodies were incubated overnight at 4°C, stained with anti- rabbit HRP-labeled polymer (Dako) and developed in aminoethylcarbazole (AEC; Sigma).

These slides were then counterstained for IgM by overnight incubation at 4°C with alkaline peroxidase (AP)-conjugated anti-IgM antibody (Southern Biotech). IgM stainings were developed in nitro blue tetrazolium chloride–5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP;

Roche, New York, NY). The images were acquired by means of a Digital Sight camera (Nikon) mounted on a Nikon Eclipse E600 microscope (Nikon).

B cell isolation and culture

Single cell suspensions of mouse spleen were subjected to hypotonic lysis and B cells were purified by depletion of magnetically labeled non-B cells using the MACS B cell isolation kit

(Miltenyi Biotec). Purified B cells from the indicated genotypes were cultured in the presence of:

20 g/mL LPS (Sigma-Aldrich) or 10ng/mL IL5 (R&D systems), 10 ng/mL IL4 (R&D systems) and 12.5 ng/mL CD40L (R&D systems). Cell density in LPS cultures was 1.0 x 106 cells/mL using RPMI with 10% FBS and in CD40 + IL5 + IL4 cultures cell density was 0.5 x 106 cells/mL using RPMI with 15% FBS.

RNA-seq

136 eGFP+ GC B cells (CD95hiPNAhi) were sorted from purified B cells isolated from relbfl/flnfkb2fl/flCγ1-Cre SRBC-immunized mice (day 7) and GC B cells were sorted from purified

B cells isolated from Cγ1-Cre control mice. Cell sorting was conducted on a FACSAria (BD) and cells were sorted into 50% FBS in PBS. Total RNA of ~60,000-400,000 cells per sample was isolated using the Nucleospin RNA XS isolation kit (Macherey-Nagel) according to the manufacturer’s protocol, and RNA integrity >8 was determined using a BioAnalyzer 2100

(Agilent Technologies). ~2-70ng of RNA per sample was submitted to the New York Genome

Center and RNA was amplified via the NuGEN Ovation RNA-Seq System V2 prior to RNA sequencing. 35-40 million 2x50 bp paired-end reads were sequenced per sample on the

HiSeq2500 instrument (Illumina). DESeq analysis was used to identify differentially expressed genes and determine fold expression changes between genotypes.

Gene set enrichment analysis (GSEA)

The GSEA tool403 available from the website of the Broad Institute was used to identify gene signatures enriched in RELB/NF-B2-proficient vs. RELB/NF-B2-deficient GC B cells. We screened the collection of signatures under the category GO_MF and the curated category

Reactome to identify signatures with significant enrichment (FDR < 25%, p ≤ 0.05).

Immunoblot analysis

Purified B cells were subjected to NP40-based lysis, separated by SDS-PAGE and blotted on nitrocellulose membranes (GE Healthcare). Samples were incubated with the following primary antibodies overnight at 4°C. Rabbit anti-p100/p52 (Cell Signaling Technologies, CST), goat anti-

IRF4 (Santa Cruz, clone M-17), rat anti-BLIMP1 (Santa Cruz, clone 6D3) and mouse anti-β- actin (Sigma, clone AC-15). Horseradish peroxidase-conjugated secondary antibodies (anti- rabbit IgG, GE Healthcare, anti-mouse IgG, Thermo Scientific, anti-goat IgG, Santa Cruz and anti-rat IgG, GE Healthcare) and ECL Western Blotting Substrate or SuperSignal West Dura

Extended Duration Substrate (Thermo Scientific) were used for detection.

137

Section 4.3: RESULTS

Combined GC B cell-specific deletion of relb and nfkb2 impairs the GC reaction.

To determine the in vivo role of the alternative NF-B subunits RELB and NF-B2 during GC B cell development, we utilized conditional relb and nfkb2 alleles generated in Chapter 2. These alleles were crossed to Cγ1-Cre mice either alone or in combination to allow deletion in GC B cells402. In Cγ1-Cre mice, the expression of Cre recombinase is linked to the transcription of the

Ig γ1 constant region segment (Cγ1)402. The expression of Cγ1-Cre is therefore induced upon

TD immunization, and the Cγ1-Cre promoter is expressed in activated B cells as early as 2-4 days after immunization. As GCs are the main source of class switched antibodies in response to TD, Cγ1-Cre is efficiently induced in GC B cells402. Expression is induced in the majority of

GC B cells and is not restricted to just those cells that undergo CSR to IgG1402. relbfl/flCγ1-Cre, nfkb2fl/flCγ1-Cre or relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-

Cre control mice were immunized with sheep red blood cells (SRBCs), a TD antigen that induces a robust GC response. 14 days following immunization, the fractions of splenic

CD95hiPNAhi GC B cells in relbfl/flCγ1-Cre and nfkb2fl/flCγ1-Cre were comparable to controls (Fig.

4.1A,B). A trend towards a reduced fraction of GC B cells was observed in nfkb2fl/flCγ1-Cre compared to nfkb2fl/+Cγ1-Cre and Cγ1-Cre control mice; however, this difference did not reach significance (Fig. 4.1B). In contrast, the fraction of splenic GC B cells in relbfl/flnfkb2fl/flCγ1-Cre was strongly reduced in comparison to relbfl/+nfkb2fl/+Cγ1-Cre and Cγ1-Cre control mice 14 days post-immunization (Fig. 4.1C). Consistent with the flow cytometric data, reduced BCL6+ GCs were observed within B cell follicles in relbfl/flnfkb2fl/flCγ1-Cre in comparison to controls at day 14 via immunohistochemical analysis (Fig. 4.2A). These findings therefore demonstrate that single ablation of either RELB or NF-B2 in GC B cells has no significant impact on the GC reaction, in agreement with previous studies that utilized adoptive transfer of bone marrow with constitutional ablation of RELB or NF-B2360,362. Instead, combined ablation of RELB and NF-

138

B2 in GC B cells strongly impaired the GC reaction, demonstrating that both alternative pathway subunits are required for GC maintenance.

To define the temporal kinetics of the GC impairment observed in relbfl/flnfkb2fl/flCγ1-Cre mice, we analyzed CD95hiPNAhi splenic GC B cells at various time-points following immunization with SRBCs. 7 days post-immunization, the fraction of CD95hiPNAhi splenic GC B cells was not significantly different between relbfl/flnfkb2fl/flCγ1-Cre mice and relbfl/+nfkb2fl/+Cγ1-Cre and Cγ1-

Cre controls (Fig. 4.2B). However, reduced fractions of GC B cells were observed from 10-12 days following immunization in relbfl/flnfkb2fl/flCγ1-Cre mice compared to relbfl/+nfkb2fl/+Cγ1-Cre mice (Fig. 4.2C). It appears therefore that GC B cell development occured normally in relbfl/flnfkb2fl/flCγ1-Cre mice up to day 7 of the GC reaction, after which GC B cells were progressively lost. This phenotype is similar to the collapse of formed GCs observed following

GC B cell-specific ablation of c-REL338 or c-MYC162,163 and is suggestive of a defect in GC maintenance in the combined absence of RELB and NF-B2. As observed following GC B cell- specific ablation of c-REL338, a defect in the maintenance of formed GCs would be predicted to result in the equal loss of CXCR4hiCD86lo DZ and CXCR4loCD86hi LZ splenic GC B cell fractions over time. While the differences in the DZ and LZ B cell fractions between relbfl/flnfkb2fl/flCγ1-Cre mice and Cγ1-Cre control mice did reach significance (Fig. 4.3A), these differences were minor and thus do not indicate a preferential loss of either DZ or LZ B cells in relbfl/flnfkb2fl/flCγ1-Cre mice. In addition, the proliferative capacity of GC B cells in relbfl/flnfkb2fl/flCγ1-Cre mice did not vary significantly from controls at day 8-12 post-immunization (Fig. 4.3B), consistent with the almost equal fractions of proliferative DZ B cells observed between relbfl/flnfkb2fl/flCγ1-Cre and control mice (Fig. 4.3A). These data suggest that RELB and NF-B2 are jointly required for the maintenance of both DZ and LZ GC subpopulations, while GC formation appears unaffected.

Identification of genes controlled by the alternative NF-B subunits, RELB and NF-B2, in

GC B cells at day 7 of the GC reaction.

139

In order to identify biological programs controlled by RELB and NF-B2 that are required for GC maintenance, we subjected GC B cells isolated at day 7 post-immunization to RNA-seq analysis. As normal fractions of GC B cells were present in relbfl/flnfkb2fl/flCγ1-Cre mice at day 7

(Fig. 4.2B), this time-point presented an opportunity to harvest GC B cells in relbfl/flnfkb2fl/flCγ1-

Cre mice prior to their loss at later time-points (Fig. 4.1C, 4.2C). We reasoned that gene expression changes that ultimately result in the loss of GC B cells would already be evident at day 7 in relbfl/flnfkb2fl/flCγ1-Cre mice. A similar strategy was successfully utilized previously to identify the genetic program controlled by c-REL in GC B cells338. At 7 days post-immunization with SRBCs, splenic CD95hiPNAhi GC B cells were sorted from 3 Cγ1-Cre control mice and splenic eGFP+CD95hiPNAhi GC B cells were sorted from 2 relbfl/flnfkb2fl/flCγ1-Cre mice. The use of eGFP positivity allowed GC B cells with ablation of RELB and NF-B2 to be specifically isolated, thus excluding GC B cells that escaped Cre-mediated recombination. RNA was extracted from these samples and subjected to RNA amplification, which was required as GC B cells comprise a very small proportion of total spleen, prior to RNA-seq analysis. Reduced transcript counts of relb and nfkb2 were identified in GC B cells from relbfl/flnfkb2fl/flCγ1-Cre mice compared to Cγ1-Cre controls (Fig. 4.4A). In addition, unsupervised clustering analysis revealed that the Cγ1-Cre samples and relbfl/flnfkb2fl/flCγ1-Cre samples clustered into two separate groups as expected (Fig. 4.4B), thereby validating the robustness of the RNA-seq data set.

Differentially expressed sequence analysis (DESeq) of RELB/NF-B2-proficient vs.

RELB/NF-B2-deficient GC B cells identified 59 transcripts with greater than 2.5-fold reduced expression and 84 transcripts with greater than 2.5-fold increased expression at a significance threshold of p<0.01. Genes were assigned to putative functional categories (Fig. 4.4C); for the identity of the corresponding genes, fold-change and p values, see Appendix B. Of note, the genetic program controlled by RELB and NF-B2 was largely mutually exclusive with the transcriptional program controlled by c-REL in GC B cells at day 7338, suggesting the different

140

NF-B transcription factors have distinct functions within the same GC context. Reduced expression of genes involved in apoptosis or proliferation was not detected in RELB/NF-B2- deficient GC B cells compared to controls. In addition, the expression of genes encoding major regulators of GC and post-GC development, including BCL6, AID, c-MYC, IRF4 and BLIMP1, remained intact. A large proportion of transcripts with reduced expression in the combined absence of RELB and NF-B2 encoded genes with functions in the immune response and metabolism (Fig. 4.4C, left panel). In contrast, the vast majority of transcripts with increased expression in the combined absence of RELB and NF-B2 were non-protein coding in nature

(Fig. 4.4C, right panel), comprising a range of RNA elements including lincRNAs, antisense transcripts and pseudogenes (see Appendix B). While miRNAs have been linked to functional roles within the GC404, the biological significance of the upregulated RNA elements identified in

RELB/NF-B2-deficient GC B cells is unclear.

The genes with metabolic functions that displayed reduced expression in RELB/NF-B2- deficient GC B cells compared to controls were primarily found to have functions in lipid and amino acid metabolism (Fig. 4.5A, left and right). Reduced expression of genes involved in ion and carbohydrate metabolism was also observed in RELB/NF-B2-deficient vs. proficient GC B cells (Fig. 4.5A); for the identity of the corresponding genes, fold-change and p values, see

Appendix B. In regards to lipid metabolism, RELB/NF-B2-deficient GC B cells demonstrated reduced expression of genes involved in lipid transport, synthesis and breakdown compared to controls (Fig. 4.5A, left). These included the gene encoding CD36, a molecule with heterogeneous functions in different cell types405. On the cell surface of endothelial cells, CD36 serves a receptor for thrombospondin and negatively regulates angiogenesis405. On phagocytes, functions associated with CD36 include the recognition of lipid molecules on apoptotic cells and pathogens, thereby facilitating uptake405. In phagocytes, CD36 can also function as a fatty acid translocase that mediates the transport of long-chain fatty acids into cells405, and this function

141 was used to designate a possible role in lipid metabolism to this gene (Fig. 4.5A, left). In B cells,

CD36 is understudied. It has been identified as an NF-B target in a murine pre-B cell line406.

CD36 expression is rapidly induced on follicular B cells following in vitro stimulation with LPS and CD40407; however, CD36-deficient B cells respond normally to a range of mitogenic stimuli in vitro408. As CD36 is a cell surface protein, we assayed for the expression of CD36 on eGFP+

GC B cells from relbfl/flnfkb2fl/flCγ1-Cre mice via flow cytometry and identified reduced protein expression in comparison to GC B cells in Cγ1-Cre mice and eGFP+ GC B cells in relbfl/+nfkb2fl/+Cγ1-Cre mice at day 8-12 post-immunization (Fig. 4.5B). The reduced expression at the protein level provides us with additional validation of our RNA-seq data set and may indicate that RELB and NF-B2-deficient GC B cells have a defect in fatty acid metabolism. This notion is further supported by the reduced expression of the genes encoding ceramide synthase

6 (CERS6), an enzyme involved in sphingolipid synthesis409, oxysterol binding protein-like 3

(OSBPL3), an intracellular lipid receptor410, and solute carrier family 27 (SLC27A4), another long-chain fatty acid transporter with acyl-CoA synthetase activity411,412. Reduced expression of genes that have functions in amino acid modifications was also identified in RELB/NF-B2- deficient GC B cells (Fig. 4.5A, right). These included the genes encoding glutathione S- transferase theta 1 and 3 (GSTT1 and GSTT3), detoxifying enzymes that catalyze the conjugation of glutathione to electrophilic compounds413, UDP-GalNAC:Polypeptide N-

Acetylgalactosaminyltransferase-Like Protein 3 (WBSCR17), an enzyme that may catalyze the transfer of N-acetyl-D-galactosamine to a serine or threonine residue414, and peptipdyl arginine deiminase, type IV (PADI4), an enzyme that catalyzes the conversion of arginine residues into citrulline415. Collectively, these gene expression changes suggest that RELB/NF-B2-deficient

GC B cells may have defective lipid and amino acid metabolism. Also, c-REL was found to control a metabolic program in GC B cells; in contrast, however, this program was mainly comprised of genes involved in glycolytic processes338.

142

In comparison to controls, reduced expression of a number of immune response genes was detected in RELB/NF-B2-deficient GC B cells, which comprised immune trafficking, antiviral, inflammation, Ig and B cell-related genes (Fig. 4.5C). Within this immune response category, genes with functions in immune response trafficking are presented in Fig. 4.5C; for the identity of the remaining genes within the immune response category and associated fold- changes and p values, see Appendix B. The immune trafficking genes included cxcl10, encoding CXCL10, an IFNγ-inducible, inflammatory chemokine that binds to the chemokine receptor CXCR3 and is primarily associated with secretion by non-B cell types including fibroblasts, monocytes and T cells416, although expression of CXCL10 in B cells has been reported417,418. CXCL10 serves as a chemotactic factor for a range of cell types, including T cells, DCs and natural killer cells and non-immune cell types416. Additional immune trafficking genes with reduced expression in RELB/NF-B2-deficient were rgs13 and rgs3 (encoding regulators of G protein signaling 13 and 3). Chemokines induce signaling through chemokine receptors that are coupled to G proteins. RGS13 and RGS3 modulate G protein-coupled signaling by chemokine receptors and are associated with inhibiting the responsiveness of chemokine receptors to chemokine ligation419. Several reports implicate RGS13 and RGS3 in B cell and GC B cell trafficking. Rgs13 expression is rapidly induced in activated B cells and is highly expressed in GC B cells420,421. RGS13 and RGS3 have been shown to inhibit signaling through the GC chemokine receptors CXCR5 and CXCR4 following expression in a cell line system420, and reduced expression of RGS13 resulted in elevated responsiveness to the GC chemokines CXCL12 and CXCL13 in a Burkitt lymphoma cell line422. In addition, RGS3 expression in a B cell line also inhibited chemotaxis towards a range of lymphoid chemokines423.

Finally, gene set enrichment analysis (GSEA)403 revealed enrichment of chemokine/chemokine receptor gene signatures in RELB/NF-B2-proficient GC B cells in comparison to RELB/NF-

143

B2-deficient GC B cells (Fig. 4.5D). These data collectively suggest that in the absence of

RELB and NF-B2, GC B cells may respond improperly to trafficking signals within the GC.

PC frequencies following GC B cell-specific deletion of alternative NF-B subunits.

We next sought to determine the in vivo role of these RELB and NF-B2 subunits during post-

GC PC development. relbfl/flCγ1-Cre, nfkb2fl/flCγ1-Cre or relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-Cre control mice were immunized with SRBCs and analyzed for splenic CD138hi PCs 14 days later (Fig. 4.6A-C). A reduced fraction of splenic PCs was observed in relbfl/flCγ1-Cre mice in comparison to relbfl/+Cγ1-Cre mice and Cγ1-Cre controls

(Fig. 4.6A). In contrast, the fraction of PCs in nfkb2fl/flCγ1-Cre was comparable to controls (Fig.

4.6B). As relbfl/flnfkb2fl/flCγ1-Cre mice have a greatly reduced fraction of GC B cells at day 14

(Fig. 4.1C), reduced GC output would be predicted. As expected, the fraction of PCs in these mice was strongly reduced in comparison to controls (Fig. 4.6C). These data demonstrate that in contrast to the effects observed on GC development, GC B cell-specific ablation of RELB alone results in a reduction in PCs. This reduction in PCs is also reflected by reduced NP- specific serum IgG1 in relbfl/flCγ1-Cre mice in comparison to relbfl/+nfkb2fl/+Cγ1-Cre and Cγ1-Cre control mice 14 days following immunization with the TD antigen NP-KLH (Fig. 4.8A).

As nfkb2fl/flCγ1-Cre displayed no PC impairment at day 14 post-immunization, we analyzed CD138hi PCs in the spleen and bone marrow 28 days following immunization with NP-

KLH, and once again PCs frequencies were comparable to controls (Fig. 4.7). However, nfkb2fl/flCγ1-Cre mice were found to have greatly reduced NP-specific serum IgG1 in comparison to nfkb2fl/+Cγ1-Cre and Cγ1-Cre control mice at 14 and 28 days post-immunization with NP-KLH (Fig. 4.8B). This reduction in serum IgG1 was somewhat counterintuitive considering that reduced PC frequencies were not detectable in nfkb2fl/flCγ1-Cre mice (Fig.

4.6B, 4.7). To further investigate this phenotype, we conducted ELISPOT analysis for NP- specific IgG1 ASCs. The results revealed a strong reduction in these ASCs in the spleen and

144 bone marrow of nfkb2fl/flCγ1-Cre mice compared to control mice 14 and 28 days post- immunization with NP-KLH (Fig. 4.8C). Therefore, despite apparently normal GC and PC development (Fig. 4.1B, 4.6B, 4.7), GC B cell-specific ablation of NF-B2 resulted in reduced antigen-specific serum IgG1 and antigen-specific IgG1 ASCs.

As PCs comprise a very small proportion of splenic and bone marrow mononuclear cells, we reasoned that plasmablasts generated in vitro would prove more amenable for use in functional assays to further investigate the reduced antigen-specific serum IgG1 and IgG1 ASCs observed in nfkb2fl/flCγ1-Cre mice. To this end, we isolated NF-B2-deficient B cells from nfkb2fl/flCD19-Cre mice and B cells from nfkb2fl/+CD19-Cre and CD19-Cre control mice in order to perform published assays that promote plasmablastic differentiation. These assays involved 3 days of stimulation with LPS or 4 days of stimulation with CD40L, IL5 and IL4191,424. While it appeared that cultures of NF-B2-deficient B cells had an increased propensity to undergo plasmablastic differentiation in comparison to control cultures in each assay (which was not an intuitive result), the difference did not reach significance (Fig. 4.9A). ELISPOT analysis revealed no significant differences in the generation of IgM ASCs in cultures of B cells from nfkb2fl/flCD19-

Cre, nfkb2fl/+CD19-Cre or control CD19-Cre mice. Therefore, the in vitro plasmablastic differentiation assays did not mimic the reduced ASCs observed in nfkb2fl/flCγ1-Cre mice in vivo

(Fig. 4.9B). These assays, however, had additional limitations. First, no processing of p100 to p52 could be detected following LPS stimulation (Fig. 4.9C), consistent with the fact that LPS is known to activate the NF-B cascade via the canonical pathway. It appears therefore that the

LPS plasmablastic differentiation assay is not suitable for the study of the alternative subunit

NF-B2 in plasmablasts. Second, the CD40 + IL4 + IL5 assay, while demonstrating limited p100

– p52 processing, was associated with poor induction of BLIMP1, a major regulator of PC development (Fig. 4.9C). Therefore, this assay also appears to have only limited value for the

145 study of PC biology. Overall, the currently available in vitro assays are not suitable for the study of NF-B2 in antibody generation/secretion.

Section 4.4: DISCUSSION

The role of alternative NF-B subunits in GC development has been obscured in mice with constitutional ablation of RELB or NF-B2 by phenotypes produced by non-hematopoietic cell types359-362. Previous work using bone marrow chimeras indicated that neither RELB or NF-B2 alone is required in hematopoietic cells to support the GC reaction360,362. In agreement, we found that GC B cell development proceeds normally in mice with GC B cell-specific ablation of either RELB or NF-B2 alone. In contrast, combined ablation of RELB or NF-B2 resulted in the progressive loss of GC B cells over time, a phenotype similar to that observed following ablation of c-REL in GC B cells338. It appears, therefore, that RELB and NF-B2 are jointly required for the maintenance of the GC reaction.

These findings revealed redundancy between the subunits of the alternative NF-B pathway and suggest that the loss of either RELB of NF-B2 alone does not disrupt the functional role of the alternative NF-B pathway within the GC. The redundancy in the absence of either RELB of NF-B2 may be explained by dimerization of the respective, remaining transcription factor with subunits of the canonical NF-B pathway. Indeed, while the classical dimer pair of the alternative pathway is RELB-p52, numerous homo- and heterodimer combinations have been described277,425. It is therefore possible that hybrid alternative/canonical

NF-B dimers, such as RELA-p52 or RELB-p50, are able to functionally compensate for the absence of the classical RELB-p52 dimer of the alternative pathway. The possibility that global dimer compositions are altered in the absence of either RELB or NF-B2 is an interesting point that could be investigated with gel shift assays. Alternatively, this question could be addressed genetically by intercrossing conditional alleles of canonical and alternative NF-B subunits. It is

146 clear however, that redundancy does not exist generally between the canonical and alternative

NF-B pathways, since the GC maintenance defect observed in the combined absence of RELB and NF-B2 is not compensated for by canonical NF-B subunits. This lack of pathway redundancy is further evidenced by the distinct gene expression programs controlled by cREL338 and RELB/NF-B2 at day 7 of the GC reaction.

GCs reach maturity at ~day 7 of the GC reaction, the time-point at which DZ/LZ polarization is thought to be complete and when selection of high-affinity GC B mutants, followed by cyclic reentry, is initiated88,107. It is clear that persistent signals are required for the maintenance of mature GCs, as the involution of established GCs has been observed upon inhibition of CD40381 and BAFF70,389 signaling as well as upon ablation of c-MYC162,163, c-REL338 and, as described here, RELB and NF-B2. In addition, a recent study showed that GC B cells within spontaneous GCs in mesenteric lymph nodes and Peyer’s patches decreased over time following inducible ablation of NIK372. As outlined in the introduction to this chapter (Section 4.1), it is likely that c-REL is activated in LZ B cells undergoing selection and promotes cyclic reentry by enabling cell growth and division upon cyclic reentry into the DZ338. Rounds of selection and cyclic reentry then sustain the GC reaction over time. Similar to c-REL-deficient GC B cells,

RELB/NF-B2-deficient GC B cells were also lost past day 7 of the GC reaction, suggesting that activation of the alternative NF-B pathway occurs within LZ B cells and plays a role during selection. Furthermore, DZ and LZ B cell subsets were lost equally over time. Such equal loss is suggestive of a role for RELB and NF-B2 in promoting cyclic reentry following selection either by ensuring the survival or growth of LZ B cells, or by enabling cycling reentry into and propagation within the DZ. A functional role in selection and cyclic reentry was further supported in the case of c-REL by the finding that c-REL-deficient GC B cells undergo reduced affinity maturation in comparison to controls338. A similar analysis of RELB/NF-B2-deficient GC B cells to determine if the selection of high-affinity BCRs is impaired within these cells would further

147 clarify whether alternative NF-B subunits are also critically involved in maintaining cycles of mutation and selection.

The mechanism by which RELB and NF-B2 promote the maintenance of the GC reaction remains unclear. It is possible that the alternative NF-B subunits are required to maintain the survival of GC B cells in response to BAFF signals within the GC. This would be consistent with the potential role of BAFF signaling in GC maintenance70,389 and the function of the alternative NF-B pathway downstream of BAFF in naïve B cells (Chapter 2). Reduced expression of anti-apoptotic genes was not detected in RELB/NF-B2-deficient GC B cells at day 7; however, this does not formally exclude the possibility that RELB/p52 heterodimers control the expression of anti-apoptotic genes at later time-points. To exclude a role for the alternative NF-B subunits in the survival of B cells within the GC, the ability of a BCL2 transgene to rescue RELB/NF-B2-deficient GC B cells could be assessed. With regard to the role of c-REL in GC B cells, the loss of c-REL-deficient GC B cells was not correlated with increased apoptosis, but rather with the failure of these cells to express a glycolytic metabolism program which likely promotes cell growth and subsequent proliferation338. This c-REL-regulated program is important for the generation of energy and building blocks for anabolic processes, and c-REL-deficient B cells stimulated in vitro displayed reduced oxygen consumption rate, ATP production and glycolytic capacity in comparison to controls338. Reduced expression of genes with metabolic functions was also observed in RELB/NF-B2-deficient GC B cells; however, these genes were primarily found to encode factors that direct fatty acid and amino acid metabolism. It is therefore possible that the two branches of the NF-B signaling cascade direct different types of cellular metabolism that may be jointly required to ensure proper GC B cell growth.

In naïve B cells, NF-B2 was found to control the expression of a number of genes that direct trafficking within B cell follicles (Chapter 2). Similarly, reduced expression of genes

148 involved in the regulation of trafficking within the GC was detected in RELB/NF-B2-deficient

GC B cells in comparison to controls. Dysregulated trafficking would be predicted to disrupt cyclic reentry and thereby prevent DZ reentry and proliferation, events that maintain the GC reaction. A trafficking defect would therefore be predicted to impair GC maintenance, consistent with the phenotype observed upon GC B cell-specific combined ablation of RELB and NF-B2.

However, a putative defect in trafficking within the GC is not reflected by significantly altered fractions of DZ and LZ B cells in relbfl/flnfkb2fl/flCγ1-Cre compared to Cγ1-Cre control mice. In addition, studies focused on the role of rgs13, a gene involved in the regulation of trafficking that is expressed at reduced levels in RELB/NF-B2-deficient GC B cells, present conflicting findings in regards to the function of RGS13 in GC development. Analysis of RGS13-deficient mice revealed enlarged GCs and elevated numbers of GC B cells, suggesting that RGS13 has a role in limiting the GC reaction421. In contrast, in a mouse model of autoimmunity, deletion of rgs13 was associated with smaller GCs, impaired SHM and affinity maturation426. Whether RGS13 has a role in limiting or maintaining the GC reaction is therefore unclear. Proper DZ/LZ trafficking within the GC also ensures the proper segregation of mutation and selection processes to enable effective affinity maturation108. Aberrant trafficking may therefore disturb affinity maturation, which is a testable hypothesis that will be addressed by investigating the capacity of mice with GC B cell-specific ablation of RELB and NF-B2 to generate high-affinity GC B cells.

Deletion of relb in GC B cells resulted in a significant reduction in PCs, suggesting that

RELB has a role in PC development. This defect was however not as severe as that observed following GC B cell-specific ablation of the canonical NF-B subunit RELA, which resulted in greatly reduced PCs338. In contrast, deletion of nfkb2 in GC B cells led to the disappearance of antigen-specific serum IgG1 and antigen-specific IgG1 ASCs, despite the occurrence of apparently normal fractions of PCs in the spleen and bone marrow. It should be noted, however, that similar to the weak expression of eGFP following recombination of the floxed nfkb2 allele

149 with CD19-Cre, eGFP is also weakly expressed in GC B cells and PCs from nfkb2fl/flCγ1-Cre mice. Due to this technical limitation, we were unable to conclusively detect a preferential loss of eGFP+ PCs in nfkb2fl/flCγ1-Cre mice compared to nfkb2fl/+Cγ1-Cre mice, and therefore a possible counterselection of nfkb2-deleted cells cannot be excluded. What is clear, however, is that on a whole population level (assessing eGFP+ and eGFP- PCs), there does not appear to be any reduction in the fraction of PCs within the spleen or bone marrow following GC B cell- specific ablation of NF-B2. This is suggestive of a functional rather than a developmental role for NF-B2 in PC biology, and as such contrasts with the role of RELA in the generation of

PCs338.

The greatly reduced levels of antigen-specific IgG1 in nfkb2fl/flCγ1-Cre mice could have several explanations. One possibility is that GC B cells expressing high affinity, antigen-specific

BCRs are not generated in the absence of NF-B2, thus selectively preventing the development of PCs secreting high affinity antibody. This putative defect could be investigated by assessing affinity maturation at the IgV sequence level within nfkb2fl/flCγ1-Cre mice. Alternatively, CSR to

IgG1 may be impaired in nfkb2fl/flCγ1-Cre mice. This possibility could be addressed by assessing the levels of other antibody isotypes within these mice. However, as B cells from mice with constitutional deletion of nfkb2 were found to be able to switch normally to all Ig isotypes in response to in vitro stimulation361, selective inhibition of CSR to IgG1 appears unlikely. A final possibility is that in the absence of NF-B2, PCs have an impaired ability to secrete antibody. A secretion phenotype would be expected to impact high and low affinity antibodies and antibodies of different isotypes equally. To address this possibility, additional experiments assaying the levels of other types of antibodies in nfkb2fl/flCγ1-Cre mice would be informative. LPS-mediated plasmablastic differentiation was used to demonstrate reduced expression of BLIMP1 in RELA-deficient B cells in comparison to controls, which may explain the PC differentiation defect observed in vivo338. In order to determine whether NF-B2-deficient

150 plasmablasts express reduced levels of factors involved in the secretion of antibody, such as

XBP1210,216-219, a plasmablastic differentiation system would be of use. However, as LPS stimulation does not activate the alternative NF-B pathway, more suitable assays will need to be devised. In the absence of such assays, experiments on native PCs may be required. This strategy is however hampered by the limited numbers of PCs in murine spleen and bone marrow. Nevertheless, determining whether CD138hi PCs express intracellular IgG1 in nfkb2fl/flCγ1-Cre mice by immunohistochemical analysis of splenic and bone marrow sections would be a feasible experiment.

On a final note, the phenotypes observed in mice with constitutive activation of the alternative NF-B pathway in GC B cells, via overexpression of NIK, do not entirely agree with those observed in the absence of the downstream alternative NF-B subunits. Overexpression of NIK had minimal impact on GC B cells315, an observation that contrasts with a possible elevation that would be predicted based on the loss of GCs in mice with GC B cell-specific ablation of RELB and NF-B2. Overexpression of NIK did however promote PC hyperplasia315, in agreement with the reduced fractions of PCs observed in the combined absence of RELB and

NF-B2. However, the reduced PCs observed in relbfl/flnfkb2fl/flCγ1-Cre mice is likely to be a direct consequence of the reduced GC B cells observed in these mice. Inconsistent phenotypes are also evident upon constitutive activation of the canonical pathway in GC B cells via the use of a constitutively active IKKβ. Rather than an elevation of GC B cells that would be predicted based on the loss of GCs in the absence of c-REL338, a slight loss of GCs is observed following constitutive canonical NF-B activation313. In addition, constitutive canonical activation in GC B cells had no impact on PC development313, a phenotype inconsistent with the loss of PCs observed upon ablation of RELA in PCs338. These discrepancies may be explained by non- physiological effects produced by forced over-activation of either pathway. In turn, genetic ablation of NF-B subunits may lead to alterations in dimer composition that produce non-

151 physiological phenotypes. Despite these potential experimental limitations, it appears that other routes of signaling may be in operation in addition to the linear canonical and alternative NF-B signaling axes which have largely been defined in cell lines systems. In the case of NIK, activation of the canonical pathway is strongly implied as PC hyperplasia in the absence of a

GC B cell phenotype observed following NIK overexpression would be consistent with activation of RELA. Interrogation of the canonical and alternative NF-B signaling routes via genetic manipulation of individual components may therefore provide a more nuanced picture of the NF-

B cascade in GC B cells and other cell types.

152

This page has been intentionally left blank.

153

Figure 4.1. Combined GC B cell-specific deletion of relb and nfkb2 impairs the GC reaction. (A-C, top) relbfl/flCγ1-Cre, nfkb2fl/flCγ1-Cre or relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-Cre control mice were analyzed via flow cytometry 14 days following immunization with sheep red blood cells (SRBCs) for CD95 and PNA expression on splenic B cells. Numbers above each gate denote the fractions of CD95hiPNAhi splenic germinal center (GC) B cells in mice of the indicated genotypes. (A-C, bottom) Summary of the frequencies of GC B cells. Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (**, P<0.01; ***,

P<0.001).

154

155

Figure 4.2. Normal fractions of GC B cells are present in relbfl/flnfkb2fl/flCγ1-Cre mice up to day 7 of the GC reaction. (A) Spleen sections from relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-Cre control mice were analyzed for the expression of

BCL6 and IgM via immunohistochemistry. 40x magnification. Data generated by Nicole Heise.

(B) Mice of the indicated genotypes were analyzed via flow cytometry 7 days following immunization with sheep red blood cells (SRBCs) for CD95 and PNA expression by splenic B cells. Numbers above each gate denote the fractions of CD95hiPNAhi splenic germinal center

(GC) B cells in mice of the indicated genotypes (top). Summary of the frequencies of GC B cells

(bottom). (C) Summary of the frequencies of CD95hiPNAhi splenic GC B cells in mice of the indicated genotypes as determined by flow cytometric analysis of CD95 and PNA expression by splenic B cells 10-12 days following immunization with SRBCs.

156

157

Figure 4.3. Normal DZ/LZ fractions and proliferation capacity of GC B cells in relbfl/flnfkb2fl/flCγ1-Cre mice. (A) relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-Cre control mice were analyzed via flow cytometry 10-12 days following immunization with sheep red blood cells (SRBCs) for the expression of CXCR4, CD86, CD95 and CD38 in splenic B cells. Numbers beside each gate indicate the percentage of

CD38loCD95hiCXCR4hiCD86lo dark zone (DZ) B cells and CD38loCD95hiCXCR4loCD86hi light zone (LZ) B cells in mice of the indicated genotypes (top). Summary of frequencies of DZ and

LZ B cells (bottom). Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05). (B) Mice of the indicated genotyes were analyzed via flow cytometry 10 days following immunization with

SRBCs for CD95 and PNA expression on splenic B cells and DNA content was assessed via

DAPI. Numbers beside each gate indicate the percentage of CD95hiPNAhi splenic germinal center (GC) B cells in G1 or G2/M and are representative of the fraction of GC B cells in G1 and

G2/M at day 8 and 12 of the GC reaction.

158

159

Figure 4.4. Identification of genes controlled by the alternative NF-B subunits, RELB and

NF-B2, in GC B cells at day 7 of the GC reaction. (A) eGFP+ germinal center (GC) B cells from relbfl/flnfkb2fl/flCγ1-Cre mice and GC B cells from Cγ1-Cre control mice 7 days following immunization with sheep red blood cells (SRBCs) were purified via flow cytometric sorting and

RNA was subjected to RNA-sequencing (RNA-seq). The mRNA expression of RELB and NF-

B2 is presented as total transcript counts for each gene. Each symbol represents a mouse.

Data are shown as mean ± standard deviation. Statistical significance was determined by

Student’s t test (*, P<0.05). (B) Unsupervised clustering analysis of transcript counts obtained from RNA-seq of RNA isolated via flow cytometry from eGFP+ GC B cells from relbfl/flnfkb2fl/flCγ1-Cre mice (red) and GC B cells from Cγ1-Cre control mice (blue). (C)

Differentially expressed sequence analysis (DESeq) of RNA from eGFP+ GC B cells from relbfl/flnfkb2fl/flCγ1-Cre mice and GC B cells from Cγ1-Cre control mice identified 59 transcripts with greater than 2.5-fold reduced expression (left) and 84 transcripts with greater than 2.5-fold increased expression (right) in the combined absence of RELB and NF-B2 at a significant threshold of p<0.01. Genes were assigned putative functional categories. For the identity of the corresponding genes, fold-change and p values, see Appendix Table B. eGFP+ identifies relb/nfkb2-deleted GC B cells from relbfl/flnfkb2fl/flCγ-Cre mice. Ub, ubiquitiation.

160

161

Figure 4.5. Reduced expression of genes with metabolic and immune trafficking functions in RELB/NF-B2-deficient GC B cells at day 7 of the GC reaction. Differentially expressed sequence analysis (DESeq) of RNA from eGFP+ GC B cells from relbfl/flnfkb2fl/flCγ1-

Cre mice and germinal center (GC) B cells from Cγ1-Cre control mice identified genes with functions in metabolism (A) and the immune response (C) with greater than 2.5-fold reduced expression in the combined absence of RELB and NF-B2 at a significant threshold of p<0.01.

Genes were assigned to putative functional categories. Genes within the lipid and amino acid categories (A) and trafficking categories (C) are presented, fold-change and padj values are indicated. For the identity of the remaining metabolic and immune response genes, fold-change and padj values, see Appendix B. (B) relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-Cre control mice were analyzed via flow cytometry 8-12 days following immunization with sheep red blood cells (SRBCs) for the expression of CD95, CD38, CD36 and eGFP in splenic B cells. The expression of CD36 in eGFP+ CD95hiPNAhi splenic germinal center

(GC) B cells from relbfl/flnfkb2fl/flCγ1-Cre (red) and relbfl/+nfkb2fl/+Cγ1-Cre (blue) mice and

CD95hiPNAhi splenic GC B cells (grey) from Cγ1-Cre control mice is presented (left). Summary of the corresponding median fluorescence intensities (MFI) (right). Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05). eGFP+ identifies relb/nfkb2-deleted GC B cells from relbfl/flnfkb2fl/flCγ-Cre and relbfl/+nfkb2fl/+Cγ1-Cre mice. (D) Chemokine/chemokine receptor gene signatures enriched in RELB/NF-B2-proficient GC B cells in comparison to RELB/NF-B2- deficient GC B cells identified by gene set enrichment analysis (GSEA) using the gene set collections Reactome (top) and GO_MF (bottom). Normalized enrichment scores (NES), p values and false discovery rates (FDRs) are indicated.

162

163

Figure 4.6. Normal PC frequencies following GC B cell-specific deletion of alternative NF-

B subunits. (A-C, top) relbfl/flCγ1-Cre, nfkb2fl/flCγ1-Cre or relbfl/flnfkb2fl/flCγ1-Cre mice and the corresponding heterozygous and Cγ1-Cre control mice were analyzed via flow cytometry 14 days following immunization with sheep red blood cells (SRBCs) for CD138 and eGFP expression of splenocytes. Numbers beside each gate denote the fractions of CD138hi splenic plasma cells (PCs) in mice of the indicated genotypes. (A-C, bottom) Summary of the frequencies of PCs. Each symbol represents a mouse. Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05; **, P<0.01). eGFP identifies relb/nfkb2-deleted PCs within the indicated genotypes.

164

165

Figure 4.7. Normal PC frequencies in nfkb2fl/flCγ1-Cre mice. nfkb2fl/flCγ1-Cre and the corresponding heterozygous and Cγ1-Cre control mice were analyzed via flow cytometry 14 and

28 days following immunization with NP-KLH for CD138 and eGFP expression of splenocytes

(top left) and bone marrow mononuclear cells (top right). Numbers beside each gate denote the fractions of CD138hi plasma cells in mice of the indicated genotypes. Summary of the frequencies of PCs (bottom). Each symbol represents a mouse. Data are shown as mean ± standard deviation. eGFP identifies nfkb2-deleted cells from nfkb2fl/flCγ1-Cre mice and nfkb2fl/+Cγ1-Cre mice.

166

167

Figure 4.8. GC B cell-specific deletion of nfkb2 results in reduced antigen-specific serum

IgG1 and antigen-specific IgG1 ASCs. (A) relbfl/flCγ1-Cre and the corresponding heterozygous and Cγ1-Cre control mice were analyzed for NP-9-specific IgG1 levels via ELISA

14 days following immunization with NP-KLH. (B) nfkb2fl/flCγ1-Cre and the corresponding heterozygous and Cγ1-Cre control mice were analyzed for NP-9-specific IgG1 levels via ELISA

14 (left) and 28 (right) days following immunization with NP-KLH. (C) nfkb2fl/flCγ1-Cre and the corresponding heterozygous and Cγ1-Cre control mice were analyzed for NP-25-specific IgG1 antibody secreting cells (ASCs) in splenocytes and bone marrow mononuclear cells via

ELISPOT analysis 28 days following immunization with NP-KLH (top). Numbers within each image indicate the number of ASCs observed for the indicated genotypes. Summary of the frequencies of NP-25-specific IgG1 ASCs (bottom). Data are shown as mean ± standard deviation. Statistical significance was determined by Student’s t test (*, P<0.05; **, P<0.01; ***,

P<0.001).

168

169

Figure 4.9. In vitro plasmablastic differentiation of NF-B2-deficient B cells. (A) NF-B2- deficient B cells were isolated from nfkb2fl/flCD19-Cre mice together with wild-type (WT) B cells from CD19-Cre control mice and induced to undergo plasmablastic differentiation via 3 days of

LPS stimulation or 4 days of CD40L + IL4 + IL5 stimulation. These cultures were analyzed via flow cytometry for the expression of CD138 and B220. Numbers beside each gate denote the fractions of CD138hi plasmablasts generated in the indicated cultures (left). Summary of the frequencies of generated plasmablasts in the indicated cultures (right). Each symbol represents a mouse. (B) d3 LPS and d4 CD40L + IL4 + IL5 plasmablast cultures of B cells from the indicated genotypes were analyzed for IgM antibody-secreting cells (ASCs) via ELISPOT. (C)

Western blot analysis of p100/p52, IRF4 and BLIMP1 protein expression in ex vivo (day 0) B cells, day 4 CD40L + IL4 + IL5 and day 3 LPS cultures of B cells from the indicated genotypes.

170

171

This page has been intentionally left blank.

172

CHAPTER 5: Conclusions

173

Two branches of the NF-B cascade have evolved over time, the canonical and alternative NF-

B pathways, and it is likely that these pathways have both overlapping and distinct functions within various cellular systems. Distinguishing these functions by studying upstream regulators of the separate NF-B signaling pathways is limited by the possibility of pathway cross-talk.

Therefore, analyzing the roles of the downstream transcription factor subunits appears to be a more effective strategy towards identifying the specific functions of the canonical and alternative

NF-B pathways. The biological roles of the alterative NF-B subunits within the B cell lineage have been obscured in constitutional knockout mice by the diverse functions of these subunits in non-B cell types280. To overcome these limitations, conditional alleles were generated to investigate the roles of the alternative NF-B subunits, RELB and NF-B2, in B cell development. These alleles allowed the identification of complex functional requirements for

RELB and/or NF-B2 in naïve B cells, GC B cells and PCs. The genetic tools generated in this study will also facilitate the investigation of the cell-intrinsic roles of RELB and/or NF-B2 in other cell types of the immune system.

A large body of work has demonstrated that BAFF signaling is critical for the maintenance of mature B cells. However, the contribution of the alternative NF-B subunits that are activated downstream of BAFF remained unclear, especially in regards to their specific target genes. We have identified critical, B cell-intrinsic roles for RELB and NF-B2 in the maintenance of mature B cells. In response to BAFF, these subunits were found to control the expression of anti-apoptotic genes, genes that direct positioning within B cell follicles and genes involved in promoting B–T cell interactions. Correct positioning within the B cell niche is likely to ensure continued exposure to BAFF and other essential signals required for the survival and growth of B cells, and optimal B–T cell engagements promote effective antigen-mediated activation. Interestingly, it is now clear that tonic BCR signaling is unlikely to signal through the canonical NF-B pathway347, and BAFF-mediated activation of this pathway may explain the

174 contribution of this signaling route to B cell maintenance. Identifying the gene targets controlled by canonical NF-B subunits in response to BAFF will therefore provide important insights into the distinct roles of the canonical and alternative NF-B pathways in the maintenance of mature

B cells. In addition, this will clarify to what extent these pathways reinforce the expression of shared, critical gene targets.

Our findings may have implications for diseases of mature B cells. For example, BIRC3

(cIAP-2), an upstream regulator of the alternative pathway, is frequently mutated in chronic lymphocytic leukemia (CLL)427, a malignancy of mature, resting B cells, and in splenic MZ lymphoma428. In addition, BAFF has been implicated in promoting autoimmunity, including systemic lupus erythematosus (SLE)429,430, and a monoclonal antibody specific to human BAFF was approved as the first targeted therapy for the treatment of SLE431,432. Elevated levels of

BAFF have also been associated with various B cell malignances including B-NHLs, MM and

CLL430. Under physiological conditions, competition for BAFF stimulation serves as a limiting factor that restricts B cell expansion; elevated levels of BAFF therefore allow unrestricted expansion. Furthermore, BAFF inhibition was found to induce apoptosis in B-NHL and CLL B cells, and BAFF stimulation resulted in the expression of several anti-apoptotic NF-B target genes433,434. Therefore, one possible consequence of mutations in B cell malignancies that lead to constitutive activation of the NF-B cascade335 is the acquisition of independence from BAFF signals which may contribute to further unrestricted survival and propagation. Experimentally this possibility is supported by the finding that BAFF-R deficiency can be rescued by constitutive canonical pathway activation353 or following the inactivation of negative regulators of the alternative NF-B pathway, TRAF2, TRAF3 or cIAP1/2356. Our results suggest that constitutive activation of RELB and NF-B2, via mutation or prolonged exposure to BAFF, may promote cell survival, distort trafficking within the lymphoid tissue, and potentially hyperactivate B cells, all of which may contribute to disease pathogenesis.

175

In contrast to the maintenance of mature B cells, the molecular events that coordinate the dynamic environment of the GC, particularly during the selection of high affinity LZ B cells, are not well understood. Highly reinforced systems of mutual repression exist between the factors governing mature B cell and GC B cell development and those that drive PC development. The means by which this balance is tipped during the selection of GC B cell clones within the LZ to induce terminal differentiation is unclear. In addition, progression through the GC reaction is not linear, as selected GC B cells can be instructed to recirculate back to the

DZ, undergo CSR or differentiate into memory B cells or PCs. These heterogeneous outcomes suggest that selected LZ B cells are subjected to a complex array of signals resulting in different developmental outcomes. In the case of PC development, a signal leading to the inhibition of

PAX5 expression or function is critical191, whereas c-MYC induction is important for cyclic reentry162,163. The affinity of the BCR435,436 and the spectrum of T cell-derived signals a GC B cell is exposed to, either in the form of physical interactions (such as CD40-CD40L119,129) or cytokines (such as IL-4 and IL-21126,136,437,438), have roles in determining cell fates. Asymmetric distribution of key regulators during cell division may also play a critical role439, and epigenetic regulation may enable rapid and reversible activation or repression of key regulators such as

BLIMP1 and IRF4, thus allowing dynamic modulation of cellular phenotypes within the

GC269,270,440.

Our work has demonstrated that activation of distinct NF-B pathways and even NF-B subunits provides an additional layer of molecular control during the critical selection phase within the GC (Fig. 5.1). In the case of the canonical subunits, c-REL is required for GC maintenance, likely via the facilitation of cyclic reentry, whereas RELA promotes PC development338. The work presented here has demonstrated that in addition to c-REL, the alternative NF-B subunits RELB and NF-B2 are also jointly required for GC maintenance.

This requirement is also likely to be associated with cyclic reentry of LZ B cells into the DZ and

176 may be centered on the establishment of a biological program that is complementary to that directed by c-REL. In addition to this joint function in GC maintenance, individual roles for RELB and NF-B2 were identified in PCs. RELB appears to be involved in PC development whereas

NF-B2 may control a functional aspect of PC biology related to the generation or secretion of high affinity IgG1. The role of RELB and NF-B2 in PCs may be related to the preferential expression of these subunits in human PCs and PC precursors. While it is difficult to precisely link receptor-mediated activation to downstream signaling events, LZ B cells are probably exposed to a spectrum of signals that collectively determine which NF-B subunits are activated in particular cells and which cellular fates are promoted. Possible signaling routes of NF-B activation in LZ B cells include stimulation by BAFF, CD40L, antigen and also exposure to particular cytokines.

The frequency of activating NF-B mutations in ABC-DLBCL and MM335 is likely to be a reflection of the physiological role of the NF-B signaling cascade in LZ B cells and PCs, respectively. During lymphomagenesis it is possible that these endogenous roles are hijacked by tumor-precursor cells and contribute to pathogenesis. Understanding the normal role of NF-

B signaling within the GC may therefore provide insights that shed new light on the pathogenesis of these diseases. Mutations that activate the canonical or alternative NF-B pathway may lead to the hyperactivation of c-REL-containing dimers or the RELB-p52 heterodimer, respectively. Our findings suggest that constitutive activation of these dimers may then enforce a metabolic program that induces inappropriate cell growth and/or locks GC B cells into a state of continuous recirculation that is independent of affinity-mediated selection and prevents GC exit. In contrast, constitutive activation of RELA- or RELB-containing dimers may lead to an overexpansion of plasmablasts and/or PCs. In regards to the role of NF-B2 in a functional aspect of normal PC biology, over-activation of p52-containing dimers may contribute to enhancing the aggressiveness of a MM precursor cell. As both normal and malignant PCs are

177 dependent on the bone marrow stroma for survival signals that activate the NF-B signaling cascade, activating NF-B mutations are also likely to enable growth independent of the bone marrow niche333.

There is a demand for novel, targeted therapies for the treatment of B-NHL and MM as current treatment regimens have broad effects and are associated with significant toxicity. As constitutive NF-B activation is observed in ABC-DLBCL and MM, these malignancies are candidates for treatment with NF-B pathway inhibitors. To this end, proteasomal inhibitors, such as bortezomib, have been successfully utilized in the clinic for the treatment of MM441.

These proteasomal inhibitors work in part by inhibiting NF-B signaling in MM cells and expression of a strong NF-B gene signature, or low levels of TRAF3 is associated with a better response to treatment320,442,443. However, proteasomal inhibitors are also associated with significant toxicity, and targeted NF-B inhibition may therefore prove to be a more effective strategy. This strategy may also be of use for the treatment of autoimmune diseases, as sequestration of BAFF leads to general B cell depletion429. Overall, it appears that inhibition of the canonical NF-B pathway may be particularly effective for the treatment of ABC-DLBCL, whereas dual inhibition of the canonical and alternative pathways may be required for effective

MM therapy. In regards to the canonical NF-B pathway, several IKKβ inhibitors have been developed and toxicity has been demonstrated in MM318,323,444. Recently, pharmacological inhibition of NIK has been shown to have efficacy towards mantle cell lymphoma that is refractory to canonical pathway inhibition following treatment with BCR inhibitors445. These resistant mantle cell lymphomas were found to have mutations that activated the alternative NF-

B pathway, suggesting the activity of this pathway was sufficient to preserve growth upon inhibition of the canonical pathway445. Inhibition of the NF-B cascade at the level of downstream transcription factor subunits as opposed to global NF-B pathway inhibition would be predicted to reduce toxicity. The validity of inhibiting NF-B subunits for the treatment of B

178 cell malignancies is supported by our work that has demonstrated that specific functions can be ascribed to single subunits or heterodimers. To provide a further rationale for this approach, the effects of NF-B subunit knockout in mouse models of lymphoma could be assessed. In regards to the feasibility of inhibiting individual NF-B subunits, a promising advance has been the development of a small molecule c-REL inhibitor446. Finally, gene targets found to be controlled by c-REL338 or RELB and NF-B2, could potentially be utilized as biomarkers to identify patients in which selective inhibition of the canonical or alternative NF-B pathway or individual subunits may be beneficial. These biomarkers could also be used to gauge responsiveness to NF-B pathway inhibitors.

179

Figure 5.1. Requirement of NF-B subunits in GC B cell and PC development. Distinct NF-

B pathways and NF-B subunits have functional roles during the selection of high-affinity light zone (LZ) B cells within the germinal center (GC). In the case of the canonical NF-B subunits, c-REL is required for GC maintenance, likely via the facilitation of cyclic reentry, whereas RELA promotes plasma cell (PC) development338. The work presented in this thesis has demonstrated that in addition to c-REL, the alternative NF-B subunits RELB and NF-B2 are also jointly required for GC maintenance. This requirement is also likely to be associated with cyclic reentry of LZ B cells into the dark zone (DZ) and may be centered on the establishment of a biological program that is complementary to that directed by c-REL. In addition to this joint function in GC maintenance, individual roles for RELB and NF-B2 were identified in PCs. RELB appears to be involved in PC development whereas NF-B2 may control a functional aspect of PC biology related to the generation or secretion of high affinity IgG1.

180

181

This page has been intentionally left blank.

182

REFERENCES

183

1 Allman, D. & Pillai, S. Peripheral B cell subsets. Current opinion in immunology 20, 149‐157, doi:10.1016/j.coi.2008.03.014 (2008). 2 Carsetti, R., Rosado, M. M. & Wardmann, H. Peripheral development of B cells in mouse and man. Immunological reviews 197, 179‐191 (2004). 3 Dorshkind, K. & Montecino‐Rodriguez, E. Fetal B‐cell lymphopoiesis and the emergence of B‐1‐ cell potential. Nature reviews. Immunology 7, 213‐219, doi:10.1038/nri2019 (2007). 4 Allman, D. et al. Resolution of three nonproliferative immature splenic B cell subsets reveals multiple selection points during peripheral B cell maturation. Journal of immunology (Baltimore, Md. : 1950) 167, 6834‐6840 (2001). 5 Nemazee, D. & Buerki, K. Clonal deletion of autoreactive B lymphocytes in bone marrow chimeras. Proceedings of the National Academy of Sciences of the United States of America 86, 8039‐8043 (1989). 6 Hartley, S. B. et al. Elimination from peripheral lymphoid tissues of self‐reactive B lymphocytes recognizing membrane‐bound antigens. Nature 353, 765‐769, doi:10.1038/353765a0 (1991). 7 Fulcher, D. A. & Basten, A. Reduced life span of anergic self‐reactive B cells in a double‐ transgenic model. The Journal of experimental medicine 179, 125‐134 (1994). 8 Cariappa, A., Chase, C., Liu, H., Russell, P. & Pillai, S. Naive recirculating B cells mature simultaneously in the spleen and bone marrow. Blood 109, 2339‐2345, doi:10.1182/blood‐2006‐ 05‐021089 (2007). 9 Lindsley, R. C., Thomas, M., Srivastava, B. & Allman, D. Generation of peripheral B cells occurs via two spatially and temporally distinct pathways. Blood 109, 2521‐2528, doi:10.1182/blood‐ 2006‐04‐018085 (2007). 10 Tanigaki, K. et al. Notch‐RBP‐J signaling is involved in cell fate determination of marginal zone B cells. Nature immunology 3, 443‐450, doi:10.1038/ni793 (2002). 11 Saito, T. et al. Notch2 is preferentially expressed in mature B cells and indispensable for marginal zone B lineage development. Immunity 18, 675‐685 (2003). 12 Tan, J. B. et al. Lunatic and manic fringe cooperatively enhance marginal zone B cell precursor competition for delta‐like 1 in splenic endothelial niches. Immunity 30, 254‐263, doi:10.1016/j.immuni.2008.12.016 (2009). 13 Kuroda, K. et al. Regulation of marginal zone B cell development by MINT, a suppressor of Notch/RBP‐J signaling pathway. Immunity 18, 301‐312 (2003). 14 Quong, M. W. et al. Receptor editing and marginal zone B cell development are regulated by the helix‐loop‐helix protein, E2A. The Journal of experimental medicine 199, 1101‐1112, doi:10.1084/jem.20031180 (2004). 15 Zhang, X. K. et al. The transcription factor Fli‐1 modulates marginal zone and follicular B cell development in mice. Journal of immunology (Baltimore, Md. : 1950) 181, 1644‐1654 (2008). 16 Huang, Z., Nie, L., Xu, M. & Sun, X. H. Notch‐induced E2A degradation requires CHIP and Hsc70 as novel facilitators of ubiquitination. Molecular and cellular biology 24, 8951‐8962, doi:10.1128/mcb.24.20.8951‐8962.2004 (2004). 17 Martin, F. & Kearney, J. F. Positive selection from newly formed to marginal zone B cells depends on the rate of clonal production, CD19, and btk. Immunity 12, 39‐49 (2000). 18 Cariappa, A. et al. The follicular versus marginal zone B lymphocyte cell fate decision is regulated by Aiolos, Btk, and CD21. Immunity 14, 603‐615 (2001). 19 Wen, L. et al. Evidence of marginal‐zone B cell‐positive selection in spleen. Immunity 23, 297‐ 308, doi:10.1016/j.immuni.2005.08.007 (2005). 20 Hampel, F. et al. CD19‐independent instruction of murine marginal zone B‐cell development by constitutive Notch2 signaling. Blood 118, 6321‐6331, doi:10.1182/blood‐2010‐12‐325944 (2011).

184

21 Mebius, R. E. & Kraal, G. Structure and function of the spleen. Nature reviews. Immunology 5, 606‐616, doi:10.1038/nri1669 (2005). 22 Nolte, M. A., Hoen, E. N., van Stijn, A., Kraal, G. & Mebius, R. E. Isolation of the intact white pulp. Quantitative and qualitative analysis of the cellular composition of the splenic compartments. European journal of immunology 30, 626‐634, doi:10.1002/1521‐4141(200002)30:2<626::aid‐ immu626>3.0.co;2‐h (2000). 23 Forster, R., Emrich, T., Kremmer, E. & Lipp, M. Expression of the G‐protein‐‐coupled receptor BLR1 defines mature, recirculating B cells and a subset of T‐helper memory cells. Blood 84, 830‐ 840 (1994). 24 Schmidt, K. N., Hsu, C. W., Griffin, C. T., Goodnow, C. C. & Cyster, J. G. Spontaneous follicular exclusion of SHP1‐deficient B cells is conditional on the presence of competitor wild‐type B cells. The Journal of experimental medicine 187, 929‐937 (1998). 25 Ansel, K. M. et al. A chemokine‐driven positive feedback loop organizes lymphoid follicles. Nature 406, 309‐314, doi:10.1038/35018581 (2000). 26 Forster, R. et al. A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs and specific anatomic compartments of the spleen. Cell 87, 1037‐1047 (1996). 27 Ngo, V. N. et al. Lymphotoxin alpha/beta and tumor necrosis factor are required for stromal cell expression of homing chemokines in B and T cell areas of the spleen. The Journal of experimental medicine 189, 403‐412 (1999). 28 Barrington, R. A., Pozdnyakova, O., Zafari, M. R., Benjamin, C. D. & Carroll, M. C. B lymphocyte memory: role of stromal cell complement and FcgammaRIIB receptors. The Journal of experimental medicine 196, 1189‐1199 (2002). 29 Heesters, B. A., Myers, R. C. & Carroll, M. C. Follicular dendritic cells: dynamic antigen libraries. Nature reviews. Immunology 14, 495‐504, doi:10.1038/nri3689 (2014). 30 Gunn, M. D. et al. Mice lacking expression of secondary lymphoid organ chemokine have defects in lymphocyte homing and dendritic cell localization. The Journal of experimental medicine 189, 451‐460 (1999). 31 Forster, R. et al. CCR7 coordinates the primary immune response by establishing functional microenvironments in secondary lymphoid organs. Cell 99, 23‐33 (1999). 32 Ansel, K. M., McHeyzer‐Williams, L. J., Ngo, V. N., McHeyzer‐Williams, M. G. & Cyster, J. G. In vivo‐activated CD4 T cells upregulate CXC chemokine receptor 5 and reprogram their response to lymphoid chemokines. The Journal of experimental medicine 190, 1123‐1134 (1999). 33 Garside, P. et al. Visualization of specific B and T lymphocyte interactions in the lymph node. Science (New York, N.Y.) 281, 96‐99 (1998). 34 Okada, T. et al. Antigen‐engaged B cells undergo chemotaxis toward the T zone and form motile conjugates with helper T cells. PLoS biology 3, e150, doi:10.1371/journal.pbio.0030150 (2005). 35 Reif, K. et al. Balanced responsiveness to chemoattractants from adjacent zones determines B‐ cell position. Nature 416, 94‐99, doi:10.1038/416094a (2002). 36 Schwab, S. R. et al. Lymphocyte sequestration through S1P lyase inhibition and disruption of S1P gradients. Science (New York, N.Y.) 309, 1735‐1739, doi:10.1126/science.1113640 (2005). 37 Pappu, R. et al. Promotion of lymphocyte egress into blood and lymph by distinct sources of sphingosine‐1‐phosphate. Science (New York, N.Y.) 316, 295‐298, doi:10.1126/science.1139221 (2007). 38 Cinamon, G. et al. Sphingosine 1‐phosphate receptor 1 promotes B cell localization in the splenic marginal zone. Nature immunology 5, 713‐720, doi:10.1038/ni1083 (2004). 39 Gunn, M. D. et al. A B‐cell‐homing chemokine made in lymphoid follicles activates Burkitt's lymphoma receptor‐1. Nature 391, 799‐803, doi:10.1038/35876 (1998).

185

40 Vora, K. A. et al. Sphingosine 1‐phosphate receptor agonist FTY720‐phosphate causes marginal zone B cell displacement. Journal of leukocyte biology 78, 471‐480, doi:10.1189/jlb.0904487 (2005). 41 Lu, T. T. & Cyster, J. G. Integrin‐mediated long‐term B cell retention in the splenic marginal zone. Science (New York, N.Y.) 297, 409‐412, doi:10.1126/science.1071632 (2002). 42 Simonetti, G. et al. IRF4 controls the positioning of mature B cells in the lymphoid microenvironments by regulating NOTCH2 expression and activity. The Journal of experimental medicine 210, 2887‐2902, doi:10.1084/jem.20131026 (2013). 43 Chen, Y. et al. Defective microarchitecture of the spleen marginal zone and impaired response to a thymus‐independent type 2 antigen in mice lacking scavenger receptors MARCO and SR‐A. Journal of immunology (Baltimore, Md. : 1950) 175, 8173‐8180 (2005). 44 Balazs, M., Martin, F., Zhou, T. & Kearney, J. Blood dendritic cells interact with splenic marginal zone B cells to initiate T‐independent immune responses. Immunity 17, 341‐352 (2002). 45 Guinamard, R., Okigaki, M., Schlessinger, J. & Ravetch, J. V. Absence of marginal zone B cells in Pyk‐2‐deficient mice defines their role in the humoral response. Nature immunology 1, 31‐36, doi:10.1038/76882 (2000). 46 Test, S. T., Mitsuyoshi, J., Connolly, C. C. & Lucas, A. H. Increased immunogenicity and induction of class switching by conjugation of complement C3d to pneumococcal serotype 14 capsular polysaccharide. Infection and immunity 69, 3031‐3040, doi:10.1128/iai.69.5.3031‐3040.2001 (2001). 47 Oliver, A. M., Martin, F., Gartland, G. L., Carter, R. H. & Kearney, J. F. Marginal zone B cells exhibit unique activation, proliferative and immunoglobulin secretory responses. European journal of immunology 27, 2366‐2374, doi:10.1002/eji.1830270935 (1997). 48 Treml, L. S. et al. TLR stimulation modifies BLyS receptor expression in follicular and marginal zone B cells. Journal of immunology (Baltimore, Md. : 1950) 178, 7531‐7539 (2007). 49 Genestier, L. et al. TLR agonists selectively promote terminal plasma cell differentiation of B cell subsets specialized in thymus‐independent responses. Journal of immunology (Baltimore, Md. : 1950) 178, 7779‐7786 (2007). 50 Rubtsov, A. V. et al. TLR agonists promote marginal zone B cell activation and facilitate T‐ dependent IgM responses. Journal of immunology (Baltimore, Md. : 1950) 180, 3882‐3888 (2008). 51 Haas, K. M., Poe, J. C. & Tedder, T. F. CD21/35 promotes protective immunity to Streptococcus pneumoniae through a complement‐independent but CD19‐dependent pathway that regulates PD‐1 expression. Journal of immunology (Baltimore, Md. : 1950) 183, 3661‐3671, doi:10.4049/jimmunol.0901218 (2009). 52 Kang, Y. S. et al. A dominant complement fixation pathway for pneumococcal polysaccharides initiated by SIGN‐R1 interacting with C1q. Cell 125, 47‐58, doi:10.1016/j.cell.2006.01.046 (2006). 53 Martin, F., Oliver, A. M. & Kearney, J. F. Marginal zone and B1 B cells unite in the early response against T‐independent blood‐borne particulate antigens. Immunity 14, 617‐629 (2001). 54 Belperron, A. A., Dailey, C. M. & Bockenstedt, L. K. Infection‐induced marginal zone B cell production of Borrelia hermsii‐specific antibody is impaired in the absence of CD1d. Journal of immunology (Baltimore, Md. : 1950) 174, 5681‐5686 (2005). 55 Garcia de Vinuesa, C., O'Leary, P., Sze, D. M., Toellner, K. M. & MacLennan, I. C. T‐independent type 2 antigens induce B cell proliferation in multiple splenic sites, but exponential growth is confined to extrafollicular foci. European journal of immunology 29, 1314‐1323, doi:10.1002/(sici)1521‐4141(199904)29:04<1314::aid‐immu1314>3.0.co;2‐4 (1999).

186

56 Amlot, P. L. & Hayes, A. E. Impaired human antibody response to the thymus‐independent antigen, DNP‐Ficoll, after splenectomy. Implications for post‐splenectomy infections. Lancet 1, 1008‐1011 (1985). 57 Kruetzmann, S. et al. Human immunoglobulin M memory B cells controlling Streptococcus pneumoniae infections are generated in the spleen. The Journal of experimental medicine 197, 939‐945, doi:10.1084/jem.20022020 (2003). 58 Castagnola, E. & Fioredda, F. Prevention of life‐threatening infections due to encapsulated bacteria in children with hyposplenia or asplenia: a brief review of current recommendations for practical purposes. European journal of haematology 71, 319‐326 (2003). 59 Cerutti, A., Cols, M. & Puga, I. Marginal zone B cells: virtues of innate‐like antibody‐producing lymphocytes. Nature reviews. Immunology 13, 118‐132, doi:10.1038/nri3383 (2013). 60 Rajewsky, K. Clonal selection and learning in the antibody system. Nature 381, 751‐758, doi:10.1038/381751a0 (1996). 61 Lam, K. P., Kuhn, R. & Rajewsky, K. In vivo ablation of surface immunoglobulin on mature B cells by inducible gene targeting results in rapid cell death. Cell 90, 1073‐1083 (1997). 62 Kraus, M., Alimzhanov, M. B., Rajewsky, N. & Rajewsky, K. Survival of resting mature B lymphocytes depends on BCR signaling via the Igalpha/beta heterodimer. Cell 117, 787‐800, doi:10.1016/j.cell.2004.05.014 (2004). 63 Grande, S. M., Bannish, G., Fuentes‐Panana, E. M., Katz, E. & Monroe, J. G. Tonic B‐cell and viral ITAM signaling: context is everything. Immunological reviews 218, 214‐234, doi:10.1111/j.1600‐ 065X.2007.00535.x (2007). 64 Gross, J. A. et al. TACI‐Ig neutralizes molecules critical for B cell development and autoimmune disease. impaired B cell maturation in mice lacking BLyS. Immunity 15, 289‐302 (2001). 65 Lentz, V. M., Cancro, M. P., Nashold, F. E. & Hayes, C. E. Bcmd governs recruitment of new B cells into the stable peripheral B cell pool in the A/WySnJ mouse. Journal of immunology (Baltimore, Md. : 1950) 157, 598‐606 (1996). 66 Miller, D. J. & Hayes, C. E. Phenotypic and genetic characterization of a unique B lymphocyte deficiency in strain A/WySnJ mice. European journal of immunology 21, 1123‐1130, doi:10.1002/eji.1830210506 (1991). 67 Sasaki, Y., Casola, S., Kutok, J. L., Rajewsky, K. & Schmidt‐Supprian, M. TNF family member B cell‐ activating factor (BAFF) receptor‐dependent and ‐independent roles for BAFF in B cell physiology. Journal of immunology (Baltimore, Md. : 1950) 173, 2245‐2252 (2004). 68 Schiemann, B. et al. An essential role for BAFF in the normal development of B cells through a BCMA‐independent pathway. Science (New York, N.Y.) 293, 2111‐2114, doi:10.1126/science.1061964 (2001). 69 Scholz, J. L. et al. BLyS inhibition eliminates primary B cells but leaves natural and acquired humoral immunity intact. Proceedings of the National Academy of Sciences of the United States of America 105, 15517‐15522, doi:10.1073/pnas.0807841105 (2008). 70 Shulga‐Morskaya, S. et al. B cell‐activating factor belonging to the TNF family acts through separate receptors to support B cell survival and T cell‐independent antibody formation. Journal of immunology (Baltimore, Md. : 1950) 173, 2331‐2341 (2004). 71 Thompson, J. S. et al. BAFF‐R, a newly identified TNF receptor that specifically interacts with BAFF. Science (New York, N.Y.) 293, 2108‐2111, doi:10.1126/science.1061965 (2001). 72 Yan, M. et al. Identification of a novel receptor for B lymphocyte stimulator that is mutated in a mouse strain with severe B cell deficiency. Current biology : CB 11, 1547‐1552 (2001). 73 Harless, S. M. et al. Competition for BLyS‐mediated signaling through Bcmd/BR3 regulates peripheral B lymphocyte numbers. Current biology : CB 11, 1986‐1989 (2001).

187

74 Gross, J. A. et al. TACI and BCMA are receptors for a TNF homologue implicated in B‐cell autoimmune disease. Nature 404, 995‐999, doi:10.1038/35010115 (2000). 75 Mackay, F. et al. Mice transgenic for BAFF develop lymphocytic disorders along with autoimmune manifestations. The Journal of experimental medicine 190, 1697‐1710 (1999). 76 Moore, P. A. et al. BLyS: member of the tumor necrosis factor family and B lymphocyte stimulator. Science (New York, N.Y.) 285, 260‐263 (1999). 77 Khare, S. D. et al. Severe B cell hyperplasia and autoimmune disease in TALL‐1 transgenic mice. Proceedings of the National Academy of Sciences of the United States of America 97, 3370‐3375, doi:10.1073/pnas.050580697 (2000). 78 Batten, M. et al. BAFF mediates survival of peripheral immature B lymphocytes. The Journal of experimental medicine 192, 1453‐1466 (2000). 79 Castro, I. et al. B cell receptor‐mediated sustained c‐Rel activation facilitates late transitional B cell survival through control of B cell activating factor receptor and NF‐kappaB2. Journal of immunology (Baltimore, Md. : 1950) 182, 7729‐7737, doi:10.4049/jimmunol.0803281 (2009). 80 Meyer‐Bahlburg, A., Andrews, S. F., Yu, K. O., Porcelli, S. A. & Rawlings, D. J. Characterization of a late transitional B cell population highly sensitive to BAFF‐mediated homeostatic proliferation. The Journal of experimental medicine 205, 155‐168, doi:10.1084/jem.20071088 (2008). 81 Stadanlick, J. E. et al. Tonic B cell antigen receptor signals supply an NF‐kappaB substrate for prosurvival BLyS signaling. Nature immunology 9, 1379‐1387, doi:10.1038/ni.1666 (2008). 82 El Shikh, M. E., El Sayed, R. M., Sukumar, S., Szakal, A. K. & Tew, J. G. Activation of B cells by antigens on follicular dendritic cells. Trends in immunology 31, 205‐211, doi:10.1016/j.it.2010.03.002 (2010). 83 Jellusova, J. et al. Context‐specific BAFF‐R signaling by the NF‐kappaB and PI3K pathways. Cell reports 5, 1022‐1035, doi:10.1016/j.celrep.2013.10.022 (2013). 84 Otipoby, K. L. et al. BAFF activates Akt and Erk through BAFF‐R in an IKK1‐dependent manner in primary mouse B cells. Proceedings of the National Academy of Sciences of the United States of America 105, 12435‐12438, doi:10.1073/pnas.0805460105 (2008). 85 Patke, A., Mecklenbrauker, I., Erdjument‐Bromage, H., Tempst, P. & Tarakhovsky, A. BAFF controls B cell metabolic fitness through a PKC beta‐ and Akt‐dependent mechanism. The Journal of experimental medicine 203, 2551‐2562, doi:10.1084/jem.20060990 (2006). 86 Schweighoffer, E. et al. The BAFF receptor transduces survival signals by co‐opting the B cell receptor signaling pathway. Immunity 38, 475‐488, doi:10.1016/j.immuni.2012.11.015 (2013). 87 Woodland, R. T. et al. Multiple signaling pathways promote B lymphocyte stimulator dependent B‐cell growth and survival. Blood 111, 750‐760, doi:10.1182/blood‐2007‐03‐077222 (2008). 88 MacLennan, I. C. Germinal centers. Annual review of immunology 12, 117‐139, doi:10.1146/annurev.iy.12.040194.001001 (1994). 89 Berek, C., Berger, A. & Apel, M. Maturation of the immune response in germinal centers. Cell 67, 1121‐1129 (1991). 90 Blink, E. J. et al. Early appearance of germinal center‐derived memory B cells and plasma cells in blood after primary immunization. The Journal of experimental medicine 201, 545‐554, doi:10.1084/jem.20042060 (2005). 91 Durandy, A., Kracker, S. & Fischer, A. Primary antibody deficiencies. Nature reviews. Immunology 13, 519‐533, doi:10.1038/nri3466 (2013). 92 Batista, F. D. & Harwood, N. E. The who, how and where of antigen presentation to B cells. Nature reviews. Immunology 9, 15‐27, doi:10.1038/nri2454 (2009). 93 Kerfoot, S. M. et al. Germinal center B cell and T follicular helper cell development initiates in the interfollicular zone. Immunity 34, 947‐960, doi:10.1016/j.immuni.2011.03.024 (2011).

188

94 Kitano, M. et al. Bcl6 protein expression shapes pre‐germinal center B cell dynamics and follicular helper T cell heterogeneity. Immunity 34, 961‐972, doi:10.1016/j.immuni.2011.03.025 (2011). 95 Qi, H., Cannons, J. L., Klauschen, F., Schwartzberg, P. L. & Germain, R. N. SAP‐controlled T‐B cell interactions underlie germinal centre formation. Nature 455, 764‐769, doi:10.1038/nature07345 (2008). 96 Choi, Y. S. et al. ICOS receptor instructs T follicular helper cell versus effector cell differentiation via induction of the transcriptional repressor Bcl6. Immunity 34, 932‐946, doi:10.1016/j.immuni.2011.03.023 (2011). 97 Baumjohann, D., Okada, T. & Ansel, K. M. Cutting Edge: Distinct waves of BCL6 expression during T follicular helper cell development. Journal of immunology (Baltimore, Md. : 1950) 187, 2089‐ 2092, doi:10.4049/jimmunol.1101393 (2011). 98 Johnston, R. J. et al. Bcl6 and Blimp‐1 are reciprocal and antagonistic regulators of T follicular helper cell differentiation. Science (New York, N.Y.) 325, 1006‐1010, doi:10.1126/science.1175870 (2009). 99 Nurieva, R. I. et al. Bcl6 mediates the development of T follicular helper cells. Science (New York, N.Y.) 325, 1001‐1005, doi:10.1126/science.1176676 (2009). 100 Yu, D. et al. The transcriptional repressor Bcl‐6 directs T follicular helper cell lineage commitment. Immunity 31, 457‐468, doi:10.1016/j.immuni.2009.07.002 (2009). 101 Gatto, D., Paus, D., Basten, A., Mackay, C. R. & Brink, R. Guidance of B cells by the orphan G protein‐coupled receptor EBI2 shapes humoral immune responses. Immunity 31, 259‐269, doi:10.1016/j.immuni.2009.06.016 (2009). 102 Pereira, J. P., Kelly, L. M., Xu, Y. & Cyster, J. G. EBI2 mediates B cell segregation between the outer and centre follicle. Nature 460, 1122‐1126, doi:10.1038/nature08226 (2009). 103 Hannedouche, S. et al. Oxysterols direct immune cell migration via EBI2. Nature 475, 524‐527, doi:10.1038/nature10280 (2011). 104 Liu, C. et al. Oxysterols direct B‐cell migration through EBI2. Nature 475, 519‐523, doi:10.1038/nature10226 (2011). 105 Kelly, L. M., Pereira, J. P., Yi, T., Xu, Y. & Cyster, J. G. EBI2 guides serial movements of activated B cells and ligand activity is detectable in lymphoid and nonlymphoid tissues. Journal of immunology (Baltimore, Md. : 1950) 187, 3026‐3032, doi:10.4049/jimmunol.1101262 (2011). 106 Green, J. A. et al. The sphingosine 1‐phosphate receptor S1P(2) maintains the homeostasis of germinal center B cells and promotes niche confinement. Nature immunology 12, 672‐680, doi:10.1038/ni.2047 (2011). 107 Victora, G. D. & Nussenzweig, M. C. Germinal centers. Annual review of immunology 30, 429‐ 457, doi:10.1146/annurev‐immunol‐020711‐075032 (2012). 108 Bannard, O. et al. Germinal center centroblasts transition to a centrocyte phenotype according to a timed program and depend on the dark zone for effective selection. Immunity 39, 912‐924, doi:10.1016/j.immuni.2013.08.038 (2013). 109 Jacob, J., Kelsoe, G., Rajewsky, K. & Weiss, U. Intraclonal generation of antibody mutants in germinal centres. Nature 354, 389‐392, doi:10.1038/354389a0 (1991). 110 William, J., Euler, C., Christensen, S. & Shlomchik, M. J. Evolution of autoantibody responses via somatic hypermutation outside of germinal centers. Science (New York, N.Y.) 297, 2066‐2070, doi:10.1126/science.1073924 (2002). 111 Muramatsu, M. et al. Class switch recombination and hypermutation require activation‐induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102, 553‐563 (2000). 112 Revy, P. et al. Activation‐induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the Hyper‐IgM syndrome (HIGM2). Cell 102, 565‐575 (2000).

189

113 Papavasiliou, F. N. & Schatz, D. G. Cell‐cycle‐regulated DNA double‐stranded breaks in somatic hypermutation of immunoglobulin genes. Nature 408, 216‐221, doi:10.1038/35041599 (2000). 114 Bross, L. et al. DNA double‐strand breaks in immunoglobulin genes undergoing somatic hypermutation. Immunity 13, 589‐597 (2000). 115 Rajewsky, K., Forster, I. & Cumano, A. Evolutionary and somatic selection of the antibody repertoire in the mouse. Science (New York, N.Y.) 238, 1088‐1094 (1987). 116 Kuppers, R., Zhao, M., Hansmann, M. L. & Rajewsky, K. Tracing B cell development in human germinal centres by molecular analysis of single cells picked from histological sections. The EMBO journal 12, 4955‐4967 (1993). 117 Allen, C. D. et al. Germinal center dark and light zone organization is mediated by CXCR4 and CXCR5. Nature immunology 5, 943‐952, doi:10.1038/ni1100 (2004). 118 Caron, G., Le Gallou, S., Lamy, T., Tarte, K. & Fest, T. CXCR4 expression functionally discriminates centroblasts versus centrocytes within human germinal center B cells. Journal of immunology (Baltimore, Md. : 1950) 182, 7595‐7602, doi:10.4049/jimmunol.0804272 (2009). 119 Victora, G. D. et al. Germinal center dynamics revealed by multiphoton microscopy with a photoactivatable fluorescent reporter. Cell 143, 592‐605, doi:10.1016/j.cell.2010.10.032 (2010). 120 Allen, C. D., Okada, T., Tang, H. L. & Cyster, J. G. Imaging of germinal center selection events during affinity maturation. Science (New York, N.Y.) 315, 528‐531, doi:10.1126/science.1136736 (2007). 121 Schwickert, T. A. et al. In vivo imaging of germinal centres reveals a dynamic open structure. Nature 446, 83‐87, doi:10.1038/nature05573 (2007). 122 Hauser, A. E. et al. Definition of germinal‐center B cell migration in vivo reveals predominant intrazonal circulation patterns. Immunity 26, 655‐667, doi:10.1016/j.immuni.2007.04.008 (2007). 123 Allen, C. D., Okada, T. & Cyster, J. G. Germinal‐center organization and cellular dynamics. Immunity 27, 190‐202, doi:10.1016/j.immuni.2007.07.009 (2007). 124 Victora, G. D. et al. Identification of human germinal center light and dark zone cells and their relationship to human B‐cell lymphomas. Blood 120, 2240‐2248, doi:10.1182/blood‐2012‐03‐ 415380 (2012). 125 Khalil, A. M., Cambier, J. C. & Shlomchik, M. J. B cell receptor signal transduction in the GC is short‐circuited by high phosphatase activity. Science (New York, N.Y.) 336, 1178‐1181, doi:10.1126/science.1213368 (2012). 126 Shulman, Z. et al. Dynamic signaling by T follicular helper cells during germinal center B cell selection. Science (New York, N.Y.) 345, 1058‐1062, doi:10.1126/science.1257861 (2014). 127 Gitlin, A. D., Shulman, Z. & Nussenzweig, M. C. Clonal selection in the germinal centre by regulated proliferation and hypermutation. Nature 509, 637‐640, doi:10.1038/nature13300 (2014). 128 Di Noia, J. M. & Neuberger, M. S. Molecular mechanisms of antibody somatic hypermutation. Annual review of biochemistry 76, 1‐22, doi:10.1146/annurev.biochem.76.061705.090740 (2007). 129 Basso, K. et al. Tracking CD40 signaling during germinal center development. Blood 104, 4088‐ 4096, doi:10.1182/blood‐2003‐12‐4291 (2004). 130 Liu, D. et al. T‐B‐cell entanglement and ICOSL‐driven feed‐forward regulation of germinal centre reaction. Nature 517, 214‐218, doi:10.1038/nature13803 (2015). 131 Toellner, K. M., Gulbranson‐Judge, A., Taylor, D. R., Sze, D. M. & MacLennan, I. C. Immunoglobulin switch transcript production in vivo related to the site and time of antigen‐ specific B cell activation. The Journal of experimental medicine 183, 2303‐2312 (1996).

190

132 Liu, Y. J. et al. Within germinal centers, isotype switching of immunoglobulin genes occurs after the onset of somatic mutation. Immunity 4, 241‐250 (1996). 133 Tafuri, A. et al. ICOS is essential for effective T‐helper‐cell responses. Nature 409, 105‐109, doi:10.1038/35051113 (2001). 134 McAdam, A. J. et al. ICOS is critical for CD40‐mediated antibody class switching. Nature 409, 102‐105, doi:10.1038/35051107 (2001). 135 Xu, Z., Zan, H., Pone, E. J., Mai, T. & Casali, P. Immunoglobulin class‐switch DNA recombination: induction, targeting and beyond. Nature reviews. Immunology 12, 517‐531, doi:10.1038/nri3216 (2012). 136 Vitetta, E. S. et al. Serological, biochemical, and functional identity of B cell‐stimulatory factor 1 and B cell differentiation factor for IgG1. The Journal of experimental medicine 162, 1726‐1731 (1985). 137 Shan, H., Shlomchik, M. & Weigert, M. Heavy‐chain class switch does not terminate somatic mutation. The Journal of experimental medicine 172, 531‐536 (1990). 138 Klein, U. & Dalla‐Favera, R. Germinal centres: role in B‐cell physiology and malignancy. Nature reviews. Immunology 8, 22‐33, doi:10.1038/nri2217 (2008). 139 De Silva, N. S. & Klein, U. Dynamics of B cells in germinal centres. Nature reviews. Immunology 15, 137‐148, doi:10.1038/nri3804 (2015). 140 Allman, D. et al. BCL‐6 expression during B‐cell activation. Blood 87, 5257‐5268 (1996). 141 Cattoretti, G. et al. BCL‐6 protein is expressed in germinal‐center B cells. Blood 86, 45‐53 (1995). 142 Dent, A. L., Shaffer, A. L., Yu, X., Allman, D. & Staudt, L. M. Control of inflammation, cytokine expression, and germinal center formation by BCL‐6. Science (New York, N.Y.) 276, 589‐592 (1997). 143 Ye, B. H. et al. The BCL‐6 proto‐oncogene controls germinal‐centre formation and Th2‐type inflammation. Nature genetics 16, 161‐170, doi:10.1038/ng0697‐161 (1997). 144 Huang, C. et al. The BCL6 RD2 domain governs commitment of activated B cells to form germinal centers. Cell reports 8, 1497‐1508, doi:10.1016/j.celrep.2014.07.059 (2014). 145 Niu, H., Cattoretti, G. & Dalla‐Favera, R. BCL6 controls the expression of the B7‐1/CD80 costimulatory receptor in germinal center B cells. The Journal of experimental medicine 198, 211‐221, doi:10.1084/jem.20021395 (2003). 146 Basso, K. et al. Integrated biochemical and computational approach identifies BCL6 direct target genes controlling multiple pathways in normal germinal center B cells. Blood 115, 975‐984, doi:10.1182/blood‐2009‐06‐227017 (2010). 147 Ci, W. et al. The BCL6 transcriptional program features repression of multiple oncogenes in primary B cells and is deregulated in DLBCL. Blood 113, 5536‐5548, doi:10.1182/blood‐2008‐12‐ 193037 (2009). 148 Shaffer, A. L. et al. BCL‐6 represses genes that function in lymphocyte differentiation, inflammation, and cell cycle control. Immunity 13, 199‐212 (2000). 149 Tunyaplin, C. et al. Direct repression of prdm1 by Bcl‐6 inhibits plasmacytic differentiation. Journal of immunology (Baltimore, Md. : 1950) 173, 1158‐1165 (2004). 150 Ding, B. B. et al. Constitutively activated STAT3 promotes cell proliferation and survival in the activated B‐cell subtype of diffuse large B‐cell lymphomas. Blood 111, 1515‐1523, doi:10.1182/blood‐2007‐04‐087734 (2008). 151 Phan, R. T. & Dalla‐Favera, R. The BCL6 proto‐oncogene suppresses p53 expression in germinal‐ centre B cells. Nature 432, 635‐639, doi:10.1038/nature03147 (2004). 152 Phan, R. T., Saito, M., Basso, K., Niu, H. & Dalla‐Favera, R. BCL6 interacts with the transcription factor Miz‐1 to suppress the cyclin‐dependent kinase inhibitor p21 and cell cycle arrest in germinal center B cells. Nature immunology 6, 1054‐1060, doi:10.1038/ni1245 (2005).

191

153 Ranuncolo, S. M. et al. Bcl‐6 mediates the germinal center B cell phenotype and lymphomagenesis through transcriptional repression of the DNA‐damage sensor ATR. Nature immunology 8, 705‐714, doi:10.1038/ni1478 (2007). 154 Ranuncolo, S. M., Polo, J. M. & Melnick, A. BCL6 represses CHEK1 and suppresses DNA damage pathways in normal and malignant B‐cells. Blood cells, molecules & diseases 41, 95‐99, doi:10.1016/j.bcmd.2008.02.003 (2008). 155 Saito, M. et al. BCL6 suppression of BCL2 via Miz1 and its disruption in diffuse large B cell lymphoma. Proceedings of the National Academy of Sciences of the United States of America 106, 11294‐11299, doi:10.1073/pnas.0903854106 (2009). 156 Klein, U. et al. Transcriptional analysis of the B cell germinal center reaction. Proceedings of the National Academy of Sciences of the United States of America 100, 2639‐2644, doi:10.1073/pnas.0437996100 (2003). 157 Liu, Y. J. et al. Mechanism of antigen‐driven selection in germinal centres. Nature 342, 929‐931, doi:10.1038/342929a0 (1989). 158 Feuillard, J., Taylor, D., Casamayor‐Palleja, M., Johnson, G. D. & MacLennan, I. C. Isolation and characteristics of tonsil centroblasts with reference to Ig class switching. International immunology 7, 121‐130 (1995). 159 Billian, G., Bella, C., Mondiere, P. & Defrance, T. Identification of a tonsil IgD+ B cell subset with phenotypical and functional characteristics of germinal center B cells. European journal of immunology 26, 1712‐1719, doi:10.1002/eji.1830260808 (1996). 160 Dang, C. V. MYC on the path to cancer. Cell 149, 22‐35, doi:10.1016/j.cell.2012.03.003 (2012). 161 Shaffer, A. L. et al. Signatures of the immune response. Immunity 15, 375‐385 (2001). 162 Dominguez‐Sola, D. et al. The proto‐oncogene MYC is required for selection in the germinal center and cyclic reentry. Nature immunology 13, 1083‐1091, doi:10.1038/ni.2428 (2012). 163 Calado, D. P. et al. The cell‐cycle regulator c‐Myc is essential for the formation and maintenance of germinal centers. Nature immunology 13, 1092‐1100, doi:10.1038/ni.2418 (2012). 164 Peled, J. U. et al. Requirement for cyclin D3 in germinal center formation and function. Cell research 20, 631‐646, doi:10.1038/cr.2010.55 (2010). 165 Cato, M. H., Chintalapati, S. K., Yau, I. W., Omori, S. A. & Rickert, R. C. Cyclin D3 is selectively required for proliferative expansion of germinal center B cells. Molecular and cellular biology 31, 127‐137, doi:10.1128/mcb.00650‐10 (2011). 166 Jacob, J., Przylepa, J., Miller, C. & Kelsoe, G. In situ studies of the primary immune response to (4‐hydroxy‐3‐nitrophenyl)acetyl. III. The kinetics of V region mutation and selection in germinal center B cells. The Journal of experimental medicine 178, 1293‐1307 (1993). 167 Vikstrom, I. et al. Mcl‐1 is essential for germinal center formation and B cell memory. Science (New York, N.Y.) 330, 1095‐1099, doi:10.1126/science.1191793 (2010). 168 Carrington, E. M. et al. BH3 mimetics antagonizing restricted prosurvival Bcl‐2 proteins represent another class of selective immune modulatory drugs. Proceedings of the National Academy of Sciences of the United States of America 107, 10967‐10971, doi:10.1073/pnas.1005256107 (2010). 169 Kurosaki, T., Kometani, K. & Ise, W. Memory B cells. Nature reviews. Immunology 15, 149‐159, doi:10.1038/nri3802 (2015). 170 MacLennan, I. C. et al. Extrafollicular antibody responses. Immunological reviews 194, 8‐18 (2003). 171 Puga, I. et al. B cell‐helper neutrophils stimulate the diversification and production of immunoglobulin in the marginal zone of the spleen. Nature immunology 13, 170‐180, doi:10.1038/ni.2194 (2012).

192

172 Hargreaves, D. C. et al. A coordinated change in chemokine responsiveness guides plasma cell movements. The Journal of experimental medicine 194, 45‐56 (2001). 173 Ellyard, J. I., Avery, D. T., Mackay, C. R. & Tangye, S. G. Contribution of stromal cells to the migration, function and retention of plasma cells in human spleen: potential roles of CXCL12, IL‐ 6 and CD54. European journal of immunology 35, 699‐708, doi:10.1002/eji.200425442 (2005). 174 Jacob, J. & Kelsoe, G. In situ studies of the primary immune response to (4‐hydroxy‐3‐ nitrophenyl)acetyl. II. A common clonal origin for periarteriolar lymphoid sheath‐associated foci and germinal centers. The Journal of experimental medicine 176, 679‐687 (1992). 175 Paus, D. et al. Antigen recognition strength regulates the choice between extrafollicular plasma cell and germinal center B cell differentiation. The Journal of experimental medicine 203, 1081‐ 1091, doi:10.1084/jem.20060087 (2006). 176 O'Connor, B. P. et al. Imprinting the fate of antigen‐reactive B cells through the affinity of the B cell receptor. Journal of immunology (Baltimore, Md. : 1950) 177, 7723‐7732 (2006). 177 Jacob, J., Kassir, R. & Kelsoe, G. In situ studies of the primary immune response to (4‐hydroxy‐3‐ nitrophenyl)acetyl. I. The architecture and dynamics of responding cell populations. The Journal of experimental medicine 173, 1165‐1175 (1991). 178 Smith, K. G., Hewitson, T. D., Nossal, G. J. & Tarlinton, D. M. The phenotype and fate of the antibody‐forming cells of the splenic foci. European journal of immunology 26, 444‐448, doi:10.1002/eji.1830260226 (1996). 179 Nutt, S. L., Heavey, B., Rolink, A. G. & Busslinger, M. Commitment to the B‐lymphoid lineage depends on the transcription factor Pax5. Nature 401, 556‐562, doi:10.1038/44076 (1999). 180 Cobaleda, C., Jochum, W. & Busslinger, M. Conversion of mature B cells into T cells by dedifferentiation to uncommitted progenitors. Nature 449, 473‐477, doi:10.1038/nature06159 (2007). 181 Cobaleda, C., Schebesta, A., Delogu, A. & Busslinger, M. Pax5: the guardian of B cell identity and function. Nature immunology 8, 463‐470, doi:10.1038/ni1454 (2007). 182 Revilla, I. D. R. et al. The B‐cell identity factor Pax5 regulates distinct transcriptional programmes in early and late B lymphopoiesis. The EMBO journal 31, 3130‐3146, doi:10.1038/emboj.2012.155 (2012). 183 Schebesta, A. et al. Transcription factor Pax5 activates the chromatin of key genes involved in B cell signaling, adhesion, migration, and immune function. Immunity 27, 49‐63, doi:10.1016/j.immuni.2007.05.019 (2007). 184 Pridans, C. et al. Identification of Pax5 target genes in early B cell differentiation. Journal of immunology (Baltimore, Md. : 1950) 180, 1719‐1728 (2008). 185 Delogu, A. et al. Gene repression by Pax5 in B cells is essential for blood cell homeostasis and is reversed in plasma cells. Immunity 24, 269‐281, doi:10.1016/j.immuni.2006.01.012 (2006). 186 Reimold, A. M. et al. Transcription factor B cell lineage‐specific activator protein regulates the gene for human X‐box binding protein 1. The Journal of experimental medicine 183, 393‐401 (1996). 187 Nera, K. P. et al. Loss of Pax5 promotes plasma cell differentiation. Immunity 24, 283‐293, doi:10.1016/j.immuni.2006.02.003 (2006). 188 Rinkenberger, J. L., Wallin, J. J., Johnson, K. W. & Koshland, M. E. An interleukin‐2 signal relieves BSAP (Pax5)‐mediated repression of the immunoglobulin J chain gene. Immunity 5, 377‐386 (1996). 189 Usui, T. et al. Overexpression of B cell‐specific activator protein (BSAP/Pax‐5) in a late B cell is sufficient to suppress differentiation to an Ig high producer cell with plasma cell phenotype. Journal of immunology (Baltimore, Md. : 1950) 158, 3197‐3204 (1997).

193

190 Lin, K. I., Angelin‐Duclos, C., Kuo, T. C. & Calame, K. Blimp‐1‐dependent repression of Pax‐5 is required for differentiation of B cells to immunoglobulin M‐secreting plasma cells. Molecular and cellular biology 22, 4771‐4780 (2002). 191 Kallies, A. et al. Initiation of plasma‐cell differentiation is independent of the transcription factor Blimp‐1. Immunity 26, 555‐566, doi:10.1016/j.immuni.2007.04.007 (2007). 192 Huang, C., Geng, H., Boss, I., Wang, L. & Melnick, A. Cooperative transcriptional repression by BCL6 and BACH2 in germinal center B‐cell differentiation. Blood 123, 1012‐1020, doi:10.1182/blood‐2013‐07‐518605 (2014). 193 Ochiai, K. et al. Plasmacytic transcription factor Blimp‐1 is repressed by Bach2 in B cells. The Journal of biological chemistry 281, 38226‐38234, doi:10.1074/jbc.M607592200 (2006). 194 Lin, L., Gerth, A. J. & Peng, S. L. Active inhibition of plasma cell development in resting B cells by microphthalmia‐associated transcription factor. The Journal of experimental medicine 200, 115‐ 122, doi:10.1084/jem.20040612 (2004). 195 Shapiro‐Shelef, M. & Calame, K. Regulation of plasma‐cell development. Nature reviews. Immunology 5, 230‐242, doi:10.1038/nri1572 (2005). 196 De Silva, N. S., Simonetti, G., Heise, N. & Klein, U. The diverse roles of IRF4 in late germinal center B‐cell differentiation. Immunological reviews 247, 73‐92, doi:10.1111/j.1600‐ 065X.2012.01113.x (2012). 197 Shaffer, A. L., Emre, N. C., Romesser, P. B. & Staudt, L. M. IRF4: Immunity. Malignancy! Therapy? Clinical cancer research : an official journal of the American Association for Cancer Research 15, 2954‐2961, doi:10.1158/1078‐0432.ccr‐08‐1845 (2009). 198 Ochiai, K. et al. Transcriptional regulation of germinal center B and plasma cell fates by dynamical control of IRF4. Immunity 38, 918‐929, doi:10.1016/j.immuni.2013.04.009 (2013). 199 Willis, S. N. et al. Transcription factor IRF4 regulates germinal center cell formation through a B cell‐intrinsic mechanism. Journal of immunology (Baltimore, Md. : 1950) 192, 3200‐3206, doi:10.4049/jimmunol.1303216 (2014). 200 Klein, U. et al. Transcription factor IRF4 controls plasma cell differentiation and class‐switch recombination. Nature immunology 7, 773‐782, doi:10.1038/ni1357 (2006). 201 Sciammas, R. et al. Graded expression of interferon regulatory factor‐4 coordinates isotype switching with plasma cell differentiation. Immunity 25, 225‐236, doi:10.1016/j.immuni.2006.07.009 (2006). 202 Bollig, N. et al. Transcription factor IRF4 determines germinal center formation through follicular T‐helper cell differentiation. Proceedings of the National Academy of Sciences of the United States of America 109, 8664‐8669, doi:10.1073/pnas.1205834109 (2012). 203 Saito, M. et al. A signaling pathway mediating downregulation of BCL6 in germinal center B cells is blocked by BCL6 gene alterations in B cell lymphoma. Cancer cell 12, 280‐292, doi:10.1016/j.ccr.2007.08.011 (2007). 204 Kallies, A. et al. Plasma cell ontogeny defined by quantitative changes in blimp‐1 expression. The Journal of experimental medicine 200, 967‐977, doi:10.1084/jem.20040973 (2004). 205 Angelin‐Duclos, C., Cattoretti, G., Lin, K. I. & Calame, K. Commitment of B lymphocytes to a plasma cell fate is associated with Blimp‐1 expression in vivo. Journal of immunology (Baltimore, Md. : 1950) 165, 5462‐5471 (2000). 206 Soro, P. G. et al. Differential involvement of the transcription factor Blimp‐1 in T cell‐ independent and ‐dependent B cell differentiation to plasma cells. Journal of immunology (Baltimore, Md. : 1950) 163, 611‐617 (1999). 207 Shapiro‐Shelef, M. et al. Blimp‐1 is required for the formation of immunoglobulin secreting plasma cells and pre‐plasma memory B cells. Immunity 19, 607‐620 (2003).

194

208 Turner, C. A., Jr., Mack, D. H. & Davis, M. M. Blimp‐1, a novel zinc finger‐containing protein that can drive the maturation of B lymphocytes into immunoglobulin‐secreting cells. Cell 77, 297‐306 (1994). 209 Shapiro‐Shelef, M., Lin, K. I., Savitsky, D., Liao, J. & Calame, K. Blimp‐1 is required for maintenance of long‐lived plasma cells in the bone marrow. The Journal of experimental medicine 202, 1471‐1476, doi:10.1084/jem.20051611 (2005). 210 Shaffer, A. L. et al. XBP1, downstream of Blimp‐1, expands the secretory apparatus and other organelles, and increases protein synthesis in plasma cell differentiation. Immunity 21, 81‐93, doi:10.1016/j.immuni.2004.06.010 (2004). 211 Shaffer, A. L. et al. Blimp‐1 orchestrates plasma cell differentiation by extinguishing the mature B cell gene expression program. Immunity 17, 51‐62 (2002). 212 Lin, K. I., Lin, Y. & Calame, K. Repression of c‐myc is necessary but not sufficient for terminal differentiation of B lymphocytes in vitro. Molecular and cellular biology 20, 8684‐8695 (2000). 213 Lin, Y., Wong, K. & Calame, K. Repression of c‐myc transcription by Blimp‐1, an inducer of terminal B cell differentiation. Science (New York, N.Y.) 276, 596‐599 (1997). 214 Fairfax, K. A. et al. Different kinetics of blimp‐1 induction in B cell subsets revealed by reporter gene. Journal of immunology (Baltimore, Md. : 1950) 178, 4104‐4111 (2007). 215 Iwakoshi, N. N., Lee, A. H. & Glimcher, L. H. The X‐box binding protein‐1 transcription factor is required for plasma cell differentiation and the unfolded protein response. Immunological reviews 194, 29‐38 (2003). 216 Todd, D. J. et al. XBP1 governs late events in plasma cell differentiation and is not required for antigen‐specific memory B cell development. The Journal of experimental medicine 206, 2151‐ 2159, doi:10.1084/jem.20090738 (2009). 217 Taubenheim, N. et al. High rate of antibody secretion is not integral to plasma cell differentiation as revealed by XBP‐1 deficiency. Journal of immunology (Baltimore, Md. : 1950) 189, 3328‐3338, doi:10.4049/jimmunol.1201042 (2012). 218 Hu, C. C., Dougan, S. K., McGehee, A. M., Love, J. C. & Ploegh, H. L. XBP‐1 regulates signal transduction, transcription factors and bone marrow colonization in B cells. The EMBO journal 28, 1624‐1636, doi:10.1038/emboj.2009.117 (2009). 219 Reimold, A. M. et al. Plasma cell differentiation requires the transcription factor XBP‐1. Nature 412, 300‐307, doi:10.1038/35085509 (2001). 220 Radbruch, A. et al. Competence and competition: the challenge of becoming a long‐lived plasma cell. Nature reviews. Immunology 6, 741‐750, doi:10.1038/nri1886 (2006). 221 Kabashima, K. et al. Plasma cell S1P1 expression determines secondary lymphoid organ retention versus bone marrow tropism. The Journal of experimental medicine 203, 2683‐2690, doi:10.1084/jem.20061289 (2006). 222 Tokoyoda, K., Egawa, T., Sugiyama, T., Choi, B. I. & Nagasawa, T. Cellular niches controlling B lymphocyte behavior within bone marrow during development. Immunity 20, 707‐718, doi:10.1016/j.immuni.2004.05.001 (2004). 223 Hauser, A. E. et al. Chemotactic responsiveness toward ligands for CXCR3 and CXCR4 is regulated on plasma blasts during the time course of a memory immune response. Journal of immunology (Baltimore, Md. : 1950) 169, 1277‐1282 (2002). 224 Odendahl, M. et al. Generation of migratory antigen‐specific plasma blasts and mobilization of resident plasma cells in a secondary immune response. Blood 105, 1614‐1621, doi:10.1182/blood‐2004‐07‐2507 (2005). 225 Kunkel, E. J. & Butcher, E. C. Plasma‐cell homing. Nature reviews. Immunology 3, 822‐829, doi:10.1038/nri1203 (2003).

195

226 Chu, V. T. & Berek, C. The establishment of the plasma cell survival niche in the bone marrow. Immunological reviews 251, 177‐188, doi:10.1111/imr.12011 (2013). 227 Nutt, S. L., Hodgkin, P. D., Tarlinton, D. M. & Corcoran, L. M. The generation of antibody‐ secreting plasma cells. Nature reviews. Immunology 15, 160‐171, doi:10.1038/nri3795 (2015). 228 Minges Wols, H. A., Underhill, G. H., Kansas, G. S. & Witte, P. L. The role of bone marrow‐derived stromal cells in the maintenance of plasma cell longevity. Journal of immunology (Baltimore, Md. : 1950) 169, 4213‐4221 (2002). 229 Sciammas, R. & Davis, M. M. Modular nature of Blimp‐1 in the regulation of gene expression during B cell maturation. Journal of immunology (Baltimore, Md. : 1950) 172, 5427‐5440 (2004). 230 Chu, V. T. et al. Eosinophils are required for the maintenance of plasma cells in the bone marrow. Nature immunology 12, 151‐159, doi:10.1038/ni.1981 (2011). 231 O'Connor, B. P. et al. BCMA is essential for the survival of long‐lived bone marrow plasma cells. The Journal of experimental medicine 199, 91‐98, doi:10.1084/jem.20031330 (2004). 232 Peperzak, V. et al. Mcl‐1 is essential for the survival of plasma cells. Nature immunology 14, 290‐ 297, doi:10.1038/ni.2527 (2013). 233 Belnoue, E. et al. APRIL is critical for plasmablast survival in the bone marrow and poorly expressed by early‐life bone marrow stromal cells. Blood 111, 2755‐2764, doi:10.1182/blood‐ 2007‐09‐110858 (2008). 234 Benson, M. J. et al. Cutting edge: the dependence of plasma cells and independence of memory B cells on BAFF and APRIL. Journal of immunology (Baltimore, Md. : 1950) 180, 3655‐3659 (2008). 235 Marsters, S. A. et al. Interaction of the TNF homologues BLyS and APRIL with the TNF receptor homologues BCMA and TACI. Current biology : CB 10, 785‐788 (2000). 236 Kawano, M. M., Mihara, K., Huang, N., Tsujimoto, T. & Kuramoto, A. Differentiation of early plasma cells on bone marrow stromal cells requires interleukin‐6 for escaping from apoptosis. Blood 85, 487‐494 (1995). 237 Cassese, G. et al. Plasma cell survival is mediated by synergistic effects of cytokines and adhesion‐dependent signals. Journal of immunology (Baltimore, Md. : 1950) 171, 1684‐1690 (2003). 238 Kuppers, R., Klein, U., Hansmann, M. L. & Rajewsky, K. Cellular origin of human B‐cell lymphomas. The New England journal of medicine 341, 1520‐1529, doi:10.1056/nejm199911113412007 (1999). 239 Stevenson, F. et al. Insight into the origin and clonal history of B‐cell tumors as revealed by analysis of immunoglobulin variable region genes. Immunological reviews 162, 247‐259 (1998). 240 Lohr, J. G. et al. Discovery and prioritization of somatic mutations in diffuse large B‐cell lymphoma (DLBCL) by whole‐exome sequencing. Proceedings of the National Academy of Sciences of the United States of America 109, 3879‐3884, doi:10.1073/pnas.1121343109 (2012). 241 Morin, R. D. et al. Frequent mutation of histone‐modifying genes in non‐Hodgkin lymphoma. Nature 476, 298‐303, doi:10.1038/nature10351 (2011). 242 Pasqualucci, L. et al. Analysis of the coding genome of diffuse large B‐cell lymphoma. Nature genetics 43, 830‐837, doi:10.1038/ng.892 (2011). 243 Zhang, J. et al. Genetic heterogeneity of diffuse large B‐cell lymphoma. Proceedings of the National Academy of Sciences of the United States of America 110, 1398‐1403, doi:10.1073/pnas.1205299110 (2013). 244 Richter, J. et al. Recurrent mutation of the ID3 gene in Burkitt lymphoma identified by integrated genome, exome and transcriptome sequencing. Nature genetics 44, 1316‐1320, doi:10.1038/ng.2469 (2012).

196

245 Schmitz, R. et al. Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature 490, 116‐120, doi:10.1038/nature11378 (2012). 246 Stevenson, F. K. et al. The occurrence and significance of V gene mutations in B cell‐derived human malignancy. Advances in cancer research 83, 81‐116 (2001). 247 Kuppers, R. & Dalla‐Favera, R. Mechanisms of chromosomal translocations in B cell lymphomas. Oncogene 20, 5580‐5594, doi:10.1038/sj.onc.1204640 (2001). 248 Kuppers, R. Mechanisms of B‐cell lymphoma pathogenesis. Nature reviews. Cancer 5, 251‐262, doi:10.1038/nrc1589 (2005). 249 Lenz, G. et al. Aberrant immunoglobulin class switch recombination and switch translocations in activated B cell‐like diffuse large B cell lymphoma. The Journal of experimental medicine 204, 633‐643, doi:10.1084/jem.20062041 (2007). 250 Pasqualucci, L. et al. Hypermutation of multiple proto‐oncogenes in B‐cell diffuse large‐cell lymphomas. Nature 412, 341‐346, doi:10.1038/35085588 (2001). 251 Khodabakhshi, A. H. et al. Recurrent targets of aberrant somatic hypermutation in lymphoma. Oncotarget 3, 1308‐1319 (2012). 252 Kotani, A. et al. Activation‐induced cytidine deaminase (AID) promotes B cell lymphomagenesis in Emu‐cmyc transgenic mice. Proceedings of the National Academy of Sciences of the United States of America 104, 1616‐1620, doi:10.1073/pnas.0610732104 (2007). 253 Pasqualucci, L. et al. AID is required for germinal center‐derived lymphomagenesis. Nature genetics 40, 108‐112, doi:10.1038/ng.2007.35 (2008). 254 Dave, S. S. et al. Molecular diagnosis of Burkitt's lymphoma. The New England journal of medicine 354, 2431‐2442, doi:10.1056/NEJMoa055759 (2006). 255 Alizadeh, A. A. et al. Distinct types of diffuse large B‐cell lymphoma identified by gene expression profiling. Nature 403, 503‐511, doi:10.1038/35000501 (2000). 256 Basso, K. & Dalla‐Favera, R. Germinal centres and B cell lymphomagenesis. Nature reviews. Immunology 15, 172‐184, doi:10.1038/nri3814 (2015). 257 Dalla‐Favera, R. et al. Human c‐myc onc gene is located on the region of chromosome 8 that is translocated in Burkitt lymphoma cells. Proceedings of the National Academy of Sciences of the United States of America 79, 7824‐7827 (1982). 258 Taub, R. et al. Translocation of the c‐myc gene into the immunoglobulin heavy chain locus in human Burkitt lymphoma and murine plasmacytoma cells. Proceedings of the National Academy of Sciences of the United States of America 79, 7837‐7841 (1982). 259 Vafa, O. et al. c‐Myc can induce DNA damage, increase reactive oxygen species, and mitigate p53 function: a mechanism for oncogene‐induced genetic instability. Molecular cell 9, 1031‐ 1044 (2002). 260 Dominguez‐Sola, D. et al. Non‐transcriptional control of DNA replication by c‐Myc. Nature 448, 445‐451, doi:10.1038/nature05953 (2007). 261 Rosenwald, A. et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large‐B‐cell lymphoma. The New England journal of medicine 346, 1937‐1947, doi:10.1056/NEJMoa012914 (2002). 262 Pasqualucci, L. & Dalla‐Favera, R. SnapShot: diffuse large B cell lymphoma. Cancer cell 25, 132‐ 132.e131, doi:10.1016/j.ccr.2013.12.012 (2014). 263 Chen, W., Iida, S., Louie, D. C., Dalla‐Favera, R. & Chaganti, R. S. Heterologous promoters fused to BCL6 by chromosomal translocations affecting band 3q27 cause its deregulated expression during B‐cell differentiation. Blood 91, 603‐607 (1998). 264 Ye, B. H. et al. Chromosomal translocations cause deregulated BCL6 expression by promoter substitution in B cell lymphoma. The EMBO journal 14, 6209‐6217 (1995).

197

265 Cattoretti, G. et al. Deregulated BCL6 expression recapitulates the pathogenesis of human diffuse large B cell lymphomas in mice. Cancer cell 7, 445‐455, doi:10.1016/j.ccr.2005.03.037 (2005). 266 Pasqualucci, L. et al. Mutations of the BCL6 proto‐oncogene disrupt its negative autoregulation in diffuse large B‐cell lymphoma. Blood 101, 2914‐2923, doi:10.1182/blood‐2002‐11‐3387 (2003). 267 Basso, K. & Dalla‐Favera, R. Roles of BCL6 in normal and transformed germinal center B cells. Immunological reviews 247, 172‐183, doi:10.1111/j.1600‐065X.2012.01112.x (2012). 268 Sneeringer, C. J. et al. Coordinated activities of wild‐type plus mutant EZH2 drive tumor‐ associated hypertrimethylation of lysine 27 on histone H3 (H3K27) in human B‐cell lymphomas. Proceedings of the National Academy of Sciences of the United States of America 107, 20980‐ 20985, doi:10.1073/pnas.1012525107 (2010). 269 Beguelin, W. et al. EZH2 is required for germinal center formation and somatic EZH2 mutations promote lymphoid transformation. Cancer cell 23, 677‐692, doi:10.1016/j.ccr.2013.04.011 (2013). 270 Caganova, M. et al. Germinal center dysregulation by histone methyltransferase EZH2 promotes lymphomagenesis. The Journal of clinical investigation 123, 5009‐5022, doi:10.1172/jci70626 (2013). 271 Muppidi, J. R. et al. Loss of signalling via Galpha13 in germinal centre B‐cell‐derived lymphoma. Nature 516, 254‐258, doi:10.1038/nature13765 (2014). 272 Cattoretti, G. et al. Targeted disruption of the S1P2 sphingosine 1‐phosphate receptor gene leads to diffuse large B‐cell lymphoma formation. Cancer research 69, 8686‐8692, doi:10.1158/0008‐5472.can‐09‐1110 (2009). 273 Pasqualucci, L. et al. Inactivation of the PRDM1/BLIMP1 gene in diffuse large B cell lymphoma. The Journal of experimental medicine 203, 311‐317, doi:10.1084/jem.20052204 (2006). 274 Tam, W. et al. Mutational analysis of PRDM1 indicates a tumor‐suppressor role in diffuse large B‐cell lymphomas. Blood 107, 4090‐4100, doi:10.1182/blood‐2005‐09‐3778 (2006). 275 Mandelbaum, J. et al. BLIMP1 is a tumor suppressor gene frequently disrupted in activated B cell‐like diffuse large B cell lymphoma. Cancer cell 18, 568‐579, doi:10.1016/j.ccr.2010.10.030 (2010). 276 Davis, R. E., Brown, K. D., Siebenlist, U. & Staudt, L. M. Constitutive nuclear factor kappaB activity is required for survival of activated B cell‐like diffuse large B cell lymphoma cells. The Journal of experimental medicine 194, 1861‐1874 (2001). 277 Hayden, M. S. & Ghosh, S. Signaling to NF‐kappaB. Genes & development 18, 2195‐2224, doi:10.1101/gad.1228704 (2004). 278 Vallabhapurapu, S. & Karin, M. Regulation and function of NF‐kappaB transcription factors in the immune system. Annual review of immunology 27, 693‐733, doi:10.1146/annurev.immunol.021908.132641 (2009). 279 Siebenlist, U., Brown, K. & Claudio, E. Control of lymphocyte development by nuclear factor‐ kappaB. Nature reviews. Immunology 5, 435‐445, doi:10.1038/nri1629 (2005). 280 Sun, S. C. The noncanonical NF‐kappaB pathway. Immunological reviews 246, 125‐140, doi:10.1111/j.1600‐065X.2011.01088.x (2012). 281 Dejardin, E. et al. The lymphotoxin‐beta receptor induces different patterns of gene expression via two NF‐kappaB pathways. Immunity 17, 525‐535 (2002). 282 Yin, L. et al. Defective lymphotoxin‐beta receptor‐induced NF‐kappaB transcriptional activity in NIK‐deficient mice. Science (New York, N.Y.) 291, 2162‐2165, doi:10.1126/science.1058453 (2001).

198

283 Coope, H. J. et al. CD40 regulates the processing of NF‐kappaB2 p100 to p52. The EMBO journal 21, 5375‐5385 (2002). 284 Qing, G., Qu, Z. & Xiao, G. Stabilization of basally translated NF‐kappaB‐inducing kinase (NIK) protein functions as a molecular switch of processing of NF‐kappaB2 p100. The Journal of biological chemistry 280, 40578‐40582, doi:10.1074/jbc.M508776200 (2005). 285 Novack, D. V. et al. The IkappaB function of NF‐kappaB2 p100 controls stimulated osteoclastogenesis. The Journal of experimental medicine 198, 771‐781, doi:10.1084/jem.20030116 (2003). 286 Claudio, E., Brown, K., Park, S., Wang, H. & Siebenlist, U. BAFF‐induced NEMO‐independent processing of NF‐kappa B2 in maturing B cells. Nature immunology 3, 958‐965, doi:10.1038/ni842 (2002). 287 Kayagaki, N. et al. BAFF/BLyS receptor 3 binds the B cell survival factor BAFF ligand through a discrete surface loop and promotes processing of NF‐kappaB2. Immunity 17, 515‐524 (2002). 288 Morrison, M. D., Reiley, W., Zhang, M. & Sun, S. C. An atypical tumor necrosis factor (TNF) receptor‐associated factor‐binding motif of B cell‐activating factor belonging to the TNF family (BAFF) receptor mediates induction of the noncanonical NF‐kappaB signaling pathway. The Journal of biological chemistry 280, 10018‐10024, doi:10.1074/jbc.M413634200 (2005). 289 Berberich, I., Shu, G. L. & Clark, E. A. Cross‐linking CD40 on B cells rapidly activates nuclear factor‐kappa B. Journal of immunology (Baltimore, Md. : 1950) 153, 4357‐4366 (1994). 290 Ramakrishnan, P., Wang, W. & Wallach, D. Receptor‐specific signaling for both the alternative and the canonical NF‐kappaB activation pathways by NF‐kappaB‐inducing kinase. Immunity 21, 477‐489, doi:10.1016/j.immuni.2004.08.009 (2004). 291 Gerondakis, S. & Siebenlist, U. Roles of the NF‐kappaB pathway in lymphocyte development and function. Cold Spring Harbor perspectives in biology 2, a000182, doi:10.1101/cshperspect.a000182 (2010). 292 Kaileh, M. & Sen, R. NF‐kappaB function in B lymphocytes. Immunological reviews 246, 254‐271, doi:10.1111/j.1600‐065X.2012.01106.x (2012). 293 Tam, W. F., Lee, L. H., Davis, L. & Sen, R. Cytoplasmic sequestration of rel proteins by IkappaBalpha requires CRM1‐dependent nuclear export. Molecular and cellular biology 20, 2269‐2284 (2000). 294 Lee, S. H. & Hannink, M. The N‐terminal nuclear export sequence of IkappaBalpha is required for RanGTP‐dependent binding to CRM1. The Journal of biological chemistry 276, 23599‐23606, doi:10.1074/jbc.M011197200 (2001). 295 He, J. Q., Saha, S. K., Kang, J. R., Zarnegar, B. & Cheng, G. Specificity of TRAF3 in its negative regulation of the noncanonical NF‐kappa B pathway. The Journal of biological chemistry 282, 3688‐3694, doi:10.1074/jbc.M610271200 (2007). 296 Liao, G., Zhang, M., Harhaj, E. W. & Sun, S. C. Regulation of the NF‐kappaB‐inducing kinase by tumor necrosis factor receptor‐associated factor 3‐induced degradation. The Journal of biological chemistry 279, 26243‐26250, doi:10.1074/jbc.M403286200 (2004). 297 Zarnegar, B. J. et al. Noncanonical NF‐kappaB activation requires coordinated assembly of a regulatory complex of the adaptors cIAP1, cIAP2, TRAF2 and TRAF3 and the kinase NIK. Nature immunology 9, 1371‐1378, doi:10.1038/ni.1676 (2008). 298 Vallabhapurapu, S. et al. Nonredundant and complementary functions of TRAF2 and TRAF3 in a ubiquitination cascade that activates NIK‐dependent alternative NF‐kappaB signaling. Nature immunology 9, 1364‐1370, doi:10.1038/ni.1678 (2008). 299 Mahoney, D. J. et al. Both cIAP1 and cIAP2 regulate TNFalpha‐mediated NF‐kappaB activation. Proceedings of the National Academy of Sciences of the United States of America 105, 11778‐ 11783, doi:10.1073/pnas.0711122105 (2008).

199

300 Varfolomeev, E. et al. c‐IAP1 and c‐IAP2 are critical mediators of tumor necrosis factor alpha (TNFalpha)‐induced NF‐kappaB activation. The Journal of biological chemistry 283, 24295‐24299, doi:10.1074/jbc.C800128200 (2008). 301 He, J. Q. et al. Rescue of TRAF3‐null mice by p100 NF‐kappa B deficiency. The Journal of experimental medicine 203, 2413‐2418, doi:10.1084/jem.20061166 (2006). 302 Grech, A. P. et al. TRAF2 differentially regulates the canonical and noncanonical pathways of NF‐ kappaB activation in mature B cells. Immunity 21, 629‐642, doi:10.1016/j.immuni.2004.09.011 (2004). 303 Senftleben, U. et al. Activation by IKKalpha of a second, evolutionary conserved, NF‐kappa B signaling pathway. Science (New York, N.Y.) 293, 1495‐1499, doi:10.1126/science.1062677 (2001). 304 Xiao, G., Harhaj, E. W. & Sun, S. C. NF‐kappaB‐inducing kinase regulates the processing of NF‐ kappaB2 p100. Molecular cell 7, 401‐409 (2001). 305 Davis, R. E. et al. Chronic active B‐cell‐receptor signalling in diffuse large B‐cell lymphoma. Nature 463, 88‐92, doi:10.1038/nature08638 (2010). 306 Jun, J. E. et al. Identifying the MAGUK protein Carma‐1 as a central regulator of humoral immune responses and atopy by genome‐wide mouse mutagenesis. Immunity 18, 751‐762 (2003). 307 Tanner, M. J., Hanel, W., Gaffen, S. L. & Lin, X. CARMA1 coiled‐coil domain is involved in the oligomerization and subcellular localization of CARMA1 and is required for T cell receptor‐ induced NF‐kappaB activation. The Journal of biological chemistry 282, 17141‐17147, doi:10.1074/jbc.M700169200 (2007). 308 Lenz, G. et al. Oncogenic CARD11 mutations in human diffuse large B cell lymphoma. Science (New York, N.Y.) 319, 1676‐1679, doi:10.1126/science.1153629 (2008). 309 Ngo, V. N. et al. Oncogenically active MYD88 mutations in human lymphoma. Nature 470, 115‐ 119, doi:10.1038/nature09671 (2011). 310 Wang, J. Q., Jeelall, Y. S., Beutler, B., Horikawa, K. & Goodnow, C. C. Consequences of the recurrent MYD88(L265P) somatic mutation for B cell tolerance. The Journal of experimental medicine 211, 413‐426, doi:10.1084/jem.20131424 (2014). 311 Compagno, M. et al. Mutations of multiple genes cause deregulation of NF‐kappaB in diffuse large B‐cell lymphoma. Nature 459, 717‐721, doi:10.1038/nature07968 (2009). 312 Kato, M. et al. Frequent inactivation of A20 in B‐cell lymphomas. Nature 459, 712‐716, doi:10.1038/nature07969 (2009). 313 Calado, D. P. et al. Constitutive canonical NF‐kappaB activation cooperates with disruption of BLIMP1 in the pathogenesis of activated B cell‐like diffuse large cell lymphoma. Cancer cell 18, 580‐589, doi:10.1016/j.ccr.2010.11.024 (2010). 314 Bushell, K. R. et al. Genetic inactivation of TRAF3 in canine and human B‐cell lymphoma. 125, 999‐1005, doi:10.1182/blood‐2014‐10‐602714 (2015). 315 Zhang, B. et al. An Oncogenic Role for Alternative NF‐kappaB Signaling in DLBCL Revealed upon Deregulated BCL6 Expression. Cell reports 11, 715‐726, doi:10.1016/j.celrep.2015.03.059 (2015). 316 Fonseca, R. et al. International Myeloma Working Group molecular classification of multiple myeloma: spotlight review. Leukemia 23, 2210‐2221, doi:10.1038/leu.2009.174 (2009). 317 Bergsagel, P. L. et al. Cyclin D dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood 106, 296‐303, doi:10.1182/blood‐2005‐01‐0034 (2005). 318 Annunziata, C. M. et al. Frequent engagement of the classical and alternative NF‐kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer cell 12, 115‐130, doi:10.1016/j.ccr.2007.07.004 (2007).

200

319 Demchenko, Y. N. et al. Classical and/or alternative NF‐kappaB pathway activation in multiple myeloma. Blood 115, 3541‐3552, doi:10.1182/blood‐2009‐09‐243535 (2010). 320 Keats, J. J. et al. Promiscuous mutations activate the noncanonical NF‐kappaB pathway in multiple myeloma. Cancer cell 12, 131‐144, doi:10.1016/j.ccr.2007.07.003 (2007). 321 Migliazza, A. et al. Heterogeneous chromosomal aberrations generate 3' truncations of the NFKB2/lyt‐10 gene in lymphoid malignancies. Blood 84, 3850‐3860 (1994). 322 Fracchiolla, N. S. et al. Structural alterations of the NF‐kappa B transcription factor lyt‐10 in lymphoid malignancies. Oncogene 8, 2839‐2845 (1993). 323 Hideshima, T. et al. Biologic sequelae of I{kappa}B kinase (IKK) inhibition in multiple myeloma: therapeutic implications. Blood 113, 5228‐5236, doi:10.1182/blood‐2008‐06‐161505 (2009). 324 O'Mahony, A., Lin, X., Geleziunas, R. & Greene, W. C. Activation of the heterodimeric IkappaB kinase alpha (IKKalpha)‐IKKbeta complex is directional: IKKalpha regulates IKKbeta under both basal and stimulated conditions. Molecular and cellular biology 20, 1170‐1178 (2000). 325 Woronicz, J. D., Gao, X., Cao, Z., Rothe, M. & Goeddel, D. V. IkappaB kinase‐beta: NF‐kappaB activation and complex formation with IkappaB kinase‐alpha and NIK. Science (New York, N.Y.) 278, 866‐869 (1997). 326 Zarnegar, B., Yamazaki, S., He, J. Q. & Cheng, G. Control of canonical NF‐kappaB activation through the NIK‐IKK complex pathway. Proceedings of the National Academy of Sciences of the United States of America 105, 3503‐3508, doi:10.1073/pnas.0707959105 (2008). 327 Malinin, N. L., Boldin, M. P., Kovalenko, A. V. & Wallach, D. MAP3K‐related kinase involved in NF‐ kappaB induction by TNF, CD95 and IL‐1. Nature 385, 540‐544, doi:10.1038/385540a0 (1997). 328 Moreaux, J., Veyrune, J. L., De Vos, J. & Klein, B. APRIL is overexpressed in cancer: link with tumor progression. BMC cancer 9, 83, doi:10.1186/1471‐2407‐9‐83 (2009). 329 Tai, Y. T. et al. Role of B‐cell‐activating factor in adhesion and growth of human multiple myeloma cells in the bone marrow microenvironment. Cancer research 66, 6675‐6682, doi:10.1158/0008‐5472.can‐06‐0190 (2006). 330 Neri, P. et al. Neutralizing B‐cell activating factor antibody improves survival and inhibits osteoclastogenesis in a severe combined immunodeficient human multiple myeloma model. Clinical cancer research : an official journal of the American Association for Cancer Research 13, 5903‐5909, doi:10.1158/1078‐0432.ccr‐07‐0753 (2007). 331 Moreaux, J. et al. BAFF and APRIL protect myeloma cells from apoptosis induced by interleukin 6 deprivation and dexamethasone. Blood 103, 3148‐3157, doi:10.1182/blood‐2003‐06‐1984 (2004). 332 Moreaux, J. et al. The level of TACI gene expression in myeloma cells is associated with a signature of microenvironment dependence versus a plasmablastic signature. Blood 106, 1021‐ 1030, doi:10.1182/blood‐2004‐11‐4512 (2005). 333 Demchenko, Y. N. & Kuehl, W. M. A critical role for the NFkB pathway in multiple myeloma. Oncotarget 1, 59‐68 (2010). 334 Staudt, L. M. Oncogenic activation of NF‐kappaB. Cold Spring Harbor perspectives in biology 2, a000109, doi:10.1101/cshperspect.a000109 (2010). 335 Klein, U. & Heise, N. Unexpected functions of nuclear factor‐kappaB during germinal center B‐ cell development: implications for lymphomagenesis. Current opinion in hematology 22, 379‐ 387, doi:10.1097/moh.0000000000000160 (2015). 336 Courtois, G. & Gilmore, T. D. Mutations in the NF‐kappaB signaling pathway: implications for human disease. Oncogene 25, 6831‐6843, doi:10.1038/sj.onc.1209939 (2006). 337 Karin, M. Nuclear factor‐kappaB in cancer development and progression. Nature 441, 431‐436, doi:10.1038/nature04870 (2006).

201

338 Heise, N. et al. Germinal center B cell maintenance and differentiation are controlled by distinct NF‐kappaB transcription factor subunits. The Journal of experimental medicine 211, 2103‐2118, doi:10.1084/jem.20132613 (2014). 339 Pasparakis, M., Schmidt‐Supprian, M. & Rajewsky, K. IkappaB kinase signaling is essential for maintenance of mature B cells. The Journal of experimental medicine 196, 743‐752 (2002). 340 Li, Z. W., Omori, S. A., Labuda, T., Karin, M. & Rickert, R. C. IKK beta is required for peripheral B cell survival and proliferation. Journal of immunology (Baltimore, Md. : 1950) 170, 4630‐4637 (2003). 341 Ruland, J. et al. Bcl10 is a positive regulator of antigen receptor‐induced activation of NF‐kappaB and neural tube closure. Cell 104, 33‐42 (2001). 342 Xue, L. et al. Defective development and function of Bcl10‐deficient follicular, marginal zone and B1 B cells. Nature immunology 4, 857‐865, doi:10.1038/ni963 (2003). 343 Hara, H. et al. The MAGUK family protein CARD11 is essential for lymphocyte activation. Immunity 18, 763‐775 (2003). 344 Ruland, J., Duncan, G. S., Wakeham, A. & Mak, T. W. Differential requirement for Malt1 in T and B cell antigen receptor signaling. Immunity 19, 749‐758 (2003). 345 Ruefli‐Brasse, A. A., French, D. M. & Dixit, V. M. Regulation of NF‐kappaB‐dependent lymphocyte activation and development by paracaspase. Science (New York, N.Y.) 302, 1581‐1584, doi:10.1126/science.1090769 (2003). 346 Kurosaki, T. Genetic analysis of B cell antigen receptor signaling. Annual review of immunology 17, 555‐592, doi:10.1146/annurev.immunol.17.1.555 (1999). 347 Srinivasan, L. et al. PI3 kinase signals BCR‐dependent mature B cell survival. Cell 139, 573‐586, doi:10.1016/j.cell.2009.08.041 (2009). 348 Mackay, F. & Schneider, P. Cracking the BAFF code. Nature reviews. Immunology 9, 491‐502, doi:10.1038/nri2572 (2009). 349 Kaisho, T. et al. IkappaB kinase alpha is essential for mature B cell development and function. The Journal of experimental medicine 193, 417‐426 (2001). 350 Yamada, T. et al. Abnormal immune function of hemopoietic cells from alymphoplasia (aly) mice, a natural strain with mutant NF‐kappa B‐inducing kinase. Journal of immunology (Baltimore, Md. : 1950) 165, 804‐812 (2000). 351 Hatada, E. N. et al. NF‐kappa B1 p50 is required for BLyS attenuation of apoptosis but dispensable for processing of NF‐kappa B2 p100 to p52 in quiescent mature B cells. Journal of immunology (Baltimore, Md. : 1950) 171, 761‐768 (2003). 352 Enzler, T. et al. Alternative and classical NF‐kappa B signaling retain autoreactive B cells in the splenic marginal zone and result in lupus‐like disease. Immunity 25, 403‐415, doi:10.1016/j.immuni.2006.07.010 (2006). 353 Sasaki, Y. et al. Canonical NF‐kappaB activity, dispensable for B cell development, replaces BAFF‐ receptor signals and promotes B cell proliferation upon activation. Immunity 24, 729‐739, doi:10.1016/j.immuni.2006.04.005 (2006). 354 Gerondakis, S. et al. Unravelling the complexities of the NF‐kappaB signalling pathway using mouse knockout and transgenic models. Oncogene 25, 6781‐6799, doi:10.1038/sj.onc.1209944 (2006). 355 Gardam, S., Sierro, F., Basten, A., Mackay, F. & Brink, R. TRAF2 and TRAF3 signal adapters act cooperatively to control the maturation and survival signals delivered to B cells by the BAFF receptor. Immunity 28, 391‐401, doi:10.1016/j.immuni.2008.01.009 (2008). 356 Gardam, S. et al. Deletion of cIAP1 and cIAP2 in murine B lymphocytes constitutively activates cell survival pathways and inactivates the germinal center response. Blood 117, 4041‐4051, doi:10.1182/blood‐2010‐10‐312793 (2011).

202

357 Xie, P., Stunz, L. L., Larison, K. D., Yang, B. & Bishop, G. A. Tumor necrosis factor receptor‐ associated factor 3 is a critical regulator of B cell homeostasis in secondary lymphoid organs. Immunity 27, 253‐267, doi:10.1016/j.immuni.2007.07.012 (2007). 358 Mills, D. M., Bonizzi, G., Karin, M. & Rickert, R. C. Regulation of late B cell differentiation by intrinsic IKKalpha‐dependent signals. Proceedings of the National Academy of Sciences of the United States of America 104, 6359‐6364, doi:10.1073/pnas.0700296104 (2007). 359 Weih, F. et al. Multiorgan inflammation and hematopoietic abnormalities in mice with a targeted disruption of RelB, a member of the NF‐kappa B/Rel family. Cell 80, 331‐340 (1995). 360 Franzoso, G. et al. Mice deficient in nuclear factor (NF)‐kappa B/p52 present with defects in humoral responses, germinal center reactions, and splenic microarchitecture. The Journal of experimental medicine 187, 147‐159 (1998). 361 Caamano, J. H. et al. Nuclear factor (NF)‐kappa B2 (p100/p52) is required for normal splenic microarchitecture and B cell‐mediated immune responses. The Journal of experimental medicine 187, 185‐196 (1998). 362 Weih, D. S., Yilmaz, Z. B. & Weih, F. Essential role of RelB in germinal center and marginal zone formation and proper expression of homing chemokines. Journal of immunology (Baltimore, Md. : 1950) 167, 1909‐1919 (2001). 363 Rickert, R. C., Roes, J. & Rajewsky, K. B lymphocyte‐specific, Cre‐mediated mutagenesis in mice. Nucleic acids research 25, 1317‐1318 (1997). 364 Gorelik, L. et al. Cutting edge: BAFF regulates CD21/35 and CD23 expression independent of its B cell survival function. Journal of immunology (Baltimore, Md. : 1950) 172, 762‐766 (2004). 365 Hu, H. et al. Noncanonical NF‐kappaB regulates inducible costimulator (ICOS) ligand expression and T follicular helper cell development. Proceedings of the National Academy of Sciences of the United States of America 108, 12827‐12832, doi:10.1073/pnas.1105774108 (2011). 366 Almaden, J. V. et al. A pathway switch directs BAFF signaling to distinct NFkappaB transcription factors in maturing and proliferating B cells. Cell reports 9, 2098‐2111, doi:10.1016/j.celrep.2014.11.024 (2014). 367 Rickert, R. C. Regulation of B lymphocyte activation by complement C3 and the B cell coreceptor complex. Current opinion in immunology 17, 237‐243, doi:10.1016/j.coi.2005.03.001 (2005). 368 Sharpe, A. H. & Freeman, G. J. The B7‐CD28 superfamily. Nature reviews. Immunology 2, 116‐ 126, doi:10.1038/nri727 (2002). 369 Busch, R., Doebele, R. C., Patil, N. S., Pashine, A. & Mellins, E. D. Accessory molecules for MHC class II peptide loading. Current opinion in immunology 12, 99‐106 (2000). 370 Kuwano, Y. et al. CD83 influences cell‐surface MHC class II expression on B cells and other antigen‐presenting cells. International immunology 19, 977‐992, doi:10.1093/intimm/dxm067 (2007). 371 Franzoso, G. et al. Requirement for NF‐kappaB in osteoclast and B‐cell development. Genes & development 11, 3482‐3496 (1997). 372 Brightbill, H. D. et al. Conditional Deletion of NF‐kappaB‐Inducing Kinase (NIK) in Adult Mice Disrupts Mature B Cell Survival and Activation. Journal of immunology (Baltimore, Md. : 1950) 195, 953‐964, doi:10.4049/jimmunol.1401514 (2015). 373 Do, R. K. et al. Attenuation of apoptosis underlies B lymphocyte stimulator enhancement of humoral immune response. The Journal of experimental medicine 192, 953‐964 (2000). 374 Tardivel, A. et al. The anti‐apoptotic factor Bcl‐2 can functionally substitute for the B cell survival but not for the marginal zone B cell differentiation activity of BAFF. European journal of immunology 34, 509‐518, doi:10.1002/eji.200324692 (2004).

203

375 Rahman, Z. S. & Manser, T. B cells expressing Bcl‐2 and a signaling‐impaired BAFF‐specific receptor fail to mature and are deficient in the formation of lymphoid follicles and germinal centers. Journal of immunology (Baltimore, Md. : 1950) 173, 6179‐6188 (2004). 376 Viatour, P. et al. NF‐ kappa B2/p100 induces Bcl‐2 expression. Leukemia 17, 1349‐1356, doi:10.1038/sj.leu.2402982 (2003). 377 Zarnegar, B. et al. Unique CD40‐mediated biological program in B cell activation requires both type 1 and type 2 NF‐kappaB activation pathways. Proceedings of the National Academy of Sciences of the United States of America 101, 8108‐8113, doi:10.1073/pnas.0402629101 (2004). 378 Ley, K. The role of selectins in inflammation and disease. Trends in molecular medicine 9, 263‐ 268 (2003). 379 Guo, F., Weih, D., Meier, E. & Weih, F. Constitutive alternative NF‐kappaB signaling promotes marginal zone B‐cell development but disrupts the marginal sinus and induces HEV‐like structures in the spleen. Blood 110, 2381‐2389, doi:10.1182/blood‐2007‐02‐075143 (2007). 380 de Villartay, J. P., Fischer, A. & Durandy, A. The mechanisms of immune diversification and their disorders. Nature reviews. Immunology 3, 962‐972, doi:10.1038/nri1247 (2003). 381 Han, S. et al. Cellular interaction in germinal centers. Roles of CD40 ligand and B7‐2 in established germinal centers. Journal of immunology (Baltimore, Md. : 1950) 155, 556‐567 (1995). 382 Niu, H., Ye, B. H. & Dalla‐Favera, R. Antigen receptor signaling induces MAP kinase‐mediated phosphorylation and degradation of the BCL‐6 transcription factor. Genes & development 12, 1953‐1961 (1998). 383 Falini, B. et al. A monoclonal antibody (MUM1p) detects expression of the MUM1/IRF4 protein in a subset of germinal center B cells, plasma cells, and activated T cells. Blood 95, 2084‐2092 (2000). 384 Fellmann, C. et al. Functional identification of optimized RNAi triggers using a massively parallel sensor assay. Molecular cell 41, 733‐746, doi:10.1016/j.molcel.2011.02.008 (2011). 385 Dickins, R. A. et al. Probing tumor phenotypes using stable and regulated synthetic microRNA precursors. Nature genetics 37, 1289‐1295, doi:10.1038/ng1651 (2005). 386 Lee, H. H., Dadgostar, H., Cheng, Q., Shu, J. & Cheng, G. NF‐kappaB‐mediated up‐regulation of Bcl‐x and Bfl‐1/A1 is required for CD40 survival signaling in B lymphocytes. Proceedings of the National Academy of Sciences of the United States of America 96, 9136‐9141 (1999). 387 Barish, G. D. et al. Bcl‐6 and NF‐kappaB cistromes mediate opposing regulation of the innate immune response. Genes & development 24, 2760‐2765, doi:10.1101/gad.1998010 (2010). 388 Cormier, F. et al. Frequent engagement of RelB activation is critical for cell survival in multiple myeloma. PloS one 8, e59127, doi:10.1371/journal.pone.0059127 (2013). 389 Rahman, Z. S., Rao, S. P., Kalled, S. L. & Manser, T. Normal induction but attenuated progression of germinal center responses in BAFF and BAFF‐R signaling‐deficient mice. The Journal of experimental medicine 198, 1157‐1169, doi:10.1084/jem.20030495 (2003). 390 Vora, K. A. et al. Cutting edge: germinal centers formed in the absence of B cell‐activating factor belonging to the TNF family exhibit impaired maturation and function. Journal of immunology (Baltimore, Md. : 1950) 171, 547‐551 (2003). 391 Tarte, K., Zhan, F., De Vos, J., Klein, B. & Shaughnessy, J., Jr. Gene expression profiling of plasma cells and plasmablasts: toward a better understanding of the late stages of B‐cell differentiation. Blood 102, 592‐600, doi:10.1182/blood‐2002‐10‐3161 (2003). 392 Liou, H. C., Sha, W. C., Scott, M. L. & Baltimore, D. Sequential induction of NF‐kappa B/Rel family proteins during B‐cell terminal differentiation. Molecular and cellular biology 14, 5349‐5359 (1994).

204

393 Garber, M. et al. A high‐throughput chromatin immunoprecipitation approach reveals principles of dynamic gene regulation in mammals. Molecular cell 47, 810‐822, doi:10.1016/j.molcel.2012.07.030 (2012). 394 Heinz, S. et al. Effect of natural genetic variation on enhancer selection and function. Nature 503, 487‐492, doi:10.1038/nature12615 (2013). 395 Jin, F. et al. A high‐resolution map of the three‐dimensional chromatin interactome in human cells. Nature 503, 290‐294, doi:10.1038/nature12644 (2013). 396 Kasowski, M. et al. Variation in transcription factor binding among humans. Science (New York, N.Y.) 328, 232‐235, doi:10.1126/science.1183621 (2010). 397 Lim, C. A. et al. Genome‐wide mapping of RELA(p65) binding identifies E2F1 as a transcriptional activator recruited by NF‐kappaB upon TLR4 activation. Molecular cell 27, 622‐635, doi:10.1016/j.molcel.2007.06.038 (2007). 398 Martone, R. et al. Distribution of NF‐kappaB‐binding sites across human chromosome 22. Proceedings of the National Academy of Sciences of the United States of America 100, 12247‐ 12252, doi:10.1073/pnas.2135255100 (2003). 399 Schreiber, J. et al. Coordinated binding of NF‐kappaB family members in the response of human cells to lipopolysaccharide. Proceedings of the National Academy of Sciences of the United States of America 103, 5899‐5904, doi:10.1073/pnas.0510996103 (2006). 400 Zhao, B. et al. The NF‐kappaB genomic landscape in lymphoblastoid B cells. Cell reports 8, 1595‐ 1606, doi:10.1016/j.celrep.2014.07.037 (2014). 401 Smale, S. T. Hierarchies of NF‐kappaB target‐gene regulation. Nature immunology 12, 689‐694, doi:10.1038/ni.2070 (2011). 402 Casola, S. et al. Tracking germinal center B cells expressing germ‐line immunoglobulin gamma1 transcripts by conditional gene targeting. Proceedings of the National Academy of Sciences of the United States of America 103, 7396‐7401, doi:10.1073/pnas.0602353103 (2006). 403 Subramanian, A. et al. Gene set enrichment analysis: a knowledge‐based approach for interpreting genome‐wide expression profiles. Proceedings of the National Academy of Sciences of the United States of America 102, 15545‐15550, doi:10.1073/pnas.0506580102 (2005). 404 Basso, K. et al. BCL6 positively regulates AID and germinal center gene expression via repression of miR‐155. The Journal of experimental medicine 209, 2455‐2465, doi:10.1084/jem.20121387 (2012). 405 Silverstein, R. L. & Febbraio, M. CD36, a scavenger receptor involved in immunity, metabolism, angiogenesis, and behavior. Science signaling 2, re3, doi:10.1126/scisignal.272re3 (2009). 406 Li, J. et al. Novel NEMO/IkappaB kinase and NF‐kappa B target genes at the pre‐B to immature B cell transition. The Journal of biological chemistry 276, 18579‐18590, doi:10.1074/jbc.M100846200 (2001). 407 Won, W. J., Bachmann, M. F. & Kearney, J. F. CD36 is differentially expressed on B cell subsets during development and in responses to antigen. Journal of immunology (Baltimore, Md. : 1950) 180, 230‐237 (2008). 408 Corcoran, L., Vremec, D., Febbraio, M., Baldwin, T. & Handman, E. Differential regulation of CD36 expression in antigen‐presenting cells: Oct‐2 dependence in B lymphocytes but not dendritic cells or macrophages. International immunology 14, 1099‐1104 (2002). 409 Mullen, T. D., Hannun, Y. A. & Obeid, L. M. Ceramide synthases at the centre of sphingolipid metabolism and biology. The Biochemical journal 441, 789‐802, doi:10.1042/bj20111626 (2012). 410 Lehto, M. et al. The OSBP‐related protein family in humans. Journal of lipid research 42, 1203‐ 1213 (2001). 411 Gimeno, R. E. Fatty acid transport proteins. Current opinion in lipidology 18, 271‐276, doi:10.1097/MOL.0b013e3281338558 (2007).

205

412 Watkins, P. A. & Ellis, J. M. Peroxisomal acyl‐CoA synthetases. Biochimica et biophysica acta 1822, 1411‐1420, doi:10.1016/j.bbadis.2012.02.010 (2012). 413 Oakley, A. Glutathione transferases: a structural perspective. Drug metabolism reviews 43, 138‐ 151, doi:10.3109/03602532.2011.558093 (2011). 414 Ten Hagen, K. G., Fritz, T. A. & Tabak, L. A. All in the family: the UDP‐GalNAc:polypeptide N‐ acetylgalactosaminyltransferases. Glycobiology 13, 1r‐16r, doi:10.1093/glycob/cwg007 (2003). 415 Mohanan, S. et al. Potential role of peptidylarginine deiminase enzymes and protein citrullination in cancer pathogenesis. Biochemistry research international 2012, 895343, doi:10.1155/2012/895343 (2012). 416 Van Raemdonck, K., Van den Steen, P. E., Liekens, S., Van Damme, J. & Struyf, S. CXCR3 ligands in disease and therapy. Cytokine & growth factor reviews 26, 311‐327, doi:10.1016/j.cytogfr.2014.11.009 (2015). 417 Kelly‐Scumpia, K. M. et al. B cells enhance early innate immune responses during bacterial sepsis. The Journal of experimental medicine 208, 1673‐1682, doi:10.1084/jem.20101715 (2011). 418 Hoff, S. T. et al. Human B cells produce chemokine CXCL10 in the presence of Mycobacterium tuberculosis specific T cells. Tuberculosis (Edinburgh, Scotland) 95, 40‐47, doi:10.1016/j.tube.2014.10.005 (2015). 419 Moratz, C., Harrison, K. & Kehrl, J. H. Regulation of chemokine‐induced lymphocyte migration by RGS proteins. Methods in enzymology 389, 15‐32, doi:10.1016/s0076‐6879(04)89002‐5 (2004). 420 Shi, G. X., Harrison, K., Wilson, G. L., Moratz, C. & Kehrl, J. H. RGS13 regulates germinal center B lymphocytes responsiveness to CXC chemokine ligand (CXCL)12 and CXCL13. Journal of immunology (Baltimore, Md. : 1950) 169, 2507‐2515 (2002). 421 Hwang, I. Y., Hwang, K. S., Park, C., Harrison, K. A. & Kehrl, J. H. Rgs13 constrains early B cell responses and limits germinal center sizes. PloS one 8, e60139, doi:10.1371/journal.pone.0060139 (2013). 422 Han, J. I., Huang, N. N., Kim, D. U. & Kehrl, J. H. RGS1 and RGS13 mRNA silencing in a human B lymphoma line enhances responsiveness to chemoattractants and impairs desensitization. Journal of leukocyte biology 79, 1357‐1368, doi:10.1189/jlb.1105693 (2006). 423 Reif, K. & Cyster, J. G. RGS molecule expression in murine B lymphocytes and ability to down‐ regulate chemotaxis to lymphoid chemokines. Journal of immunology (Baltimore, Md. : 1950) 164, 4720‐4729 (2000). 424 Knodel, M., Kuss, A. W., Berberich, I. & Schimpl, A. Blimp‐1 over‐expression abrogates IL‐4‐ and CD40‐mediated suppression of terminal B cell differentiation but arrests isotype switching. European journal of immunology 31, 1972‐1980, doi:10.1002/1521‐ 4141(200107)31:7<1972::aid‐immu1972>3.0.co;2‐t (2001). 425 Shih, V. F., Tsui, R., Caldwell, A. & Hoffmann, A. A single NFkappaB system for both canonical and non‐canonical signaling. Cell research 21, 86‐102, doi:10.1038/cr.2010.161 (2011). 426 Wang, J. H. et al. Extension of the germinal center stage of B cell development promotes autoantibodies in BXD2 mice. Arthritis and rheumatism 65, 2703‐2712, doi:10.1002/art.38059 (2013). 427 Rossi, D. et al. Disruption of BIRC3 associates with fludarabine chemorefractoriness in TP53 wild‐ type chronic lymphocytic leukemia. Blood 119, 2854‐2862, doi:10.1182/blood‐2011‐12‐395673 (2012). 428 Rossi, D. et al. Alteration of BIRC3 and multiple other NF‐kappaB pathway genes in splenic marginal zone lymphoma. Blood 118, 4930‐4934, doi:10.1182/blood‐2011‐06‐359166 (2011).

206

429 Vincent, F. B., Morand, E. F., Schneider, P. & Mackay, F. The BAFF/APRIL system in SLE pathogenesis. Nature reviews. Rheumatology 10, 365‐373, doi:10.1038/nrrheum.2014.33 (2014). 430 Ng, L. G., Mackay, C. R. & Mackay, F. The BAFF/APRIL system: life beyond B lymphocytes. Molecular immunology 42, 763‐772, doi:10.1016/j.molimm.2004.06.041 (2005). 431 Navarra, S. V. et al. Efficacy and safety of belimumab in patients with active systemic lupus erythematosus: a randomised, placebo‐controlled, phase 3 trial. Lancet 377, 721‐731, doi:10.1016/s0140‐6736(10)61354‐2 (2011). 432 Furie, R. et al. A phase III, randomized, placebo‐controlled study of belimumab, a monoclonal antibody that inhibits B lymphocyte stimulator, in patients with systemic lupus erythematosus. Arthritis and rheumatism 63, 3918‐3930, doi:10.1002/art.30613 (2011). 433 He, B. et al. Lymphoma B cells evade apoptosis through the TNF family members BAFF/BLyS and APRIL. Journal of immunology (Baltimore, Md. : 1950) 172, 3268‐3279 (2004). 434 Kern, C. et al. Involvement of BAFF and APRIL in the resistance to apoptosis of B‐CLL through an autocrine pathway. Blood 103, 679‐688, doi:10.1182/blood‐2003‐02‐0540 (2004). 435 Phan, T. G. et al. High affinity germinal center B cells are actively selected into the plasma cell compartment. The Journal of experimental medicine 203, 2419‐2424, doi:10.1084/jem.20061254 (2006). 436 Smith, K. G., Light, A., Nossal, G. J. & Tarlinton, D. M. The extent of affinity maturation differs between the memory and antibody‐forming cell compartments in the primary immune response. The EMBO journal 16, 2996‐3006, doi:10.1093/emboj/16.11.2996 (1997). 437 Linterman, M. A. et al. IL‐21 acts directly on B cells to regulate Bcl‐6 expression and germinal center responses. The Journal of experimental medicine 207, 353‐363, doi:10.1084/jem.20091738 (2010). 438 Zotos, D. et al. IL‐21 regulates germinal center B cell differentiation and proliferation through a B cell‐intrinsic mechanism. The Journal of experimental medicine 207, 365‐378, doi:10.1084/jem.20091777 (2010). 439 Barnett, B. E. et al. Asymmetric B cell division in the germinal center reaction. Science (New York, N.Y.) 335, 342‐344, doi:10.1126/science.1213495 (2012). 440 Hatzi, K. et al. A hybrid mechanism of action for BCL6 in B cells defined by formation of functionally distinct complexes at enhancers and promoters. Cell reports 4, 578‐588, doi:10.1016/j.celrep.2013.06.016 (2013). 441 McConkey, D. J. & Zhu, K. Mechanisms of proteasome inhibitor action and resistance in cancer. Drug resistance updates : reviews and commentaries in antimicrobial and anticancer chemotherapy 11, 164‐179, doi:10.1016/j.drup.2008.08.002 (2008). 442 Mitsiades, N. et al. Molecular sequelae of proteasome inhibition in human multiple myeloma cells. Proceedings of the National Academy of Sciences of the United States of America 99, 14374‐14379, doi:10.1073/pnas.202445099 (2002). 443 Chng, W. J. et al. Molecular dissection of hyperdiploid multiple myeloma by gene expression profiling. Cancer research 67, 2982‐2989, doi:10.1158/0008‐5472.can‐06‐4046 (2007). 444 Romagnoli, M. et al. Canonical nuclear factor kappaB pathway inhibition blocks myeloma cell growth and induces apoptosis in strong synergy with TRAIL. Clinical cancer research : an official journal of the American Association for Cancer Research 13, 6010‐6018, doi:10.1158/1078‐ 0432.ccr‐07‐0140 (2007). 445 Rahal, R. et al. Pharmacological and genomic profiling identifies NF‐kappaB‐targeted treatment strategies for mantle cell lymphoma. Nature medicine 20, 87‐92, doi:10.1038/nm.3435 (2014).

207

446 Shono, Y. et al. A small‐molecule c‐Rel inhibitor reduces alloactivation of T cells without compromising antitumor activity. Cancer discovery 4, 578‐591, doi:10.1158/2159‐8290.cd‐13‐ 0585 (2014).

208

APPENDICES

209

Appendix A Differentially expressed sequence analysis (DESeq) of NF-B2-proficient vs. NF-B2-deficient BAFF-stimulated B cells

Reduced expression in nfkb2-deleted B cells padj<0.05 FOLD FUNCTIONAL CATEGORY GENE CHANGE padj value ANTI-APOTOTIC bcl2 1.325543529 0.02945239 B CELL MARKERS fcer2a (cd23) 1.325732056 0.017623548 CYTOKINE il16 1.174198435 0.015844786 il12a 1.364670688 0.025448272 CYTOSKELETON psd 2.710416969 3.31E-50 reep3 1.413310723 5.22E-08 ccdc99 1.639556254 0.000630591 dynlt3 1.28487809 0.000651276 tuba1b 1.310452963 0.013300371 tubb5 1.183371423 0.016273855 lamb3 1.457520181 0.006795493 DNA REPAIR bre 1.254578558 0.005088337 neil1 1.231318507 0.011716652 lig1 1.266981951 0.041050674 IMMUNE RESPONSE - INNATE tmem173 1.494749042 2.57E-15 irak3 1.331124105 0.000166854 ticam2 1.605212254 0.001376318 map3k8 1.324359846 0.001711831 gbp5 1.311710295 0.003818353 lrrc33 1.211129325 0.006204436 IMMUNE RESPONSE - T-dependent h2-dmb2 1.315573363 8.73E-07 h2-oa 1.249701592 0.003787994 icosl 1.279887924 5.89E-06 ctss 1.267684422 0.000133023 cd83 1.192126349 0.008264942 h2-aa 1.279103195 0.014128882 METABOLISM chst10 1.646260235 1.42E-20 b4galnt1 1.342557059 3.86E-08 pgap1 1.414539953 1.20E-06 ugcg 1.280810535 2.33E-05 prdx6 1.316376842 9.93E-05 faah 1.664599634 0.000220923 sdhaf1 1.332327141 0.001090296 naaa 1.35377873 0.001711831

210

kynu 1.285099975 0.003828254 alkbh6 1.257097276 0.013127776 hsd17b11 1.234976001 0.016890233 oat 1.219954321 0.018450351 sgms1 1.2008314 0.025188693 kmo 1.173450616 0.040117349 MISCELLANEOUS lpcat1 1.679481466 1.39E-18 syngr2 1.368374306 7.05E-10 gpr132 1.418653018 4.29E-09 p2ry10 1.416996618 1.04E-07 gpr174 1.351249124 1.05E-07 srgn 1.344730774 2.31E-07 stk17b 1.284992045 1.31E-06 tmem123 1.266805908 7.34E-06 prcp 1.322126349 7.34E-06 cnp 1.352586628 3.02E-05 ppp1r16b 1.28170883 0.000139812 panx1 1.369383341 0.000220731 prss12 1.6583446 0.000562756 smap2 1.222848398 0.000762389 filip1l 1.223693318 0.000850199 zfp36l1 1.217865092 0.000910526 fam69a 1.26830035 0.002361236 bmp2k 1.264286133 0.006693482 smagp 1.457531848 0.010647072 ttpal 1.191433902 0.012988025 fam111a 1.248348619 0.013757943 rdm1 1.304869816 0.018049232 sesn1 1.184896876 0.01922234 tgif1 1.182398998 0.025188693 plek 1.241242607 0.028269871 cytip 1.1701271 0.029403356 scn8a 2.147250624 0.030652721 csrp1 1.307827093 0.031092574 zdhhc18 1.167138677 0.031345903 tmem8 1.26459962 0.036128199 hmgn1 1.186759232 0.036523669 hpse 1.33345198 0.038481626 snx8 1.167525968 0.04194061 gtpbp1 1.184101515 0.047488424 ms4a4c 1.412947428 3.99E-10 rbpms 1.440213131 0.028769288 PROLIFERATION mcm5 1.374548993 6.03E-06

211

pola1 1.436176686 0.011209285 cables1 1.44033209 0.027286066 mcm2 1.239161005 0.03791968 RIBOSOME rps15a 1.371883866 8.30E-05 rps24 1.222974504 0.000685617 rpl32 1.284748379 0.000845661 rps15 1.272048419 0.001263385 rps3a 1.219342319 0.001961946 rps5 1.315793758 0.001987104 rpsa 1.27300757 0.002992526 rps17 1.327365118 0.003865064 rps11 1.198842614 0.003865064 rpl21 1.313838181 0.006942322 rps25 1.299329208 0.009376262 rps20 1.301480356 0.017351649 rps18 1.284337556 0.017623548 rplp1 1.25196849 0.019607973 rps4x 1.242016261 0.03328255 rpl22 1.27621509 0.040756683 RNA snrnp25 1.531122318 1.31E-11 dhx57 1.246873827 0.000629503 dhx40 1.204614275 0.018834808 znrd1 1.275235877 0.03328255 ddx25 1.687054793 0.041050674 SIGNALLING stat4 1.479017416 1.60E-14 tifa 1.405759008 8.02E-07 cyth1 1.27202071 2.77E-05 pde4b 1.268933849 4.17E-05 rapgef1 1.255319991 5.03E-05 grm6 1.626442754 5.90E-05 sfn 1.280482713 7.70E-05 rasgef1b 1.284530883 0.000419028 ppp3ca 1.21579356 0.006090489 mapk11 1.293431975 0.0064556 gng10 1.296843924 0.006871586 stat6 1.182310174 0.011645639 dennd4a 1.266048973 0.016890233 gpr18 1.269401893 0.019567736 mapk12 1.378862817 0.021083551 grap 1.272861725 0.025740469 gbp4 1.218677047 0.027353694 rras2 1.193438225 0.030652721 rasa3 1.169639635 0.035721026

212

syk 1.154232114 0.04580206 pik3cd 1.183538141 0.012281839 TRAFFICKING sell 1.471424122 1.49E-15 gpr183 (ebi2) 1.491272082 1.27E-12 cxcr5 1.185785738 0.009919532 ccr7 1.609781954 3.95E-10 TRANSCRIPTION pou2af1 (oca-b) 1.376137691 3.62E-11 ets2 1.517653974 6.08E-07 trim24 1.242155151 0.008863833 spic 1.64995705 0.017174723 ets1 1.163928278 0.022668769 TRANSPORT abcc1 1.361944309 6.92E-07 slc20a1 1.265175303 0.000132885 slc2a3 1.256328465 0.013127776 UNKNOWN rnf157 1.439456186 5.40E-08 fam169b 1.304064186 1.70E-06 1300014i06rik 1.434671289 8.30E-05 i730030j21rik 1.660975421 0.001727066 kbtbd11 1.273509517 0.002091503 arrdc3 1.247105275 0.002091503 zfp958 1.345445679 0.004537295 fam107b 1.193584764 0.004563295 6230427j02rik 1.795926088 0.005214902 pqlc1 1.274168472 0.012594903 ly6a 1.214844256 0.013580634 crisp3 1.385696438 0.016312525 zfp608 1.212734532 0.020015354 zmat2 1.200039092 0.032788432 mir682 1.316900195 0.037378097 - nfkb2 2.534237408 3.72E-88

Increased expression in nfkb2-deleted B cells padj<0.05 FOLD FUNCTIONAL CATEGORY GENE CHANGE padj value APOPTOSIS bmf 1.65972061 7.24E-18 bbc3 (puma) 1.602398362 0.002756872 B-CELL-specific fcrl5 (irta2) 1.55685032 7.61E-07 CYTOKINES il10 3.968080839 0.000229336 il12b 3.256747297 0.027089435 CYTOSKELETON myo1g 1.39419205 8.88E-10

213

ccdc88b 1.877179269 1.96E-09 vim 1.530263792 8.73E-07 ssh3 1.389174726 0.000716074 ppl 1.791479862 0.002556438 tubb2b 1.636812511 0.003624001 sspn 1.956492802 0.004133041 epb4.1l2 1.24688001 0.005611603 vasp 1.202044796 0.008882303 lrrc16a 1.289955402 0.010647072 col11a2 1.544285219 0.030448929 DNA REPAIR polm 1.446524167 0.004511346 INTRACELLULAR TRAFFICKING cplx2 2.310715937 1.63E-11 stxbp1 1.854395487 1.99E-09 ehd2 1.520200238 1.29E-05 itsn1 1.380361577 0.001792806 IMMUNE CELL-specific cd93 1.718718609 5.09E-08 rag1 2.611717576 0.009856109 cd44 1.304220533 0.000845661 IMMUNE RESPONSE - ANTIGEN PRES. h2-q4 1.18160423 0.027089435 lilrb3 1.266567534 0.031177751 h2-t24 1.551205223 0.000522479 lrmp 1.556443325 0.001917319 IMMUNE RESPONSE - INNATE tlr12 1.85666738 0.016883121 METABOLISM mgll 3.306638136 1.12E-06 scd2 1.337186265 3.14E-06 pld4 1.555203766 3.60E-06 pfkfb3 1.335451739 0.000522938 serinc5 1.53084655 0.000562756 gad1 1.379001693 0.000651276 adssl1 1.524260273 0.003865064 gns 1.185790081 0.015916005 ephx1 1.196044916 0.021921341 aldh3b1 1.408427183 0.022668769 man2a2 1.211480894 0.04837014 padi2 1.698509786 0.002091503 MISCELLANEOUS myof 1.97537815 1.12E-06 ski 1.342692476 4.51E-06 cacna1e 2.180152448 1.30E-05 cd36 1.614495279 3.52E-05 ceacam1 1.484156469 6.23E-05 p2rx7 2.522779678 0.000139812

214

psap 1.229355515 0.000280845 sema4b 1.276660723 0.00028676 chst3 1.236295424 0.000341418 akap12 1.628510684 0.000373941 mfhas1 1.551160076 0.0004705 ermap 2.108700363 0.000486286 slfn5 1.39007718 0.000716191 bhlhe41 2.211166293 0.000721515 sbk1 1.250090186 0.000796129 ddit4 1.548451661 0.000850199 anxa6 1.299845861 0.001601099 cacna1h 2.216175621 0.002091503 pitpnm2 1.328399418 0.00250882 chrnb1 1.482908083 0.003501811 bcl7a 1.219872147 0.008189727 plac8 1.201274585 0.008189727 capn2 1.341838252 0.008884271 ankrd13b 2.826201546 0.010647072 midn 1.22986858 0.012988025 gramd4 1.207224294 0.015141179 pea15a 1.311059957 0.015546891 leprotl1 1.317667921 0.017623548 ldlrap1 1.265427008 0.020044034 fam126a 1.258812468 0.020078026 tmc6 1.202302472 0.022160316 nrp2 2.966739515 0.025568096 ret 1.718560278 0.025821009 vwf 3.14323754 0.02843997 plau 4.506826964 0.029247347 mreg 1.336153784 0.031177751 txndc5 1.228409254 0.031370276 vangl2 1.813726173 0.031858297 ankrd13d 1.365886221 0.03414065 entpd1 1.297766431 0.03414065 mgst1 1.63195404 0.041148798 trio 1.370346122 0.041388395 cd300lf 2.021806619 0.047199353 sorcs2 1.943011395 0.047488424 gpr56 2.224602872 5.46E-05 ndrg1 2.078787078 8.90E-05 alpl 1.618002936 0.012281839

215

csda 1.251669977 0.012623083 PROLIFERATION cdkn1a 1.917659738 0.015844786 PROT DEG nedd4 1.779161016 1.36E-06 asb2 2.426063337 6.86E-06 znrf3 1.921626815 0.020078026 spsb1 2.342476744 0.040268915 PROT STAB hspa1a (hsp70-1) 5.695433618 0.003575741 hspa1b (hsp70- 1b) 5.653293421 0.003785318 hsph1 (hsp105) 1.394889811 0.006204436 RNA akap17b 1.657912082 0.003192827 zfp385a 1.696129503 0.046728948 SIGNALING csf2rb 1.685757509 0.007818925 hck 1.427269462 0.044854085 sh2d3c 1.20554822 0.049821404 wnt10a 2.270625494 5.98E-06 ptprs 1.419476773 2.02E-05 pde4a 1.54521352 8.30E-05 pde4c 3.584870496 0.00032063 prex1 1.259246403 0.000850199 rgs16 2.409556442 0.001130216 nfkbid (ikbdelta) 1.321156627 0.001442443 rasal1 1.357421117 0.001987104 prkcc 2.41388695 0.004511346 mcf2l 3.12026425 0.009778409 tbc1d9 2.015899948 0.010133594 notch1 1.452631168 0.011023733 camk2n1 1.70615537 0.013314237 sbf2 1.428543916 0.01631898 rasgrp1 1.186950546 0.023282283 card11 1.202173451 0.027089435 fgd2 1.415367791 0.03214068 rgs2 1.413159148 0.035292485 ptger4 1.31817624 0.035657449 arhgap22 3.17048857 0.047917149 gadd45b 1.274776673 0.013314237 pik3c2b 1.292350717 0.0363481 dusp10 1.466706856 0.037795143 inppl1 1.303723978 0.039449574 TRAFFICKING plxnd1 1.67076192 5.36E-09 s1pr3 2.007543312 0.010647072 ccr1 1.846591874 0.017231401

216

TRANSCRIPTION hivep3 2.401702315 1.74E-10 e2f2 1.517881193 4.29E-09 egr1 1.572770367 3.95E-06 cbx6 1.489462934 4.16E-06 myb 1.882719151 5.34E-06 hdac9 1.419631399 5.54E-05 pou2f2 (oct-2) 1.285403536 7.28E-05 zbtb32 1.70449426 0.000850199 tle3 1.447015761 0.006796405 klf10 1.457623253 0.013314237 id2 2.012132476 0.01631898 mafg 1.557442984 0.017555605 egr3 1.617609745 0.017623548 arid3a 1.272936652 0.020970289 tfeb 1.248220816 0.041050674 egr2 1.366652935 0.047277274 TRANSPORT atp2a3 1.30898498 6.08E-07 atp1b1 1.927424598 4.24E-05 slco4a1 1.83499065 0.001045242 slc12a4 1.477917824 0.001711831 abcb4 1.771772685 0.009726483 slc7a7 1.589285148 0.010647072 sfxn3 1.220610506 0.013830368 slc29a3 1.271817491 0.023865009 orai2 1.231216397 0.02619518 UNKNOWN fam167a 1.543222633 4.69E-08 e330020d12rik 1.611926172 4.19E-07 myadm 1.914808864 1.12E-06 rapgefl1 2.86823894 8.30E-05 ccdc123 1.377680746 0.000950196 heatr6 1.27480024 0.002091503 2700081o15rik 1.504156524 0.003865064 fam53b 1.281283446 0.004176585 ahdc1 1.31907436 0.008189727 emid1 1.676734373 0.009376262 4930506m07rik 2.981996994 0.013314237 lrrc32 2.803024746 0.01634561 tm6sf1 1.384744771 0.016890233 cobll1 1.281385277 0.016890233 zfp395 1.265067085 0.017623548 gm6307 2.902272473 0.023301272

217

zcchc24 1.332098962 0.040756683 2210403k04rik 1.592064594 0.041161547 efr3b 6.413563779 0.04136992 endod1 1.326848769 0.043648057

218

Appendix B Differentially expressed sequence analysis (DESeq) of RELB/NF-B2-proficient vs. RELB/NF-B2-deficient GC B cells at day 7 of the GC reaction

Reduced expression in relb/nfkb2-deleted GC B cells >2.5 fold, p<0.01 FOLD FUNCTIONAL CATEGORY TRANSCRIPT CHANGE p value CYTOSKELETON cep290 3.701368327 0.001250872 fmn1 6.77867816 0.002083243 nebl 3.408183756 0.002700872 pls3 3.639672747 0.009879553 cdc42ep4 9.822813676 0.0005552 IMMUNE RESPONSE oasl2 3.56672381 0.00213694 ifit1 3.538140389 0.009510845 ephb2 10.74595395 0.000774539 ighg2c 8.537256103 0.004010517 tlr3 3.442985706 0.001035165 cxcl10 4.015628084 0.00643676 rgs3 2.885035238 0.009433817 rgs13 3.184725593 1.44E-06 METABOLISM gstt1 8.83364123 1.77E-05 gstt3 3.001044657 0.000196394 padi4 3.956754531 0.009296115 wbscr17 20.48998638 0.00060861 chst1 6.041301989 0.004207427 mocos 5.674386583 0.005944022 as3mt 2.643569946 1.58E-06 osbpl3 2.551941535 0.003505027 cers6 2.567999041 0.001065751 slc27a4 2.652195026 0.004516612 cd36 3.252357531 3.74E-10 NON-PROTEIN CODING gm14963 27.12136716 0.000116433 gm12063 7.718110564 0.003664288 OTHER olfr1444 8.242565212 0.002878901 nrep 11.9868501 0.003761633 trp73 5.053304329 0.009116683 adam15 5.588725201 0.009544978 hps6 2.547747033 0.005642318 stx11 3.230499654 0.007398995 enpp4 8.546181084 7.78E-06

219

POSTTRANSLATIONAL neurl1a 19.24256824 2.73E-06 RNA gemin7 2.699870842 0.000506399 SIGNALLING rhobtb1 4.174352932 0.004952224 ptpn14 2.866098295 0.005009919 ptpn13 7.525755869 0.001412894 evc 4.966383366 0.000100413 wls 4.286310822 0.000270604 daam1 3.173809654 0.001281587 TRANSCRIPTION hoxa7 28.03543672 0.000109421 tox2 10.1999987 0.000144835 hmga1-rs1 7.134385975 0.005625708 zfp365 2.695290407 0.006027941 hic1 6.826539202 0.002277476 UNKNOWN gm10800 59.2378127 1.36E-08 gm21738 17.17454777 6.78E-07 gm10801 24.19518446 3.49E-05 a930033h14rik 13.3365832 0.000117682

gm10717 18.58433817 0.000145953 sh3bgrl 2.647161401 0.000293758 ccdc120 5.445959362 0.00172023 fam135a 6.437804933 0.005610792 gm9825 2.794255416 0.006924743 paqr7 6.794428153 0.007168583 4930506m07rik 3.870257186 0.007304508

- relb 3.872632342 0.001113217 - nfkb2 2.93907889 0.001172224

Increased expression in relb/nfkb2-deleted GC B cells >2.5 fold, p<0.01 FOLD FUNCTIONAL CATEGORY TRANSCRIPT CHANGE p value CYTOSKELETON dnm1 28.84816562 1.71E-06 tppp 6.031130636 0.000340554 col27a1 3.005588623 0.004863368 IMMUNE RESPONSE iigp1 8.369047288 0.009720592 METABOLISM slc15a2 5.230059989 6.09E-08 cp 4.031121943 0.002349171 ltf 15.26600973 0.000597746

220

mgll 2.645647937 0.001999474 NON-PROTEIN CODING gm26789 5.0514664 8.62E-16 gm26625 4.09937526 9.00E-10 gm26749 4.178514721 3.36E-07 gm11932 3.581939544 5.75E-07 gm17477 7.342866239 8.25E-07 gm4279 4.806875857 9.35E-07 gm16153 2.503956747 2.49E-05 gm11721 6.337946518 3.94E-05 gm26882 2.868542527 0.000121993 gm12543 3.211250557 0.000139233 gm26518 2.857975964 0.00020212 gm16316 3.850261574 0.000445266 gm12064 2.967931907 0.00046968 gm23346 3.2731059 0.000853068 dleu2_1 2.533891205 0.000893282 gm17388 4.649648388 0.001055935 gm13171 5.594264975 0.001063637 gm13543 4.955253692 0.001080368 gm15478 3.132276391 0.001116323 gm15897 4.591754099 0.001322378 gm17195 3.071474284 0.001583323 gm12734 3.19714221 0.001613551 gm11052 2.912597908 0.001772564 gm15824 4.182238288 0.001805796 rp23-93m1.3 2.766231152 0.002223009 a730011c13rik 2.560051398 0.002292352

gm15342 2.890674769 0.00242165 gm15334 4.256464348 0.002454759 gm26531 2.738309967 0.002463093 gm15765 2.643931266 0.002782114 0610039k10rik 11.74433708 0.002876952

gm15154 5.406874136 0.002980496 gm14052 3.02174149 0.0029832 gm20717 3.690336855 0.003063738 4933426d04rik 5.357351955 0.003378546

gm25820 6.338011442 0.003555737 gm20682 3.684991448 0.003602044 rp23-224i23.1 7.417608209 0.003806439

221

1700099i09rik 3.275967106 0.004075714 gm16305 3.307957241 0.004241203 gm11805 2.983369408 0.005019858 gm22304 4.347490289 0.006110111 gm25848 2.569071108 0.006148293 gm13510 3.983865544 0.006522137 rpl31-ps6 4.09498902 0.007240531 gm14335 2.63348781 0.007470624 4933426k07rik 2.524163198 0.007472611

gm13571 3.331657237 0.007629546 gm13570 3.54909018 0.007968129 gm14853 4.61051573 0.007981705 mir5120 2.856865965 0.008087406 gm26836 2.592935332 0.008418244 gm11655 4.266930256 0.008922059 4930565n06rik 2.591519262 0.009337

gm15367 6.818256329 0.009380237 OTHER angpt2 2.900169285 0.001541543 lrrc18 2.729916644 0.004545494 efcab12 5.841862587 0.008085481 h2afy3 3.412721398 0.002619242 snap25 72.08175121 1.51E-07 sec61a2 2.938930987 0.001541862 kdelr3 5.756452542 0.004405829 POSTTRANSLATIONAL (Ub) trim2 2.819847322 0.001544851 SIGNALLING camk2n1 8.271429708 0.003738905 pla2g4c 2.933583458 0.008957931 TRANSCRIPTION adnp 3.308903224 0.002100786 TRANSPORT scn8a 2.985559464 0.001391662 atp1a3 8.434327762 0.001444815 slc22a23 4.575176512 0.00433535 UNKNOWN 4933408b17rik 5.410317807 0.000242984

UNKNOWN gm6104 3.31093544 0.000559432 UNKNOWN 2610203c22rik 6.299653481 0.000815857

UNKNOWN xkr6 5.22397547 0.002744925 UNKNOWN fam196b 2.514140993 0.004613704 UNKNOWN gm17227 5.075807117 0.005449014 UNKNOWN gm9754 7.642938157 0.006785963

222