Fcγ Receptors in the Immune Response

Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1102

_____________________________ _____________________________

Fcγ Receptors in the Immune Response

BY

TERESITA DIAZ DE STÅHL

ACTA UNIVERSITATIS UPSALIENSIS

UPPSALA 2001

Dissertation for the Degree of Doctor of Philosophy (Faculty of Medicine) in Pathology presented at Uppsala University in 2001

A BSTRACT

Díaz de Ståhl, T. 2001. Fcγ Receptors in the Immune Response. Acta Universitatis Upsaliensis.

Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1102. 58 pp.

Uppsala. ISBN 91-554-5175-6

Circulating immune complexes play an important role in the modulation of antibody responses and in the pathogenesis of immune diseases. This thesis deals with the in vivo regulatory properties of antibodies and their specific Fc receptors.

The immunosuppressive function of IgG is used clinically, to prevent rhesus-negative women from becoming sensitized to rhesus-positive erythrocytes from the fetus. The mechanism behind this regulation is poorly understood but involvement of a receptor for IgG, FcγRII, has been suggested. It is shown in this thesis that IgG and also IgE induce immunosuppression against sheep erythrocytes to a similar extent both in mice lacking all the known Fc receptors as in wild-type animals. These findings imply that antibody-mediated suppression of humoral responses against particulate antigens is Fc-independent and that the major operating mechanism is masking of epitopes.

Immunization with soluble antigens in complex with specific IgG leads to an augmentation of antibody production. The cellular mechanism behind this control is examined here and it is found that the capture of IgG2a immune complexes by a bone marrow-derived cell expressing FcγRI (and FcγRIII) is essential. An analysis of the ability of IgG3 to mediate this regulation indicated that, in contrast, this subclass of IgG augments antibody responses independently of FcγRI (and FcγRIII). These findings suggest that distinct mechanisms mediate the enhancing effect of different subclasses of antibodies.

Finally, the contribution of FcγRIII was studied in the development of collagen- induced arthritis (CIA), an animal model for rheumatoid arthritis in humans. It was discovered that while DBA/1 wild-type control mice frequently developed severe CIA, with high incidence, FcγRIII-deficient mice were almost completely protected, indicating a crucial role for FcγRIII in CIA.

The results presented here help to understand how immune complexes regulate immune responses in vivo and show that Fc receptors for IgG, if involved, could be new targets for the treatment of immune complex-related disorders.

Key words: Receptors, Fc; Receptors, IgG; Receptors, IgE; Immune regulation; In vivo animal models; Mouse; Rh prophylaxis; Rheumatoid arthritis and Transgenic/knockout.

Teresita Díaz de Ståhl, Department of Genetics and Pathology, Rudbeck Laboratory, Uppsala University, SE-751 85, Uppsala, Sweden

© Teresita Díaz de Ståhl 2001

ISSN 0282-7476 ISBN 91-554-5175-6

Printed in Sweden by Lindbergs Grafiska HB, Uppsala 2001

To my family

This thesis is based on the following papers, which are referred to in the text by their Roman numerals:

I Karlsson, M.C.I., Wernersson, S., Díaz de Ståhl, T., Gustavsson, S., and Heyman, B. Efficient IgG-mediated suppression of primary antibody responses in Fcγ receptor-deficient mice.

Proc Natl Acad Sci U S A, 1999. 96(5): p. 2244-2249.

II Karlsson, M.C.I., Díaz de Ståhl, T., and Heyman, B.

IgE-mediated suppression of primary antibody responses in vivo.

Scand J Immunol 2001; 53: p. 381-385.

III Díaz de Ståhl, T. and Heyman, B.

IgG2a-mediated enhancement of antibody responses is dependent on FcRγ ⁺ bone marrow-derived cells.

Scand J Immunol 2001; 54: p. 495-500.

IV Díaz de Ståhl, T., Dahlström, J. and Heyman, B.

IgG3 augments Ab responses in FcRγ-deficient mice.

Manuscript.

V Díaz de Ståhl, T., Andrén, M., Martinsson, P., Verbeek, J.S., Kleinau, S.

FcγRIII-deficient mice are highly protected to collagen-induced arthritis.

Submitted.

v

TABLE OF CONTENTS

TABLE OF CONTENTS ... VI ABBREVIATIONS...VII

INTRODUCTION ...1

T HE I MMUNE S YSTEM ... 1

Antibodies ... 1

B cells and T cells... 2

Antigen presenting cells ... 3

Follicular dendritic cells and germinal centers ... 4

The complement system ... 6

F C R ECEPTORS ... 7

Fc Receptors for IgG ... 7

Fc Receptors for IgE ... 11

F EEDBACK R EGULATION BY A NTIBODIES ... 12

Mechanisms behind antibody-mediated suppression ... 12

Mechanisms behind antibody-mediated enhancement ... 13

A UTOIMMUNITY ... 16

Rheumatoid arthritis... 17

Collagen-induced arthritis ... 18

THE PRESENT INVESTIGATION...19

S PECIFIC A IMS ... 19

E XPERIMENTAL M ODELS ... 20

Mice ... 21

Adoptive transfer ... 21

R ESULTS AND D ISCUSSION ... 23

(I) Involvement of Fc γ Rs in IgG-mediated suppression ... 23

(II) Involvement of FcRs in IgE-mediated suppression... 25

(III) FcR γ ⁺ bone marrow-derived cells in IgG2a-mediated enhancement... 27

(IV) Fc γ Rs in IgG3-mediated enhancement ... 30

(V) The role of Fc γ RIII in collagen-induced arthritis ... 32

SUMMARY...34

GENERAL DISCUSSION...35

ACKNOWLEDGEMENTS ...38

REFERENCES ...40

ABBREVIATIONS

Ab(s) Antibody (Antibodies) Ag(s) Antigen (Antigens) APC Antigen presenting cells β2m β2-microglobulin BCR B cell receptor

BM Bone marrow

BSA Bovine serum albumin CFA Complete Freund’s adjuvant CIA Collagen-induced arthritis CII Collagen type II

CVF Cobra venom factor DC Dendritic cells

ELISA Enzyme-linked immunosorbent assay ELISPOT Enzyme-linked immunospot assay FcεR Fc receptors for IgE

FcγR Fc receptors for IgG

FcRγ The common gamma subunit of immunoglobulin FcRs FcRn Nenonatal FcR

FcRs Fc receptors

FDC Follicular dendritic cells HRBC Horse red blood cells

IFA Incomplete Freund’s adjuvant

Ig Immunoglobulin

ITAM Immunoreceptor tyrosine-based activation motif ITIM Immunoreceptor tyrosine-based inhibitory motif i.v. Intravenous

KLH Keyhole limpet hemocyanin mAb Monoclonal antibody

MHC Major histocompatibility complex

OVA Ovalbumin

PBS Phosphate buffered saline PFC Plaque forming cell assay RA Rheumatoid arthritis

Rh Rhesus

SRBC Sheep red blood cells

TCR T cell receptor

TNP 2,4,6-trinitrophenyl

INTRODUCTION

T HE I MMUNE S YSTEM

Immunity is the state of protection against infections. It comprises both innate and acquired immunity, terms which refer to the non-specific and specific immune response, respectively. The role of the immune system is to protect the organism against foreign pathogens, toxins and cancer. To achieve this goal, a perfect coordination between the cells and molecules of the immune system is needed. This is done by controlled and complicated cooperation between components of humoral and cellular immunity. The key molecular components of humoral immunity are the antibodies (Abs), but cytokines and complement also play an important role. The cellular components of the immune system are B and T cells, monocytes, macrophages, dendritic cells (DC), follicular dendritic cells (FDC), neutrophils, eosinophils, basophils and mast cells. This thesis deals with the regulatory properties of Abs.

Antibodies

Abs, also known as immunoglobulins (Igs), are glycoproteins present in B cell membranes and secreted by plasma cells. They bind to a foreign substance, the antigen (Ag) and neutralize it or promote its elimination. They consist of two pairs of identical heavy chains and light chains held together by disulfide bonds (Fig. 1). The amino-terminal ends

of the heavy and light chains form two analogous sites for recognition of the Ag. This portion of the molecule, known as the variable region, differs between different Ab molecules and confers antigenic specificity on the Ab. The remaining part of the heavy chain is called the constant region. There are five different types of heavy chain, defined by the constant regions, which determine the class (or isotype) of the Ab: IgM, IgD, IgG, IgE and IgA. Minor differences in the amino acid sequences of the heavy chain classify the IgG isotype in four subclasses having different biological activity: IgG1, IgG2, IgG3 and IgG4 in humans and IgG1, IgG2a, IgG2b and IgG3 in mice [21]. The

Light chain

Heavy chain

Ag binding

Fc

Figure 1. Antibody structure Variable

regions Constant

regions

carboxyl-terminal ends of both heavy chains together form the Fc region. There are many receptors, known as Fc receptors (FcRs), that bind to the Fc portion of Abs.

Digestion of IgG with papain results in two identical Fab fragments (Fragment antigen binding) and a third Fc fragment (Fragment crystallizable), originally observed to crystallize readily, which contains no Ag-binding activity. Alternatively, digestion with pepsin generates one composite unit of two Fab-like fragments, designated F(ab’) 2 and multiple digested Fc fragments. Fab and F(ab’) 2 fragments retain Ag-binding capacity but do not bind FcRs.

B cells and T cells

B cells are bone marrow- (BM) derived cells. They have a membrane bound Ig, associated with a heterodimer called Ig-α/Ig-β, forming the B cell receptor (BCR), which recognizes the Ag [17, 107, 166]. Both B cells and T cells recognize discrete sites on the Ag called antigenic determinants or epitopes. BCRs are activated by Ag- induced cross-linking. When this happens, specific B cells start to proliferate and produce soluble Abs. This process requires the help of activated T cells. B cells finally develop either into plasma cells producing large amounts of Abs or into memory B cells, which are long lived cells that can respond rapidly in subsequent encounters with the Ag [77].

T cells are also BM-derived cells but unlike B cells, which mature in the BM, T cells

mature in the thymus [204]. They express a membrane receptor called the T cell

receptor (TCR) [107], associated with the signal-transducing CD3 complex. Activation

of T cells is dependent on interaction with Ag presenting cells (APC). Unlike B cells,

which recognize Ag in its native form, T cells recognize the Ag after it has been

processed. The Ag is cleaved into small peptides which associate with specific protein

complexes, major histocompatibility complex (MHC) class I or class II molecules, on

the surface of APC [72]. Naïve (unprimed) T cells are activated when they encounter

and recognize Ag presented on MHC molecules in combination with appropriate co-

stimulatory signals delivered by the APC. When activated, T cells start to proliferate

and develop either into effector or long lived memory T cells. Effector T cells include

cytotoxic and helper T cells which are characterized by the presence of membrane-

associated glycoproteins, known as CD8 and CD4 respectively. CD8 ⁺ T cells identify

Ag presented on MHC class I molecules on target cells and CD4 ⁺ T cells recognize Ag

on MHC class II on APC [51, 180]. Activated T cytotoxic cells react by killing the

The Immune System 3

infected or altered cell, these T cells are therefore particularly important in destroying virus infected or tumor cells. Activated T helper cells instead, secrete small soluble proteins called cytokines with important regulatory functions in the progress of an immune response. T helper cells have been divided into two functional classes, according to the cytokines that they produce: Th1 secrete IL-2 and IFNγ, and are responsible for activating cell-mediated functions while Th2 cells secrete IL-4, IL-5, IL- 6 and IL-10, and are important in the activation of B cells.

Antigen presenting cells

Antigen presenting cells (APC) are specialized cells capable of processing the Ag and presenting it, in association with MHC class II molecules, to T helper cells. They have a number of co-stimulatory surface molecules that are important in T cell activation.

Professional APC include DC, macrophages and B cells.

Dendritic cells

DC are BM-derived leukocytes [182], which capture Ag in the tissues by phagocytosis or endocytosis and migrate to lymphoid organs where they present the Ag to T cells [11, 32]. DC are more potent APC than macrophages and B cells, both of which have to be activated before being able to prime T cells. The potency of DC is facilitated by the constitutive expression of MHC II molecules and by a high level of co-stimulatory (principally B7) and cell adhesion molecule expression [180, 207]. Further, a low level of sialic acid on the cell surface decreases repellent electrostatic forces and favors the interaction with T cells [179].

Once primed by DC, T cells play a role in further increasing the activating capacity of

DC. Ligation of CD40 (a membrane glycoprotein which regulates cell proliferation and

programmed cell death) on DC by CD40L (a ligand for CD40 which is expressed on T

helper cells upon activation) induces DC maturation. Activated T cells can then in turn

promote B cell activation, both by releasing T cell-derived cytokines and by direct

intercellular contact [34]. During maturation, DC undergo several changes. Immature

DC have a high level of endocytic activity, a high proliferation rate, low motility, an

organized cytoskeleton and MHC class II expression is largely restricted to cytoplasmic

vesicles. Co-stimulatory molecules as CD40 and B7, however, are expressed at low

levels. Lipopolysaccharides or cytokines such as TNF-α or IL-1β induce the maturation

of DC, decreasing the endocytic activity and the proliferative rate, depolarizing the

cytoskeleton, increasing motility, and augmenting the expression of MHC class II, CD40 and B7 on the cell surface. Immune complexes have also been reported to induce some of the modifications characteristic of DC maturation [164]. After functional maturation, DC die by apoptosis [220].

Macrophages

Macrophages are derived from hematopoietic stem cells in BM. Monocytes, their precursors, circulate in the blood for some hours before migrating to the tissues where they differentiate into macrophages. Macrophages are phagocytic cells and can also function as APC. They must, however, be activated before being able to activate T cells.

Phagocytosis of bacteria or bacterial products (as lipopolysaccharides) activates macrophages, up-regulating the expression of MHC class II and the B7 co-stimulatory molecules. Activated macrophages are common activators of memory or effector T cells. Macrophages have a number of FcRs and complement receptors that facilitate binding and enhance phagocytosis of Ag when it is coated with Abs or complement components. When murine macrophages were used as APC for soluble Ag mixed with specific IgG, the complexes were endocytosed via FcRs and T cells were activated using several hundred-fold lower doses of Ag than when Ag was used alone [127]. The role of macrophages in priming T cells is still uncertain because they usually do not express the co-stimulatory molecules necessary to activate naive T cells [83].

B cells

B cells can also present Ag on MHC class II molecules, which they constitutively express, and activate effector or memory T cells. The contribution of resting B cells to priming CD4 ⁺ T cells is controversial. Naive B cells probably do not express sufficient levels of co-stimulatory molecules and/or MHC class II/peptide complexes to activate resting T cells [143] and therefore must be activated first.

Follicular dendritic cells and germinal centers

The activation of B and T cells occurs in the lymphoid follicles of spleen, lymph nodes

and other peripheral lymphoid tissues. In the absence of activation lymphoid follicles

(called primary follicles), comprise a network of FDC and small resting B cells. Ag or

Ag/Ab complexes are transported into lymphoid follicles either alone or on Ag-

transporting cells. DC or macrophages present the Ag to T cells, activating them and

causing them to proliferate. Naive B cells that bind the Ag become activated and, with

The Immune System 5

help from appropriate T cells, also start to proliferate. These proliferating B cells together with FDC, macrophages and T cells develop into secondary follicles.

Secondary follicles consist of a mantle area (concentrically packed B cells) and a germinal center. The germinal center is a specialized microenvironment favorable for the interaction of B cells, T cells and FDC [115, 124] (Fig. 2). Three zones can be distinguished:

(i) the dark zone with activated and proliferating B cells called centroblasts, which are characterized by the absence of membrane Ig

(ii) the basal light zone with a lot of FDC and centrocytes that express membrane Ig (iii) the apical light zone where centrocytes differentiate either to memory B cells or

plasma cells.

FDC have long cytoplasmic extensions and retain immune complexes for long periods of time. They express both FcRs for IgG and IgE and complement receptors. Immune complexes on FDC efficiently cross-link BCR and stimulate B cell activation. In addition, FDC provide nonspecific co-stimulatory signals that augment B cell proliferation and Ab production and seem to have a role in rescuing B cells from apoptosis [197]. During class switch and affinity maturation of Abs in the basal light zone, many centrocytes are generated. Those that cannot bind to Ag on FDC, or bind with low affinity, die by apoptosis and are phagocytosed by the specialized tingible- body macrophage. In the apical light zone, centrocytes that bind Ag with high affinity

FDC

Centrocyte Th

Macrophage Apoptosis Affinity Maturation IC

Centroblast (Activated B cell) Memory B cell

Plasma Cell APICAL LIGHT ZONE

DARK ZONE BASAL LIGHT ZONE

Figure 2. Structure of the germinal center

further differentiate either into small memory B cells or plasma cells secreting Abs. The plasma cells migrate to the medulla of the node or to the BM where they produce Abs [57]. Some memory B cells stay in the follicular mantle and others recirculate in the body.

The complement system

Among the important functions of IgM and IgG immune complexes are the activation of the complement system. The complement system consists of a series of serum and membrane proteins that interact in an enzymatic cascade following activation, to induce the clearance of pathogenic Ags and the generation of inflammatory products. The complement system is a major mediator of innate immunity and links the innate with the adaptive immune response [28]. Three different pathways known as the classical (induced by IgM and IgG immune complexes), lectin and alternative pathways can give rise to this activation. The activation of the complement system culminates in the formation of the membrane attack complex (MAC) that produces a lytic pore on the target cell surface. The activation of the intermediate components C4, C3 and C5 generates both soluble products (C4a, C3a and C5a), with potent inflammatory properties and membrane bound products (C3b and its degradation components, iC3b and C3dg), that subsequently bind complement receptors (designated CR1-CR4) and contribute in the localization or clearance of the Ag [29].

Complement receptor type 1 (CR1, also known as CD35) and complement receptor type 2 (CR2 also known as CD21) have important roles in the initiation and regulation of the immune response through Ag trapping, B cell activation and Ig-class switching [27, 89].

They are predominantly expressed on B cells, FDC and activated granulocytes [27]. The

receptor mediating the effects of complement on Ab responses in vivo is most likely

CR2 expressed on B cells [45], where it is found associated in the B cell co-receptor

(CR2/CD19/TAPA-1) [64]. Co-ligation of the BCR and the CR2/CD19/TAPA-1

complex lowers the threshold for B cell signaling [30]. In mice, CR1 and CR2 are

alternatively spliced products derived from the same gene (Cr2) [118]. CR3 (CD18/11b)

and CR4 (CD18/11c) are leukocyte integrins that mediate adhesion but also

phagocytosis of opsonized particles.

Fc Receptors 7

Mice deficient in complement components C3 and C4 or in Cr2 (lacking both CR1 and CR2) have been generated; they have a profound defect in their Ab response to T- dependent Ags [3, 66, 133].

Despite its activating functions, C1q, C1r, C1s, C2 or C4 deficiency is the major predisposing factor for lupus [50, 135, 170]. Complement therefore appears to have a role in the induction of B cell tolerance and in promoting the clearance of apoptotic cells, reducing the chance that autoreactive B cells become activated [26, 28].

F C R ECEPTORS

Fc receptors (FcRs) comprise a group of transmembrane and soluble glycoproteins that bind the Fc portion of Abs. They can consist of a single chain (the alpha subunit), which determines affinity and isotype specificity, or be associated with molecules that regulate ligand specificity, internalization and signaling [157]. The affinity depends on the type of FcR involved, the Ab isotype and if they bind to Ab alone or in complex with Ag.

The cross-linking of FcRs on immunocompetent cells can trigger allergic hypersensitivity reactions, release of inflammatory mediators, Ab-dependent cellular toxicity, phagocytosis, endocytosis, and immune complex clearance. FcRs also provide a link between the cellular and humoral immune response and they are involved in allergy, autoimmunity and inflammation. Most hematopoietic cells express FcRs and the nature of responses depends primarily on the cell type involved. There are FcRs for all classes of Igs, FcγR for IgG, FcεR for IgE, FcαR for IgA and possibly Fc∂R for IgD and FcµR for IgM [101, 160, 202]. A receptor able to bind IgM and IgA, Fcα/µR, has recently been identified [175].

Fc Receptors for IgG

Fc receptors for IgG (FcγRs) are a family of receptors with activating or inhibitory functions [157-161]. They are FcγRI (CD64), FcγRII (CD32) and FcγRIII (CD16) (Fig.

3) [101], where the Roman number is based on historical precedent. FcγRs are members

of the Ig gene superfamily, with similar Ig-like extracellular regions but differing in

their cytoplasmic domain and membrane associated molecules [160, 163]. FcγRI

consists of three Ig-like extracellular domains and FcγRII and FcγRIII have two [4, 46,

101, 172]. While the first two domains are homologous in all three FcγRs, the third

domain (closest to the cell surface) is different and confers FcγRI, its high affinity characteristic for monomeric IgG.

The various FcγRs bind the distinct IgG subclasses with different affinity. In mice, FcγRI is the high affinity receptor for monomeric IgG2a and it also binds complexes of IgG2a, IgG2b, IgG1 and IgG3 [46, 71, 73, 101, 157]. Murine FcγRII and FcγRIII are the low affinity receptors and bind complexes of IgG2a, IgG1 and IgG2b [46, 73, 101, 157].

The affinity of four IgG isotype-switch variants (derived from 4C8 IgM anti-erythrocyte auto-Ab) has been recently defined as follows: IgG2a > IgG2b > IgG3/IgG1 for murine FcγRI, and IgG2a > IgG1 > IgG2b > IgG3 for FcγRIII [67]. Human FcγRI binds human IgG1≥IgG3>IgG4>>>IgG2 [7]. Human FcγRII and FcγRIII bind complexes of IgG1 and IgG3 preferentially over IgG2 and IgG4 [203, 211].

Figure 3. Murine Fcγ receptors for IgG

α α γ γ β α

Fc γRII B1, B1' B2

γ γ

Fc γRI Fc γRIII

ITAM Immunoreceptor Tyrosine-based Activation Motif ITIM Immunoreceptor Tyrosine-based Inhibitory Motif

α

Inhibitory sequence for phagocytosis

s s s s s s

s s s

s s s s s

s s

s s s s

s

s

Fc Receptors 9

FcγRI and FcγRIII are activating receptors which elicit cell activation, calcium mobilization, endocytosis, phagocytosis and Ag presentation [46] when aggregated at the cell surface [46]. They share some similarities with other activating receptors such as BCR and TCR:

• in being composed of several chains [46, 165]

• in having a ligand binding subunit, the α chain [46, 165]

• in containing a conserved motif, the immunoreceptor tyrosine-based activation motif (ITAM), in the cytoplasmic tail of the transducing molecule, the γ chain (FcRγ) [24, 165, 214]

FcRγ is an essential component required for both receptor assembly and signal transduction of murine FcγRI, FcγRIII and FcεRI and can also be found associated with the TCR complex [189]. Mice deficient in FcRγ (FcRγ ^-/- ) created by targeted disruption of the FcRγ subunit have been generated [189]. They lack FcγRI, FcγRIII and FcεRI which results in immunocompromised mice, with macrophages lacking the ability to phagocytose Ab-coated particles, defects in natural killer cell-mediated Ab-dependent cytotoxicity and in mast cell-mediated allergic responses [189]. The β subunit found associated with FcγRIII in mast cells [116] and with FcεRI also contains ITAM but cannot signal by itself; it amplifies the signaling through the γ chain [58].

The relative contribution of FcγRI vs. FcγRIII in immune reactions in vivo is difficult to discriminate. A prominent role for FcγIII in inflammatory and anaphylactic responses was though indicated when mice deficient in FcγRIII showed impaired IgG-dependent anaphylaxis and Arthus reaction [85].

In mice there is only one FcγRI whereas in humans three isoforms, FcγRIA, FcγRIB and FcγRIC, are found [63], where A, B and C designate structurally related receptors derived from different genes within the same group. Human FcγRIA (the FcγRI detected on cell membranes) is the homologous to murine FcγRI. Human FcγRIII is present in two isoforms, FcγRIIIA and FcγRIIIB [149, 163, 169, 177]. FcγRIIIA is a transmembrane receptor, while FcγRIIIB is linked to the membrane through a glycophosphatidylinositol linkage [117, 174, 177]. FcγRIIIB lacks a cytoplasmic tail and is not associated with the FcRγ subunit. Mouse FcγRIII is translated from a single gene and its membrane and intracellular domains are homologous to human FcγRIIIA.

Murine FcγRI is expressed in monocytes, macrophages and there are some indications

that is also expressed on DC. Human FcγRI is also found in neutrophils, eosinophils and

DC. FcγRIII is expressed on macrophages, mast cells, DC, neutrophils and natural killer cells [46, 101, 164].

In humans, FcγRII is found on the majority of leukocytes [101] and several isoforms have been identified, FcγRIIA, FcγRIIB and FcγRIIC. FcγRIIA and FcγRIIC are single- chain, low affinity, activating receptors that contain an extracellular ligand-binding domain and ITAMs in the cytoplasmic domain. In mice, the only FcγRII that has been identified is the inhibitory FcγRIIB (Fig. 3). Activating FcγRII is not found in mice.

FcγRIIB in both mice and humans is a single chain inhibitory receptor, containing an immunoreceptor tyrosine-based inhibitory motif (ITIM) in the cytoplasmic tail [16].

FcγRIIB negatively regulates cell activation when clustered with activating counterparts on the cell surface by the recruitment of 5’-inositol-phosphatases, and the inhibition of ITAM-triggered calcium mobilization and cellular proliferation [5, 46, 126]. FcγRIIB is the only FcR that expresses an ITIM motif. Murine FcγRIIB exists in four isoforms generated by (cell type-specific) alternative splicing of the transmembrane and cytoplasmic exons. FcγRIIB1 (B1 and B1’) are expressed in B cells and FcγRIIB2 and B3 in macrophages [162]. FcγRIIB1 and B1’ differ from FcγRIIB2 by the insertion in their cytoplasmic tail of 47 and 19 amino acids, respectively, that inhibit endocytosis and Ag presentation [6]. The absence of endocytic FcγRIIB2 in B cells restricts Ag presentation to molecules recognized by BCR. FcγRIIB3 is a soluble receptor released by macrophages. It lacks transmembrane and cytoplasmic region [193].

Mice deficient in FcγRIIB (FcγRII ^-/- ) have been generated; they display augmented humoral and anaphylactic responses demonstrating that FcγRII acts as a general negative regulator of immune complex-triggered activation in vivo [190].

The MHC class I-related neonatal FcR, FcRn (Fig. 4), is the only FcR that binds all IgG subclasses. It is expressed on vascular endothelial cells in both mice and humans [18], on intestinal epithelial cells in mice [176] and in the placenta

in humans [75]. FcRn plays a role in Figure 4. The neonatal Fc receptor

β2m ^{s s}

s s

s s

FcRn

Fc Receptors 11

IgG homeostasis [74], regulating its level in serum [106] and its transport from mother to fetus across the placenta and across the intestinal epithelium of suckling mice [75, 104]. FcRn consists of an α chain associated with β2-microglobulin (β2m). As β2m is a constitutive component of FcRn, mice deficient in β2m lack the receptor. The absence of FcRn leads to a 10-fold faster IgG-catabolism and to correspondingly lower IgG levels [106] in the mutant mice.

Fc Receptors for IgE

FcεRI and FcεRII are the high and respective low affinity receptors for IgE (Fig. 5).

FcεRI, a member of the Ig gene superfamily is a multimeric receptor found on the cell surface. In mice it is present on mast cells and basophils whereas in humans FcεRI is also found on monocytes, macrophages, eosinophils, platelets and DC [109].

FcεRI contains four polypeptide chains:

one α, one β and two disulfide-linked FcRγ chains. The cross-linking of FcεRI (by binding of the Ag to the FcεRI- bound IgE), transduces an activating signal to the cell through the ITAMs contained in FcRγ chains, triggering mast cell and basophil degranulation, with release of inflammatory mediators.

FcεRI is responsible for linking pathogen- or allergen-specific IgE with cellular immunologic effector functions.

FcγRIIB and FcγRIII on mast cells are potent regulators of IgE-mediated responses [47, 201].

FcεRII, known as CD23, is the low affinity receptor for IgE and is the only

FcR that is not a member of the Ig gene

superfamily. It belongs to the C-type animal lectin family [12, 40]. It has a short N- terminal cytoplasmic tail and a C-terminal extracellular domain displaying a coiled structure that favors aggregation in trimmers [56] (Fig. 5). Murine CD23 is

Figure 5. Murine Fcε receptors for IgE

α β

ITAM Immunoreceptor Tyrosine- based Activation Motif

s s

γ

Fc ε ^{RI Fc} ε ^RII

s s s

s

constitutively expressed on B cells [156] and FDC [125]. Human CD23 is found in two isoforms derived from the same gene. CD23a expression is limited to B cells while CD23b is also found on T cells, FDC, eosinophils, platelets, macrophages and natural killer cells [40, 52]. Both human and murine CD23 can be released as soluble receptors.

Soluble murine CD23 has much lower affinity for IgE than the soluble human isoform.

Mice deficient in CD23 can mount normal Ab responses, with normal germinal center formation and normal in vitro B cell proliferative responses. Ag-specific IgE-mediated enhancement of Ab responses however, is severely impaired in these animals [69].

F EEDBACK R EGULATION BY A NTIBODIES

Abs administered in complex with the Ag have the capacity to regulate the out-coming specific immune response [88, 90]. The regulatory effects are often dramatic and small doses of syngenic Abs can result in several hundred–fold enhancement or in almost complete suppression [42, 61, 112]. The regulatory effects of the Ab are specific for the Ag to which they bind and the immune response to simultaneously administered unrelated Ags, that do not bind the Ab, are neither up- nor down-regulated. Many factors such as the Ab class or subclass and affinity, the nature of the Ag, administration route and dose, influence this regulation. The net effect is therefore difficult to predict for any given situation. A requirement for feedback regulation by Abs to work is that the Ab has to be physically linked to the Ag. This occurs in complexes of Ag-specific Abs together with Ag, or hapten-specific Abs together with hapten-carrier-conjugate, where haptens, e.g. TNP (2,4,6-trinitrophenyl) are small organic molecules coupled to the antigenic protein (the carrier). IgM, IgG and IgE are capable to feedback regulate Ab responses.

Mechanisms behind antibody-mediated suppression

The suppressive effect of Abs (reviewed in [90]) has primarily been described for IgG

[87, 139] and against particulate Ags or Ags with high epitope density, able to form

large Ag/IgG complexes [95, 155, 218]. Primary IgM and IgG production is severely

impaired by passively administered IgG, often <1% of control responses [23, 87, 95,

138]. Both passively administered and endogenously produced IgG can down-regulate

Ab responses [95]. Suppression has been correlated to the affinity and number of Ab

molecules bound to the Ag, but not to complement activation, hemagglutination, or Ab

Feedback Regulation by Antibodies 13

isotype [92, 95, 218]. Intravenously, intraperitoneally and subcutaneously (in this case using soluble Ag in adjuvant) administered Ag/IgG complexes can induce suppression [23, 87, 200, 210]. IgM, which generally induces enhancement against particular Ags, may induce suppression, when administered in high doses [23, 33, 139, 212].

The mechanism behind Ab-mediated suppression has been a subject of controversy over the years and different mechanisms have been proposed (Fig. 6). Ab binding to the Ag may cover and hide the epitopes, making B cells unable to reach and recognize the Ag (Fig. 6A). A rapid removal of Ag/Ab complexes, mediated by FcγR-dependent phagocytosis, preventing B cells from encountering the Ag, is another possible explanation (Fig. 6B). Finally, a direct negative effect on B cells, mediated by the cross- linking of BCR on specific B cells with the inhibitory FcγRIIB, could account for Ab- mediated suppression (Fig. 6C). The masking of epitopes is the only proposal that does not require an intact Fc portion of the Ab (or involvement of FcR). The mechanism proposed for Ab-mediated suppression derived from our findings will be discussed below.

Figure 6. Possible mechanisms for Ab-mediated suppression

Mechanisms behind antibody-mediated enhancement

IgG-mediated enhancement

IgG Abs administered along with the Ag can, under certain circumstances, considerably stimulate the primary immune response against the Ag. This is demonstrated by the

B cell

Ag/Ab BCR

FcγRIIB Ag/Ab -

Phagocytic cell B cell FcγR

Ag/Ab BCR

B Rapid elimination of Ag/Ab by phagocytic cells

C Cross-linking of BCR with FcγRIIB on B cells

A Masking of epitopes

B cells unable to see

the Ag

augmented Ag-specific IgG and IgM production, as well as by the increased number of Ag-specific B cells and memory B cells, compared to responses after immunization with Ag alone [42, 61, 112, 194, 217-219]. Enhancement induced by passively administered IgG imitates the action of endogenous IgG Abs in primary responses or the action of Abs produced during primary responses on secondary responses, which is certainly of physiological relevance [196, 206]. Many factors such as the nature and dose of the Ag, the Ab subclass and the molar Ag:Ab ratio used, may affect the magnitude of the response, which sometimes can be enhanced by up to 1000-fold [42].

Enhancement has been reported against a number of soluble Ags, e.g. keyhole limpet hemocyanin (KLH) [42], tetanus toxoid [119, 183] and bovine serum albumin (BSA) [78] but a modest up-regulation of the response to low doses of a particulate Ag, sheep red blood cells (SRBC), has also been reported [53, 62]. IgG-mediated enhancement has been found in mice given specific IgG1, IgG2a or IgG2b together with Ag [215]. The efficiency of IgG to enhance does not correlate with its affinity [42, 218]. Moreover enhancement is non-epitope specific, meaning that the regulation is not restricted to the epitope to which the Ab binds [42, 78, 217-219]. Complexes formed in vivo or in vitro have the same stimulatory property [42] and IgG-mediated enhancement is effective over a wide range of Ag:Ab ratios [42] and occurs whatever the route of injection [42, 216].

The mechanism behind IgG-mediated enhancement of Ab responses is not well understood. The adjuvant effect of IgG cannot be explained merely by aggregation of the Ag because heat-aggregated Ag cannot induce B cell priming as effectively as Ag/Ab complexes [111]. Enhancement seems to require Fc-mediated functions of IgG molecules [218] and F(ab’) 2 fragments are less effective than intact IgG [111]. The deposition of the Ag on lymphoid follicles [124, 195] has been proposed as an important step in the maturation of B cell responses in germinal centers. During somatic hypermutation and affinity maturation, competition for the Ag determines which B cells are selected to survive and become secondary Ab secreting cells or memory B cells.

IgG-mediated enhancement of Ab responses to soluble Ag could be explained by rapid

B cell priming, facilitated by immune complexes trapped on FDC in lymphoid follicles

[61]. As FDC express CR1 and CR2 [110], FcγRII [153] and probably FcγRIII [86], the

deposition of immune complexes in FDC could be mediated via complement receptors,

FcRs or both. The role of complement has been a subject of controversy. Enhancement

was first correlated with IgG complement activation properties [218] and the adjuvant

effects of IgG in promoting follicular trapping, as well as in generating memory B cells,

Feedback Regulation by Antibodies 15

was found to be abolished in mice deprived of C3 [62, 111, 146]. In contrast, a mutant IgG2a mAb, unable to activate complement, as well as a non-complement-activating IgG1 could efficiently enhance the humoral response [219]. Additionally, a complement activating IgG2a could enhance Ab responses in mice depleted of C3 [219] and recently IgG2a-mediated enhancement was shown to be unimpaired in mice deficient for CR1/2 [9]. A role for FcγRs in IgG-mediated augmentation of Ab responses by the abrogation of IgG1- or IgG2a-mediated enhancement of Ab responses in mice deficient in FcRγ chain (FcRγ ^-/- ) has recently been suggested [215]. A better understanding of the involvement of FcγRs and complement receptors in IgG-mediated enhancement is one of the aims of this thesis and the postulated mechanism derived from our data will be discussed below.

IgE-mediated enhancement

Specific IgE Abs can also induce an increase of Ab responses. As early as 7 days after primary immunization with Ag/IgE complexes, a peak in IgG serum levels can be observed [216]. Ag/IgE complexes also stimulate the primary IgM, IgG and IgE responses and induction of memory [79]. IgE-mediated enhancement is dependent on CD23 [79, 94] and is absent in mice deficient for this receptor [69, 80]. The need of T cells for this effect has also been demonstrated [79] and by using adoptive transfer systems, B cells have been suggested as the effector cells, responsible for IgE-mediated enhancement [80]. Increased CD23-mediated uptake of Ag/IgE complexes by B cells, followed by an efficient presentation for T helper cells, or a direct activation of B cells, initiated by cross-linking of CD23 with BCR, may be the operating mechanisms [80].

IgM-mediated enhancement

Specific IgM has the ability to enhance humoral responses to particulate Ags [53, 87,

91, 95, 212]. Natural serum IgM provides the initial response to foreign Ag and plays a

regulatory role in accelerating the production of high-affinity IgG [14, 60]. Most of the

studies on the enhancement of Ab responses by IgM have been done using SRBC as Ag

[44, 151] but IgM has also been found to be a potent “adjuvant” in vaccination against

malaria in mice [84]. With few exceptions [61], IgM does not enhance Ab responses

against soluble Ags. The effect of IgM on Ab responses is observed in vivo and to

suboptimal doses of Ag [54, 87, 122, 151, 212]. IgM induces a polyclonal increase of

Ag-specific IgM-producing B cells. This is not due to a non-specific "adjuvant" effect

of IgM, as similar effects on an ongoing response to horse red blood cells (HRBC) in

mice given HRBC, SRBC and IgM anti-SRBC Abs are not detected [44, 96]. IgG

responses as well as induction of memory are also augmented by IgM [39, 91, 96]. The ability of IgM to exert its immunostimulatory capacity on Ab production also requires T cells [39, 44, 54]. Mutant mice in which B cells are incapable of secreting IgM have impaired IgG responses to suboptimal doses of T cell-dependent Ags [14].

Activation of C3 by IgM Abs, as well as localization of Ag in the spleen has been suggested as necessary steps in IgM-mediated enhancement of Ab responses. According to this, depletion of C3 by treatment with cobra venom factor (CVF) severely impairs the capacity of IgM to induce an enhanced anti-SRBC response in mice. However, C5 deficiency did not alter the ability of IgM to potentiate the response, demonstrating that complement factors acting upstream of C5 are the crucial ones [93]. Additionally, a point mutated IgM Ab, unable to activate complement was impaired in its capacity to enhance Ab responses against SRBC [93]. Interestingly, it has recently been shown that IgM- but not IgE- or IgG2a-mediated enhancement of Ab responses is impaired in CR1/2-deficient mice (CR1/2 ^-/- ) [9]. It is hypothesized that co-ligation of the BCR and co-receptor by Ag/IgM/C3 complexes causes positive signaling in Ag-specific B cells responsible for the effects of IgM on Ab responses [9, 27, 88].

Despite the role of IgM in enhancing Ab responses to low doses of foreign Ags, it seems to have a protective effect in autoimmunity [19, 152, 154] and deficiency in IgM in mice leads to increased propensity to the development of IgG Abs to auto-Ags [15, 60]. This reminds the paradoxical effects of complement in human autoimmune diseases and suggests that autoreactive IgM may bind to auto-Ags and promote their removal through complement-mediated processes [15].

A UTOIMMUNITY

The response of the immune system against self-components is called autoimmunity.

Mechanisms of self-tolerance protect a normal individual against self-reactive

leukocytes, but when this regulation is disturbed, self-reactive clones of T or B cells can

generate cell-mediated or humoral responses to self-Ags [205]. When reactive

lymphocytes or Abs bind to the target Ag, they mediate direct cellular damage and/or

induce inflammatory responses, with deleterious consequences. Autoimmune diseases

(reviewed in [49]) are poorly understood disorders. Both genetics and environmental

factors contribute to the manifestation of these diseases. Rheumatoid arthritis (RA),

systemic lupus erythematosus, insulin-dependent diabetes mellitus and multiple

Autoimmunity 17

sclerosis are examples of autoimmune diseases in humans. In RA the chronic inflammation and destruction of the joints can lead to invalidity. In systemic lupus erythematosus, auto-Abs to specific tissue Ags like DNA or nuclear proteins are made, producing different symptoms. In diabetes, the insulin-producing cells of the pancreas are destroyed, resulting in decreased production of insulin and higher levels of glucose.

In multiple sclerosis, inflammation and demyelination of the central nervous system take place, ultimately causing paralysis.

Rheumatoid arthritis

RA is a severe systemic autoimmune disease of unknown etiology. It is a relatively common disorder affecting about 1% of the population, most often women between 40 and 60 years old. The diagnosis is based on clinical criteria defined by the American Rheumatism Association [10]. HLA-DR genes are associated with the susceptibility to RA [140, 181]. Chronic inflammation of the joints is the major symptom, with synovial hyperplasia and massive cellular infiltration, which leads to degradation of the cartilage and bone and ultimately to destruction of the affected joints (reviewed in [121]). Many pro-inflammatory cytokines including TNF-α and IL-1, chemokines and growth factors are found in diseased joints. Additionally an abnormal production of Abs, [65] usually reactive with the Fc region of IgG (known as rheumatoid factors) [65], and with collagen type II (CII) are often seen in RA patients [192]. These Abs bind to the Ag, forming immune complexes that are subsequently deposited in the joints and locally activate the complement system. C5a and C3a are released in abundance inducing the secretion of histamine and migration of polymorphonuclear leukocytes towards the articular cavity, which contribute in the perpetuation of the inflammatory process. The binding of immune complexes to FcRs may also contribute to inflammation in RA [59]

and a key role for FcγRIIIa in e.g. induction of TNF-α production [2] and monocyte chemoattractactant protein-1 expression [128] has been suggested. Additionally, FcγR- polymorphism has been associated with risk factors for rheumatic diseases [141].

Despite the advances in the understanding of RA, the causative agent(s) and regulatory mechanisms contributing to the development of the disease remain poorly understood.

In this regard animal models are important tools for the study of RA, permitting the

examination of e.g. molecular receptors, inflammatory mediators or genetic factors in

the promotion of the disease. Findings made in animal models usually lead to similar

results in humans.

Collagen-induced arthritis

Collagen-induced arthritis (CIA) is an animal model for RA that allows us to conveniently analyze the contribution of different factors in the development of the disease [8]. Intradermal injections of homologous or heterologous CII in adjuvants induces the development of CIA in susceptible rats [199], mice [43] and primates [31].

CIA is similar to RA with regard to the spread of affected joints and cellular and humoral responses [97]. The susceptibility to CIA in mice is primarily controlled by MHC genes (H-2 complex) [223] and limited to H-2 ^q and H-2 ^r haplotypes [221].

Synergy between cell-mediated and humoral responses is required to induce arthritis [43, 173]. During the course of CIA, a T cell response to CII, predominantly of Th1 type is followed by a destructive inflammatory response [130] characterized by synovial hyperplasia, mononuclear cell infiltration, pannus formation, tissue injury and cartilage and bone destruction [43]. High levels of circulating CII-reactive Abs also accompany development of CIA, which play a major role in the immunopathology of CIA [99]. The effector mechanism by which Abs contribute to arthritis development is not well understood, but the deposition of Abs in the joints, crossreactive with CII, could be the pathogenic factor that triggers inflammation. FcγRs, displaying both activating and inhibitory functions, play an important regulatory role in the progress of CIA. Mice with the MHC haplotype H-2 ^b are normally resistant to development of CIA. However, if these mice lack the inhibitory FcγRII they become sensitive [226]. Moreover, normally susceptible H-2 ^q mice, develop a more severe condition when lacking FcγRII [113].

Interestingly, activating FcγRs also seem to be involved since mice deficient in FcRγ are

completely protected from arthritis [113]. Moreover, injection of rheumatoid factor-like

complexes selectively induced Ab production with rheumatoid factor activity in wild-

type but not in FcRγ ^-/- mice [144]. The data support the idea that FcγRII inhibits

whereas FcγRI (and FcγRIII) favors the development of CIA. The significance of

FcγRIII involvement in the development of CIA will be discussed below.

THE PRESENT INVESTIGATION

The aim of this work was to study how immune complexes can regulate the immune response in vivo. The involvement of Fc receptors and complement in this regulation has been investigated by using mice selectively lacking one or several of these molecules.

S PECIFIC A IMS

I Characterize the mechanism behind antibody-mediated suppression of immune responses

II Analyze the suppressive effects of IgE and the participation of Fc receptors in this regulation

III Investigate the role of FcRγ ⁺ -follicular dendritic vs. FcRγ ⁺ -bone marrow- derived cells in IgG2a-mediated enhancement

IV Examine the ability of IgG3 immune complexes to regulate in vivo humoral responses

V Determine the contribution of FcγRIII in the development of collagen-

induced arthritis in mice

E XPERIMENTAL M ODELS

Immunizations and assays for antibody responses

Immunization: Mice were immunized intravenously (I-IV) with Ag alone (control group) or with preformed Ag/Ab complexes (experimental group); no conventional adjuvants were used in the immunization and the reagents were given in physiological salt solutions. For the suppression model (I-II) the Ag used was either SRBC or SRBC- TNP (2,4,6-trinitrophenyl coupled to SRBC). HRBC were used as a specificity control.

In enhancement experiments (III-IV) the Ag used was BSA-TNP (2,4,6-trinitrophenyl coupled to BSA) or OVA-TNP (2,4,6-trinitrophenyl coupled to ovalbumin). Unrelated Ags, OVA or KLH ware used as specificity controls. Mouse Abs used were either polyclonal IgG anti-SRBC (I) or mAbs against TNP of IgG (IgG1, IgG2a, IgG2b and IgG3) or IgE isotypes (II-IV). For induction of CIA (V), mice were immunized with bovine CII in complete Freund's adjuvant (CFA), and boosted with bovine CII emulsified in incomplete Freund's adjuvant (IFA), injected intradermally at the base of the tail.

Assays: Serum Ab responses and Ab-forming cells were quantified by ELISA, Plaque- Forming Cell (PFC) Assay or Enzyme-Linked Immunospot (ELISPOT). In ELISA the Ab serum levels were analyzed by SRBC-, HRBC-, BSA-, OVA- or KLH-specific ELISA. The measurement of injected TNP-specific mAbs was therefore avoided. In CIA, Ab serum levels were analyzed by bovine CII-specific ELISA.

Depletion of C3: Mice were injected intraperitoneally with four doses of 100 µl of 100 U/ml CVF over 24 hrs. Treatment with CVF induces the formation of a more stable C3 convertase, which transiently consumes C3. Levels of C3 in sera were assayed before and after CVF treatment, by radial immunodiffusion using polyclonal goat anti-mouse C3 antiserum.

Statistical analysis: Statistical differences between groups were calculated by Student’s

t-test, p values are presented as: not significant, p>0.05; *, p<0.05; **, p<0.01; ***,

p<0.001. Stimulation index was calculated as the geometric mean of the experimental

group divided by that of the control group.

Experimental Models 21

Mice

Mice used in this work were CBA/J (H-2 ^k ), BALB/c (H-2 ^d ), DBA/1 (H-2 ^q ) and different knockouts (Table I) with their corresponding wild-type controls. Mice of all MHC haplotypes respond to SRBC. Mice of haplotypes H-2 ^d , H-2 ^k , H-2 ^p , H-2 ^q and H-2 ^s are high responders when immunized with soluble Ag in complex with specific IgG or IgE.

Class II molecules in mice are encoded by genes of two regions: I-A and I-E. C57BL/6 and 129/Sv (H-2 ^b ), the most common strains used for generation of knockouts, have an I-A ^b -linked low responsiveness in Ab-mediated up-regulation of immune responses [78]. In enhancement experiments, therefore, mutant mice backcrossed into responder strains were used (Fig. 7). Though the optimal strains would have been fully congenic strains, the most important gene locus for the studied Ab responses (I-A) was similar in wild-type and knockout animals. Susceptibility to arthritis in mice is primarily limited to H-2 ^q and H-2 ^r haplotypes [221], so for CIA, the knockout mice used were backcrossed into DBA/1, H-2 ^q background (Fig. 7).

Figure 7. Breeding scheme to backcross the mutant, non-responder H-2

mice into responder H-2

haplotypes.

Adoptive transfer

To generate adoptively transferred mice, single cell suspensions of BM cells were harvested by flushing out the cells from femurs and tibias of donor mice with PBS.

Recipient mice were irradiated (600 rad) 24 hrs before being reconstituted with 1 x 10 ⁷ BM cells in 0.2 ml PBS injected intravenously. Irradiation destroys all leukocytes but not FDC [102, 105] and donor FDC do not normally develop in irradiated mice after

- / - H- 2 b

+ / + H - 2 *

+ / - H - 2 b / *

n

.. ..

The mutant founder animals, usually in H-2

non-responder background, were backcrossed to responder strains (indicated as H-2*) for one or more (2...n) generations. Heterozygous (+/-) mice from the n-backcross progeny were intercrossed and homozygous, mutant and wild-type mice, with the desired haplotype as identified by PCR, were obtained.

x

2

BM or spleen cell reconstitution [76, 102, 103, 129]. In the two models created, either BM cells from FcRγ ^+/+ were transferred into FcRγ ^-/- mice (Table IIA) or BM cells from FcRγ ^-/- were transferred into FcRγ ^+/+ mice (Table IIB). FcRγ ^+/+ animals reconstituted with FcRγ ^+/+ BM cells and FcRγ ^-/- animals reconstituted with FcRγ ^-/- BM cells were used as positive and negative controls, respectively. Six weeks after the transfer the different groups were immunized.

Mutant mice Deficiency Founder ^* Paper ^‡ H2-A (n ^§ )

I H-2A ^b

II, III H-2A ^k (1) FcRγ ^-/- FcγRI, FcγRIII, FcεRI [189]

IV H-2A ^q (5)

I H-2A ^b

FcγRII ^-/- FcγRII [190]

II, IV H-2A ^k (5) FcRγ ^-/- x FcγRII ^-/- FcγRI, FcγRII, FcγRIII, FcεRI [ ^† ] I H-2A ^b

FcγRIII ^-/- FcγRIII [85] V H-2A ^q (5)

CD23 ^-/- CD23 [69] II H-2A ^k (10)

β2m ^-/- FcRn [114] I H-2A ^b

CR1/2 ^-/- CR1, CR2 [133] IV H-2A ^q (2)

Table I. Fc receptors or complement receptors ablated in knockout mice.

Reference number where founder animals are described. ^† FcRγ ^-/- x FcγRII ^-/- mice generated in the laboratory of Dr. J. Ravetch at the Memorial Sloan Kettering Cancer Institute, were purchased from Taconic Farms. ^‡ Number of the paper in which mutant mice were used. ^§ n: number of backcrossed- intercrossed generations, as explained in Fig. 7.

Donor Recipient Chimera

Phenotype Phenotype FDC Leukocytes

A +/+ -/- -/- +/+

B -/- +/+ +/+ -/-

Table II Generation of chimeric mice

Results and Discussion 23

R ESULTS AND D ISCUSSION

(I) Involvement of FcγRs in IgG-mediated suppression

Although IgG-mediated suppression of Ab responses is well documented, the mechanism behind this regulation has been elusive. Here by using mice lacking one or several FcγRs the involvement of these receptors in IgG-mediated suppression of Ab responses is studied.

Suppression in Fc γ RII ^-/- mice

As briefly discussed in the introduction of this thesis, three mechanisms for IgG- mediated suppression have been proposed (reviewed in [90]). One of them postulates that cross-linking of BCR and FcγRIIB on B cells is the operating mechanism, negatively regulating cell activation. To test this hypothesis, FcγRII ^-/- and FcγRII ^+/+

wild-type controls were immunized with SRBC alone or together with specific IgG. We found that IgG induced the same degree of suppression of SRBC responses in wild-type as in knockout animals, indicating no involvement of FcγRII in this regulation.

Suppression in FcR γ ^-/- mice

Another postulated mechanism is a rapid elimination of IgG-coated Ag from the system, making B cells unable to recognize the Ag. FcγRI and FcγRIII (both absent in FcRγ ^-/- mice) could be the receptors mediating this rapid endocytosis and prompt elimination of the Ag. When FcRγ ^-/- and FcRγ ^+/+ wild-type controls were tested, IgG induced the same degree of suppression in wild-type as in knockout animals, suggesting no involvement of FcγRI and/or FcγRIII in this regulation.

Suppression in FcR γ ^-/- x Fc γ RII ^-/- mice

To exclude redundancy between the different FcγRs, double knockout FcRγ ^-/- x FcγRII ^-/- mice, lacking FcγRI, FcγRII and FcγRIII were tested. SRBC-specific IgG induced the same degree of suppression of SRBC responses in wild-type as in knockout animals.

Suppression in β 2m ^-/- mice

β2m is required for expression of MHC class I molecules and β2m ^-/- mice are deficient

in cytotoxic T cells [114, 228]. β2m is also a constitutive component of FcRn and mice

deficient in β2m lack FcRn [106]. The long life of IgG in serum is attributed to the

FcRn-mediated protection of all subclasses of IgG and therefore IgG protection is

disrupted in β2m ^-/- mice. Additionally, FcRn has been found in human monocytes, intestinal macrophages, and DC suggesting that FcRn could have FcγRs-related functions in these cells [227].

To test if FcRn was involved in IgG-mediated suppression, β2m ^-/- mice, as well as wild- type controls, were tested in our model. The same degree of suppression was induced in wild-type as in knockout animals, indicating no involvement of FcRn in this regulation.

Suppression by F(ab’) 2

To investigate the ability of F(ab’) 2 fragments to suppress Ab responses, SRBC-TNP were injected alone or together with F(ab’) 2 fragments of anti-TNP specific mAb IgG2a.

F(ab’) 2 fragments shown to be nearly as efficient as intact IgG2a in suppressing the Ab response against SRBC.

Suppression by IgE

Monoclonal IgE anti-TNP was also shown to be suppressive when administered together with SRBC-TNP. This added strong support to the existence of suppression independent of the Fc portion of IgG.

Priming of T cells

To test if the suppression of B cell responses also abolished the priming of T cells, a hapten carrier system was used. Mice were first immunized with either SRBC-TNP alone or SRBC-TNP/IgG. A few mice from each group were tested to confirm that the primary SRBC-response was suppressed >90% by IgG. Spleen cells from the respective groups or from unimmunized mice were transferred to irradiated syngenic recipients that were subsequently immunized with 4-hydroxy-5-iodo-3-nitro-phenacetyl-SRBC (SRBC-NIP). The IgG anti-NIP response, reflecting the number of primed SRBC- specific T helper cells, was enhanced (compared to recipients into which unprimed cells were transferred) to the same degree in groups transferred with cells primed with SRBC/TNP/IgG or with SRBC-TNP only. This shows that although IgG can suppress primary responses, the priming of T cells was intact.

The mechanism behind feedback suppression

The mechanism behind Ab-mediated suppression has been difficult to elucidate.

Contradictory results are found in the literature respect whether the Fc-portion of IgG is

needed or not and if suppression is epitope specific or not.

Results and Discussion 25

The unexpected finding that IgG was able to efficiently suppress Ab responses in mice deficient for all the known FcγRs suggested that IgG inhibits through Fc-independent mechanisms. Ab molecules covering the Ag may prevent B cells from binding and responding to Ag. When mice are given 4 x 10 ⁶ SRBC-TNP together with 10 µg IgG (4 x 10 ¹³ molecules), the Ab response against SRBC is suppressed. The number of IgG molecules available per erythrocyte (molar ratio IgG/SRBC-TNP) is 10 ⁷ ; probably enough to cover the entire surface of the erythrocyte (with a high degree of TNP coupling) and shield the Ag from the B cells by steric hindrance.

An interesting finding was that when injected together with SRBC-TNP, TNP-specific F(ab’) 2 fragments could suppress the response to all SRBC determinants. This showed that non-epitope-specific suppression could be obtained, without involvement of the Fc portion of IgG.

According to our hypothesis, depending on the epitope density, suppression will be epitope- or non-epitope-specific. When IgG binds to TNP-epitopes present at high density, non-epitope-specific suppression is obtained, that is, IgG sterically hinders the recognition by B cells of all (specific and non-specific) epitopes on the surface of SRBC. On the other hand, when IgG binds to an epitope that is not so abundant, only epitope-specific suppression would be expected.

We conclude that the operating mechanism behind feedback suppression of IgG, is most likely to be masking of antigenic determinants. B cells are unstimulated with a complete lack of primary Ab response, but phagocytosis/endocytosis of the Ag and presentation to T cells can take place normally.

(II) Involvement of FcRs in IgE-mediated suppression

The ability of IgE to enhance Ag-specific Ab responses is well documented. This regulation is mediated by CD23 (FcεRII, the low affinity receptor for IgE) and is abrogated in mice genetically deficient for this receptor [69, 79]. The suppressive ability of IgE, previously described in (I) however, is less understood. IgE is known to bind FcεRI, CD23, FcγRII and FcγRIII [109, 120, 191], therefore it is possible that these receptors could be involved in this regulation. The cross-linking of FcεRI leads to mast cell degranulation, an effect that is positively or negatively regulated by FcγRIII [201]

or FcγRII [47] respectively. Moreover, co-cross-linking of FcγRII and BCR can down-

regulate B cell activation [137, 168, 208]. A regulatory role for CD23 has also been reported and while mice overexpressing this receptor have impaired IgG and IgE responses [148, 198], mice deficient in CD23 have higher IgE levels [224].

Additionally, cross-linking of CD23 or co-cross-linking of CD23 and BCR inhibits B cell proliferation [25, 123].

In this paper (II) IgE-mediated suppression of Ab responses is further characterized and the involvement of FcRs in this regulation is studied in mice deficient for several FcRs.

IgE-mediated suppression of IgM responses

The suppressive effect of IgE and IgG was compared by immunizing mice with IgE or IgG2b anti-TNP mAb together with SRBC-TNP and HRBC (a non-cross-reacting Ag, used as a specificity control). The number of spleen B cells producing IgM anti-SRBC five days later, revealed that both IgE and IgG2b were able to suppress SRBC responses (>90%). The suppression was Ag-specific since no suppression of HRBC responses could be detected.

IgE-mediated suppression of IgG responses in FcR γ ^-/- , Fc γ RII ^-/- and CD23 ^-/- mice

Knockouts and wild-type mice were immunized with SRBC-TNP alone or together with IgE anti-TNP. IgE was able to induce equally efficient suppression of SRBC responses in FcRγ ^-/- and in FcγRII ^-/- as in wild-type controls. This rules out ITIM-mediated negative inhibition via Fc γ RII by immune complex-mediated co-cross-linking of this receptor and the BCR, as well as phagocytosis-mediated elimination of the Ag by these receptors, as the effector mechanism for IgE-mediated suppression of Ab responses.

IgE could also induce suppression in CD23 ^-/- mice, although it was less efficient than in

wild-type animals. Murine CD23 is expressed only on B cells and FDC [125, 156],

implying that any suppressive effect mediated by IgE, via CD23, should most likely

take place at the level of B cells. Engagement of CD23 on B cells in vitro using Ag/IgE

complexes has been reported to result in a suppressed response [25, 123]. The

suppression was seen with a high level of co-cross-linking of CD23 and the BCR

whereas a low level induced an enhanced response [25]. Under our experimental

conditions SRBC-TNP/IgE complexes due to their large size and the high density of

TNP-residues, may be able to efficiently cross-link CD23 on the B-cell-surface and

thereby produce negative signals. The exact mechanisms and requirements for the

opposite effects induced by CD23 on immune responses remain unknown. An attractive

Results and Discussion 27

model is that a low degree of cross-linking of CD23 induces increased Ag presentation and that a high degree of cross-linking overrides this and induces the suppressive effect [41].

The mechanism behind IgE-mediated suppression

In summary, IgE could induce efficient suppression of IgG responses in mice lacking all the known FcRs for IgE, suggesting that Fc-mediated effects are of minor importance and that the major mechanism behind IgE-mediated suppression, under the experimental conditions used, is masking of epitopes. This “neutralization” of the Ag, may prevent Ag-specific B cells from recognizing and responding to the Ag. A minor contribution of CD23 was though suggested, by the less efficient suppression obtained in CD23 ^-/- mice.

(III) FcRγ+ bone marrow-derived cells in IgG2a-mediated enhancement It is known that immunization with a soluble Ag in complex with specific IgG can lead to an augmentation of Ag-specific Ab production [42, 61, 112], an effect that is called Ab-mediated enhancement. Previous findings from our laboratory showed that enhancement after immunization with BSA-TNP in complex with TNP-specific mAb IgG2a or IgG1 was abrogated in FcRγ ^-/- mice (lacking both FcγRI and FcγRIII), but not in FcγRIII ^-/- animals (selectively deficient in FcγRIII) [215]. Thus suggesting an involvement of FcγRI (alone or together with FcγRIII) in this regulation [215]. The cellular mechanism behind IgG2a-mediated enhancement of Ab responses is further characterized in this paper.

Isotype profile in IgG2a-mediated enhancement

In the analysis of the isotype profile of the Ab responses after immunization with immune complexes, we found that predominantly IgG1, but also IgG2a, responses were up-regulated. IgG1 is generally accepted as a Th2 isotype and IgG2a as a Th1 isotype [108, 136, 145]. The massive production of IgG1 after immunization with Ag/IgG2a complexes indicates a Th2 profile.

T cells in IgG2a-mediated enhancement

To exclude the possibility that the enhancing ability of IgG was a consequence of T

helper cell-independent B cell activation, the ability of IgG2a to induce enhancement in

BALB/c and BALB/c nude mice, which lack functional T helper cells was examined. A

significant increase of BSA-specific IgG was observed in BALB/c but not in nude mice,