University of Groningen B cells in ANCA-associated vasculitides von Borstel, Anouk

(1)

University of Groningen

B cells in ANCA-associated vasculitides

von Borstel, Anouk

DOI:

10.33612/diss.93537940

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

von Borstel, A. (2019). B cells in ANCA-associated vasculitides: from pathogenic players to biomarkers. University of Groningen. https://doi.org/10.33612/diss.93537940

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

B cells in ANCA-associated Vasculitides

From Pathogenic Players to Biomarkers

Anouk von Borstel

(3)

The research presented in this thesis was financially supported by:

The printing of this thesis was financially supported by:

Cover Design: Nikki Vermeulen - Ridderprint Lay-out: Nikki Vermeulen - Ridderprint Printing: Ridderprint - www.ridderprint.nl ISBN: 978-94-034-1866-7 (printed version) 978-94-034-1865-0 (electronic version) Copyright © 2019, Anouk von Borstel

(4)

B cells in ANCA-associated Vasculitides

From Pathogenic Players to Biomarkers

PhD thesis

to obtain the degree of PhD at the University of Groningen

on the authority of the Rector Magnificus Prof. C. Wijmenga

And in accordance with the decision by the College of Deans. This thesis will be defended in public on Wednesday 11 September 2019 at 12:45 hours

by

Anouk von Borstel

born on 7 May 1991 in Luzern, Switzerland

(5)

Supervisors Prof. P. Heeringa Prof. C.A. Stegeman Co-supervisors Dr. J.S.F. Sanders Dr. W.H. Abdulahad Assessment Committee Prof. N.A. Bos

Prof. L.B. Hilbrands Prof. J.D. Laman

(6)

Voor Olaf

(7)

Paranymphs Dr. Olaf Perdijk Gerjan Dekkema

(8)

1

General Introduction and

Aims of this Thesis

(11)

10 | Chapter 1

Anti-Neutrophil Cytoplasmic Autoantibody-associated Vasculitides

Anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis (AAV) is a rare, but severe autoimmune disease that affects 10-20 individuals per million annually1_.

AAV usually affects elderly individuals with a peak-age of disease onset between 64-75 years1_{. The disease can be fatal if left untreated. AAV is a form of small vessel vasculitis}

that is characterized by inflammation of small- to medium-sized blood vessels2_{and the}

presence of circulating ANCA. Necrotizing inflammation, thickening, and scarring of the blood vessel walls that disturbs normal blood flow are characteristics of vascular inflammation in AAV. Obstruction of blood vessels will result in diminished oxygen exchange and can eventually lead to tissue necrosis and severe organ failure. AAV can manifest itself in all small- and medium-sized blood vessels in the body, but has a predilection for the upper airways, kidneys, lungs, and skin and often affects multiple organs or tissues simultaneously.

AAV can be divided, based on clinical and pathological symptoms, into granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), and eosinophilic GPA2_{. GPA}

patients show, often in addition to necrotizing vasculitis and/or glomerulonephritis, extravascular necrotizing granulomatous inflammation, whereas MPA patients lack (extravascular) granulomatous inflammation. In addition, while frequent in GPA, destructive upper airway involvement in MPA is absent. ANCA in AAV are directed against proteinase (PR) 3 or myeloperoxidase (MPO), which are enzymes found in granules of neutrophils and monocytes3_{. Although anti-PR3- and anti-MPO-ANCA can}

be found across the spectrum of AAV, ANCA of GPA patients are mainly directed against PR34_{, whereas MPA patients more often present with ANCA directed against MPO}5_.

AAV Pathogenesis

The etiology of AAV is unknown but is considered to be multifaceted involving genetic predisposition, environmental exposure (e.g. infections, toxic exposures) and acquired alterations in the immune system. Two genome-wide association studies in AAV patients found single nucleotide polymorphisms in the genes coding for the major histocompatibility complex (MHC) II related to the disease. These associations segregated along the ANCA specificity, with human leucocyte antigen (HLA) DP associated with PR3-ANCA positivity and HLA-DQ with MPO-ANCA positivity6,7_{. This}

indicates that antigen (Ag) presentation by specific major histocompatibility complex (MHC)-II molecules might predispose individuals to develop specific ANCA and AAV. Other single nucleotide polymorphisms associated with AAV susceptibility include mutations in the PRTN3 and SERPINA1 genes6_{. PRTN3 encodes the ANCA-Ag PR3, and}

SERPINA1 encodes a null allele of the enzymatic inhibitor α1-antitrypsin8_{, allowing}

(12)

1

General Introduction and Aims of this Thesis | 11

The current line of thought is that Staphylococcus aureus infection or exposure is an important driver for onset and relapse of AAV. The majority (± 60%) of GPA patients are nasal carriers of S. aureus9_{. This nasal presence of S. aureus correlates with higher relapse}

rates9,10_{that can be prevented by oral antibiotic treatment}11_{. To date, the exact role of S.}

aureus in the disease pathogenesis is unknown but it has been proposed that S. aureus

infections may induce tolerance breakdown to self-Ags in susceptible individuals and contribute to triggering of disease relapses. With respect to tolerance breakdown to self-Ags by S. aureus, at least two mechanisms have been put forward. First, it has been reported that S. aureus contains proteins highly homologous to a complementary (c) form of human PR312_{. cPR3 is the protein translated from the antisense DNA strand of the}

PR3 gene. Immunization of mice with human cPR3 or a synthetic homolog resulted in the production of antibodies (Abs) directed against cPR3 and, via idiotype-anti-idiotype responses, ANCA directed against PR3, indicating that these homologous proteins can contribute to breakdown of tolerance12_{. However, the relevance of this finding}

remains controversial since in the original study only seven out of 34 PR3-ANCA positive patients presented with cPR3 Abs12_{, and increased reactivity against cPR3 could not be}

confirmed in an independent cohort of AAV patients13_{. Second, other S. aureus derived}

proteins, e.g. extracellular adherence protein (Eap) and staphylococcal peroxidase inhibitor (SPIN), could also indirectly contribute to autoimmune disease onset14_{. Eap}

and SPIN are recently identified enzyme inhibitors produced by S. aureus that target and form complexes with PR3 and MPO, respectively15,16_{. Importantly, natural (low-affinity)}

Abs directed against PR3/MPO (natural ANCAs) have been demonstrated in sera of healthy controls (HCs)17,18_{, indicative of the presence of natural ANCA-producing B cells.}

Hypothetically, these Eap:PR3 or SPIN:MPO complexes may be recognized by those natural ANCA-producing B cells and become internalized. Since B cells are professional Ag-presenting cells, processing of these complexes followed by presentation of bacterial peptides by MHC-II molecules may activate EAP- or SPIN-specific T helper (Th) cells. In turn, these EAP- or SPIN-specific Th cells provide help to the natural ANCA-producing B cell, allowing isotype switching and affinity maturation, leading to the production of high affinity pathogenic IgG ANCA and thus initiation of autoimmune disease19_{. In}

summary, several mechanisms have been proposed by which S. aureus may cause a breakdown of tolerance in AAV patients but direct evidence pinpointing S. aureus as the instigator of autoimmunity in AAV is lacking.

S. aureus also seems to play a role in triggering relapses in AAV patients. Using similar

mechanisms as described for disease onset, Eap:PR3 or SPIN:MPO complexes could also play a role in inducing disease flares via reactivation of PR3- or MPO-specific B cells. In addition, S. aureus-derived constituents such as CpG motifs and superAgs may also reactivate the autoinflammatory response in AAV patients20,21_{. This notion is supported}

by studies showing that peripheral blood mononuclear cells (PBMCs) derived from GPA

(13)

12 | Chapter 1

patients stimulated with CpG and IL-222_{or IL-21 and B cell activating factor (BAFF) trigger}

in vitro PR3-ANCA production. In addition, superAgs derived from S. aureus may activate

autoreactive immune cells by forcing cell contact between T- and B cells or T cells and Ag-presenting cells24_{. Thus, S. aureus-derived constituents may contribute to tolerance}

breakdown and triggering inflammatory responses in AAV patients in various ways. The pathogenesis of AAV involves a complex interplay between several immune cells. As described before, it is believed that the AAV pathogenesis is induced by (super) Ags of S. aureus at the respiratory epithelium. Ag-presenting cells present (S. aureus-derived) Ags to naïve T cells and produce interleukin (IL) 23, resulting in naïve T cell activation and generation of effector Th17 cells. These Th17 cells secrete their signature cytokine IL-17 that activates innate immune cells such as macrophages to produce the pro-inflammatory cytokines IL-1β and tumor necrosis factor (TNF) α. IL-1β and TNFα prime neutrophils to express adhesion molecules and enzymes such as PR3 and MPO on their surface3_{. PR3 and MPO are also released upon neutrophil priming and are}

sampled by dendritic cells and presented to Th cells. Th cells play an important role in AAV, as demonstrated by the dominant ANCA isotypes (i.e. IgG1 and IgG4)25_{that are}

only formed after isotype switching of the autoreactive B cell interacting with a Th cell. Next to these effector T cells, regulatory T cells (Tregs) are considered to fundamentally contribute to the AAV pathogenesis. These cells are crucial for sustaining tolerance but have been found to be defective in function in AAV26_{. Moreover, perturbations in the}

distribution of circulating Th cell subsets in GPA patients have been reported25,27,28_{. For}

example, an increase in a subset of memory CD4+_{T cells, termed effector memory T}

(T_EM) cells, has been detected in peripheral blood of GPA patients in remission compared to HCs29_{. Interestingly, these circulating CD4}+_T

EM cells decrease during active disease

which could be consistent with their migration towards inflamed tissues (e.g. the kidneys) where these cells may contribute to tissue damage30_{. Moreover, aberrant}

Th cell responses have been reported in GPA in terms of skewing towards a Th17 cell phenotype as demonstrated by an increased circulating T_EM17 cell frequency31_and

serum IL-17A levels32_{, and an expanded circulating population of PR3-specific Th17}

cells20,32_{. Collectively, these observations indicate that effector Th cells, mainly T} EM17

cells, play an important role in the AAV disease pathogenesis.

Th cells also have an important function in the activation of B cells. In the AAV pathogenesis, help of these Th cells is needed for B cell activation and differentiation, ultimately leading to the formation of ANCA-producing plasma cells. The interaction between Th- and B cells takes place in germinal centers of secondary lymphoid organs. The T-B cell interplay triggers B cells to undergo isotype switching and somatic hypermutation to increase their Ab affinity33_{. Ag-specific plasmablasts (i.e. the precursors}

of plasma cells)/plasma cells typically migrate to the bone marrow after finishing the germinal center reaction and are rarely found in the circulation34_{. However, recent}

(14)

1

studies suggest that PR3-specific plasmablasts can be detected in the circulation of GPA patients35_{. Plasmablasts and plasma cells need several growth factors to survive and}

provide long-term humoral memory. One of these growth factors is B cell activating factor (BAFF) which is found in increased abundance in serum of GPA patients36_{. This}

has led to the hypothesis that increased B cell survival in the periphery of AAV patients enhances ANCA production initiating the effector phase of AAV that ultimately causes tissue injury. In the effector phase of the disease, release of pro-inflammatory cytokines, due to for example infection with S. aureus, primes neutrophils to express PR3 and MPO on their cell surface and induces expression of adhesion molecules on both neutrophils and endothelial cells. These adhesion molecules allow neutrophils to bind to endothelial cells of small blood vessels37_{. Next, ANCA bind to PR3 or MPO on the surface of the}

adherent neutrophils inducing full-blown neutrophil activation leading to the release of their granular contents and production of reactive oxygen species causing endothelial cell injury. Furthermore, upon neutrophil degranulation, the ANCA-Ags PR3 and MPO are deposited in the blood vessel wall which may further activate the Ag-specific cellular immune response and ultimately result in the formation of granulomas38_{. More recently,}

ANCA have been found to induce the formation of neutrophil extracellular traps (NETs)39_.

NET formation is a form of regulated cell death termed NETosis in which neutrophils discharge nuclear chromatin decorated with granule proteins forming an extracellular web to trap and kill bacteria. Since NETs are covered with various potentially injurious and immune stimulating proteins, including the ANCA Ags PR3 and MPO, it has been proposed that ANCA-induced NET formation contributes to the vessel wall damage and promotes the autoimmune response in AAV.

B cells in AAV

The most convincing evidence that B cells are crucially involved in the pathogenesis of AAV comes from studies involving B cell depletion as a treatment for AAV. In clinical trials, B cell depletion in peripheral blood using rituximab was shown to be an effective therapy for remission induction and maintenance in AAV patients40,41_{. Rituximab is a}

mouse-human chimeric monoclonal Ab that targets CD20 on B cells, thereby effectively depleting all B cells, except CD20-_{plasma cells. Depletion of circulating B cells in AAV}

patients by rituximab results in a reduction of total IgG42_{and also ANCA}40_{levels. This}

indicates that rituximab indirectly reduces (auto)Ab levels by depleting other B cells or directly via depletion of ANCA-producing plasma cell precursors. However, B cells possess multiple other functions that can potentially contribute to AAV pathogenesis, such as Ag presentation43_{and cytokine production}44_{. In the context of AAV, little is}

known about the involvement of those Ab-independent properties of B cells in disease pathogenesis. Therefore, it is of interest to further investigate B cell distribution and

(15)

14 | Chapter 1

function in AAV. Analyses of the circulating B cell repertoire of GPA patients revealed several alterations in B cell subset distribution when compared to HCs. One of these alterations is that GPA patients show decreased circulating regulatory B cell (Breg) frequencies45–47_{. Like Tregs, these cells play a role in maintaining immune tolerance}48_.

Importantly, their exact role and function in GPA pathogenesis is not yet elucidated. Other alterations in the circulating B cell compartment of GPA patients that have been reported include higher naïve B cell and lower memory B cell frequencies compared to HCs46_{. Additionally, circulating B cells of active GPA patients are in a heightened}

activated state compared to circulating B cells of GPA patients in remission and HCs28_.

It is currently unknown why these B cells of GPA patients are more activated. In other autoimmune diseases (i.e. primary Sjögren's syndrome (SS) and rheumatoid arthritis (RA)) important differences in the B cell receptor (BCR) signaling pathway have been demonstrated to be the basis of increased B cell activation49_.

In both SS and RA patients, protein levels of Bruton’s tyrosine kinase (BTK) in B cells are increased and related to increased B cell activation in these patients50_{. BTK is}

critically involved in the BCR signaling pathway. The BCR signaling pathway is initiated upon Ag binding and cross-linking of two BCRs triggering an intracellular signaling cascade that starts with the activation of the Src tyrosine kinase Lyn (Figure 1). Lyn phosphorylates the immunoreceptor tyrosine-based activation motifs (ITAMs) of Iga and Igβ BCR co-receptors (also known as CD79). Spleen tyrosine kinase (SYK) becomes activated upon binding to these phosphorylated ITAMs, and this protein continues to phosphorylate B cell linker protein (BLNK) and BTK. The direct downstream target of BTK is phosphatidylinositol-specific phospholipase C (PLC) γ2. PLCγ2 phosphorylation occurs when both BTK and PLCγ2 bind to phosphorylated tyrosines on BLNK. Eventually, PLCγ2 induces the mitogen-activated protein kinase (MAPK) pathway, resulting in the activation of the extracellular signal-regulated kinase (ERK). However, PLCγ2 also induces transformation of phosphatidylinositol-bisphosphate (PIP2) to inositol trisphosphate (IP3) and diacylglycerol (DAG), resulting in increased cytoplasmic calcium levels. These increased calcium levels allow DAG to ultimately activate S6 kinase, which induces nuclear factor κ-light chain enhancer of activated B cells (NF-ĸB) activation51_{. NF-ĸB is}

a transcription factor that regulates DNA transcription resulting in cytokine production and mediating cell survival52_{. NF-ĸB activation also induces B cell survival, differentiation,}

cytokine production and Ab secretion51_{(Figure 1). Since B cells of AAV patients persist in}

an activated state and play an important role in the disease pathogenesis, protein levels or phosphorylation of molecules in the BCR signaling pathway are likely to be increased. Therefore, characterization of the BCR signaling cascade in AAV may contribute to our understanding of B cell activation in AAV pathogenesis and could lead to novel and more specific treatment targets.

(16)

1

BLNK CD79 P P ERK SYK P BTKPLCɣ P P P S6 PIP2 IP3 AKT

Survival & Differentiation Cytokine production P P Iga Igb Lyn CD19 P Ca2+ Ca2+ Ca2+ Ca2+ DAG NF-κB (auto)Antibody Production

Figure 1. BCR signaling cascade. Simplified overview of the BCR signaling cascade after the BCR becomes cross-linked by Ag. Lyn phosphorylates ITAMs of the BCR as well as Iga and Igβ (CD79). Phosphorylated CD79 mediates activation of SYK, which continues to phosphorylate CD19, BTK, and BLNK. Next, PLCγ is phosphorylated that activates the MAPK pathway, involving ERK and increasing intracellular calcium levels. PLCγ also induces PIP2 conversion into IP3 and DAG, which activate the AKT pathway in which S6 is involved. Eventually, both pathways result in NF-ĸB activation that ultimately results in Ab production, B cell survival, differentiation, and cytokine production.

Therapies Used in AAV

Since the introduction of cyclophosphamide and glucocorticoids, the prognosis for patients with AAV has significantly improved53_{. More recently, rituximab was}

introduced to treat AAV patients and found to be as efficacious as cyclophosphamide in inducing remission54,55_{. After induction therapy with rituximab or cyclophosphamide,}

patients switch to maintenance therapy consisting of either azathioprine (AZA), methotrexate, or mycophenolate mofetil (MMF) combined with corticosteroids, which are tapered over time56_{. Additionally, a randomized controlled trial compared the}

effectiveness of rituximab and AZA maintenance therapy after induction of remission with cyclophosphamide, demonstrating a lower relapse rate in patients treated with rituximab54_{. A second rituximab maintenance study demonstrated that the intensity of}

the B cell depleting therapy can be guided by monitoring circulating peripheral blood B cells and the level of ANCA57_.

Current induction and subsequent maintenance therapies successfully induce remission in most patients and reduce subsequent relapse rate. However, AAV patients are still at risk for cumulative tissue damage resulting from disease relapses and toxic side effects of immunosuppressive treatment58_{. Patients who receive MMF as maintenance}

(17)

16 | Chapter 1

therapy have a higher relapse risk compared to AZA-treated patients59_{. Both treatments}

are non-specific and target proliferation of leucocytes (AZA) and lymphocytes (MMF)60_{. This discrepancy between the relapse rates in AAV patients treated with}

these immunosuppressive drugs and the effects of these therapies on immune cell functioning remains poorly understood. Therefore, there is a need to develop more specific treatments to limit side effects. As described before, B cell depletion using rituximab might be more effective for remission maintenance54_{. However, depleting}

all circulating B cells (including Bregs) seems an over-rigorous method and long-term effects of this therapy are to date unclear. Thus, there is still a need to further understand the pathogenic mechanisms involved in AAV in order to develop more specific and effective treatment.

Predicting Disease Relapses

To date, treatment strategies for AAV are not completely effective in preventing disease relapses. During the course of their disease, ±60% of GPA patients experience disease relapses compared to around 25% of MPA patients61_{. Each episode of active disease}

in AAV patients causes, in addition to renewed treatment toxicity, additional tissue injury and can eventually result in severe organ failure. Therefore, prediction of disease relapses is pivotal for patient care62_{, so that relapses can be treated at an early stage or}

even prevented. To predict ensuing AAV disease relapses, a (bio)marker, i.e. a measurable and validated indicator correlated to an upcoming disease relapse, is needed. These markers can include a single marker, a set of (bio)markers or a method to assess the patients’ immune status. Previously, multiple studies aimed at identifying biomarkers that are related to future disease relapses in AAV. Markers to predict disease relapses proposed so far include pulmonary involvement during active disease, PR3-ANCA positivity63_{and chronic nasal carriage of S. aureus}9_{. However, none of these markers}

proved to be accurate indicators for disease relapses. Other suggested biomarkers for AAV are the serum ANCA titer and changes herein during follow-up. Although in some studies correlations between future relapses and persistence or a rise in ANCA titers have been reported64–66_{, other studies could not confirm such associations}67_{. Indeed, a}

meta-analysis by Tomasson et al. showed that rises in serum ANCA titer over time only weakly predicted disease relapses and were ineffective in predicting disease relapses for all AAV patients68_{. ANCA titers are determined by immunofluorescent techniques,}

and recently novel and more quantitative methods for measuring PR3-ANCA levels have been introduced (e.g. Phadia ImmunoCAP analyzer). Recently, variation in in vitro PR3-ANCA production was assessed as a potentially sensitive approach for predicting disease relapses in GPA patients. In this cell culture assay, PBMCs are cultured in the presence of CpG, BAFF and IL-21 and PR3-ANCA levels are measured in supernatants

(18)

1

using the Phadia analyzer. Prospective monitoring of in vitro PR3-ANCA levels turned out to be very variable between patients and did not aid in predicting future disease relapses better than ANCA titers21_{. Thus, there is an urgent need for further prospective}

studies to discover markers or develop methods to more accurately identify those patients at risk for relapse.

Aims of this Thesis

The research presented in this thesis focuses on the role of B cells in the disease pathogenesis of AAV (Figure 2), with a major focus on GPA. B cells play a crucial role in the GPA pathogenesis as demonstrated by the efficacy of B cell depletion therapy. However, the exact involvement of regulatory and effector B cell functions in the disease pathogenesis is not yet fully understood. To further expand our understanding of B cell activation, function and response to treatment in AAV, studies were performed on B cells from GPA patients to map the BCR signaling pathway and assess the effects of BCR activation and commonly used immunosuppressive drugs on B cell function. In addition, the B cell phenotype distribution and its relation to disease relapses in AAV was explored.

Besides the possible pathogenic role that B and T cells can exert in AAV, these cells can also play an important role in the suppression of inflammation. It has been proposed that the suppressive function of Tregs and Bregs could be decreased in AAV, thereby allowing continuation of autoimmune inflammation. Therefore, our aim in Chapter 2 was to review the literature on Tregs and Bregs focusing on the role both regulatory immune cell subsets could play in the AAV disease pathogenesis. We propose that an imbalance between regulatory and effector functions underlies the pathogenic process in AAV, and that this imbalance is responsible for decreased suppression by Bregs and Tregs.

The exact involvement of Bregs and potential aberrancies in Breg-mediated suppression in the disease pathogenesis of AAV is not yet elucidated. A decreased Breg frequency has been demonstrated in the circulation of AAV patients, but no clear functional deviance of Bregs has been established in AAV. Since Th17 cells play an important role in the AAV pathogenesis, we hypothesized in Chapter 3 that Bregs of GPA patients are less effective at suppression of the Th17 response. To address this hypothesis, we first assessed the frequency of CD24hi_CD38hi_{Bregs and Th17 cells in the circulation of GPA}

patients and subsequently studied the functional effect of Bregs on Th17 cell expansion

in vitro (Figure 2).

(19)

18 | Chapter 1

+

_{Future disease relapse?}

T-cell IL-17 IFNɣ Breg IL-17 IFNɣ

.

B-cell IL-10 regulatory eﬀector

.

Standard Immunosuppressive Medication BTK P IL-10 Pro-inﬂammatory cytokines ANCA Levels? BTK Blockade

.

Circulating B-cells Breg Blood donation GPA patient T-cell Pro-inﬂammatory cytokines

Figure 2. Knowledge gaps pertaining to the role of B cells in AAV. B cells are considered crucial players in the AAV pathogenesis but their exact role is not completely understood. In this thesis, several aspects concerning the contribution of B cells to the AAV pathogenesis were investigated including functional implications (e.g. eff ector and regulatory functions) in disease development (Chapters 2, 3 and 4), as treatment target (Chapters 4 and 5) and as a dynamic circulating immune cell population related to relapses (Chapter 6). See the text for a more detailed explanation. Although B cells seem to persist in an activated state in GPA, little is known about the status of the BCR signaling pathway (Figure 1). In other autoimmune diseases, it has been demonstrated that BCR sensitivity and signaling is increased and correlates with autoAb levels. Possibly, deviations in this pathway could contribute to GPA pathogenesis as well (Figure 2). Therefore, in Chapter 4, we studied the BCR signaling pathway in B cells from GPA patients focusing on the levels and phosphorylation status of BTK, a tyrosine kinase essential in downstream BCR signaling (Figure 1). Protein levels and phosphorylation status of BTK, as well as various other signaling molecules up- and downstream of BTK, were determined in B cells from active and remission GPA patients and HCs. In addition, we also investigated the eff ects of inhibition of

(20)

1

BTK phosphorylation on in vitro B cell function to assess whether BTK blockade could constitute a novel treatment option in GPA.

Currently, several immunosuppressive drugs are used to treat GPA patients. AZA and MMF are two drugs frequently used to maintain remission in AAV patients. Interestingly, MMF-treated patients have been reported to relapse more frequently than AZA-treated patients which may be due to differential effects of these drugs on immune cell function. In Chapter 5, we hypothesized that MMF, in contrast to AZA, inhibits the regulatory function of B cells, and that this might be the underlying mechanism for increased relapse rates in MMF-treated GPA patients. To this end, ex vivo and in vitro effects of AZA and MMF on B cell phenotype and function were studied.

After remission induction in GPA, immunosuppressive drugs are tapered. During tapering or ceasing immunosuppressive therapies, the majority (±60%) of GPA patients experience a disease relapse. Relapses significantly increase morbidity and contribute to mortality in AAV. So far, some markers predicting future disease relapses have been identified, however, none proved to be sufficiently reliable to identify patients at risk for future disease relapse. Since B cells are important in the AAV pathogenesis, both as Ab- and cytokine producing cells, we hypothesized that plasmablasts, i.e. the precursors of (ANCA-producing) plasma cells, are related to future disease relapses. The research described in Chapter 6 assessed frequencies of several B cell subsets, including plasmablasts, and differences herein as possible (bio)marker for future disease relapses (Figure 2).

Finally, in Chapter 7, we summarize and discuss the results presented in this thesis. We propose how these findings contribute to the understanding of the GPA pathogenesis and could aid future research on novel treatments and predictors of disease relapses.

(21)

(22)

Cellular Immune Regulation in

the Pathogenesis of

ANCA-associated Vasculitides

2

a_{Department of Internal Medicine, Division of Nephrology,}

University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, the

Netherlands b_{Department of Rheumatology and Clinical}

Immunology, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen,

the Netherlands c_{Department of Pathology and Medical}

Biology, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, the

Netherlands Published:

Autoimmunity Reviews 2018, 17(4): 413-421

Anouk von Borstela_{, Jan Stephan Sanders}a_{, Abraham}

Rutgersb_{, Coen A. Stegeman}a_{, Peter Heeringa}c_and

Wayel H. Abdulahadb,c

(23)

22 | Chapter 2

Abstract

Anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides (AAV) are systemic autoimmune diseases characterized by necrotizing inflammation of small- to medium-sized blood vessels, affecting primarily the lungs and kidneys. Both animal and human studies show that the balance between inflammatory- and regulatory T- and B cells determines the AAV disease pathogenesis. Recent evidence shows malfunctioning of the regulatory lymphocyte compartment in AAV. In this review we summarize the immune regulatory properties of both T- and B cells in patients with AAV and discuss how aberrations herein might contribute to the disease pathogenesis.

Key Messages

• In AAV the suppressive capacity of the immune system is diminished by either a functional or numerical aberrancy

• The underlying mechanisms mediating decreased suppressive function remain to be elucidated

(24)

2

Cellular Immune Regulation in AAV | 23

Introduction

Anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides (AAV) are characterized by systemic inflammation of small- to medium-sized blood vessels69_{. A}

hallmark of AAV is the presence of ANCA70_{, predominantly directed against the enzymes}

proteinase 3 (PR3)71_{and myeloperoxidase (MPO)}72_{. Both enzymes are stored in both}

primary granules of neutrophils and lysosomes of monocytes and are released into the extracellular environment upon cell activation3_{. Based on clinical and histopathological}

features, AAV is divided into three disease categories: granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA) and eosinophilic GPA (EGPA)2_{. GPA is characterized}

by pauci-immune necrotizing granulomatous inflammation, particularly in the upper airways and the kidneys73,74_{. In MPA necrotizing vasculitis is also common, however,}

without granulomatous inflammation74_{. ANCA are most commonly directed against}

PR34_{and MPO}5_{in GPA and MPA respectively. In contrast to GPA and MPA, EGPA is, in}

addition to small vessel vasculitis, typically characterized by eosinophilia and asthma and is an infrequent AAV disease type2_{. Because EGPA is histopathologically and}

clinically different from GPA and MPA, this disease subtype will not be further discussed in this review.

Although the pathogenesis of AAV has not been fully elucidated, the pathogenic potential of ANCA in the early effector phase of the disease is fairly well established (reviewed in 75_{). The generally accepted mechanism includes exposure of neutrophils to}

pro-inflammatory cytokines (e.g. interleukin (IL)-1β and tumor necrosis factor (TNF)-α) causing translocation of the ANCA-antigens (Ags) to the cell surface. Following binding of ANCA, neutrophils are fully activated, resulting in the production of reactive oxygen species and release of their granular contents, which include the ANCA-Ags PR3 and MPO. More recently, an important role for alternative complement pathway activation in the disease induction has been established and C5a/C5a receptor interactions have been demonstrated to be important primers of neutrophils for activation by ANCA (reviewed in 76_).

AAV are autoimmune diseases implicating a breach of tolerance to self, but why autoimmunity to ANCA-Ags develops and why the immune system fails to suppress autoreactive lymphocytes in these patients is unknown. In homeostasis, our immune system is well balanced and regulatory mechanisms exist to dampen effector immune responses to limit tissue damage and maintain tolerance. Autoreactivity usually does not lead to autoimmune disease since regulatory cells efficiently suppress the immune response against autoAgs. Failure of immune suppression due to defects of regulatory immune cells might thus cause a break in self-tolerance, leading to autoimmune disease. Consequently, resolution of chronic inflammation often requires immunosuppressive medication in order to regain control over the autoreactive response55_{. Two immune}

regulatory cell types essential for peripheral tolerance are regulatory T cells (Tregs) and

(25)

24 | Chapter 2

regulatory B cells (Bregs). Tregs were first discovered by Sakaguchi and co-workers, who showed that this T cell subset was critical in preventing autoimmunity77_{. More recently,}

immunosuppressive capacity has been demonstrated in the B cell lineage as well and the existence of so-called Bregs has been proposed78_.

In recent years, evidence has emerged that the delicate balance between immune effector responses and immune regulation in AAV is disturbed and malfunctioning of immune regulatory cells in AAV patients has been described. In this review, we will discuss the current knowledge on the role of Tregs and Bregs in the AAV pathogenesis and how aberrations in their immune regulatory properties might contribute to disease progression.

T cell Involvement in AAV

AAV is generally considered an autoantibody (autoAb)-mediated disease due to the central role of ANCA in the effector phase of the disease. However, this does not exclude an important pathogenic role for T cell responses in AAV as well. The onset of AAV is postulated to arise from repeated exposure to an (super)Ag, resulting in persistent T cell activation. Malfunctioning of the Treg compartment results in failure to suppress the activation of autoAg-specific T cells79_{. Indeed, activated T cells can be readily}

detected in inflammatory lesions in the affected organs of AAV patients, particularly in the granulomas, which are a characteristic histopathological feature of GPA. Moreover, in AAV patients alterations in circulating T cell subsets and increased levels of soluble factors indicative of T cell activation in the serum have been described36,80,81_{. For}

example, increased levels of soluble T cell activation markers such as sCD25 and sCD30 can be detected in the plasma of GPA patients when compared to healthy controls (HCs)36_{. Also, increased numbers of persistently activated T cells (i.e. T cells with high}

HLA-DR expression) are present in AAV patients and are positively associated with disease severity80,81_{. This strongly suggests that the T cell compartment in AAV is in a}

persistently activated state.

In GPA, the contribution of CD4+_{T helper (Th) cells to the immune pathology might}

depend on whether the disease is localized or generalized. Initially, research in GPA patients focused on the disturbed balance between Th1 and Th2 cells. Studies revealed that CD4+_{T cells of GPA patients with active disease displayed a profound}

Th1 cytokine profile, characterized by increased interferon (IFN) γ production but normal IL-4 production82,83_{. Subsequently it was found that Th1-associated markers in}

the circulation and in nasal granulomatous lesions dominate in patients with localized disease, while Th2-associated markers are more prominent in patients with generalized disease84,85_{. More recently, it was demonstrated that in GPA patients Ag-specific Th17}

(26)

2

GPA patients elevated serum IL-17 levels have been found20_{, which remained elevated}

when patients entered remission86_{. Moreover, in vitro stimulation of PBMCs from GPA}

patients with PR3 resulted in increased IL-17 and RORγT (i.e. transcription factor (TF) of Th17 cells) expression20_.

In summary, in AAV persistent T cell activation, especially of Th17 cells, is well established and likely contributes to the autoimmune pathogenesis of AAV. However, the mechanisms that initiate and maintain T cell activation in AAV remain unclear. Interestingly, emerging evidence indicates that in AAV Tregs do not function properly, suggesting that disturbed immune regulation is an important factor in persistent T cell activation in AAV. Following a brief introduction on Tregs and their mechanisms of action, the existing evidence for Treg malfunctioning in AAV will be discussed.

Regulatory T cells: Phenotype and Mechanism of Action

The immunosuppressive ability of T cells was first discovered >40 years ago87,88_{. However}

not until 1995, a subset of circulating CD4+_{Th cells characterized by the expression of the}

IL-2 receptor α-chain (i.e. CD25) was identified and shown to be critical for the prevention of autoimmunity77_{. Depletion of CD4}+_CD25+_{T cells in wild type (WT) mice and transfer}

of the remaining T cells into syngeneic athymic nude mice induced autoimmunity and multi-organ injury77_{. Interestingly, transfer of the CD4}+_CD25+_{T cell subset into the same}

mice prevented the development of autoimmunity. Later, similar CD4+_{suppressor T cells}

or Tregs were also found in humans. In co-cultures of CD4+_CD25hi_{T cells with CD4}+_CD25

-responder T cells, it was shown that the CD4+_CD25hi_{T cell subset was potent in directly}

suppressing proliferation of and cytokine production by CD4+_CD25-_{responder T cells}89_.

These results provided the first evidence for the existence of human Tregs and identified a phenotype whereby these cells can be detected.

Subsequent research identified a key marker expressed in Tregs, which distinguishes them from other CD4+_{T cells: the TF FoxP3. FoxP3 encodes the scurfin protein and was}

first detected in the scurfy mouse strain. These mice have X-linked recessive mutations in the FoxP3 gene, leading to mortality of hemizygous male mice at the age of 16-25 days. These hemizygous male mice show excess lymphoproliferation combined with infiltration of CD4+_{T cells in multiple organs that overproduce cytokines}90_{. A}

mutated form of FoxP3 was later also described in patients with X-linked syndrome (i.e. IPEX), a disease characterized by immunodysregulation, polyendocrinopathy, and enteropathy91–93_{. Additional research in both scurfy mice and IPEX patients showed}

that CD4+_CD25+_{Tregs were absent. Interestingly, retroviral transduction of FoxP3 into}

naïve CD25-_CD4+_{T cells induced T cells that phenotypically and functionally resembled}

potent Tregs94_{. These results indicate that FoxP3 is necessary for both Treg development}

and function. According to the expression level of FoxP3 in naïve and memory CD4+

Th cells, different Treg subsets have been described by Miyara and co-workers95_{. In}

(27)

26 | Chapter 2

total, three subpopulations of FoxP3+_CD4+_{T cells in the peripheral blood of humans}

were identified: CD45RA+_FoxP3lo_{resting Tregs (rTregs), CD45RA}-_FoxP3hi_{activated Tregs}

(aTregs) and IFNγ- and IL-2-secreting Tregs that are CD45RA-_FoxP3lo_CD25+_{. The first two}

Treg subpopulations (i.e. rTregs and aTregs) had potent suppressive function in vitro, whereas cytokine-secreting CD45RA-_FoxP3lo_CD25+_{Tregs did not suppress effector}

T cells95_{. This indicates that expression of FoxP3 alone is insufficient to identify truly}

suppressive Tregs. Accordingly, different isoforms of FoxP3 have been investigated in human Tregs96_{. The first isoform represents the full-length isoform, and the second}

represents an isoform lacking the exon 2. It has been shown that different FoxP3-isoforms impact Treg function and lineage commitments. More specifically, the full-length FoxP3 (FoxP3fl), but not the isoforms lacking exon 2 (FoxP3D2), interact with the Th17 TF RORγt and inhibits the expression of genes that define Th17 cells97–99_.

Interestingly, a reciprocal relationship in the development of Tregs and Th17 cells has been described100–102_{. In line with this, it has been demonstrated that, under neutral}

conditions in vitro, TGFβ can shift the balance towards functional FoxP3+_{Tregs, whereas}

in the context of an inflammatory cytokine milieu (i.e. IL-1β, IL-2, IL-15) functional Tregs convert into IL-17-producing, non-functional Tregs. Thus, assessment of FoxP3-isoforms in combination with the expression of CD25 and CD45RA are essential to delineate the truly suppressive Tregs.

The exact mechanisms by which Tregs regulate immune responses are not fully understood. It is assumed that these cells use different mechanisms to exert their function, namely secretion of anti-inflammatory cytokines, cytolysis, metabolic disruption and modulation of dendritic cell (DC) maturation and/or function (reviewed in 103_{) (Figure 1). Tregs are potent producers of anti-inflammatory cytokines such as IL-10,}

IL-35 and TGFβ, all of which can suppress responses of activated immune cells (reviewed in 103_{). Cytolysis is a mechanism in which Tregs release enzymes that induce apoptosis}

of effector cells (e.g. granzyme A or B). Tregs are equipped with a third mechanism, namely metabolic disruption. An example of metabolic disruption by Tregs is IL-2 deprivation that induces cell death in effector cells. Another example is that Tregs can inhibit effector cell function via transfer of soluble mediators (e.g. cyclic AMP) via gap junctions into target cells (e.g. effector T cells and DCs)103_{. In target cells, cAMP induces}

inducible cAMP early repressor (ICER), which is thought to inhibit cell proliferation and IL-2 production104_{. Lastly, Tregs can modulate DC maturation and function via physical}

interaction through inhibitory surface molecules such as lymphocyte activation gene (LAG) 3 and cytotoxic T lymphocyte-associated Ag (CTLA)-4. When DCs sense these molecules, production of indoleamine 2,3-dioxygenase (IDO) is initiated, which in turn suppresses other immune cells103_{. The mechanism by which Tregs initiate suppression is}

dependent on the Ag dose. A high Ag dose elicits potent inhibition via induction of FAS-mediated apoptosis or T cell receptor (TCR) ligation, which both result in T cell anergy,

(28)

2

whereas a low Ag dose only induces secretion of IL-10 or TGF-β105_{. Tregs can prevent}

autoimmunity by inhibiting the activation of T effector cells via both inhibitory co-stimulation and secretion of anti-inflammatory cytokines (reviewed in 106_{). In conclusion,}

Tregs employ multiple mechanisms to exert their immune suppressive effect and play a pivotal role in the suppression of autoimmune responses.

Figure 1. Overview of immune suppressive mechanisms used by Tregs and Bregs. Bregs and Tregs inhibit autoreactive cells and Th17 cells via production of anti-inflammatory cytokines (e.g. IL-10, IL-35, TGFβ) and induce apoptosis of autoreactive cells via production of granzymes. Bregs also exert suppressive function via direct cell contact mediated via CD40-CD40L and CD80/86-CTLA-4 interaction. Tregs can suppress autoreactive cells via CTLA4-CD80/86-mediated

induction of IDO, which is an immune suppressive molecule produced by DCs. In addition, Tregs

can indirectly inhibit activated autoreactive cells via IL-2 deprivation or via transfer of soluble mediators (e.g. cyclic AMP) through gap junctions. The suppressive capacity of Treg cells can be enhanced by IL-10 secreted by Bregs. In AAV, aberrancies in both Tregs and Bregs contribute to the disease pathogenesis. The suppressive function of Tregs in AAV seems to be diminished. Due to decreased (functional) IL-2R expression, less IL-2 can be bound mediating decreased suppressive cytokine production (i.e. IL-10, TGFβ and IL-35) that inhibit autoreactive cells. Moreover, the relative ratio of the full length FoxP3 isoform (FoxP3fl) over the FoxP3 isoforms lacking the exon 2 (FoxP3D2) may impact Treg function and lineage commitments. Lower ratio of FoxP3fl/FoxP3D2 isoforms will favor conversion of Tregs into IL-17-secreting cells, which may contribute to the AAV disease pathogenesis. Furthermore, a lower Breg frequency was found in AAV, which seems so far the most important reason for decreased suppressive function by Bregs.

(29)

28 | Chapter 2

Regulatory T cells in AAV

In AAV defective immune suppression of Tregs may contribute to persistent immune cell activation and development of chronic autoimmune inflammation. To date, most research on immune regulation in AAV has focused on the aberrant function and/or altered frequency of Tregs. Although most studies agree on the fact that Treg-mediated immune suppression is impaired in AAV, some controversy exists on whether this defect is caused by numerical or functional changes or both. These aspects will be discussed in more detail in the next paragraphs. Key publications on Tregs in AAV are summarized in Table 1.

Table 1. Key publications on Tregs in AAV.

Authors, Journal & Year Main Conclusions Reference

Marinaki et al.

Clin. Exp. Immunol. (2005) - Persistent state of T cell activation, accompanied by increased T cell proliferation - Decreased IL-10 production of PBMCs after in vitro

stimulation

80

Abdulahad et al. Arthr. Rheum. (2007)

- GPA patients have an expanded frequency of CD4+_CD25hi

Tregs

- CD4+_CD25hi_{Tregs in GPA are functionally defect}

26

Klapa et al.

Clin. Exp. Rheumatol. (2010)

Decreased frequency FoxP3+_{Tregs in GPA.} 107

Morgan et al. Immunology (2010)

Relative deficiency (1) and functional impairment (2) of FoxP3+ _{Tregs in GPA.}

108

Free et al.

Arthr. Rheum. (2013)

CD4+_CD25hi_CD127lo_FoxP3+_{Tregs with the FoxP3D2 isoform}

have diminished suppressive function.

109

Wilde et al. Dis. Markers (2014)

CD122 expression is decreased on CD4+_CD25+_FoxP3+_Tregs

in AAV, leading to decreased suppressive function.

110

Zhao et al.

Rheumatology (2014)

B cell depletion normalizes balance in the T cell compartment.

111

Numerical and Functional Alterations of Tregs in AAV

There is some discrepancy in literature regarding Treg proportions in AAV: Our group previously assessed the frequency of Tregs in peripheral blood of GPA patients in remission and observed that the frequency of CD25hi_FoxP3+_{Tregs was increased in}

AAV patients compared to HCs26_{, as did others}81,109_{. In contrast, others have reported}

a decreased CD4+_CD25hi_{Treg frequency in GPA patients}86,107,108,111_{. In one study, the}

decrease in the CD4+_CD25hi_{Treg frequency was associated with the conventional}

immunosuppressive therapy the patients received (i.e. CYC + pred or MMF/MTX/AZA + pred)111_{. Interestingly, in the same study it was observed that the circulating Treg}

frequency in patients receiving B cell depletion therapy (i.e. Rituximab (RTX)) was similar to HCs, whereas it was significantly increased compared to conventionally treated patients111_{. However, in this study, FoxP3 expression within the CD4}+_CD25hi

(30)

2

T cell compartment was not assessed, which is important since it identifies activated Tregs as demonstrated by Miyara et al.95_{. Another study found that the percentage of}

Tregs significantly expanded during extended remission (≥1 year), independently of immunosuppressive treatment. The increase in Treg frequency was accompanied by an increased Th2 frequency. The authors speculated that decreased Treg numbers in active AAV allow expansion of Th17 cells during remission, whereas immunosuppressive medication inhibits Th17 cells and allows Treg expansion resulting in a shift in Th cell balance86_{. Combined, these results indicate that immunosuppressive medication}

restores the Treg frequency of AAV patients in remission.

Although there is no consensus concerning Treg frequency in AAV, all studies conducted so far, but one112_{, report a reduced suppressive function of Tregs in the}

majority of AAV patients26,107–109_{. Previously, our group showed that despite an increased}

frequency of FoxP3+_{Tregs in GPA patients, their suppressive function was diminished as}

demonstrated by increased proliferation of responder T cells in co-cultures with Tregs from AAV patients93_{, which was subsequently confirmed by others}107–109_.

Together, these results indicate that the discrepancy in the literature with respect to Treg frequencies in AAV patients may, at least in part, be explained by differences in markers or gating strategies used to identify Tregs between the studies. Nevertheless, most studies agree on the fact that Treg function is diminished in AAV.

Possible Underlying Causes for Treg Malfunctioning in AAV

Although most studies demonstrate diminished suppressive function of Tregs in AAV patients, the exact mechanisms underlying impaired Treg function are unknown. The first mechanism that might induce decreased suppression by Tregs is the apparent plasticity of these cells in a pro-inflammatory microenvironment. Evidence exists of Tregs converting into Th1 cells in diabetes patients113,114_{. Two studies involving type 1}

diabetes patients showed that Th1-like Tregs expressed Tbet and CXCR3 besides FoxP3 and produced IFNγ, indicating that these Tregs lost their suppressive function113,114_.

Interestingly, in 2008 it was demonstrated that CD25hi_FoxP3+_{T cells could convert into}

IL-17-producing T cells when stimulated with allogeneic Ag presenting cells (APCs)100_{. This}

observation was later also confirmed by others102,115_{. Subsequent studies in rheumatoid}

arthritis (RA)116_{and cancer}117_{demonstrated that Th17-like Tregs expressing FoxP3 were}

specifically more abundant at inflammatory sites. Also in AAV there seems to be increased skewing of Tregs towards IL-17-producing cells at sites of inflammation118_{, which was}

hypothesized to be due to increased IL-6 and TGFβ production in the inflammatory microenvironment119_{. These findings suggest that the impaired suppressive capacity of}

Tregs could be due to extensive exposure to TGFβ and IL-6 at inflammatory sites. Such a mechanism may be operative in AAV as well, given the reported increase in Th17 levels and diminished Treg function in AAV patients20_{. IL-6 seems to be of great importance as}

(31)

30 | Chapter 2

it was recently demonstrated that Tregs lose their suppressive function, accompanied by decreased Helios expression, when exposed to IL-6120_{. Helios is a TF expressed in}

Tregs that can be induced in vitro by TGFβ stimulation, and has previously shown to support suppressive function of Tregs. Interestingly, in RA patients treatment with Tocilizumab, a monoclonal antibody (Ab) blocking the IL-6-receptor, was associated with an increased frequency of circulating Helios+_FoxP3+_CD4+_{T cells compared to HCs}

or RA patients that did not receive Tocilizumab120_{. Moreover, forced Helios expression}

in murine Tregs enhanced their suppressive function in conjunction with increased expression of Treg markers (i.e. CD103, GITR, GARP, FR4 and IL-10)120_{. Collectively, the}

observations described above suggest that the plasticity of Tregs may contribute to an amplification loop in autoimmune diseases such as AAV. In AAV a pro-inflammatory environment with high levels of IL-6 converts Tregs into IL-17-producing T cells lose their suppressive function and promote ongoing effector cell activation. Moreover, since these non-suppressive Th17-like Tregs do express FoxP3, it again emphasizes the need for additional markers to identify genuine suppressive Tregs.

As a second mechanism, Wilde and colleagues proposed that perhaps the diminished suppressive capacity of Tregs in AAV is related to reduced responsiveness to IL-2. They demonstrated significantly decreased expression of the β-chain of the IL-2 receptor (IL-2R; CD122) on activated Th cells and Tregs in AAV patients compared to HCs110_{. IL-2}

is a cytokine that all T cells need for their functioning under both homeostatic and inflammatory conditions (reviewed in 121_{). If Tregs are less responsive to IL-2, Tregs}

possibly cannot exert their suppressive function.

Aberrant expression in FoxP3–isoforms could be a third explanation for Treg malfunctioning in AAV. As mentioned before, the FoxP3fl, but not FoxP3D2, interacts with RORγt and inhibits the expression of genes that define Th17 cells. Thus, FoxP3D2 may result in a dominant expression of RORγt in Tregs, ensuing production of IL-17 and skewing of Tregs towards pathogenic Th17 cells. Interestingly, the FoxP3D2 isoform in Th cells was increased in both active and inactive AAV patients compared to HCs. In addition, the frequency Tregs with FoxP3fl was decreased in the same patients in comparison to HCs. A positive correlation was found between decreased suppressive capacity of Tregs of AAV patients and the frequency of exon 2-deficient Tregs109_{. Taken}

together, these results indicate that expression of the splice variant of FoxP3 may be responsible for decreased suppressive function.

A fourth explanation focuses on epigenetic aberrancies in Tregs. The CpG motifs in the FoxP3 promoter, also called Treg-specific demethylated region (TSDR), are usually demethylated which allows transcription of the FoxP3 gene122_{. In resting conventional}

T cells the same motifs are only weakly demethylated. Acetylated histones were more strongly associated with the FoxP3 promoter in Tregs compared to conventional T cells122,123_{, which might makes the FoxP3 promoter more accessible to RNA polymerase.}

(32)

2

This would result in transcription of FoxP3 specifically in Tregs and thereby activating their suppressive function. Strikingly, a study confirmed the existence of an inactive Treg population that lost both FoxP3 expression and their suppressive function, which was associated with a demethylated TSDR indicating commitment to the Treg lineage. Intriguingly, FoxP3 expression, and therewith suppressive function, could be reinforced via TCR stimulation in these inactive Tregs124_{. Further studies are necessary to confirm}

whether histones are less acetylated in AAV, causing increased methylation of the FoxP3 promoter and resulting in more inactive Tregs.

In summary, several mechanisms have been proposed to mediate decreased suppressive function of Tregs including IL-6 mediated Treg-Th17 conversion, reduced IL-2 responsiveness, increased expression of FoxP3 splice variants and the FoxP3 promotor methylation status. All of these may be involved in the reported reduced suppressive capacity of Tregs in AAV as well. However, the exact mechanism of their impairment in AAV remains to be determined.

B cell Involvement in AAV

B cells play an important role in the pathogenesis of AAV as they are the precursors of plasma cells that produce ANCA. However, B cells exert multiple other functions including Ag presentation and production of a variety of pro- and anti-inflammatory cytokines. These properties of B cells suggest that these cells may contribute to the pathogenic and immune regulatory processes in AAV in an Ab-independent manner as well40,41_.

Evidence for an Ab-independent pathogenic role of B cells in AAV was first described in 1999. In this study it was shown that the frequency of activated B cells, identified as B cells with high expression levels of CD38, was significantly increased during active disease compared to patients in remission and HCs28_{. Interestingly, no correlation}

between activated B cells and ANCA levels was found, whereas the proportion of activated B cells did significantly correlate with disease activity28_.

Additional evidence for an Ab-independent role of B cells in AAV comes from two major clinical trials that demonstrated that treatment with the B cell-depleting Ab RTX is as efficacious in inducing disease remission in AAV as standard immunosuppressive therapy40,41_{. RTX is a chimeric monoclonal anti-CD20 Ab that depletes B- but not plasma}

cells. Upon RTX treatment, ANCA in the circulation were detectable in patients in remission and only decreased after six months after start of RTX treatment40,41_indicating

that clinical improvement in RTX-treated patients can precede the reduction in autoAb titers. Moreover, relapses did occur in RTX-treated AAV patients, but only after B cell repopulation40_{. This indicates that besides eliminating the precursors of Ab-generating}

cells, the therapeutic effects of RTX also involve modulation of Ab-independent

(33)

32 | Chapter 2

properties of B cells.

The Ab-independent functions of B cells and their potential role in autoimmune disease pathogenesis have gained considerable interest in recent years. In particular, the fact that B cells can produce pro- and anti-inflammatory cytokines that can shape both T cell and innate immune responses and act as drivers or regulators of (auto)immune responses respectively, is increasingly recognized125_{. Moreover, in mice and humans}

specific B cell subsets termed Bregs, have been identified that display immune regulatory properties. Bregs are defined by their capacity to suppress pathological immunity primarily via provision of IL-10 and in mouse models of various autoimmune disorders IL-10-producing B cells have been shown to suppress disease development78,126_{. This has}

led to the concept that in autoimmune disorders such as AAV, the balance between B effector cells and Bregs may be disturbed, driving the pathological autoimmune response.

Regulatory B cells: Phenotype and Mechanism of Action

The theory that B cells are able to suppress immune responses was first postulated by both Katz and Neta in 1974127,128_{. They showed that adoptive transfer of B}

cell-depleted splenocytes were unable to suppress delayed type hypersensitivity in guinea pigs, which suggested that B cells may suppress the activity of T cells in delayed skin hypersensitivity. However, the phenotype and the mechanism of action of these suppressor B cells remained uncharacterized. In 2002, studies in animal models of colitis and experimental autoimmune encephalitis (EAE) showed that a subset of B cells could suppress inflammation and that IL-10 is a hallmark of these suppressive B cells78,129_.

In 2008, Yanaba and co-workers reintroduced the paradigm of suppressor B cells in mice by identifying a subset of peripheral B cells expressing surface CD5 and a high level of CD1d, which were termed Bregs130_{. Bregs were also identified in humans and}

characterized by the production of IL-10 upon in vitro stimulation131,132_{. Since in vitro}

identification of human Bregs via IL-10 is labor intensive (i.e. PBMC isolation and cell culture is needed), there was a clear need for immunophenotypical surface markers that classify Bregs. The search for a specific Breg immunophenotype has so far led to multiple studies that claimed to have identified (different) Breg-specific markers131–134_.

Firstly, Bregs have been identified in the immature or transitional B cell subset131_{. These}

Bregs were characterized by high expression of both CD24and CD38. Flores-Borja et al. showed that this CD24hi_CD38hi_{Breg subset successfully inhibited the differentiation of}

naïve T cells into both Th1 and Th17 cells. Production of IL-10 was mainly found in the CD24hi_CD38hi_{Breg compartment, and these Bregs were potent suppressors of multiple}

activated cell types and played an important role in activating Tregs135_{. Additional}

surface markers have been reported to be expressed by immature Bregs to identify truly suppressive Bregs that include CD1d134_{and CD5}131_{. However, the exact phenotype by}

(34)

2

which suppressor CD24hi_CD38hi _{Bregs can be recognized is not yet elucidated.}

In addition to the immature/transitional Bregs, a second Breg subset was identified within the CD27+_{memory B cell compartment}132_{. These memory CD24}hi_CD27+_Bregs

were also enriched for IL-10 production and suppressed cytokine production in CD4+

T cells in vitro132_{. It seems that the immunosuppressive property of Bregs is mainly}

mediated via IL-10 secretion (Figure 1).

In addition to IL-10, studies have shown that activated Bregs can express other immune regulatory cytokines including IL-35 and TGFβ. Studies in EAE mice showed that B cells lacking IL-35 expression mediated exacerbated EAE and these mice lost the ability to recover from the disease136_{. Another study demonstrated that adoptive transfer of}

IL-35-producing Bregs inhibited experimental uveitis137_{. It has also been shown that B cells}

down-regulate pathogenic autoimmune responses via expression of TGFβ138_{. Treatment}

of prediabetic non-obese diabetic mice with activated TGFβ-secreting Bregs inhibited spontaneous T cell autoreactivity to β cell autoAgs, enhanced mononuclear cell apoptosis in the peripheral lymphoid tissue, and temporarily impaired the function of APCs. TGFβ-secreting Bregs have also been identified in humans, where this subset was characterized by high expression of CD25, a signature marker of Tregs. TGFβ produced by these Bregs promoted FoxP3 and CTLA-4 expression in T cells139_.

All the aforementioned findings demonstrate that Bregs mediate their suppressive effect through the expression of immune regulatory cytokines. In addition to expression of immunomodulatory cytokines, a new cytotoxic marker in Bregs was identified: granzyme B(GzmB) production133_{. GzmB production by these Bregs seems to be an}

additional mechanism of Breg-mediated suppression to exert their cytotoxic activity. Besides the secretion of cytokines and cytotoxic proteins, Bregs need direct cell-cell interaction to suppress other immune cells. It was shown that direct cell-cell contact of Bregs with T cells, mediated via CD40-CD40L interaction, was a crucial pathway for IL-10 induction in B cells and also induced CTLA-4 and FoxP3 expression in T cells78_.

Interestingly, IL-10 production by Bregs also impacts the suppressive function by Tregs. It has been shown that Tregs lacking the IL-10 receptor do not suppress activated Th17 cells140_{, indicating that IL-10 signaling in Tregs is crucial for maintaining certain}

immunosuppressive functions of this subset. This could be an additional mechanism of Bregs by which they impact the suppressive function of Tregs.

Altogether, Bregs play an important role in suppressing immune responses and multiple mechanisms of suppression have been demonstrated. Besides IL-10, Bregs can exert their suppressive function via TGFβ- and IL-35 secretion and the expression of cytotoxic proteins. Multiple studies clearly show that Bregs are needed for both the suppression of activated immune cells and induction of differentiation of naïve cells into regulatory cells. However, more research is needed to clarify the exact underlying mechanism for each Breg subset.

(35)

34 | Chapter 2

Regulatory B cells in AAV

Studies involving Bregs in autoimmune diseases have focused predominantly on abnormalities in their numbers (based on their surface markers) and/or function (based on in vitro IL-10 expression and suppression of immune cells in co-culture). In this section we will review studies on Bregs in AAV. Key publications on B cell suppression in AAV are summarized in Table 2.

Table 2. Key publications on Bregs in AAV.

Authors, Journal & Year Patients Breg subset studied Main conclusion Reference

O’Dell Bunch et al.

Ann. Rheum. Dis. (2013) PR3- & MPO-AAV CD5

+_{B cells} _{A decreased CD5}+_{B cell}

frequency after B cell depletion shortens the time to relapse in AAV.

141

Wilde et al.

Ann. Rheum. Dis. (2013) PR3- & MPO-AAV IL-10

+_{B cells} _{IL-10 producing B cells are}

diminished in AAV.

142

Lepse et al.

Rheumatology (2014) PR3- & MPO-AAV CD24

hi_CD38hi_&

CD24hi_CD27+_Bregs In AAV, the B cell balance _{is altered, but B cell}

function is not impaired reflected as the inhibition of TNFα production from monocytes.

46

Todd et al.

Rheumatology (2014) GPA & MPA CD24

hi_CD38hi_Bregs _{Regulatory B cells are}

numerically but not functionally deficient in AAV.

45

Aybar et al.

Clin. Exp. Immunol. (2015) PR3- & MPO-AAV CD5

+_CD24hi_CD38hi

Bregs Reduced CD5+_CD24hi_CD38hi_and

IL-10+_{regulatory B cells}

in active AAV permit increased circulating autoAbs.

47

Unizony et al.

Arthr. Rheum. (2015) AAV CD5

+_{B cells} _{The percentage CD5}+_B

cells correlates inversely with disease activity after RTX therapy.

143

Using the phenotypical markers CD24 and CD38 to distinguish circulating Bregs, contradictory results were found in patients with AAV. The study by Todd and coworkers demonstrated a significant reduction in the frequency of circulating immature CD24hi_CD38hi_{Bregs in both MPO- and PR3-AAV patients with quiescent disease, whereas}

during active disease only in PR3-AAV patients a decreased Breg frequency was found45_.

Another study by Aybar and coworkers showed that the frequency of circulating CD24hi_CD38hi _{Bregs, additionally defined by CD5, was decreased during active AAV and}

normalized during disease remission47_{. In accordance we found a decreased frequency}

of circulating CD24hi_CD38hi_{Bregs in GPA patients with active disease, whereas no}

difference was observed in quiescent patients in comparison to HCs46_{. In addition we}

University of Groningen B cells in ANCA-associated vasculitides von Borstel, Anouk