Doctoral thesis from the Department of Molecular Biosciences The Wenner-‐Gren Institute, Stockholm University Stockholm, Sweden
Activation, adhesion and motility of B lymphocytes in health and disease
Natalija Gerasimčik
Stockholm, 2013
Cover: Scanning electron microscopy of Cdc42 knockout B cells by Natalija Gerasimčik
© Natalija Gerasimčik, Stockholm 2013 ISBN 978-‐91-‐7447-‐704-‐7
Printed in Sweden by Universitetsservice AB, Stockholm 2013 Distributor: Stockholm University Library
“B cells means the Best” (PhD, Associate Professor Mikael Karlsson)
To my Parents
Summary
B cells can be activated by T cell-‐dependent stimuli, such as CD40 ligation and cytokines, which induce extensive proliferation, class switch recombination and somatic hypermutation.
Epstein-‐Barr virus (EBV) can also induce B cell activation by mimicking T cell help through its main oncoprotein, latent membrane protein 1 (LMP-‐1). It is regulated by another EBV-‐encoded protein, EBV nuclear antigen 2 (EBNA-‐2), which is absent in Hodgkin and Burkitt lymphomas. We have studied LMP-‐1 induction by cytokines in vitro and shown that LMP-‐1 is induced through the transcription factor signal transducer and activator of transcription (STAT6) and a newly defined high-‐affinity STAT6-‐binding site.
When IL-‐4 is added together with lipopolysaccharide (LPS) or α-‐CD40 to B cells, it induces homotypic round and tight aggregates in vitro, whereas LPS alone does not induce such morphological changes. I describe here attempts to identify the molecules that regulate these responses.
I have shown that the Rho GTPase Cdc42 controls the spreading of B cells, whereas two other molecules in the same family, Rac1 and Rac2, control homotypic adhesion. Further, I have shown by conditional deletion of Cdc42 in B cells that it is important in the humoral immune response.
Dock10 is a guanosine nucleotide exchange factor (GEF) for Cdc42. It is expressed through all differentiation stages of B cell development. However, targeted deletion of Dock10 in B cells does not result in an aberrant phenotype. Furthermore, by studying conditional knockout mice for Dock10, Cdc42, Rac1 and Rac2, I have elucidated the mechanism of cytoskeletal changes during B cell activation, leading to adhesion and motility.
My results may lead to a better understanding of normal B cell activation and of EBV infection, which is associated with many human tumours and may help to understand cancer development and progression in B cells.
List of Publications
This thesis is based on the results presented in the following paper and manuscripts:
I. Kis, L.L., Gerasimcik, N., Salamon, D., Persson, E.K., Nagy, N. Klein, G., Severinson, E. and Klein, E. The STAT6 signaling pathway activated by the cytokines IL-‐4 and IL-‐13 induces expression of the Epstein-‐Barr virus-‐encoded protein LMP-‐1 in absence of EBNA-‐2: implications for the type II EBV latent gene expression in Hodgkin lymphoma. Blood 2011; 117:165-‐174.
II. Gerasimcik, N., Dahlberg, C., Baptista, M., Westerberg, L. and Severinson, E. B cells devoid of the Rho GTPase Cdc42 coordinate the actin and microtubule cytoskeleton less effectively and form an extrafollicular antibody response.
Manuscript.
III. Gerasimcik, N., Baptista, M., Westerberg, L. and Severinson, E. The guanine nucleotide exchange factor Dock10: expression and function in B lymphocytes.
Manuscript.
IV. Gerasimcik, N. and Severinson, E. Investigation of the role of the small Rho GTPases Rac1 and Rac2 in B cell activation. Preliminary results.
Publication not included in this thesis:
Cernysiov V., Gerasimcik N., Mauricas M., Girkontaite I. Regulation of T-‐cell-‐independent and T-‐cell-‐dependent antibody production by circadian rhythm and melatonin.
International Immunology 2010; 22(1):25-‐34.
List of abbreviations
APRIL A proliferation-‐inducing ligand Arp2/3 Actin-‐related proteins 2 and 3 B Basic region
BAFF(R) B cell activating factor (receptor) Bcl-‐6 B cell lymphoma 6
BCR B cell receptor BL Burkitt lymphoma Btk Bruton tyrosine kinase
Cdc42 Cell division control protein 42 CD40L CD40 ligand
CDM Ced-‐5/Dock180/Myoblast city
CFSE 6-‐Carboxyfluorescein succinimidyl ester cHL Classical Hodgkin lymphoma
CD Cluster of differentiation CIP4 Cdc42-‐interacting protein 4 CLL Chronic lymphocytic leukaemia CSR Class switch recombination CZH CDM-‐zizimin homology Dbl Diffuse B cell lymphoma DC Dendritic cell
DH Dbl-‐homology
DHR Dock homology region Dock Dedicator of cytokinesis DZ Dark zone
EBER EBV-‐encoded RNA Ebi2 EBV-‐induced receptor 2 EBNA EBV-‐nuclear antigen EBV Epstein Barr virus
EMSA Electrophoretic mobility shift assay EDL1 EcoRI D leftward 1
F-‐BAR Fes/CIP4 homology-‐Bin/Amphyphysin/Rvsp FCH Fes/CIP4 homology
FDC Follicular dendritic cell FOB Follicular B cell
GAP GTPase activating protein GBD GTPase-‐binding domain GC Germinal centre
γc Common gamma chain
GDI Rho GTP dissociation inhibitor
GEF Guanosine nucleotide exchange factor GL Germline transcripts
GTPase Guanine triphosphatase HL Hodgkin lymphoma
HRS Hodgkin Reed-‐Sternberg cells swIg Isotype switched immunoglobulin
IL Interleukin
ICAM-‐1 Intercellular adhesion molecule-‐1
ITAM Immunoreceptor tyrosine-‐based activation motif iNKT Invariant natural killer cell
JAK Janus kinase
JNK c-‐Jun N-‐terminal kinase LCL Lymphoblastoid cell line
LFA-‐1 Leukocyte function-‐associated antigen-‐1 LMP-‐1 Latent membrane protein 1
LPS Lipopolysaccharide
LRS LMP-‐1 regulatory sequences LZ Light zone
MAPK Mitogen-‐activated protein kinase MZB Marginal zone B cell
NF-‐κB Nuclear factor kappa beta NK Natural killer cell
NPC Nasopharyngeal carcinoma PC Plasma cell
PD-‐1 Programmed death-‐1
PHA Phytohemagglutinin
PI3K Phosphatidylinositol 3-‐kinase PRD Proline-‐rich domain
Rac Ras-‐related C3 botulinum toxin substrate RICH Rho GAP-‐interacting with CIP4 homologues SEB Staphylococcus enterotoxin B
SHM Somatic hypermutation SH3 Src homology 3
STAT Signal transducer and activator of transcription Syk Spleen tyrosine kinase
S1P Sphingosine 1 phosphate T1, T2 Transitional 1 or 2 B cells MZP Marginal zone precursor B cells TFR Follicular T regulatory cell TFH Follicular T helper cell TLR Toll-‐like receptor TNF Tumour necrosis factor TR Terminal repeat
TRAF TNF-‐receptor associated factor
VCA Veprolin-‐, central-‐, acidic region domain VCAM-‐1 Vascular cell adhesion mediator-‐1 V(D)J Variable, Diversity, Joining
VLA-‐4 Very late antigen-‐4
WASP Wiskott-‐Aldrich syndrome protein WH1 WASP homology domain
WIP WH1-‐interacting protein
Contents
Summary...4
List of publications ...5
List of abbreviations ...6
Contents ...8
Introduction...9
Early B cell differentiation ... 9
Late B cell differentiation ...10
B1 and B2 cells and their localization ... 10
Transitional B cells ... 11
Chemokine guidance of follicular and marginal zone B cells... 12
Marginal zone B cells and their activation ... 13
Follicular B cells and their further differentiation... 14
Plasma cell decisions ... 15
Memory B cells ... 16
The germinal centre reaction...17
T cell help: CD40-CD40L interaction...21
When activation leads to disease: the Epstein-Barr virus...22
Signalling similarity between LMP-1 and CD40: Mimicking T cell help...25
Interleukin-4 and its signalling pathway ...26
Other interleukins...28
Regulation of B cell adhesion and motility ...29
Guanine nucleotide exchange factors... 31
Small Rho GTPases, their effectors and effects... 33
The F-‐BAR protein CIP4 ... 36
Regulation in vitro...36
Matherials and Methods ...39
Tamoxifen preparation and administration for Mb1-Cre-ERT2 induction ...39
Results and Discussions...40
Paper I ...40
Paper II...42
Paper III ...44
Part IV (preliminary results)...46
Conclusions and perspectives ...48
Acknowledgements...51
References...53
Introduction
Early B cell differentiation
B lymphocytes develop in the foetal liver during embryogenesis and in the bone marrow in adults (Hardy and Hayakawa, 2001; Mackay et al., 2010). During development in the bone marrow, murine B cells express the chemokine receptor CXCR4 in order to ensure that the cells are maintained, and thus attracted to reticular stromal cells that express the ligand CXCL12 (Allende et al., 2010; Mackay et al., 2010).
In early development, the murine common lymphoid progenitor (CLP) cells differentiate to pro-‐B cells (Mackay et al., 2010). Upon IL-‐7 stimulation, pro-‐B cells are induced to rearrange the immunoglobulin heavy (IgH) chain V(D)J gene segments (Herzog et al., 2009;
Mackay et al., 2010). The recombination-‐activating gene products 1, 2 (Rag1 and Rag2) generate double-‐stranded DNA breaks between recombinational signal sequences (RSS) that flank the V, D and J gene segments, and join cleaved ends by non-‐homologous end joining (NHEJ). After rearrangement is complete, cells become pre-‐B cells, expressing the Igμ heavy (H) chain. Igμ associated with surrogate light (L) chains (VpreB and λ5), Igα (CD79a or Mb-‐1) and Igβ subunits, form the pre-‐B cell receptor (pre-‐BCR) complex (Hardy and Hayakawa, 2001; Herzog et al., 2009; Kurosaki et al., 2010; Gonzalez et al., 2011). This is one of the checkpoints for B cells – only B cells with a properly functioning pre-‐BCR can mature further (Herzog et al., 2009). Later, when V(D)J recombination of another IgH allele is suppressed, B cells start to produce their surface receptors with a single specificity.
Several kinases become activated upon pre-‐BCR signalling – the Src-‐family protein kinase Lyn and the spleen tyrosine kinase (Syk), inducing phosphorylation of the immunoreceptor tyrosine-‐based activation motifs (ITAMs) on the cytoplasmic parts of Igα/Igβ. Syk activates phosphatidylinositol 3-‐kinase (PI3K), which regulates survival, proliferation and differentiation (Herzog et al., 2009). In addition, Bruton tyrosine kinase (Btk) activates the mitogen-‐activated protein kinase (MAPK) cascade, leading to activation of nuclear factor κB (NF-‐κB) (Pieper et al., 2013).
Signalling via pre-‐BCR leads to λ5 downregulation. Rag1 and Rag2 subsequently induce Ig light chain (L) gene rearrangement (Hardy and Hayakawa, 2001; Herzog et al., 2009). After successful light chain rearrangement, cells start to express a functional B cell receptor.
After all maturation steps, cells that are highly reactive to the self-‐antigens die by apoptosis. However, B cells that have low reactivity to self-‐antigens are allowed to leave the bone marrow and enter the periphery (Pieper et al., 2013).
Positively selected immature B cells downregulate expression of their receptor CXCR4, resulting in disruption of CXCR4-‐CXCL12 interaction, and are released from the bone marrow to the blood (Mackay et al., 2010). This process is also regulated by a family of G protein coupled receptors – the sphingosine 1 phosphate (S1P1) receptors (Allende et al., 2010). Later, these immature cells will differentiate into B2 conventional B cells, which are involved in the adaptive immune response.
Late B cell differentiation
B1 and B2 B cells and their localization
B1 B cells arise from progenitors that differ from those from which conventional B2 B cells arise, and migrate to the peritoneal and pleural cavities (Montecino-‐Rodriguez and Dorshkind, 2012). They can be divided into two subsets – B1a (CD5+) and B1b (CD5-‐). B1 cells are part of the innate immune system, able to recognize self-‐antigens and carbohydrates. They are responsible for the early reaction to an antigen with IgM responses (Hardy and Hayakawa, 2001; Montecino-‐Rodriguez and Dorshkind, 2012).
However, the B1b population can switch to IgA production and has a high rate of somatic hypermutations in its VH regions (Roy et al., 2009).
B2 B cells, in the form of immature transitional B cells that have left the bone marrow migrate to the secondary lymphoid organs such as the spleen and lymph nodes. The spleen has a specialized structure that promotes the appropriate immune responses against many different blood-‐borne antigens, and consists of two compartments: red and white pulp. In the red pulp destruction of erythrocytes takes place. The white pulp contains white blood cells: B and T lymphocytes, macrophages, dendritic cells, and other cells (Oracki et al., 2010).
Transitional B cells
Transitional B cells enter the red pulp of the spleen from the blood through the marginal sinus. They then enter the follicles, which are surrounded by the marginal zone. Here B cells acquire IgD expression and complete their maturation (Oracki et al., 2010). There are different ways to further sub-‐divide transitional B cells, into two or more distinct populations.
According to one classification, there are two types of transitional B cells, which are defined by their expressions of the complement receptor CD21 and of the low-‐affinity Fcε receptor (FcεRII or CD23). Early immature (T1) B cells are defined as IgMhighIgDnegCD23-‐
CD21low, whereas late immature (T2) B cells are defined as IgMhighIgDnegCD23+CD21high (Loder et al., 1999; Carsetti et al., 2004; Allman and Pillai, 2008). According to another classification there are three transitional stages of immature B cells: T0, T1 and T2. All three stages express CD93 and IgM, but differ in their expressions of IgD and CD23 (Henderson et al., 2010). T0 transitional B cells (IgD-‐CD23-‐) migrate from the bone marrow via the bloodstream to the red pulp of the spleen, but are unable to enter the white pulp before they mature into the T1 (IgD+CD23-‐) and T2 (IgD+CD23+) stages (Henderson et al., 2010). For transitional T1 and T2 B cells to be able to enter the white pulp of the spleen, the GTPases Rac1 and Rac2, as well the integrins leukocyte function-‐associated antigen-‐1 (LFA-‐1), very late antigen-‐4 (VLA-‐4) and chemokine receptors, are required (Henderson et al., 2010). In mice, whose B cells lack these Rac GTPases, the transitional B cells accumulate in the blood. In addition, in the absence of the tyrosine kinase, Syk, B cell development is arrested at the same stage as in the absence of Rac1 and Rac2 (Henderson et al., 2010), and this leads to the disappearance of the most mature cells (Schweighoffer et al., 2013). One more population of transitional B cells, T3 (CD23+IgMlo), has been suggested, but appears to be an anergic population in the spleen, and these cells do not mature further (Vossenkämper and Spenser, 2011). Eventually, some of the naïve transitional cells home to the marginal zone (MZ) and become marginal zone B cells (MZB), while the majority differentiate into follicular B cells (FOB) (Radbruch et al., 2006; Vossenkämper and Spenser, 2011).
All naïve B cells must encounter antigen and become activated via the BCR, which is a second checkpoint essential for their survival (Reth, 1994; Reth and Wienands, 1997).
Other signals are also necessary for cell survival. One such signal is the binding of B cell-‐
activating factor (BAFF) to its receptor (BAFF-‐R). Both receptors induce signalling through the transcription factor NF-‐κB, although the B cell receptor signalling leads to the classical pathway, while BAFF signalling leads to alternate pathways. In addition, PI3K, Btk and many other molecules play important roles in this signalling (Mackay et al., 2010; Pieper et al., 2013). The BCR signal-‐strength model, which describes the fates of follicular and marginal zone B cells, has been proposed: Intermediate BCR signalling to self-‐antigens, together with Btk signalling, results in T2 B cell development into follicular B cells. Weak signal via BCR and poor Btk signalling give rise to T2-‐marginal zone precursor (T2-‐MZP) cells, which, after Notch2 engagement by its ligand delta-‐like 1 (DLL1) on the epithelial cells, leads to differentiation into marginal zone B cells (Allman and Pillai, 2008; Pillai and Cariappa, 2009; Cerutti et al., 2013).
Chemokine guidance of follicular and marginal zone B cells
Follicular B cells. Mature B cells express the chemokine receptor CXCR5 and high levels of integrins LFA-‐1 and VLA-‐4. After B cells encounter an antigen, they become attracted to the follicles by their response to the chemokine CXCL13, which is secreted by follicular dendritic cells (FDCs) (Goodnow et al., 2010; Pereira et al., 2010; Cyster, 2010).
Intercellular adhesion molecule-‐1 (ICAM-‐1) and vascular cell adhesion mediator-‐1 (VCAM-‐
1), the ligands for LFA-‐1 and VLA-‐4, respectively, also play important roles in this interaction between B cells and FDCs, promoting their cell-‐to-‐cell contacts (Harwood and Batista, 2010). In addition to CXCL13, the cellular Epstein-‐Barr virus-‐induced receptor 2 (Ebi2, also known as GPR183) and its ligand 7α,25-‐dihydroxycholesterol (7α,25-‐OHC) guide naïve and activated B cells to the outer follicular niche(s) of secondary lymphoid organs (Pereira et al., 2009; Gatto et al., 2009; Gatto and Brink, 2013). However, germinal centres (GC) are still formed in their normal locations when Ebi2 is absent (Green et al., 2011). Moreover, S1P2 is important for inhibiting the response of GC B cells to chemoattractants, and helps to confine these cells to the middle of the follicle (Green et al., 2011). The chemokine receptor CCR7 is expressed by naïve T cells and (at a low level) in mature B cells, but its expression is greatly upregulated after encountering an antigen (Goodnow et al., 2010; Pereira et al., 2010). CCR7 helps B cells to move towards the T cell zone, where Ebi2 helps to distribute them in the border between the T cell area and the midline of the follicular B cell zone. Here, after B cell interaction with T cells through CD40
engagement, expression of Ebi2 is upregulated, whereas CCR7 is downregulated (Pereira et al., 2010; Kelly et al., 2011; Gatto and Brink, 2013). The shuttling of the follicular B cells between the interfollicular and outer follicular regions is very important for FOB proliferation and GC formation (Gatto and Brink, 2013).
Marginal zone B cells. The cannabinoid receptor 2 (CR2) guides and positions MZB cells to the marginal zone and prevents their elution to the blood (Basu et al., 2011; Muppidi et al., 2011). Sphingosine 1 phosphate (S1P), which comes to the marginal zone via the blood stream, binds to the sphingosine 1 phosphate receptor 1 (S1P1) or sphingosine 1 phosphate receptor 3 (S1P3) on marginal zone B cells. It interferes with signals from CXCL13, and retains MZB cells in the marginal zone (Cerutti et al., 2013). The adhesive interactions between MZ B cells and stroma cells also play an important role in the retention of the former cells. In this case, LFA-‐1 and VLA-‐4 expressed on B cells interact with ICAM-‐1 and VCAM-‐1 on stromal cells. Later, marginal zone macrophages retain MZB cells by the macrophage receptor with collagenous structure (MARCO) (Pillai and Cariappa, 2009; Cerutti et al., 2013). However, marginal zone B cells are highly motile and migrate constantly between the MZ and follicles (Cinamon et al., 2008; Arnon et al., 2012). The chemokine CXCR5 is required for migration into the follicle. To return to the marginal zone, MZB cells again use S1P1 and S1P3, which are responsible for the attraction and retention of these cells. Shuttling of marginal zone B cells between the MZ and follicles ensures that they can capture an antigen more efficiently, and deliver more of it to the FDCs (Cinamon et al., 2008). In addition, Ebi2 is essential for activated MZB cell movements into the extrafollicullar areas during the primary immune response (Gatto et al., 2009).
Marginal zone B cells and their activation
Marginal zone B lymphocytes (MZB) are a minor population of the conventional B cells localized in the outer zone of the splenic white pulp. These cells can be identified as IgMhiIgDloCD23-‐CD21hiCD1dhi and are the first to respond to blood-‐borne pathogens. High levels of CD1d in marginal zone B cells make possible antigen (lipid) presentation by these cells to invariant natural killer cells (iNKT) (Pillai and Cariappa, 2009; Cerutti et al., 2013).
MZ B cells possess polyreactive BCR and can therefore bind many microbial patterns.
These cells are a link between innate and adaptive immune systems. MZ B cells express high levels of Toll-‐like receptors (TLRs) in the same way as other types of cells, such as
dendritic cells or macrophages (Cerutti et al., 2013).
After antigen encounter, MZB cells migrate to the extra-‐follicular areas between the T cell zone and the red pulp of the spleen, rapidly proliferate and differentiate into plasmablasts with low-‐affinity antibodies (Oracki et al., 2010; Mackay et al., 2010; Vinuesa et al., 2010;
Cerutti et al., 2013). However, marginal zone B cells are very diverse and can also generate long-‐lived plasma cells with high-‐affinity antibodies, which can be achieved both along pathways that are T cell-‐dependent and along those that are T cell-‐independent. In addition, MZB cells can undergo class switch recombination, to produce IgG and some IgA (Chappell et al., 2012; Puga et al., 2012; Cerutti et al., 2013).
Follicular B cells and their further differentiation
Follicular B cells (FOB) are the major population of the B cell pool in the spleen, which home to the follicles. They can be identified as IgMloIgDhiCD23+CD21intCD1dlo (Cerutti et al., 2013). Upon binding an antigen that is presented on FDCs, follicular B cells migrate and localize in the interfollicular zone, at the boundary between B cell follicles and the T cell zone, where ligands of CCR7, CCL21 and CCL19, are expressed (Oracki et al., 2010;
Goodnow et al., 2010; Pereira et al., 2010; Cyster 2010; Kerfood et al., 2011). Here, B cells receive the necessary stimulation signals from T helper cells. Major histocompatibility complex (MHC) class II-‐antigen peptides on B cells interact with the T cell receptor (TCR) on T cells, while CD40 on B cells interacts with CD40L on T cells and provide additional stimulation from cytokines secreted by T cells. After they receive these signals, B cells can undergo one of two fates. Either they continue migration to the extra-‐follicular areas, where they differentiate into short-‐lived plasma cells located in the extra-‐follicular foci where they produce early IgM and IgG, or they re-‐enter the follicles, proliferate and form germinal centres. This occurs as early as Day 4 after infection, and is discussed below in more detail. In this case, LFA-‐1 and ICAM-‐1 interaction helps them to survive by preventing apoptosis (Allen et al., 2007; Vinuesa et al., 2010; Mackay et al., 2010; Oracki et al., 2010;
Kurosaki, 2010; Cyster, 2010; Gatto and Brink, 2010; Harwood and Batista, 2010; Kerfood et al., 2011; Chu and Berek, 2012). However, B cells cannot enter the germinal centre and differentiate, until Ebi2 expression has been downregulated by the transcriptional repressor B cell lymphoma 6 (Bcl-‐6) (Chan et al., 2010; Goodnow et al., 2010; Pereira et al., 2010; Victora and Nussenzweig, 2012). While GC B cells downregulate Ebi2, they maintain
expression of CXCR5, which keeps them in the follicle, where CXCL13 is expressed (Chan et al., 2010; Gatto and Brink, 2010).
The antibody-‐producing GC B cells have high specificity for an antigen and differentiate either into long-‐lived and non-‐dividing plasma cells or into memory B cells (Good-‐Jacobson and Shlomchik, 2010; Mackay et al., 2010; Vinuesa et al., 2010; Yoshida et al., 2010; Chu and Berek, 2012).
Plasma cell decisions
The location at which cells receive activation signals will later determine the ability of plasmablasts to migrate to specific locations (Radbruch et al., 2006; Mackay et al., 2010).
There are two ways for B cells to become plasma cells. After activation, B cells either differentiate into short-‐lived plasma cells, which are found in the extrafollicular areas of secondary lymphoid tissue, or go through the germinal centre reaction and become long-‐
lived plasma cells, which migrate to the bone marrow (Shapiro-‐Shelef and Calame, 2005;
Oracki et al., 2010; McHeyzer-‐Williams et al., 2012). It has been suggested that the affinity of the BCR for an antigen regulates the capacity of the B cells to present the antigen to follicular T helper cells (McHeyzer-‐Williams et al., 2012). With increased help from follicular T helper cells, the fate of B cells is directed towards the germinal centre reaction.
Moreover, the follicular T helper cells direct B cell commitment towards either non-‐GC or GC plasma cells and determine the class of antibody produced (Schwickert et al., 2011;
McHeyzer-‐Williams et al., 2012).
It has, however, been suggested that the plasma cell pool in the bone marrow contains not only long-‐lived cells, but also short-‐lived cells (Bortnick and Allman, 2013). In addition, recent observations suggest that B cells that have responded to a T cell-‐independent antigen, such as lipopolysaccharide (LPS), are also able to generate long-‐lived plasma cells, even though they are not able to maintain a germinal centre response (Bortnick and Allman, 2013).
Due to affinity maturation in the germinal centres, the long-‐lived plasma cells produce IgG antibodies with high affinity and with hypermutated variable regions (Radbruch et al., 2006; Chu and Berek, 2012). Plasma cells that are terminally differentiated, non-‐dividing
and are secreting antibodies can be identified by surface expression of CD138 (Syndecan-‐
1) (Smith et al., 1996).
For differentiation into long-‐lived plasma cells, the transcriptional repressor B lymphocyte-‐
induced maturation protein-‐1 (Blimp-‐1), which represses both Bcl-‐6 and Pax5, is essential (Shapiro-‐Shelef and Calame et al., 2005; Oracki et al., 2010; Chu and Berek, 2012). During differentiation into plasma cells, B cells downregulate CXCR5 and upregulate Ebi2 and CXCR4. This allows plasmablasts or plasma cell precursors to home to the bone marrow in response to CXCL12, which is produced by stromal cells (Chan et al., 2010; Gatto and Brink, 2010; Mackay et al., 2010; Pereira et al., 2010; Yoshida et al., 2010). Additionally, S1P1 receptors are also important for the egress of lymphocytes from the secondary lymphoid organs (Allende et al., 2010; Pereira et al., 2010; Cyster, 2010). S1P1 blockage causes plasmablast accumulation in the spleen, thereby inhibiting the migration of plasmablasts to the bone marrow (Yoshida et al., 2010).
Proper homing is critical for the survival of long-‐lived plasma cells, since failure to enter the bone marrow may compromise long-‐lived humoral immunity (Bortnick and Allman, 2013). In the bone marrow, the interaction of APRIL and/or BAFF, which are produced by stromal cells, with BAFF-‐R, which is present on plasma cells, is essential for plasma cells to be sustained as long-‐lived plasma cells and thus keep antibody titres high for a long time without the need of the memory B cell pool to be activated (Ahuja et al., 2008; Allman and Pillai, 2008).
Memory B cells
Memory B cells play an essential role in long-‐term immunity maintenance. They can be defined as antigen-‐primed cells that express high-‐affinity antibodies, which can quickly differentiate into plasma cells during antigen recall. Memory B cells may remain in a resting state long after stimulation, and do not need antigen or T cell help for survival (Klein and Dalla-‐Favera, 2008; Shlomchik and Weisel, 2012). Memory B cells do not secrete and can be generated via germinal centres either as IgM+ or as isotype-‐switched (swIg+) types. The latter make up more than 95% of the cells. They can also be generated by a GC-‐
independent pathway with non-‐mutated receptors (Pape et al., 2011; Taylor et al., 2012).
The B cell receptors on swIg+ memory B cells have higher affinity than that of the IgM+
memory B cells (Pape et al., 2011). Highly mutated GC-‐derived memory B cells with either IgM+ and swIg+ receptors express the surface receptor CD73.
Early memory B cells can be detected even before the formation of GCs, and they most probably come from the same precursors as the cells of the germinal centre, since they express the memory B cell markers CD38, Bcl-‐2 and CCR6, together with the GC markers GL7 and CD95 (FAS) (Taylor et al., 2012a; Taylor et al., 2012b).
A key player that determines whether the precursors will differentiate into memory cells or enter the germinal centre reaction is CD40. In mice treated with anti-‐CD40 antibodies, the germinal centre differentiation was completely blocked, while generation of GC-‐
independent memory B cells was not affected (Erickson et al., 2002; Taylor et al., 2012).
After antigen challenge, IgM+ memory B cells proliferate and differentiate via the GC reaction, but swIg+ can quickly generate large amounts of the plasma cells without entering GCs (Dogan et al., 2009; Pape et al., 2011). This rapid expansion of plasma cells requires help from T cells, most probably T follicular helper memory cells (Taylor et al., 2012). The function of IgM+ memory B cells has not been fully elucidated since they respond poorly.
They might be important for re-‐infection when the pathogen has mutated. In this case, memory B cells can enter germinal centres, mutate their receptors, and produce a high-‐
affinity response. This contrasts with swIg+ cells, which cannot re-‐enter the germinal centres (Taylor et al., 2012).
Comparing the long-‐lived plasma cell response with the memory B cell response, Purtha et al. (2011) have shown by studying virus infections in mice that the polyclonal pool of swIg memory B cells can recognize and neutralize mutated pathogens equally well as wild-‐type pathogens. Long-‐lived plasma cells, however, were specific only for the original pathogen.
The Germinal Centre reaction
The germinal centre (GC) (Fig. 1) is the structure within the follicle in which B cells rapidly proliferate in response to T cell-‐dependent antigen stimulation. Shortly after the germinal centre is formed, it starts to resolve into two functionally distinct compartments – the dark zone (DZ) and the light zone (LZ). In the latter, B cells undergo class switch recombination
and somatic hypermutation. Only B cells with high-‐affinity receptors for the antigen will be selected (in the LZ) and differentiate either to long-‐lived memory B cells or antibody-‐
producing plasma cells (MacLennan, 1994; Hauser et al., 2007; Schwickert et al., 2007;
Kurosaki, 2010; Vinuesa et al., 2010; Gatto and Brink, 2010; Gonzalez et al., 2011).
The germinal centre compartmentalisation into DZ and LZ is mediated by opposing gradients of CXCL12 and CXCL13 (Gatto and Brink, 2010; Victora and Nussenzweig, 2012).
The DZ is densely packed with proliferating B cells, whereas the LZ is populated with B cells, follicular T helper cells (TFH) and follicular dendritic cells (FDC). The DZ is located close to the T cell zone, while the LZ is located close to the marginal sinus, where antigens enter the tissue (Hauser et al, 2007; Allen et al., 2007; Gatto and Brink, 2010; Cyster, 2010;
Victora and Nussenzweig, 2012).
Figure 1. The dynamic Germinal Centre model (from Victora and Nussenzweig, 2012).
FDCs are stromal cells that accumulate an antigen and form a network in the LZ of the germinal centre. Ablation of FDCs from the germinal centre leads to its disappearance (Vinuesa et al., 2010; Wang et al., 2011). These cells express high levels of integrin ligands (VCAM-‐1 and ICAM-‐1), and they catch and retain antigens in the form of immune complexes through Fc and complement receptors (Allen et al., 2007; Allen and Cyster, 2008; Hauser et al., 2007; Gatto and Brink, 2010). FDCs are the source of the CXCL13
chemokine, the ligand for CXCR5 and which is expressed on T and B cells (Vinuesa et al., 2010). Immune complexes, presented on FDCs, are strong stimuli of B cells. B cells activation results in the expression of activation-‐induced cytidine deaminase (AID), and the induction of class switch recombination (CSR) and somatic hypermutation (SHM) (Allen and Cyster, 2008; Victora and Nussenzweig, 2012).
TFH cells are a minor population (5-‐20%) in the LZ of the GC, but are crucial for the induction of GC responses by providing survival signals to GC B cells (Gatto and Brink, 2010). They can be distinguished by their high expression of programmed death-‐1 (PD-‐1).
Also, TFH cells express CD40L and produce high amounts of IL-‐4 and IL-‐21, which are essential for GC B cell survival and differentiation. They are thus essential for the full development and maintenance of mature germinal centres (Goodnow et al., 2010; Vinuesa et al., 2010; Hauser et al., 2010). Additionally, follicular T regulatory (TFR) cells control the GC reaction and the humoral immune response, since the absence of these cells leads to a greater GC reaction, but with only few antigen-‐specific GC B cells (Chung et al., 2011;
Linterman et al., 2011; Wollenberg et al., 2011).
The dark and light zone compartments of the germinal centre have different gene expression patterns. DZ cells have upregulated expressions of genes involved in mitosis, whereas genes that control lymphocyte activation, cell surface receptors and regulators of apoptosis are elevated in LZ cells (Victora et al., 2010). B cells in both zones have similar DNA synthesis levels, but cell division occurs mainly in the DZ of the germinal centre (Victora et al., 2010). Several markers have been identified that can be used in flow cytometry to distinguish DZ and LZ B cells: CXCR4hiCD83loCD86lo (DZ cells) and CXCR4loCD83hiCD86hi (LZ cells) (Victora et al., 2010). In addition, GL7 and CD95 (FAS) can be used as markers of germinal centres (Taylor et al., 2012).
During the affinity maturation process, germinal centre B cells move continuously within the LZ zone, searching for FDCs that carry an antigen, or within the DZ during division and mutation. They also move between these compartments (Hauser et al., 2007; Allen et al., 2007). It has been shown that 50% of cells from the DZ can migrate to the LZ in 6 hours in vivo, whereas cells from the LZ are less motile: only 15% migrate into the DZ during the same period (Victora et al., 2010; Victora and Nussenzweig, 2012).
There are several hypotheses about how cells move within the GC. The dynamic germinal centre model (Fig. 1) has recently been proposed, in which B cells in the LZ capture an antigen and compete for T cell help. Those B cells that present peptide-‐MHC class II to TFH cells obtain an activation signal and migrate to the DZ, where they rapidly divide. On the other hand, LZ cells that fail to be selected undergo apoptosis. In addition, this selection process is highly synchronized with the cycling between the DZ and LZ (Victora et al., 2010;
Victora and Nussenzweig, 2012). This agrees with the finding that germinal centre B cells move preferentially from the DZ to the LZ (Beltman et al., 2011).
Meyer-‐Hermann et al. have presented another, mathematical model (called LEDA), in which they combine the previous theory with an experimental mathematical approach to integrate large amounts of data. This model predicts a cyclic re-‐entry path for the B cells that have been positively selected on FDCs and can successfully compete for help from TFH cells. The more peptide-‐MHC B cells express, the more they will divide, but fewer mutations will be introduced. Cells that do not compete will die by apoptosis. The model was confirmed, since B cells enter the S phase when still in the LZ, just after they are selected, and immediately move to the DZ, just before the G2/M cell cycle phases. The model predicts that intracellular antigen is distributed to the daughter cells unequally in the dark zone, which is compatible with the asymmetric cell division described by Thaunat et al. (2012). LEDA predicts that those daughter cells that retain the antigen will enter final differentiation. In addition, the model predicts that the plasmablasts will exit via the DZ, where they must first divide and then leave the GC towards the T cell zone. Daughter cells that did not have a sufficient amount of an intracellular antigen will return to the LZ for more cycles (Meyer-‐Hermann et al., 2012). The model predicts that this is the mechanism by which the long-‐lived plasma cells are generated, but not memory B cells. Generation of memory B cells remains to be investigated separately, because it has different dynamics. In summary, Meyer-‐Herman et al. (2012) have analysed previous results and present new details of the B cell selection process. They also propose new models for division and the pathway of GC exit.
In the majority of cases, GCs are formed in T cell-‐dependent B cell responses, but certain T cell-‐independent antigens, such as bacterial polysaccharides, can also induce GC formation (Sverremark and Fernandez, 1998; Good-‐Jacobson and Shlomchik, 2010; Oracki et al., 2010; Vinuesa et al., 2010). A T cell-‐independent GC response is very poor, the structures
are short-‐lived and somatic hypermutation of the IgV region cannot take place (Vinuesa et al., 2010; McHeyzer-‐Williams et al., 2012). Interestingly, T cells seem to be required, at least for the maintenance of the GCs, even when induced by a T cell-‐independent antigen (Sverremark and Fernandez, 1998; Vinuesa et al., 2000). There is growing evidence that cells of the innate immune system present antigens to B cells and induce a fast and highly diverse antibody response, also providing survival signals to plasma cells (Cerutti et al., 2012).
T cell help: CD40-CD40L interaction
The CD40-‐CD40L interaction plays a critical role in the development of humoral and cellular immune responses. It is involved in the activation and proliferation of B cells, the formation of germinal centres, antibody production, isotype switching, somatic hypermutation, and the generation of memory B cells and plasma cells (Elgueta et al., 2009;
Graham et al., 2010).
CD40L (also known as CD154) belongs to the TNF family, and is mainly expressed on activated T cells (Elgueta et al., 2009; Kurosaki et al., 2010; Vinuesa et al., 2010; Graham et al., 2010). It is expressed also on monocytes, macrophages, platelets, mast cells, basophils, eosinophils, epithelial cells and NK cells (Elgueta et al., 2009; Graham et al., 2010). CD40L is a very important co-‐stimulus for the initiation of the GC reaction, and plays an important role in T cell-‐independent GC formation. In the latter case, CD40L is expressed on non-‐T cells (Vinuesa et al., 2010; Cerutti et al., 2012).
CD40 belongs to the tumour necrosis factor receptor (TNF-‐R) family. It was first found as a cell surface antigen, restricted to human urinary bladder carcinomas and B cells (Paulie et al., 1985). CD40 is constitutively expressed on almost all B cells (not in plasma cells), DCs, platelets, monocytes and macrophages (Kurosaki et al., 2010; Vinuesa et al., 2010; Graham et al., 2010). It is expressed also on non-‐hematopoietic cells such as fibroblasts, epithelial and endothelial cells (Elgueta et al., 2009). CD40 is associated with TNF receptor-‐
associated factors (TRAFs), and it initiates NF-‐κB, c-‐Jun N-‐terminal kinase (JNK) and p38 signalling pathways (Kurosaki et al., 2010; Graham et al., 2010). B cell activation via BCR and CD40 leads to inhibition of the transcription factor Bcl-‐6, a critical regulator of GCs (Klein and Dalla-‐Favera, 2008; Kurosaki et al., 2010). In T helper cells, Bcl-‐6 expression