Maturation, Reactivity and Therapy Induced Changes of Memory T-Cells

Maturation, reactivity and therapy induced changes of memory T cells in rheumatic autoimmune diseases

Ruth Fritsch-Stork

Maturation, reactivity and therapy induced changes of memory T cells in rheumatic autoimmune diseases

Ontwikkeling, reactiviteit en therapie geïnduceerde veranderingen van memory T cellen in reumatische auto-immuunziekten. (met een samenvatting in het Nederlands)

Proefschrift

ter verkrijging van de graad doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op dinsdag 29 april 2014 des morgens te 10:30.

door

Ruth Fritsch-Stork

Geboren op 21 Augustus 1968 te Wenen Promotoren: prof. dr. J.W.J. Bijlsma prof. dr. F.P.J.G. Lafeber

Contents

Chapter 1 General introduction 1

Chapter 2 Stepwise differentiation of CD4 memory T cells defined by 19 expression of CCR7 and CD27

Chapter 3 Abnormal differentiation of memory T cells in systemic lupus 29 erythematosus

Chapter 4 Characterisation of cellular and humoral autoimmune responses to 49 histone H1 and core histones in human systemic lupus erythematosus

Chapter 5 The spliceosomal autoantigen heterogeneous nuclear 57 ribonucleoprotein A2 (hnRNP-A2) is a major T cell autoantigen in patients with systemic lupus erythematosus

Chapter 6 Characterization of autoreactive T cells to the autoantigens 73 heterogeneous nuclear ribonucleoprotein A2 (RA33) and filaggrin in patients with rheumatoid arthritis

Chapter 7 Potential predictors of glucocorticoid-response in rheumatoid 83 arthritis. A role for endoplasmic reticulum aminopeptidase 2 (ERAP2) and leucocyte-specific transcript 1 (LST1)?

Chapter 8 Glucocorticoid-responsiveness correlates with an interferon 97 signature in rheumatoid arthritis patients

Chapter 9 Summary and general discussion 135

Nederlandse samenvatting 157

Dankwoord / Acknowledgements 165

List of publications 171

Curriculum vitae 177

Chapter 1

General introduction

1 Immunological memory and CD4+ or CD8+ cells acquire altered capacity memory T cells: to produce cytokines, different migratory capability, and diminished activation requirements that allow them to generate an In the description of the Athenian plague by effective secondary response on rechallenge Thucydides in his “History of the with their specific Ag; in short: they become Peloponnesian War” the concept of memory cells, and are identified by the immunological memory entered the literary expression of CD45RO. world. (1) “But those that were recovered

had much compassion both on them that died As memory cells are crucial in our defense and on them that lay sick, as having both against pathogens, the exact mechanisms of known the misery themselves and now no lymphocyte maturation and differentiation more subject to the danger. For this disease from naïve to terminally differentiated never took any man the second time so as to memory cells need to be fully understood. be mortal. And these men were both by Our perception of this differentiation others counted happy, and they also pathway has been influenced by a model, themselves, through excess of present joy, which distinguishes central memory T cells conceived a kind of light hope never to die of (T ) from effector memory T cells (T ) by any other sickness hereafter.“ Approximately CM EM the phenotypic expression of the cysteine 2500 years later the pathogen is still chemokine receptor (CCR) CCR7. (3; figure unknown. Some scholars suspect typhus to 1) The ligands of CCR7 are CCL19 (also be the reason for the epidemic, killing known as Epstein Barr Virus-induced approximately a third of the Athenians, molecule 1 ligand) and CCL21 (also called including their leader Pericles. (2) Similar to secondary lymphoid tissue chemokine). the characterization of the pathogen, These markers are predominantly expressed scientific comprehension of man’s by high endothelial venules (HEV) of immunological memory has deepened, and secondary lymphoid organs and parenchymal our insight is still expanding. cells of the T cell zones of lymph nodes.

Through the interaction of CCR7 on T - Immunological memory is conferred by the CM cells and the respective ligands, the homing adaptive immune system, in which T cells of T cells to secondary lymphoid organs is play a central role. According to the CM achieved. (4-6) Here they can encounter their expression of glycoproteins on their cell- specific antigen, presented by antigen surface, T cells are divided into the CD4+ or presenting cells such as dendritic cells (DC). to the CD8+ lineage. Whereas the main Notably, T only show a low capacity to function of CD8+ cytotoxic T cells is the CM secrete cytokines. In contrast, CCR7- elimination of virally infected or tumorous negative T display effector functions, such cells, CD4+ helper T cells assist several other EM as vigorous cytokine production or immune cells in their tasks, among which cytotoxicity, which they can exert after their antibody (Ab)-production by B cells. T cells migration to inflamed peripheral tissue. (7) In originate in the bone marrow and mature in line with this model, an enrichment of T the thymus, where their antigen (Ag)- CM was demonstrated in tonsils, whereas most T- specificity as well as lineage commitment to cells isolated from nonlymphoid tissue such CD4+ or CD8+ T cells are determined. as the lung, skin, or intestine lacked the Thereafter they re-enter the circulation as expression of CCR7 and thus were thought to recent thymic emigrants (RTE) and be T . (8) eventually mature into naïve CD4+ or CD8+ EM

T cells in the periphery. Naïve T cells are This view, however was challenged by a commonly characterized by the cell surface study showing the majority of cytokine- expression of CD45RA. After the encounter secreting human T cells to express CCR7 and of their respective specific antigen, naïve

2 thus being capable of lymphoid organ Although the division into TCM and TEM has homing. (9) Nevertheless, these authors did advanced our knowledge of lymphocyte find a higher frequency of polarized (T maturational steps, each of these subsets still helper) Th1 or Th2 cells in the CCR7- contain a mixture of different T cells. As T- negative TEM subset. Additionally, the cells, in particular CD4+ T-helper cells, are presence of “pre-Th1” and “pre-Th2” cells in able to produce an array of cytokines (i.e. the TCM compartment secreting low amounts bioactive proteins), by which they influence of interferon gamma (IFN-γ) or interleukin different parts of the immune systems, they (IL-)4, with the potential to differentiate into can also be divided according to the “key” Th1 and Th2 effector memory T cells by cytokines they produce. antigenic stimulation or in response to homeostatic cytokines, were described. (10)

Figure 1: The maturation of T cells. T cells originate in the thymus, where their lineage- and antigen- specificity are determined. After migration to the blood they circulate as naïve, CCR7+CD45RA+CD45RO- cells until they encounter their antigen and become CD45RA-CD45RO+ memory T cells. These are distinguished by their expression of CCR7, a chemokine receptor. CCR7 expressing central memory cells (TCM) circulate to secondary lymphoid tissue, whereas CEM migrate to inflamed tissue. The exact relationship between them had not been fully elucidated. (see text)

3 Th1 cells, e.g., produce IFN, and hereby (19) The incidence and prevalence of SLE activate macrophages to fight intracellular varies considerably across the countries and bacteria; Th2 cells in contrast produce IL-4, were found to be the lowest in Finland IL-5 and IL-13 and contribute to the host (prevalence: 28/100,000 persons) and the UK defense against parasites. As both, Th1 and (prevalence: all races 26/100,000 persons), Th2, are implicated in pathological processes and highest in Italy (prevalence: 71/100,000 (e.g. Th1 in some autoimmune diseases and persons) and Spain (prevalence: all races Th2 in allergy), a deviation from the 91/100,000 persons). (20) SLE is programmed differentiation of these cells as characterized by a spectrum of organ therapeutic approach in certain diseases manifestations ranging from musculoskeletal seems desirable. For this purpose, a more and muco-cutaneous symptoms to life- precise knowledge of the steps toward threatening involvement of kidney, brain or terminal differentiation is of importance. lung. The 5-year survival of SLE patients has improved in the past few decades from less In the attempt to find surface markers for than 50% in the 1950s to more than 90% in different developmental stages of T cells, the 1990s. (21,22) phenotypic expression of some costimulatory molecules has been employed. The diagnosis of SLE is based on the One of these is CD27, a member of the tumor presence of clinical symptoms and laboratory necrosis factor receptor (TNFR) family, parameters, which were defined in American which is expressed by naive CD4+ and College of Rheumatology (ACR) criteria. CD8+ T cells as well as by most memory T- (22,23) Although drafted as classification- cells. (11-13) Initial up-regulation of CD27 criteria, they are widely used as diagnostic expression upon T cell receptor (TCR) tools among rheumatologists and achieve a engagement is followed by irreversible loss sensitivity of 89% and a specificity of around of this surface molecule after repeated 90% for the diagnosis of SLE. (24) antigenic stimulation. (14-15) The loss of CD27 correlates with an increase of IL-4 The etiology of SLE is multifactorial, production and the accumulation of these including genetic, hormonal, immunologic, cells has been reported in patients with and environmental influences. However, the chronic allergic diseases or rheumatoid precise mechanisms remain elusive. A arthritis, and ageing. (13,16) prominent role is ascribed to a dys-regulated apoptosis and clearance of apoptotic Although the combination of CCR7 and material. (25) A delay or inability to clear CD27 had been useful in the definition of a apoptotic cells allows the latter to enter the maturational pathway of CD8 T cells, (17,18) stage of secondary necrosis, resulting in the the differentiation steps of CD4 memory T- release of intracellular danger signals, which cells characterized by expression of CCR7 crucially contribute to the breaking of and CD27 had not yet been examined. immune tolerance. (26-28) These danger signals comprise nucleosomes consisting of double-stranded (ds)DNA and histones. Systemic lupus erythematosus (SLE) These particles are in turn recognized by specific auto-antibodies (auto-Abs), which Systemic lupus erythematosus (SLE) is the are a hallmark of the disease. Consequently, prototypical autoimmune disease, which immune-complexes are formed, setting in predominantly affects women in their motion several pro-inflammatory childbearing years. The prevalence and mechanisms as complement activation or severity is dependent on ethnic background interferon (IFN) type I production by and is generally higher in Hispanics, African plasmacytoid dendritic cells (pDC). Americans, Asians, and Indian Americans.

4 Rheumatoid arthritis (RA) increased expression of adhesion molecules and chemokines. Adaptive and innate Rheumatoid arthritis is the most common immune pathways consequently synergize systemic auto-immune disease with a and induce tissue remodelling and damage. prevalence of approximately 0.5 -1% Furthermore, positive feedback loops worldwide and, as many auto-immune between leukocytes, synovial fibroblasts, diseases, affects more women than men. (29) chondrocytes, and osteoclasts, together with The disease is marked by joint inflammation, the molecular products of damage, perpetuate ultimately leading to articular cartilage and the pathological processes of rheumatoid periarticular bone destruction. Additionally, arthritis. (reviewed in 29) extra-articular manifestations affecting skin, lung, hart, and the vascular system can occur, (Memory) T cells in SLE and RA although the frequency of these complications has been reduced by better The original idea that effector memory CD4 treatment options. (30) T cells (TEM) come in two flavours, (36) Th1 The diagnosis of RA relies on the clinical and Th2 cells, has been gradually expanded feature of swollen joints, combined with over the last decade. Additional subsets have serologic changes. (31) These include, apart been identified and characterized by master from inflammation markers, specific regulators (T-bet for Th1, GATA3 for Th2, antibodies like the rheumatoid factor (RF) or RORt for Th17, Bcl6 for TFH (follicular anti-citrullinated proteins-Ab (ACPA). helper cells) and Foxp3 for Treg (regulatory Pathogenesis of rheumatoid arthritis is T cells)). The function of most subsets is multifaceted and involves genetic as well as determined, as described above, by the environmental triggers. Genetic influence is cytokine profile they produce and the hereby implicated by twin studies, with concordance targeted aspects of the immune system. rates of 15 to 30% among monozygotic twins Whereas Th1 cells are implicated in host and 5% among dizygotic twins. (32) defense against intracellular pathogens Susceptibility for RA has long been through the secretion of IFN, Th2 cells associated with a gene locus in the HLA- influence the defense mechanisms against DRB1 region in RF- and ACPA- positive parasites, and mainly secrete IL-4, IL-5 and patients, (33) whereby alleles of HLA-DRB1 IL-13. The more recently defined Th17 cells containing a common amino acid motif, act through the production of their key termed the shared epitope, confer particular cytokines IL-17A, IL-17F and IL-21 and susceptibility. Environmental influences are induce the immunological reaction against diverse and include smoking, infectious extracellular pathogens. (37) Another factors, and hormonal disturbances. Smoking considerably new “kid on the block” is the T and HLA-DRB1 alleles synergistically follicular helper cell (TFH) subset. These cells increase one’s risk of having ACPA, (34) are, due to the secretion of IL-21, and the possibly due to increased posttranslational localization in the germinal centers, crucial modifications of mucosal proteins by for providing B cell help by promoting class- smoking. These are mediated through switching. (38) The described effector T cell peptidyl arginine deiminase, type IV subsets are kept under control by regulatory (PADI4), which is induced by smoking or an T cells (Treg), which are commonly defined altered oral flora, (34,35) and result in by their master regulator Foxp3. (see quantitative or qualitative alteration of extensive review Ref. 39) This classification citrullination of mucosal proteins, creating is far from absolute and far from complete, neo-epitopes. Loss of tolerance to such neo- with new subsets waiting to be identified; epitopes can result in ACPA production. furthermore, a growing body of evidence Arthritis in RA is the result of influx of suggests a high degree of plasticity in those leukocytes into the joint mediated by already defined subsets. (40)

5 (Memory) T cells in SLE Beyond changes in the effector T cells, also the quantity and function of regulatory T The pathognomonic lupus auto-Abs to cells have been investigated with chromatin components show contradictory conclusions. The finding of a immunoglobulin class switching and affinity decrease in number and/or function of Treg maturation, which are typical features of a T in SLE by some authors (56,57) was disputed cell-dependent Ag-driven immune response. by others. (58,59) This discrepancy is partly (41,42) The role of T cells in SLE is further due to the lack of a reliable and specific Treg underlined by several genetic risk loci marker until recently, especially under involving T cell signalling or T cell function conditions of T cell activation as present in such as STAT4, PTPN22 or OX40L, (43-45) SLE. Also the different therapeutic regimen which have been confirmed in several of the investigated patient population could genome wide association studies (GWAS). contribute to the inconsistencies described. (reviewed in 46) Contrasting the interest given to the examination of functional T cell subsets in Several abnormalities in lupus T cells have SLE, not much attention has been devoted to been described ranging from intrinsic the maturational status of memory T cells in peculiarities to a disturbed composition of SLE and its possible implications regarding the T cell compartment. The former are disease pathogenesis. In this respect, an reviewed in Ref. 47 and include the rewiring increase of the T cell-receptor (TCR) in lupus T cells. of memory CD4+ and a decrease of naive This is brought about by the alteration of the CD4+ T cells in patients with longstanding TCR composition, which ultimately leads to SLE have been observed. (60, 61) However, an amplification of early signalling events, the differentiation status of the CD45RO+ and, among other consequences, results in an memory subsets had not been analyzed. impaired IL-2 production by lupus T cells. Thus, how the above mentioned various (48) abnormalities may influence overall T cell function in vivo in SLE has not yet been SLE has traditionally been viewed as a Th2 determined. driven disease, which was suggested by the prominent role of antibodies and the therefore suspected underlying T cell help T cells in RA from Th2 cells. In fact, a higher level of cytokines attributed to Th2 cells, such as IL- Disease specific antibodies as the rheumatoid 4, IL-6 and IL-10 had been detected in lupus factor or ACPA are hallmark of RA, patients (49-51) and a predominance of Th2 comparable to the importance of Ab against cells was seen after stimulation of peripheral dsDNA in SLE. Also these are often class- blood leukocytes. (52) However, also switched and have undergone some affinity contradictory reports have been published, maturation, which is indicative of an showing an increased Th1/Th2 ratio, which underlying T cell process. (62) The role of T- was particularly seen in patients with lupus cells in RA is further substantiated by the nephritis. (53) After the discovery of Th17 as immunogenetic association of RA with the a pro-inflammatory subset of T cells, these shared epitope, but also with minor were also examined in SLE and found to be susceptibility loci (e.g. PTPN22 or CTLA-4) increased. (54,55) Interestingly, also a subset involved in T cell stimulation and of double negative (CD4-CD8-) T cells, differentiation. (63,64) which is rare in healthy controls, was increased in SLE patients and produced IL- Traditionally, RA has been seen as Th1 17. (55) mediated disease, marked by a preponderance of IFN over IL-4 in synovial

6 tissue and fluid, especially in the established directed to single histones, to histone–DNA phases of RA. (65-67) However, with the complexes, or to complexes of histones. (80) discovery of Th17 cells, the latter have In particular, histone H2A–H2B dimers increasingly been recognized as driving frequently elicit humoral anti-histone factor of RA-inflammation. (68-69) The responses, especially in drug-induced lupus cocktail of transforming growth factor erythematosus (LE). (81) Anti-histone Abs (TGF)β and IL-1β, IL-6, IL-21, and IL-23 can be found in up to 75% of SLE sera, while derived from synovial macrophages and DC their prevalence in other rheumatic diseases provides a milieu that supports Th17 usually does not exceed 20%. (82) Like anti- differentiation and suppresses differentiation dsDNA Abs, anti-histone Abs also correlate of regulatory T cells. This leads to a shift in with SLE disease activity, particularly Abs to T cell homeostasis toward inflammation. In histone H1, which appear to be more specific turn, IL-17A, synergizes with TNF-α derived for SLE and may better correlate with disease from macrophages to promote activation of activity than Abs to core histones. (83,84) Of fibroblasts and chondrocytes. This new note, anti-H1 Abs are responsible for the LE- insight has led to the targeting of IL-17 in cell phenomenon, which was originally clinical trials. (70) In contrast to the widely included in the ACR classification criteria for accepted contribution of Th17 to RA- SLE. (85) pathogenesis, data regarding the number of Tregs in the circulation of RA patients In order to elucidate the underlying compared to healthy individuals are pathological mechanisms leading to auto-Ab inconclusive and contradictory. Decreased production, the auto-reactivity of T cells (71-73) as well as increased (74,75) numbers against chromatin structures has been of circulating Treg in RA patients have been investigated as well. reported compared to healthy individuals. In this context, histones are considered the Furthermore, regulatory T cells that are major T cell Ags in SLE for the generation of detected in tissues from patients with RA autoimmune responses against chromatin appear to have limited functional capability. components: thus, histone-specific T cell (76) This imbalance between Th17 and clones augmented the production of anti- regulatory T cells may also reflect local dsDNA, anti-single-stranded (ss)DNA and presence of TNF-α, which has been shown to anti-histone Abs in murine models of SLE. hamper regulatory T cell function, as anti- (86) Furthermore, they promoted anti- TNF therapy enhances suppressive Treg dsDNA production by autologous B cells in function via IL-10 and TGF-. (77) human SLE. (87,88) However, details of the histone-specific T cell response in human SLE have not been fully Auto-reactive T cells in SLE elucidated. In particular, it is not clear which histone constitutes the major auto-Ag in T Antinuclear auto-Abs are a hallmark of SLE, cell activation. and are directed against chromatin components, either single structures thereof The search for new serological markers and or against the complete nucleosome. Among auto-antibodies has led to the identification these auto-Abs, IgG Abs to (double-stranded of a number of additional auto-antigens, DNA) dsDNA are a pathognomonic feature among them the Ro and La proteins, (89) as of SLE and can be directly pathogenic. (78) well as components of the spliceosome. This However, dsDNA may not be the first target large and highly dynamic complex contains of the auto-Ab response; rather, auto-Abs to many evolutionarily conserved proteins, such other chromatin components, and to histones as the U1–70 kD RNP or the Sm proteins, in particular, may precede anti-dsDNA Ab (90, 91) which are recognized by auto-Ab in formation.(79) Anti-histone Abs can be approximately 30% of SLE patients. About

7 20 years ago, the heterogeneous nuclear filaggrin), fibrinogen, fibronectin, collagen, ribonucleoprotein (hnRNP-)A2, another and vimentin. The common denominator of component of the spliceosome, was these antigens are the citrullinated residues characterized as a novel target of auto-Abs in present in all these proteins after deimination systemic autoimmune diseases. (92) by (peptidyl arginine deiminase) PAD4, Although initially described to occur mainly which can be recognized by ACPAs. in patients with RA, auto-Abs to hnRNP-A2 Consequently, ACPAs are highly cross- (originally termed anti-RA33) were later reactive and the original (cellular) target detected also in 20% to 30% of patients with triggering the antibody response has not been SLE as well as in 40% of patients with mixed identified. connective tissue disease, usually in Up to 20% of patients are seronegative to association with other anti-spliceosomal either RF and ACPA, which still warrants the auto-Abs. (93,94) The auto-Ab response search for new auto-Abs and their respective against hnRNP-A2 shows some differences auto-antigen to use as a biomarker for in epitope recognition among patients with diagnosis, disease progression and/or RA, SLE and mixed connective tissue therapeutical response. During the last disease. (95) HnRNP-A2 has a predominant decades this has led to the identification of a nuclear localization and exerts multiple number of candidate proteins including functions, including regulation of alternative collagen, cartilage link protein, and gp39 splicing, transport of mRNA, and regulation (103,104) up to the only recently defined of translation. (96-98) There have been only humoral targets PAD4 and BRAF. (105) few reports about the T cell response to Among those auto-Ag is hnRNP-A2 (106) spliceosomal antigens, mainly focusing on described above as auto-Ag for SLE as well. the T cell reactivity to small nuclear Auto-Abs to hnRNP-A2 can be detected in ribonucleoprotein antigens. (99) HnRNP-A2 about one-third of RA-patients and in 20- specific CD4+ T cell clones (TCCs) derived 30% of SLE patients (see above) and are a from one mixed connective tissue disease marker of early disease. (107) However, as B patient and two SLE patients have been cell-epitopes differ between these diseases, characterised, and were attributed a possible (107) also the underlying T cell reactivity pathogenic role in SLE. (100) However, as may not be the same. Furthermore, the presence of auto-reactive T cells is not autoimmunity to hnRNP-A2 has been limited to patients, but has repeatedly been described in TNF-transgenic mice, (108) observed in healthy individuals as well, which develop arthritis spontaneously, and (87,101,102) the search for possible seems to be the driving event in pristane- differences in auto-antigen specific cellular induced arthritis in rats, (109) underlining the reactivity between patients and healthy importance to define the cellular reactivity of controls can give insight into the hnRNP-A2 in RA. pathogenesis of the respective autoimmune disease. Glucocorticoid resistance in RA-

Auto-reactive T cells in RA patients

As T cells are central in the disease process Auto-antibodies have been a hallmark of RA of several auto-immune diseases, among as well, as marked by the inclusion of RF and which also RA, they often constitute a target ACPA into the new classification criteria. for therapy. Although the armamentarium of (31) Whereas the target of RF is treatment options in RA is diverse, unmistakably the Fc fraction of therapeutical intervention often relies on imunoglobulins, the targets of ACPAs are glucocorticoids (GC), which are known to manifold and include α-enolase, keratin (or influence T cells survival and function. Due

8 to the inhibition of radiographic damage seen AP-1 or STAT5 could contribute to GC- after long-term treatment, (110) GC have resistance in RA. (118) However, no gained recognition as a disease modifying predictors of such GC-resistance in RA- anti-rheumatic drug (DMARD). GC-pulse patients have been established thus far. therapy (≥ 250mg prednisone equivalent) is usually applied in patients with high disease activity because of its fast anti-inflammatory Outline of this thesis: action to bridge the lag time until (conventional) DMARDs deploy their effect. The studies combined in this thesis GC influence 10-20% of all genes, (111, 112) illuminate different aspects of T cells in two which is mediated by the glucocorticoid typical systemic auto-immune diseases, lupus receptor (GR). The GR is a member of the erythematosus (SLE) and Rheumatoid nuclear receptor family, and has several arthritis (RA). isoforms, dependent on alternative splicing, Different aspects, such as the maturational with GR the alpha isoform (GR) and the status (first part of the thesis), the auto- beta isoform (GR being the most important. reactive potential (second part), and the Whereas GR is the biologically active form, differential reaction to therapy with GR is acting as dominant negative inhibitor glucocorticoids (third part) were studied and of GR, affecting the transcriptional activity discussed. of the latter. After binding of the ligand to the GR the GR/GC complex translocates to the nucleus, where it binds to promoter First part: memory T cell subsets in regions of GC-responsive genes. The induced health and disease transcription through binding to positive glucocorticoid response elements (GRE) is In the first part, the maturational pathway of called trans-activation and includes the memory T cells in healthy persons was induction of anti-inflammatory genes (e.g. examined as a starting point. The lipocortin or IB). In contrast, trans- differentiation steps leading to operational repression denominates reduced transcription memory T cells have been the focus of (e.g. of IL-1, IL-8 or TNF), which can be research, and generated the model of central elicited by either binding to negative GRE or and effector memory cells, distinct in their via protein-protein interaction between the expression of CCR7. However, this model is GC/GR complex and transcription factors, not sufficient, as especially the CCR7 such as AP-1, NFB or IRF-3. (113) negative TEM are a heterogeneous population. As in earlier studies the use of the co- Despite this general anti-inflammatory stimulatory protein CD27 had been used activity of GC, about one third of RA successfully to distinguish T cells according patients does not respond to GC adequately. to their maturational state and preferential .(114) GC-resistance has also seen observed cytokine production, the combination of in other inflammatory diseases as systemic these two surface markers seemed a lupus erythematosus and asthma. (115,116) promising approach to elucidate the Several mechanisms underlying GC- maturation of memory T cells in more detail. resistance in general have been described, The maturational relation of the subsets affecting every step in the above described defined by these phenotypical markers and working mechanism. (113) In RA, an their function was examined in Chapter 2. increased expression of the GC-isoform, Several T cell abnormalities are known to blocking the binding of GC-GR to the play a role in SLE; however, only little is GRE, has been reported in GC-resistant known of the composition and associated patients. (117) Also an increased cellular maturational status of the T cell level of transcription factors such as NFB, compartment. Therefore, in Chapter 3 the

9 different memory T cell subsets as found in differential response of T cells to GC, healthy controls, identifying a maturational marked by a different gene expression pathway, were studied in SLE patients. profile. This difference could lead to the identification of biological markers predicting the response to GC before the Second part: auto-reactive T cells in administration of therapy. Accordingly, in SLE and RA Chapter 7 the gene expression profile of T cells and monocytes before start of GC- The second part of the thesis describes the therapy was examined to find such markers. different auto-reactive properties of T cells in An additional aspect of the differential SLE and RA. The hallmark of SLE is the clinical response to GC is a possibly humoral reactivity to chromatin. The underlying differential GC-working underlying cellular reactivity has been mechanisms in GC-Responders compared to examined in mice and men and identified Non-Responders. This was examined in histones as prime antigens for T cells. Chapter 8 by comparing the change in gene However, the contribution of the single expression profile elicited by GC-therapy in histones to this reactivity has not been T cells and monocytes of GC-Responders elucidated; equally unknown was the type of versus Non-Responders. T cells reacting to histones. Thus, in Chapter 4, the T cell target was specified and, by Finally Chapter 9 embodies a summary of generating T cell clones, the nature of this the findings obtained in this thesis and a auto-reactivity studied. general discussion hereof. As antibodies to ribonucleoproteins are also a prominent feature of SLE, in chapter 5 the cellular reactivity to hnRNP-A2, the target of References auto-antibodies in approximately 20% of SLE patients was investigated. Again, the 1) The Peloponnesian War, by goal was not only to define the auto-antigen, Thucydides, the complete Hobbes but also to characterize the possibly translation, with notes and a new pathogenic auto-reactive T cells. introduction by David Grene, HnRNP-A2 is an antigen, which was University of Chicago Press, 1989; originally discovered in RA-patients, with Thucydides on the plague of Athens; approximately one third of these displaying p115-8. auto-Ab against this part of the spliceosome. 2) Gomme AW, edited by A Andrewes Consequently, in Chapter 6 the and KJ Dover.1981. An Historical characterization of the T cell response in RA Commentary on Thucydides, Volume 5. to this auto-antigen was carried out. Book VIII Oxford University Press. 3) Sallusto F, Lenig D, Forster R, Lipp M, Third part: T cells and the response Lanzavecchia A. 1999. Two subsets of to Glucocorticoids in RA memory T lymphocytes with distinct homing and effector functions.

Nature;401:708-12. The third part of this thesis is dedicated to the 4) Baekkevold ES, Yamanaka T, role of T cells in the differential response to Palframan RT, Carlsen HS, Reinholt therapy in RA. Glucocorticoids (GC) are a FP, von Andrian UH, Brandtzaeg P, mainstay of therapy in RA, but Haraldsen G. 2001. The CCR7 ligand approximately one third of the patients do not elc (CCL19) is transcytosed in high respond adequately. As T cells play a role in endothelial venules and mediates T cell RA-pathogenesis and constitute an important recruitment. J Exp Med; 193:1105-12. target of GC-action, the difference in clinical response to GC should be reflected in a

10 5) Tanabe S, Lu Z, Luo Y, Quackenbush on subsets of mature T lymphocytes. J EJ, Berman MA, Collins-Racie LA, Mi Immuno;151:2426-30. S, Reilly C, Lo D, Jacobs KA, Dorf 13) Tortorella C, Schulze-Koops H, ME. 1997. Identification of a new Thomas R, Splawski JB, Davis LS, mouse β-chemokine, thymus-derived Picker, LJ, Lipsky PE. 1995. chemotactic agent 4, with activity on T Expression of CD45RB and CD27 lymphocytes and mesangial cells. J identifies subsets of CD4+ memory T Immunol;159:5671-9. cells with different capacities to induce 6) Gunn MD, Tangemann K, Tam C, B cell differentiation. J Immunol; Cyster JG, Rosen SD, Williams LT. 155:149-62. 1998. A chemokine expressed in 14) Hintzen RQ, Jong J, Lens SM, van Lier lymphoid high endothelial venules RAW. 1994. CD27: marker and promotes the adhesion and chemotaxis mediator of T cell activation. Immunol of naïve T cells. Proc Natl Acad Sc. Today;15:307-11. USA; 95:258-63. 15) De Jong R, Brouwer M, Hooibrink B, 7) Masopust D, Vezys V, Marzo AL, van der Pauw-Kraan T, Miedema F, Lefrancois L. 2001. Preferential van Lier RA. 1992. The CD27− subset localization of effector memory cells in of peripheral blood memory CD4+ nonlymphoid tissue. Science; lymphocytes contains functionally 291:2413-7. differentiated T lymphocytes that 8) Campbell JJ, Murphy KE, Kunkel EJ, develop by persistent antigenic Brightling CE, Soler D, Shen Z, stimulation in vivo. Eur J Immunol; Boisvert J, Greenberg HB, Vierra MA, 22:993-9. Goodman SB, Genovese MC, Wardlaw 16) Nijhuis EW, Remarque EJ, Hinloopen AJ, Butcher EC, Wu L. 2001. CCR7 B, van der Pouw-Kraan T, van Lier Expression and memory T cell RA, Ligthart GJ, Negelkerken L. 1994. diversity in humans. J Immuol; Age related increase in the fraction of 166:877-84. CD27−CD4+ T cells and IL-4 9) Kim CH, Rott L, Kunkel EJ, Genovese production as a feature of CD4+ T cell MC, Andrew DP, Wu L, Butcher EC. differentiation in vivo. Clin Exp 2001. Rules of chemokine receptor Immuno.;96:528-34. association with T cell polarization in 17) Rufer N, Zippelius A, Batard P, Pittet vivo. J Clin Invest;108:1331-9. M, Kurth I, Corthesy P, Cerottini JC, 10) Rivino L, Messi M, Jarrossay D, Leyvraz S, Roosne E, Nabholz M, Lanzavecchia A, Sallusto F, Geginat J. Romero P. 2003. Ex vivo 2004. Chemokine receptor expression characterization of human CD8+ T identifies pre-T helper (Th)1, pre-Th2, subsets with distinctive replicative and nonpolarized cells among human history and partial effector functions. CD4+ central memory T cells. J Exp Blood;102:1779-87. Med;200:725-35. 18) Tomiyama HT, Matsuda M,Takiguchi. 11) van Lier RAW, Borst J, Vroom TM, 2002. Differentiation of human CD8+ T Klein H, van Mourik P, Zeijlemaker cells from a memory to WP, Melief CJ. 1987. Tissue memory/effector phenotype. J distribution and biochemical and Immunol;168:5538-50. functional properties of Tp55 (CD27), 19) Alarcon GS. Of ethnicity, race and a novel T cell differentiation antigen. J lupus. 2001. Lupus;10: 594-6 Immunol;139:1589-96. 20) Danchencko N, Satia JA, Anthony MS. 12) Hintzen RQ, de Jong R, Lens Epidemiology of systemic lupus SM,Brouwer M, Baars P, van Lier RA. erythematosus: a comparison of world- 1993. Regulation of CD27 expression

11 wide disease burden. 2006. Lupus;15: apoptosis with pathogenic implications. 308-18 Apoptosis;13:463–82 21) Kellum RE, Haserick JR, Systemic 28) Matzinger, P. 2002.The danger model: lupus erythematodes. A statistical a renewed sense of self. Science; evaluation of mortality based on a 296:301–5. consecutive series of 299 patients. 29) The pathogenesis of rheumatoid 1964. Arch Intern Med; 113:200-7. arthritis.2011.McInnes IB, Schett G. 22) Cervera R, Khamashta MA, Font J, New Engl J Med.; 365:2205-19. Sebastiani GD, Gil A, Lavilla P, Mejía, 30) Turesson C. 2013.Extra-articular JC, Aydintug AO, Chwalinska- rheumatoid arthritis. Curr Opin Sadowska H, de Ramón E, Fernández- Rheumatol;25:360-6 Nebro A, Galeazzi M, Valen M, 31) Aletaha D, Neogi T, Silman AJ, Mathieu A, Houssiau F, Caro N, Alba Funovits J, Felson DT, Bingham CO P, Ramos-Casals Manuel, Ingelmo M, 3rd, Birnbaum NS, Burmester GR, Hughes GRV; European Working Party Bykerk VP, Cohen MD, Combe B, on Systemic Lupus Erythematosus. Costenbader KH, Dougados M, Emery 2003. Morbidity and mortality in P, Ferraccioli G, Hazes JM, Hobbs K, systemic lupus erythematosus during a Huizinga TW, Kavanaugh A, Kay J, 10-year period: a comparison of early Kvien TK, Laing T, Mease P, Ménard and late manifestations in a cohort of HA, Moreland LW, Naden RL, Pincus 1,000 patients. Medicine (Baltimore); T, Smolen JS, Stanislawska-Biernat E, 82:299-308. Symmons D, Tak PP, Upchurch KS, 23) Tan EM, Cohen AS, Fries JF, Masi Vencovsky J, Wolfe F, Hawker G. AT, McShane DJ, Rothfield NF, 2010. 2010 Rheumatoid arthritis Schaller JG, Talal N, Winchester RJ. classification criteria: an American 1982. The 1982 revised criteria for the College of Rheumatology/ European classification of systemic lupus League Against Rheumatism erythematosus. Arthritis Rheum; collaborative initiative. Ann Rheum 25:1271-7 Dis;69:1580-8. 24) Hochberg MC, for the Diagnostic and 32) MacGregor AJ, Snieder H, Rigby AS, Therapeutic Criteria Committee of the Koskenvuo M, Kaprio J, Aho K, American College of Rheumatology. Silman AJ. 2000. Characterizing the 1997. Updating the American College quantitative genetic contribution to of Rheumatology revised criteria for rheumatoid arthritis using data from the classification of systemic lupus twins. Arthritis Rheum;43:30-7. erythematosus [letter]. Arthritis 33) Gregersen PK, Silver J, Winchester RJ. Rheum;40: 1725 1987. The shared epitope hypothesis: 25) Tutuncu ZN & Kalunian KC in Dubois' an approach to understanding the Lupus Erythematosus (eds Wallace, D. molecular genetics of susceptibility to J. & Hahn, B. H.) 16–20 (Lippincott rheumatoid arthritis. Arthritis Williams & Wilkins, Philadelphia, Rheum;30:1205-13. 2007) 34) Klareskog L, Stolt P, Lundberg K, 26) Muñoz LE, Lauber K, Schiller M, Källberg H, Bengtsson C, Grunewald J, Manfredi AA, Herrmann M. 2010. The Rönnelid J, Harris HE, Ulfgren AK, role of defective clearance of apoptotic Rantapää-Dahlqvist S, Eklund A, cells in systemic autoimmunity. Nat Padyukov L, Alfredsson L. 2006.A Rev Rheumatol;6:280-9 new model for an etiology of 27) Silva TM, do Vale A & dos Santos NM rheumatoid arthritis: smoking may 2008.Secondary necrosis in trigger HLA-DR (shared epitope)- multicellular animals: an outcome of restricted immune reactions to

12 autoantigens modified by citrullination. Gregersen PK.2007. STAT4 and the Arthritis Rheum;54: 38-46. risk of rheumatoid arthritis and 35) Wegner N, Wait R, Sroka A, Eick S, systemic lupus erythematosus N Engl J Nguyen KA, Lundberg K, Kinloch A, Med;357:977-86. Culshaw S, Potempa J, Venables PJ. 44) Orozco G, Sanchez E, Gonzalez-Gay 2010.Peptidylarginine deiminase from MA, Lopez-Nevot MA, Torres B, Caliz Porphyromonas gingivalis citrullinates R, Ortego-Centeno N, Jiménez-Alonso human fibrinogen and α-enolase: J, Pascual-Salcedo D, Balsa A, de implications for autoimmunity in Pablo R, Nuñez-Roldan A, González- rheumatoid arthritis. Arthritis Escribano MF, Martín J. 2005. Rheum;62:2662-72. Association of a functional single- 36) Mosmann TR, Coffman RL.1989. TH1 nucleotide polymorphism of PTPN22, and TH2 cells: different patterns of encoding lymphoid protein lymphokine secretion lead to different phosphatase, with rheumatoid arthritis functional properties. Annu Rev and systemic lupus erythematosus. Immunol;7:145-73. Arthritis Rheum; 52:219–24. 37) Miossec P, Korn T, Kuchroo VK. 45) Cunninghame Graham DS, Graham 2009. Interleukin-17 and type 17 helper RR, Manku H, Wong AK, Whittaker T cells. N Engl J Med;361:888-898 JC, Gaffney PM, Moser KL, Rioux JD, 38) Crotty S. 2011. Follicular helper CD4 Altshuler D, Behrens TW, Vyse TJ. T cells (TFH). Annu Rev Immunol; 2008.Polymorphism at the TNF 29:621-63 superfamily gene TNFSF4 confers 39) Sakaguchi S, Yamaguchi T, Nomura T, susceptibility to systemic lupus Ono M.2008. Regulatory T cells and erythematosus. Nat Genet;40:83-9. immune tolerance. Cell. 2008 May 46) Cui Y, Sheng Y, Zhang X. 2013. 30;133:775-87. Genetic susceptibility to SLE: recent 40) Nakayamada S, Takahashi H, Kanno progress from GWAS. J Autoimmun; Y, O'Shea JJ.2012. Helper T cell 41:25-33. diversity and plasticity. Curr Opin 47) Crispin JC, Liossis SN, Kis-Toth K, Immunol;2012;24:297-302. Lieberman LA, Kyttaris VC, Juang YT, 41) van Es JH, Gmelig Meyling FH, van de Tsokos GC. 2010. Pathogenesis of Akker WR, Aanstoot H, Derksen RH, human systemic lupus erythematosus: Logtenberg T. 1991. Somatic recent advances. Trends Mol mutations in the variable regions of a Med;16:47-57. human IgG anti-doublestranded DNA 48) Katsiari CG, Kyttaris VC, Juang YT, autoantibody suggest a role for antigen Tsokos GC. 2005. Protein phosphatase in the induction of systemic lupus 2A is a negative regulator of IL-2 erythematosus. J Exp Med;173:461–70. production in patients with systemic 42) Winkler TH, Fehr H, Kalden JR. lupus erythematosus. J Clin Analysis of immunoglobulin variable Invest;115:3193–204. region genes from human IgG anti- 49) Ogawa N, Itoh M, Goto Y. 1992. DNA hybridomas. 1992.Eur J Immuno Abnormal production of B cell growth ;22:1719–28. factor in patients with systemic lupus 43) Remmers EF, Plenge RM, Lee AT, erythematosus. Clin Exp Graham RR, Hom G, Behrens TW, de Immunol;89:26–31. Bakker PI, Le JM, Lee HS, Batliwalla 50) Linker-Israeli M, Deans RJ, Wallace F, Li W, Masters SL, Booty MG, DJ, Prehn J, Ozeri-Chen T, Klinberg Carulli JP, Padyukov L, Alfredsson L, JR. 1991. Elevated levels of Klareskog L, Chen WV, Amos CI, endogenous IL-6 in systemic lupus Criswell LA, Seldin MF, Kastner DL,

13 erythematosus: a putative role in CD25high T cell population in patients pathogenesis. J Immunol;147:117–23. with systemic lupus erythematosus 51) Llorente L, Richaud-Patin Y, Wijdenes treated with glucocorticoids. Ann J, Alcocer-Varela J, Maillot M, Rheum Dis;65:1512–7. Durand-Gasselin I, Fourrier BM, 59) Zhang B, Zhang X, Tang FL, Zhu LP, Galanaud P, Emilie D. 1993. Liu Y, Lipsky PE. 2008. Clinical Spontaneous production of interleukin- significance of increased CD4+CD25- 10 by B lymphocytes and monocytes in Foxp3+ T cells in patients with new- systemic lupus erythematosus. Eur onset systemic lupus erythematosus. Cytokine Netw;4:421–30. Ann Rheum Dis; 67:1037–40. 52) M. Funauchi, S. Ikoma, H. Enomoto, 60) Gordon C, Matthews N, Schlesinger and A. Horiuchi. 1998. Decreased Th1- BC, Akbar AN, Bacon PA, Emery P, like and increased Th2-like cells in Salmon M. 1996. Active systemic systemic lupus erythematosus. Scand J lupus erythematosus is associated with Rheumatology; 27: 219–24. the recruitment of naive/resting T cells. 53) Akahoshi M, Nakashima H, Tanaka Y, Br J Rheumatol;35:226–30. Kohsaka T, Nagano S, Ohgami E, 61) Neidhart M, Pataki F, Michel BA, Fehr Arinobu Y, Yamaoka K, Niiro H, K. CD45 isoforms expression on CD4+ Shinozaki M, Hirakata H, Horiuchi T, and CD8+ peripheral blood T- Otsuka T, Niho Y. 1999. Th1/Th2 lymphocytes is related to auto-immune balance of peripheral T helper cells in processes and hematological systemic lupus erythematosus. Arthritis manifestations in systemic lupus Rheum;42:1644-8. erythematosus. Schweiz Med 54) Shah K, Lee WW, Lee SH, Kim SH, Wochenschr;126:1922–5. Kang SW, Craft J, Kang I.2010. 62) Suwannalai P, van de Stadt LA, Radner Dysregulated balance of Th17 and Th1 H, Steiner G, El-Gabalawy HS, Zijde cells in systemic lupus erythematosus. CM, van Tol MJ, van Schaardenburg Arthritis Res Ther; 12:R53. D, Huizinga TW, Toes RE, Trouw 55) Crispín JC, Oukka M, Bayliss G, LA.2012. Avidity maturation of anti- Cohen RA, Van Beek CA, Stillman IE, citrullinated protein antibodies in Kyttaris VC, Juang YT, Tsokos GC. rheumatoid arthritis. Arthritis 2008. Expanded double negative T Rheum.;64:1323-8. cells in patients with systemic lupus 63) Begovich AB, Carlton VE, Honigberg erythematosus produce IL-17 and LA, Schrodi SJ, Chokkalingam AP, infiltrate the kidneys. J Immunol; Alexander HC, Ardlie KG, Huang Q, 181:8761-6. Smith AM, Spoerke JM, Conn MT, 56) Valencia X, Yarboro C, Illei G, Lipsky Chang M, Chang SY, Saiki RK, PE.2007. Deficient CD4+CD25high T Catanese JJ, Leong DU, Garcia VE, regulatory cell function in patients with McAllister LB, Jeffery DA, Lee AT, active systemic lupus erythematosus. J Batliwalla F, Remmers E, Criswell LA, Immunol;178:2579-88. Seldin MF, Kastner DL, Amos CI, 57) Bonelli M, Savitskaya A, von Dalwigk Sninsky JJ, Gregersen PK. 2004. A K, Steiner CW, Aletaha D, Smolen JS, missense single-nucleotide poly- Scheinecker C. 2008. Quantitative and morphism in a gene encoding a protein qualitative deficiencies of regulatory T tyrosine phosphatase (PTPN22) is cells in patients with systemic lupus associated with rheumatoid arthritis. erythematosus (SLE). Int Am J Hum Genet;75:330-7. Immunol;20:861–8. 64) Remmers EF, Plenge RM, Lee AT, 58) Suarez A, Lopez P, Gómez J, Gutiérrez Graham RR, Hom G, Behrens TW, de C.2006. Enrichment of CD4+ Bakker PI, Le JM, Lee HS, Batliwalla

14 F, Li W, Masters SL, Booty MG, CD25brightCD4+ regulatory T cells Carulli JP, Padyukov L, Alfredsson L, are enriched in inflamed joints of Klareskog L, Chen WV, Amos CI, patients with chronic rheumatic Criswell LA, Seldin MF, Kastner DL, disease. Arthritis Res. Ther; 6,R335– Gregersen PK. 2007.STAT4 and the 46. risk of rheumatoid arthritis and 72) Lawson CA, Brown AK, Bejarano V, systemic lupus erythematosus. N Engl Douglas SH, Burgoyne CH, Greenstein J Med;357:977-86. AS, Boylston AW, Emery P, Ponchel 65) Kanik KS, Hagiwara E, Yarboro C, F, Isaacs JD.2006. Early rheumatoid Schumacher HR, Wilder RL, Klinman arthritis is associated with a deficit in DM.1998. Distinct patterns of cytokine the CD4+CD25high regulatory T cell secretion characterize new onset population in peripheral blood. synovitis versus chronic rheumatoid Rheumatology (Oxford);45:1210–7. arthritis. J Rheumatol;25:16-22. 73) Toubi E, Kessel A, Mahmudov Z, 66) Steiner, G., M. Tohidast-Akrad, G. Hallas K, Rozenbaum M, Rosner I. Witzmann, M. Vesely, A. Studnicka- 2005 Increased spontaneous apoptosis Benke, A. Gal, M. Kunaver, P. Zenz, of CD4+CD25+ T cells in patients with and J. S. Smolen. 1999. Cytokine active rheumatoid arthritis is reduced production by synovial T cells in by infliximab. Ann NY Acad rheumatoid arthritis. Rheumatology Sci;1051:506–14. (Oxford);38:202-13. 74) van Amelsfort JM, Jacobs KM, Bijlsma 67) Ronnelid J, Berg L, Roberg S, Nilsson JW, Lafeber FP, Taams LS.2004 A, Albertsson K, Klareskog L. CD4(+)CD25(+) regulatory T cells in 1998.Production of T-cell cytokines at rheumatoid arthritis: differences in the the single-cell level in patients with presence, phenotype, and function inflammatory arthritides: enhanced between peripheral blood and synovial activity in synovial fl uid compared to fluid. Arthritis Rheum; 50: 2775– 85. blood. Br J Rheumatol;37:7-14. 75) Han GM, O'Neil-Andersen NJ, Zurier 68) Chabaud M, Fossiez F, Taupin JL, RB, Lawrence DA. 2008 Miossec P.1998. Enhancing effect of CD4+CD25high T cell numbers are IL-17 on IL-1-induced IL-6 and enriched in the peripheral blood of leukemia inhibitory factor production patients with rheumatoid arthritis. Cell by rheumatoid arthritis synoviocytes Immunol;253: 92–101. and its regulation by Th2 cytokines. J 76) Behrens F, Himsel A, Rehart S, Immunol;161:409-14. Stanczyk J, Beutel B, Zimmermann 69) Miossec P, Korn T, Kuchroo VK. SY, Koehl U, Möller B, Gay S, 2009. Interleukin-17 and type 17 helper Kaltwasser JP, Pfeilschifter JM, T cells. N Engl J Med;361:888-98. Radeke HH.2007. Imbalance in 70) Genovese MC, Van den Bosch F, distribution of functional autologous Roberson SA, Bojin S, Biagini IM, regulatory T cells in rheumatoid Ryan P, Sloan-Lancaster J. 2010. arthritis. Ann Rheum Dis;66:1151-6. LY2439821, a humanized anti- 77) Nadkarni S, Mauri C, Ehrenstein interleukin-17 monoclonal antibody,in MR.2007.Anti-TNF-alpha therapy the treatment of patients with induces a distinct regulatory T cell rheumatoid arthritis: a phase I population in patients with rheumatoid randomized, double-blind, placebo- arthritis via TGF-beta. J Exp controlled, proof-of-concept study. Med;204:33-39. [Erratum, J Exp Med Arthritis Rheum;62:929-39. 2007;204:205.] 71) Cao D, van Vollenhoven R, Klareskog L, Trollmo C, Malmström V.2004.

15 78) Ebling FM, Hahn BH. Pathogenic chromatin peptides reveal a major subsets of antibodies to DNA.1989. Int autoepitope that primes pathogenic T Rev Immunol;5:79–95. and B cells of lupus. J Immunol; 79) Burlingame RW, Rubin RL, Balderas 168:2530–7. RS, Theofilopoulos AN. 1993. Genesis 87) Voll RE, Roth EA, Girkontaite I, Fehr and evolution of antichromatin H, Herrmann M, Lorenz HM, Kalden autoantibodies in murine lupus JR. 1997. Histonespecific Th0 and Th1 implicates T-dependent immunization clones derived from systemic lupus with self antigen. J Clin Invest; erythematosus patients induce double- 91:1687–96. stranded DNA antibody production. 80) Burlingame RW, Boey ML, Arthritis Rheum;40:2162–71. Starkebaum G, Rubin RL.1994. The 88) Rajagopalan S, Mao C, Datta SK. central role of chromatin in 1992. Pathogenic autoantibody- autoimmune responses to histones and inducing  T helper cells from DNA in systemic lupus erythematosus. patients with lupus nephritis express J Clin Invest;94:184–92. unusual T cell receptors. Clin Immunol 81) Rubin RL, Bell SA, Burlingame RW. Immunopathol;62:344–50. 1992. Autoantibodies associated with 89) Alspaugh MA, Tan EM. 1975. lupus induced by diverse drugs target a Antibodies to cellular antigens in similar epitope in the (H2A-H2B)- Sjogren's syndrome. J Clin Invest; DNA complex. J Clin Invest;90:165– 55:1067-73. 73. 90) Reyes PA, Tan EM. 1977. DNA- 82) Shoenfeld Y, Segol O. 1989. Anti- binding property of Sm nuclear histone antibodies in SLE and other antigen. J Exp Med;145:749-54. autoimmune diseases. Clin Exp 91) Holyst MM, Hill DL, Hoch SO, Rheumatol;7:265–71. Hoffmann RW. 1997. Analysis of T 83) Schett G, Smolen J, Zimmermann C, and B cell responses against U small Hiesberger H, Hoefler E, Fournel S, nuclear ribonucleoprotein 70-kD, B Muller S, Rubin RL, Steiner G. 2002. and D polypeptides among patients The autoimmune response to chromatin with systemic lupus erythematosus and antigens in systemic lupus mixed connective tissue disease. erythematosus: autoantibodies against Arthritis Rheum; 40:1493-503. histone H1 are a highly specific marker 92) Steiner G, Hartmuth , Skriner K, for SLE associated with increased Maurer-Fogy I, Sinski A, Hassfeld W, disease activity. Lupus;11:704–15. Barta A, Smolen JS.1992. Purification 84) Schett G, Rubin RL, Steiner G, and partial sequencing of the nuclear Hiesberger H, Muller S, Smolen J. autoantigen RA33 shows that it is 2000. The lupus erythematosus cell indistinguishable from the A2 protein phenomenon: comparative analysis of of the heterogeneous nuclear antichromatin antibody specificity in ribonucleoprotein complex. J Clin lupus erythematosus cell-positive and - Invest; 90:1061-6. negative sera. Arthritis Rheum;43:420– 93) Hassfeld W, Steiner G, Studnicka- 8. Benke A, Skriner K, Graninger W, 85) Schett G, Steiner G, Smolen JS. 1998. Fischer I, Smolen JS.1995. Nuclear antigen histone H1 is primarily Autoimmune response to the involved in lupus erythematosus cell spliceosome: an immunologic link formation. Arthritis Rheum;41:1446– between rheumatoid arthritis, mixed 55. connective tissue disease, and systemic 86) Kaliyaperumal A, Michaels MA, Datta lupus erythematosus. Arthritis Rheum; SK. 2002. Naturally processed 38:777-85.

16 94) Steiner G, Skriner K, Smolen JS.1996. M.2005.Ex vivo analysis of Autoantibodies to the A/B proteins of desmoglein 1-responsive T-helper (Th) the heterogeneous nuclear 1 and Th2 cells in patients with ribonucleoprotein complex: novel tools pemphigus foliaceus and healthy for the diagnosis of rheumatic diseases. individuals. Exp Dermatol;14:586-92. Int Arch Allergy Immunol;111:314-9. 102) Diaz-Villoslada P, Shih A, Shao L, 95) Skriner K, Sommergruber WH, Genain CP, Hauser Tremmel V, Fischer I, Barta A, Smolen SL.1999.Autoreactivity to myelin JS, Steiner G.1997. Anti-hnRNP-A2 antigens: myelin/oligodendrocyte autoantibodies are directed to the RNA glycoprotein is a prevalent autoantigen. binding region of the A2 protein of the J Neuroimmunol;99:36-43. heterogeneous nuclear ribonucleo 103) Smolen, JS, and G Steiner. 1998. Are protein complex: differential epitope autoantibodies active players or recognition in rheumatoid arthritis, epiphenomena? Curr Opin systemic lupus erythematosus, and Rheumatol;10:201-6. mixed connective tissue disease. J Clin 104) Cope, AP, and G Sonderstrup. 1998. Invest;100:127-35. Evaluating candidate autoantigens in 96) Weighardt F, Biamonti G, Riva S.1996. rheumatoid arthritis. Springer Semim The roles of heterogeneous nuclear Immunopathol ;20:23-39. ribonucleoproteins (hnRNP) in RNA 105) Auger I, Charpin C, Balandraud N, metabolism. Bioessays;18:747-56. Martin M, Roudier J.2012. 97) Krecic AM, Swanson MS.1999. Autoantibodies to PAD4 and BRAF in HnRNP complexes: composition, rheumatoid arthritis. Autoimmun structure, and function. Curr Opin Cell Rev;11:801-3 Biol;11:363-71. 106) Hassfeld W, Steiner G, Graninger 98) Mayeda A, Munroe SH, Caceres JF, G.,Witzmann G, Schweitzer H, and Krainer AR.1994. Function of Smolen JS. 1993. Autoantibody to the conserved domains of hnRNP A1 and nuclear antigen RA33: a marker for other hnRNP A/B proteins. EMBO early rheumatoid arthritis. Br J J;13:5483-95. Rheumatol;32:199-203. 99) Talken BL, Bailey CW, Reardon SL, 107) Skriner K, Sommergruber K, Tremmel Caldwell CW, Hoffman RW. 2001. WH, Fischer V, Barta A, Smolen JS Structural analysis of TCRalpha and and Steiner G.1997. Anti-A2/RA33 beta chains from human T-Cell clones autoantibodies are directed to the RNA specific for small nuclear binding region of the A2 protein of the ribonucleoprotein polypeptides Sm-D, heterogeneous nuclear Ribonucleo Sm-B and U1–70 kDa: TCR protein complex. Differential epitope complementarity determining region 3 recognition in rheumatoid arthritis, usage appears highly conserved. Scand systemic lupus erythematosus, and J Immunol;54:204-10. mixed connective tissue disease. J Clin 100) Greidinger EL, Gazitt T, Jaimes KF, Invest.;100:127–35. Hoffman RW.2004. Human T cell 108) Hayer S, Tohidast-Akrad M, clones specific for heterogeneous Haralambous S, Jahn-Schmid B, nuclear ribonucleoprotein A2 Skriner K, Trembleau S, Dumortier H, autoantigen from connective tissue Pinol-Roma S, Redlich K, Schett G, disease patients assist in autoantibody Muller S, Kollias G, Smolen J, Steiner production. Arthritis Rheum;50:2216- G. 2005.Aberrant expression of the 22. autoantigen heterogeneous nuclear 101) Gebhard KL, Veldman CM, Wassmuth ribonucleoprotein-A2 (RA33) and R, Schultz E, Schuler G, Hertl spontaneous formation of rheumatoid

17 arthritis-associated anti-RA33 auto systemic lupus erythematosus. Arthritis antibodies in TNF-a transgenic mice. J Rheum;41:823-30. Immunol;175:8327–36. 116) Schwartz HJ, Lowell FC, and Melby 109) Hoffmann MH, Tuncel J, Skriner K, JC. 1968. Steroid resistance in Tohidast-Akrad M, Turk B, Pinol- bronchial asthma. Ann Intern Roma S, Serre G Schett G, Smolen JS, Med;69:493-9. Holmdahl R, Steiner G..2007. The 117) Kozaci DL, Chernajovsky Y, Chikanza rheumatoid arthritis-associated auto IC. 2007. The differential expression of antigen hnRNP-A2 (RA33) is a major corticosteroid receptor isoforms in stimulator of autoimmunity in rats with corticosteroid-resistant and -sensitive pristane-induced arthritis. J Immunol; patients with rheumatoid arthritis. 179:7568–76. Rheumatology (Oxford.; 46:579-85. 110) Hoes JN, Jacobs JW, Buttgereit F. and 118) Chikanza IC and Kozaci DL. 2004. Bijlsma JW. 2010. Current view of Corticosteroid resistance in rheumatoid glucocorticoid co-therapy with arthritis: molecular and cellular DMARDs in rheumatoid arthritis. Nat perspectives. Rheumatology (Oxford); Rev Rheumatol;6, 693–702. 43:1337-45. 111) Galon J, Franchimont D, Hiroi N, Frey G, Boettner A, Ehrhart-Bornstein M, O'Shea JJ, Chrousos GP, Bornstein SR. 2002. Gene profiling reveals unknown enhancing and suppressive actions of glucocorticoids on immune cells. FASEB J;16:61-71. 112) Ehrchen J, Steinmüller L, Barczyk K, Tenbrock K, Nacken W, Eisenacher M, Nordhues U, Sorg C, Sunderkötter C, Roth J. 2007.Glucocorticoids induce differentiation of a specifically activated, anti-inflammatory subtype of human monocytes. Blood.;109:1265- 74. 113) Barnes PJ, Adcock IM. 2009. Glucocorticoid resistance in inflammatory diseases. Lancet; 373: 1905-17. 114) van Schaardenburg D, Valkema R, Dijkmans BA, Papapoulos S, Zwinderman AH, Han KH, Pauwels EK, Breedveld FC. 1995. Prednisone treatment of elderly-onset rheumatoid arthritis. Disease activity and bone mass in comparison with chloroquine treatment. Arthritis Rheum;38:334-42. 115) Seki M, Ushiyama C, Seta N, Abe K, Fukazawa T, Asakawa J, Takasaki Y, Hashimoto H. 1998.Apoptosis of lymphocytes induced by glucocorticoids and relationship to therapeutic efficacy in patients with

Chapter 2

Stepwise Differentiation of CD4 Memory T Cells Defined by Expression of CCR7 and CD27

RD Fritsch1 X Shen2 GP Sims1 KS Hathcock2 RJ Hodes2,3 PE Lipsky1

1Autoimmunity Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases 2Experimental Immunology Branch, National Cancer Institute 3National Institute on Aging, National Institutes of Health, Bethesda, Maryland

The Journal of Immunology. 2005;175:6489-649.

19 Stepwise Differentiation of CD4 Memory T Cells Deﬁned by Expression of CCR7 and CD271

Ruth D. Fritsch,2* Xinglei Shen,2† Gary P. Sims,* Karen S. Hathcock,† Richard J. Hodes,†‡ and Peter E. Lipsky3*

To study the steps in the differentiation of human memory CD4 T cells, we characterized the functional and lineage relationships of three distinct memory CD4 subpopulations distinguished by their expression of the cysteine chemokine receptor CCR7 and the TNFR family member CD27. Using the combination of these phenotypic markers, three populations were defined: the CCR7؉CD27؉, the CCR7؊CD27؉, and the CCR7؊CD27؊ population. In vitro stimulation led to a stepwise differentiation from naive to CCR7؉CD27؉ to CCR7؊CD27؉ to CCR7؊CD27؊. Telomere length in these subsets differed significantly CCR7؉CD27؉ > CCR7؊CD27؉ > CCR7؊CD27؊), suggesting that these subsets constituted a differentiative pathway with) progressive telomere shortening reflecting antecedent in vivo proliferation. The in vitro proliferative response of these populations declined, and their susceptibility to apoptosis increased progressively along this differentiation pathway. Cytokine secretion showed a differential functional capacity of these subsets. High production of IL-10 was only observed in CCR7؉CD27؉, whereas IFN-␥ was produced by CCR7؊CD27؉ and to a slightly lesser extent by CCR7؊CD27؊ T cells. IL-4 secretion was predominantly conducted by CCR7؊CD27؊ memory CD4 T cells. Thus, by using both CCR7 and CD27, distinct maturational stages of CD4 memory T cells with different functional activities were defined.

he ability to mount adaptive immune responses is a hall- only minimal capacity to secrete cytokines and to migrate prefer- mark of the immune system and involves the Ag-induced entially to the T cell areas of lymphoid organs, where they can be activation and expansion of naive T cells and their dif- restimulated by Ag. In contrast, CCR7-negative T are thought T EM ferentiation into memory cells. This process is essential for the to have acquired different migratory capabilities and effector func- evolution of an effective immune response because memory cells tions, such as cytokine production or cytotoxicity, which they can acquire altered capacity to produce cytokines, different migratory exert after their migration to inﬂamed peripheral tissue (5). In line

capability, and diminished activation requirements that allow them with this model, an enrichment of TCM was demonstrated in ton- to generate an effective secondary response on rechallenge with sils, whereas most T cells isolated from nonlymphoid tissue such specific Ag. Current understanding of the differentiation pathway as the lung, skin, or intestine lacked the expression of CCR7 and from naive cells through terminally differentiated memory cells is thus were thought to be TEM (6). incomplete. A model has been proposed (1), which distinguishes Notably, however, the use of CCR7 as a marker distinguishing central memory T cells (T )4 from effector memory T cells CM recirculating TCM from cytokine-secreting TEM has been chal- (TEM) by the phenotypic expression of the cysteine chemokine lenged. Thus, it has been reported that the majority of cytokine- receptor CCR7. CCR7 enables cells to home to secondary lym- secreting human T cells reside in the CCR7-expressing subset and phoid organs through interaction with its ligands, CCL19 (also as a result may be capable of lymphoid organ homing (7). Nev- known as EBV-induced molecule 1 ligand) and CCL21 (also ertheless, these authors did find a higher frequency of polarized called secondary lymphoid tissue chemokine), which are predom- Th1 or Th2 cells in the CCR7-negative TEM subset. Furthermore, inantly expressed by high endothelial venules (HEV) of secondary new data indicate the presence of “pre-Th1” and “pre-Th2” cells in lymphoid organs and parenchymal cells of the T cell zones of ␥ the TCM compartment, which secrete low amounts of IFN- or lymph nodes (2–4). TCM, expressing CCR7, are believed to have IL-4, and can be induced to differentiate into Th1 and Th2 effector memory T cells by antigenic stimulation or in response to homeostatic cytokines (8). *Autoimmunity Branch, National Institute of Arthritis and Musculoskeletal and Skin In addition to CCR7, phenotypic expression of other cell surface Diseases (NIAMS), †Experimental Immunology Branch, National Cancer Institute, and ‡National Institute on Aging, National Institutes of Health, Bethesda, MD 20892 proteins, especially costimulatory molecules, has been used to Received for publication May 17, 2005. Accepted for publication September 1, 2005. study the differentiation of memory T cells. In this respect, CD27, The costs of publication of this article were defrayed in part by the payment of page a member of the TNFR family, has also been used to delineate charges. This article must therefore be hereby marked advertisement in accordance stages of T cell differentiation (9–11). CD27 is expressed by naive with 18 U.S.C. Section 1734 solely to indicate this fact. CD4 and CD8 T cells as well as by most memory T cells and has 1 This work was supported by a Max Kade Foundation Postdoctoral Research Ex- change Grant (to R.D.F.). been assigned a costimulatory function. Initial up-regulation of CD27 expression upon TCR engagement is followed by irrevers- 2 R.D.F. and X.S. contributed equally to this work. ible loss after repeated antigenic stimulation (12, 13). CD27 ex- 3 Address correspondence and reprint requests to Dr. Peter E. Lipsky, National In- stitute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of pression in CD4 memory T cells distinguishes two populations Health, Building 10, Room 6D47C, Bethesda, MD 20892. E-mail address: whose distinct properties have been characterized (11). The loss of [email protected] CD27 correlates with an increase of IL-4 production. Accumula- 4 Abbreviations used in this paper: TCM, central memory T cell; TEM, effector memory T cell; HEV, high endothelial venule; PI, propidium iodide; FISH, fluorescence tion of these cells has been reported in patients with chronic al- in situ hybridization; Tr1, T regulatory 1. lergic diseases, rheumatoid arthritis, and in aged donors (11, 14).

Copyright © 2005 by The American Association of Immunologists, Inc. 0022-1767/05/$02.00 Although the combination of CCR7 and CD27 has been used in Cell cycle progression of T cell subsets after stimulation was examined an attempt to define a maturational pathway of CD8 T cells (15, by DNA staining with propidium iodide (PI). 16), the differentiation steps of CD4 memory T cells characterized by expression of CCR7 and CD27 have not been defined. Telomere length measurement Therefore, the purpose of this study was to characterize the mat- Telomere length of five individuals (age range: 43–72 years; mean: 55.0 Ϯ urational pattern of CD4 memory T cells by expression of CCR7 4.9) was measured by flow-fluorescence in situ hybridization (FISH) (17). ϩ and CD27. Isolated T cells were incubated in PBS 1% BSA. Cells were then incubated in the dark for 15 min in a hybridization mixture containing either 0.3 ␮g/ml FITC-labeled peptide nucleic acid-telomere FISH probe (Ap- Materials and Methods plied Biosystems) or without probe for autofluorescence background mea- Subjects surement. Cells were then heated at 86°C for 15 min and hybridized in the dark at room temperature for 1.5 h. Cell bodies were then gently washed The investigated population consisted of 42 healthy individuals (50% and resuspended. LDS 751 (Exciton), a DNA binding dye, was added to the Ϯ women, age: 40.6 2.5 years), whose blood was provided by the Depart- resuspended cells (10 ng/ml), and fluorescence intensities of the telomere ment of Transfusion Medicine, Clinical Center, National Institutes of probe and LDS were measured by flow cytometry using a FACSCalibur Health. Informed consent was obtained from all subjects. (BD Biosciences). Single lymphocytes were identified by forward scatter PBMC were obtained by centrifugation of heparinized blood over Fi- ϩ and by total DNA content, and these cells were gated for telomere binding coll-Hypaque (Amersham Biosciences). T cells or CD4 T cells were neg- fluorescence. Quantitation of telomere length was performed by interpo- atively selected on a magnetic column using the AutoMACS (Miltenyi lating the fluorescence signal from the peptide nucleic acid telomere probe Biotec) and the respective isolation kits for Pan T cell or CD4 isolation ϩ to a standard curve generated by beads with known fluorescence units (Miltenyi Biotec). In some cases, CD4 T cells were further subjected to (Bangs Laboratories). Variations in hybridization conditions between ex- positive selection for CD45RO-expressing cells by the CD45RO kit and periments were normalized by comparison to an aliquot of a control cell AutoMACS. All procedures were conducted according to the manufactur- preparation run with each experiment. er’s instruction. The purity of the resultant population was analyzed by Ͼ flow cytometry and was generally 95%. Telomerase activity Flow cytometry Naive CD4 T cells and sorted CD4 memory T cell subsets were stimulated ␮ For phenotypic analysis, T cells were incubated for 30 min at 4°C with a overnight with 10 ng/ml PMA plus 1.34 M ionomycin followed by a FITC-coupled mAb against CCR7 (clone no. 150503; R&D Systems), PE- 3-day culture period in medium supplemented with IL2. Telomerase ac- conjugated anti-CD27 mAb (clone no. M-T271; BD Pharmingen), Per- tivity was assayed three days later using the TRAPeze Telomerase Detec- CPCy5.5-coupled anti-CD4 mAb (clone no. SK3; BD Biosciences), and tion kit (Chemicon/Serologicals) following the manufacturer’s instruc- allophycocyanin-conjugated anti-CD45RO mAb (clone no. UCHL-1; BD tions. Telomerase activity in serial dilutions of cell lysates was detected by Pharmingen). Abs of the appropriate IgG isotypes were used as negative its ability to extend telomeric repeats on a substrate, the template strand controls. Naive T cells served as positive control for staining with anti- primer. These telomerase products were then amplified by a two-step PCR CCR7 and anti-CD27. For chemokine receptor analysis, CD4ϩCD45ROϩ for 31 cycles and run on a 12% PAGE gel. DNA bands were visualized T cells were isolated by AutoMACS and stained with the above-mentioned using SYBR-green (Molecular Probes) and a PhosphorImager (Molecular FITC-anti-CCR7, PE-anti-CD27, and PerCPCy5.5-anti-CD4, as well as Dynamics). A competitive internal standard was used to control for the with allophycocyanin-conjugated mAb against CXCR4 (clone no. 12G5; efficiency of the PCR amplification as well as to detect the presence of Taq BD Biosciences) or a biotinylated mAb against CXCR5 (clone no. RF8B2; polymerase inhibitors. Telomerase activity was calculated as the intensity BD Pharmingen), followed by streptavidin-coupled allophycocyanin. Cells of telomerase products generated divided by the intensity of the internal were analyzed using a FACSCalibur flow cytometer (BD Biosciences). The standard. A heat-inactivated lysate and a buffer-only lysate were used as acquired data were analyzed using Flow Jo software (Tree Star). negative controls, and TSR8 control template and EL-4 lysate were used as a positive control. To control for day-to-day variation, one sample from Cell separation by flow cytometry each experimental group was repeated in a single assay. These values were then used as a reference to normalize the other samples from the same For cytokine analysis, proliferation assays, assessment of telomere length, experimental day. and telomerase activity, cells were sorted using a MoFlo flow cytometer (DakoCytomation). Anti-CCR7-FITC, anti-CD27-PE, anti-CD4-Per- Cytokine analysis CPCy5.5, and anti-CD45RO-allophycocyanin were in some cases supplemented with anti-CD3-PE-Cy7 to avoid purification steps on the Au- FACS-sorted CD4 memory T cell subsets (1 ϫ 104/well) were activated toMACS and the inherent cell loss. Postsort controls of the respective with PMA/ionomycin overnight, and supernatants were collected after sorted populations routinely exhibited Ͼ95% purity. 20 h. Cytokines in supernatants were measured by Cytometric bead array (CBAkit-II; BD Biosciences) according to the manufacturer’s manual us- Cell culture and cell counting assays ing a FACSCalibur (BD Biosciences). To normalize individual variation, the cytokine secretion of CCR7Ϫ/CD27ϩ and CCR7Ϫ/CD27Ϫ was ex- After an initial overnight activation with 10 ng/ml PMA (Sigma-Aldrich) ϩ ϩ and 1.34 ␮M ionomycin (Calbiochem), sorted memory T cell subsets were pressed as a percentage of the secretion observed in the CCR7 /CD27 washed and maintained in 50 U/ml IL-2 (Biological Research Branch, Di- population. vision of Cancer Treatment and Diagnosis, National Cancer Institute) in U-bottom 96-well-plates (Costar). Culture medium consisted of Ultra Cul- Statistical methods ture serum-free medium (BioWhittaker). After 1 wk, cells were stimulated Ϯ with anti-CD3 (64.1: 1 ␮g/ml) and cultured for another week, after which Values are expressed as mean SEM, and for statistical analysis Student’s they were activated again with PMA/ionomycin as described above. Phe- t test was applied. notypic characterization was performed immediately after the last stimulation and after an additional 9 days of incubation in IL-2. Alternatively, Results cells were stained 3 days after each respective stimulation and after an additional resting period of 4 days. Characterization of memory CD4 T cells For analysis of actual cell numbers, 5 ϫ 104 cells from each population Using naive CD45ROϪ CD4 T cells that uniformly express CCR7 tested were stimulated with PMA/ionomycin overnight and then cultured and CD27 as a positive control and isotype-matched control mAb with 50 U/ml IL-2. Cell numbers were assessed by flow cytometry using as a negative control, we could identify three major subsets of SPHERO beads (Spherotech) after an additional 72 h of culture. Dead cells ϩ ϩ Ϫ ϩ were excluded by staining with 7-aminoactinomycin D (Molecular Probes). memory CD4ϩT cell: the CCR7 /CD27 , the CCR7 /CD27 , The proliferative capacity is expressed as the ratio of cells recovered from and the CCR7Ϫ/CD27Ϫ. Within CD4ϩ T cells of healthy individ- ϩ ϩ the various subsets to the cells recovered from the CCR7 /CD27 subset. uals, the majority of cells expressed both CCR7 and CD27 Alternatively, proliferation of T cell subsets was measured after stimulation Ϯ Ϯ with anti-CD3 mAb. A total of 3 ϫ 104 cells of each respective subset were (mean SEM: 69.8 2.1%). Memory T cells that did not express incubated in triplicate with plate-bound anti-CD3 (64.1: see above), and CCR7 but were CD27 positive accounted for 21.5 Ϯ 1.8%, ϩ Ϫ proliferation was measured after 72 h by [3H]thymidine incorporation. whereas only 5.6 Ϯ 0.6% of memory CD4 T cells were CCR7 / FIGURE 1. Percentage of memory CD4 T cell subsets in normal subjects. Memory CD4 T cells were stained for the expression of CCR7 and CD27, and three major populations were assessed by comparison to the negative (isotype control staining) and positive control (i.e., naive T cells), namely CCR7ϩ/CD27ϩ, CCR7Ϫ/CD27ϩ, and CCR7Ϫ/CD27 Ϫ (A). The bars in B are mean (ϮSEM) of 42 samples from different normal subjects.

CD27Ϫ (Fig. 1). The observed patterns of CCR7 and CD27 ex- CD27ϩ and CCR7Ϫ/CD27ϩ CD4 memory T cell subsets (Fig. pression in healthy individuals did not correlate with the propor- 2A). A portion of CCR7ϩ/CD27ϩ memory CD4 T cells (mean Ϯ tion of CD4ϩCD45ROϩ total memory cells and did not correlate SEM: 30.6 Ϯ 4.3%) became CCR7Ϫ/CD27ϩ, whereas approxi- signiﬁcantly with the age of the donors (age range: 8–72 years; mately one-half (mean Ϯ SEM: 46.7 Ϯ 5.0%) became negative for mean age Ϯ SEM: 40.6 Ϯ 2.5). both surface markers. Furthermore, the majority (mean Ϯ SEM: Ϯ Ϫ ϩ ϩ 80.5 6.9%) of the CCR7 /CD27 cell population lost expres- Phenotype of CD4 memory T cells after repeated stimulation sion of CD27 and became CCR7Ϫ/CD27Ϫ after repeated stimula- To examine the relationship between the CD4 memory T cell sub- tion (Fig. 2A). To examine the stability of the phenotypic change, sets, the effect of in vitro stimulation on cell surface phenotype was cells were also assessed after an additional resting period. As can assessed. Repeated stimulation altered the phenotype of CCR7ϩ/ be seen in Fig. 2A, the altered phenotype remained stable during

FIGURE 2. Phenotypic changes of CD4 memory T cell subsets after repeated stimulation. CD4ϩ memory T cells from healthy individuals were sorted according to their phenotypic expression of CCR7 and CD27. Subsets were stimulated three times with anti-CD3 or PMA/ionomycin over the course of 3 wk and subsequently assessed for their phenotype. After a resting period of 9 days in medium supplemented with IL-2, cells were again stained for the surface expression of CCR7 and CD27. The displayed pattern was seen in six healthy individuals; one representative example is shown in A. B, The mean percentage of CCR7Ϫ/CD27Ϫ cells in the respective populations (CCR7ϩ/CD27ϩ: ϩ/ϩ; CCR7Ϫ/CD27ϩ: Ϫ/ϩ; and CCR7Ϫ/CD27Ϫ: Ϫ/Ϫ) derived from three healthy individuals 3 days af- The Jour ter each stimulation. the resting period of 9 days. CD4ϩ T cells appeared to alter their ulations, except for the CCR7Ϫ/CD27Ϫ subset. Proliferative ca- phenotype in a stepwise manner following activation (Fig. 2B). pacity decreased from naive cells to CCR7ϩ/CD27ϩ (Fig. 3A), Ϫ ϩ After one round of stimulation, the majority of the CCR7 /CD27 further declining in the CCR7Ϫ/CD27ϩ T cells. The lowest cell Ϫ Ϫ converted to a CCR7 /CD27 phenotype (mean: 77.0 Ϯ 1.8%). In recovery was noted in CCR7Ϫ/CD27Ϫ memory T cells (Fig. 3B). ϩ ϩ contrast, Ͼ75% of CCR7 /CD27 cells lost CCR7 after the first Stimulation with anti-CD3 demonstrated a similar pattern, with Ϫ Ϫ stimulation and then progressively converted to CCR7 /CD27 . a significant decrease in proliferative capacity from naive CD4 T ϩ ϩ The bulk of naive cells retained their CCR7 /CD27 phenotype cells to CCR7ϩ/CD27ϩ with the CCR7Ϫ/CD27Ϫ exhibiting the after one stimulation (68.4 Ϯ 9.6%), after which they increasingly least proliferation (Fig. 3C). The difference in proliferation was Ϫ Ϫ became CCR7 /CD27 with additional stimulation (Fig. 2B). Fi- Ͻ Ϫ Ϫ statistically significant between all subsets ( p 0.03). nally, the CCR7 /CD27 subset retained its phenotype despite To gain more insight into the balance of cell death and prolif- repeated stimulation and did not (re)-express either CCR7 or CD27 eration manifested by the subsets, cell cycle analysis was con- (Fig. 2B). These results clearly identify a differentiation pathway ducted. A representative example is shown in Fig. 4A. Immediately of CD4ϩ memory T cells in which CCR7ϩ/CD27ϩ cells mature to Ϫ ϩ Ϫ Ϫ ex vivo, most cells from each subset were in G0-G1, and at most, CCR7 /CD27 cells and finally to CCR7 /CD27 terminally dif- Ϫ Ϫ 2% (CCR7 /CD27 ) were in G -M (Fig. 4B). Upon stimulation, ferentiated memory T cells. 2 each of the subsets progressed into the cell cycle to varying degrees as noted in Fig. 4, C and D. Naive and CCR7ϩ/CD27ϩ ϩ Proliferative capacity and cell survival of the CD4 memory T memory T cells showed a similar pattern of cell cycle progression cell subsets ϳ with 30% of the cells in S-G2-M and only a small percentage Ϫ ϩ Ϫ The subsets were assessed for the capacity to proliferate after in becoming apoptotic. In contrast, CCR7 /CD27 and CCR7 / Ϫ vitro stimulation. Initially, cells were stimulated with PMA/iono- CD27 subsets exhibited similar cell progression patterns that ϩ ϩ mycin so that intrinsic differences in proliferative capacity could were different from those in naive and CCR7 /CD27 cells. A be assessed separately from potential changes in TCR signaling. rather small percentage of these cells was stimulated to enter

As shown in Fig. 3, A and B, 3 days after in vitro stimulation with S-G2-M, whereas a sizeable portion of the cells become apoptotic PMA/ionomycin, the number of cells detected increased in all pop- (Fig. 4D).

FIGURE 3. Differences in survival and proliferation of CD4 memory subsets. CD4ϩ (CD45ROϪ) naive and (CD45RObright) memory T cells of seven healthy individuals were sorted according to their expression of CCR7 and CD27. Viability and cell number of each subset were assessed 72 h after PMA/ionomycin stimulation by the use of 7-aminoactinomycin D and counting beads and the flow cytometer. Alternatively, proliferation was induced by immobilized anti-CD3 and measured by [3H]thymidine incorporation. A, A typical example of the different proliferation profiles of the subtypes. A total of 5 ϫ 104 cells (dashed line) was stimulated with PMA/ionomycin, and the cell number was assessed by FACS analysis. All subpopulations showed an increased cell number after stimulation with PMA/ionomycin. B, The mean (Ϯ SEM) number of live cells (derived from seven healthy individuals) of the indicated subsets after the 72 h incubation with PMA/ionomycin normalized to the CCR7ϩ/CD27ϩ cell count (naive CD4 T cells: 141.6 Ϯ 37.1%; CCR7Ϫ/ CD27ϩ T cells: 63.4 Ϯ 8.5%; and CCR7Ϫ/CD27Ϫ: 61.4 Ϯ 7.6%). The difference in the number of cells between CCR7ϩ/CD27ϩ and all other subsets reached statistical significance. (૾, p Ͻ 0.01; ૾૾, p Ͻ 0.0003). C, The different degrees of proliferation exhibited by the CD4 memory cell subsets upon stimulation with anti-CD3. A total of 3 ϫ 104 T cells of each subset was stimulated with immobilized anti-CD3 for 72 h, and proliferation was measured by [3H]thymidine incorporation. Proliferation declined from naive CD4 T cells to CCR7ϩ/CD27ϩ and further to CCR7Ϫ/CD27ϩ and was lowest in the CCR7Ϫ/CD27Ϫ subset. The difference in proliferation was statistically significant between all subsets (p Ͻ 0.03). Data from one representative experiment are shown. FIGURE 4. Cell cycle analysis of memory subsets ex vivo and after in vitro stimulation. A, Cell cycle of the respective memory T cell subsets was assessed immediately ex vivo using PI staining. Subsets from one representative subject ex vivo are shown. B, Analysis of memory subsets from four healthy individuals ex vivo. C, A representative example of changes in cell cycle stage of the different subsets after stimulation with PMA/ionomycin overnight followed by 24 h incubation with IL-2-supplemented medium using PI staining is shown. D, Analysis of memory subsets from four healthy ϩ ϩ Ϫ ϩ individuals after stimulation. The differences in the percentage of cells in G0-G1 and apoptosis between naive or CCR7 /CD27 and either CCR7 /CD27 or CCR7Ϫ/CD27Ϫ were all statistically significant (p Ͻ 0.01). The difference between CCR7ϩ/CD27ϩ and CCR7Ϫ/CD27Ϫ was also statistically significant ϭ in G2-M (p 0.01).

Telomere lengths in memory CD4 T cell subsets Inducible telomerase activity in memory CD4 subsets Previous studies have indicated that telomeres shorten significantly As telomere length is influenced by the activity of telomerase, as T cells divide during in vivo differentiation from naive to mem- inducible telomerase activity was examined in CD4 T cell subsets. ory cells. To test the lineage relationship between memory CD4 The populations derived from each normal donor showed very subsets suggested by the in vitro experiments, we examined the ex consistent patterns of inducible telomerase activity, which was vivo telomere length of the respective subsets by flow-FISH. As higher in the naive population than any of the memory subsets shown in Fig. 5A, consistent differences in telomere length were ( p Ͻ 0.002; Fig. 6A). Within the CD4 memory subsets, the observed among these populations. Naive CD4 T cells of five CCR7ϩ/CD27ϩ subset had higher activity than either the CCR7Ϫ/ healthy individuals (age range: 43–72 years; mean: 55.0 Ϯ 4.9; CD27ϩ or CCR7Ϫ/CD27Ϫ subsets (both p Ͻ 0.002), both of 60% female) had significantly longer telomeres than any of the which had similarly low telomerase activity levels (Fig. 6B). memory CD4 subsets. Moreover, subsets within the memory population had significant differences in telomere length, with longest Differences in cytokine secretion in CD4 T cell memory subsets telomeres in the CCR7ϩ/CD27ϩ subset, followed by the CCR7Ϫ/ The functional properties of the memory T cell subsets were ex- CD27ϩ subset, whereas CCR7Ϫ/CD27Ϫ cells had the shortest amined by measuring cytokine secretion after stimulation (Fig. 7). telomeres. IFN-␥ and IL-4, the main cytokines associated with mature Th1 Thus, the finding of progressive telomere shortening is consis- and Th2 cells, respectively, showed a differential secretion pattern tent with the developmental sequence suggested by the phenotypic in CD4 memory subsets. The greatest IFN-␥ secretion was ob- changes after in vitro stimulation. served in CCR7Ϫ/CD27ϩ and less in the CCR7ϩ/CD27ϩ and

FIGURE 5. Telomere length in CD4 naive and memory T cell subsets. Telomere length was measured by flow-FISH in each of the sorted populations. A, Representative sample of relative telomere lengths from one do- nor. B, Mean telomere lengths of subsets from five healthy individuals. Samples from the same donor are connected by dashed lines. Mean values of telomere length for each subset are represented by the horizontal lines. Telomere length of each subset is significantly different from the others (p Ͻ 0.025). FIGURE 6. Telomerase activity in CD4 naive and memory T cell subsets. A, Telomerase activity of (PMA/ionomycin) activated cells is calculated as the intensity of the telomerase product generated (TPG) divided by the PCR amplification of the internal control (IC) as shown in a representative experiment. B, Com- parison of inducible telomerase activity between subsets. The telomerase activity in CD4 T cells decreased from naive cells to CCR7ϩ/CD27ϩ to CCR7Ϫ/CD27ϩ and CCR7Ϫ/CD27Ϫ (૾, p Ͻ 0.02).

CCR7Ϫ/CD27Ϫ populations. In contrast, as already reported for shown in Fig. 8, the expression of CXCR4 decreased from CD27Ϫ cells (13), CCR7Ϫ/CD27Ϫ cells showed the highest IL-4 CCR7ϩ/CD27ϩ Ͼ CCR7Ϫ/CD27ϩ Ͼ CCR7Ϫ/CD27Ϫ. As with production of the three populations, whereas the IL-4 secretion of CXCR4, a decrease in CXCR5 expression was observed from the CCR7Ϫ/CD27ϩ subset was comparable to the CCR7ϩ/CD27ϩ. CCR7ϩ/CD27ϩ to CCR7Ϫ/CD27ϩ and CCR7Ϫ/CD27Ϫ. The ex- Secretion of the anti-inflammatory cytokine, IL-10, was greatest in pression of CXCR4 and CXCR5 in the CCR7ϩ/CD27ϩ was sig- the CCR7ϩ/CD27ϩ population, decreased in CCR7Ϫ/CD27ϩ nificantly ( p Ͻ 0.05) higher than in the CCR7Ϫ/CD27ϩ subset. cells, and was even less in the CCR7Ϫ/CD27Ϫ. IL-2 production showed an expected decrease from CCR7ϩ/CD27ϩ to CCR7Ϫ/ CD27ϩ and CCR7Ϫ/CD27Ϫ. Discussion The ability to separate stages of development by expression of the Differences in chemokine receptor expression in CD4 T cell chemokine receptor CCR7, as well as the TNFR family member memory subsets CD27, has provided a more precise delineation of stages of CD4 To investigate the trafficking potential of the CD4 memory T cell memory cell maturation. Based on in vitro studies, telomere length subsets, the expression of CXCR4 and CXCR5 was examined. As measurement, and assessment of changes in functional capacities,

FIGURE 7. Characterization of cytokine secretion in memory subsets. Subsets from eight individuals were examined. Secretory capacity is expressed as percentage of the secretion observed in the CCR7ϩ/CD27ϩ subset. The signiﬁcant differences (p Ͻ 0.05) between the respective subsets are indicated by ૾. marked decrease in proliferative capacity appeared to be characteristic of the entire CCR7-/CD27- subset, and not a unique property of the CD28- fraction. Progressive decrease in telomere length from naive CD4 T cells to the CCR7ϩ/CD27ϩ memory CD4 T cell population and further to the CCR7Ϫ/CD27ϩ subset and the CCR7Ϫ/CD27Ϫ CD4 memory cells substantiated the model of stepwise CD4 T cell maturation and expanded upon the known correlation of telomere shortening and maturation from the naive to memory stage (23) in CD4 T cells. These results have clearly delineated the CCR7Ϫ population into two subsets based on the expression of CD27 and also established the lineal relationship between the previously desig- ϩ Ϫ nated CCR7 TCM and the previously designated CCR7 TEM. Previous data have not clearly demonstrated whether TCM and TEM are independent subsets or stages of maturation along a con- tinuum of differentiation. The current data clearly support the latter conclusion. The intrinsic in vitro proliferative capacity of the memory subsets differed. This was demonstrated initially with PMA/ionomycin stimulation to bypass possible variations in TCR signaling and conﬁrmed by stimulation with anti-CD3. In previous reports, memory CD45RO T cells have been shown to exhibit a higher turnover rate and a relatively short lifespan compared with naive T

cells (24–26); furthermore, recently TEM of healthy individuals were found to manifest a higher in vivo turnover rate compared

with TCM and naive CD4 T cells (27). These results are consistent

with the telomere length analysis conducted here. If TEM had an increased in vivo turnover than TCM, it would be anticipated that FIGURE 8. Characterization of chemokine receptor expression in mem- they would have shorter telomeres, as we found in the CCR7Ϫ ory subsets. Subsets from four healthy individuals were examined for their subsets, indicating that the T had undergone more rounds of in expression of the chemokine receptors, CXCR4 and CXCR5. Significant EM vivo proliferation than T . However, the high turnover of T (p Ͻ 0.05) differences in chemokine receptor expression between the re- CM EM spective subsets are indicated by ૾. shown in the aforementioned study differed from the diminished in vitro proliferative capacity of both CCR7Ϫ subsets shown here. These results suggest that there might be in vivo constraints on ϩ including cytokine secretion, we propose a model of the develop- proliferation by CCR7 TCM, perhaps related to increased require- 3 ϩ ϩ ment from naive CD4 T cells CCR7 /CD27 memory CD4 T ments for productive activation. Alternatively, TEM may have cells 3 CCR7Ϫ/CD27ϩ 3 to the terminally differentiated greater in vivo exposure to Ag presented by the appropriate APC. CCR7Ϫ/CD27Ϫ CD4 memory T cells. Importantly, the in vivo experiments did not examine the potential In normal individuals, the majority (70%) of memory T cells proliferative capacity as was examined here by cell cycle analysis, found in blood bear both CCR7 and CD27 on their surfaces, ϳ20% cell counting by FACS beads and thymidine incorporation. To- express only CCR7, and only a very small proportion (5%) of gether, these data are most consistent with the conclusion that pro- memory cells do not express either markers. An even lower per- liferating TCM eventually give rise to TEM with diminished in vivo centage (3.0 Ϯ 0.3%, data not shown) exhibited the CCR7ϩ/ and in vitro proliferative potential. The endogenous signals that CD27Ϫ phenotype. We did not investigate this subset and its func- may contribute to the somewhat increased in vivo proliferation of tion is currently unknown. TEM remain to be delineated. In vitro culture experiments demonstrated a stepwise conversion The shortening of telomere length in dividing cells is counter- of CCR7ϩ/CD27ϩ into a CCR7Ϫ/CD27ϩ population and the pro- acted by telomerase, an enzyme that is able to extend telomeres by gression of the CCR7Ϫ/CD27ϩ into a CCR7Ϫ/CD27Ϫ subset. Be- synthesis of terminal telomeric repeats. Although regulation of te- cause the induced phenotypes were maintained after a resting pe- lomerase activity in lymphocytes is not completely understood, riod, a transient change of expression of either marker by recent decreased induction of telomerase activity has been observed with stimulation (10, 18) seems unlikely. The conclusion that there is a both repeated in vitro stimulation, as well as in vivo differentiation stepwise down-regulation of CCR7 and then CD27 during CD4 from naive to memory cells, and in parallel with telomere length memory T cells maturation is consistent with the finding that stim- shortening (28, 29). In line with these findings, we observed de- ulated naive CD4 T cells rapidly down-regulated CCR7, whereas creases in induced telomerase activity in the CCR7Ϫ/CD27ϩ and CD27 only declined at a later round of proliferation (19). A similar CCR7Ϫ/CD27Ϫ subsets with shorter telomeres. However, as te- finding was recently observed in bulk cultures of alloreactive lomerase activity might be influenced by stimulation and subse- CD45ROϩ memory T cells (20). quent entry into the cell cycle (27), the lower proportion of in As the down-regulation of CD27 exhibited similarities to the vitro-stimulated cycling cells in the CCR7Ϫ/CD27ϩ and CCR7Ϫ/ loss of CD28 with replicative senescence (21), we examined the CD27Ϫ compared with naive or CCR7ϩ/CD27ϩ memory cell sub- expression of CD28 by the various CD4ϩ memory T cell subsets. sets could account at least in part for the reduced telomerase ac- In accordance with previous findings (22), the CD28- cells were tivity of the CCR7Ϫ subsets. confined to the CCR7-/CD27- subset, constituting about half of Another feature of maturation and differentiation of cells is the this population (data not shown). Thus, the CD28- memory cells acquisition of certain functional properties such as cytokine secre- represent a portion of the CCR7-/CD27- subset. Notably, the tion. In this respect, a difference in cytokine secretion between TCM and TEM has been postulated (1) but also disputed in human expression in the more mature stages of CD4 memory T cells as well as murine models (5, 7), whereas the loss of CD27 is might again represent a decreased inclination of the majority of commonly associated with higher IL-4 production (13, 30). We these cells to circulate through lymphoid tissue as these cells dif- found a clear pattern of cytokine secretion of the CD4ϩ T cell ferentiate and is in accordance with recently published findings (8). subsets. We observed increased IFN-␥ and decreased IL-2 secre- Altogether, the data presented in this study lead to a more de- tion by CCR7Ϫ/CD27ϩ and CCR7Ϫ/CD27Ϫ cells compared with tailed view of human CD4 memory T cell maturation. We propose the CCR7ϩ/CD27ϩ population. This result is consistent with the a model of differentiation, in which the sequential stages can be conclusion that there is an association of “effector” cytokine se- distinguished by the phenotypic expression of CCR7 and CD27. ϩ cretion with the loss of CCR7 expression. However, the TEM pop- The results clearly indicate a stepwise conversion of CCR7 / ulation, as defined by CCR7Ϫ only, was not homogeneous, and the CD27ϩ T cells via the stage of CCR7Ϫ/CD27ϩ to a terminally subdivision into CCR7Ϫ/CD27ϩ and CCR7Ϫ/CD27Ϫ cells re- differentiated CCR7Ϫ/CD27Ϫ state. The relationship between vealed different cytokine secretion patterns. Whereas the produc- these subsets was demonstrated by the stepwise unidirectional dif- tion of IL-2 and IFN-␥ was somewhat greater in the CCR7Ϫ/ ferentiation in vitro, as well as differences in telomere length, cell CD27ϩ compared with the CCR7Ϫ/CD27Ϫ cells, IL-4 production cycle analysis, and proliferative capacity. Furthermore, as the cells was predominantly conducted only by CCR7Ϫ/CD27Ϫ memory T mature, they acquire different functional properties illustrated by a cells. change in cytokine secretion and chemokine receptor expression. These findings are consistent with previous results that IFN-␥ Thus, the defined subsets represent different maturational stages of secretion is initiated early after activation and then increases with CD4 memory T cells with unique functional properties successive cell cycles, whereas IL-4 production only appears after several cell divisions (31). The current data indicate that the Acknowledgments CCR7Ϫ/CD27Ϫ population constitutes the most terminally differ- We thank James Simone, Derek Hewgill, and Liusheng He of the Flow entiated state of memory cells, enriched in IL-4-producing cells Cytometry Facility of the National Institute of Arthritis and Musculoskel- and IFN-␥ secretors. The CCR7Ϫ/CD27ϩ subset represents an ear- etal and Skin Diseases Office of Science and Technology for carrying out lier maturational stage consisting of cells that have already been the cell sorting. biased toward Th1 cells but also of cells that are not yet irrevers- Disclosures ibly committed and might become Th2 cells in the following mat- The authors have no financial conflict of interest. urational step to CCR7Ϫ/CD27Ϫ cells. Interestingly, the pattern of IL-10 secretion was dissociated from IL-4 production, as produc- References tion of this anti-inflammatory cytokine was predominantly con- 1. Sallusto, F., D. Lenig, R. Forster, M. Lipp, and A. Lanzavecchia. 1999. Two fined to the CCR7ϩ/CD27ϩ cells. subsets of memory T lymphocytes with distinct homing and effector functions. ϩ ϩ Nature 401: 708–712. The enhanced IL-10 production by CCR7 /CD27 cells might 2. Baekkevold, E. S., T. Yamanaka, R. T. Palframan, H. S. Carlsen, F. P. Reinholt, indicate that a pool of T regulatory 1 (Tr1) cells resides within the U. H. von Andrian, P. Brandtzaeg, and G. Haraldsen. 2001. The CCR7 ligand elc CCR7ϩ/CD27ϩ subset. Tr1 cells suppress immune responses and (CCL19) is transcytosed in high endothelial venules and mediates T cell recruitment. J. Exp. Med. 193: 1105–1112. are known to express CCR7 (reviewed in Ref. 32). Because Tr1 3. Tanabe, S., Z. Lu, Y. Luo, E. J. Quackenbush, M. A. Berman, cells display a low proliferative capacity, it is somewhat surprising L. A. Collins-Racie, S. Mi, C. Reilly, D. Lo, K. A. Jacobs, and M. E. Dorf. 1997. ␤ that they were prominently found in the CCR7ϩ/CD27ϩ subset Identification of a new mouse -chemokine, thymus-derived chemotactic agent 4, with activity on T lymphocytes and mesangial cells. J. Immunol. 159: 5671–5679. that exhibited the greatest proliferative capacity. 4. Gunn, M. D., K. Tangemann, C. Tam, J. G. Cyster, S. D. Rosen, and The loss of IL-10-producing cells in the CCR7Ϫ/CD27ϩ and L. T. Williams. 1998. A chemokine expressed in lymphoid high endothelial Ϫ Ϫ venules promotes the adhesion and chemotaxis of naı¨ve T cells. Proc. Natl. Acad. CCR7 /CD27 T cell populations is consistent with the conclu- Sci. USA 95: 258–263. sion that TEM, which enter tissue, are devoid of this population of 5. Masopust, D., V. Vezys, A. L. Marzo, and L. Lefrancois. 2001. Preferential regulatory cells so as to ensure maximal capacity to respond to localization of effector memory cells in nonlymphoid tissue. Science 291: 2413–247. pathogens peripherally. The current results are compatible with the 6. Campbell, J. J., K. E. Murphy, E. J. Kunkel, C. E. Brightling, D. Soler, Z. Shen, possibility that regulation of immune responses by Tr1 cells may J. Boisvert, H. B. Greenberg, M. A. Vierra, S. B. Goodman, et al. 2001. CCR7 largely occur in secondary lymphoid organs. Expression and memory T cell diversity in humans. J. Immuol. 166: 877–884. 7. Kim, C. H., L. Rott, E. J. Kunkel, M. C. Genovese, D. P. Andrew, L. Wu, and Another characteristic of cell maturation and differentiation is E. C. Butcher. 2001. Rules of chemokine receptor association with T cell polar- the change in trafficking potential. The central role of CXCR4 in ization in vivo. J. Clin. Invest. 108: 1331–1339. 8. Rivino, L., M. Messi, D. Jarrossay, A. Lanzavecchia, F. Sallusto, and J. Geginat. the trafficking and maintenance of hemopoietic cells has been de- 2004. Chemokine receptor expression identifies pre-T helper (Th)1, pre-Th2, and scribed previously (33, 34). It is likely that the tendency to recir- nonpolarized cells among human CD4ϩ central memory T cells. J. Exp. Med. culate to lymphoid organs through the interaction of this receptor 200: 725–735. 9. van Lier, R. A. W., J. Borst, T. M. Vroom, H. Klein, P. van Mourik, with its ligand, CXCL12 (stromal cell-derived factor 1), declines W. P. Zeijlemaker, and C. J. Melief. 1987. Tissue distribution and biochemical with maturational state. In this respect, the unique homing poten- and functional properties of Tp55 (CD27), a novel T cell differentiation antigen. tial of the terminally differentiated CCR7Ϫ/CD27Ϫ subset has J. Immunol. 139: 1589–1596. Ϫ 10. Hintzen, R. Q., R. de Jong, S. M. Lens, M. Brouwer, P. Baars, and R. A. van Lier. been addressed previously (6) and the predominance of CCR7 / 1993. Regulation of CD27 expression on subsets of mature T lymphocytes. J. Im- CD27Ϫ T cells in gut, lung, and liver has been demonstrated. Sim- munol. 151: 2426–2430. ϩ 11. Tortorella, C., H. Schulze-Koops, R. Thomas, J. B. Splawski, L. S. Davis, ilar to CXCR4, a reduction of CXCR5 cells was observed in the L. J. Picker, and P. E. Lipsky. 1995. Expression of CD45RB and CD27 identifies more differentiated T cell subsets. The ligand for CXCR5, subsets of CD4ϩ memory T cells with different capacities to induce B cell dif- CXCL13, is a chemokine expressed in B cell follicles and serves ferentiation. J. Immunol. 155: 149–162. 12. Hintzen, R. Q., J. Jong, S. M. Lens, and R. A. W. van Lier. 1994. CD27: marker to attract follicular helper T cells that have an enhanced capacity to and mediator of T cell activation. Immunol. Today 15: 307–311. provide help to B cells (35) It is possible that a fraction of the 13. De Jong, R., M. Brouwer, B. Hooibrink, T. Van der Puw-Kraan, F. Miedema, and Ϫ ϩ CXCR5ϩCCR7ϩ/CD27ϩ cells represents follicular helper cells, R. A. van Lier. 1992. The CD27 subset of peripheral blood memory CD4 Ϫ lymphocytes contains functionally differentiated T lymphocytes that develop by although previous studies have shown that CD27 memory T cells persistent antigenic stimulation in vivo. Eur. J. Immunol. 22: 993–999. have an enhanced capability of providing B cell help (11). This 14. Nijhuis, E. W., E. J. Remarque, B. Hinloopen, T. van der Pouw-Kraan, R. A. van Lier, G. J. Ligthart, and L. Negelkerken. 1994. Age related increase in makes it more likely that the follicular helper cells may reside the fraction of CD27ϪCD4ϩ T cells and IL-4 production as a feature of CD4ϩ T within the CCR7ϪCD27Ϫ population. The decrease in CXCR5 cell differentiation in vivo. Clin. Exp. Immunol. 96: 528–534. 15. Rufer, N., A. Zippelius, P. Batard, M. Pittet, I. Kurth, P. Corthesy, J.-C. Cerottini, 24. Mackay, C. R., W. L. Marston, and L. Dudler. 1990. Naı¨ve and memory T cells S. Leyvraz, E. Roosne, M. Nabholz, and P. Romero. 2003. Ex vivo character- show distinct pathways of lymphocyte recirculation. J. Exp. Med. 171: 801–817. ization of human CD8ϩ T subsets with distinctive replicative history and partial 25. Michie, C. A., A. McLean, A. Alcock, and P. C. L. Beverley. 1992. Lifespan of effector functions. Blood 102: 1779–1787. human lymphocyte subsets defined by CD45 isoforms. Nature 360: 264–265. 16. Tomiyama, H., T. Matsuda, and M. Takiguchi. 2002. Differentiation of human 26. Tough, D. F., and J. Sprent. 1994. Turnover of naive- and memory-phenotype T CD8ϩ T cells from a memory to memory/effector phenotype. J. Immunol. 168: cells. J. Exp. Med. 179: 1127–1135. 5538–5550. 27. Macallan, D. C., D. Wallace, Y. Zhang, C. de Lara, A. Worth, H. Ghattas, 17. Hathcock, K. S., R. J. Hodes, and N. P. Weng. 2004. Analysis of telomere length G. Griffin, P. C. L. Beverley, and D. F. Tough. 2004. Rapid turnover of effector and telomerase activity. In Current Protocols in Immunology. J. E. Coligan, memory CD4 T cells. J. Exp. Med. 200: 255–260. A. M. Kruisbeck, D. H. Margulies, E. M. Shevach, and W. Strober, eds. Greene 28. Weng, N. P., L. D. Palmer, B. L. Levine, H. C. Lane, C. H. June, C. H., and Publishing Associates and Wiley-Interscience, New York. pp. 10.30.1–10.30.27. R. J. Hodes. 1997. Tales of tails: regulation of telomere length and telomerase 18. Langenkamp, A., K. Nagata, K. Murphy, L. Wu, A. Lanzavecchia, and activity during lymphocyte development, differentiation, activation, and aging. F. Sallusto. 2003. Kinetics and expression patterns chemokine receptors in human Immunol. Rev. 160: 43–54. CD4ϩ T lymphocytes primed by myeloid or plasmacytoid dendritic cells. Eur. 29. Buchkovich, K. J., and C. W. Greider. 1996. Telomerase regulation during entry J. Immunol. 33: 474–482. into the cell cycle in normal human T cells. Mol. Biol. Cell 7: 1443–1454. 30. Elson, L. H., S. Shaw, R. A. W. van Lier, and T. B. Nutman. 1994. T cell 19. Ma, C. S., P. D. Hodgkin, and S. G. Tangye. 2004. Automatic generation of subpopulation phenotypes in filarial infections: CD27 negativity defines a pop- lymphocyte heterogeneity: division dependent changes in the expression of ulation greatly enriched for Th2 cells. Int. Immunol. 6: 1003–1009. CD27, CCR7 and CD45 by activated human naive CD4 T cells are independently 31. Bird, J. J., D. R. Brown, A. C. Mullen, N. H. Moskowitz, M. A. Mahowald, regulated. Immunol. Cell Biol. 82: 67–74. J. R. Sider, T. F. Gajewski, C. R. Wang, and S. L. Reiner. 1998. Helper T cell 20. Nikolaeva, N. E., E. M. Uss, M. van Leeuwen, R. A. W. van Lier, and ϩ ϩ differentiation is controlled by the cell cycle. Immunity 9: 229–237. I. J. M. ten Berge. 2004. Differentiation of human alloreactive CD4 and CD8 32. Roncarolo, M. G., R. Bacchetta, C. Bordignon, S. Narula, and M. K. Levings. T cells in vitro. Transplantation 78: 815–824. 2001. Type 1 T regulatory cells. Immunol. Rev. 182: 68–79. 21. Vallejo, A. N., J. C. Brandes, C. M. Weyand, and J. J. Goronzy. 1999. Modu- 33. Nagasawa, T. 2002. A chemokine, SDF-1/PBSF, and its receptor, CXC chemo- lation of CD28 expression: distinct regulatory pathways during activation and kine receptor 4, as mediators of hematopoiesis. Int. J. Hematol. 72: 408–411. replicative senescence. J. Immunol. 162: 6572–6579. 34. Scimone, M. L., T. W. Felbinger, I. B. Mazo, J. V. Stein, U. H. von Andrian, and 22. Amyes, E., C. Hatton, D. Montamat-Sicotte, N. Gudgeon, A. B. Rickinson, W. Weninger. 2004. CXCL12 mediates CCR7-independent homing of central ϩ A. J. McMichael, and M. F. C. Callan. 2003. Characterization of the CD4 T cell memory cells, but not naı¨ve T cells in peripheral lymph nodes. J. Exp. Med. 199: response to Epstein-Barr virus during primary and persistent infection. J. Exp. 1113–1120. Med. 198: 903–911 35. Breitfeld, D., L. Ohl, E. Kremmer, J. Ellwart, J., F. Sallusto, M. Lipp, and 23. Weng, N. P., B. L. Levine, C. H. June, and R. J. Hodes. 1995. Human naive and R. Forster. 2000. Follicular B helper T cells express CXC chemokine receptor 5, memory T lymphocytes differ in telomeric length and replicative potential. Proc. localize to B cell follicles, and support immunoglobulin production. J. Exp. Med. Natl. Acad. Sci. USA 92: 11091–11094. 192: 1545–1552.

Chapter 3

Abnormal Differentiation of Memory T Cells in Systemic Lupus Erythematosus

RD Fritsch X Shen GG Illei CH Yarboro C Prussin KS Hathcock RJ Hodes PE Lipsky

National Institutes of Health, Bethesda, Maryland

Arthritis & Rheumatism. 2006;54(7):2184-2197.

29 Abnormal Differentiation of Memory T cells in Systemic Lupus Erythematosus

Ruth D. Fritsch, Xinglei Shen, Gabor G. Illei, Cheryl H. Yarboro, Calman Prussin, Karen S. Hathcock, Richard J. Hodes, and Peter E. Lipsky

Objective. The chemokine receptor CCR7 and the tumor necrosis factor receptor family member CD27 define 3 distinct, progressively more differentiated maturational stages of CD4 memory subpopulations in healthy individuals: the CCR7+,CD27+, the CCR7-, CD27+, and the CCR7-,CD27- populations. The goal of this study was to examine maturational disturbances in CD4 T cell differentiation in systemic lupus erythematosus (SLE), using these phenotypic markers. Methods. Phenotypic analysis by flow cytometry, in vitro stimulation experiments, telomere length measurement, and determination of inducible telomerase were carried out. Results. In SLE patients, significant increases of CCR7-,CD27- and CCR7-,CD27+ and a reduction of CCR7+,CD27+ CD4 memory T cells were found. In vitro stimulation of SLE T cells showed a stepwise differentiation from naive to CCR7+,CD27+ to CCR7- ,CD27+ to CCR7-,CD27-; telomere length and inducible telomerase decreased in these subsets in the same progressive sequence. The in vitro proliferative response of these populations progressively declined as their susceptibility to apoptosis increased. Interestingly, a significant reduction in inducible telomerase was noted in SLE naive and CCR7+,CD27+ CD4+ memory T cells. Additionally, SLE CCR7-,CD27+ and CCR7-, CD27- CD4 memory T cells proliferated poorly in response to in vitro stimulation and underwent significantly more apoptosis than their normal counterparts. Finally, expression of CXCR4 was significantly reduced in all SLE subsets compared with normal. Conclusion. Together these data indicate an increased degree of in vivo T cell stimulation in SLE, resulting in the accumulation of terminally differentiated memory T cells with a decreased proliferative capacity and an increased tendency to undergo apoptosis upon stimulation.

Systemic lupus erythematosus (SLE) is a activation (2,3,6), altered signaling capability prototypical autoimmune disease of the T cell receptor (2,3,7), and diminished characterized by a spectrum of organ interleukin-2 (IL-2) production (8). How manifestations (1). Although a wide array of these various abnormalities may influence T intrinsic T cell defects in SLE patients has cell function in vivo in SLE has not been been reported (for review, see refs. 2 and 3), determined. Therefore, we undertook an ex little is known about the maturational status vivo analysis of memory T cell subsets to of memory T cells and its possible assess the overall status of T cell function in implications regarding disease pathogenesis. SLE. An increase of CD45RO+ cells of the CD4+ Previously, we proposed a model of stepwise subset and a decrease of naive CD4 T cells in CD4 memory T cell differentiation (9), that patients with longstanding SLE have been expands the concept of central memory T observed (4,5), but the differentiation status cells (TCM) and effector memory T cells of the CD45RO+ memory subsets has not (TEM) (10). TCM and TEM are distinguished by been analyzed completely. A variety of the phenotypic expression of CCR7, enabling abnormalities in SLE T cells have been cells to home to secondary lymphoid organs reported, including a reduced threshold of through interaction with its ligands, CCL19

and CCL21 (11). TCM, expressing CCR7, are dose glucocorticoids (n=25; mean ±SEM believed to have only minimal cytokine- daily dosage 3.5±0.6 mg). Eight patients secreting capacity and to migrate (21.6%) did not receive any immuno preferentially to the T cell areas of lymphoid suppressive medication. Disease activity was organs, where they can be restimulated by determined by the SLE Disease Activity antigen. In contrast, CCR7-TEM are thought Index (SLEDAI) (21). Twenty-one patients to have acquired different migratory had inactive disease (SLEDAI score <4) and capabilities and effector functions, such as 16 patients had active SLE (SLEDAI score cytokine production or cytotoxicity, which ≥4). The mean±SEM SLEDAI score was they can exert after their migration to 3.9±1.6, and the mean ±SEM disease inflamed peripheral tissue (12–14). duration was 11.2±1.5 years. The control In addition to CCR7, we used the expression population consisted of 42 healthy of CD27, a member of the tumor necrosis individuals (9). Additionally, 11 patients factor (TNF) receptor family (15–19), to with chronic allergies, mostly manifested by delineate stages of T cell differentiation. allergic asthma (5 women and 6 men; mean CD27 is expressed by naive CD4 and CD8 T ±SEM age 41.7±3.1 years) and 5 female cells as well as by most memory T cells. The patients with RA (age 48.2±5 years) were loss of CD27 correlates with an increase of examined. Informed consent was obtained IL-4 production, and accumulation of these from all patients and controls. cells in patients with chronic allergic T cells or CD4+ T cells were negatively diseases, in patients with rheumatoid arthritis selected on a magnetic column using an (RA), and in aged donors has been reported AutoMACS (Miltenyi Biotec, Auburn, CA). (17,20). In some cases, CD4+ T cells were further The use of the combination of CCR7 and subjected to positive selection for CD45RO- CD27 as markers of the differentiation of expressing cells with the AutoMACS. The memory T cells led to the proposed model in purity of the resultant population was which CCR7+,CD27+ memory T cells analyzed by flow cytometry and was mature in a stepwise manner to CCR7-, generally >95%. CD27+ memory T cells and then to CCR7- ,CD27-memory T cells (9). This phenotypic Flow cytometry. differentiation was associated with alteration For phenotypic analysis, T cells were in proliferative activity, telomere length, incubated for 30 minutes at 4°C with a telomerase activity, and cytokine production. fluorescein isothiocyanate (FITC)–coupled We therefore used a similar approach in the monoclonal antibody (mAb) against CCR7 present study to assess the differentiation (clone 150503; R&D Systems, Minneapolis, status of CD4 memory T cells in patients MN), phycoerythrin (PE)–conjugated anti- with SLE. CD27 mAb (clone M-T271; BD PharMingen, San Diego, CA), peridin chlorophyll protein– Cy5.5 (PerCP–Cy5.5)– Materials and methods coupled anti-CD4 mAb (clone SK3; BD Biosciences, San Jose, CA), PerCP–Cy5.5– Patients and controls. coupled anti-CD8 mAb (clone SK1; BD Peripheral blood from 37 patients with SLE Biosciences), and allophycocyanin (APC)– (89% women; mean±SEM age 39±2 years) conjugated anti-CD45RO mAb (clone classified according to the revised criteria of UCHL-1; BD PharMingen). For chemokine the American College of Rheumatology (1) receptor analysis, CD4+, CD45RO+ T cells was obtained. Patients were treated with a were isolated with an AutoMACS and variety of drugs (hydroxychloroquine n=18, stained with the above-mentioned FITC–anti- cyclophosphamide n=6, mycophenolate CCR7, PE–anti- CD27, and PerCP–Cy5.5– mofetil n=3, cyclosporine n=3) and/or low- anti-CD4, as well as with APC-conjugated

mAb against CXCR3 (clone 1C6) and were activated again with PMA/ionomycin as CXCR4 (clone12G5) (both from BD described above. Phenotypic characterization Biosciences); mAb against CXCR5 and was performed immediately after the last CCR4 (clones RF8B2 and 1G1; BD stimulation and after an additional 9 days of PharMingen) were biotinylated and detected incubation in IL-2. For analysis of actual cell with streptavidin-coupled APC. For CD31 numbers, 5x104 cells from each population examination, cells were stained with FITC– tested were stimulated with PMA/ionomycin anti-CD31 (clone WM59; BD PharMingen) overnight and then cultured with 50 units/ml along with the abovementioned anti-CD4 and IL-2. Cell numbers were assessed by flow anti-CD45RO mAb. Antibodies of the cytometry using SPHERO beads appropriate IgG isotypes were used as (Spherotech, Libertyville, IL) after an negative controls. Cells were analyzed using additional 72 hours of culture. Dead cells a FACSCalibur flow cytometer (Becton were excluded by staining with 7– Dickinson, San Jose, CA). The data were actinomycin D (Molecular Probes, Eugene, analyzed with Flow Jo software (Tree Star, OR). The proliferative capacity is expressed Ashland, OR). as the ratio of cells recovered from the various subsets to cells recovered from the Cell separation by flow cytometry. CCR7+,CD27+ subset. Additionally, a For proliferation assays, assessment of kinetic analysis of proliferation of the CD4 telomere length, and telomerase activity, memory T cell subsets after PMA/ cells were sorted using a MoFlo flow ionomycin stimulation was carried out. cytometer (Dako Cytomation, Colorado Proliferation was measured by 3H-thymidine Springs, CO), FITC–anti-CCR7, PE–anti- incorporation and expressed as counts per CD27, PerCP–Cy5.5–anti-CD4, APC–anti- minute. Cell cycle progression of T cell CD45RO, and PE– Cy7–anti-CD3. Postsort subsets after stimulation was examined by populations routinely exhibited >95% purity. DNA staining with propidium iodide.

Cell culture and cell enumeration. Telomere length measurement. Because T cell receptor (TCR) signaling has Telomere length was measured in samples been shown to be disturbed in SLE (for from 7 of the SLE patients (ages 32–57 review, see refs. 2 and 3), we opted to use a years, mean ±SEM 42.3±3.1) by Flow– stimulation protocol that employed phorbol fluorescence in situ hybridization (FISH) myristate acetate (PMA; Sigma, St. Louis, (9,22). Isolated T cells were incubated in the MO) and ionomycin (Calbiochem, La Jolla, dark for 15 minutes in a hybridization CA) to avoid an influence of TCR signaling mixture either containing 0.3 µg/ml FITC- abnormalities in SLE T cells. After an initial labeled peptide nucleic acid telomere FISH overnight activation with 10 ng/ml PMA and probe (Applied Biosystems, Foster City, 1.34 µM ionomycin, sorted memory T cell CA) or without probe for autofluorescence subsets were maintained in IL-2 (50 units/ml; background measurement. Cells were then Biological Research Branch, Division of heated at 86°C for 15 minutes, hybridized Cancer Treatment and Diagnosis, National in the dark at room temperature for 1.5 Cancer Institute, Frederick, MD) in U- hours, and then gently washed and bottomed 96-well plates (Costar, Cambridge, resuspended. LDS 751 (Exciton, Dayton, MA). Culture medium consisted of Ultra OH), a DNA binding dye, was added to the Culture serum-free medium (BioWhittaker, resuspended cells (10 ng/ml), and Walkersville, MD). After 1 week, cells were fluorescence intensities of the telomere stimulated with anti-CD3 (64.1:1 µg/ml) and probe and LDS 751 were measured by flow cultured for another week, after which they cytometry using a FACSCalibur.

Figure 1. Percentage of CD4 memory T cell subsets in representative patients with systemic lupus erythematosus (SLE), allergy, and rheumatoid arthritis (RA), and a healthy control. CD4 memory T cells were stained for the expression of CCR7 and CD27, and 3 major populations, i.e., CCR7+,CD27+, CCR7-,CD27+, and CCR7-,CD27-, were assessed by comparison with a negative control (isotype control staining) and a positive control (naive T cells). PE = phycoerythrin; FITC = fluorescein isothiocyanate.

Telomerase activity. Naive CD4 T cells and sorted CD4 memory T cell subsets were stimulated overnight with 10 ng/ml PMA/1.34 µM ionomycin, followed by a 3-day culture in medium supplemented with IL-2. Telomerase activity was assayed using the TRAPeze Telomerase Detection Kit (Chemicon Serologicals, Temecula, CA) as described (9). Stimulated cells from each sample were suspended at the same density of live cells, and telomerase activity in serial dilutions

of cell lysates was detected by the ability Figure 2. Mean and SEM percentages of the to extend telomeric repeats on a substrate, CCR7+,CD27+, CCR7-, CD27+, and CCR7-, the TS primer. These telomerase products CD27- subsets among CD4 memory T cells in 42 were then amplified by 2-step polymerase normal controls, 5 patients with RA, 11 patients chain reaction (PCR) for 31 cycles, and run with allergy, and 37 patients with SLE (21 with inactive SLE and 16 with active SLE; no difference on a 12% polyacrylamide electrophoresis in the distribution between patients with active SLE gel. DNA bands were visualized using and those with inactive SLE was observed). SYBR Green (Molecular Probes) and a ✧=P<0.007; ✧✧=P<0.00003 versus normal PhosphorImager (Molecular Dynamics, controls. See Figure 1 for definitions. Sunnyvale, CA). A competitive internal standard was used to control for the efficiency of the PCR amplification. for 21.5±1.8%, whereas only 5.6±0.6% of CD4+ memory T cells were CCR7-,CD27- Statistical analysis. (Figures 1 and 2). Values are expressed as mean ± SEM. Similarly, in SLE patients, the CCR7+, Student’s 2-tailed t-test was used to CD27+ subtype was the most abundant determine statistical significance. Where subset in CD4 memory T cells (54.1±2.8%) applicable, nonparametric tests were (Figures 1 and 2). However, the frequency of performed (Wilcoxon rank test); for this subset in SLE patients was significantly correlation analysis, Spearman’s less than in normal controls (P<0.00003). correlation test was used. The CCR7-, CD27+ population accounted for 29.1±2.0% of CD4 memory T cells, whereas the frequency of CCR7-, CD27- Results cells among the CD4 memory T cells was 13.9±2.3%, both significantly greater than in Characterization of CD4 memory T cells healthy controls (P<0.007 and P<0.002, in patients with SLE. respectively). Patients with active SLE and Using naive CD45RO- CD4 T cells that those with inactive SLE showed a similar uniformly express CCR7 and CD27 as a distribution of these subsets (CCR7+,CD27+ positive control and isotype-matched control 54.4±4.4% in active SLE, 53.2±3.6% in mAb as a negative control, we identified 3 inactive SLE; CCR7-,CD27+ 30.4±2.5% in major subsets of CD4+ memory T cells in active SLE, 29.4±3.0% in inactive SLE; normal subjects: CCR7+, CD27+, CCR7-, CCR7-, CD27- 12.8±3.8% in active SLE, CD27+, and CCR7-,CD27-. As previously 14.2±3.3% in inactive SLE). No significant reported (9), the majority of CD4 memory T correlation (by Spearman’s rank test) could cells in normal controls expressed both be detected between the increase in CCR7 and CD27 (mean±SEM 69.8±2.1%). frequency of CCR7-, CD27- CD4 memory CCR7-, CD27+ memory T cells accounted cells and disease activity, disease duration,

age, ethnicity, or organ manifestations percentage of CCR7-, CD27- CD4 cells and (skin/mucosa, kidney, joint, central nervous the absolute CD4+, CD45RO+ cell counts system). (R2=0.198, P<0.02). Medication Although there was no correlation between (hydroxychloroquine, cyclophosphamide, the percentage of the CCR7-, CD27- subset glucocorticoids) did not influence the and absolute CD4 T cell numbers, there was distribution of these CD4 memory subsets. a positive but modest correlation between the

Figure 3. Percentage of CD8 memory T cell subsets in patients with SLE and healthy controls. CD8+, CD45RO+ T cells were stained for the expression of CCR7 and CD27, and 3 major populations, i.e., CCR7+,CD27+, CCR7-,CD27+, and CCR7-,CD27-, were assessed as described in Figure 1. A: Representative results from a SLE patient and a healthy control. B: Mean and SEM results in 13 controls and 23 patients with SLE (14 with inactive SLE and 9 with active SLE; no difference in the distribution between patients with active SLE and those with inactive SLE was observed). ✧=P<0.001 versus controls. See Figure 1 for definitions.

Figure 4. Phenotypic changes of CD4 memory T cell subsets from SLE patients after repeated stimulation. CD4+ memory T cells from SLE patients were sorted according to their phenotypic expression of CCR7 and CD27. Subsets were stimulated 3 times with anti-CD3 or phorbol myristate acetate/ionomycin over the course of 3 weeks, and their phenotype subsequently assessed. After a resting period of 9 days in medium supplemented with interleukin-2, cells were again stained for surface expression of CCR7 and CD27. The displayed pattern was seen in 7 SLE patients; 1 representative example is shown. APC = allophycocyanin (see Figure 1 for other definitions).

As a control to determine the impact of percentage in healthy individuals periodic immunostimulation on the (48.4±5.0%). Notably, a striking increase in distribution of the memory T cell subsets, we the CCR7-,CD27- population among studied 11 patients with allergy. The CD45RO+ CD8 cells was observed in SLE percentages of the CD4 memory cell subsets patients, with 18.6±3.3% manifesting this in these patients were similar to those phenotype, compared with 5.4±1.1% in observed in normal controls (Figures 1 and healthy individuals (P<0.001 by Student’s t- 2). As an additional disease control, we test) (Figure 3). As found in CD4 memory studied 5 RA patients. The percentages of the cells, patients with active SLE and those with memory cell subsets in RA patients were also inactive SLE did not differ significantly in similar to those observed in normal controls the distribution of memory subsets among (Figures 1 and 2). CD45RO+ CD8 cells (CCR7+,CD27+ The differential expression of CCR7 and 30.8±6.7% in active SLE, 36.1±5.8% in CD27 on CD8+,CD45RO+ T cells was also inactive SLE, CCR7-,CD27+ 53.3±5.9% in investigated. Compared with healthy active SLE, 42.4±5.0% in inactive SLE; individuals, the percentage of CCR7+, CCR7-, CD27- 15.7±5.3% in active SLE, CD27+ cells was slightly decreased in SLE 20.4±4.4% in inactive SLE). Similar to the patients (mean±SEM 34.1±4.3%, versus findings with CD4 T cells, there was no 42.5±5.6% in normal subjects), although the correlation between the frequency of CCR7- difference did not reach statistical ,CD27- CD8 memory cells and age, significance. The CCR7-,CD27+ population ethnicity, medication, organ manifestations, amounted to 46.6±3.9% of the CD45RO+, disease duration, disease activity, or absolute CD8 T cells in SLE patients, similar to the CD8+ and CD8+, CD45RO+ cell numbers.

Figure 5. Differences in survival and proliferation of CD4 memory T cell subsets in normal controls and systemic lupus erythematosus (SLE) patients. CD4+ (CD45RO-) naive and (CD45RObright) memory T cells from 3 SLE patients were sorted according to their expression of CCR7 and CD27. Viability and cell number of each subset 72 hours after stimulation with phorbol myristate acetate (PMA)/ionomycin were assessed using 7– actinomycin D (7-AAD), counting beads, and a flow cytometer. A: Typical example of the different proliferation profiles of the subsets. Cells (5x104) (dashed line) were stimulated and cell number was assessed as described above. All subpopulations except for CCR7-,CD27- memory T cells showed an increased number of both total and viable cells after stimulation with PMA/ ionomycin. B: Mean±SEM number of live cells (derived from 3 SLE patients) of the indicated subsets after the 72-hour incubation, normalized to the CCR7+,CD27+ cell count (i.e., CCR7+,CD27+ = 100%). ✧=P<0.002. C: Mean and SEM percentage of viable, 7-AAD–negative cells. ✩=P<0.05; ✩✩= P<0.006; ✧ = P<0.03 versus controls; ✧✧ = P<0.0003 versus controls. D: Kinetic analysis of proliferation of the memory subsets 48 hours and 72 hours after PMA/ionomycin stimulation. Proliferation was measured by 3H-thymidine incorporation. Results of 1 representative experiment, performed in triplicate, are shown. Each symbol represents 1 well (all from the same donor) (with some overlap of symbols); bars represent the mean. ✧=P<0.05; ✩=P<0.0005.

Phenotype of CD4+ memory T cells from additional resting period. The altered SLE patients after repeated stimulation. phenotype remained stable during the resting To examine the relationship between the period of 9 days. Finally, the CCR7-,CD27- CD4 memory T cell subsets, the effect of in subset retained its phenotype despite vitro stimulation on cell surface phenotype repeated stimulation and did not (re)express was assessed. Repeated stimulation altered either CCR7 or CD27. Importantly, the the phenotypes of CCR7+,CD27+ and results in SLE patients were not different CCR7-,CD27+ CD4 memory T cell subsets from those previously observed in normal (Figure 4): 21.1±11.4% (mean±SEM) of subjects (9). CCR7+,CD27+ CD4 memory T cells became CCR7-,CD27+, whereas 57.0±10.5% became Proliferative capacity and cell survival of negative for both surface markers. the CD4+ memory T cell subsets from Furthermore, 66.6±7.1% of the CCR7- SLE patients. ,CD27+ cell population became CCR7- As shown in Figure 5A, 3 days after in vitro ,CD27- after repeated stimulation. To stimulation, the number of cells detected examine the stability of the phenotypic increased in all populations except the change, cells were also assessed after an CCR7-,CD27- subset.

Figure 6. Cell cycle analysis of memory subsets ex vivo and after in vitro stimulation. A: Cell cycle of the respective memory T cell subsets was assessed immediately ex vivo, using propidium iodide (PI) staining. Subsets from 1 systemic lupus erythematosus (SLE) patient are shown. B: Analysis of memory subsets from 5 SLE patients ex vivo, showing a significantly higher proportion of CCR7-,CD27+ cells in the S/G2/M phase of the cell cycle in SLE patients than in normal controls (✧ =P <0.04 versus controls). The increased proportion of CCR7-,CD27- cells in the S/G2/M phase (4.9%) did not reach statistical significance. Values shown are the mean and SEM in SLE patients; values in controls (percentages) were as follows: G0/G1 naive cells 96.1±1.2, CCR7+,CD27+ subset 96.7±0.8, CCR7-,CD27+ subset 94.8± 1.6, CCR7-,CD27- subset 96.1±1.0; apoptotic (apo) naive cells 2.7±1.0, CCR7+,CD27+ subset 1.2±0.8, CCR7-, CD27+ subset 0.8±0.3, CCR7-,CD27- subset 0.8±0.3; S/G2/M naive cells 0.75±0.23, CCR7+,CD27+ subset 1.3±0.2, CCR7-,CD27+ subset 1.2±0.2, CCR7-,CD27- subset 2.0±0.7. In SLE patients, the difference in the percentage of cells in G0/G1 and S/G2/M between naive cells or cells of the CCR7+,CD27+ subset and cells of either the CCR7-,CD27+ or the CCR7-,CD27- subset was significant (both P<0.006). C: Analysis of memory subsets from 5 SLE patients after stimulation. The differences in the percentage of cells in G0/G1, S/G2/M, and apoptosis between naive cells or cells of the CCR7+,CD27+ subset and cells of either the CCR7-,CD27+ or the CCR7-,CD27- subset was significant (P<0.01). Except for a significantly higher proportion of apoptotic CCR7-,CD27+ cells in SLE patients than in controls (✧=P=0.05), responses of the cell subsets from SLE patients and controls were similar. Values shown are the mean and SEM in SLE patients; values in controls (percentages) were as follows: G0/G1 naive cells 68.1±6.2, CCR7+,CD27+ subset 63.2±3.7, CCR7-,CD27+ subset 38.7±4.7, CCR7-,CD27- subset 35.0±7.2; apoptotic naive cells 3.5±1.6, CCR7+,CD27+ subset 6.8±2.3, CCR7-,CD27+ subset 46.8±7.5, CCR7-,CD27- subset 66.5±9.6; S/G2/M naive cells 29.0±8.6, CCR7+,CD27+ subset 31.9±4.7, CCR7-,CD27+ subset 18.8±4.1, CCR7-,CD27- subset 9.1±2.2.

The proliferative capacity was highest in (P<0.002) (Figure 5B), which reflected the naive T cells and decreased in significantly (P<0.02) reduced absolute CCR7+,CD27+ cells (Figure 5A), further number of viable cells surviving in the declining in the CCR7-,CD27+ T cells. The cultures of SLE T cells (Figure 5C). The lowest cell recovery was noted in CCR7-, mean number of viable cells in cultures of CD27- memory T cells. SLE CCR7-,CD27- CD4 T cells declined Notably, the relative proliferative capacity of from a starting number of 5x104 per culture CCR7-,CD27- cells from SLE patients was well to 2.3±0.5x104, compared with significantly reduced compared with controls 5.3±1.0x104 in normal donor cell cultures.

Kinetic analysis of proliferation, measured ,CD27- at both time points. Although the by 3H-thymidine incorporation, showed differences in proliferation between subsets comparable results with incubation periods of were significant (P<0.05) at both time points, 48 and 72 hours. The subsets exhibited a there was a more pronounced difference after similar decrease in proliferation from 72 hours (Figure 5D). CCR7+, CD27+ to CCR7-,CD27+ to CCR7-

Figure 7. Telomere length, telomerase activity, and CD31 expression in CD4 T cell subsets. Telomere length was measured by Flow–fluorescence in situ hybridization in each of the sorted populations. A: Telomere lengths (in arbitrary units [AU]) of CD4 memory subsets from 7 systemic lupus erythematosus (SLE) patients. Each symbol (connected by dashed lines) represents an individual patient; bars represent the mean. Telomere length in subsets derived from SLE patients did not differ significantly from that in subsets derived from controls (control values [mean±SEM AU] naive cells 13,326±1,662, CCR7+,CD27+subset 10,807±1,273, CCR7-,CD27+subset 7,520±655, CCR7-,CD27-subset 6,552±625). In SLE patients, the telomere length of each subset was significantly different from the others (P<0.025). B: Mean±SEM inducible telomerase activity in SLE patients and controls. Telomerase activity in CD4 T cell subsets from SLE patients was significantly less than that in normal controls (✧=P<0.02 versus controls). C: CD31 expression in CD45RO-naive T cells from 15 controls and 14 age-matched SLE patients. No significant difference in expression of this marker by naive CD4 T cells from SLE patients versus controls was detected.

Additionally, cell cycle analysis was carried (P<0.009). The activity of the out. Immediately ex vivo, most cells from CCR7+,CD27+ subset was significantly each subset derived from normal subjects different from that of the naive and the were in G0/G1, and at most 2% (CCR7-, CCR7-,CD27+ populations (P<0.009) CD27-) were in G2/M (9). In cells obtained (Figure 7B), and as in normal donors, the from SLE patients, the percentage of cells in activity of the CCR7-,CD27+ subset was not G2/M was increased in the CCR7-,CD27+ significantly different from that of the subset (7.4%; P<0.04) and in the CCR7-, CCR7-,CD27- subset. CD27- subset (5.0%; P=0.1), whereas in naive and CCR7+,CD27+ subsets it was similar to that in normal subjects (Figure 6A). Upon stimulation, each of the subsets progressed into the cell cycle to varying degrees, as noted previously (9) (Figure 6C). It is notable that upon stimulation, SLE- derived CCR7-,CD27+ cells were significantly more prone to become apoptotic than were those from normal subjects (P=0.05).

Telomere lengths in CD4 memory T cell subsets from normal donors and SLE patients. As shown in Figure 7A, consistent differences in telomere length were observed among the populations. In 7 SLE patients (5 women and 2 men; mean age 42.3 years), naive T cells had significantly longer telomeres than any of the memory subsets and the order of telomere lengths in each subset was the same as in normal controls (Figure 7A). Comparison between SLE patients and normal donors (9) did not reveal significant differences in the average telomere length of any of the CD4 T cell populations.

Inducible telomerase activity in CD4

memory T cell subsets. Inducible telomerase activity in CD4+ T cell Figure 8. Characterization of chemokine receptor subsets from SLE patients in general expression in memory cell subsets from systemic followed the same trend as in normal donors lupus erythematosus (SLE) patients and normal (9), i.e., naive CD4+ T cells had the greatest controls. Subsets from 4–8 controls, 4 patients with active SLE, and 4 patients with inactive SLE were inducible telomerase activity, followed by examined. ✧=P< 0.05 versus controls; ✩=P< 0.05 memory CCR7+,CD27+ > CCR7-,CD27+ = between the respective subsets derived from controls; CCR7-,CD27- (Figure 7B). The activity O=P<0.05 between the respective subsets derived manifested by naive CD4+ T cells was from SLE patients. significantly different from that of either the CCR7-,CD27+ or the CCR7-,CD27- subset

In addition, the mean inducible telomerase cell subsets in either controls or SLE activity was significantly decreased in the patients, and there were no differences SLE naive CD4+ and CCR7+,CD27+ between the subsets derived from controls memory T cells compared with those from and those from SLE patients. normal donors (P<0.02) (Figure 7B). The comparison of inducible telomerase in naive CD4 T cells of 8 additional normal donors Discussion and 8 additional SLE patients confirmed these results (data not shown). Using the chemokine receptor CCR7 and the Since maturation of naive CD4+ T cells can TNF receptor family member CD27, we were be assessed by expression of CD31 (23) and able to assess abnormalities of CD4 memory a higher proportion of more mature, cell maturation in SLE. Whereas the “antigen-experienced” CD31-naive CD4 T distribution of CD4 memory T cell subsets in cells might lead to the reduced inducible patients with allergy or RA did not differ telomerase activity observed in SLE patients, significantly from normal, a distinct we examined CD31 expression of naive difference could be observed in the blood- CD4 T cells derived from SLE patients and derived CD4 memory T cells of SLE from normal controls. As seen in Figure 7C, patients. The proportion of CCR7+,CD27+ the proportion of CD31-bearing naïve CD4 T cells was significantly reduced, whereas a cells was not significantly different between significant increase of both the CCR7- SLE patients and age-matched controls ,CD27- and CCR7-,CD27+ subsets was (mean±SEM 65.5±3.3% and 72.0±1.9%, observed, independent of disease activity and respectively; P=0.1), and thus did not provide organ manifestation. Interestingly, a similar a basis for the observed decrease in increase in CCR7-,CD27- CD8+,CD45RO+ telomerase activity in naive SLE CD4 cells. cells was also observed (Figures 1 and 2), possibly indicating a similar mechanism in Differences in chemokine receptor CD8 T cells. Using somewhat different expression in CD4 memory T cell subsets. phenotypic markers, viral antigen–specific As shown in Figure 8, the expression of CD4 memory and CD8 T cells have been CXCR4 decreased from CCR7+,CD27+ to found to reside in different compartments CCR7-,CD27+ to CCR7-,CD27-, and defined by CD27 expression (24). The significantly lower expression of CXCR4 in current results indicate that in vivo immuno- all subsets was detected in cells from SLE stimulation in SLE patients results in the patients than in corresponding populations expansion of cells in both compartments, from controls. Similarly, a decrease in suggesting that a response to a single antigen CXCR5 expression from CCR7+,CD27+ to is unlikely to account for the results. CCR7-,CD27+ to CCR7-,CD27- was To investigate whether the abnormalities in observed. Notably, the expression of CXCR5 SLE were limited to the distribution of cells, was significantly higher in the CCR7-, we examined a number of functional CD27- subset from patients with active SLE activities. Similar to findings in cells from in comparison with the subset derived from normal individuals (9), in vitro experiments patients with inactive SLE (mean±SEM with SLE CD4 memory T cells demonstrated 34.1±1.6% and 19.5±1.9%, respectively; a stepwise conversion of CCR7+,CD27+ into P<0.03). CXCR3 expression did not differ a CCR7-,CD27+ population and the significantly among subsets in SLE patients. progression of the CCR7-,CD27+ into a Interestingly, the proportion of CXCR3- CCR7-,CD27-subset (Figure 4). Since the expressing CCR7-,CD27+ cells was higher in change in phenotype was not reversible after SLE patients than in normal controls a resting period, the decreased expression of (P<0.05). Finally, no differences were found both CCR7 and CD27 by SLE memory T in CCR4 expression among the memory T cells did not appear to be the result of recent

in vivo stimulation (18,25), but rather autoantigens as proposed remains to be reflected a change in the differentiation status determined, but clearly the current findings of memory T cells. should permit more precise analysis of these Examination of the proliferative capacity of perturbations. the memory subsets from SLE patients Additionally, a progressive decrease in revealed a significant decrease from naive to telomere length from naive CD4 T cells to CCR7+,CD27+ to CCR7-,CD27+ to CCR7-, the CCR7+,CD27+ CD4 memory T cell CD27-, similar to the findings in normal population and further to the CCR7-,CD27+ subjects (9) (Figure 5). Although the subset and the CCR7-,CD27- CD4 memory maturational relationship between the cells of SLE patients was found, comparable memory T cell subsets appeared to be with the shortening of telomere length of maintained in SLE patients, significant CD4 memory T cells in normal subjects (9). deficiencies in cell survival (Figures 5 and In contrast to the differences in proliferative 6) and proliferation of the CCR7-,CD27+ and capacity and apoptosis, however, in no subset CCR7-, CD27-subsets compared with their did the telomere length differ significantly counterparts from healthy controls were between normal subjects and SLE patients. apparent. Two major abnormalities were This is consistent with previous reports of observed. First, the proportion of CCR7-, normal telomere length observed in CD27+ in the S/G2/M phase immediately ex peripheral blood mononuclear cells (PBMCs) vivo was significantly increased in SLE from SLE patients, unseparated CD4 T cells patients, and second, the CCR7-subsets from older SLE patients (28,29), and freshly showed a higher inclination to undergo isolated CD4+ T cells from SLE patients apoptosis upon in vitro stimulation and a (30). Moreover, the similarity in telomere decreased proliferative capacity. These length between normal and SLE CD4 results clearly show that a higher proportion memory subsets indicates that this is unlikely of CCR7-,CD27+ CD4 memory T cells are to be the explanation for the increased cycling in vivo, suggesting an increased apoptosis of the SLE CCR7-,CD27- memory level of in vivo stimulation in these patients. T cell subset. Whether the increased likelihood of Telomerase, an enzyme that extends undergoing apoptosis is secondary to in vivo telomeres by synthesis of terminal telomeric stimulation or reflects the previously repeats, is activated in dividing cells and can reported dysregulation of cell death in SLE compensate for the telomere shortening that remains unclear (26,27). It is notable, occurs with cell division. Repeated in vitro however, that the enhanced tendency for stimulation as well as in vivo differentiation apoptosis was noted only in the CCR7-, from naive to memory cells has been linked CD27+ subset, suggesting that it is an to a decreased induction of telomerase, abnormality resulting from in vivo paralleled by telomere length shortening (31– differentiation and not a characteristic of all 33). Taken together, the reduced capacity for SLE memory T cells. The data do not permit cell proliferation and diminished induction a conclusion concerning whether there is an of telomerase activity in the CCR7-,CD27+ intrinsic abnormality in SLE T cells or, and CCR7-,CD27- compared with naive or alternatively, abnormalities in the external CCR7+,CD27+ memory cell subsets in influences on SLE T cells. However, it is normal subjects and SLE patients (9) are clear that there is an increased accumulation consistent with a history of more extensive of mature memory T cells in vivo in SLE cell replication and progression toward cell patients and specific abnormalities in senescence in these populations. individual subsets, especially those with a Notably, however, the level of inducible more differentiated phenotype. Whether telomerase activity in naive and these relate to a lower threshold of T cell CCR7+,CD27+ CD4+ memory T cells was activation (2,3,6) or greater exposure to reduced in CD4 T cell populations derived

from SLE patients compared with controls. telomerase activity. An increase in serum Since repeated in vitro stimulation has been TNF levels is characteristic of SLE (36), and associated with reduced inducible telomerase TNF is known to regulate telomerase (37). activity (for review, see ref. 32), the Whether chronic exposure to TNF alters decreased activity observed in SLE patients subsequent in vitro induction of telomerase is may therefore reflect a history of increased T unknown, but it could contribute to the cell activation and cell division in vivo. This abnormalities noted. possibility is consistent with reports of Another feature of maturation and increased telomerase activity in unstimulated differentiation of cells is the acquisition of PBMCs (28,29) and in freshly isolated certain functional properties such as cytokine unstimulated CD4+ and CD8+ T cells as well secretion. In normal subjects, we found a as CD19+ B cells from SLE patients (30). As clear pattern of cytokine secretion by the mentioned above, we found a greater CD4+ T cell subsets (9). It is notable that, percentage of CCR7-,CD27+ CD4 memory T whereas some of the SLE subsets differed in cells in the S/G2/M phase of the cell cycle in their ability to proliferate and responded to SLE patients compared with the same subset stimulation with increased apoptosis, relative derived from normal individuals, in cytokine production by all populations was accordance with this hypothesis. Therefore, comparable between SLE patients and the diminished inducible telomerase activity controls (data not shown). of the CD4 memory subsets could be a An additional indication of maturation in reflection of the increased in vivo activation. CD4 memory T cell subsets in normal In contrast, the reduced telomerase activity subjects is their progressively decreasing observed in naive CD4 T cells of SLE expression of CXCR4 and CXCR5 (9) patients was unlikely to be the result of (Figure 8). CXCR3 and CCR4 have been persistent stimulation. This finding was suggested to be markers for in vitro polarized unexpected, since there was no increase in Th1 and Th2 cells, respectively. In this study cycling naive T cells. Moreover, there was no we found no differential expression of either increase in “antigen-experienced” chemokine receptor in the defined peripherally expanded naive CD4 T cells maturational stages of memory T cells in characterized by the lack of CD31 expression normal individuals. These findings are (23,34) in the SLE patients examined. There consistent with the conclusion that CXCR3 are a number of alternative possible and CCR4, expressed in the respective explanations for the diminished inducible polarized cells, are also present on less telomerase activity noted in SLE CD4 T committed and less differentiated cells (38). cells. First, there could be a genetic Furthermore, these results are in accordance abnormality in telomerase expression causing with previous reports that expression of reduced functionality of this enzyme. these cell surface markers does not represent Consistent with this, a newly described SLE a reliable tool to distinguish in vivo– susceptibility locus comprises the generated Th1 from Th2 cells (39). chromosomal region of the hTERT gene, the Comparable with the situation in normal enzymatically active component of controls, an inverse correlation of CXCR4 telomerase (35). It is unknown whether the expression and maturational stage of memory polymorphism alters the expression of this T cells was observed in SLE patients. enzyme. The uniform decrease in telomerase However, CXCR4 expression was reduced in expression in SLE patients compared with all memory T cell subsets from SLE patients. healthy controls makes this an unlikely Since CXCR4 expression is down-regulated explanation, but it warrants further and CXCR3 is up-regulated upon TCR examination. engagement (25,40), the reduced level of Second, the different cytokine milieu in CXCR4 and increased expression of CXCR3 patients with SLE might also alter inducible in the CCR7-,CD27+ SLE subset may again

reflect in vivo activation. Interestingly, an [review]. Trends Immunol increase of the CXCR4 ligand stromal cell– 2003;24:259–63. derived factor 1 (SDF-1) in the blood of SLE 3) Kong PL, Odegard JM, Bouzahzah patients has been reported (41). Since F, Choi JY, Eardley LD, Zielinski chemokine receptor engagement is known to CE, et al. Intrinsic T cell defects in down-regulate receptor expression (42), systemic autoimmunity [review]. Ann chronic exposure to SDF-1 is another N Y Acad Sci 2003;987:60–7. potential explanation for this finding. 4) Gordon C, Matthews N, Schlesinger In summary, the current data clearly indicate BC, Akbar AN, Bacon PA, Emery P, a stepwise conversion of CCR7+,CD27+ T et al. Active systemic lupus cells via the stage of CCR7-,CD27+ to a erythematosus is associated with the terminally differentiated CCR7-,CD27-state recruitment of naive/resting T cells. in SLE patients as well as in normal subjects. Br J Rheumatol 1996;35:226–30. However, the distribution of subsets was 5) Neidhart M, Pataki F, Michel BA, altered in SLE, with a significant increase in Fehr K. CD45 isoforms expression CCR7-,CD27+ and CCR7-,CD27- CD4 on CD4+ and CD8+ peripheral blood memory T cells. Moreover, SLE CCR7-, T-lymphocytes is related to auto- CD27+ T cells were cycling to a higher immune processes and hematological degree in vivo, and manifested a greater manifestations in systemic lupus inclination to undergo apoptosis after in vitro erythematosus. Schweiz Med stimulation. Together, these results provide Wochenschr 1996;126:1922–5. insight into the status of memory T cell 6) Vassilopoulos D, Kovacs B, Tsokos activation in SLE and demonstrate that GC. TCR/CD3 complex-mediated memory T cell differentiation is significantly signal transduction pathway in T cells heightened in this disease, presumably and T cell lines from patients with because of chronic in vivo stimulation. systemic lupus erythematosus. J Immunol 1995;155:2269–81. 7) Krishnan S, Farber DL, Tsokos GC. T Acknowledgments cell rewiring in differentiation and disease. J Immunol 2003;171:3325– The authors wish to thank James Simone, 31. Derek Hewgill, and Liusheng He of the 8) Katsiari CG, Kyttaris VC, Juang YT, Flow Cytometry Facility of the National Tsokos GC. Protein phosphatase 2A is Institute of Arthritis and Musculoskeletal and a negative regulator of IL-2 production Skin Diseases Office of Science and in patients with systemic lupus Technology for carrying out the cell sorting erythematosus. J Clin Invest 2005;115: 3193–204. 9) Fritsch RF, Shen X, Hathcock KS, References Hodes RJ, Lipsky PE. Stepwise differentiation of CD4 memory T 1) Tan EM, Cohen AS, Fries JF, Masi cells defined by expression of CCR7 AT, McShane DJ, Rothfield NF, et al. and CD27. J Immunol 2005;175:6489– The 1982 revised criteria for the 97. classification of systemic lupus 10) Sallusto F, Lenig D, Forster R, Lipp M, erythematosus. Arthritis Rheum Lanzavecchia A. Two subsets 1982;25:1271–7. ofmemory T lymphocytes with distinc 2) Tsokos GC, Nambiar MP, Tenbrock homing and effector functions. Nature K, Yuang-Taung J. Rewiring the T- 1999;401:708–12. cell: signaling defects and novel 11) Baekkevold ES, Yamanaka T, prospects for the treatment of SLE Palframan RT, Carlsen HS, Reinholt

FP, von Andrian UH, et al. The CCR7 van Lier RA.The CD27-subset of ligand elc (CCL19) is transcytosed in peripheral blood memory CD4+ high endothelial venules and mediates lymphocytes contains functionally T cell recruitment. J Exp Med differentiated T lymphocytes that 2001;193:1105–12. develop by persistent antigenic 12) Tanabe S, Lu Z, Luo Y, Quackenbush stimulation in vivo. Eur J Immunol EJ, Berman MA, Collins- Racie LA, et 1992;22:993–9. al. Identification of a new mouse - 20) Nijhuis EW, Remarque EB, Hinloopen chemokine, thymus-derived B, van der Pouw-Kraan T, van Lier chemotactic agent 4, with activity on T RA, Ligthart GJ, et al. Age related lymphocytes and mesangial cells. J increase in the fraction of CD27-CD4+ Immunol 1997;159:5671–9. T cells and IL-4 production as a feature 13) Gunn MD, Tangemann K, Tam C, of CD4+ T cell differentiation in vivo. Cyster JG, Rosen SD, Williams LT. A Clin Exp Immunol 1994;96:528–34. chemokine expressed in lymphoid 21) Bombardier C, Gladman DD, Urowitz high endothelial venules promotes the MB, Caron D, Chang CH, and the adhesion and chemotaxis of naive T Committee on Prognosis Studies in cells. Proc Natl Acad Sci U S A SLE. Derivation of the SLEDAI: a 1998;95:258–63. disease activity index for lupus 14) Masopust D, Vezys V, Marzo AL, patients. Arthritis Rheum 1992;35:630– Lefrancois L. Preferential localization 40. of effector memory cells in 22) Hathcock KS, Hodes RJ, Weng NP. nonlymphoid tissue. Science Analysis of telomere length and 2001;291:2413–7. telomerase activity. In: Coligan JE, 15) Van Lier RAW, Borst J, Vroom TM, Kruisbeck AM, Margulies DH, Klein H, van Mourik P, Zeijlemaker Shevach EM, Strober W, editors. WP, et al. Tissue distribution and Current protocols in immunology. biochemical and functional properties New York: Greene Publishing of Tp55 (CD27), a novel T cell Associates and Wiley. Interscience; differentiation antigen. J Immunol 2004. p. 10.30.1–27. 1987;139:1589–96. 23) Kimming S, Przybylski GK, Schmidt 16) Hintzen RQ, de Jong R, Lens SM, CA, Laurisch K, Mowes B, Radbruch Brouwer M, Baars P, van Lier RA. A, et al. Two subsets of naive T helper Regulation of CD27 expression on cells with a distinct T cell receptor subsets of mature T-lymphocytes. J excision circle content in human adult Immunol 1993;151:2426–30. peripheral blood. J Exp Med 17) Tortorella C, Schulze-Koops H, 2002;195:789–794. Thomas R, Splawski JB, Davis LS, 24) Amyes E, Hatton C, Montamat-Sicotte Picker LJ, et al. Expression of D, Gudgeon N, Rickinson AB, CD45RB and CD27 identifies subsets McMichael AJ, et al. Characterization of CD4+ memory T cells with of the CD4+ T cell response to different capacities to induce B cell Epstein-Barr virus during primary and differentiation. J Immunol 1995;155: persistent infection. J Exp Med 149–62. 2003;198:903–911. 18) Hintzen RQ, Jong J, Lens SM, van Lier 25) Langenkamp A, Nagata K, Murphy K, RA. CD27: marker and mediator of T- Wu L, Lanzavecchia A, Sallusto F. cell activation. Immunol Today 1994; Kinetics and expression patterns of 15:307–11. chemokine receptors in human CD4+ 19) De Jong R, Brouwer M, Hooibrink B, T lymphocytes primed by myeloid or van der Pouw-Kraan T, Miedema F,

plasmacytoid dendritic cells. Eur J activation. J Immunol 1997;158:3215– Immunol 2003;33:474–82. 20. 26) Emlen W, Niebur JA, Adera R. 34) Thiel A, Schmitz J, Miltenyi S, Accelerated in vitro apoptosis of Radbruch A. CD45RA-expressing lymphocytes from patients with memory/effector Th cells committed to systemic lupus erythematosus. J production of interferon lack Immunol 1994;152:3685–92. expression of CD31. Immunol Lett 27) Silvestris F, Grinello D, Tucci M, 1997;57:189–92. Cafforio P, Dammaco F. Enhancement 35) Nath SK, Namjou B, Garriott CP, of T cell apoptosis correlates with Frank S, Joslin PA, Kilpatrick J, et al. increased serum levels of soluble Fas Linkage analysis of SLE susceptibility: (CD95-Apo-I) in active lupus. Lupus confirmation of SLER1 at 5p15.3. 2003;12:8–14. Genes Immun 2004;5:209–14. 28) Kurosaka D, Yasuda J, Yoshida K, 36) Aringer M, Stummvoll GH, Steiner G, Yokoyama T, Ozawa Y, Obayashi Y, Koller M, Steiner CW, Hofler E, et al. et al. Telomerase activity and telomere Serum interleukin-15 is elevated in length of peripheral blood mononuclear systemic lupus erythematosus. cells in SLE patients. Lupus 2003;12: Rheumatology (Oxford) 2001;40:876– 591–9. 81. 29) Honda M, Mengesha E, Albano S, 37) Akiyama M, Yamada O, Hideshima T, Nichols WS, Wallace DJ, Metzger A, Yanagisawa T, Yokoi K, Fujisawa K, et al. Telomere shortening and et al. TNFa induces rapid activation decreased replicative potential, and nuclear translocation of telomerase contrasted by continued proliferation of in human lymphocytes. Biochem Bio telomerase-positive CD8+CD28 (lo) T phys Res Commun 2004;316:528–32. cells in patients with systemic lupus 38) Rivino L, Messi M, Jarrossay D, erythematosus. Clin Immunol 2001; Lanzavecchia A, Sallusto F, Geginat J. 99:211–21. Chemokine receptor expression 30) Klapper W, Moosig F, Sotnikova A, identifies Pre-T helper (Th)1, Pre-Th2, Qian W, Schroder JO, Parwaresch R. and nonpolarized cells among human Telomerase activity in B and T CD4+ central memory T cells. J Exp lymphocytes of patients with systemic Med 2004;200:725–35. lupus erythematosus. Ann Rheum Dis 39) Nanki T, Lipsky PE. Lack of 2004;63:1681–3. correlation between chemokine 31) Weng NP, Palmer LD, Levine BL, receptor and T(h)1/T(h)2 cytokine Lane HC, June CH, Hodes RJ. Tales of expression by individual memory T tails: regulation of telomere length and cells. Int Immunol 2000;12:1659–67. telomerase activity during lymphocyte 40) Bermejo M, Martin-Serrano J, Oberlin development, differentiation, E, Pedraza MA, Serrano A, Santiago activation, and aging. Immunol Rev B, et al. Activation of blood T 1997;160:43–54. lymphocytes downregulates CXCR4 32) Buchkovich KB, Greider CW. expression and interferes with Telomerase regulation during entry into propagation of X4 HIV strains. Eur J the cell cycle in normal human T cells. Immunol 1998;28:3192–204. Mol Biol Cell 1996;7:1443–54. 41) Eriksson C, Eneslatt K, Ivanoff J, 33) Weng N, Levine BL, June CH, Hodes RantapaaDahlqvist S, Sundqvist KG. RJ. Regulation of telomerase RNA Abnormal expression of chemokine template expression in human T receptors on T-cells from patients with lymphocyte development and systemic lupus erythematosus. Lupus 2003;12:766–74.

42) Signoret N, Oldrige J, Pelchen- Matthews A, Klasse PJ, Tran T, Brass LF, et al. Phorbol esters and SDF-1 induce rapid endocytosis and down- modulation of the chemokine receptor CXCR4. J Cell Biol 1997;139:651– 664.

Chapter 4

Characterisation of Cellular and Humoral Autoimmune Responses to Histone H1 and Core Histones in Human Systemic Lupus Erythaematosus

GH Stummvoll1 RD Fritsch2 B Meyer1 E Hoefler3 M Aringer1 JS Smolen1,3,4 G Steiner1,4

1Division of Rheumatology,Internal Medicine III, Medical University of Vienna, Vienna,Austria 2Department of Rheumatology and Clinical Immunology, University Medical Center Utrecht, Utrecht, The Netherlands 3Second Department of Medicine, Hietzing Hospital, Vienna, Austria 4Ludwig Boltzmann Institute for Rheumatology and Balneology, Rheumatology Cluster of the Ludwig Boltzmann Society, Vienna, Austria

Annals of the Rheumatic Diseases. 2009;68;110-116.

49 Characterisation of cellular and humoral autoimmune responses to histone H1 and core histones in human systemic lupus erythaematosus G H Stummvoll,1 R D Fritsch,2 B Meyer,1 E Hoefler,3 M Aringer,1 J S Smolen,1,3,4 G Steiner1,4 c Supplementary tables 1 and 2 ABSTRACT responses, especially in drug-induced lupus and supplementary figs 1–3 are Objective: To address key aspects of anti-histone erythaematosus (LE).56 Anti-histone Abs can be published online only at http:// autoimmunity in systemic lupus erythaematosus (SLE), ard.bmj.com/content/vol68/ found in a variety of diseases, including inflamma- issue1 we performed a detailed characterisation of cellular and tory hepatic, malignant, infectious and, of course, humoral autoreactivity to histone H1 and the four core rheumatic autoimmune diseases.7–9 They may be 1 Division of Rheumatology, Internal Medicine III, Medical histones H2A, H2B, H3, H4 in patients with SLE and present in up to 75% of SLE sera, while their University of Vienna, Vienna, healthy controls. prevalence in other rheumatic diseases usually does Austria; 2 Department of Methods: Peripheral blood mononuclear cells of 41 not exceed 20%.10 11 Like anti-dsDNA Abs, anti- Rheumatology and Clinical patients with SLE and 28 healthy controls were exposed histone Abs also correlate with SLE disease activ- Immunology, University Medical 12 13 Center Utrecht, Utrecht, The to individual histones and proliferation was measured by ity, particularly Abs to histone H1, which appear 3 Netherlands; 3 Second [ H]-thymidine incorporation. H1-reactive T cell clones to be more specific for SLE and may better correlate Department of Medicine, were obtained by limiting dilution. Cytokines and total IgG with disease activity than Abs to core histones.10 Of Hietzing Hospital, Vienna, in culture supernatants was measured by ELISA, and 4 note, it is anti-H1 Abs that are responsible for the LE Austria; Ludwig Boltzmann autoantibodies to histones were determined by ELISA and cell phenomenon, which was originally included Institute for Rheumatology and Balneology, Rheumatology immunoblotting. into the American College of Rheumatology (ACR) Cluster of the Ludwig Boltzmann Results: Proliferative responses to H1 were more classification criteria for SLE.14 15 Society, Vienna, Austria frequent and more pronounced in cell cultures from AutoAbs to chromatin components show immu- patients with SLE (p,0.002), while among the core noglobulin class switching and affinity maturation, Correspondence to: G Steiner, Division of histones only the response to H2A was increased in which are typical features of a T cell-dependent Rheumatology, Internal Medicine patient cultures (p,0.01). All histones elicited a Th1-like Ag-driven immune response.16–18 Histones are con- III, Medical University of Vienna, cytokine response in patients and controls (high interferon sidered the major T cell antigens (Ags) in SLE for Wa¨hringer Gu¨rtel 18, A-1090 Vienna, Austria; guenter. (IFN)c and tumour necrosis factor (TNF)a, no interleukin the generation of autoimmune responses against [email protected] (IL)4) with H1 inducing the highest levels of TNFa. chromatin components: thus, histone-specific T cell However, H1 stimulated production of IgG and anti- clones augmented the production of anti-dsDNA, GHS and RDF contributed histone antibodies only in cell cultures derived from anti-single-stranded (ss) DNA and anti-histone Abs equally to this work patients with SLE. H1-specific T cell clones from patients in murine models of SLE19 and also promoted anti- Accepted 9 March 2008 and controls showed a CD4+CD28+ phenotype and a Th1 dsDNA production by autologous B cells in human Published Online First cytokine profile. Anti-histone antibodies were detected in SLE;320however, details of the histone-specific T cell 28 March 2008 51% of patients with SLE, were primarily directed to H1, response in human SLE have not been fully H3 and H4, and predominantly of the IgG2 subtype. elucidated. In particular, it is not clear which histone Conclusions: Histone H1 constitutes a major B cell and constitutes the major autoAg in T cell activation. T cell autoantigen in SLE, triggering a proinflammatory Th1 Considering current Ag-specific therapeutic response and driving autoantibody production. This approaches,21 and in order to specify the impact suggests that histone H1 may be of considerable of individual histones in T cell activation, we relevance for the pathogenesis of SLE. performed a detailed analysis of the cellular and humoral responses against histone H1 and the core histones H2A, H2B, H3 and H4 in patients with Antinuclear autoantibodies are a hallmark of sys- SLE and healthy controls. The results suggest that temic lupus erythaematosus (SLE). The vast major- histone H1 may indeed be the most important ity of patients with SLE develop autoantibodies single histone autoAg at the B and T cell level. (autoAbs) against chromatin components that may be directed against single structures or against the complete nucleosome. Among these autoAbs, IgG PATIENTS AND METHODS Abs to double-stranded (ds)DNA are a pathogno- Patients and controls monic feature of SLE and can be directly pathogenic.1 A total of 41 patients with SLE fulfilling the ACR However, dsDNA may not be the first target of the criteria for SLE14 and 28 healthy individuals were autoAb response; rather, autoAbs to other chroma- analysed; demographic details of patients and tin components, and to histones in particular, may controls are given in the supplementary material. precede anti-dsDNA Ab formation.2 In the course of routine blood testing, 40 ml of Anti-histone Abs can be directed to single peripheral blood was drawn after obtaining histones, to histone–DNA complexes, or to com- informed consent from each patient. Disease plexes of histones.34 In particular, histone H2A– activity was measured using the SLE disease H2B dimers frequently elicit humoral antihistone activity index (SLEDAI), the European Consensus lupus activity measurement (ECLAM) and the SLE activity monoclonal Ab (R&D Systems, Minneapolis, Minnesota, index score (SIS).22–25 USA). In some experiments the autoAg hnRNP-A2 was used as an unrelated basic control-Ag (pI = 9.3) at 0.35 mg/ml as 29 T cell stimulation assays described (see supplementary material). Peripheral blood mononuclear cells (PBMCs) were cultured for 5 days in triplicate in the presence of Ags (105 cells/well) as Detection of cytokines described previously.26 During the last 16 h of culture, 0.5 mCi/ ELISA assays for IFNc and IL4 were obtained from Endogen well tritiated thymidine ([3H]-TdR; Amersham Biosciences, (Woburn, Massachusetts, USA), the ELISA for tumour necrosis Freiburg, Germany) was added, and the incorporated radio- factor (TNF)a was from Biosource (Nivelles, Belgium). activity was measured by scintillation counting and expressed Detection limits were 9 pg/ml for interferon (IFN)c, 4 pg/ml as counts per min (cpm). Results are given as stimulation index for interleukin (IL)4 and 15 pg/ml for TNFa, respectively. If not (SI) defined by the ratio of (mean cpm obtained in cultures stated otherwise, cytokines were measured in supernatant (SN) incubated with Ag):(mean cpm obtained in cultures incubated after 24 h of incubation as described.26 30 in the absence of Ag). An SI >2 was regarded a positive D response, if, at the same time, the cpm ((mean cpm obtained T cell lines and clones in cultures with Ag)–(mean cpm obtained in cultures without H1-specific T cell lines were generated using a previously Ag)) was .1000 cpm.32627 established protocol.26 30 31 In brief, 26106 PBMCs were stimulated with histone H1 for 5 days in 24-well plates. On day 5, Antigens 20 U/ml IL2 was added and the culture was continued for Highly purified calf thymus histones H1, H2A, H2B, H3, H4 7 days. To generate T cell clones (TCCs), T cell lines were (Roche Diagnostic GmbH, Mannheim, Germany) were used restimulated with histone H1 and, after 2 more days, viable throughout the study. Purity was confirmed by sodium dodecyl T cell blasts were separated and seeded by limiting dilution sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and (0.5 cells/well) in 96 well plates together with 16105 irradiated Coomassie Blue staining. For each SLE patient and healthy allogeneic PBMCs as feeder cells. Growing microcultures were control (HC), the individual reactivity of PBMCs to histone Ags then expanded at weekly intervals in the presence of feeder cells was tested at four different concentrations (ie, 0.7, 1.25, 2.5 and and IL2. 5 mg/ml). The concentration yielding the highest SI was used in The specificity of TCCs was assessed in proliferation assays subsequent experiments (see supplementary material). In the (as detailed above): 26104 T cell blasts were stimulated with H1, majority of patients and HCs 2.5 or 5 mg/ml were found to be H2A, H2B, H4 and TT Ags, respectively, in the presence of 105 optimum concentrations. The control Ag tetanus toxoid (TT) autologous irradiated PBMCs. For surface phenotyping, fluor- was obtained from Pasteur Merieux Connaught (Willowdale, escence-labelled monoclonal Abs against CD3, CD4, CD8, Ontario, Canada) and used at a concentration of 0.5 U/ml as CD28, TCRa/b, TCRc/d (BD Biosciences, Palo Alto, described previously.26 28 When indicated, PBMC cultures were California, USA) were used. Isotype-matched Abs to irrelevant also stimulated with 1 mg/ml pokeweed mitogen (PWM; Sigma Ags (Dako, Glostrup, Denmark, or BD Biosciences) served as Aldrich, Vienna, Austria) and/or with 1 mg/ml anti-CD3 negative controls.

Table 1 Cellular proliferation and cytokine production of peripheral blood mononuclear cells (PBMCs) in response to histone antigens H1 H2A H2B H3 H4 TT

Proliferation: SLE (n = 41): Positive response 44%* 50% 60% 26% 23% 66% SI 2.3 (1.5)* 2.9 (2.4)* 2.8 (1.7) 1.9 (1.7) 1.5 (0.8) 4.9 (5.3) Counts/min (bgd 1535 (1326)) 7466 (2815) 3746 (3240) 4549 (4745) 4133 (7811) 2556 (2571) 5951 (7609) Correlation with H1 response p,0.001, r = 0.82 p = 0.01, r = 0.59 p = NS, r = NS p,0.01, r = 0.53 p,0.05, r = 0.41 HC (n = 28): Positive response 11%* 26% 37% 26% 21% 85% SI 1.5 (0.4)* 1.5 (0.8)* 2.2 (1.6) 1.7 (1.2) 1.4 (0.7) 4.7 (4.3) Counts/min (bgd 2842 (2441)) 4161 (3413) 4536 (4338) 5810 (4537) 4071 (2887) 4030 (3371) 12590 (9478) Correlation with H1 response p = NS, r = NS p = NS, r = NS p = NS, r = NS p,0.05, r = 0.51 p = NS, r = NS Cytokine production: IFNc (pg/ml, mean (SEM)): SLE 69 (50) 20 (7) 56 (25) 17 (9) 41 (25) NP HC 66 (25) 50 (38) 40 (0.3) 11 (6) 39 (35) NP SLE+HC 68 (33) 25 (9), p = NS{ 53 (20), p = NS{ 16 (7), p = NS{ 41 (20), p = NS{ TNFa (pg/ml, mean (SEM)): SLE 630 (135) 259 (108) 526 (141) 190 (67) 276 (109) NP HC 912 (192) 263 (31) 457 (75) 248 (107) 339 (151) NP SLE+HC 734 (112) 260 (80), p,0.01{ 510 (109), p = NS{ 200 (54), p,0.001{ 283 (88), p,0.01{ In patients with SLE cellular reactivity to H1 and H2A was significantly elevated and H1 responses correlated with responses to H2A, H2B and H4. Similar amounts of IFNc and TNFa were secreted by PBMCs of patients with SLE and HC but highest TNFa levels were observed after stimulation with H1. Values are given as mean (SD) unless stated otherwise. *Denotes a significant difference (p,0.01) in cellular reactivity between patients with SLE and HC; {denotes a significant difference vs H1-induced IFNc and TNFa production. Ag, antigen; bgd, background proliferation without Ag; HC, healthy control; IFN, interferon; SEM, standard error of the mean; SI, stimulation index; SLE, systemic lupus erythaematosus; TNF, tumour necrosis factor; TT, tetanus toxoid; NP, not performed; NS, not significant. Levels of total anti-histone and anti-chromatin Abs in serum and culture SN were measured by ELISA (Inova Diagnostics, San Diego, California, USA). Sensitivity and specificity of both ELISAs was determined by measuring Ab levels in 420 sera from our serum bank (see supplementary material). Results are given as arbitrary units/ml (U/ml) employing the reference sera provided by the manufacturer. Since sera were diluted 1:100 and culture SN 1:10, 10 U/ml in SN would correspond to 1 U/ml in serum. Anti-dsDNA Abs were measured by radioimmunoassay (Ortho-Clinical Diagnostics Inc., Rochester, New York, USA), total IgG in culture SN was measured by ELISA (Alpco Diagnostics, Windham, New York, USA).

T cell/B cell interactions and formation of anti-histone autoantibodies PBMCs (105 cells/well) obtained from patients with SLE, rheumatoid arthritis (RA) and HCs were cultured in the presence or absence of histone H1 as described above. Unspecific polyclonal stimuli such as PWM and anti-CD3 monoclonal Ab served as positive controls. After 7 days of culture,33 the SN were analysed for total IgG and anti-histone Abs by ELISA.

Statistical analysis All group results are expressed as mean (SD) unless stated otherwise. The Student t test, or Fisher exact test (two-tailed) Figure 1 Proliferative responses to histone H1 and core histones in were used for the comparison of group values and discriminatory patients with systemic lupus erythaematosus (SLE) and healthy controls parameters, where appropriate. If logarithmic transformation (HCs). Peripheral blood mononuclear cells (PBMCs) from 41 patients resulted in Gaussian distributions, such transformation was with SLE and 28 HCs were cultured in the presence or absence of performed. The Welch correction was applied if variances were 3 histone antigens (Ags) and proliferation was measured by [ H]thymidine significantly different. Where appropriate, p values corrected by incorporation. A positive response was defined as a stimulation index the Bonferroni method are given. Pearson and Spearman (SI).2 and D counts/min (cpm).1000 cpm. Results obtained at the correlation coefficients were calculated for variables following optimum histone concentration are shown, determined for each histone in all patients and HCs as detailed in the Patients and methods section; and not following Gaussian distributions, respectively. typical dose–response curves are shown in the supplementary material. For patients with SLE, the proliferative response ranged from 1535 RESULTS (1326) cpm (mean (SD)) in the absence of any antigen to 7466 (2815) cpm after H1 stimulation to 21 540 (22 830) cpm after polyclonal Increased proliferative response to histones H1 and H2A in stimulation with phytohaemagglutinin (PHA). (A) Response to histone patients with SLE H1. A positive response was seen in 44% of patients with SLE as Among 41 patients with SLE analysed for cellular reactivity to compared to 11% of HCs (p,0.01) and the strength of the proliferative single histones, 18 (44%) showed a distinct T cell response (ie, response was significantly higher in SLE cultures (mean SI (SD)): 2.3 SI.2) to histone H1, as compared to only 3 responders (11%) (1.5) vs 1.5 (0.4), p,0.002). (B) Response to core histones. Frequencies among HCs (p,0.01, fig 1A), and responses of patients with of responders did not differ between SLE (black triangles) and HCs (white SLE were significantly higher than those of healthy individuals triangles) and a significant difference in the mean stimulation index (SI) (p,0.002, table 1). was observed only for H2A (mean SI: 2.9 (2.4) vs 1.5 (0.8), p,0.01). All four core histones elicited responses in PBMCs of patients TT, tetanus foxoid. with SLE and HCs, but the frequency of responders was similar in both groups (fig 1B) and significant group differences between patients and HCs were seen only for H2A (p,0.01, Detection of autoantibodies table 1). Interestingly, in patients with SLE, but not in controls, Serological reactivities to single histones were determined by H1 responses correlated significantly with responses to the core immunoblotting as described previously,32 employing highly histones H2A, H2B and H4 (table 1). purified histones separated by SDS-PAGE and blotted onto nitrocellulose membranes. Blotted histones were incubated for IFNc and TNFa, but not IL4 is secreted in response to histone 40 min with sera diluted 1:100 and bound antibodies were Ags visualised using an alkaline-phosphatase conjugated secondary The cytokines IFNc, TNFa and IL4 were measured in SN of SLE Ab. Affinity-purified rabbit Abs specific for the individual and control PBMCs after stimulation with H1, H2A, H2B, H3 histones (Upstate Charlottesville, Virginia, USA) served as and H4. All five histones elicited production of IFNc and TNFa, positive controls and were later replaced by a human serum while IL4 was not detectable (table 1). Cytokine levels in SN of containing autoAbs to all five histones. For a subset of sera, the patients with SLE and HCs showed a similar pattern, with the IgG subclasses of anti-histone Abs were analysed by immuno- highest values measured upon stimulation with H1. blotting using alkaline-phosphatase conjugated secondary anti- Remarkably, TNFa levels exceeded IFNc levels by about 10- bodies specific for IgG1, IgG2, IgG3 and IgG4, respectively (The fold, and TNFa production induced by H1 was significantly Binding Site, Birmingham, UK). higher than the production elicited by H2A, H3, or H4 (table 1). Figure 3 Immunoblot analysis of anti-histone autoantibodies. All sera of patients with systemic lupus erythaematosus (SLE) and healthy controls (HCs) were investigated by immunoblotting for the presence of autoantibodies (autoAbs) to individual histones. Representative immunoblots of 32 patients with SLE (A) and 8 HCs (B) are shown. A human serum reactive with all five histones was used as reference (R). Sera considered positive are marked by an asterisk (*). Of note, SLE sera 3, 18, 23 and HC serum 5 showed a negative result in the anti-histone ELISA, all other immunoblot-positive sera were also positive by ELISA. Patients with anti-H1 Abs were also reactive with histones H3 and/or H4 (prevalence 51% and 49%, respectively), while reactivities to H2A and H2B appeared weaker in general (prevalence 35% and 38%, respectively). Figure 2 H1-specific T cell clones (TCCs). Seven TCCs specific for H1 could be raised by limiting dilution. Four TCCs were generated from one systemic lupus erythaematosus (SLE) patient (A) and three from two healthy individuals (B). The clones were re-exposed to H1, H2A, H2B and Anti-histone antibodies are predominantly of the IgG2 subclass H4 antigens (Ags) and to tetanus toxoid (TT) as a control Ag. All clones and mainly directed to histones H1, H3 and H4 showed proliferative responses against H1, but not against core histones Using ELISA, autoAbs to total histones were detected in 41% of or the control Ag TT. Further details on proliferation, surface phenotype SLE sera, but not in sera from HCs (p,0.001). Anti-histone Abs and cytokine profile of individual TCCs are presented in table 2 and correlated with anti-chromatin Abs (r = 0.75, p,0.001), which supplementary material. were present in 41% of SLE sera, and anti-dsDNA Abs (r = 0.54, p,0.001), but not with total IgG levels. In addition, there was a significant correlation between anti-histone Abs and disease H1-specific T cell clones display a Th1 phenotype activity for all three activity scores (for SIS: r = 38, p,0.05). Seven H1-specific TCCs could be generated from one SLE patient When reactivities to individual histones were analysed by and two HCs (table 2). These clones did not crossreact with other immunoblotting all positive ELISA results were confirmed. In histones or TT (fig 2 and supplementary material). All TCCs addition, four patients with SLE and two HCs who had been expressed CD3 and were of CD4+CD82TCRa/b+/+ phenotype negative by ELISA were positive in the immunoblot assay; of (table 2). Similar to PBMCs in the primary culture, all clones note, all four SLE sera contained Abs to dsDNA and one serum produced high amounts of IFNc,withSNlevelsexceeding was additionally positive in the anti-chromatin assay, in 3000 pg/ml in six of seven TCCs, whereas IL4 was not detectable. contrast to the sera from two HCs. Thus, 51% of SLE sera Furthermore, all clones secreted TNFa with particularly high and 7% of sera from HCs showed anti-histone reactivity on levels produced by the TCCs derived from patients with SLE immunoblots. Almost all positive SLE sera were reactive with (table 2). Thus, all H1-specific TCCs displayed features compa- H1, H3 and H4, while reactions to H2A and H2B occurred less tible with a proinflammatory Th1 phenotype. frequently (fig 3). In all but 1 of 12 SLE sera selected for IgG Seven TCCs specific for H1 were established from one SLE subclass analysis anti-histone Abs were of the IgG2 type, while patient (n = 4) and two HCs (n = 3). All TCCs showed a IgG1 anti-histone Abs were detected in only 1 serum (not CD4+CD82CD28+ phenotype, expressed T cell receptor (TCR) shown) further supporting the concept of a Th1-driven ab and produced high amounts of IFNc and TNFa but no IL4. autoimmune response. Table 2 Phenotyping of histone H1 specific T cell clones (TCCs) TCC Diagnosis Surface phenotyping SI H1 cpm Bgd cpm IFNc pg/ml IL4 pg/ml TNFa pg/ml

TA04 SLE CD4+CD82CD28+ 2.5 1638 655 .3000 ,4 .1385 TA08 SLE CD4+CD82CD28+ 2.3 1730 752 .3000 ,4 .1385 TA30 SLE CD4+CD82CD28+ 4.4 1688 379 .3000 ,4 .1385 TA89 SLE CD4+CD82CD28+ 2.5 1245 498 .3000 ,4ND RI89 HC CD4+CD82CD28+ 2.0 12794 6397 339 ,4 538 SC04 HC CD4+CD82CD28+ 11.5 6820 593 .3000 ,4 496 SC104 HC CD4+CD82CD28+ 9.7 5810 599 .3000 ,4 478 Bgd, background; cpm, counts/min; IFN, interferon; IL, interleukin; HC, healthy control; ND, not determined; SI, stimulation index; SLE, systemic lupus erythaematosus; TNF, tumour necrosis factor.

Association between T cell and B cell responses: histone H1 interesting that, despite the relatively high prevalence of induces anti-histone Ab production in PBMCs of patients with autoAbs to H3 and H4 in sera of patients with SLE, cellular SLE reactivity to these histones was not different from healthy In patients with SLE, the presence of anti-H1 Abs was linked to controls. Moreover, TNFa secretion elicited by H3 and H4 in increased cellular reactivity to H1: thus, 75% of patients the primary culture assays was considerably lower than TNFa showing increased T cell responsiveness to H1 had anti-H1 secretion induced by H1. Thus, it is possible that the Abs as compared to 38% of non-responders (p,0.05). By autoimmune response to H3 and H4 may have less pathogenic contrast, such an association was not seen for the core histones, consequences, or might even exert protective functions. This including H2A. We therefore sought to investigate the effect of assumption is supported by observations made in experimental H1-mediated T cell activation on IgG production and autoAb lupus in which peptides derived from H3 or H4, in contrast to formation by measuring total IgG and anti-histone Abs in the H1-specific peptides, suppressed or delayed lupus symptoms SN of PBMCs from patients with SLE or RA and HCs after when applied to lupus-prone mice.35 36 7 days of cultivation (fig 4). Baseline IgG levels were similar in The data obtained in our system of cultured PBMCs suggest all cultures and polyclonal stimulation by anti-CD3 and PWM that H1-specific T cell help for autologous B cells led to the induced only small increases of IgG, which were most observed increases in anti-histone Ab levels that were only seen pronounced in SLE cultures (p = not significant). However, in SLE cultures, but not in cultures from patients with RA or IgG levels increased upon stimulation with H1 only in SLE HCs. Several reports underline the ability of histone-specific Th cultures and levels were significantly higher (p = 0.02) than in cells to trigger the production of autoAbs (such as anti-dsDNA), H1-stimulated cultures from HCs (fig 4A). Baseline levels of in human and murine SLE.337 Although major T and B cell anti-histone Abs were similar in SLE and RA cultures and at the epitopes have been reported not only in H1 (but also in H2B, lower detection limit in HC cultures. Stimulation with H1 H3, H4), H1 seems particularly important in promoting lupus- increased anti-histone Ab levels above baseline again only in SLE like features in mice. Thus, a peptide derived from histone H1 cultures (p,0.04) resulting in a significant difference (p = 0.02) (H122–42) stimulated autoimmune Th cells, which, in turn, when compared to H1-stimulated PBMCs from HCs (fig 4B). Of augmented the production of pathogenic antinuclear Abs, and note, all patients with SLE had serum Abs to histones and/or this H1 peptide was more potent than core histones in dsDNA, in contrast to patients with RA and HCs. The control Ag accelerating nephritis in lupus-prone mice.19 37 38 TT stimulated IgG production to a similar extent as anti- The occurrence of autoreactive T cells in healthy individuals CD3+PWM but had no effect on production of anti-histone Abs. is in line with previous investigations on potent autoAgs3 36 39–41 and further supports the idea that also peripheral tolerance DISCUSSION mechanisms are disturbed in SLE.42 In contrast to other 29 43 44 Antinucleosomal autoimmune responses seem to play a pivotal reports, no phenotypic differences were seen between role in the pathogenesis of SLE, but it is still unclear which TCCs derived from patients with SLE and HCs including histone(s) form the primary target(s) for histone-specific expression of cell surface markers and cytokine profile. autoreactive T cells. To address this question, we performed a Although they were derived from only one patient and two detailed investigation on autoimmune reactions to all five HCs, it is noteworthy that they were highly specific for H1 and histones at the T cell and additionally also at the B cell level. closely mirrored the cytokine pattern observed in PBMCs after All histone Ags elicited proliferative responses in PBMC H1 stimulation with pronounced production of IFNc and cultures from patients with SLE and HCs, but significant TNFa. Thus, these clones may at least partially reflect the differences were found only for H1 and H2A. Interestingly, complex pathological situation in SLE. cellular reactivity to H1 correlated significantly with cellular Some authors have, mainly based on serum cytokine levels, responses to histones H2A, H2B and H4 suggesting that H1- considered SLE a Th2-mediated disease,45 but the issue has specific T cells may at least partially drive immune responses to remained controversial since other studies reported elevated Th1 other histone Ags, presumably due to the mechanism of cytokines,46 or a balanced intracellular Th1/Th2 ratio in intermolecular epitope spreading as has been described for peripheral SLE lymphocytes.47 In a subgroup of patients with spliceosomal autoAgs in SLE.34 severe renal SLE, however, the intracellular cytokine ratio was In line with previous observations, histones H1, H3 and H4 shifted towards a Th1 phenotype, which was also reflected by appeared to be major Ags of the humoral response in SLE, while histopathological features in the inflamed kidneys.47 48 Abs to H2A and H2B were less frequently found.10 However, Experiments in murine lupus support these findings: A shift concerning these two histones, H2A–H2B dimers may rather towards Th1 predominance was associated with disease form the preferred target of autoAbs, especially in drug-induced acceleration,49 and lupus-like disease after immunisation with LE,56 an issue that was not addressed in our study. It is a peptide mimotope of dsDNA was Th1-mediated.50 including amelioration of lupus nephritis,52 it is tempting to speculate that histone specific T cells, and particularly those directed to H1, could be important stimulators of inflammation in SLE. This further supports the concept of a histone-driven Th1 (or Th0) response in human and murine SLE,338which is in accordance with our observation that anti-histone Abs were predominantly of the IgG2 subclass. In summary, we conclude that histone H1 constitutes an important B and T cell autoAg of probable pathogenic relevance in SLE. Thus, in search of Ag-specific treatment strategies the H1-directed autoimmune response may well constitute a relevant target.

Funding: The work was supported by the Center for Molecular Medicine (CeMM) of the Austrian Academy of Sciences, Vienna, Austria. Competing interests: None declared. Ethics approval: Ethics apprioval was obtained Present address for MA: Clinic III, Medical Faculty ‘‘Carl Gustav Jarus’’ of the Technical University Dresden, Dresden, Germany.

REFERENCES 1. Ebling FM, Hahn BH. Pathogenic subsets of antibodies to DNA. Int Rev Immunol 1989;5:79–95. 2. Burlingame RW, Rubin RL, Balderas RS, Theofilopoulos AN. Genesis and evolution of antichromatin autoantibodies in murine lupus implicates T-dependent immunization with self antigen. J Clin Invest 1993;91:1687–96. 3. Voll RE, Roth EA, Girkontaite I, Fehr H, Herrmann M, Lorenz HM, et al. Histone- specific Th0 and Th1 clones derived from systemic lupus erythematosus patients induce double-stranded DNA antibody production. Arthritis Rheum 1997;40:2162–71. 4. Burlingame RW, Boey ML, Starkebaum G, Rubin RL. The central role of chromatin in autoimmune responses to histones and DNA in systemic lupus erythematosus. JClin Invest 1994;94:184–92. 5. Rubin RL, Bell SA, Burlingame RW. Autoantibodies associated with lupus induced by diverse drugs target a similar epitope in the (H2A-H2B)-DNA complex. J Clin Invest 1992;90:165–73. 6. Rubin RL, Burlingame RW, Arnott JE, Totoritis MC, McNally EM, Johnson AD. IgG but not other classes of anti-[(H2A-H2B)-DNA] is an early sign of procainamide- induced lupus. J Immunol 1995;154:2483–93. 7. Konikoff F, Swissa M, Shoenfeld Y. Autoantibodies to histones and their subfractions in chronic liver diseases. Clin Immunol Immunopathol 1989;51:77–82. 8. Costa O, Monier JC. Antihistone antibodies detected by ELISA and immunoblotting in systemic lupus erythematosus and rheumatoid arthritis. J Rheumatol 1986;13:722–5. 9. Garzelli C, Incaprera M, Bazzichi A, Manunta M, Rognini F, Falcone G. Epstein-Barr virus-transformed human B lymphocytes produce natural antibodies to histones. Immunol Lett 1994;39:277–82. 10. Schett G, Smolen J, Zimmermann C, Hiesberger H, Hoefler E, Fournel S, et al. The Figure 4 Induction of anti-histone antibody (Ab) production by histone autoimmune response to chromatin antigens in systemic lupus erythematosus: H1 in peripheral blood mononuclear cell (PBMC) cultures from patients autoantibodies against histone H1 are a highly specific marker for SLE associated with systemic lupus erythaematosus (SLE). PBMCs from eight patients with increased disease activity. Lupus 2002;11:704–15. 11. Shoenfeld Y, Segol O. Anti-histone antibodies in SLE and other autoimmune with SLE, six patients with rheumatoid arthritis (RA) and six healthy diseases. Clin Exp Rheumatol 1989;7:265–71. controls (HCs) were cultured in the presence of histone H1, or the 12. Gioud M, Kaci MA, Monier JC. Histone antibodies in systemic lupus erythematosus. polyclonal stimuli anti-CD3 and pokeweed mitogen (PWM), or the control A possible diagnostic tool. Arthritis Rheum 1982;25:407–13. antigen (Ag) tetanus toxoid (TT). After 7 days, total IgG and anti-histone 13. Schett G, Rubin RL, Steiner G, Hiesberger H, Muller S, Smolen J. The lupus Abs were measured in the SN by ELISA. All patients with SLE had serum erythematosus cell phenomenon: comparative analysis of antichromatin antibody Abs to histones and/or dsDNA, in contrast to controls subjects. (A) specificity in lupus erythematosus cell-positive and -negative sera. Arthritis Rheum 2000;43:420–8. Induction of IgG production. IgG baseline levels were similar in all 14. Hochberg MC. Updating the American College of Rheumatology revised criteria for cultures and moderate but insignificant increases were observed upon the classification of systemic lupus erythematosus. Arthritis Rheum 1997;40:1725. stimulation with anti-CD3+PWM or in the presence of TT. Upon exposure 15. Schett G, Steiner G, Smolen JS. Nuclear antigen histone H1 is primarily involved in to histone H1, IgG production increased in SLE patient derived cultures lupus erythematosus cell formation. Arthritis Rheum 1998;41:1446–55. (#, p = 0.02 as compared to HC cultures), whereas IgG levels in 16. van Es JH, Gmelig Meyling FH, van de Akker WR, Aanstoot H, Derksen RH, cultures derived from patients with RA or HCs did not change. (B) Logtenberg T. Somatic mutations in the variable regions of a human IgG anti-double- stranded DNA autoantibody suggest a role for antigen in the induction of systemic Induction of anti-histone Ab production. Baseline levels were similar in lupus erythematosus. J Exp Med 1991;173:461–70. SLE and RA cultures and very low in HC cultures. Only in SLE cultures did 17. Winkler TH, Fehr H, Kalden JR. Analysis of immunoglobulin variable region genes histone H1 induce a significant increase of anti-histone Ab levels from from human IgG anti-DNA hybridomas. Eur J Immunol 1992;22:1719–28. baseline (*, p,0.04) and H1-induced anti-histone Ab levels were 18. Kasaian MT, Ikematsu H, Balow JE, Casali P. Structure of the VH and VL segments significantly higher than in H1-stimulated HC cultures (#, p = 0.02). of monoreactive and polyreactive IgA autoantibodies to DNA in patients with systemic lupus erythematosus. J Immunol 1994;152:3137–51. 19. Kaliyaperumal A, Michaels MA, Datta SK. Naturally processed chromatin peptides Interestingly, H1 stimulation did not only induce IFNc but reveal a major autoepitope that primes pathogenic T and B cells of lupus. J Immunol also TNFa production, which exceeded IFNc production 2002;168:2530–7. 51 20. Rajagopalan S, Mao C, Datta SK. Pathogenic autoantibody-inducing c/d T helper severalfold. Since TNFa correlates with SLE disease activity, cells from patients with lupus nephritis express unusual T cell receptors. Clin Immunol and since its therapeutical blockade results in clinical benefit, Immunopathol 1992;62:344–50. 21. Schiffenbauer J, Hahn B, Weisman MH, Simon LS. Biomarkers, surrogate markers, 35. Suen JL, Chuang YH, Tsai BY, Yau PM, Chiang BL. Treatment of murine lupus using and design of clinical trials of new therapies for systemic lupus erythematosus. nucleosomal T cell epitopes identified by bone marrow-derived dendritic cells. Arthritis Arthritis Rheum 2004;50:2415–22. Rheum 2004;50:3250–9. 22. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH. Derivation of the 36. Kaliyaperumal A, Michaels MA, Datta SK. Antigen-specific therapy of murine lupus SLEDAI. A disease activity index for lupus patients. The Committee on Prognosis nephritis using nucleosomal peptides: tolerance spreading impairs pathogenic Studies in SLE. Arthritis Rheum 1992;35:630–40. function of autoimmune T and B cells. J Immunol 1999;162:5775–83. 23. Vitali C, Bencivelli W, Mosca M, Carrai P, Sereni M, Bombardieri S. Development of 37. Datta SK. Major peptide autoepitopes for nucleosome-centered T and B cell a clinical chart to compute different disease activity indices for systemic lupus interaction in human and murine lupus. Ann NY Acad Sci 2003;987:79–90. erythematosus. J Rheumatol 1999;26:498–501. 38. Kaliyaperumal A, Mohan C, Wu W, Datta SK. Nucleosomal peptide epitopes for 24. Smolen JS. Clinical and serologic features: incidence and diagnostic approach. In: nephritis-inducing T helper cells of murine lupus. J Exp Med 1996;183:2459–69. Smolen JS, Zielinski CC, eds. Systemic lupus erythematosus. Berlin, Germany: 39. Gebhard KL, Veldman CM, Wassmuth R, Schultz E, Schuler G, Hertl M. Ex vivo Springer Verlag, 1987:171–96. analysis of desmoglein 1-responsive T-helper (Th) 1 and Th2 cells in patients with 25. Bencivelli W, Vitali C, Isenberg DA, Smolen JS, Snaith ML, Sciuto M, et al. Disease pemphigus foliaceus and healthy individuals. Exp Dermatol 2005;14:586–92. activity in systemic lupus erythematosus: report of the Consensus Study Group of the 40. Ota K, Matsui M, Milford EL, Mackin GA, Weiner HL, Hafler DA. T-cell recognition of European Workshop for Rheumatology Research. III. Development of a computerised an immunodominant myelin basic protein epitope in multiple sclerosis. Nature clinical chart and its application to the comparison of different indices of disease 1990;346:183–7. activity. The European Consensus Study Group for Disease Activity in SLE. Clin Exp 41. Hoffman RW, Takeda Y, Sharp GC, Lee DR, Hill DL, Kaneoka H, et al. Human T cell Rheumatol 1992;10:549–54. clones reactive against U-small nuclear ribonucleoprotein autoantigens from 26. Fritsch R, Eselbock D, Skriner K, Jahn-Schmid B, Scheinecker C, Bohle B, et al. connective tissue disease patients and healthy individuals. J Immunol Characterization of autoreactive T cells to the autoantigens heterogeneous nuclear 1993;151:6460–9. ribonucleoprotein A2 (RA33) and filaggrin in patients with rheumatoid arthritis. 42. Mudd PA, Teague BN, Farris AD. Regulatory T cells and systemic lupus J Immunol 2002;169:1068–76. erythematosus. Scand J Immunol 2006;64:211–18. 27. Jahn-Schmid B, Radakovics A, Luttkopf D, Scheurer S, Vieths S, Ebner C, et al. Bet 43. Wucherpfennig KW, Zhang J, Witek C, Matsui M, Modabber Y, Ota K, et al. Clonal v 1142–156 is the dominant T-cell epitope of the major birch pollen allergenand expansion and persistence of human T cells specific for an immunodominant myelin important for cross-reactivity with Bet v 1-related food allergens. J Allergy Clin basic protein peptide. J Immunol 1994;152:5581–92. Immunol 2005;116:213–19. 44. Oliveira EM, Bar-Or A, Waliszewska AI, Cai G, Anderson DE, Krieger JI, et al.A. 28. Scheinecker C, Machold KP, Majdic O, Hocker P, Knapp W, Smolen JS. Initiation of CTLA-4 dysregulation in the activation of myelin basic protein reactive T cells may the autologous mixed lymphocyte reaction requires the expression of costimulatory distinguish patients with multiple sclerosis from healthy controls. J Autoimmun molecules B7-1 and B7-2 on human peripheral blood dendritic cells. J Immunol 2003;20:71–81. 1998;161:3966–73. 29. Fritsch-Stork R, Mullegger D, Skriner K, Jahn-Schmid B, Smolen JS, Steiner G. The 45. Houssiau FA, Lefebvre C, Vanden Berghe M, Lambert M, Devogelaer JP, Renauld spliceosomal autoantigen heterogeneous nuclear ribonucleoprotein A2 (hnRNP-A2) is JC. Serum interleukin 10 titers in systemic lupus erythematosus reflect disease a major T cell autoantigen in patients with systemic lupus erythematosus. Arthritis activity. Lupus 1995;4:393–5. Res Ther 2006;8:R118. 46. Wong CK, Ho CY, Li EK, Lam CW. Elevation of proinflammatory cytokine (IL-18, IL- 30. Fritsch R, Bohle B, Vollmann U, Wiedermann U, Jahn-Schmid B, Krebitz M, et al. 17, IL-12) and Th2 cytokine (IL-4) concentrations in patients with systemic lupus Bet v 1, the major birch pollen allergen, and Mal d 1, the major apple allergen, erythematosus. Lupus 2000;9:589–93. cross-react at the level of allergen-specific T helper cells. J Allergy Clin Immunol 47. Akahoshi M, Nakashima H, Tanaka Y, Kohsaka T, Nagano S, Ohgami E, et al. Th1/ 1998;102:679–86. Th2 balance of peripheral T helper cells in systemic lupus erythematosus. Arthritis 31. Del Prete GF, De Carli M, Mastromauro C, Biagiotti R, Macchia D, Falagiani P, et al. Rheum 1999;42:1644–8. Purified protein derivative of Mycobacterium tuberculosis and excretory-secretory 48. Masutani K, Akahoshi M, Tsuruya K, Tokumoto M, Ninomiya T, Kohsaka T, et al. antigen(s) of Toxocara canis expand in vitro human T cells with stable and opposite Predominance of Th1 immune response in diffuse proliferative lupus nephritis. Arthritis (type 1 T helper or type 2 T helper) profile of cytokine production. J Clin Invest Rheum 2001;44:2097–106. 1991;88:346–50. 49. Takahashi S, Fossati L, Iwamoto M, Merino R, Motta R, Kobayakawa T, et al. 32. Hassfeld W, Steiner G, Studnicka-Benke A, Skriner K, Graninger W, Fischer I, et al. Imbalance towards Th1 predominance is associated with acceleration of lupus-like Autoimmune response to the spliceosome. An immunologic link between rheumatoid autoimmune syndrome in MRL mice. J Clin Invest 1996;97:1597–604. arthritis, mixed connective tissue disease, and systemic lupus erythematosus. 50. Khalil M, Inaba K, Steinman R, Ravetch J, Diamond B. T cell studies in a peptide- Arthritis Rheum 1995;38:777–85. induced model of systemic lupus erythematosus. J Immunol 2001;166:1667–74. 33. Greidinger EL, Gazitt T, Jaimes KF, Hoffman RW. Human T cell clones specific 51. Aringer M, Feierl E, Steiner G, Stummvoll GH, Hofler E, Steiner CW, et al. Increased for heterogeneous nuclear ribonucleoprotein A2 autoantigen from connective bioactive TNF in human systemic lupus erythematosus: associations with cell death. tissue disease patients assist in autoantibody production. Arthritis Rheum Lupus 2002;11:102–8. 2004;50:2216–22. 52. Aringer M, Graninger WB, Steiner G, Smolen JS. Safety and efficacy of tumor 34. Monneaux F, Muller S. Key sequences involved in the spreading of the systemic necrosis factor a blockade in systemic lupus erythematosus: an open-label study. autoimmune response to spliceosomal proteins. Scand J Immunol 2001;54:45–54. Arthritis Rheum 2004;50:3161–9.

Chapter 5

The Spliceosomal Autoantigen Heterogeneous Nucleas Ribonucleoprotein A2 (hnRNP-A2) is a Major T Cell Autoantigen in Patients with Systemic Lupus Erythematosus

RD Fritsch-Stork1 D Müllegger1,2 K Skriner1,3 B Jahn-Schmid4 JS Smolen1,5 G Steiner1,2,5

1Division of Rheumatology, Department of Internal Medicine III, Medical University of Vienna, Austria 2Center of Molecular Medicine (CeMM) of the Austrian Academy of Sciences, Vienna, Austria 3Charité University Medicine Berlin, Department of Rheumatology and Clinical Immunology, Humboldt University and Free University, Berlin, Germany 4Institute of Pathophysiology, Medical University of Vienna, Austria 5Ludwig Boltzmann Institute for Rheumatology and Balneology, Vienna, Austria

Arthritis Research & Therapy 2006,8:R118.

57 The spliceosomal autoantigen heterogeneous nuclear ribonucleoprotein A2 (hnRNP-A2) is a major T cell autoantigen in patients with systemic lupus erythematosus

Ruth Fritsch-Stork1, Daniela Müllegger1,2, Karl Skriner1,3, Beatrice Jahn-Schmid4, Josef S Smolen1,5 and Günter Steiner1,2,5

Abstract

A hallmark of systemic lupus erythematosus (SLE) is the appearance of autoantibodies to nuclear antigens, including autoantibodies directed to the heterogeneous nuclear ribonucleoprotein A2 (hnRNP-A2), which occur in 20% to 30% of SLE patients as well as in animal models of this disease. To investigate the underlying cellular reactivity and to gain further insight into the nature and potential pathogenic role of this autoimmune response we characterized the T cell reactivity against hnRNP-A2 in patients with SLE in comparison to healthy controls. Cellular proliferation of peripheral blood T cells to hnRNP-A2 was determined by [3H]thymidine incorporation and T cell clones (TCCs) specific for hnRNP-A2 were grown by limiting dilution cloning; IFNγ, IL-4 and IL-10 in culture supernatants were measured by ELISA. Bioactivity of culture supernatants was determined by incubation of anti-CD3/anti-CD28 stimulated peripheral blood CD4+ T cells with supernatants of TCCs. Stimulation assays performed with peripheral blood mononuclear cells of 35 SLE patients and 21 healthy controls revealed pronounced proliferative responses in 66% of SLE patients and in 24% of the controls, which were significantly higher in SLE patients (p<0.00002). Furthermore, hnRNP-A2 specific TCCs generated from SLE patients (n=22) contained a relatively high proportion of CD8+ clones and mostly lacked CD28 expression, in contrast to TCCs derived from healthy controls (n=12). All CD4+ TCCs of patients and all control TCCs secreted IFNγ and no IL-4. In contrast, CD8+ TCCs of patients secreted very little IFNγ, while production of IL-10 did not significantly differ from other T cell subsets. Interestingly, all CD8+ clones producing IL-10 in large excess over IFNγ lacked expression of CD28. Functional assays showed a stimulatory effect of the supernatants derived from these CD8+CD28- hnRNP-A2 specific TCCs that was similar to that of CD4+CD28+ clones. Taken together, the pronounced peripheral T cell reactivity to hnRNP-A2 observed in the majority of SLE patients and the distinct phenotype of patient-derived CD8+ TCCs suggest a role for these T cells in the pathogenesis of SLE.

Introduction etiopathogenesis of SLE still remains elusive. The presence of autoantibodies (autoAbs) to Systemic lupus erythematosus (SLE) is an nuclear antigens in virtually all SLE patients autoimmune disease characterized by a wide constitutes the most characteristic serological spectrum of multi-organ manifestations, and feature of this disorder. Among the numerous genetic, hormonal, environmental and autoantigens described, the nucleosome immunoregulatory factors are known to represents an important target structure of the contribute to expression of the disease [1]. autoimmune attack, leading to the formation However, in spite of the considerable of a wide array of autoAbs directed to accumulated knowledge, the detailed single histones,

58 Table 1

Demographic characteristics of systemic lupus erythematosus patients

Patient Sex Age (years) ECLAM score Medication 1 Female 25 1.5 HQ,P 2 Female 25 0.5 0 3 Female 52 2.5 0 4 Female 24 1.5 C,P 5 Female 29 0 HQ 6 Female 43 1 0 7 Female 29 0.5 HQ,P 8 Female 38 1.5 HQ,Aza,P 9 Female 34 0 MTX,P 10 Female 60 0 MTX,P 11 Female 52 1 P 12 Female 48 2 HQ,P 13 Female 43 0 0 14 Female 62 1 HQ,P 15 Male 33 0 HQ 16 Female 34 2 Aza,P 17 Female 37 3 0 18 Male 30 0 Ara,P 19 Female 45 4 0 20 Female 31 1 0 21 Female 24 6 Aza,P 22 Female 30 1 0 23 Female 29 1.5 MTX,HQ 24 Female 49 2 0 25 Female 28 2 HQ 26 Female 24 1.5 P 27 Male 45 1 HQ,P 28 Female 22 3.5 HQ,P 29 Female 26 0 HQ 30 Female 29 1 P 31 Female 24 2 0 32 Female 61 2 P 33 Female 43 5.5 MTX,P 34 Female 33 1 0 35 Male 44 1 C

0, none; Aza, azathioprin; C, cyclophosphamide; ECLAM, European Consensus Lupus Activity Measurement; HQ, hydroxychloroquine; MTX, metothrexate; P, prednisolone.

59 histone-DNA complexes and double-stranded [20,21]. Interestingly, most of the T cell (ds)DNA (reviewed in [2,3]). The search for clones (TCCs) raised from SLE patients were new serological markers has led to the of the Th1 or Th0 subset [21,22]. However, identification of a number of additional there have been only few reports about the T autoantigens, among them the Ro and La cell response to spliceosomal antigens, proteins [4], as well as components of the mainly focusing on the T cell reactivity to spliceosome. This large and highly dynamic small nuclear ribonucleoprotein antigens complex contains many evolutionarily [23]. In a recent report by Greidinger and conserved proteins, such as the U1–70 kD colleagues, the authors characterized hnRNP- RNP or the Sm proteins [5,6], which are A2 specific CD4+ TCCs derived from one targeted by approximately 30% of SLE mixed connective tissue disease patient and patients. two SLE patients, and attributed a possible pathogenic role to these T cells in SLE [24]. About 15 years ago, the heterogeneous However, the presence of autoreactive T cells nuclear ribonucleoprotein (hnRNP-)A2, is not limited to patients, but has repeatedly another component of the spliceosome, was been observed also in healthy individuals characterized as a novel target of autoAbs in [21,25,26]; thus, the search for possible systemic autoimmune diseases [7]. Although differences in autoantigen specific cellular initially described to occur mainly in patients reactivity between patients and healthy with rheumatoid arthritis (RA), autoAbs to controls might give insight into the hnRNP-A2 (originally termed anti-RA33) pathogenesis of the respective autoimmune were later detected also in 20% to 30% of disease. patients with SLE as well as in 40% of Recently, we were able to characterize the patients with mixed connective tissue cellular response against hnRNP-A2 in disease, usually in association with other patients with RA [27]. We observed that anti-spliceosomal autoAbs [8,9]. The autoAb approximately half of the RA patients harbor response against hnRNP-A2 shows some T cells against hnRNP-A2. In accordance differences in epitope recognition among with the perception of RA as an patients with RA, SLE and mixed connective inflammatory, Th1 type systemic tissue disease [10]. Interestingly, autoAbs to autoimmune disease, all generated TCCs hnRNP-A2 have also been detected in were of the Th1 subtype, as defined by their several animal models of RA and SLE pre-dominant IFNγ secretion. [11,12]. HnRNP-A2 has a predominant nuclear localization and exerts multiple To gain more insight into the role of functions, including regulation of alternative autoantigen specific T cells in SLE, we splicing, transport of mRNA and regulation investigated spontaneous T cell responses to of translation [13-17]. hnRNP-A2 in patients with SLE and in healthy control subjects and characterized Most of the autoAbs in SLE patients have TCCs specific for this antigen. The data undergone immunoglobulin class switching obtained suggest that hnRNP-A2 may and affinity maturation. This and the constitute an important T cell autoantigen in association of HLA-DR subtypes with the patients with SLE, indicating a potential role presence of certain autoAbs indicates an for it in the pathogenesis of this disorder. antigen-driven immune response, emphasizing the role of T cells in SLE [18,19]. The cellular aspect of the immune Materials and methods response to DNA in its various forms has been extensively studied, including the Patients and controls characterization of T cells inducing anti- Peripheral blood from 35 patients with SLE dsDNA autoAb production by B cells (31 female, 4 male, mean age 36.7±3.4 years,

60 for demographic characteristics see Table 1) Germany). Cytokines were measured by classified according to the revised criteria of ELISA (BioSource, Fleurus, Belgium) in the American College of Rheumatology [1] supernatants of TCCs after 24 hours was drawn into heparinized test tubes. incubation with 0.35 µg/ ml hnRNP-A2 or Informed consent was obtained from all 0.5 U/ml tetanus toxoid as control antigen. patients. Most patients were treated with Detection limits were 9 pg/ml for IFNγ, 4 immunomodulatory drugs (n=21) and/or low- pg/ml for IL-4, and 8 pg/ml for IL-10. dose glucocorticoids (n=15). Nine patients did not receive any medication. Disease T cell stimulation assays activity was determined by European Peripheral blood mononuclear cells (PBMCs) Consensus Lupus Activity Measurement were isolated from heparinized blood of SLE (ECLAM) score [28]. While 87% of the patients and controls by centrifugation on patients had moderately active disease Ficoll Hypaque (Pharmacia). After washing (ECLAM score <3), 5 patients had active and counting cells were either immediately SLE with an ECLAM score ≥3. The control used or frozen in RPMI medium containing population consisted of 21 healthy 10% dimethylsulfoxide and20% fetal calf individuals (10 female and 11 male, mean serum. Cells were cultured in the presence of age 31.8±1.7 years). the antigens for 5 days at 37°C in triplicate in 96-well plates (Costar, Cambridge, MA, Antigens USA) in a total volume of 200 µl Recombinant full-length hnRNP-A2 was (105cells/well). used in all experiments. The cDNA encoding the antigen [15] was cloned into the pET-30 LIC vector (Novagen, Madison, WI, USA) Figure 1 and expressed as His-tagged fusion protein as described [27]. Purification from bacterial lysates was achieved by Ni-NTA affinity chromatography (Quiagen, Hilden, Germany) followed by Polymyxin B Sepharose adsorption (BioRad, Hercules, CA, USA) and anion exchange chromatography on DEAE Sepharose (Pharmacia, Uppsala, Sweden) essentially as described [27]. Endotoxin content was determined by the lympholytic amoebocyte lysate assay (BioWhittaker, Verviers, Belgium). Using this procedure, a more than 99% pure, Proliferative responses of peripheral blood endotoxin-free preparation was obtained. The mononuclear cells (PBMCs) to heterogeneous nuclear optimum concentration for proliferation ribonucleoprotein (hnRNP)-A2 in systemic lupus assays was found to be 0.35 µg/ml. Tetanus erythematosus (SLE) patients and healthy controls. toxoid as control antigen was obtained from Proliferation was measured by [3H]thymidine Pasteur Merieux Connaught (Willowdale, incorporation in PBMCs of 35 SLE patients and 21 controls in the presence or absence of hnRNP-A2. A Ontario, Canada) and used at a concentration stimulation index ≥3.0 and a ∆cpm >1,000 cpm was of 0.5 U/ml as previously described [29]. considered a positive response (represented by the dashed line). The mean stimulation index (SI) was Dectection of antibodies and cytokines 6.7±2.3 for SLE patients, and 2.3±0.2 for controls AutoAbs to hnRNP-A2 were detected by (indicated by the solid bars). The difference between SLE patients and controls was highly significant immunoblotting as described [8,10], (p<0.00002). HnRNP-A2 seropositive SLE patients employing the recombinant antigen and are indicated by crosses, and hnRNP-A2 seronegative additionally by ELISA (IMTEC, Berlin, SLE patients by diamonds.

61 Culture medium consisted of Ultra Culture and pulsing with [3H]TdR for an additional serum-free medium (Biowhittaker, 16 hours, cells were harvested and the Wakersville, MD, USA) containing 2 mM incorporated radioactivity was measured by glutamine and 0.02 mM 2-mercaptoethanol scintillation counting. Production of IL-4, supplemented with 100 U/ml penicillin IFNγ and IL-10 was measured by ELISA in /streptomycin (Life Technologies, Paisley, supernatants collected after 24 hours of UK). Phytohemagglutinin (Life incubation as described [27]. Technologies) and IL-2 (Roche Molecular Biochemicals, Mannheim, Germany) were For phenotyping, cloned T cells (0.5 to used as polyclonal stimuli. During the last 16 1×105) were washed twice with ice-cold hours of culture 0.5 µCi per well [3H]TdR FACS buffer (phosphate-buffered saline, 5% (Amersham-Pharmacia Biotech Europe, fetal calf serum, 0.01% NaN3) and incubated Freiburg, Germany) was added and the for 30 minutes at 4°C with a fluorescein incorporated radioactivity was measured by isothiocyanate (FITC) or phycoerythrin- scinitillation counting. Results were conjugated monoclonal Ab (mAb) (BD expressed as stimulation index (SI) defined Pharmingen, San Diego, USA). Anti-T cell as the ratio of mean counts per minute (cpm) receptorα/β, anti-CD4 and anti- CD28 mAbs obtained in cultures with antigen to mean were FITC-conjugated, and anti-T cell cpm obtained in cultures incubated in the receptor γ/δ and anti-CD8 mAbs were absence of antigen. An SI ≥3.0 and a ∆cpm phycoerythrin-conjugated. Antibodies of the >1,000 (mean cpm obtained in cultures with appropriate IgG isotypes were used as antigen minus mean cpm obtained in cultures negative controls. Afterwards cells were incubated in the absence of antigen) was washed again in FACS buffer and analyzed regarded as a positive response. with a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA). The Antigen-specific T cell lines and clones acquired data were analyzed using Flow-Jo Antigen specific T cell lines were obtained software (Tree Star, Inc., Ashland, OR, using a previously established protocol [30]. USA). In brief, 2×106 PBMCs were stimulated with 0.35 µg/ml hnRNP-A2 for 5 days in 24-well Stimulation assays with supernatants from flat-bottomed culture plates. On the 5th day T cell clones. of culture, IL-2 was added at 20 U/ml and the To investigate a possible functional culture was continued for an additional 7 difference between the CD4+CD28+ and days. To generate TCCs, T cell lines were CD8+CD28- TCCs, purified CD4+ T cells restimulated with hnRNP-A2 and after 2 from two SLE patients and two healthy more days viable T cells were separated by controls (105 cells/well) were stimulated with Ficoll/Hypaque and seeded in limiting platebound anti-CD3 mAb (clone OKT3, dilution (0.5 cells/well) in 96-wells plates. T Janssen-Cilag, Saundertown, UK) and anti- cells were cultured in the presence of 1×105 CD28 mAb (clone 15E8, Caltag, irradiated (5,000 rad) allogeneic PBMCs as Burlingame, CA, USA) and incubated in 'feeder cells', 0.5 µg/ml phytohemagglutinin duplicates with either 50 µl supernatant and 20 U/ml IL-2 in medium containing 1% derived from CD4+CD28+ or CD8+CD28- heat-inactivated human AB serum. Growing TCCs, respectively. After 24 hours, cells microcultures were then expanded at weekly were pulsed with 0.5 µCi per well [3H]TdR intervals with fresh feeder cells in the and proliferation was measured after an presence of IL-2. The specificity of TCCs additional 16 hours. Incubation with IL-10 was assessed by proliferation assays alone (0.5 to 20 ng/ml; Insight Biotechnology incubating 2×104 T cells with 0.35 µg/ml Ltd, Wembley, UK) was used as control. hnRNP-A2 in the presence of 105 autologous irradiated PBMCs. After 48 hours incubation

62 Statistical analysis 21 age and sex matched healthy controls. Unless stated otherwise, SI are expressed as Significant responses (defined as SI ≥3, mean±SEM of the respective cohort ∆cpm ≥1,000) were observed in 66% of the analyzed. Student's t-test was used to SLE patients and 24% of the controls (Figure determine differences between controls and 1). In SLE patients, the SI ranged from 0.5 to SLE patients. A p value <0.05 was regarded 18 (median SI=4.4) with a mean SI of significant. For assessment of correlations 6.7±2.3, wheras the SI was much lower in the the Pearson's correlation coefficient was control group, ranging from 0.7 to 4.1 used. (median SI=2.0) with a mean SI of 2.3±0.2 (p<0.00002). Moreover, a SI>4 was seen in Results 18 SLE patients (52%) but in only one control subject (5%).

Proliferative responses to hnRNP-A2 in SLE patients and healthy controls As seen in Figure 1, T cell responses of SLE Cellular reactivity against hnRNP-A2 was patients did not correlate with the presence of investigated by measuring proliferation of anti-hnRNP-A2 autoAbs, which were PBMCs obtained from 35 SLE patients and detected in 20% of the SLE patients but not

Figure 2

Phenotype, proliferation and cytokine production of heterogeneous nuclear ribonucleoprotein (hnRNP)-A2 specific T cell clones (TCCs) of systemic lupus erythematosus (SLE) patients and controls. TCCs were generated by limiting dilution and their specificity was assessed by measuring proliferation in response to hnRNP-A2. Cytokine production of TCCs was measured in the supernatants after 24 hours of stimulation with autologous antigen presenting cells and hnRNP-A2. (a) Proliferation of CD4+ and CD8+ TCCs of SLE patients. (b) Proliferation and IFNγ production of TCCs of SLE patients. (c) Proliferation of CD4+ and CD8+ TCCs of healthy controls (HCs). (d) Proliferation and IFNγ production of TCCs of healthy controls. There was no correlation between proliferative capacity and cytokine secretion.

63 Figure 3 Figure 4

Cytokine production by heterogeneous nuclear ribonucleoprotein (hnRNP)-A2 specific T cell clones (TCCs) from systemic lupus erythematosus (SLE) patients and healthy controls (a) IFNγ production. Whereas IFNγ production of CD4+ and CD8+ TCCs derived from healthy controls was similar (mean ± SEM of 533 ± 148 for CD4+ versus 376 ± 312 for CD8+; p = not significant), IFNγ production by CD4+ TCCs from SLE patients was significantly higher than IFNγ production by patient derived CD8+ TCCs (500 ± 124 versus 40 ± 23 pg/ml; p < 0.003). (b) Ratio of IL-10/IFNγ secretion in hnRNP-A2 specific TCCs from SLE patients and controls. IL-10 secretion was measured in supernatants of 11 TCCs from SLE patients and 5 TCCs from healthy controls. IL-10 secretion largely exceeded IFNγ production in CD8+CD28- TCCs, which were all derived from SLE patients. FACS analysis of four heterogeneous nuclear ribonucleoprotein (hnRNP)-A2 specific T cell clones (TCCs) derived from systemic lupus erythematosus in healthy controls, in accordance with (SLE) patients. Left panels: clones were stained with previous observations [8,10]. There was also fluorescein isothiocyanate (FITC)- or phycoerythrin- conjugated mAbs against CD4 or CD8, respectively. no correlation with anti-dsDNA antibodies or Right panels: clones were stained with a FITC- disease activity as measured with the conjugated anti-CD28 mAb and a phycoerythrin- ECLAM score. conjugated control mAb. Top panel: FACS profile of a CD4+ CD28- TCC. Second panel: FACS profile of a The response to tetanus toxoid as control CD4+CD28+ TCC. Third panel: FACS profile of a CD8+ CD28- TCC. Bottom panel: FACS profile of a antigen was measured in 21 of the SLE CD8+ CD28+ TCC patients and 16 of the healthy controls. Although stimulation elicited by this recall antigen was somewhat higher in SLE patients than in controls (6.2±1.1 versus 5.0±1.4), the induced by hnRNP-A2, the response of SLE difference was not significant. Of note, while patients to tetanus toxoid was slightly but in controls the stimulation elicited by tetanus significantly lower than the response to toxoid was considerably higher than the one hnRNP-A2 (Table 2).

64 Generation of hnRNP-A2 reactive T cell Proliferation of the 12 TCCs derived from clones control individuals was comparable to T cell lines specific for hnRNP-A2 were patient-derived clones, with SIs ranging from established from the peripheral blood of 8 3.6 to 37 and a mean SI of 9.0±2.8 (Figure SLE patients and 4 healthy controls. Using 2c). Phenotyping, on the other hand, revealed the limiting dilution technique, between 1 some differences. The majority of TCCs and 6 antigen specific TCCs per individual (n=8; 67%) were CD4+/CD8-, while 3 clones (34 TCCs in total) could be raised. Twenty- (25%) were CD4-/CD8+ and one TCC was two TCCs were derived from SLE patients: CD4-/CD8-. In contrast to SLE TCCs, all 13 (59%) of them were CD4+/CD8- and thus clones secreted IFNγ, though in highly belonged to the helper T cell subset, while 9 varying amounts (Figure 2d). Whereas in were CD4-/CD8+. The TCCs showed high SLE patients IFNγ secretion was variability in their degree of proliferation, significantly higher in CD4+ than in CD8+ with SIs ranging from 3.0 to 26.0 and a mean TCCs (p<0.003), no difference in cytokine SI of 6.2±1.2. (Figure 2a). Analysis of the production was found in CD4+ and CD8+ cytokine secretion pattern revealed TCCs from healthy controls (Figure 3a). pronounced IFNγ production by 10 of the Additionally, IFNγ secretion of CD8+ SLE clones, while the remaining 12 clones TCCs was markedly lower than in CD8+ produced very little, if any, IFNγ. Cytokine control TCCs (Figure 3a); however, due to production did not correlate with the degree the small number of CD8+ control clones of proliferation as even poorly proliferating (n=3), the difference between patient and TCCs secreted high amounts of IFNγ and control TCCs (39.5±29.4 pg/ml versus vice versa (Figure 2b), whereas IL-4 was not 376±312 pg/ml) did not reach statistical detectable at all. significance.

Figure 5

Proliferative responses of CD4+ T cells upon incubation with supernatants from T cell clones (TCCs). CD4+ T cells of two healthy controls (HC-1, HC-2) and two systemic lupus erythematosus (SLE) patients (SLE-1, SLE- 2) were stimulated with platebound anti-CD3/anti-CD28 mAbs. Cells were incubated for 24 hours with the supernatants of a CD4+CD28+ and a CD8+CD28- TCC derived from an SLE patient, and proliferation was measured after an additional 16 hours. The increase in proliferation was statistically significant for all supernatants examined (indicated by a star).

65 Expression of the costimulatory molecule populations. To address this question, CD4+ CD28 was measured in six CD4+ and seven T cells (obtained from two SLE patients and CD8+ TCCs from SLE patients and in five two healthy controls) were stimulated with TCCs from healthy subjects (Figure 4): while anti-CD3 and anti-CD28 mAbs and all control TCCs examined (four CD4+ and subsequently incubated with supernatants one CD8+) expressed CD28, only three SLE from two CD8+CD28- and two CD4+ TCCs were CD28+ (two CD4+ and one CD28+ TCCs derived from SLE patients CD8+ TCC). Of note, all CD28+ clones and (Figure 5). All supernatants caused a also the four CD4+CD28- SLE TCCs significant increase of the anti-CD3/anti- produced IFNγ, whereas this cytokine was CD28 induced proliferative response: barely detectable in the supernatants of the supernatants from CD8+CD28- clones six CD8+CD28- TCCs; on the other hand, all enhanced proliferation by 452±103% clones secreted comparable amounts of IL- (p<0.001), and CD4+CD28+ derived 10. Thus, IL-10 was produced in 6- to 1,000- supernatants enhanced proliferation by fold excess over IFNγ by the CD8+CD28- 522±161% (p<0.02). Since the CD8+CD28- TCCs while all other clones examined (i.e. clones produced IL-10 in large excess over CD8+CD28+, CD4+CD28+ and CD4+ IFNγ, we examined the effect of IL-10 on CD28-) secreted comparable amounts of proliferation: a modest reduction of IFNγ and IL-10 (Figure 3b). proliferation was reproducibly seen, which, however, was not statistically significant Stimulation assays with supernatants of T (16.6±6.6%, p=0.07). Therefore, the cell clones pronounced stimulatory capacity of these As pronounced differences in the cytokine supernatants cannot be attributed to IL-10 or secretion pattern of SLE patient derived IFNγ. Unfortunately, the TCCs were short- CD4+CD28+ and CD8+CD28- TCCs had lived, limiting the number of experiments, been observed, we were interested in the and additional studies will be required to functional properties of these two further investigate this issue.

Table 2 Proliferative responses to hnRNP-A2 and the control antigen tetanus toxoid in SLE patients and healthy controls Proliferative response to

hnRNP-A2 Tetanus toxoid P value SLE patients (n = 21) Healthy 7.5±1.2 6.2±1.1 P<0.05 controls (n = 16) 2.5±0.6 5.0±1.4 P<0.002 P value P<0.000002 NS Data are given as mean stimulation index ±SEM. Student's paired t-test was used to calculate statistical significances. hnRNP, heterogeneous nuclear ribonucleoprotein; NS, not significant; SLE, systemic lupus erythematosus.

Discussion for the presence of autoreactive T cells to hnRNP-A2 to better understand the role of In previous investigations we characterized this autoimmune response in the the humoral response to hnRNP-A2 in SLE pathogenesis of SLE. The data obtained show patients and some clinical implications the existence of a pronounced cellular thereof [8,10,31,32]. In this study we response to hnRNP-A2 in the majority of examined SLE patients and healthy controls SLE patients, which was far more vigorous

66 than in healthy controls. In contrast, the indicate a special pathological role of this T response to the control antigen tetanus toxoid cell subset because this phenotype appeared was similar in patients and controls, to be restricted to SLE patient derived TCCs. demonstrating the specificity of the immune In previous reports, CD4+ T cells in SLE response towards hnRNP-A2. Moreover, the were shown to be of predominantly Th1 or response of SLE patients to hnRNP-A2 was Th0 subtypes [21,24]. Furthermore, of similar magnitude, or even slightly higher, autoreactive CD4+ TCCs from SLE patients than the response to tetanus toxoid, which is and healthy controls showed similar cytokine a recall antigen eliciting a pronounced production [22]. Whereas the relatively high response in the majority of individuals tested. proportion of hnRNP-A2 specific CD8+ TCCs might mirror the common The occurrence of autoreactive T cells in understanding of a generally increased healthy individuals is a common finding that CD8:CD4 ratio in SLE patients [33-35], there has previously been reported for antigens are only anecdotal reports about autoreactive associated with SLE [21] as well as other CD8+ TCCs in SLE [6] and their role is not autoimmune diseases like pemphigus entirely clear and may be quite complex foliaceus [25] and multiple sclerosis [26]. indeed. Thus, on the one hand, the rather low Thus, the presence of hnRNP-A2 IFNγ secretion of patient derived CD8+ autoreactive T cells in healthy controls by TCCs might be due to inherent abnormalities itself was not surprising; a striking differ- of CD8+ (suppressor) T cell function in SLE, ence, however, was the significantly higher as shown recently by Filaci and colleagues strength of the cellular immune response [36]. On the other hand, these TCCs secreted observed in SLE patients, which may be considerable amounts of IL-10 and, in considered an indication for pathogenic contrast to the TCCs derived from healthy involvement of these autoreactive cells controls, were CD28 negative. Recently, a and/or a lack of counter-regulation in SLE new T suppressor population was defined, patients. which could be generated in vitro from CD8+CD28- T cells [37]. This subset exerted Interestingly, no correlation of cellular its regulatory and suppressive function in a reactivity to hnRNP-A2 with the appearance cell contact independent manner via secretion of the respective autoAbs in SLE patients of IL-10, which is in line with previous could be observed. Thus, hnRNP-A2 appears reports attributing a regulatory function to to be a predominant T cell antigen, while the CD8+CD28- T cells [38-40]. An increase of generation of autoAbs might represent a these cells in patients with SLE might thus bystander phenomenon occurring in a constitute an effort of counter-regulation subgroup of patients. However, recent data within the disturbed immune system of SLE indicate that in SLE patients the humoral patients. response to hnRNP-A2 fluctuates considerably and increases during flares. Surprisingly, though, the supernatants of the Therefore, the prevalence of these autoAbs hnRNP-A2 specific CD8+CD28- TCCs may be considerably higher than previously derived from SLE patients did not inhibit assumed (unpublished observation). proliferation of CD4+ T cells, but instead enhanced anti-CD3/anti-CD28 induced Although the cellular reactivity to hnRNP-A2 stimulation, similar to supernatants derived appeared to be primarily a Th1 response, we from CD4+CD28+ TCCs. As IL-10 alone observed a relatively high percentage of had an antiproliferative effect, the increased CD8+ TCCs in SLE patients. Of particular content of IL-10 in the supernatants of interest, most of these clones lacked CD28 CD8+CD28- TCCs was obviously expression and produced neither IFNγ nor counteracted by other yet unknown IL-4, but did produce IL-10. This may stimulatory components of the supernatant,

67 which remain to be identified. Of note, the SLE, playing a presumably pivotal role in the lack of CD28 expression is also a hallmark of pathogenesis of this disease, even though the senescent T cells, which are known to be reasons for this are still far from being fully increased in various autoimmune diseases understood. and chronic inflammatory conditions and may show a rather aggressive phenotype [41]. Therefore, it is conceivable that the Conclusion CD8+CD28- TCC may be derived from the senescent T cell pool of SLE patients; this Our data reveal a pronounced T cell response also correlates better with our data. against the autoantigen hnRNP-A2 in the majority of SLE patients in contrast to a Elevated IL-10 serum levels have been scarcer and lower response in healthy reported in SLE patients and correlated with subjects. Although this autoreactivity seemed clinical and serological disease activity, to be mainly conferred by CD4+ T cells especially anti-DNA antibody titres [42]. showing a Th1 phenotype, a newly defined Several cell types have been implicated in the hnRNP-A2 specific CD8+CD28- T cell- production of IL-10 [43] and augmented IL- subset was observed in SLE patients. This 10 secretion has been linked to autoAb subpopulation showed a predominant IL-10 production in a model where PBMCs from secretion and a lack of IFNγ or IL-4 patients with SLE were transferred into mice production. Our data suggest involvement of with severe combined immunodieficiency both populations in the complex and still [44]. So, altogether, the CD8+CD28- T cell incompletely understood pathogenesis of subpopulation might enhance the SLE, and warrant further investigations into inflammatory process in SLE patients by the cellular aspects of this autoimmune stimulation of B cells via IL-10. response. Unfortunately, because of the short life-span of the clones, additional functional assays could not be performed and remain the Competing interests subject of future investigations. The authors declare that they have no Finally, the question remains why hnRNP- competing interests. A2 is such a preferred target of the T cell response in SLE, while autoAbs occur less frequently and may represent an Authors' contributions epiphenomenon of ongoing T cell autoreactivity. On the other hand, autoAb RF participated in the design of the study, titers increase during disease flares worked with the T cell cultures and drafted (unpublished observation) and, importantly, the manuscript. DM performed the T cell autoAbs to hnRNP-A2 are, together with cloning and T cell assays. KS prepared the anti-Sm antibodies, among the earliest recombinant antigen. BJS participated in the detectable autoAbs in MRL/lpr mice, where design of the study. JS and GS conceived of they were found to precede anti-dsDNA the study, participated in its design and autoAbs [11]. HnRNP-A2 is a coordination and helped to write the multifunctional protein that, in the nucleus, manuscript. All authors read and approved partially colocalizes with spliceosomal the final manuscript. complexes [13,14], a preferred autoimmune target in SLE. Thus, the data obtained in the course of this study further strengthen the view of the spliceosome being one of the predominant autoimmune target structures in

68 Acknowledgements Barta A, Smolen JS: Purification and partial sequencing of the nuclear We thank Dr Alexander Ploner for his help in autoantigen RA33 shows that it is statistics and Elisabeth Höfler for expert indistinguishable from the A2 protein technical help with determination of of the heterogeneous nuclear autoantibodies and Karolina von Dalwigk for ribonucleoprotein complex. J Clin performing functional assays with Invest 1992, 90:1061-6. supernatants of TCCs. The work was 8) Hassfeld W, Steiner G, Studnicka- supported by CeMM, the Center of Molec- Benke A, Skriner K, Graninger W, ular Medicine of the Austrian Academy of Fischer I, Smolen JS: Autoimmune Sciences. response to the spliceosome: an immunologic link between rheumatoid arthritis, mixed connective tissue References disease, and systemic lupus erythematosus. Arthritis Rheum 1995, 1) Tan EM, Cohen AS, Fries JF, Masi 38:777-85. AT, McShane DJ, Rothfield NF, 9) Steiner G, Skriner K, Smolen JS: Schaller JG, Talal N, Winchester RJ: Autoantibodies to the A/B pro-teins of The 1982 revised criteria for the the heterogeneous nuclear classification of systemic lupus ribonucleoprotein complex: novel tools erythematosus (SLE). Arthritis Rheum for the diagnosis of rheumatic diseases. 1982, 25:1271-7. Int Arch Allergy Immunol 1996, 2) Pisetsky DS: Anti-DNA and 111:314-9. autoantibodies. Curr Opin Rheumatol 10) Skriner K, Sommergruber WH, 2000, 12:364-8. Tremmel V, Fischer I, Barta A, Smo- 3) Schett G, Rubin RL, Steiner G, len JS, Steiner G: Anti-hnRNP-A2 Hiesberger H, Muller S, Smolen J: The autoantibodies are directed to the RNA lupus erythematosus cell phenomenon: binding region of the A2 protein of the comparative analysis of antichromatin heterogeneous nuclear ribonucleo- antibody specificity in lupus protein complex: differential epitope erythematosus cell-positive and - recognition in rheumatoid arthritis, negative sera. Arthritis Rheum 2000, systemic lupus erythematosus, and 43:420-8. mixed connective tissue disease. J Clin 4) Alspaugh MA, Tan EM: Antibodies to Invest 1997, 100:127-35. cellular antigens in Sjogren's 11) Dumortier H, Monneaux F, Jahn- syndrome. J Clin Invest 1975, 55:1067- Schmid B, Briand JP, Skriner K, 73. Cohen PL, Smolen JS, Steiner G, 5) Reyes PA, Tan EM: DNA-binding Muller S: B and T cell responses to the property of Sm nuclear antigen. J Exp spliceosomal heterogeneous nuclear Med 1977, 145:49-754. ribonucleoproteins A2 and B1 in 6) Holyst MM, Hill DL, Hoch SO, normal and lupus mice. J Immunol Hoffmann RW: Analysis of T and B 2000, 165:2297-305. cell responses against U small nuclear 12) Hayer S, Tohidast-Akrad M, ribonucleoprotein 70-kD, B and D Haralambous S, Jahn-Schmid B, polypeptides among patients with Skriner K, Trembleau S, Dumortier H, systemic lupus erythematosus and Pinol-Roma S, Redlich K, Schett G, et mixed connective tissue disease. al.: Aberrant expression of the Arthritis Rheum 1997, 40:1493-503. autoantigen hnRNP-A2/B1 (RA33) and 7) Steiner G, Hartmuth , Skriner K, spontaneous formation of rheumatoid Maurer-Fogy I, Sinski A, Hassfeld W, arthritis associated anti-RA33 autoanti-

69 bodies in TNFα transgenic mice. J erythematosus patients induce double- Immunol 2005, 175:8327-36. stranded DNA antibody production. 13) Weighardt F, Biamonti G, Riva S: The Arthritis Rheum 1997,40:2162-72. roles of heterogeneous nuclear 22) Converso M, Bertero MT, Vallario A, ribonucleoproteins (hnRNP) in RNA Caligaris-Cappio F: Analysis of T-cell metabolism. Bioessays 1996, 18:747- clones in systemic lupus 56. erythematosus. Haematologica 2000, 14) Krecic AM, Swanson MS: HnRNP 85:118-23. complexes: composition, structure, and 23) Talken BL, Bailey CW, Reardon SL, function. Curr Opin Cell Biol 1999, Caldwell CW, Hoffman RW: Structural 11:363-71. analysis of TCRalpha and beta chains 15) Mayeda A, Munroe SH, Caceres JF, from human T-Cell clones specific for Krainer AR: Function of conserved small nuclear ribonucleoprotein domains of hnRNP A1 and other polypeptides Sm-D, Sm-B and U1–70 hnRNP A/B proteins. EMBO J 1994, kDa: TCR complementarity 13:5483-95. determining region 3 usage appears 16) Carson JH, Kwon S, Barbarese E: RNA highly conserved. Scand J Immunol trafficking in myelinating cells. Curr 2001, 54:204-10. Opin Neurobiol 1998, 8:607-12. 24) Greidinger EL, Gazitt T, Jaimes KF, 17) Kamma H, Horiguchi H, Wan L, Hoffman RW: Human T cell clones Matsui M, Fujiwara M, Fujimoto M, specific for heterogeneous nuclear Yazawa T, Dreyfuss G: Molecular ribonucleoprotein A2 autoantigen from characterization of the hnRNP A2/B1 connective tissue disease patients assist proteins: tissue-specific expression and in autoantibody production. Arthritis novel iso-forms. Exp Cell Res 1999, Rheum 2004,50:2216-22. 246:399-411. 25) Gebhard KL, Veldman CM, Wassmuth 18) Smolen JS, Klippel JH, Penner E, R, Schultz E, Schuler G, Hertl M: Ex Reichlin M, Steinberg AD, Chused vivo analysis of desmoglein 1- TM, Scherak O, Graninger W, Hartter responsive T-helper (Th) 1 and Th2 E, Zielinski CC, et al.: HLA-DR cells in patients with pemphigus antigens in systemic lupus foliaceus and healthy individuals. Exp erythematosus: association with Dermatol 2005, 14:586-92. specificity of autoantibody reponses to 26) Diaz-Villoslada P, Shih A, Shao L, nuclear antigens. Ann Rheum Dis Genain CP, Hauser SL: Autoreactivity 1987, 46:457-462. to myelin antigens: 19) Stephens HA, McHugh NJ, Maddison myelin/oligodendrocyte glycoprotein is PJ, Isenberg DA, Welsh KI, Panayi a prevalent autoantigen. J GS: HLA class II restriction of Neuroimmunol 1999,99:36-43. autoantibody production in patients 27) Fritsch R, Eselbock D, Skriner K, Jahn- with systemic lupus erythematosus. Schmid B, Scheinecker C, Bohle B, Immunogengetics 1991, 33:276-80. Tohidast-Akrad M, Hayer S, 20) Mohan C, Adams S, Stanik V, Datta Neumuller J, Pinol-Roma S, et al.: SK: Nuclesosome: a major immunogen Characterization of autoreactive T cells for pathogenic autoantibody-inducing to the autoantigens heterogeneous T cells of lupus. J Exp Med 1993, nuclear ribonucleoprotein A2 (RA33) 177:1367-81. and filaggrin in patients with 21) Voll RE, Roth EA, Girkontaite I, Fehr rheumatoid arthritis. J Immunol H, Herrmann M, Lorenz HM, Kalden 2002,169:1068-76. JR: Histone-specific Th0 and Th1 28) Bencivelli W, Vitali C, Isenberg DA, clones derived from systemic lupus Smolen JS, Snaith ML, Sciuto M,

70 Bombardieri S: Disease activity in erythematosus. Arthritis Rheum 1983, systemic lupus erythematosus: report of 26:745-50. the Consensus Study Group of the 35) Maeda N, Sekigawa I, Matsumoto M, European Workshop for Rheumatology Hashimoto H, Hirose S: Relationship Research. III. Development of a between CD4+/CD8+ T cell ratio and computerised clinical chart and its T cell activation in systemic lupus application to the comparison of erythematosus. Scand J Rheumatol different indices of disease activity. 1999, 28:166-70. The European Consensus Study Group 36) Filaci G, Bacilieri S, Fravega M, for Disease Activity in SLE. Clin Exp Monetti M, Contini P, Ghio M, Setti Rheumatol 1992, 10:549-54. M, Puppo F, Indiveri F: Impairment of 29) Scheinecker C, Machold KP, Majdic O, CD8+ T suppressor cell function in Hocker P, Knapp W, Smolen JS: patients with active Systemic Lupus Initiation of the autologous mixed Erythmatosus. J Immunol 2001, lymphocyte reaction requires the 166:6452-7. expression of costimulatory molecules 37) Filaci G, Fravega M, Negrini S, B7-1 and B7-2 on human peripheral Procopio F, Fenoglio D, Rizzi M, blood dendritic cells. J Immunol 1998, Brenci S, Contini P, Olive D, Ghio M, 161:3966-73. et al.: Nonantigen specific CD8+ T 30) Fritsch R, Bohle B, Vollmann U, suppressor lymphocytes originate from Wiedermann U, Jahn-Schmid B, CD8+CD28- T cells and inhibit both Krebitz M, Breiteneder H, Kraft D, T-cell proliferation and CTL function. Ebner C: Bet v 1, the major birch Hum Immunol 2004, 65:142-56. pollen allergen, and Mal d 1, the major 38) Liu Z, Tugulea S, Cortesini R, Suciu- apple allergen, cross-react at the level Foca N: Specific suppression of T of allergen-specific T helper cells. J helper alloreactivity by allo-MHC class Allergy Clin Immunol 1998, 102:679- I restriced CD8+CD28- T cells. Int 86. Immunol 1998, 10:775-83. 31) Isenberg DA, Steiner G, Smolen JS: 39) Colovai AI, Liu Z, Ciubotariu R, Clinical utility and serological Lederman S, Cortesini R, Suciu-Foca connections of anti-RA33 antibodies in N: Induction of xenoreactive CD4+ T systemic lupus erythematosus. J cell anergy by suppressor CD8+CD28 Rheumatol 1994, 21:1260-1263. Tcells. Transplantation 2000,69:1304- 32) Richter-Cohen M, Steiner G, Smolen 10. JS, Isenberg DA: Erosive arthritis in 40) Najafian N, Chitnis T, Salama AD, Zhu systemic lupus erythematosus: analysis B, Benou C, Yuan X, Clarkson MR, of a distinct clinical and serological Sayegh MH, Khoury SJ: Regulatory subgroup. Br J Rheumatol functions of CD8+CD28- T cells in an 1998,37:421-4. auoimmune disease mode. J Clin 33) Smolen JS, Chused TM, Leiserson Invest 2003, 112:1037-48. WN, Reeves JP, Alling D, Stein- berg 41) Vallejo AN: CD28 extinction in human AD: Heterogeneity of T cells: altered functions and the immunoregulatory T-cell subsets in program of T-cell senescence. systemic lupus erythematosus. Am J Immunol Rev 2005,205:158-69. Med 1982, 72:783-90. 42) Houssiau FA, Lefebvre C, Vanden 34) Bakke AC, Kirkland PA, Kitridou RC, Berghe M, Lambert M, Devogelaer JP, Quismorio FP Jr, Rea T, Ehresmann Renauld JC: Serum IL-10 titres in GR, Horwitz AD: T lymphocyte systemic lupus erythematosus reflect subsets in systemic lupus disease activity. Lupus 1995, 4:393-5.

71 43) Llorente L, Richaud-Patin Y, Fior R, Alcocer-Varela J, Wijdenes J, Fourrier BM, Galanaud P, Emilie D: In vivo production of interleukin-10 by non-T cells in rheumatoid arthritis, Sjogren's syndrome, and systemic lupus erythematosus. Arthritis Rheum 1994, 37:1647-56. 44) Llorente L, Zou W, Levy Y, Richaud- Patin Y, Wijdenes J, Alcocer- Varela J, Morel-Fourrier B, Brouet JC, Alarcon-Segovia D, Galanaud P, Emilie D: Role of interleukin-10 in the B lymphocyte hyperactivity and autoantibody production of human systemic lupus erythematosus. J Exp Med 1995, 181:839-44.

Chapter 6

Characterization of Autoreactive T Cells to the Autoantigens Heterogeneous Nuclear Ribonucleoprotein A2 (RA33) and Filaggrin in Patients with Rheumatoid Arthritis

RD Fritsch1 D Eselböck1 K Skriner1,2 B Jahn-Schmid1,3 C Scheinecker1 B Bohle3 M Tohidast-Akrad5 S Hayer1 J Neumüller4 S Pinol-Roma6 JS Smolen1,5 G Steiner1,2

1Division of Rheumatology, Department of Internal Medicine III 2,3,4 Institutes of 2Medical Biochemistry, 3Pathophysiology, and 4Histology, University of Vienna, Vienna,Austria; 5Ludwig Boltzmann Institute for Rheumatology and Balneology, Vienna, Austria 6Department of Cell Biology and Anatomy, Mount Sinai School of Medicine, New York, NY 10029

The Journal of Immunology. 2002, 169:1068–1076.

73 Characterization of Autoreactive T Cells to the Autoantigens Heterogeneous Nuclear Ribonucleoprotein A2 (RA33) and Filaggrin in Patients with Rheumatoid Arthritis1

Ruth Fritsch,* Daniela Eselbo¨ck,* Karl Skriner,*† Beatrice Jahn-Schmid,*‡ Clemens Scheinecker,2* Barbara Bohle,‡ Makiyeh Tohidast-Akrad,¶ Silvia Hayer,* Josef Neumu¨ller,§ Seraﬁn Pinol-Roma,ʈ Josef S. Smolen,*¶ and Gu¨nter Steiner3*†

The role of autoimmune reactions in the pathogenesis of rheumatoid arthritis (RA) is poorly understood. To address this issue we have investigated the spontaneous T cell response to two well-characterized humoral autoantigens in RA patients and controls: 1) the heterogeneous nuclear ribonucleoprotein A2, i.e., the RA33 Ag (A2/RA33), and 2) filaggrin in unmodified and citrullinated forms. In stimulation assays A2/RA33 induced proliferative responses in PBMC of almost 60% of the RA patients but in only 20% of the controls (patients with osteoarthritis or psoriatic arthritis and healthy individuals), with substantially stronger responses in RA patients (p < 0.00002). Furthermore, synovial T cells of seven RA patients investigated were also clearly responsive. In contrast, responses to filaggrin were rarely observed and did not differ between RA patients and controls. Analysis of A2/RA33- induced cytokine secretion revealed high IFN-␥ and low IL-4 production in both RA and control PBMC, whereas IL-2 production was mainly observed in RA PBMC (p < 0.03). Moreover, A2/RA33-specific T cell clones from RA patients showed a strong Th1 phenotype and secreted higher amounts of IFN-␥ than Th1 clones from controls (p < 0.04). Inhibition experiments performed with mAbs against MHC class II molecules showed A2/RA33-induced T cell responses to be largely HLA-DR restricted. Finally, immunohistochemical analyses revealed pronounced overexpression of A2/RA33 in synovial tissue of RA patients. Taken together, the presence of autoreactive Th1-like cells in RA patients in conjunction with synovial overexpression of A2/RA33 may indicate potential involvement of this autoantigen in the pathogenesis of RA. The Journal of Immunology, 2002, 169: 1068–1076.

heumatoid arthritis (RA)4 is a chronic inflammatory joint Already in early disease stages the inflamed rheumatoid syno- disease of unknown etiology characterized by hyperpla- vial membrane is characterized by massive infiltration of T cells sia of the synovial membrane, degradation of cartilage, and other immune cells including B cells, macrophages, and mast R ϩ and bone erosion. Overexpression of proinflammatory cytokines cells (3–5). The majority of synovial T cells are CD4 memory such as IL-1 and TNF-␣ is considered to drive the destructive cells that typically express activation markers such as HLA-DR, processes, but the causes for this disregulated cytokine production CD69, or CD40L but, remarkably, may be deficient in CD28 ex- are unknown. Although it is generally assumed that RA belongs to pression (6). Cytokine analyses performed in situ revealed a pre- the group of systemic autoimmune diseases, the pathogenetic role dominance of IFN-␥ over IL-4-producing T cells, although the of cellular and humoral autoimmune reactions is still incompletely numbers of cytokine-secreting T cells were found to be relatively understood (1, 2). low (7–9). Furthermore, the T cell repertoire of RA patients shows features of clonal expansion (6) and most T cell clones (TCC) obtained from synovial tissue or fluid could be functionally attrib- *Division of Rheumatology, Department of Internal Medicine III, and Institutes of uted to the Th1 subset (10–12), which is generally considered to †Medical Biochemistry, ‡Pathophysiology, and §Histology, University of Vienna, Vi- constitute a driving force of pathologic autoimmune reactions (13– enna, Austria; ¶Ludwig Boltzmann Institute for Rheumatology and Balneology, Vi- ʈ enna, Austria; and Department of Cell Biology and Anatomy, Mount Sinai School of 15). However, in vitro, synovial T cells generally show some fea- Medicine, New York, NY 10029 tures of anergy such as reduced responsiveness to mitogens or Received for publication July 19, 2001. Accepted for publication May 10, 2002. recall Ags, which has been attributed to the effects of chronic The costs of publication of this article were defrayed in part by the payment of page TNF-␣ exposure and oxidative stress in the joint (16, 17). charges. This article must therefore be hereby marked advertisement in accordance The abundance of T cells and the association of RA suscepti- with 18 U.S.C. Section 1734 solely to indicate this fact. bility with particular MHC class II alleles suggests an important if 1 This work was supported by the Interdisciplinary Cooperation Project of the Aus- trian Federal Ministry for Education, Science, and Culture and in part by Austrian not pivotal role of T cells in the pathogenesis of RA, which has National Bank Grant 7021. been widely discussed in recent years (2, 18–22). Moreover, the 2 Current address: Laboratory of Immunology, National Institute of Allergy and In- presence of high-titer rheumatoid factor (RF) and other autoanti- fectious Diseases, National Institutes of Health, Bethesda, MD 20892. bodies (Aab) in the sera of RA patients is considered a further 3 Address correspondence and reprint requests to Dr. Gu¨nter Steiner, Division of indication for involvement of autoimmune processes, although di- Rheumatology, Department of Internal Medicine III, University Hospital of Vienna, Wa¨hringer Gu¨rtel 18-20, A-1090 Vienna, Austria. E-mail address: Guenter.Steiner@ rect evidence for a role of Aab in the pathophysiology of RA is akh-wien.ac.at scarce (23). Based on these observations, an autoimmune model of 4 Abbreviations used in this paper: RA, rheumatoid arthritis; RF, rheumatoid factor; RA pathogenesis has been suggested in which autoreactive T cells Aab, autoantibody; hnRNP, heterogeneous nuclear ribonucleoprotein; Fil, filaggrin; are initially activated by Ags presented via disease-associated AFA, anti-Fil Aab; OA, osteoarthritis; PSA, psoriatic arthritis; cFil, citrullinated Fil; TT, tetanus toxoid; SFMC, synovial fluid mononuclear cell; SI, stimulation index; HLA molecules (2). These T cells subsequently activate autoreac- SN, supernatant; TCC, T cell clone; CRP, C-reactive protein. tive B cells, macrophages, and synoviocytes, which may further

Copyright © 2002 by The American Association of Immunologists, Inc. 0022-1767/02/$02.00 stimulate T cell responses. Thus, a vicious circle is induced and tigation patients were treated with nonsteroidal anti-inflammatory drugs maintained that leads to chronic inflammation and finally results in (n ϭ 32), disease-modifying antirheumatic drugs (n ϭ 34), and/or low- ϭ irreversible destruction of the joint. dose glucocorticoids (n 18). While 37 patients had moderately active disease with fewer than five swollen joints and C-reactive protein (CRP) The search for Ags triggering pathogenic T cell responses has Յ2 mg/dl, 13 patients had active RA (i.e., more than six swollen joints and led to the identification of a number of joint-specific candidate CRP Ͼ2 mg/dl). Synovial fluids were obtained during routine joint tapping proteins including collagen, cartilage link protein, and gp39 (23, from seven of the patients using heparinized tubes. The control population Ϯ 24). However, even though arthritis may be induced in susceptible consisted of 18 patients with OA (all female, mean age 59.2 12 years), 11 patients with PSA (four male, seven female, mean age 45 Ϯ 17 years), mouse and rat strains by immunization with these Ags, their patho- and 21 healthy individuals (10 female, 11 male, mean age 39.3 Ϯ 13.4 genetic role has remained elusive. In addition to joint-specific Ags, years), including two otherwise healthy persons with osteoporosis. Four of several other proteins have been identified as targets of Aab of the PSA patients had active disease with more than two swollen joints patients with RA. Interestingly, even though these Ags are more or and/or elevated serum CRP levels. HLA-DR genotyping as assessed by less ubiquitously expressed, the Aab directed to them seem to be more PCR revealed that 66% of the RA patients and 30% of the tested controls carried HLA-DR4 and/or DR1. specific for RA than anti-collagen Aab or RF (23). These include the following: Aab to the heterogeneous nuclear ribonucleoprotein (hnRNP) A2, also known as the RA33 Ag (A2/RA33) (25); anti- Antigens keratin Aab, which recognize the cytokeratin-aggregating protein Recombinant full-length hnRNP-A2/RA33 was used throughout this study. filaggrin (Fil) (26); anti-Sa Aab, whose target moiety is still unknown The cDNA encoding the Ag (45) was cloned by ligation-independent clon- (27); and Aab to the H chain binding stress protein (28). ing into the pET-30 LIC vector (Novagen, Madison, WI) and expressed as Aab to A2/RA33 can be detected in about one-third of RA pa- His-tagged fusion protein. Purification from bacterial lysates was achieved by nickel-nitrilotriacetic acid affinity chromatography (Qiagen, Hilden, tients and may also occur in patients with systemic lupus erythem- Germany) followed by polymyxin B Sepharose adsorption (Bio-Rad, Her- atosus or mixed connective tissue disease, usually in association cules, CA) and anion exchange chromatography on DEAE Sepharose with anti-spliceosomal Aab (such as anti-Sm or anti-U1RNP) (Pharmacia Biotech, Uppsala, Sweden). Endotoxin content was determined which are not found in RA (29, 30). A2/RA33 is an abundant by the lympholytic amebocyte lysate assay (BioWhittaker, Verviers, Bel- Ͼ mRNA binding protein that, like other hnRNP proteins, shows a gium). By this procedure a 99% pure, endotoxin-free preparation was obtained. The optimum concentration for proliferation assays was found to modular structure consisting of two conserved RNA binding do- be 0.35 ␮g/ml. mains and a glycine-rich auxiliary domain assumed to be involved Recombinant His-tagged Fil and cFil were a kind gift from Dr. A. Union in interactions with other proteins (31, 32). It has a predominant (Innogenetics, Ghent, Belgium) and were used at a concentration of 2.5 ␮ nuclear localization and exerts multiple functions, including reg- g/ml. In addition, two synthetic cFil-derived peptides as well their unmodified isoforms were provided and used at 2.5 ␮g/ml. These peptides ulation of alternative splicing and transport of mRNA (33, 34). were selected on the basis of recognition by AFA (46) and had the fol- A2/RA33 appears to be ubiquitously expressed, although the level lowing sequences (X ϭ citrulline, i.e., deiminated arginine): HSAS of expression may greatly vary between different tissues (35). QDGQDTIXGHPGSS (filpep 1) and DSGHXGYSGSQASQDNEGH Anti-keratin or anti-Fil Aab (AFA), respectively, can be de- (filpep 2). tected by indirect immunofluorescence or immunoblotting in 40– Tetanus toxoid (TT) as control Ag was obtained from Pasteur Merieux Connaught (Willowdale, Ontario, Canada) and used at a concentration of 50% of RA patients (36). They are rather specific for RA and 0.5 U/ml as previously described (47). recognize the citrullinated form of Fil (cFil), which is generated by posttranslational deimination of certain arginine residues (37). Fil is exclusively expressed in terminally differentiated epithelial Ab detection cells, where it is involved in aggregation of cytokeratin filaments RF was determined by nephelometry. Anti-A2/RA33 Aab were detected by and presumably plays a role during apoptosis (38). Remarkably, ELISA (IMTEC, Berlin, Germany) and by immunoblotting as described when citrullinated peptides were used instead of the whole protein, (29). AFA were detected by an ELISA (Eurodiagnostics, Arnhem, The Ͼ Netherlands) using cFil-derived peptides (39). RF, anti-A2/RA33 Aab, and 60% of the patients were found to be reactive, even with peptides AFA were present in 57, 22, and 50% of the RA patients, respectively. not derived from Fil (39, 40). This has led to speculations on the existence of other citrullinated targets, and recently strong evidence has been provided that such proteins are indeed present in T cell stimulation assays synovial tissue of RA patients (41). PBMC or synovial fluid mononuclear cells (SFMC) were isolated from Anti-A2/RA33 and AFA are already present in early stages of fresh heparinized blood or synovial fluid samples of RA patients and con- the disease (40, 42–44) and are predominantly of the IgG isotype, trols by centrifugation on Ficoll-Hypaque (Pharmacia Biotech). After which is indicative of T cell-driven processes. To gain more in- washing and counting, cells were either immediately used or frozen in RPMI medium containing 10% DMSO and 20% FCS. Cells were cultured sight into such processes and their potential pathogenetic role, we for 5 days at 37°C in triplicate in 96-well plates (Costar, Cambridge, MA) investigated the spontaneous T cell responses to A2/RA33 and Fil in a total of 200 ␮l (105 cells/well) in the presence of the Ags. Culture in patients with RA and in control subjects including patients with medium consisted of Ultra Culture serum-free medium (BioWhittaker, osteoarthritis (OA) and psoriatic arthritis (PSA) and healthy per- Walkersville, MD) containing 2 mM glutamine and 0.02 mM 2-ME supplemented with 100 U/ml penicillin/streptomycin (Life Technologies, Pais- sons. The data obtained suggest that A2/RA33 may constitute an ley, U.K.). PHA (Life Technologies) and IL-2 (Roche Molecular Bio- important T cell autoantigen in patients with RA, whereas Fil does chemicals, Mannheim, Germany) were used as polyclonal stimuli. During not seem to be a major T cell autoantigen and therefore presum- the last 16 h of culture, 0.5 ␮Ci/well [3H]TdR (Amersham Biosciences ably does not drive the humoral autoimmune response to Europe, Freiburg, Germany) was added and the incorporated radioactivity citrullinated Ags. was measured by scintillation counting. Results were expressed as stimulation index (SI) defined as the ratio of mean cpm obtained in cultures with Ag to mean cpm obtained in cultures incubated in the absence of Ag. SI Ն2 Materials and Methods and ⌬cpm Ͼ1000 (mean cpm obtained in cultures with Ag minus mean Patients and controls cpm obtained in cultures incubated in the absence of Ag) was regarded as a positive response. Peripheral blood from 50 patients with RA (37 female, 13 male, mean age Cytokines were measured by ELISA (BioSource, Fleurus, Belgium) in 55 Ϯ 17.7 years) classified according to the established criteria of the supernatants (SN) after a 24-h incubation with A2/RA33 or 0.5 U/ml TT as American College of Rheumatology was drawn into heparinized test tubes. control Ag. Detection limits were 5 pg/ml for IL-2, 4 pg/ml for IL-4, and Informed consent was obtained from all patients. At the time of the inves- 9 pg/ml for IFN-␥. Blocking experiments Inhibition of Ag-induced T cell activation was investigated by incubating PBMC with A2/RA33 in the presence or absence of 5 ␮g of mAb specific for HLA-DR (clone G46-6) or HLA-DR/DP/DQ (clone TU¨ 39), respectively (BD PharMingen, San Diego, CA) known to block MLR. An isotype-matched mAb to ␤-galactosidase was used as control Ab (Zymed Laboratories, San Francisco, CA). Proliferation was measured by [3H]TdR uptake and IFN-␥ production was measured by ELISA. T cell lines and clones Ag-specific T cell lines were obtained using a previously established protocol (48, 49). In brief, 2 ϫ 106 PBMC were stimulated with A2/RA33 for 5 days in 24-well flat-bottom culture plates. On day 5 of culture 20 U/ml IL-2 was added and the culture was continued for an additional 7 days. To generate TCC, T cell lines were restimulated with A2/RA33 and after 2 more days viable T cell blasts were separated by Ficoll-Hypaque and seeded in limiting dilution (0.5 cells/well) in 96-well plates. T cell blasts FIGURE 1. Proliferative responses of PBMC from RA patients and were cultured in the presence of 1 ϫ 105 irradiated (5000 rad) allogeneic controls to A2/RA33. Proliferation was measured by [3H]TdR incorpora- PBMC as feeder cells, 0.5 ␮g/ml PHA, and 20 U/ml IL-2 in medium tion in PBMC from 50 RA patients and 50 controls (18 patients with OA, containing 1% heat-inactivated human AB sera. Growing microcultures 11 patients with PSA, and 21 healthy subjects) in the presence or absence were then expanded at weekly intervals with fresh feeder cells in the pres- of A2/RA33. A SI Ն2 and a ⌬cpm Ͼ1000 cpm were considered a positive ence of IL-2. The specificity of TCC was assessed by proliferation assays response; the cut-off value is indicated by a solid line. Mean SI was 4 Ϯ 3.5 incubating 2 ϫ 104 T cell blasts with A2/RA33 (0.35 ␮g/ml) in the pres- 5 for RA patients and 1.5 Ϯ 1.1 for controls (p Ͻ 0.00002); no significant ence of 10 autologous irradiated PBMC. After a 48-h incubation and puls- Ϯ ing with [3H]TdR for additional 16 h, cells were harvested and the incor- differences were seen among the three control groups (mean SI was 1.6 Ϯ Ϯ porated radioactivity was measured by scintillation counting. Production of 1.2 in OA patients, 1.2 1.5 in PSA patients, and 1.5 0.8 in healthy IL-4 and IFN-␥ was measured by ELISA in SN collected after 24 h of subjects). incubation. For phenotyping, cloned T cells (5 ϫ 104–105) were washed twice with ice-cold FACS buffer (PBS, 5% FCS, 0.01% NaN3) by centrifugation for 5 min at 1000 ϫ g. The washed cells were incubated for 30 min at 4°C with SI did not differ among the three control groups (see Fig. 1). More- ␣␤ a FITC- or PE-conjugated mAb (BD PharMingen). Anti-TCR and anti- over, SI Ն4 were seen in 18 RA patients (maximum SI ϭ 15) but CD4 mAb were FITC conjugated, and anti-TCR␥␦ and anti-CD8 mAb ϭ were PE conjugated. Afterward, cells were washed again in FACS buffer in only two controls (maximum SI 5.8). and analyzed with a FACScan flow cytometer (BD Biosciences, Franklin In RA patients the T cell response did not correlate with the Lakes, NJ) supported by PC Lysis software (BD Biosciences). presence of anti-A2/RA33 Aab, which were detected in 22% of the RA patients but in neither healthy controls nor patients with OA or Immunohistochemistry PSA, confirming previous observations (29, 44). Thus, T cell re- Immunohistochemical analysis of synovial tissue was performed as de- activity to A2/RA33 was seen in 8 of the 11 (73%) anti-A2/RA33 scribed previously (9). To assess expression of A2/RA33, cryostat sections Aab-positive patients and also in 18 of the 39 (46%) anti- from synovial tissue of five patients with RA and three patients with OA were incubated for 60 min with the anti-A2/RA33 mAb 10D1 (50). After A2/RA33-negative patients. Although 73% of the reactive RA pa- rinsing and blocking of endogenous peroxidase, sections were incubated tients vs 57% of the nonreactive ones carried HLA-DR4 and/or for 30 min with a biotinylated horse anti-mouse IgG Ab followed by in- DR1, this association did not reach the level of statistical cubation with the Vectastain@ABC reagent (Vector Laboratories, Burlin- significance. game, CA) for another 30 min using diaminobenzidine (Sigma-Aldrich, St. Louis, MO) as substrate leading to brown staining of A2/RA33-expressing Proliferative responses of synovial T cells to A2/RA33 cells. Finally, slides were counterstained with hematoxylin (Merck, Darm- stadt, Germany). To investigate expression of A2/RA33 in macrophages, To address whether cellular reactivity against A2/RA33 was also tissues were double stained with 10D1 and an anti-CD68 mAb (DAKO, present in the synovial compartment, proliferation was measured Glostrup, Denmark) using an alkaline phosphatase-based detection system in SFMC obtained from seven RA patients; proliferation of the (DAKO) leading to blue staining of the CD68-positive cells. Additional double stainings were performed with an anti-CD3 mAb (BD PharMingen) corresponding PBMC was determined in parallel. In six patients, and the fibroblast-specific mAb ASO2 (Dianova, Hamburg, Germany). Iso- both SFMC and PBMC proliferated significantly, while in one type-matched mAb (DAKO) served as negative controls. patient only SFMC were responsive (Fig. 2). Of note, with one Statistical analysis Unless stated otherwise, SI or cytokine concentrations, respectively, are indicated as mean Ϯ SD. Student’s t test was used to determine differences between groups. Where appropriate, Bonferroni corrections were done. A p value Ͻ0.05 was regarded as significant. Results Proliferative responses of PBMC to A2/RA33 Cellular reactivity against A2/RA33 was investigated by measuring proliferation of PBMC obtained from 50 RA patients and 50 controls, including 18 patients with OA, 11 patients with PSA, and 21 healthy subjects. As shown in Fig. 1, pronounced responses were observed in 58% of the RA patients and in 20% of the con- FIGURE 2. Proliferative responses of SFMC from RA patients to A2/ trols (six healthy subjects, three OA patients, and one PSA pa- RA33. SFMC and PBMC obtained from seven RA patients were cultured Ϯ ϭ tient). In RA patients mean SI was 4 3.5 (median SI 2.6), in parallel in the presence or absence of A2/RA33 and proliferation was whereas in controls mean SI amounted to 1.5 Ϯ 1.1 (median SI ϭ determined by measuring [3H]TdR incorporation. Except for patient 1, 1.1). This difference was highly significant ( p Ͻ 0.00002), while SFMC always showed a stronger response than the corresponding PBMC. FIGURE 3. Cytokine production of PBMC from RA patients and controls in response to A2/RA33. PBMC from 15 RA patients and eight controls were incubated in the presence or absence of A2/RA33, and secretion of IL-2, IFN-␥, and IL-4 was measured in culture SN by ELISA. Back- ground levels were generally below the detection limit. Mean values are indicated by solid lines. A significant difference between RA patients and controls was seen only for IL-2.

exception, the SI of SFMC was always higher than the SI of the was obtained with the second anti-HLA class II mAb (data not corresponding PBMC, although the difference was not significant. shown). In contrast, an isotype-matched control Ab had no significant effect (Fig. 4, C and D). Cytokine secretion by PBMC in response to A2/RA33 To further characterize A2/RA33-induced T cell responses, the T cell reactivity to Fil cytokines IL-2, IL-4, and IFN-␥ were measured by ELISA in cul- The proliferative responses to both Fil and cFil were investigated ture SN of PBMC from 15 responsive RA patients and eight con- in PBMC of 19 RA patients and 20 healthy controls. As shown in trol subjects stimulated with A2/RA33 or the control Ag TT, re- Fig. 5, proliferative responses to either form of Fil were observed spectively (Fig. 3). These investigations revealed significantly in only four RA patients and two controls: three patients responded higher IL-2 production by PBMC of RA patients (10.3 Ϯ 9.1 vs to both forms and the fourth patient recognized only Fil. Interest- 4.1 Ϯ 3.3 pg/ml, p Ͻ 0.03), with IL-2 detectable in 10 SN of RA ingly, cFil always elicited a lower response than Fil, indicating that patients and in three SN of controls (Fig. 3A). A similar result was arginine deimination may affect T cell recognition. Of the two obtained with IL-4, but here the difference between the two groups controls, one responded to Fil and the other one responded to cFil. did not reach the level of statistical significance (Fig. 3B). In con- Mean SI of PBMC of RA patients was 1.7 Ϯ 1.4 for Fil and 1.1 Ϯ trast to IL-2 and IL-4, high IFN-␥ levels were measured in virtu- 0.7 for cFil, respectively, which did not significantly differ from ally all SN of both RA patients and controls (Fig. 3C). Levels were the SI values obtained in controls (1.1 Ϯ 0.6 for Fil and 1.3 Ϯ 0.6 similar in both groups ( p ϭ NS) and were an order of magnitude for cFil, respectively). In contrast, the proliferative responses to higher than IL-4 levels. However, when A2/RA33-induced IFN-␥ A2/RA33 of these 19 RA patients were significantly higher, with production was compared with TT-induced production, a remark- a mean SI of 4.1 Ϯ 3.5 ( p Ͻ 0.02 vs Fil and p Ͻ 0.002 vs cFil, able difference became apparent: thus, in SN of RA patients IFN-␥ respectively). levels were markedly higher upon A2/RA33 stimulation than upon To learn whether citrullinated peptides could elicit a better T stimulation with TT (484 Ϯ 379 vs 117 Ϯ 97 pg/ml, p Ͻ 0.003), cell response, two synthetic cFil-derived peptides harboring major whereas PBMC from controls responded to both Ags in a compa- B cell epitopes (46) as well as their corresponding unmodified rable manner (875 Ϯ 672 vs 1376 Ϯ 1918 pg/ml, p ϭ NS). There- isoforms were used in proliferation assays involving PBMC from fore, TT-induced IFN-␥ production was significantly higher in seven RA patients, three of whom had responded to the whole controls than in RA patients ( p Ͻ 0.02). Concerning IL-4, A2/ RA33 induced a stronger response than TT only in RA patients, although compared with IFN-␥ the difference was less pronounced ( p Ͻ 0.05). However, with respect to IL-2, no difference between the two stimulants was seen in RA patients, whereas in controls TT induced a stronger IL-2 response than A2/RA33 ( p Ͻ 0.01). It is noteworthy that this response was comparable to the IL-2 response induced in RA patients by either A2/RA33 or TT, respectively. Taken together, these data suggested that A2/RA33-responsive T cells belong predominantly to the Th1 subset in both RA patients and controls. However, in RA patients these cells secreted IL-2 and appeared to be expanded as indicated by increased IFN-␥ production upon A2/RA33 stimulation relative to TT-induced stimulation.

MHC restriction of the A2/RA33-induced T cell response To investigate whether A2/RA33-induced T cell activation was dependent on Ag presentation by MHC class II molecules, two FIGURE 4. ␥ mAbs directed to HLA-DR or to HLA-DR/DP/DQ, respectively, Suppression of A2/RA33-induced IFN- production and proliferation by an anti-HLA-DR mAb. PBMC from eight RA patients and were used in inhibition experiments performed with PBMC of three controls were incubated with Ag in the presence of a mAb against eight responsive patients and three responsive controls (Fig. 4). As HLA-DR (A and B) or an isotype-matched control mAb (C and D). IFN-␥ shown in Fig. 4A, addition of the anti-HLA-DR Ab led to a marked secretion (A and C) was measured by ELISA, and proliferation (B and D) ␥ reduction of IFN- secretion in all patients and controls (43–90% was determined by measuring [3H]TdR incorporation (cpm). The anti- decrease, p Ͻ 0.02). Proliferation was inhibited to an even higher HLA-DR mAb caused signiﬁcant reduction of IFN-␥ secretion and prolif- degree (63–94% decrease, p Ͻ 0.0003) (Fig. 4B). A similar result eration, whereas the control mAb had no signiﬁcant effect. Table II. Phenotype, proliferation, and cytokine production of A2/ RA33-reactive TCC derived from RA patients and healthy controls

IFN-␥ IL-4 Clone Phenotype SI (pg/ml) (pg/ml)

RA patientsa ϩ Ϫ HWRA.10 CD4 CD8 3.7 Ͼ2000 ϩ Ϫ HWRA.33 CD4 CD8 3.4 Ͼ2000 ϩ Ϫ HWRA.43 CD4 CD8 4.0 Ͼ2000 ϩ Ϫ HWRA.91 CD4 CD8 3.8 Ͼ2000 ϩ Ϫ HWRA.14 CD4 CD8 29.0 1014 5 ϩ ϩ HWRA.151 CD4 CD8 2.8 318 ϩ Ϫ KBRA.113 CD4 CD8 3.0 1033 15 ϩ Ϫ KBRA.15 CD4 CD8 3.4 136 ϩ Ϫ KBRA.110 CD4 CD8 9.4 59 ϩ Ϫ FIGURE 5. Proliferative responses of PBMC from RA patients and KBRA.107 CD4 CD8 3.1 48 3 controls to Fil. Proliferation was measured by [ H]TdR incorporation in LLRA.114 ND 47.0 Ͼ2000 PBMC from 19 RA patients and 20 controls in the presence or absence of LLRA.75 ND 2.3 33 ϩ Ϫ either Fil or cFil. A SI Ն2 and a ⌬cpm Ͼ1000 cpm was considered a SSRA.130 CD4 CD8 3.3 Ͼ2000 ϩ Ϫ SSRA.41 CD4 CD8 3.1 Ͼ2000 positive response. The cut-off value is indicated by a solid line. Mean SI of ϩ Ϫ Ϯ Ϯ WWRA.175 CD4 CD8 2.2 Ͼ2000 PBMC of RA patients was 1.7 1.4 for Fil and 1.1 0.7 for cFil; mean ϩ Ϫ SI of PBMC of controls was 1.1 Ϯ 0.6 for Fil and 1.3 Ϯ 0.6 for cFil. These KRRA.2 CD4 CD8 2.8 Controlsb differences were not significant. ϩ Ϫ RMHC.200 CD4 CD8 2.6 Ͼ2000 ϩ Ϫ RMHC.80 CD4 CD8 2.3 154 ϩ ϩ RMHC.126 CD4 CD8 3.6 Ͼ2000 12 ϩ ϩ RMHC.140 CD4 CD8 2.7 Ͼ2000 molecule, and 13 controls (Table I). Among the three responsive ϩ ϩ RMHC.116 CD4 CD8 2.3 Ͼ2000 patients, only patient A responded selectively to the citrullinated Ϫ ϩ RMHC.87 CD4 CD8 4.7 Ͼ2000 ϩ Ϫ form of peptide 1 (and to both forms of peptide 2), whereas pa- BSHC.102 CD4 CD8 37.0 39 ϩ Ϫ tients B and C responded only to the unmodified peptides. One of BSHC.83 CD4 CD8 5.5 36 ϩ Ϫ the four nonresponders (patient D) was reactive with both forms of BSHC.99 CD4 CD8 2.1 12 Ϫ ϩ BSHC.16 CD4 CD8 4.5 53 peptide 2 and one of the 13 controls responded to unmodified ϩ Ϫ WiWiHC.11 CD4 CD8 19.0 360 peptide 1. Thus, these data confirmed that deimination can affect T ϩ Ϫ WiWiHC.41 CD4 CD8 2.2 68 cell recognition. a Sixteen TCC were derived from six RA patients. Mean SI was 7.9 Ϯ 12.3; ϩ Ϫ median SI was 3.4. All but one clone showed a CD4 CD8 phenotype (clone A2/RA33-reactive TCC ϩ ϩ HWRA.151 was CD4 CD8 ). The majority of clones showed a pronounced Th1 phenotype (IFN-␥ Ն100 pg/ml) and secreted significantly more IFN-␥ than T cell lines specific for A2/RA33 were established from the pe- ϩ Ϫ CD4 CD8 control clones ( p Ͻ 0.04). ripheral blood of six RA patients and three healthy controls. Using b Twelve TCC were derived from three healthy controls. Mean SI was 7.4 Ϯ 10.4; ϩ Ϫ the limiting dilution technique, between one and six Ag-specific median SI was 3.2. Seven TCC showed a CD4 CD8 phenotype, two clones were Ϫ ϩ CD4 CD8 (RMHC.87 and BSHC.16), and three clones were double positive TCC per individual could be generated (Table II). Sixteen TCC ϩ Ϫ (RMHC.126, 140, and 116). Only three CD4 CD8 clones (RMHC.200, RMHC.80, were derived from RA patients, the majority of which (i.e., 13 of and WiWiHC.11) showed a pronounced Th1 phenotype (IFN-␥ Ն100 pg/ml). ϩ Ϫ the 14 TCC analyzed) were CD4 CD8 and thus belonged to the Th cell subset. The TCC showed high variability in their proliferative responses, with SI ranging from 2.2 to 47 and a mean SI of ϩ ϩ were CD4 CD8 (all from donor RM). Although all TCC pro- 7.9 Ϯ 12.3. Analysis of the cytokine secretion pattern, in contrast, duced IFN-␥ (and no IL-4), only five of them secreted high revealed high IFN-␥ production (i.e., Ն100 pg/ml) by the majority amounts (Ͼ1 ng/ml), which, interestingly, were all derived from of the clones with levels in the SN exceeding 1 ng/ml in 10 of donor RM. However, only one of these high producers them, while IL-4 was generally not detectable. Cytokine produc- ϩ Ϫ (RMHC.200) showed a Th1 phenotype (CD4 CD8 ), while the tion did not correlate with proliferation because even poorly pro- Ϫ ϩ remaining four clones were either CD4 CD8 or double positive. liferating TCC secreted large amounts of IFN-␥. ϩ Ϫ Among the six remaining CD4 CD8 clones only two secreted Proliferation of the 12 TCC derived from control individuals IFN-␥ Ͼ100 pg/ml (RMHC.80 and WiWiHC.11). Thus, Th1 was comparable to patient-derived clones with SI ranging from 2.1 clones from controls produced significantly less IFN-␥ than Th1 to 37 with a mean SI of 7.4 Ϯ 10.4. In contrast, phenotyping clones from RA patients ( p Ͻ 0.04). revealed some differences because only seven TCC were ϩ Ϫ Ϫ ϩ CD4 CD8 , while two clones were CD4 CD8 and three clones Expression of A2/RA33 in synovial tissue Expression at the protein level was studied in synovial tissue from five RA and three OA patients using a mAb against A2/RA33 (Fig. Table I. Proliferative responses of PBMC from RA patients to Fil- derived peptidesa 6). These investigations revealed the autoantigen to be abundantly expressed in RA synovial tissue, particularly in the lining layer, Patient cFil Fil cfilpep 1 filpep 1 cfilpep 2 filpep 2 where macrophage-like type A synoviocytes are the predominating cell population, and in the sublining area (Fig. 6A). Double stain- A36 4.1 1.7 5.7 5.8 ing experiments confirmed abundant expression by CD68-positive B22.3 1.8 2.2 1.8 2.2 macrophages (Fig. 6, B and C) and CD68-negative fibroblast-like C 0.8 4 1.4 3.5 1.2 2.1 D 0.9 1.9 1.2 1.4 3.2 2.8 synoviocytes, as well as by endothelial cells, whereas A2/RA33 expression was scarcely detectable in T and B cells (data not a Two synthetic peptides (for sequences refer to Materials and Methods) containing dominant B cell epitopes were employed in citrullinated (cfilpep) and unmodified shown). Remarkably, A2/RA33, which in cultured cells shows a (filpep) forms. SI Ն 2 are underlined. predominant nuclear localization (51), was detected not only in the FIGURE 6. Immunohistochemical analysis of A2/RA33 expression in synovial tissue. Cryosections from a patient with RA (A–C and F) and a patient with OA (D and E) were stained with a mAb against A2/RA33 (A–E) or an isotype- matched control mAb (F). Cells were counterstained with either hematoxylin (A, D, and F) or an anti-CD68 mAb (B, C, and E) to identify synovial macrophages. Cells expressing A2/RA33 are stained brown; CD68-bearing cells are blue. Pronounced expression of A2/ RA33 is visible in the lining layer and in the sublining areas of RA tissue (A–C), whereas in OA tissue many fewer cells are stained (D and E). Note both nuclear and cytoplasmic localization of A2/ RA33 in RA synovial cells. Magnifica- tion: 100-fold (B, E, and F), 200-fold (A and D), and 400-fold (E).

nucleus but also in the cytoplasm of RA synovial cells (Fig. 6, A bition experiments with mAb against MHC class II molecules. and C). In contrast, in OA synovial tissue only few cells stained Studies are currently in progress to investigate interaction of A2/ positive and the degree of cellular expression appeared to be gen- RA33 peptides with RA-associated HLA-DR molecules. erally weaker (Fig. 6, D and E). Cytokine analysis revealed abundant IFN-␥ secretion in response to A2/RA33 by the majority of patient and control PBMC, Discussion whereas production of IL-4 was considerably lower or undetect- The two Ags investigated in this study, A2/RA33 and cFil, repre- able. The data obtained with A2/RA33-specific TCC bolster these sent well-characterized target structures of the humoral autoim- observations because almost all TCC derived from RA patients mune response of RA patients. Because the factors driving the Aab showed a strong Th1 phenotype that was more pronounced than response to these two proteins are unknown, we were interested to that of control-derived Th1 clones. Taken together, these data in- search for the presence of autoreactive T cells in blood and syno- dicate a strong preponderance of Th1 cells among A2/RA33-reac- vial fluid of RA patients, assuming that autoreactive T cells might tive T cells, which is in agreement with previous observations on have more pathogenetic relevance than Aab, whose contribution to Th1:Th2 ratios in RA patients (9–12, 53). Remarkably, IFN-␥ pro- the pathophysiology of RA is still unclear. The data obtained in the duction of PBMC of RA patients was much higher in response to course of this study seem to support such an assumption, because A2/RA33 than to TT, whereas PBMC from control subjects re- T cell responses to A2/RA33 were observed in peripheral blood of sponded to both Ags in a comparable manner. Therefore, in RA ϳ60% of the RA patients and in all synovial fluids investigated patients A2/RA33-reactive T cells appeared to be more prevalent and thus occurred more frequently than anti-A2/RA33 Aab. Fur- in relation to TT-responsive T cells, suggesting Ag-driven expan- thermore, cellular responses to A2/RA33 were seen in only 20% of sion. This conclusion is supported by the observation that A2/ the controls and were significantly weaker than in RA patients, RA33-induced IL-2 production appeared to be largely restricted to even in patients suffering from PSA, an inflammatory and destruc- PBMC of RA patients. tive arthropathy showing some similarities with RA, particularly The presence of autoreactive T cells to A2/RA33 in some with respect to TNF-␣ production (52). healthy persons was not unexpected and is consistent with obser- Although no significant correlation was found between the A2/ vations repeatedly made by other investigators in both humans and RA33 response and the MHC susceptibility alleles HLA-DR4 and animal models (54–60). Under normal physiological conditions DR1, it should be noted that Ͼ70% of the reactive patients carried autoreactive T cells are under tight control and may even play the shared epitope. Importantly, this autoimmune response ap- beneficial roles (61, 62). The lack of significant IL-2 secretion as peared to be largely HLA-DR restricted as demonstrated by inhi- well as the low proliferative capacity of control PBMC in response to A2/RA33 may indicate an anergic state preventing these cells ity of the T cells that actually drive the Aab production to citrul- from becoming expanded. Furthermore, TCC derived from con- linated targets. trols were phenotypically more heterogeneous, with some of them At the present time it is unclear whether any of the autoantigens ϩ ϩ being CD4 CD8 , indicating the existence of immature subpopu- recognized by RA patients or any of the Ags eliciting chronic lations of A2/RA33-reactive cells in the blood of healthy persons. arthritis in animal models (such as cartilage Ags or the mycobac- Thus, it will be particularly interesting to see whether there exist terial 65-kDa heat shock protein) are of relevance in the patho- differences in epitope recognition between RA patients and healthy genesis of RA (23, 24, 74). This is particularly true for type II subjects as has been observed in organ-specific autoimmune dis- collagen, even though collagen-induced arthritis is the most widely ease (63). used model of RA and even though autoimmunity to collagen II The pronounced overexpression and cytoplasmic localization of occurs early in the course of RA (75, 76). An interesting novel A2/RA33 observed in synovial membranes of RA patients was candidate Ag is the glycolytic enzyme glucose-6 phosphate striking and was not seen in tissue of OA patients, who also did not isomerase, which has been identified as the disease-inducing au- ϫ show a significant autoimmune response to A2/RA33. Remark- toantigen in the KRN NOD model of RA (77, 78). This finding ably, abundant cytoplasmic expression was particularly seen in demonstrates once more the potential pathogenetic importance of synovial macrophages and may form a prerequisite for presenta- autoimmune reactions to ubiquitously expressed targets that are tion of this autoantigen to autoreactive T cells. It is well known not joint specific. Similar to the observations made for A2/RA33 or that the stressful conditions in inflamed tissues lead to activation, citrullinated fibrin, glucose-6 phosphate isomerase was recently overexpression, aberrant localization, or deposition of numerous described to be present in a high concentration in synovial tissue of molecules, some of which may function as autoantigens (64). At RA patients who also had anti-glucose-6 phosphate isomerase Aab present the molecular mechanisms leading to aberrant localization in their serum (79). This is suggestive of pathogenetic involvement of A2/RA33 are unknown, but, interestingly, a similar observation also in human RA; however, this remains to be proven. has recently been made in pathological tissue of lung cancer pa- Taken together, the frequent presence of A2/RA33-reactive tients (65). Thus, under conditions of chronic cellular stress trans- Th1-like cells in the blood of RA patients, their presence in the location of A2/RA33 to the cytoplasm seems to occur in which this synovial compartment in conjunction with the observed aberrant protein may exert a hitherto unknown function. A2/RA33 is a shut- expression of A2/RA33 in the inflamed synovium, and the occurrence of Aab to A2/RA33 early in the course of RA (42–44) sug- tling protein involved in transport of certain mRNAs and possibly gest that this protein may be an important autoantigen in RA. also in regulation of their translation as suggested for myelin basic Moreover, A2/RA33 Aab are among the earliest detectable Aab in protein or glucose transporter 1 (66, 67). Like other shuttling pro- MRL/lpr lupus mice (80), which, in addition to anti-DNA and teins, A2/RA33 is known to undergo posttranslational modifica- other systemic lupus erythematosus-specific Aab, also produce RF tions such as phosphorylation or methylation (68, 69). Therefore, and may develop erosive arthritis (81), and have recently been it is conceivable that the cytoplasmic form differs from the nuclear found to occur also in TNF-␣ transgenic mice (Ref. 82 and S. one and might in fact be more immunogenic, taking into consid- Hayer, B. Jahn-Schmid, D. Plows, K. Skriner, H. Erlandsson-Har- eration that posttranslational modifications may lead to the forma- ris, S. Haralambous, F. Monneaux, S. Trembleau, G. Schett, W. tion of neoepitopes and render a protein autoantigenic (70). Thus, van Venrooij, et al., manuscript in preparation), which represent a it may well be possible that the Aab are primarily directed to post- well-established model of RA (83). Although the significance of translationally modified (i.e., the cytoplasmic) forms of A2/RA33 these observations is not yet clear, they may be considered as and escape detection in our assays where Ag is used that is either additional indications for pathogenetic relevance of this autoim- bacterially expressed or derived from cultured cells in which ex- mune response. Further studies in patients and experimental ani- pression is largely restricted to the nucleus (51). To address this mals will show whether A2/RA33 is indeed part of the inflamma- issue, A2/RA33 must be isolated from human synovial tissue and tory and deleterious cascade of events characteristic of RA. subjected to a detailed biochemical and immunological analysis. Although Aab to the second Ag investigated, Fil, can be more Acknowledgments frequently found in RA patients than anti-A2/RA33 Aab (26, 36, We thank Dr. Ann Union for generously providing us with Fil and Fil 39), cellular reactivity to Fil was seen in only few RA patients. peptides, Elisabeth Ho¨fler for expert determination of anti-A2/RA33 and Moreover, these responses were substantially weaker than those AFA, and Walter Knapp and the other members of the Interdisciplinary against A2/RA33, which could be detected in all patients reactive Cooperation Project for their interest in our work and stimulating to Fil. Thus, these data suggest that Fil may not be the T cell Ag discussions. driving the humoral autoimmune response to citrullinated Ags. Nevertheless, B cells secreting AFA have been detected in syno- References vial tissue and fluid of RA patients (71, 72), and citrullinated pro- 1. Feldmann, M., F. Brennan, and R. N. Maini. 1996. Rheumatoid arthritis. Cell teins apparently different from Fil (which is not expressed in the 85:307. 2. Weyand, C. M., and J. J. Goronzy. 1999. HLA polymorphisms and T cells in joint) have been demonstrated to be present in the synovium of RA rheumatoid arthritis. Int. Rev. Immunol. 18:37. patients (41, 73). Remarkably, fibrin was recently identified as one 3. Schumacher, H. R., Jr., B. B. Bautista, R. E. Krauser, A. K. Mathur, and E. P. Gall. 1994. Histological appearance of the synovium in early rheumatoid of the major citrullinated proteins in the joint and, importantly, arthritis. Semin. Arthritis Rheum. 23:3. citrullinated fibrin was recognized by purified AFA (41). Thus, it 4. Ramanujam, T., M. Luchi, H. R. Schumacher, S. Zwillich, C. P. Chang, is conceivable that T and B cells responding to fibrin-derived (cit- P. E. Callegari, J. M. Von Feldt, Q. Fang, D. B. Weiner, and W. V. Williams. 1995. Detection of T cell receptors in early rheumatoid arthritis synovial tissue. rullinated) epitopes are present in the synovial compartment and Pathobiology 63:100. may be prominent players in the RA autoimmune orchestra. How- 5. Tak, P. P., T. J. Smeets, M. R. Daha, P. M. Kluin, K. A. Meijers, R. Brand, ever, it should be borne in mind that T cells driving the anti- A. E. Meinders, and F. C. Breedveld. 1997. Analysis of the synovial cell infiltrate in early rheumatoid synovial tissue in relation to local disease activity. Arthritis citrulline B cell response may not necessarily be directed to cit- Rheum. 40:217. rullinated epitopes or Ags. The observation that cFil generally 6. Weyand, C. M., and J. J. Goronzy. 1999. T-cell responses in rheumatoid arthritis: systemic abnormalities—local disease. Curr. Opin. Rheumatol. 11:210. induced a weaker response indicates that deimination may affect T 7. Simon, A. K., E. Seipelt, and J. Sieper. 1994. Divergent T-cell cytokine patterns cell recognition, and future studies will have to show the specific- in inflammatory arthritis. Proc. Natl. Acad. Sci. USA 91:8562. 8. Klimiuk, P. A., J. J. Goronzy, J. Bjornsson, R. D. Beckenbaugh, and 37. Girbal-Neuhauser, E., J. J. Durieux, M. Arnaud, P. Dalbon, M. Sebbag, C. M. Weyand. 1997. Tissue cytokine patterns distinguish variants of rheumatoid C. Vincent, M. Simon, T. Senshu, C. Masson-Bessiere, C. Jolivet-Reynaud, et al. synovitis. Am. J. Pathol. 151:1311. 1999. The epitopes targeted by the rheumatoid arthritis-associated antifilaggrin 9. Steiner, G., M. Tohidast-Akrad, G. Witzmann, M. Vesely, A. Studnicka-Benke, autoantibodies are posttranslationally generated on various sites of (pro)filaggrin A. Gal, M. Kunaver, P. Zenz, and J. S. Smolen. 1999. Cytokine production by by deimination of arginine residues. J. Immunol. 162:585. synovial T cells in rheumatoid arthritis. Rheumatology (Oxford) 38:202. 38. Ishida-Yamamoto, A., H. Tanaka, H. Nakane, H. Takahashi, Y. Hashimoto, and 10. Miltenburg, A. M., J. M. van Laar, R. de Kuiper, M. R. Daha, and H. Iizuka. 1999. Programmed cell death in normal epidermis and loricrin kera- F. C. Breedveld. 1992. T cells cloned from human rheumatoid synovial mem- toderma: multiple functions of profilaggrin in keratinization. J. Investig. Derma- brane functionally represent the Th1 subset. Scand. J. Immunol. 35:603. tol. Symp. Proc. 4:145. 11. Quayle, A. J., P. Chomarat, P. Miossec, J. Kjeldsen Kragh, O. Forre, and 39. Schellekens, G. A., B. A. de Jong, F. H. van den Hoogen, L. B. van de Putte, and J. B. Natvig. 1993. Rheumatoid inflammatory T-cell clones express mostly Th1 W. J. van Venrooij. 1998. Citrulline is an essential constituent of antigenic de- but also Th2 and mixed (Th0-like) cytokine patterns. Scand. J. Immunol. 38:75. terminants recognized by rheumatoid arthritis-specific autoantibodies. J. Clin. 12. Dolhain, R. J., A. N. van der Heiden, N. T. ter Haar, F. C. Breedveld, and Invest. 101:273. A. M. Miltenburg. 1996. Shift toward T lymphocytes with a T helper 1 cytokine- 40. Schellekens, G. A., H. Visser, B. A. de Jong, F. H. van den Hoogen, J. M. Hazes, secretion profile in the joints of patients with rheumatoid arthritis. Arthritis F. C. Breedveld, and W. J. van Venrooij. 2000. The diagnostic properties of Rheum. 39:1961. rheumatoid arthritis antibodies recognizing a cyclic citrullinated peptide. Arthritis 13. Romagnani, S. 1997. The Th1/Th2 paradigm. Immunol. Today 18:263. Rheum. 43:155. 14. Miossec, P. 2000. Are T cells in rheumatoid synovium aggressors or bystanders? 41. Masson-Bessiere, C., M. Sebbag, E. Girbal-Neuhauser, L. Nogueira, C. Vincent, Curr. Opin. Rheumatol. 12:181. T. Senshu, and G. Serre. 2001. The major synovial targets of the rheumatoid arthritis-specific antifilaggrin autoantibodies are deiminated forms of the ␣- and 15. Elson, C. J., and R. N. Barker. 2000. Helper T cells in antibody-mediated, organ- ␤ specific autoimmunity. Curr. Opin. Immunol. 12:664. -chains of fibrin. J. Immunol. 166:4177. 16. Cope, A. P., M. Londei, N. R. Chu, S. B. Cohen, M. J. Elliott, F. M. Brennan, 42. Cordonnier, C., O. Meyer, E. Palazzo, M. de Bandt, A. Elias, P. Nicaise, T. Haim, R. N. Maini, and M. Feldmann. 1994. Chronic exposure to tumor necrosis factor M. F. Kahn, and G. Chatellier. 1996. Diagnostic value of anti-RA33 antibody, (TNF) in vitro impairs the activation of T cells through the T cell receptor/CD3 antikeratin antibody, antiperinuclear factor and antinuclear antibody in early complex: reversal in vivo by anti-TNF antibodies in patients with rheumatoid rheumatoid arthritis: comparison with rheumatoid factor. Br. J. Rheumatol. 35: arthritis. J. Clin. Invest. 94:749. 620. 17. Maurice, M. M., H. Nakamura, E. A. van der Voort, A. I. van Vliet, F. J. Staal, 43. Hassfeld, W., G. Steiner, W. Graninger, G. Witzmann, H. Schweitzer, and P. P. Tak, F. C. Breedveld, and C. L. Verweij. 1997. Evidence for the role of an J. S. Smolen. 1993. Autoantibody to the nuclear antigen RA33: a marker for early altered redox state in hyporesponsiveness of synovial T cells in rheumatoid ar- rheumatoid arthritis. Br. J. Rheumatol. 32:199. thritis. J. Immunol. 158:1458. 44. Hueber, W., W. Hassfeld, J. S. Smolen, and G. Steiner. 1999. Sensitivity and 18. Firestein, G. S., and N. J. Zvaifler. 1990. How important are T cells in chronic specificity of anti-Sa autoantibodies for rheumatoid arthritis. Rheumatology (Ox- rheumatoid synovitis? Arthritis Rheum. 33:768. ford) 38:155. 19. Panayi, G. S., J. S. Lanchbury, and G. H. Kingsley. 1992. The importance of the 45. Skriner, K., W. H. Sommergruber, V. Tremmel, I. Fischer, A. Barta, J. S. Smolen, T cell in initiating and maintaining the chronic synovitis of rheumatoid arthritis. and G. Steiner. 1997. Anti-A2/RA33 autoantibodies are directed to the RNA Arthritis Rheum. 35:729. binding region of the A2 protein of the heterogeneous nuclear ribonucleoprotein complex: differential epitope recognition in rheumatoid arthritis, systemic lupus 20. Smolen, J. S., M. Tohidast Akrad, A. Gal, M. Kunaver, G. Eberl, P. Zenz, erythematosus, and mixed connective tissue disease. J. Clin. Invest. 100:127. A. Falus, and G. Steiner. 1996. The role of T-lymphocytes and cytokines in 46. Union, A., L. Meheus, R. L. Humbel, K. Conrad, G. Steiner, H. Moereels, rheumatoid arthritis. Scand. J. Rheumatol. 25:1. H. Pottel, G. Serre, and F. De Keyser. 2002. Identification of citrullinated rheu- 21. Fox, D. A. 1997. The role of T cells in the immunopathogenesis of rheumatoid matoid arthritis-specific epitopes in natural filaggrin relevant for antifilaggrin arthritis. Arthritis Rheum. 40:598. autoantibody detection by LINE immunoassay. Arthritis Rheum. 46:1185. 22. Breedveld, F. C., and C. L. Verweij. 1997. T cells in rheumatoid arthritis. 47. Scheinecker, C., K. P. Machold, O. Majdic, P. Hocker, W. Knapp, and Br. J. Rheumatol. 36:617. J. S. Smolen. 1998. Initiation of the autologous mixed lymphocyte reaction re- 23. Smolen, J. S., and G. Steiner. 1998. Are autoantibodies active players or epiphe- quires the expression of costimulatory molecules B7-1 and B7-2 on human pe- nomena? Curr. Opin. Rheumatol. 10:201. ripheral blood dendritic cells. J. Immunol. 161:3966. 24. Cope, A. P., and G. Sonderstrup. 1998. Evaluating candidate autoantigens in 48. Del Prete, G. F., M. De Carli, C. Mastromauro, R. Biagiotti, D. Macchia, rheumatoid arthritis. Springer Semin. Immunopathol. 20:23. P. Falagiani, M. Ricci, and S. Romagnani. 1991. Purified protein derivative of 25. Steiner, G., K. Hartmuth, K. Skriner, I. Maurer-Fogy, A. Sinski, E. Thalmann, Mycobacterium tuberculosis and excretory-secretory antigen(s) of Toxocara ca- W. Hassfeld, A. Barta, and J. S. Smolen. 1992. Purification and partial sequenc- nis expand in vitro human T cells with stable and opposite (type 1 T helper or ing of the nuclear autoantigen RA33 shows that it is indistinguishable from the type 2 T helper) profile of cytokine production. J. Clin. Invest. 88:346. A2 protein of the heterogeneous nuclear ribonucleoprotein complex. J. Clin. In- 49. Fritsch, R., B. Bohle, U. Vollmann, U. Wiedermann, B. Jahn-Schmid, vest. 90:1061. M. Krebitz, H. Breiteneder, D. Kraft, and C. Ebner. 1998. Bet v 1, the major birch 26. Simon, M., E. Girbal, M. Sebbag, V. Gomes-Daudrix, C. Vincent, G. Salama, and pollen allergen, and Mal d 1, the major apple allergen, cross-react at the level of G. Serre. 1993. The cytokeratin filament-aggregating protein filaggrin is the tar- allergen-specific T helper cells. J. Allergy Clin. Immunol. 102:679. get of the so-called “antikeratin antibodies,” autoantibodies specific for rheuma- 50. Marcu, A., B. Bassit, R. Perez, and S. Pinol-Roma. 2001. Heterogeneous nuclear toid arthritis. J. Clin. Invest. 92:1387. ribonucleoprotein complexes from Xenopus laevis oocytes and somatic cells. Int. 27. Despres, N., G. Boire, F. J. Lopez Longo, and H. A. Menard. 1994. The Sa J. Dev. Biol. 45:743. system: a novel antigen-antibody system specific for rheumatoid arthritis. 51. Satoh, H., H. Kamma, H. Ishikawa, H. Horiguchi, M. Fujiwara, Y. T. Yamashita, J. Rheumatol. 21:1027. M. Ohtsuka, and K. Sekizawa. 2000. Expression of hnRNP A2/B1 proteins in 28. Blaess, S., A. Union, J. Raymackers, F. Schumann, U. Ungethu¨m, S. Mu¨ller- human cancer cell lines. Int. J. Oncol. 16:763. Steinbach, F. De Keyser, J. M. Engel, and G. R. Burmester. 2001. The stress 52. Partsch, G., G. Steiner, B. F. Leeb, A. Dunky, H. Broll, and J. S. Smolen. 1997. protein BiP is overexpressed and is a major B- and T-cell target in rheumatoid Highly increased levels of tumor necrosis factor-␣ and other proinflammatory arthritis. Arthritis Rheum. 44:971. cytokines in psoriatic arthritis synovial fluid. J. Rheumatol. 24:518. 29. Hassfeld, W., G. Steiner, A. Studnicka Benke, K. Skriner, W. Graninger, 53. Berner, B., D. Akca, T. Jung, G. A. Muller, and M. A. Reuss-Borst. 2000. Anal- ϩ ϩ I. Fischer, and J. S. Smolen. 1995. Autoimmune response to the spliceosome: an ysis of Th1 and Th2 cytokines expressing CD4 and CD8 T cells in rheumatoid immunologic link between rheumatoid arthritis, mixed connective tissue disease, arthritis by flow cytometry. J. Rheumatol. 27:1128. and systemic lupus erythematosus. Arthritis Rheum. 38:777. 54. Villoslada, P., K. Abel, N. Heald, R. Goertsches, S. L. Hauser, and C. P. Genain. 30. Steiner, G., K. Skriner, and J. S. Smolen. 1996. Autoantibodies to the A/B pro- 2001. Frequency, heterogeneity and encephalitogenicity of T cells specific for teins of the heterogeneous nuclear ribonucleoprotein complex: novel tools for the myelin oligodendrocyte glycoprotein in naive outbred primates. Eur. J. Immunol. diagnosis of rheumatic diseases. Int. Arch. Allergy Immunol. 111:314. 31:2942. 31. Weighardt, F., G. Biamonti, and S. Riva. 1996. The roles of heterogeneous nu- 55. Kuwana, M., C. A. Feghali, T. A. Medsger, Jr., and T. M. Wright. 2001. Auto- clear ribonucleoproteins (hnRNP) in RNA metabolism. Bioessays 18:747. reactive T cells to topoisomerase I in monozygotic twins discordant for systemic 32. Krecic, A. M., and M. S. Swanson. 1999. hnRNP complexes: composition, struc- sclerosis. Arthritis Rheum. 44:1654. ture, and function. Curr. Opin. Cell Biol. 11:363. 56. Rohowsky-Kochan, C., D. Molinaro, and S. D. Cook. 2000. Cytokine secretion 33. Mayeda, A., S. H. Munroe, J. F. Caceres, and A. R. Krainer. 1994. Function of profile of myelin basic protein-specific T cells in multiple sclerosis. Mult. Scler. conserved domains of hnRNP A1 and other hnRNP A/B proteins. EMBO J. 6:69. 13:5483. 57. Diaz-Villoslada, P., A. Shih, L. Shao, C. P. Genain, and S. L. Hauser. 1999. 34. Carson, J. H., S. Kwon, and E. Barbarese. 1998. RNA trafficking in myelinating Autoreactivity to myelin antigens: myelin/oligodendrocyte glycoprotein is a prev- cells. Curr. Opin. Neurobiol. 8:607. alent autoantigen. J. Neuroimmunol. 99:36. 35. Kamma, H., H. Horiguchi, L. Wan, M. Matsui, M. Fujiwara, M. Fujimoto, 58. Dosch, H., R. K. Cheung, W. Karges, M. Pietropaolo, and D. J. Becker. 1999. T. Yazawa, and G. Dreyfuss. 1999. Molecular characterization of the hnRNP Persistent T cell anergy in human type 1 diabetes. J. Immunol. 163:6933. A2/B1 proteins: tissue-specific expression and novel isoforms. Exp. Cell Res. 59. Hertl, M., and R. Riechers. 1999. Analysis of the T cells that are potentially 246:399. involved in autoantibody production in pemphigus vulgaris. J. Dermatol. 26:748. 36. Vincent, C., M. Simon, M. Sebbag, E. Girbal-Neuhauser, J. J. Durieux, 60. Hoffman, R. W. 2001. T cells in the pathogenesis of systemic lupus erythema- A. Cantagrel, B. Fournie, B. Mazieres, and G. Serre. 1998. Immunoblotting de- tosus. Front. Biosci. 6:D1369. tection of autoantibodies to human epidermis filaggrin: a new diagnostic test for 61. Schwartz, M., and I. R. Cohen. 2000. Autoimmunity can benefit self-mainte- rheumatoid arthritis. J. Rheumatol. 25:838. nance. Immunol. Today 21:265. 62. van Eden, W., R. van der Zee, A. G. Paul, B. J. Prakken, U. Wendling, 73. Baeten, D., I. Peene, A. Union, L. Meheus, M. Sebbag, G. Serre, E. M. Veys, and S. M. Anderton, and M. H. Wauben. 1998. Do heat shock proteins control the F. De Keyser. 2001. Specific presence of intracellular citrullinated proteins in balance of T-cell regulation in inflammatory diseases? Immunol. Today 19:303. rheumatoid arthritis synovium: relevance to antifilaggrin autoantibodies. Arthritis 63. Tejada-Simon, M. V., J. Hong, V. M. Rivera, and J. Z. Zhang. 2001. Reactivity Rheum. 44:2255. pattern and cytokine profile of T cells primed by myelin peptides in multiple 74. Joe, B., and R. L. Wilder. 1999. Animal models of rheumatoid arthritis. Mol. sclerosis and healthy individuals. Eur. J. Immunol. 31:907. Med. Today 5:367. 64. Schett, G., M. Tohidast Akrad, G. Steiner, and J. Smolen. 2001. The stressed 75. Luross, J. A., and N. A. Williams. 2001. The genetic and immunopathological synovium. Arthritis Res 3:80. processes underlying collagen-induced arthritis. Immunology 103:407. 65. Fielding, P., L. Turnbull, W. Prime, M. Walshaw, and J. K. Field. 1999. Heter- 76. Cook, A. D., M. J. Rowley, I. R. Mackay, A. Gough, and P. Emery. 1996. ogeneous nuclear ribonucleoprotein A2/B1 up-regulation in bronchial lavage Antibodies to type II collagen in early rheumatoid arthritis: correlation with dis- specimens: a clinical marker of early lung cancer detection. Clin. Cancer Res. ease progression. Arthritis Rheum. 39:1720. 5:4048. 77. Korganow, A. S., H. Ji, S. Mangialaio, V. Duchatelle, R. Pelanda, T. Martin, 66. Hoek, K. S., G. J. Kidd, J. H. Carson, and R. Smith. 1998. hnRNP A2 selectively C. Degott, H. Kikutani, K. Rajewsky, J. L. Pasquali, C. Benoist, and D. Mathis. binds the cytoplasmic transport sequence of myelin basic protein mRNA. Bio- 1999. From systemic T cell self-reactivity to organ-specific autoimmune disease chemistry 37:7021. via immunoglobulins. Immunity 10:451. 78. Matsumoto, I., A. Staub, C. Benoist, and D. Mathis. 1999. Arthritis provoked by 67. Hamilton, B. J., R. C. Nichols, H. Tsukamoto, R. J. Boado, W. M. Pardridge, and linked T and B cell recognition of a glycolytic enzyme. Science 286:1732. W. F. Rigby. 1999. hnRNP A2 and hnRNP L bind the 3ЈUTR of glucose trans- 79. Schaller, M., D. R. Burton, and H. J. Ditzel. 2001. Autoantibodies to GPI in porter 1 mRNA and exist as a complex in vivo. Biochim. Biophys. Acta 261:646. rheumatoid arthritis: linkage between an animal model and human disease. Nat. 68. Duncan, K., J. G. Umen, and C. Guthrie. 2000. A putative ubiquitin ligase re- Immunol. 2:746. quired for efficient mRNA export differentially affects hnRNP transport. Curr. 80. Dumortier, H., F. Monneaux, B. Jahn-Schmid, J. P. Briand, K. Skriner, Biol. 10:687. P. L. Cohen, J. S. Smolen, G. Steiner, and S. Muller. 2000. B and T cell responses 69. Nichols, R. C., X. W. Wang, J. Tang, B. J. Hamilton, F. A. High, to the spliceosomal heterogeneous nuclear ribonucleoproteins A2 and B1 in nor- H. R. Herschman, and W. F. Rigby. 2000. The RGG domain in hnRNP A2 affects mal and lupus mice. J. Immunol. 165:2297. subcellular localization. Exp. Cell Res. 256:522. 81. Gilkeson, G. S., P. Ruiz, A. J. Pritchard, and D. S. Pisetsky. 1991. Genetic control 70. Doyle, H. A., and M. J. Mamula. 2001. Post-translational protein modifications of inflammatory arthritis and glomerulonephritis in congenic lpr mice and their F1 in antigen recognition and autoimmunity. Trends Immunol. 22:443. hybrids. J. Autoimmun. 4:595. 71. Masson-Bessiere, C., M. Sebbag, J. J. Durieux, L. Nogueira, C. Vincent, 82. Schett, G., S. Hayer, M. Tohidast-Akrad, B. J. Schmid, S. Lang, B. Turk, E. Girbal-Neuhauser, R. Durroux, A. Cantagrel, and G. Serre. 2000. In the rheu- F. Kainberger, S. Haralambous, G. Kollias, A. C. Newby, et al. 2001. Adenovi- matoid pannus, anti-filaggrin autoantibodies are produced by local plasma cells rus-based overexpression of tissue inhibitor of metalloproteinases 1 reduces tis- and constitute a higher proportion of IgG than in synovial fluid and serum. Clin. sue damage in the joints of tumor necrosis factor ␣ transgenic mice. Arthritis Exp. Immunol. 119:544. Rheum. 44:2888. 72. Reparon-Schuijt, C. C., W. J. van Esch, C. van Kooten, G. A. Schellekens, 83. Kollias, G., E. Douni, G. Kassiotis, and D. Kontoyiannis. 1999. On the role of B. A. de Jong, W. J. van Venrooij, F. C. Breedveld, and C. L. Verweij. 2001. tumor necrosis factor and receptors in models of multiorgan failure, rheumatoid Secretion of anti-citrulline-containing peptide antibody by B lymphocytes in arthritis, multiple sclerosis and inflammatory bowel disease. Immunol. Rev. rheumatoid arthritis. Arthritis Rheum. 44:41. 169:175.

Chapter 7

Potential predictors of Glucocorticoid-response in Rheumatoid arthritis. A role for endoplasmic reticulum aminopeptidase 2 (ERAP2) and leucocyte-specific transcript 1 (LST1)?

RD Fritsch-Stork SC Cardoso M Groot-Koerkamp J Broen A Conception B Giovannone A Ploner FPJG Lafeber JWJ Bijlsma

Department of Rheumatology & Clinical Immunology, University Medical Center Utrecht, Utrecht, The Nederlands

Submitted

Potential predictors of Glucocorticoid-response in Rheumatoid arthritis A role for endoplasmic reticulum aminopeptidase 2 (ERAP2) and leucocyte- specific transcript 1 (LST1)?

Ruth D Fritsch-Stork, MD; Sandra CS Cardoso, MSc; Marian JA Groot-Koerkamp, BSc; Jasper Broen, MD, PhD; Arno N Concepcion, BSc; Barbara Giovannone, PhD; Alexander Ploner, PhD; Floris PJ Lafeber, PhD; Johannes WJ Bijlsma, MD, PhD

This work was supported by the Dutch Arthritis Foundation.

Abstract

Objective: Glucocorticoids (GC) are a cornerstone of Rheumatoid arthritis (RA)-therapy, although a third of RA-patients do not respond adequately. As monocytes and T-cells play an important role in RA-pathogenesis, the gene expression of these cells before pulse therapy with methylprednisolone was evaluated in order to find potential predictors for clinical response to GC. Methods: Patients were treated with 3x 1000mg Methylprednisolone in 5 days. Before start of treatment CD14- and CD4-positive cells were separated by MACS sorting. At day 5 response was determined by the European League against Rheumatism response criteria using the DAS28. Labeled cRNA from 5 GC-Responders and 5 Non-Responders was hybridized to Agilent 4x44K microarray chips, and differentially expressed genes were identified via mixed-model analysis of variance based on permutation-based false disvovery rates to account for multiple testing. Selected genes were validated by quantitative real-time PCR (qPCR). Results: Eight genes were differentially regulated in CD14+ monocytes and 4 genes in CD4+T-cells of GC-Responders compared to Non-Responders. Significant up-regulation of ERAP2 in monocytes and CD4+cells and LST1 and FAM26F in CD4+T-cells of GC- Responders was validated by qPCR and correlated with DAS28 at day 5. LST1 and FAM26F expression were also correlated with smoking. Conclusion: As ERAP2 and LST1 are implicated in autoimmune diseases, their increased expression in GC-Responders before GC-therapy may constitute not only a potential predictor for the clinical response to GC in RA, but also warrants further investigation to elucidate their role in the inflammatory process of RA.

Introduction the genetic association with certain HLA- proteins indicate a prominent role for T- 1 Rheumatoid arthritis (RA) is a systemic auto- cells. immune disease, affecting approximately Besides T-cells, the pathogenetic importance 0.5-1% of the general population. The of macrophages is established in RA and disease is marked by joint inflammation reflected in their abundance in RA-synovium leading to cartilage and bone destruction. and their function as antigen presenting cells Several observations, such as the occurrence and main producers of inflammatory 1 of class switched autoantibodies, the cytokines. characterization of auto-reactive T-cells and

84 As we are moving from conventional disease Medical Center and patients signed informed modifying drugs (DMARDs) to biological consent for the study protocol. therapies and treatment with small molecules, Glucocorticoids (GC) have Treatment and clinical measurements: remained a cornerstone of RA-therapy. Due Patients were given three doses of 1000mg to the inhibition of radiographic damage methyl-prednisolone on alternate days. (reviewed in Ref. 2), GC-therapy has gained Disease activity was assessed by DAS28 recognition as a DMARD. The rapid anti- before start and at the end of treatment (+12- inflammatory effect of GC is used in pulse 24 hours after the third dose). Treatment therapy (≥ 250mg prednisone equivalent response was defined according to the i.v.), given to patients with high disease EULAR response criteria. Good responders burden as bridging until the effect of a were classified by a DAS28<3.2 and a ΔDAS DMARD becomes apparent. Pulse treatment of ≥1.2. Considering the intensity of results in rapid amelioration (48hours) treatment, patients not fulfilling the criteria clinically and serologically, lasting for a for good response were considered Non- mean of 10 weeks, although induced Responders. remission for more than 42 weeks has been reported.3 RNA of CD14+Monocytes and CD4+T- cells: About one third of RA patients do not Peripheral blood mononuclear cells (PBMC) respond to GC adequately4, a phenomenon from patients were obtained before therapy which is also seen in other inflammatory from peripheral Li-heparinized blood by diseases as systemic lupus erythematosus or Ficoll-PaqueTM Plus density gradient asthma. Although several mechanisms centrifugation (GE-Healthcare, Bio-Sciences underlying GC resistance have been AB, Uppsala, Sweden). CD14+monocytes described, no predictors of GC sensitivity in and CD4+T-cells were isolated by positive RA-patients have been established.5 During (CD14+cells) or negative (CD4+cells) the last decade gene expression profiling has selection by AutoMACS Pro Separator been employed in RA to analyze and predict (Miltenyi Biotec, Bergisch Gladbach, the response to (mainly biological) therapy.6 Germany) and the respective isolation kits The aim of this study was to find biomarkers for CD14 or CD4 isolation (Miltenyi Biotec). for the response to GC in RA using gene Purity of the viable population assessed by expression profiling of T-cells and flow cytometry (BD FACS Calibur, Franklin monocytes from RA-patients before the first Lakes, New Jersey, U.S.) was generally administration of GC-pulse therapy. >90%. RNA was isolated using the RNeasy Mini Kit from QIAGEN (Venlo, The Netherlands). Total RNA concentration was Materials and methods measured by Nanodrop spectrophotometer (Nanodrop Technologies, Wilmington, DE, Patients: USA). RNA purity and integrity were Patients, fulfilling the revised American verified by lab-on-chip technology (Agilent College of Rheumatology (ACR) criteria for 2100 Bioanalyzer, Santa Clara, CA, USA). rheumatoid arthritis7 and scheduled by their All procedures were performed according to treating rheumatologist for GC-pulse therapy, manufacturers’ instruction. with stable DMARD medication and no GC- treatment during the previous 4 weeks were Microarray RNA profiling: recruited. From the first 5 GC-Responders and 5 Non- The study was approved by the medical responders meeting the quality requirements ethics committee of the Utrecht University RNA of both cell subsets was evaluated. Human whole genome gene expression

85 microarrays v1 (Agilent Technologies, Hochberg) were considered significantly Belgium) with 41000 homo sapiens 60-mer changed. probes in a 4x44K layout were used. For other statistical analyses SPSS version Universal Human Reference RNA 20.0 (SPSS, Chicago, IL, USA) and Graph (Stratagene) was used as a common Pad Prism (version 5) were used. P values reference. were calculated using unpaired Students T- RNA amplification and labeling were tests and χ-square tests, as appropriate; performed on an automated system (Caliper reported correlations are Spearman’s. P Life Sciences NV/SA, Belgium) with 2 g values < 0.05 are considered significant. total RNA. Hybridizations were done on a HS4800PRO system with QuadChambers Results (Tecan Benelux B.V.B.A.) using 1000 ng labeled cRNA per channel.8 All patient Patients: samples were labeled with cy5, reference Patient characteristics and clinical cRNA was labeled with cy3. Hybridized manifestations of the 5 GC-Responders and 5 slides were scanned on an Agilent scanner Non-Responders are given in Table 1. They (G2565BA) at 100% laser power, 30% PMT. did not differ significantly in age, gender, concomitant methotrexate (MTX)- Quantitative real-time PCR: medication, disease duration, sero-positivity Quantitative real-time PCR (qPCR) was for anti-citrullinated-protein-antibodies performed on total RNA samples using an (ACPA) or Rheumatoid factor (RF) or BioRad MyiQ real time PCR Detection disease activity at the start of therapy. There System (BioRad, Hercules, CA, USA). RNA were significantly more active smokers in the was reverse transcribed into cDNA using a Non-Responders. In response to treatment, BioRad iScript cDNA synthesis kit (BioRad). the DAS28 differed significantly between For the reaction mixture, iQ SYBR Green GC-Responders and Non-Responders at day Supermix kit (BioRad) was used, with 5μM 5 (p<0.0009). of each primer and 20ng of the cDNA sample. For the target genes ERAP2, JAK3, Microarray RNA profiling: TRIM33, TNRC6B, FAM26F and LST1 and Using a FDR multiple testing correction, 13 the housekeeping gene glyceraldehyde-3- probes were found differentially regulated in phosphate dehydrogenase (GAPDH) the PCR CD14+cells, corresponding to 8 known primers are shown in supplementary table 1. genes, of which three were up-regulated in Expression levels of target genes were GC- Responders compared to Non- calculated as deltaCT relative to GAPDH Responders. Five known and 4 unknown (ΔCT= CTtarget gene –CTGAPDH) with a higher genes were down-regulated. Four known deltaCT corresponding to a lower quantity of genes were differentially expressed in mRNA. CD4+cells; all were up-regulated in GC- Responders. Genes are listed with their fold Statistical analysis: change and adjusted p-values in Microarray analysis: supplementary table 2 (CD14+cells) and After automated data extraction using supplementary table 3 (CD4+cells). A heat Imagene 8.0 (BioDiscovery), Loess map displaying the expression level of the normalization was performed on mean spot- respective genes in individual patients is intensities. Data were analysed using shown in Figure 1A (CD14+cells) and Figure MAANOVA.9 P-values were determined by 2A (CD4+cells). a permutation F2-test, in which residuals were shuffled 10000 times across probes. Genes with p<0.05 after FDR-based correction (False discovery rate; Benjamini-

86 GC-Responders Non-Responders significance

Female (%) 3/5 (60) 3/5(60) ns

Age (mean±SEM) 55.6±8.1 49±9.6 ns

Concomitant MTX (%) 3/5 (60) 2/5 (40) ns

Disease duration mean±SEM 12.8±6 8.6±3.9 ns

Erosive disease (%) 2/5 (40) 4/5 (80) ns

RF positive (%) 4/5 (80) 5/5 (100) ns

ACPA positive (%) 3/5 (60) 3/5 (60) ns

Current Smoking (%) 0/5 (0) 3/5 (60) p<0.05

DAS28 day0 mean±SEM 6.6±0.5 6.3±0.7 ns

DAS28 day5 mean±SEM 2.6±0.2 4.0±0.2 p<0.001

Table 1: Patient characteristics of Glucocorticoid (GC)- Responders and GC-Non-Responders Patients characteristics for gender, concomitant methotrexate medication, erosive disease, antibody positivity and smoking are given as percentage. Age, disease duration and disease activity at baseline and after therapy are given as mean with standard error. For smoking, antibody status, erosive disease and medication χ-square test was used, Student’s T-test was employed for the remaining items. MTX: Methotrexate; RF: rheumatoid factor; ACPA: Anti-citrullinated-protein-antibodies; DAS28: disease activity score.

Figure 1

Figure 1: Differentially expressed genes in CD14+cells from Glucocorticoid (GC)-Responders versus GC- Non-Responders: 1A: Heat map of gene expression in individual patients. Transcript levels are expressed as median centered ratios across all samples. Scale bar (log2 ratio): increased (red), decreased (green) or identical (black) ratio.

87 1B: To verify the expression of selected genes in CD14+cells of GC-Responders and Non-Responders quantitative PCR was performed. ΔCT (CT gene of interest – CT GAPDH) levels are given as quantification of gene expression with higher ΔCT corresponding to a lower gene expression. P<0.05 (using unpaired Students T-test) was considered statistically significant. *: p<0.05; **: p<0.01. 1C: To corroborate the predictive potential of ERAP2 for response to GC, correlation of ERAP2 expression with the DAS28 measured after therapy is depicted. A higher ΔCT signifying a lower gene expression, is correlated with a higher DAS28 at the end of therapy. 1D: The correlation of TNRCB6 with smoking, measured by number of cigarette pack years is depicted. A correlation of lower expression and higher number of pack years is seen.

Quantitative real-time PCR To examine the predictive potential of the The genes differently expressed between GC- genes, we analyzed their correlation with Responders and Non-Responders were DAS28 at day 5. The correlation proved tested by quantitative real-time PCR (Figure significant for ERAP2 in CD14+cells 1B and Figure 2B). The difference in gene (p<0.05) and LST1 and FAM26F in expression in CD14-cells was validated for CD4+cells (both: p<0.02) and is shown in ERAP2 (p<0.007; with a 6.6 fold change Figure 1C and 2C. (FC) between groups) and TRIM33 (p< 0.04; As a link between RA and smoking has been FC: 1.4); TNRC6B was expressed 1.3 fold in established, and smoking is seen as a GC-Responders compared to Non- possible cause of GC-resistance in asthma, Responders (p=0.057) and JAK3 1.2 fold we analyzed the correlation of cigarette pack (not significant). In CD4+T-cells three genes years with the genes validated by qPCR. We were validated. Difference between GC- found a significant correlation between the Responders and Non-Responders was 4.4- expression of TNRC6B (CD14+cells), LST1 fold (p<0.006) for FAM26F; 2.5-fold and FAM26F (CD4+cells) and pack years. (p<0.009) for LST1 and 4.9-fold (p<0.02) for (Figure 1D and 2D) ERAP2 respectively.

88 Figure 2

89 Figure 2

Figure 2: Differentially expressed genes in CD4 cells from Glucocorticoid (GC)-Responders versus GC- Non-Responders: 2A: Heat map of gene expression in individual patients. Transcript levels are expressed as median centered ratios across all samples. Scale bar (log2 ratio): increased (red), decreased (green) or identical (black) ratio. 2B: Quantitative PCR results of CD4+T-cells. ΔCT (CT gene of interest – CT GAPDH) levels are given as quantification of gene expression with higher ΔCT corresponding to lower gene expression. P<0.05 (using unpaired Students T-test) was considered statistically significant. *: p<0.05; **: p<0.01. 2C: To corroborate the predictive potential of FAM26F and LST1 for response to GC, correlation of gene expression with the DAS28 measured after therapy is depicted. A higher ΔCT, signifying a lower gene expression, at baseline is correlated with a higher DAS28 at the end of therapy for both genes. 2D: The correlation of FAM26F and LST1 with smoking, measured by number of cigarette pack years is depicted. A correlation of lower expression and higher number of pack years is seen.

DISCUSSION 2B. ERAP2 is suggested to be a gene duplication of ERAP1 (endoplasmic In this pilot-study gene expression profiling reticulum aminopeptidase 1). In several was used to identify potential predictors for genome-wide association studies both genes the clinical response to GC. The focus was have been discovered as susceptibility loci set on two key players in RA-pathogenesis, for autoimmune diseases (reviewed in Ref. monocytes and CD4+T-cells, which are 10). known targets of GC-action, in order to avoid Their function is the enzymatic removal of possible noise by cells of lesser importance amino acids from N-termini of peptides or to the disease process or GC-mediated effects proteins, by which they interfere in several in RA. biological processes. Among these is the Eight genes were identified that were trimming of peptides to optimal length for differentially expressed in monocytes and 4 their binding into the MHC class I groove, 11 genes in CD4+Tcells from GC-Responders which affects antigen-presentation. ERAP2 versus Non-Responders. Of those, three preferably targets basic residues as Arginine showed correlation with DAS28 at day 5 and or Lysine. The nature of the connection were up-regulated more than 2-fold in GC- between ERAP2 expression, possibly Responders. resulting in altered presentation of arginine The most striking gene was ERAP2 residues (which becomes citrulline after (endoplasmic reticulum aminopeptidase 2), deimination) in the MHC-I by RA- which was up-regulated in both cell types of monocytes or T-cells and its influence on GC-Responders, as seen in Figures 1B and

90 RA-pathogenesis and GC-response remains hitherto uncharacterized mechanistic to be analyzed. connection. Additionally, a role for ERAP1 in the An interesting aspect of the role of LST1 and proteolytic cleavage of cytokine receptors FAM26F in GC-response is their correlation TNFRI, IL6Rand IL1RII has been with cigarette pack years. The association of demonstrated.12,10 These soluble cytokine smoking and risk of RA is well-known, and receptors act as decoy for their respective in the last decade a role of smoking in GC- inflammatory cytokines. Considering that resistance of asthma-patients was postulated. TNF and IFN induce the expression of In our patients, there was a significantly both peptidases this suggests ERAP1 to be higher proportion of smokers in the group of the center of an inflammatory negative Non-Responders, suggesting an influence of feedback loop. As ERAP2 has originated as smoking also in GC-resistance in RA- duplicate from ERAP1 and displays close patients. In this light the function of LST1 functional similarity and homology, cytokine and FAM26F and their correlation with pack receptor shedding by ERAP2 seems possible. years warrants further investigation. However, only one publication examined a One anti-inflammatory effect of GC is the correlation of ERAP1 and ERAP2 down-regulation of IL-2 signaling, which polymorphisms prevalent in Ankylosing includes a decrease in JAK3 expression in T- 15 Spondylitis patients with these serum cells. Consequently, a differential baseline cytokine receptors, and found no such expression of JAK3 between GC-Responders association. In the present study the serum and Non-Responders seems possible, and level of TNFRI, IL6Rand IL1RII were was seen in our microarray experiments. analyzed, but no correlation with ERAP2 Unfortunately, this difference did not seem to gene expression at baseline (data not shown) be statistically significant in the validation was found. However, taking into account the process by qPCR. multiple factors influencing soluble receptors Altogether we identified several genes in in serum, the lack of this correlation does not CD14+ and CD4+cells, which are preclude an influence of ERAP2 in receptor significantly up-regulated in GC-Responders shedding on a cellular level. compared to Non-Responders and might In CD4+T-cells, a 4-fold difference in gene serve as predictors of GC-response, although expression between GC-Responders and exact evaluation of the predictive power Non-Responders was seen in FAM26F, the would have to rely on bigger studies. The function of which has not been reported most striking of these genes are ERAP2 and including the search of databases for gene LST1. Their previously reported link to expression of unknown genes (Atlas Data or autoimmune diseases implies a possible role G2SBC database). in the pathogenesis of RA and GC-response LST1, the leucocyte-specific transcript 1, and warrants further investigation. constitutes a susceptibility locus in RA.13 It is expressed in mononuclear cells and regulates T-cell proliferation. Increased expression of REFERENCES LST1 has been reported in blood and synovial tissue from RA-patients in 1) McInnes IB, Schett G. The pathogenesis association with a higher degree of synovial of rheumatoid arthritis. N Engl J Med. inflammation. Lipopolysaccharide and 2011;365:2205-19. interferon- are known to induce LST1 2) Hoes, JN, Jacobs, JW, Buttgereit, F & expression in monocytes.14 In our study we Bijlsma, JW. Current view of found a 2.5 fold increased expression of glucocorticoid co-therapy with DMARDs LST1 in CD4-T-cells from GC-Responders in rheumatoid arthritis. Nat. Rev. compared to Non-Responders, implicating a Rheumatol. 2010; 6: 693–702.

91 3) Forster PJ, Grindulis KA, Neumann V, 10) Fierabracci A, Milillo A, Locatelli F, Hubball S, McConkey B. High-dose Fruci D. The putative role of endoplasmic intravenous methylprednisolone in reticulum aminopeptidases in rheumatoid arthritis. Ann Rheum Dis. autoimmunity: insights from genomic- 1982;41:444-6. wide association studies. Autoimmun 4) van Schaardenburg D, Valkema R, Rev. 2012;12:281-8. Dijkmans BA, Papapoulos S, 11) Saveanu L, Carroll O, Lindo V, Del Val Zwinderman AH, Han KH et al. M, Lopez D, Lepelletier Y, et al. Prednisone treatment of elderly-onset Concerted peptide trimming by human rheumatoid arthritis. Disease activity and ERAP1 and ERAP2 aminopeptidase bone mass in comparison with complexes in the endoplasmic reticulum. chloroquine treatment. Arthritis Rheum. Nat Immunol. 2005;6:689-97 1995;38:334-42. 12) Cui X, Rouhani FN, Hawari F, Levine SJ. 5) Barnes PJ, Adcock IM. Glucocorticoid Shedding of the type II IL-1 decoy resistance in inflammatory diseases. receptor requires a multifunctional Lancet. 2009 May 30;373:1905-17. aminopeptidase, aminopeptidase 6) Smith SL, Plant D, Eyre S, Barton A. The regulator of TNF receptor type 1 potential use of expression profiling: shedding. J Immunol. 2003;171:6814-9. implications for predicting treatment 13) Kilding R, Iles MM, Timms JM, response in rheumatoid arthritis. Ann Worthington J, Wilson AG. Additional Rheum Dis. 2013;72:1118-24. genetic susceptibility for rheumatoid 7) Arnett FC, Edworthy SM, Bloch DA, arthritis telomeric of the DRB1 locus. McShane DJ, Fries JF, Cooper NS, et al. Arthritis Rheum. 2004;50:763-9. The American Rheumatism Association 14) Mulcahy H, O'Rourke KP, Adams C, 1987 revised criteria for the classification Molloy MG, O'Gara F. LST1 and NCR3 of rheumatoid arthritis. Arthritis Rheum. expression in autoimmune inflammation 1988;31:315-24. and in response to IFN-gamma, LPS and 8) Roepman P, Wessels LF, Kettelarij N, microbial infection. Immunogenetics. Kemmeren P, Miles AJ, Lijnzaad P et al. 2006;57:893-903. An expression profile for diagnosis of 15) Bianchi M, Meng C, Ivashkiv LB. lymph node metastases from primary Inhibition of IL-2-induced Jak-STAT head and neck squamous cell carcinomas. signaling by glucocorticoids. Proc Natl Nat Genet. 2005;37:182-6 Acad Sci U S A. 2000;97:9573-8. 9) Wu, H, Kerr, MK, Cui, X, and Churchill, GA. 2003. MAANOVA: A software package for the analysis of spotted cDNA Acknowledgements: microarray experiments, in Parmigiani, We are thankful for the use of the microarray G, Garrett, ES, Irizarry, RA, and Zeger, facilities within the department Molecular SL, eds., The Analysis of Gene Cancer Research. Expression Data: Methods and Software,

Springer, New York.

92 Supplementary Tables

Table 1 Sequence (5' - 3')

Forward Reverse

ERAP2 F’ TGGATGGGACCAACTCATTACA ERAP2 R’ TGCACCAACTAGCTGAAACAC

FAM26F F’ CACCCGATGCCTATCTCCAG FAM26F R’ TTTGCTGCCACTCTTTCATGC

JAK3 F’ TCTCAAGGAGCAGGGTGAGT JAK3 R’ GTAGGCAGGCCTTGTAGCTG

LST1 F’ AGGAACTTGAGGCAAGTCACC LST1 R’ CAGCCTCTGCAGAGATGCATAGT

TNRC6B F’ GCAGTTGTTACAGAACCAGAGAA TNRC6B R’ GCACTCACCATTCGAGCCA

TRIM33 F’ ATGTGGAGAGTGGCTATGTAAGA TRIM33 R’ GGGCGTTGACCAGATGCTC

GAPDH F’ AGAAGGCTGGGGCTCATTT GAPDH R’ GAGGCATTGCTGATGATCTTG

Primer sequences Sequences of forward and reverse primers for genes of interest and the house keeping gene glyceraldehyde-3- phosphate dehydrogenase (GAPDH) are given.

93 Table 2: Differentially expressed genes in CD14+cells between GC-Responders and Non- Responders geneSymbol bioSeq Description Ratio / fold FDR FWER

change

TNRC6B trinucleotide repeat containing 6B 1.82 0.001 0.002

ERAP2 endoplasmic reticulum 2.48 0.001 0.003

aminopeptidase 2

DNAJC25- DNAJC25-GNG10 readthrough 0.65 0.003 0.01

GNG10

AC091046.1 putative uncharacterized protein 0.69 0.004 0.021

ENSP00000332379, RP11-

123C21.1

JAK3 janus kinase 3 0.58 0.004 0.02

MDM1 mdm1 nuclear protein homolog 0.69 0.011 ns

(mouse)

ERAP2 endoplasmic reticulum 3.61 0.011 ns

aminopeptidase 2

TRIM33 tripartite motif-containing 33 1.48 0.027 ns

AC136475.1 RP326C3.11, no protein product 0.67 0.027 ns

AC006946.15 putative uncharacterized protein 0.69 0.027 ns

ENSP00000394032

TMEM106A transmembrane protein 106A 0.71 0.034 ns

TSPAN18 tetraspanin 18 0.70 0.034 ns

GK glycerol kinase 0.62 0.040 ns

List of genes with ratios / fold change (GC-Responders versus Non-Responders) and p values, after multiple testing correction using FDR (false discovery rate) or FWER (family wise error rate). A p value< 0.05 is considered significant. ns: non significant

94 Table 3: Differentially expressed genes in CD4+cells between GC-Responders and Non- Responders geneSymbol bioSeqDescription Ratio / fold FDR FWER

change

FAM26F family with sequence 3.07 0.006 0.004

similarity 26, member F

LST1 leukocyte-specific transcript 1.87 0.008 0.015

1 protein (Protein B144)

ERAP2 endoplasmic reticulum 3.47 0.009 0.027

aminopeptidase 2

TNRC6B trinucleotide repeat 1.69 0.023 ns

containing 6B

List of genes with fold change (GC-Responders versus Non-Responders) and p values, after multiple testing correction using FDR (false discovery rate) or FWER (family wise error rate). A p value< 0.05 is considered significant. ns: non-significant

Chapter 8

Glucocorticoid-responsiveness correlates with an Interferon signature in Rheumatoid Arthritis Patients

RD Fritsch-Stork J Broen SC Cardoso M Groot-Koerkamp JAG van Roon K van der Wurff-Jacobs TRDJ Radstake FJPG Lafeber JWJ Bijlsma

Department of Rheumatology & Clinical Immunology, University Medical Center Utrecht, Utrecht, The Nederlands

Submitted

Glucocorticoid-responsiveness correlates with an Interferon signature in Rheumatoid Arthritis Patients

Ruth D Fritsch-Stork, MD; Jasper Broen, MD, PhD; Sandra CS Cardoso, MSc; Marian JA Groot-Koerkamp, BSc, Joel van Roon, PhD; Kim van der Wurff-Jacobs, BSc; Timothy Radstake, MD, PhD; Floris PJ Lafeber, PhD; Johannes WJ Bijlsma, MD, PhD

ABSTRACT

Background: Rheumatoid arthritis (RA) is driven by monocytes and T-cells, both influenced by glucocorticoids (GC). Although a cornerstone of RA-therapy, one third of RA-patients do not respond adequately to GC and the mechanisms of resistance remain unclear. Objective: To find differences in GC-working mechanism in GC-responding versus GC- resistant RA-patients by gene expression profiling. Methods: Patients received 3x 1000mg i.v. methylprednisolone. Before start (T0) and 24 hours after the first administration (T24), CD14+ monocytes and CD4+ T-cells were MACS- isolated. At day 5, response was determined by DAS28. Labeled cRNA from 5 GC- Responders and 5 non-Responders was hybridized to Agilent 4x44K microarray chips. Differentially expressed genes between T0 and T24 were identified via fixed-model analysis of variance based on permutation-based false discovery rates and a >1.5 ratio cut-off. Gene Ontology was used for pathway analysis; Transcription-factors were identified via TRANSFAC. Relevant genes were validated by qPCR. Results: After 24 hours, 48 genes were exclusively changed in GC-Responders’ monocytes (CD4+T-cells: 19), 253 exclusively in non-Responders (CD4+T-cells: 104) and 104 genes in both (CD4+T-cells: 18). In both cell-types, a more pronounced down-regulation of interferon related genes was seen in non-Responders. In CD4+ T-cells this finding was validated by qPCR. Several relevant transcription-factors were identified, including LEF1 in both cell types. Conclusion: GC treatment resulted in stronger suppression of the interferon signature in non- Responders. This might be disease-specific to RA, but warrants further investigation of GC- mechanism in several auto-immune diseases known for an interferon signature.

4-6 Introduction of several auto-reactive T-cells. Additionally, macrophages are central Rheumatoid arthritis (RA) is a systemic auto- mediators of synovitis, and are abundant in 7 immune disease marked by joint RA-synovial tissue. They secrete inflammation, cartilage and bone destruction. inflammatory cytokines and interact with T- Susceptibility for RA has been associated cells, thereby propagating the inflammatory with a gene locus in the HLA-DRB1 region process. Blocking this cellular cooperation and other loci (e.g. PTPN22 or CTLA-4) by a CTLA-4 fusion protein (Abatacept) has 8 involved in T-cell stimulation and proved therapeutically successful. differentiation, suggesting a central role for T-cells in RA-pathogenesis.2,3 This role is Glucocorticoids (GC) remain a cornerstone further corroborated by the presence of class- of RA-therapy and are applied extensively 9 switched antibodies and the characterization since their first use in 1948, in spite of

98 possible side effects. Due to the inhibition of Gene expression profiling has been used to radiographic damage seen after long-term predict GC-response in GC-sensitive and treatment,10 GC have gained recognition as a GC-resistant asthma-patients using peripheral disease modifying anti-rheumatic drug blood mononuclear cells (PBMC), yielding a (DMARD). GC-pulse therapy (≥ 250mg list of 15 genes that could predict the prednisone equivalent) is usually applied in response with 84% accuracy.20 Although not patients with high disease activity because of yet used to investigate GC-resistance in RA, its fast anti-inflammatory action to bridge the analysis and prediction of response to lag time until (conventional) DMARDs biological therapy has been carried out with deploy their effect. Clinical amelioration is cDNA-microarrays in RA, mostly on PBMC, seen within 48hours, accompanied by but no consistent pattern was found.21 serological changes.11 In some patients the induced remission is sustained for more than In this study we investigated the difference 42 weeks; the mean duration of beneficial in working mechanism of GC in GC- effects is approximately 10 weeks.11-12 responsive and GC-resistant RA-patients, Furthermore, it has been established recently concentrating on two cell-types crucial to that the response to GC in the first days can RA-pathogenesis. To this end gene predict disease activity at 3 months.13 expression profiling of CD14+monocytes and CD4+T-cells before and 24 hours after The (genomic) effects of GC are elicited start of therapy were compared and revealed through ligation of GC with the differential gene regulation of interferon Glucocorticoid Receptor (GR) alpha isoform related genes between GC-Responders and (GR). After translocation to the nucleus, the non-Responders. GR-GC complex binds to Glucocorticoid response elements (GRE) inducing gene transcription of anti-inflammatory genes (e.g. Materials and methods lipocortin or IB). Another important anti- inflammatory mechanism is the inhibition of Patients: transcription of inflammatory genes via Patients fulfilling the revised American protein-protein interaction between the College of Rheumatology (ACR) criteria for GC/GR complex and transcription factors, rheumatoid arthritis were scheduled by their such as AP-1, NFB or IRF-3. treating rheumatologist for GC-pulse therapy.22 Only patients with stable DMARD About one third of RA patients do not medication and no GC-treatment during the respond to GC adequately,14 and GC- previous 4 weeks were recruited. The study resistance has also been observed in other was approved by the medical ethics inflammatory diseases as systemic lupus committee of the University Medical Center erythematosus and asthma.15-16 Several Utrecht, with patients giving written mechanisms underlying glucocorticoid informed consent. resistance in general have been described, affecting every step in the above described Treatment and clinical measurements: working mechanism.17 In GC-resistant RA- Patients were given three doses of methyl- patients, increased expression of the GC- prednisolone i.v., 1000mg each, on alternate isoform, blocking the binding of GC-GR days. As measurement of disease activity, to the GRE, has been reported.18 An DAS2823 was determined before start and at increased cellular level of transcription the end of pulse treatment (day 5). Clinical factors such as NFB, AP-1 or STAT5 could response was defined according to the also contribute to GC-resistance in RA.19 European League against Rheumatism However, no predictors of GC-resistance in (EULAR) response criteria using the absolute RA-patients have been established so far. DAS28 value and the DAS28. Good

99 responders were classified by a DAS28<3.2 B.V.B.A.) using 1000 ng of labeled cRNA (equivalent to low disease activity) and a per channel.25 All patient samples were DAS of ≥1.2. Patients not fulfilling these labeled with cy5, reference cRNA was criteria were considered non-Responders. labeled with cy3. Hybridized slides were scanned on an Agilent scanner (G2565BA) at Isolation of RNA from monocytes and 100% laser power, 30% PMT. CD4+T-cells: PBMC from patients were obtained at Quantitative real-time PCR: baseline (T0) and 24 hours after start of pulse Quantitative real-time PCR was performed therapy (T24) by density centrifugation of on total RNA using BioRad MyiQ real time Li-heparinized blood over Ficoll-PaqueTM PCR Detection System (BioRad, Hercules, Plus (GE-Healthcare, Bio-Sciences AB, CA, USA). RNA was reverse transcribed into Uppsala, Sweden). CD14+monocytes and cDNA using a BioRad iScript cDNA CD4+T-cells were isolated by positive synthesis kit (BioRad). For the reaction (CD14) or negative (CD4) selection using an mixture, iQ SYBR Green Supermix kit AutoMACS Pro Separator with the (BioRad) was used, with 5 μM of each respective isolation kits (αCD14 microbeads: primer and 20ng of the cDNA sample; PCR Ref 130-050-201; CD4+T-cell isolation kit II: primers are shown in supplementary table 1 Ref 130-091-155; all: Miltenyi Biotec, and 2. Expression levels of target genes were Bergisch Gladbach, Germany). The purity of calculated as deltaCT relative to GAPDH the resultant viable population, analyzed by (CT= CTtarget gene –CTGAPDH) with a higher flow cytometry (BD FACS Calibur, Franklin deltaCT corresponding to a lower quantity of Lakes, New Jersey, U.S.), was generally mRNA. >90%. RNA was isolated using the RNeasy Mini Kit from Qiagen (Qiagen, Venlo, The Statistical analysis: Netherlands). Total RNA concentration was After automated data extraction using measured by Nanodrop spectrophotometry Imagene 8.0 (BioDiscovery), Loess (Nanodrop Technologies, Wilmington, DE, normalization was performed on mean spot- 26 USA); RNA purity and integrity was verified intensities. Logarithmized expression using lab-on-chip technology (Agilent 2100 values were analyzed using paired and 27 Bioanalyzer, Santa Clara, CA, USA). All unpaired MAANOVA, as appropriate. In a procedures were conducted according to the fixed effect analysis, sample, array and dye manufacturer’s instruction. effects were modeled. P-values were determined by a permutation F2-test, in Microarray RNA profiling: which residuals were shuffled 10000 times RNA of CD14+ and CD4+ cells of five GC- globally. Genes with p<0.05 after family Responders and five non-Responders were wise error correction (FWER) were analyzed on human whole genome gene considered statistically significantly changed. expression microarrays v1 (Agilent For other statistical analyses SPSS version Technologies, Belgium) representing 41000 20.0 (SPSS, Chicago, IL, USA) was used. homo sapiens 60-mer probes in a 4x44K Associations between gene expression and layout. Universal human reference RNA qPCR results on the one hand and patient (Stratagene) was used as a common characteristics on the other hand were tested reference. using paired and unpaired Student t-tests as RNA amplification and labeling were well as χ-square tests, as appropriate. performed on an automated system (Caliper Reported correlations are Spearman’s. P Life Sciences NV/SA, Belgium) with 2 g of values <0.05 were considered statistically total RNA from each sample.24 significant. Hybridizations were done on a HS4800PRO system with QuadChambers (Tecan Benelux

100 Functional annotation: methotrexate (MTX) on the expression of Gene ontology analyses were performed GR-isoforms,30 concomitant use of MTX was using the Functional Annotation tool of analyzed and did not differ between groups. the Database for Annotation, Visualization However, there were more active smokers and Integrated Discovery (DAVID ) v6.7, among non-Responders (3 versus none). following the guidelines postulated by Huang After treatment (day 5) DAS28 differed et al.28 A functional annotation chart was significantly in GC-Responders and non- generated based on the GO terms responders (p<0.001). GOTERM_BP_FAT, GOTERM_CC_FAT, GOTERM_MF_FAT and on clustering of genes in pathways derived from BIOCARTA Differences in Gene expression profile in and KEGG. As a background-list we used the response to GC-treatment in AGILENT human genome content and CD14+monocytes and CD4+T-cells of GC- results were only considered significant if Responders and non-Responders: they withstood the Bonferroni correction for We analyzed cDNA microarrays from multiple testing. monocytes of 5 GC-Responders and 5 non- Responders before and 24 hours after GC- Analyses of transcription-factor (TF) pulse therapy-start. Using paired profiles: MAANOVA and a FWER multiple testing We investigated whether significantly correction we found 200 probes to be differentially expressed genes shared differentially regulated in CD14+cells of GC- common transcription factor binding sites, Responders. To concentrate on relevant which might imply that the up-regulation of changes we added an additional cutoff of a these genes are driven by a single TF. To this 1,5 fold up- or down-regulation. This aim we utilized the TRANSFAC database resulted in a list of 183 differentially (version 7.4, http://www.gene-regulation. regulated probes ( supplementary table 3). In com). Each significantly over-expressed non-Responders 666 probes were identified gene-set was annotated with a TRANSFAC by FWER correction resulting in a list of 491 record. Only TFs regulating more than 10 probes after additional cut-off of 1,5 fold genes that were significantly upregulated in change. (supplementary table 4). Assessing the microarray results were further analyzed. these lists further, 48 individual genes were Expression levels of TFs derived from the exclusively changed in GC-Responders, 253 microarray data were investigated by genes were exclusively altered in non- correlation with the target genes’ expression Responders, whereas 104 genes were level in the microarray data with Spearman’s changed in GC-Responders as well as non- Rank coefficient. 29 Responders (Figure 1A).

Due to a technical problem, one CD4+cell Results sample of T24 had to be eliminated and the analysis of CD4+T-cells was performed on 4 Patients: GC-Responders and 5 non-Responders. Patient characteristics and clinical Applying the same criteria as in monocytes, a manifestations of the 5 GC-Responders and 5 list of 50 probes in GC-Responders was non-Responders are given in Table 1. The generated (46 probes after 1,5 fold up- or mean age was 55.6(±8.1 SEM) years in GC- down-regulation cut-off; supplementary table Responders and 49.9(±9.6) in non- 5). In non-Responders 277 probes were Responders; mean disease duration was identified with FWER correction resulting in 12.8(±6), and 8.6(±3.9) years; neither 167 probes after application of the cut-off. difference was statistically significant. (supplementary table 6). Concerning the possible influence of

101 Nineteen individual genes were exclusively Responders, whereas 18 genes were changed changed in CD4+T-cells of GC-Responders, in GC-Responders as well as non-Responders 104 genes were exclusively altered in non- (Figure 1B).

GC-Responders non-Responders Age (yrs) 69 50 32 49 78 37 76 61 51 20 Gender m f f f m m f f m f MTX yes yes no no yes yes no no no yes DAS d0 5.83 7.41 8.1 6.72 5.17 7.32 6.95 4.27 8.11 4.84 ESR d0 (mm/Hg) 41 41 66 19 6 34 53 98 102 54 DAS d5 ** 2.26 2.72 3.15 2.2 2.54 4.68 4.32 3.81 3.84 3.46 ESR d5 6 13 34 5 2 13 15 71 60 31 Current smoking * no no no no no yes yes no yes no RF pos pos neg pos pos pos pos pos pos pos ACPA pos pos neg neg pos pos pos neg neg pos Disease duration (yrs) 34 17 0 7 6 10 13 20 0 0 Previous biological no yes no no yes no yes yes no no Erosive disease yes yes no no no yes yes yes no yes

Table 1: Patient characteristics of Glucocorticoid-Responders and non-Responders M: male; F: female MTX: Methotrexate; RF: rheumatoid factor; ACPA: Anti-citrullinated-protein-antibodies; DAS28: disease activity score. DAS28 d0: disease activity at baseline; DAS28 d5: disease activity after pulse therapy. ** p<0.01; * p<0.04. Statistically significant different characteristics between patient groups are written in bold.

Figure 1

Figure 1: Differentially expressed genes in CD14+monocytes and CD4+T-cells of GC-Responders and non- Responders in cDNA microarrays between T0 and T24 Using paired MAANOVA, FWER multiple testing correction and a cut-off of 1.5 fold change, the depicted number of differentially expressed genes were found in GC-responders, non-Responders or both. Figure 1A: genes found in CD14 monocytes. Figure 1B: genes found in CD4+T-cells.

Pathway analysis shows a difference in into the pathways: immune response, Interferon related genes between GC- defense response, and negative regulation Responders and non-Responders: of defense response. The genes from non- As seen in table 2a, the genes differentially Responders were distributed into the GO- regulated in CD14+cells of GC- terms: immune response, response to virus, Responders between T0 and T24 were defense response, response to wound clustered according Gene Ontology (GO) healing, and inflammatory response, as

102 listed in Table 2b. It is noteworthy that the Pathway analysis of CD4+T-cells are response to virus was highly and very shown in table 3a. Only the calcium ion significantly enriched in non-Responders binding pathway was overrepresented in exclusively (table 2c). Additionally, GO CD4+T-cells of GC-Responders. The pathways represented in both groups were respective genes turned out to be composed differently. Apart from genes differentially regulated in non-Responders (e.g.: VSIG4, CXCL10, IL1R2, as well (supplementary list 6). In non- SERPING1) regulated by GC in both Responders genes were clustered into the groups, most pathways also contained pathways: response to virus, IL-22 soluble several genes exclusively regulated in non- Receptor signaling pathway, and immune Responders. The majority of these non- response, as listed in Table 3b. As in Responder–exclusive genes were monocytes, the interferon related genes interferon related, (see table 2c, genes held were highly and very significantly enriched in Italic), indicating a differential in non-Responders, which was reflected in interferon signature in monocytes in the the GO-pathway ”response to virus” as response to GC between GC-Responders well as in other pathways. and non-Responders.

Table 2A GO-TERM Fold Number p-value Genes enrichment of genes (Bonferroni) Immune Response 3.8 23 1.2 E-4 OAS3, CD1E, CD55, CD97, SLAMF7, VSIG4, ADA, BPI, CCL2, CXCL10, C5AR1, ERAP2, GBP1, GBP4, GBP5, IFI44L, IL1R2, IL18, PPARG, SERPING1, TAP1, TNFSF10, TNFSF13B Defense Response 3.6 20 2.5 E-3 OAS3, SLAMF7, VSIG4, BPI, CXCL10, GBP1, GBP4, GBP5, IFI44L, IL1R2, SERPING1 Negative 17.1 6 3.1 E-2 ADA, GPX1, PPARG, SERPING1, Regulation of UACA Defense Response

Table 2A: GO-terms of differentially expressed genes in CD14+cells between T0 and T24 in GC-Responders. Genes tested by qPCR are written in bold. Interferon-related genes are depicted in Italic.

Table 2B: GO-TERM Fold Number p-value Genes enrichment of genes (Bonferroni) Immune Response 3.8 53 2.1 E-13 OAS1, OAS2, OAS3, OASL, CLEC4D, CLEC5A, CD55, DDX58, DHX58, FYB, GTPBP1, GCH1, POU2AF1, SLAMF7, ST6GAL1, VSIG4, AIM2, APOBEC3G, AQP9, BPI, CCR5, CXCL10, CXCL9, CR1, CNIH, GBP1, GBP2, GBP3, GBP4, GBP5, IFITM2, IFITM3, IFIH1, IFI6, IFI35, IFI44L, IL1RN, IL1R2, IL18, IL2RG, IL23A, HLA-DQA2, PSMB9, PTPRC, P2RY14, RSAD2, SERPING1, HLA- DQA1, TAP1, TRIM22, TNFSF10, TNFSF13B, RELB

103 Response to Virus 8.4 19 1.5 E-8 DDX58, ISG15, APOBEC3G, EIF2AK2, IFIH1, IRF7, IFI16, IFI35, IFI44, IL23A, MX1, MX2, PLSCR1, PTPRC, RSAD2, STAT1, STAT2, TRIM22, TRIM5 Defense Response 3.0 38 1.1 E-5 CLEC5A, CD44, CD48, CD55, DDX58, DHX58, GCH1, NMI, PXK, SLAMF7, VSIG4, ADORA3, APOBEC3G, BPI, CHST2, CCR5, CXCL10, CXCL9, CR1, FN1, HIF1A, IRF7, IL1RN, IL23A, LIPA, LYZ, MX1, MX2, PLA2G7, PTPRC, RSAD2, RNASE6, SERPING1, SIGLEC, TAP1, TPST1 Response to 2.7 31 2.8 E-3 CD36, CD44, CD55, NMI, PXK, Wounding VSIG4, ADORA3, CHST2, CCR5, CXCL10, CXCL9, CR1, CYSLTR1, DTNBP1, FN1, GPX1, HIF1A, IRF7, IL1RN, IL23A, LIPA, LYZ, NRG1, NRP1, PLA2G7, PLSCR1, P2RY12, SERPING1, SIGLEC1, SOD2, TPST1 Inflammatory 3.2 23 4.7 E-3 CD44, CD55, NMI, PXK, VSIG4, Response ADORA3, CHST2, CCR5, CXCL10, CXCL9, CR1, CYSLTR1, FN1, HIF1A, IRF7, IL1RN, IL23A, LIPA, LYZ, PLA2G7, SERPING1, SIGLEC1,

TPST1

Table 2B: GO-terms of differentially expressed genes in CD14+cells between T0 and T24 in non-Responders. Genes tested by qPCR are written in bold. Interferon-related genes are depicted in Italic.

Table 2C: GO-TERM Fold Number of p-value Genes enrichment genes (Bonferroni) Response to virus 10.9 17 4.1 E-9 DDX58, ISG15, APOBEC3G, EIF2AK2, IFIH1, IRF7, IFI16, IFI35, IFI44, IL23A, MX2, PLSCR1, PTPRC, RSAD2, STAT2, TRIM22, TRIM5 Immune response 3.8 37 8.4 E-9 OAS1, OAS2, OASL, CLEC4D, CLEC5A, DDX58, DHX58, FYB, GTPBP1, GCH1, POU2AF1, ST6GAL1, AIM2, APOBEC3G, AQP9, CCR5, CXCL9, CR1, CNIH, GBP2, GBP3, IFITM2, IFITM3, IFIH1, IFI6, IFI35, IL1RN, IL2RG, IL23A, HLA- DQA2, PSMB9, PTPRC, P2RY14, RSAD2, HLA-DQA1, TRIM22, RELB Defense Response 2.7 24 3.8 E-2 CLEC5A, CD48, DDX58, DHX58, GCH1, NMI, PXK, ADORA3, APOBEC3G, CCR5, CXCL9, CR1, CYSLTR1, HIF1A, IFIH1, IRF7, IL1RN, IL23A, LYZ, MX2, PTPRC, RSAD2, SIGLEC1, TPST1

Table 2C: GO-terms of differentially expressed genes in CD14+cells between T0 and T24 in non-Responders exclusively. Genes tested by qPCR are written in bold. Interferon-related genes are depicted in Italic.

104 Table 3A: GO-TERM Fold Number p-value Genes enrichment of genes (Bonferroni) Calcium ion 3.7 12 1.4 E-4 CD97, S100A12, S100A8, S100A9, binding AIF1, ANXA3, CACNA1I, LDLR, MMP8, MPO, PADI4, STAT1

Table 3A: GO-terms of differentially expressed genes in CD4+cells between T0 and T24 in GC-Responders.

Table 3B: GO-TERM Fold Number p-value Genes enrichment of genes (Bonferroni) Response to virus 15.2 12 2.6 E-7 ISG15, BST2, EIF2AK2, IFIH1, IRF7, IRF9, IFI44, LCN2, MX1, MX2, RSAD2, STAT1 IL-22 soluble 40.3 4 2.4 E-3 JAK3, STAT1, STAT3, SOCS3 Receptor Signaling Pathway Immune response 3.7 18 4.8 E-3 OAS1, OAS2, OAS3, OASL, CLEC4E, BST2, CRISP3, GBP1, IFITM2, IFITM3, IFIH1, IFI44L, LTF, LTB, PGLYRP1, RSAD2, RELB, VIPR1

Table 3B: GO-terms of differentially expressed genes in CD4+cells between T0 and T24 in non-Responders.

Table 3C: GO-TERM Fold Number p-value Genes enrichment of genes (Bonferroni) Response to virus 16.5 11 7.3 E-7 ISG15, BST2, EIF2AK2, IFIH1, IRF7, IRF9, IFI44, LCN2, MX1, MX2, RSAD2, Immune response 4.3 18 3.7 E-4 OAS1, OAS2, OAS3, OASL, CLEC4E, BST2, CRISP3, GBP1, IFITM2, IFITM3, IFIH1, IFI44L, LTF, LTB, PGLYRP1, RSAD2, RELB, VIPR1

Table 3C: GO-terms of differentially expressed genes in CD4+cells between T0 and T24 in non-Responders exclusively.

Quantitative real-time PCR validates Responders and 36.19 (p<0.01) in non- genes and GO-pathways Responders) and also the down-regulation of To validate the findings from the cDNA CXCL10 conforms to earlier reports.32 microarryas, several genes were selected for Adenosine deaminase (ADA) is up-regulated qPCR testing. Selection was based on the in inflammation and over-expressed in RA,33 magnitude of fold change/ratio seen in the and was found down-regulated in GC- microarrays and the aim to validate all Responders (ratio: 0.51; p=0.05; non- represented pathways. The qPCR-results of Responders: ratio: 0.66; p=0.08). monocytes of GC-Responders are shown in In addition to the genes mentioned above, Figure 2A. Almost all genes tested confirmed genes exclusively found in monocytes of the microarray results and Gene Ontology non-Responders by pathway analysis, analysis. Some of the genes tested are well- containing mostly down-regulated interferon known targets of GC-action, others have not (IFN) related genes (IFIH1, IRF7, ISG15, yet been reported. The up-regulation of the STAT1, CXCL10, RSAD2) were also tested known GC-target gene IL-1R231 was robust by qPCR (Figure 2B), and were all (fold change (FC): 129.88 (p<0.001) in GC- significantly down regulated by GC.

105

Figure 2: Gene expression in CD14+monocytes of GC-Responders and non-Responders at T0 and T24 by qPCR. CT (CT gene of interest – CT GAPDH) levels are given as quantification of gene expression. Higher CT correspond to a lower gene expression. Figure 2A: GC-Responders. Figure 2B: non-Responders. P<0.05 (paired Students T-test) was considered statistically significant. *: p<0.05; **: p<0.01.

Figure 3: Gene expression in CD4+T-cells of GC-Responders and non-Responders at T0 and T24 by qPCR. CT (CT gene of interest – CT GAPDH) levels are given as quantification of gene expression. Higher CT correspond to a lower gene expression. Figure 3A: GC-Responders. Figure 3B: non-Responders. P<0.05 (paired Students T test) was considered statistically significant. *: p<0.05; **: p<0.01.

Validation of CD4+ cells microarray results in RA, and attributed a pro-inflammatory by qPCR are shown in Figure 3A and 3B, potential.34 However, although increased confirming microarray results. gene expression of the calgranulins S100A8 Different members of the S100 family of and S100A9 have been reported after GC- Calcium-binding proteins are over-expressed therapy, this is the first report of GC induced

106 up-regulation of Calgranulin C (S100A12; Different expression of interferon related Figure 3A; Responders: FC: 8.45; p<0.01; genes in monocytes and CD4+T-cells of non-Responders: FC 7.62; p<0.01). GC-Responders and non-Responders measured by qPCR: STAT1, the signal transducer and activator of The observed trend of a stronger suppression transcription 1, has multiple functions, of IFN related genes in non-Responders, did among which Ca-ion binding. Nevertheless, not reach statistical significance in the main function is signal transduction, monocytes (Figure 4A). However, the down- following IFN or IFN ligation of the regulation of these genes in CD4+T-cells was respective cell-surface Receptors. The down- significantly stronger in non-Responders than regulation of STAT1 induced by GC was in GC-Responders and reached statistical seen in GC-Responders (ratio: 0.29; p<0.002) difference for IFIH1, IRF7 and IFI44L and non-Responders (ratio: 0.23; p<0.00008) (Figure 4B). Supplementary Figure 1 depicts and accompanied by down-regulation of a scheme of the common pathway of these other IFN related genes, exclusively found in genes. non-Responders (IFIH1, IRF7, IFI44L, ISG15,and RSAD2, Figure 3B)

Figure 4: Change in qPCR gene expression of Interferon related genes between T0 and T24 of CD14+monocytes and CD4+T-cells compared between GC-Responders and non-Responders. (DCT= CT T24 – CTT0 of CD14+monocytes (Figure 4A; including CD55, a non-interferon related gene, which was significantly less suppressed in non-Responders than GC-Responders after 24hours) and CD4+T-cells (Figure 4B). P<0.05 (unpaired Students T-test) was considered statistically significant. *: p<0.05; **: p<0.01.

Analyses of transcription-factor (TF) In monocytes, the transcription-factors were profiles LEF1, NFAT-1, IRF-1 and NFkB1 (Table Several transcription-factors were found in 4a); in CD4+ cells, LEF1 and SP1 were both cell types from non-Responders, which found (Table 4b). are able to influence multiple genes from the Testing identified transcription-factors for generated gene lists described above. correlation in gene expression values found

107 in microarrays with the ones from their target Taken together, these findings statistically genes, several correlations were significant. bind the up-regulation of genes in both The most striking correlations for monocytes and CD4+T-cells of non- CD14+cells were between LEF1 and DDR1, responders to a distinct transcription-factor between NFAT-1 and NT5C2 and between profile and might indicate that a specific NFkB1 and IL1RN (all: Spearmans Rho: 1.0; transcription-factor network might actually p<0.01), furthermore between IRF-1 and orchestrate the processes underlying non- NT5C3 (Rho: 0.9; p=0.037). For CD4+cells Respondership. Noteworthy, LEF1, a T-cell these were the correlations of LEF1 and receptor enhancer and regulator of the WNT SH3BP5 (Rho: 0.9; p=0.037), and SP1 and pathway, is significantly correlated with MAP2 (Rho: 0.975; p<0.01). multiple genes in both CD4+ and CD14+ cells from non-Responders.

Table 4A: Transcription-factors regulating multiple significant gene targets in the CD14+cells from non- Responders Transcription Significantly correlated target gene, which are differentially expressed between T0 and factor T24 FYN, CD226, P2RY12, LSCR1, LIMK2, BNIP2, DNAJA1, ALDH6A1, NRG1, LEF1 SLITRK2, MYO10, LMNB1, POU2AF1, MAF, TBC1D8, GCH1, IL13RA1, KLHL6, PXK, PIM1, C20orf27, KLHL5, STMN4, MAT2A, DDR1, MT2A APEX1, PLAGL2, CASP8 LDHB, LIMK2, DNAJA1, PER2, PDE4B, NRG1, KCNMA1, IL23A, GTPBP1, NFAT-1 POU2AF1, HIF1A, NT5C2, TBC1D8, HESX1, NR3C1, IL1RN, LPL, LIMS1, PIM1, FYB, LAP3 , ADAMTSL3, CNIH, PLAGL2, IL2RG DDX58, IFIT2, CASP7, ADORA3, IFI44, PSMB9, NRG1, NT5C3, ADAR, UBE2L6, IRF-1 IFI35

NFkB1 SLAMF8, RELB, XPO6, IL23A, JAK3, NR3C1, IL1RN, CCND2, DDR1, MT2A

Table 4B: Transcription-factors regulating multiple significant gene targets in the CD4+cells from non- Responders Transcription Significantly correlated target gene factor MAP2, CNP, PTMA C14orf132, LTF, TMSB4X, DHRS9, DDIT4, SH3BP5, PIM2, LEF1 TMSL4, STAT3, OGT MAP2, SESN1, BTG1, RELB, PTMA, C14orf132, NOLC1, TPM2, FBXO36, SH3BP5, SP1 PIM2, CALR, GPR37, OGT

Table 4: Transcription-factors regulating multiple significant gene targets in the CD14+monocytes and CD4+T- cells from non-Responders. Transcription-factors and genes with shared common transcription factor binding sites derived from the lists of differentially expressed genes are shown for CD14+monocytes (Table 4A) and CD4+T-cells (Table 4B). Genes, which are significantly correlated, are kept in bold.

108 Discussion: that blockade of TNF results in reduced interferon response gene expression in The aim of this study was to identify RA.42,43 mechanisms specific for clinical response to In our study this interplay is further GC in RA by examining the differences in complicated by the influence of GC, which GC-induced gene expression in monocytes are known to regulate TNF as well as and T-cells between GC-Responders and interferons, and an expected down-regulation non-Responders. A more pronounced down- of interferon response genes was seen in both regulation of interferon related genes was cell types. Although present in all patients, seen in non-Responders in both cell types, this was more prominent in non-Responders, and was validated by qPCR for T-cells. which was confirmed in qPCR experiments Moreover, several relevant transcription- in CD4+T-cells and monocytes. Due to the factors were identified, including LEF1 in interplay of interferons and TNF it cannot both cell types. be excluded that this enhanced down- The interferon signature is a relatively new regulation in non-Responders is in fact factor in RA-pathogenesis and is observed in caused by effects of GC on TNF production approximately half of RA-patients.35 and signaling. In our transcription factor Interferons are pleiotropic cytokines, which analysis, both IRF-1 as well as NFkB1 were not only confer an antiviral state, but are also found to be regulated by GC in non- involved in lymphoid differentiation, Responders (Table 4a), making a distinction homeostasis, tolerance, and memory; both between a direct effect of GC on IFN and an type I (IFN) and type II interferons indirect via TNF difficult. However, (IFN) are known to contribute to RA- whereas an up-regulation of IFN signature pathogenesis. denominated a worse response after TNF- The interferon signature has also been blockade, a stronger down-regulation of implicated in therapy response to biologicals. interferon related genes was seen in non- An increased baseline expression of IFN Responders to GC. The differential related genes was reported to be predictive of expression of interferon related genes thus worse clinical response to Rituximab in two seems to be important in the response of RA trials, including one using cDNA microarray patients to several therapeutic approaches. analysis,36,37 whereas an increase after This warrants further investigation in order to Rituximab treatment, reaching the levels of find the optimal treatment combination (GC non-Responders, correlated with a good and / or specific biologicals) for individual clinical response.38 In contrast, higher pre- patients. treatment type I interferon activity indicated The influence of GC on specific parts of the a favorable response toTNF-blokkade.39 complement system, mirrored by up- Interestingly, a different study, analyzing the regulation of VISG4, a complement Receptor interferon response gene activity one month of Ig superfamily and the down-regulation of after start of anti-TNF therapy, showed a SERPING1, the C1 esterase inhibitor, has not correlation between non-respondership and been reported to our knowledge before. It is an increase of the interferon signature.40 known that steroids affect CD55(DAF), These results reflect the complex interplay of another component of the complement TNF and type I interferons in general and in system, although this has mainly been RA in specific. The original idea of a mutual reported for progesterone. Recently an negative regulation came forth from the (unexpected) down-regulation of CD55 by observation that TNF-blockade induced an dexamethasone has been suggested, which is increased interferon signature.41 However, confirmed by our findings of a decreased this paradigm has been challenged by the expression after 24 hours (GC-Responders: discovery that TNF incubation induces IRF1 ratio 0.12; p<0.01; non-Responders: ratio: and STAT1 in primary macrophages, and 0.23; p<0.01).44

109 In patients with asthma, NFkB1 had a 81% in monocytes and CD4+T-cells of GC- accuracy in predicting the GC-response.20 In sensitive and -resistant RA-patients based on our microarray results, NFkB1 was identified interferon related genes. As this has not been as a transcription factor of importance in seen in comparable studies in asthma or ALL monocytes from non-Responders. In patients, it is possible that the resistance monocytes of GC-Responders and non- mechanisms may be disease-specific or may Responders, the actual expression values of depend on common pathogenic patterns as an NFkB1 were slightly, non-significantly interferon signature. Considering the latter, down-regulated after 24 hours (ratio 0.8). our findings might lead to new insight into Interestingly, expression in CD4+cells was treatment-options of several auto-immune down-regulated (ratio 0.29; p=ns) in non- diseases with a dominant role of the Responders, but slightly up-regulated in GC- interferon signature. Responders (FC: 1.26; p=ns). The relevance of this finding is not clear, especially as it did not withstand FWER correction. Another Acknowledgements: noteworthy finding was that RELB was significantly down-regulated in monocytes We are thankful for the use of the microarray and CD4+T-cells of non-Responders only, facilities within the department of Molecular which implies a differential role of NFkB in Cancer Research. Furthermore we would like GC-response also in RA-patients. to thank Alexander Ploner for critical reading of the manuscript. Another study characterized the microarray gene expression in bronchoalveolar macrophages of asthma patients, finding a References higher expression of LPS induced genes in GC-resistant patients, which the authors also recently described in PBMC. 45, 46 In our 1) Gregersen PK, Silver J, Winchester RJ. direct analysis between GC-Responders and The shared epitope hypothesis: an non-Responders 24 hours after start of approach to understanding the molecular therapy, but also before therapy (data not genetics of susceptibility to rheumatoid shown) these were not differentially arthritis. Arthritis Rheum 1987;30:1205- expressed. 13. 2) Begovich AB, Carlton VE, Honigberg In infant acute lymphoblastic leukemia LA, Schrodi SJ, Chokkalingam AP, (ALL) cells, GC-resistance is partly caused Alexander HC, et al. A missense single- through increased expression of S100A8 and nucleotide polymorphism in a gene S100A9 and consecutively hampered Ca-flux encoding a protein tyrosine phosphatase to the mitochondria, which is necessary for (PTPN22) is associated with rheumatoid GC-induced apoptosis.47 In our data, arthritis. Am J Hum Genet 2004;75:330- calgranulins are induced upon GC-exposure 7. in all patients and show no significant 3) Remmers EF, Plenge RM, Lee AT, difference between non-Responders and GC- Graham RR, Hom G, Behrens TW, et al. Responders in CD4+T-cells. STAT4 and the risk of rheumatoid The discrepancy between the findings of arthritis and systemic lupus these mentioned studies and the present one erythematosus. N Engl J Med suggest a disease-specific GC-resistance 2007;357:977-86. pattern in RA-patients. 4) Fritsch R, Eselböck D, Skriner K, Jahn- Schmid B, Scheinecker C, Bohle B et al. Altogether this study describes for the first Characterization of autoreactive T cells to time a difference in GC-working mechanism the autoantigens heterogeneous nuclear

110 ribonucleoprotein A2 (RA33) and 14) van Schaardenburg D, Valkema R, filaggrin in patients with rheumatoid Dijkmans BA, Papapoulos S, arthritis. J Immunol. 2002 Jul Zwinderman AH, Han KH, et al. 15;169:1068-76. Prednisone treatment of elderly-onset 5) Cantaert T, Brouard S, Thurlings RM, rheumatoid arthritis. Disease activity and Pallier A, Salinas GF, Braud C, et al. bone mass in comparison with Alterations of the synovial T cell chloroquine treatment. Arthritis Rheum. repertoire in anti-citrullinated protein 1995;38:334-42. antibody-positive rheumatoid arthritis. 15) Seki M, Ushiyama C, Seta N, Abe K, Arthritis Rheum. 2009;60:1944-56. Fukazawa T, Asakawa J, et al. Apoptosis 6) Feitsma AL, van der Voort EI, Franken of lymphocytes induced by KL, el Bannoudi H, Elferink BG, glucocorticoids and relationship to Drijfhout JW et al. Identification of therapeutic efficacy in patients with citrullinated vimentin peptides as T cell systemic lupus erythematosus. Arthritis epitopes in HLA-DR4-positive patients Rheum. 1998;41:823-30. with rheumatoid arthritis. Arthritis 16) Schwartz HJ, Lowell FC, and Melby JC. Rheum. 2010;62:117-25. Steroid resistance in bronchial asthma. 7) McInnes IB, Schett G. The pathogenesis Ann. Intern. Med.1968; 69:493-499. of rheumatoid arthritis.N Engl J Med. 17) Barnes PJ, Adcock IM. Glucocorticoid 2011;365:2205-19. resistance in inflammatory diseases. 8) Buch MH, Vital EM, Emery P. Abatacept Lancet. 2009;373:1905-17. in the treatment of rheumatoid arthritis. 18) Kozaci DL, Chernajovsky Y, Chikanza Arthritis Res Ther 2008;10:S5. IC. The differential expression of 9) Hench, P. Effects of cortisone in the corticosteroid receptor isoforms in rheumatic diseases. Lancet 1950;483–4. corticosteroid-resistant and -sensitive 10) Hoes JN, Jacobs JW., Buttgereit F & patients with rheumatoid arthritis. Bijlsma JW. Current view of Rheumatology (Oxford). 2007;46:579- glucocorticoid co-therapy with DMARDs 85. in rheumatoid arthritis. Nat. Rev. 19) Chikanza IC and Kozaci DL. Rheumatol.2010; 6, 693–702. Corticosteroid resistance in rheumatoid 11) Forster PJ, Grindulis KA, Neumann V, arthritis: molecular and cellular Hubball S, McConkey B. High-dose perspectives. Rheumatology (Oxford) intravenous methylprednisolone in 2004;43:1337-45. rheumatoid arthritis. Ann Rheum Dis. 20) Hakonarson H, Bjornsdottir US, Halapi 1982;41:444-6. E, Bradfield J, Zink F, Mouy M et al. 12) Jacobs JW, Geenen R, Evers AW, van Profiling of genes expressed in peripheral Jaarsveld CH, Kraaimaat FW, Bijlsma blood mononuclear cells predicts JW. Short term effects of corticosteroid glucocorticoid sensitivity in asthma pulse treatment on disease activity and patients. Proc Natl Acad Sci U S A. the wellbeing of patients with active 2005;102:14789-94. rheumatoid arthritis. Ann Rheum Dis. 21) Smith SL, Plant D, Eyre S, Barton A. The 2001 Jan;60:61-4. potential use of expression profiling: 13) de Jong PH, Quax RA, Huisman M, implications for predicting treatment Gerards AH, Feelders RA, de Sonnaville response in rheumatoid arthritis. Ann PB, et al. Response to glucocorticoids at Rheum Dis. 2013;72:1118-24. 2 weeks predicts the effectiveness of 22) Arnett FC, Edworthy SM, Bloch DA, DMARD induction therapy at 3 months: McShane DJ, Fries JF, Cooper NS, et al. post hoc analyses from the tREACH The American Rheumatism Association study. Ann Rheum Dis. 2012 Oct 31. 1987 revised criteria for the classification [Epub ahead of print]

111 of rheumatoid arthritis. Arthritis Rheum. isoforms in normal human peripheral 1988;31:315-24. mononuclear cells and human 23) Prevoo ML, van 't Hof MA, Kuper HH, lymphocyte cell lines in vitro. Mol van Leeuwen MA, van de Putte LB, van Immunol. 2007;44:2115-23. Riel PL. Modified disease activity scores 31) Re F, Muzio M, De Rossi M, Polentarutti that include twenty-eight-joint counts. N, Giri JG, Mantovani A, et al. The type Development and validation in a II “Receptor” as a decoy target for prospective longitudinal study of patients interleukin 1 in polymorphonuclear with rheumatoid arthritis. Arthritis leukocytes: characterization of induction Rheum 1995;38:44-8. by dexamethasone and ligand binding 24) Roepman P, Wessels LF, Kettelarij N, properties of the released decoy receptor. Kemmeren P, Miles AJ, Lijnzaad P, et al. J Exp Med. 1994;179:739–43. An expression profile for diagnosis of 32) Hu X, Li WP, Meng C, Ivashkiv LB. lymph node metastases from primary Inhibition of IFN-gamma signaling by head and neck squamous cell carcinomas. glucocorticoids.J Immunol. Nat Genet.2005;37:182-6. 2003;170:4833-9. 25) van de Peppel J, Kemmeren P, van Bakel 33) Sari RA, Taysi S, Yilmaz O, Bakan N. H, Radonjic M, van Leenen D, Holstege Correlation of serum levels of adenosine FC. Monitoring global messenger RNA deaminase activity and its isoenzymes changes in externally controlled with disease activity in rheumatoid microarray experiments. EMBO Rep. arthritis. Clin Exp Rheumatol. 2003;4:387-93. 2003;21:87-90. 26) Yang YH, Dudoit S, Luu P, Lin DM, 34) Chen YS, Yan W, Geczy CL, Brown Peng V, Ngai J,et al. Normalization for MA, Thomas R. Serum levels of soluble cDNA microarray data: a robust receptor for advanced glycation end composite methodaddressing single and products and of S100 proteins are multiple slide systematic variation. associated with inflammatory, Nucleic Acids Res. 2002; 30:15. autoantibody, and classical risk markers 27) Wu H, Kerr MK, Cui X, and Churchill of joint and vascular damage in GA. 2003. MAANOVA: A software rheumatoid arthritis. Arthritis Res Ther package for the analysis of spotted cDNA 2009;11: R39. microarray experiments, in Parmigiani, 35) van der Pouw Kraan TC, Wijbrandts CA, G., Garrett, E.S., Irizarry, R.A, and van Baarsen LG, Voskuyl AE, Zeger, S.L., eds., The Analysis of Gene Rustenburg F, Baggen JM, et al. Expression Data: Methods and Software, Rheumatoid arthritis subtypes identified Springer, New York. by genomic profiling of peripheral blood 28) Huang da W, Sherman BT, Lempicki cells: assignment of a type I interferon RA. Systematic and integrative analysis signature in a subpopulation of patients. of large gene lists using DAVID Ann Rheum Dis 2007; 66:1008–14. bioinformatics resources. Nat Protoc 36) Raterman HG, Vosslamber S, de Ridder 2009;4:44-57. S, Nurmohamed MT, Lems WF, Boers 29) Wingender E. "The TRANSFAC project M, et al. The interferon type I signature as an example of framework technology towards prediction of non-response to that supports the analysis of genomic rituximab in rheumatoid arthritis patients. regulation". Brief. Bioinformatics 2008; Arthritis Res Ther. 2012;14:R95. 9: 326–32. 37) Thurlings RM, Boumans M, Tekstra J, 30) Goecke IA, Alvarez C, Henríquez J, van Roon JA, Vos K, van Westing DM, Salas K, Molina ML, Ferreira A et al. et al. Relationship between the type I Methotrexate regulates the expression of interferon signature and the response to glucocorticoid receptor alpha and beta

112 rituximab in rheumatoid arthritis patients. response genes. Nat Immunol. Arthritis Rheum. 2010;62(12):3607-14. 2008;9:378-87. 38) Vosslamber S, Raterman HG, van der 43) Gordon RA, Grigoriev G, Lee A, Pouw Kraan TC, Schreurs MW, von Kalliolias GD, Ivashkiv LB. The Blomberg BM, Nurmohamed MT, et al. interferon signature and STAT1 Pharmacological induction of interferon expression in rheumatoid arthritis type I activity following treatment with synovial fluid macrophages are induced rituximab determines clinical response in by tumor necrosis factor α and counter- rheumatoid arthritis.Ann Rheum Dis. regulated by the synovial fluid 2011;70:1153-9. microenvironment.Arthritis Rheum. 39) Mavragani CP, La DT, Stohl W, Crow 2012;64:3119-28. MK. Association of the response to tumor 44) Nowicki S, Rana T, Singh R, Nowicki B. necrosis factor antagonists with plasma Dexamethasone treatment downregulates type I interferon activity and interferon- DAF and decrease intracellular survival beta/alpha ratios in rheumatoid arthritis od DR+ E. coli in HeLa cells. In: patients: a post hoc analysis of a Proceedings of the 5th international predominantly Hispanic cohort. Arthritis conference on complement therapeutics, Rheum. 2010;62:392-401. abstract 18, Rhodes, 22–27 June 2011. 40) van Baarsen LG, Wijbrandts CA, 45) Goleva E, Hauk PJ, Hall CF, Liu AH, Rustenburg F, Cantaert T, van der Pouw Riches DW, Martin RJ, et al. Kraan TC, Baeten DL, et al. Regulation Corticosteroid-resistant asthma is of IFN response gene activity during associated with classical antimicrobial infliximab treatment in rheumatoid activation of airway macrophages. J arthritis is associated with clinical Allergy Clin Immunol. 2008;122:550-9. response to treatment. Arthritis Res Ther. 46) Goleva E, Jackson LP, Gleason M, Leung 2010;12:R11. DY. Usefulness of PBMCs to predict 41) Palucka AK, Blanck JP, Bennett L, clinical response to corticosteroids in Pascual V, Banchereau J. Cross- asthmatic patients. J Allergy Clin regulation of TNF and IFN-alpha in Immunol. 2012 ;129:687-93. autoimmune diseases. Proc Natl Acad Sci 47) Spijkers-Hagelstein JA, Schneider P, U S A. 2005;102:3372-7. Hulleman E, de Boer J, Williams O, 42) Yarilina A, Park-Min KH, Antoniv T, Hu Pieters R, et al. Elevated X, Ivashkiv LB. TNF activates an IRF1- S100A8/S100A9 expression causes dependent autocrine loop leading to glucocorticoid resistance in MLL- sustained expression of chemokines and rearranged infant acute lymphoblastic STAT1-dependent type I interferon- leukemia. Leukemia. 2012;26:1255-65.

113 Supplementary material

Supplementary Table 1: Sequence (5' - 3') Forward Reverse ADA F’ GCCTTCGACAAGCCCAAAGTA ADA R’ CTCTGCTGTGTTAGCTGGGAG CEP350 F’ ATCGTGTGGAATTTCGTGAACC CEP350 R’ TCCGTTCTTCTCGACTGCCTA DICER1 F’ AATGCTGAAACTGCAACTGAC DICER1 R’ TCCACAATCCACCACAATC LIPA F’ TCTGGACCCTGCATTCTGAG LIPA R’ CACTAGGGAATCCCCAGTAAGAG LNPEP F’ ACCAGATGTGGTGGATTTAGCC LNPEP R’ CAGTTGCACTGTTCCGAAGG MYCBP2 F’ GGGGACGGATTCTACCCAG MYCBP2 R’ ATTGAGCGCAGCGGTATAAAT ISG15 F’ CGCAGATCACCCAGAAGATCG ISG15 R’ TTCGTCGCATTTGTCCACCA IFIH1 F’ TCACAAGTTGATGGTCCTCAAGT IFIH1 R’ CTGATGAGTTATTCTCCATGCCC IRF7 F’ GCCTGGACACTGGTTCAACA IRF7 R’ CAAGGAGCCACTCTCCGAAC STAT1 F’ AAAGGAAGCACCAGAGCCAAT STAT1 R’ TCCGAGACACCTCGTCAAAC RSAD2 F’ RSAD2 R’ CCACGGCCAATAAGGACATT GTGGTTCCAGAATTATGGTGAGTATTT VSIG4F’ TGGATGACCGGAGCCACTAC VSIG4 R’ ACTTGGTTGCCATCAGGAGTCT SERPING1 F’ CCAAGATGCTATTCGTTGAACCC SERPING1 R’ TGGTGGCTGAATTGGTTGTTG CXCL10 F’ TTCCTGCAAGCCAATTTTGT CXCL10 R’ ATGGCCTTCGATTCTGGATT CD55 F’ TGAATCGAGATGTCCATAGTCAA CD55 R’ CGGATCATTTATTTCTATTTAGGAA IL1R2 F’ ATGACACCCACATAGAGAGCGC IL1R2 R’ GTGCAAATCCTCTCTTGTGACAG ERAP2 F’ TGGATGGGACCAACTCATTACA ERAP2 R’ TGCACCAACTAGCTGAAACAC FAM26F F’ CACCCGATGCCTATCTCCAG FAM26F R’ TTTGCTGCCACTCTTTCATGC

Supplementary Table 1: Primers used in q-PCR of CD14+monocytes

Supplementary Table 2: Sequence (5' - 3') Forward Reverse GAPDH F’ AGAAGGCTGGGGCTCATTT GAPDH R’ GAGGCATTGCTGATGATCTTG ERAP2 F’ TGGATGGGACCAACTCATTACA ERAP2 R’ TGCACCAACTAGCTGAAACAC FAM26F F’ CACCCGATGCCTATCTCCAG FAM26F R’ TTTGCTGCCACTCTTTCATGC CNB2 F’ GCATTATCATCCTTCTAAGGTAGCA CCNB2 R’ TGTAATACTGCTGCTTTAAGTTCCA ADA F’ GCCTTCGACAAGCCCAAAGTA ADA R’ CTCTGCTGTGTTAGCTGGGAG CHEK1 F’ ATATGAAGCGTGCCGTAGACT CHEK1 R’ TGCCTATGTCTGGCTCTATTCTG LNPEP F’ ACCAGATGTGGTGGATTTAGCC LNPEP R’ CAGTTGCACTGTTCCGAAGG TPM4 F’ ACGGTTGCAAAACTGGAAAA TPM4 R’ TTGGCTCTGGATGGAAAATC FABP5 F’ TTGGTTCAGCATCAGGAGTG FABP5 R’ CCTGTCCAAAGTGATGATGG ISG15 F’ CGCAGATCACCCAGAAGATCG ISG15 R’ TTCGTCGCATTTGTCCACCA IFIH1 F’ TCACAAGTTGATGGTCCTCAAGT IFIH1 R’ CTGATGAGTTATTCTCCATGCCC IRF7 F’ GCCTGGACACTGGTTCAACA IRF7 R’ CAAGGAGCCACTCTCCGAAC STAT1 F’ AAAGGAAGCACCAGAGCCAAT STAT1 R’ TCCGAGACACCTCGTCAAAC

114 S100A12 F’ AAAGGAGCTTGCAAACACCATC S100A12 R’ CAGGCCTTGGAATATTTCATCAA RSAD2 F’ RSAD2 R’ CCACGGCCAATAAGGACATT GTGGTTCCAGAATTATGGTGAGTATTT IFI44L F’ CCGAGCGGTATAGGATATATTCTGTT IFI44L R’ TGCTCCTTCTGCCCCATCTA CDCA7 F’ TGTTTTCTAGCGCACGCTTAC CDCA7 R’ CTCATCCCGAGAGTCATCCTC Supplementary Table 2: Primers used in q-PCR of CD4+T-cells

Supplementary Table 3: reporterId Fold change / Ratio pval.FWER pval.BH pval.ttest geneSymbol HSAD040888 0.171147547 0 0 0 CDKN1C HSAD021898 0.271791908 0 0 0 SEL1L3 HSAD033155 0.277122651 0 0 0 ARID5B HSAD001824 0.345574637 0 0 1.11E-16 STAT1 HSAD028888 0.361356102 0 0 7.11E-15 IFI44L HSAD035570 0.36726142 0 0 0 ANKRD22 HSAD029251 0.375783684 0 0 4.44E-16 BPI HSAD031902 0.421834962 0 0 1.11E-16 XAF1 HSAD026388 0.42824604 0 0 0 MYB HSAD036741 0.512063905 0 0 4.88E-14 CD36 HSAD024002 0.531231679 0 0 1.11E-16 TNNT1 HSAD025893 0.552598531 0 0 2.01E-13 NEURL HSAD027318 0.553467456 0 0 1.11E-16 RAB37 HSAD029340 0.562351001 0 0 2.64E-13 TAP1 HSAD011565 0.57819631 0 0 1.07E-13 RPH3A HSAD013804 0.584320922 0 0 0 FAM134B HSAD026667 0.591875897 0 0 3.15E-13 MTMR11 HSAD011135 0.605864402 0 0 8.65E-14 PPA1 HSAD021998 0.619422455 0 0 3.80E-13 RAB24 HSAD038212 0.621805281 0 0 2.66E-15 OXCT1 HSAD027890 1.552242291 0 0 8.12E-14 CEACAM21 HSAD021036 1.576476579 0 0 1.18E-13 RAB13 HSAD008255 1.612306212 0 0 2.78E-15 UACA HSAD002947 1.614641594 0 0 1.91E-14 SLC2A9 HSAD027852 1.627518747 0 0 9.20E-14 MRVI1 HSAD000047 1.699630246 0 0 1.54E-14 C5AR1 HSAD033733 1.73766184 0 0 3.33E-15 PLD3 HSAD005563 1.82907652 0 0 1.38E-13 SLC25A25 HSAD011990 1.872036449 0 0 1.87E-13 C3AR1 HSAD012264 1.886575261 0 0 6.53E-14 SIRPD HSAD039322 2.22864313 0 0 1.93E-13 MRC1L1 HSAD014973 2.982177737 0 0 1.96E-13 GPR34 HSAD021104 3.078466773 0 0 0 TCN2 HSAD005548 3.105555431 0 0 0 ADAMTS2 HSAD034232 3.155367014 0 0 4.46E-13 GPER HSAD008218 3.429808393 0 0 0 HTRA1 HSAD037876 3.720894991 0 0 0 ADAMTS2 HSAD023017 15.90449965 0 0 0 VSIG4 HSAD010189 25.3072126 0 0 0 IL1R2 HSAD021559 27.67465985 0 0 0 IL1R2 HSAD001671 0.455324613 1.00E-04 2.58E-06 9.12E-13 EMR1 HSAD006308 0.59026634 1.00E-04 2.58E-06 1.18E-12 TNFSF13B HSAD001906 0.614336351 1.00E-04 2.58E-06 5.39E-13 TNFSF10

115 HSAD014434 0.661198477 1.00E-04 2.58E-06 8.43E-13 MXD3 HSAD000393 0.671234506 1.00E-04 2.58E-06 7.01E-13 AL049778.1 HSAD029622 2.573596388 1.00E-04 2.58E-06 6.20E-13 LHFPL2 HSAD005114 2.638835236 1.00E-04 2.58E-06 1.31E-12 SHMT1 HSAD014496 4.008532785 1.00E-04 2.58E-06 7.25E-13 AMPH HSAD004899 4.369828609 1.00E-04 2.58E-06 7.87E-13 ALOX15B HSAD009557 0.581392275 2.00E-04 4.96E-06 1.71E-12 HK2 HSAD033186 0.602921867 2.00E-04 4.96E-06 1.83E-12 FAM134B HSAD005650 0.298275068 3.00E-04 7.30E-06 2.67E-12 LGALS2 HSAD023366 0.276487455 4.00E-04 9.37E-06 4.43E-12 GBP1 HSAD002112 0.5606343 4.00E-04 9.37E-06 4.27E-12 ADAM8 HSAD009156 0.552715338 5.00E-04 1.15E-05 6.33E-12 ERAP2 HSAD014113 0.486636033 6.00E-04 1.35E-05 7.01E-12 ARID5B HSAD011522 0.4877897 7.00E-04 1.43E-05 1.04E-11 EPSTI1 HSAD000473 0.564509369 7.00E-04 1.43E-05 8.83E-12 NFE2L3 HSAD006073 0.574271853 7.00E-04 1.43E-05 1.50E-11 HSAD002960 0.650660996 7.00E-04 1.43E-05 1.10E-11 SYTL3 HSAD002662 2.333257189 7.00E-04 1.43E-05 1.49E-11 MARVELD1 HSAD007885 4.523250837 7.00E-04 1.43E-05 1.16E-11 AMPH HSAD037001 1.621388594 0.001 2.01E-05 2.02E-11 ADAMTS2 HSAD000423 0.610025732 0.0012 2.32E-05 2.29E-11 CD55 HSAD033316 1.717105315 0.0012 2.32E-05 2.46E-11 RNASE6 HSAD007491 0.427713041 0.0013 2.32E-05 3.76E-11 MX1 HSAD016444 0.494446971 0.0013 2.32E-05 3.77E-11 PSME2 HSAD005171 0.546104682 0.0013 2.32E-05 3.67E-11 STEAP4 HSAD024184 0.595450047 0.0013 2.32E-05 3.44E-11 TCTEX1D2 HSAD029308 0.643040214 0.0013 2.32E-05 3.60E-11 BATF3 HSAD015825 1.704852338 0.0013 2.32E-05 3.14E-11 TMEM106A HSAD031104 0.275032066 0.0015 2.60E-05 5.02E-11 CXCL10 HSAD020993 1.925639019 0.0015 2.60E-05 5.23E-11 BCAT1 HSAD034215 0.670009515 0.0016 2.66E-05 6.06E-11 PTGS1 HSAD017478 0.68478032 0.0016 2.66E-05 5.92E-11 RAB24 HSAD026637 1.70650153 0.0016 2.66E-05 6.05E-11 C3orf78 HSAD005462 0.626271528 0.0017 2.76E-05 6.38E-11 CD55 HSAD012018 1.583701625 0.0017 2.76E-05 6.47E-11 NUDT16 HSAD010054 0.295327272 0.0019 2.93E-05 6.90E-11 GBP1 HSAD034492 0.368308841 0.0019 2.93E-05 6.87E-11 WARS HSAD014942 0.600270458 0.0019 2.93E-05 6.78E-11 TMEM8B HSAD001512 1.900184721 0.0019 2.93E-05 6.68E-11 IL18 HSAD005336 0.59372381 0.0021 3.09E-05 8.61E-11 CCL2 HSAD025288 1.524557963 0.0021 3.09E-05 9.17E-11 EPS8 HSAD038010 1.740076095 0.0021 3.09E-05 8.25E-11 C21orf7 HSAD017515 0.32719147 0.0023 3.23E-05 1.25E-10 OAS3 HSAD013624 0.53437824 0.0023 3.23E-05 1.35E-10 HEG1 HSAD012774 0.680718956 0.0023 3.23E-05 9.94E-11 FAM122C HSAD028345 1.649485163 0.0023 3.23E-05 1.25E-10 UACA HSAD009573 0.578673213 0.0024 3.30E-05 1.56E-10 RABGAP1L HSAD028960 1.716888368 0.0026 3.54E-05 1.78E-10 LIPA HSAD018227 1.637796785 0.0028 3.77E-05 2.11E-10 RNASE4 HSAD020029 1.587089703 0.0032 4.26E-05 2.69E-10 GPX1 HSAD016096 0.452139335 0.004 5.11E-05 4.07E-10 SERPING1 HSAD019654 0.655194323 0.004 5.11E-05 4.12E-10 PSME1 HSAD025050 0.669240695 0.0041 5.13E-05 4.74E-10 AC104297.1 HSAD040777 0.702678867 0.0041 5.13E-05 4.52E-10 SYTL3

116 HSAD007418 0.298407514 0.0043 5.33E-05 5.23E-10 CHST2 HSAD014414 1.553627503 0.0047 5.72E-05 6.24E-10 RNF141 HSAD024195 0.715684306 0.0048 5.73E-05 6.43E-10 FGFRL1 HSAD015603 2.049285694 0.0048 5.73E-05 6.33E-10 METTL7A HSAD038430 0.410106117 0.0051 5.92E-05 7.64E-10 STAT1 HSAD012893 0.53804729 0.0051 5.92E-05 7.39E-10 PSME2 HSAD020751 1.647543442 0.0051 5.92E-05 7.33E-10 TOP1MT HSAD009107 1.81492536 0.0055 6.27E-05 1.02E-09 FOXH1 HSAD004525 0.696223238 0.0056 6.32E-05 1.08E-09 ADHFE1 HSAD019684 1.505905136 0.0059 6.55E-05 1.17E-09 NET1 HSAD025933 0.544974747 0.006 6.60E-05 1.22E-09 SH2D1B HSAD002717 0.434624057 0.0061 6.65E-05 1.28E-09 OLFM1 HSAD006367 0.564008309 0.0066 7.13E-05 1.42E-09 MOV10 HSAD014767 0.462666281 0.0069 7.27E-05 1.52E-09 XAF1 HSAD001075 0.564954963 0.0069 7.27E-05 1.52E-09 CD44 HSAD040207 0.545779322 0.0074 7.61E-05 1.64E-09 AKT3 HSAD032464 0.580417735 0.0074 7.61E-05 1.79E-09 NA HSAD007358 1.651776467 0.0074 7.61E-05 1.65E-09 FN1 HSAD033597 0.664037265 0.0076 7.73E-05 1.85E-09 C20orf111 HSAD026985 0.347084583 0.0077 7.73E-05 1.91E-09 USP18 HSAD013391 0.494386962 0.0081 7.94E-05 2.02E-09 SERPING1 HSAD012545 0.582366054 0.0081 7.94E-05 2.12E-09 MRAS HSAD016872 0.673071481 0.0081 7.94E-05 2.10E-09 GOLM1 HSAD027857 1.535949143 0.0082 7.98E-05 2.24E-09 MYO1E HSAD020221 1.534346216 0.0084 8.11E-05 2.36E-09 C1orf115 HSAD012346 0.65992011 0.0085 8.14E-05 2.43E-09 CD97 HSAD031466 0.662342206 0.0097 9.22E-05 3.14E-09 GPR27 HSAD023174 1.509480616 0.0105 9.84E-05 3.68E-09 HSAD010650 2.569347034 0.0112 0.000104148 4.19E-09 TMOD1 HSAD030234 1.649137341 0.0118 0.000108926 4.63E-09 DEXI HSAD017484 0.374372726 0.013 0.000118335 5.80E-09 SLAMF7 HSAD006096 0.598005496 0.013 0.000118335 5.77E-09 NID1 HSAD028177 0.684716917 0.013 0.000118335 5.74E-09 CD55 HSAD031977 1.725777249 0.0134 0.000121084 6.28E-09 AC093627.10 HSAD021458 0.433050469 0.0137 0.000122043 6.59E-09 DUSP5 HSAD020122 1.555889561 0.0137 0.000122043 6.59E-09 KCTD12 HSAD040770 0.577484306 0.014 0.000122134 6.93E-09 ADA HSAD036439 3.097707298 0.014 0.000122134 6.77E-09 RNASE1 HSAD003337 1.515248587 0.0152 0.000132487 8.45E-09 HMOX1 HSAD011298 1.654899381 0.0154 0.000133301 8.52E-09 MMP19 HSAD039080 1.777359801 0.0167 0.00014344 9.55E-09 MAFB HSAD033383 0.436641605 0.017 0.000144053 1.01E-08 PLA2G7 HSAD038700 1.692224897 0.017 0.000144053 1.02E-08 PIK3IP1 HSAD035948 0.504574293 0.0172 0.000144769 1.07E-08 RABGAP1L HSAD030879 0.646776659 0.0175 0.000146303 1.13E-08 INSIG1 HSAD005939 0.438747666 0.0178 0.000146863 1.16E-08 LY6E HSAD002849 0.375265197 0.019 0.000155464 1.24E-08 CMPK2 HSAD026289 0.559353521 0.0191 0.000155464 1.25E-08 EPSTI1 HSAD011810 1.527351454 0.0208 0.000168087 1.42E-08 RAPH1 HSAD019200 0.701982681 0.0211 0.000169416 1.53E-08 ZG16B HSAD014509 1.546065036 0.0214 0.000170728 1.56E-08 C3orf78 HSAD035752 0.622697369 0.0216 0.000170962 1.58E-08 AIG1 HSAD035600 0.533274768 0.0217 0.000170962 1.58E-08 SCO2 HSAD017273 0.333885831 0.0224 0.000175344 1.67E-08 IFIT1

117 HSAD003448 3.066164331 0.0239 0.000185842 1.89E-08 HSAD024304 0.66750967 0.0241 0.000187015 1.92E-08 PSPH HSAD000664 0.607679785 0.0244 0.000188174 2.04E-08 RHOC HSAD005780 0.550491342 0.0251 0.00019082 2.14E-08 HSAD032985 1.948304567 0.0251 0.00019082 2.15E-08 AL355922.3 HSAD027197 0.488688096 0.026 0.000196393 2.28E-08 RP11-792D21.2 HSAD003881 0.629698881 0.0284 0.000211263 2.84E-08 EVL HSAD004298 1.860771323 0.0284 0.000211263 2.83E-08 AHR HSAD013271 0.726501352 0.0292 0.000215352 2.94E-08 SLC44A1 HSAD022191 1.522301853 0.0293 0.000215352 2.94E-08 NPL HSAD027266 1.626749906 0.0297 0.000217003 3.02E-08 CD1E HSAD019382 0.58649609 0.0311 0.00022578 3.23E-08 SFMBT2 HSAD015590 0.491985303 0.0321 0.000229745 3.48E-08 BPI HSAD007399 1.593487595 0.0321 0.000229745 3.41E-08 SEZ6L HSAD034534 0.603060776 0.0333 0.00023686 3.75E-08 TTC7B HSAD032221 0.690095758 0.0342 0.000240491 4.00E-08 TUBB3 HSAD039868 2.222283425 0.0342 0.000240491 3.92E-08 NFIA HSAD026711 0.630194741 0.035 0.000244683 4.22E-08 RP11-379B18.2 HSAD007715 1.599484456 0.0357 0.000250196 4.59E-08 MYO7A HSAD037246 1.556052066 0.036 0.000250225 4.69E-08 ANG HSAD008016 0.531881273 0.0373 0.000257614 4.92E-08 IFI27 HSAD003255 1.555746722 0.0374 0.000257614 4.97E-08 PPARG HSAD039716 0.697468103 0.0408 0.000278753 5.84E-08 SMPDL3A HSAD038063 1.543201645 0.0409 0.000278753 5.87E-08 PNOC HSAD011878 0.681483877 0.0419 0.000283072 6.18E-08 ITGAX HSAD027056 0.727243496 0.043 0.000287127 6.36E-08 ACER3 HSAD012357 1.766144481 0.043 0.000287127 6.36E-08 TEX2 HSAD009357 0.382273474 0.0432 0.000287127 6.40E-08 GK HSAD008487 0.392615343 0.0437 0.00028888 6.52E-08 GBP5 HSAD020921 1.716020953 0.0464 0.000303771 6.95E-08 CDKN2B HSAD025849 0.743791273 0.0465 0.000303771 7.01E-08 CRY1 HSAD022762 0.465143199 0.0483 0.000313634 7.37E-08 GBP4

Supplementary Table 3: Differentially expressed genes between T24 and T0 of CD14+monocytes of GC-Responders, after FWER multiple testing correction and a cut-off of 1.5 up- or down-regulation

Supplementary Table 4: reporterId Fold change / Ratio pval.FWER pval.BH pval.ttest geneSymbol HSAD024737 0.18 0 0 0 GK HSAD037946 0.18 0 0 0 RSAD2 HSAD010770 0.19 0 0 0 IFITM2 HSAD040888 0.20 0 0 0 CDKN1C HSAD038203 0.20 0 0 0 GK HSAD009357 0.21 0 0 0 GK HSAD009269 0.22 0 0 0 HERC5 HSAD026985 0.22 0 0 0 USP18 HSAD002849 0.22 0 0 0 CMPK2 HSAD017273 0.22 0 0 0 IFIT1 HSAD017515 0.22 0 0 0 OAS3 HSAD024773 0.23 0 0 0 IFIT1P1 HSAD035570 0.23 0 0 0 ANKRD22 HSAD033155 0.24 0 0 0 ARID5B HSAD031741 0.25 0 0 0 IFIT2 HSAD010054 0.26 0 0 0 GBP1

118 HSAD023366 0.26 0 0 0 GBP1 HSAD031104 0.27 0 0 0 CXCL10 HSAD021898 0.28 2.00E-04 1.17E-06 3.10E-13 SEL1L3 HSAD015250 0.28 9.00E-04 3.72E-06 1.35E-10 RSAD2 HSAD007835 0.28 1.00E-04 7.00E-07 2.16E-14 ISG15 HSAD028888 0.29 0 0 0 IFI44L HSAD007418 0.29 0 0 0 CHST2 HSAD032558 0.29 1.00E-04 7.00E-07 8.33E-15 SOCS3 HSAD016096 0.29 0 0 0 SERPING1 HSAD033383 0.30 0 0 0 PLA2G7 HSAD033860 0.30 0 0 0 IFIT3 HSAD007491 0.30 0 0 0 MX1 HSAD039566 0.31 0 0 0 IFIT2 HSAD016448 0.32 0 0 0 PLSCR1 HSAD031199 0.32 0 0 0 PLAC8 HSAD014767 0.32 0 0 0 XAF1 HSAD031902 0.32 0 0 0 XAF1 HSAD025239 0.33 0 0 0 SAMD9L HSAD029251 0.33 0 0 7.77E-16 BPI HSAD013391 0.33 0 0 0 SERPING1 HSAD001824 0.34 0 0 0 STAT1 HSAD017484 0.34 0 0 0 SLAMF7 HSAD005939 0.34 0 0 0 LY6E HSAD008487 0.35 0 0 0 GBP5 HSAD012881 0.35 3.00E-04 1.74E-06 1.13E-12 USP18 HSAD027197 0.35 0 0 4.44E-16 RP11-792D21.2 HSAD022657 0.36 0.0024 8.53E-06 8.24E-10 BCL2A1 HSAD005650 0.36 9.00E-04 3.72E-06 1.22E-10 LGALS2 HSAD034492 0.36 0 0 0 WARS HSAD011522 0.37 0 0 0 EPSTI1 HSAD038430 0.37 0 0 0 STAT1 HSAD005054 0.37 0 0 0 DDX58 HSAD001026 0.38 0 0 0 MX2 HSAD013968 0.39 0 0 0 DDX60 HSAD018475 0.39 0 0 0 PLAC8 HSAD035143 0.39 0 0 0 OASL HSAD022520 0.39 0 0 0 MX2 HSAD030034 0.40 0.0118 3.13E-05 2.22E-08 PIM1 HSAD037515 0.40 2.00E-04 1.17E-06 1.29E-13 ST20 HSAD006902 0.40 0 0 0 AL035670.1 HSAD025527 0.41 0 0 0 OAS2 HSAD036349 0.41 0.0089 2.55E-05 1.14E-08 PID1 HSAD038088 0.41 0 0 0 XAF1 HSAD004683 0.41 0 0 0 CLEC5A HSAD031014 0.41 0.0044 1.44E-05 2.69E-09 HESX1 HSAD010299 0.41 0 0 0 OAS1 HSAD011565 0.41 0 0 0 RPH3A HSAD029454 0.42 0 0 0 DDX60 HSAD016074 0.42 0 0 0 LAP3 HSAD002717 0.42 2.00E-04 1.17E-06 4.45E-13 OLFM1 HSAD022762 0.43 4.00E-04 1.97E-06 1.55E-12 GBP4 HSAD026289 0.43 0 0 0 EPSTI1 HSAD022456 0.43 0 0 0 IRF7 HSAD029678 0.43 4.00E-04 1.97E-06 3.38E-12 OAS2

119 HSAD006072 0.44 0 0 0 SAMD9 HSAD031089 0.44 0 0 0 HSAD026206 0.44 0 0 0 IFI44 HSAD004864 0.44 4.00E-04 1.97E-06 1.01E-11 NA HSAD019220 0.45 0.0062 1.89E-05 6.35E-09 LIMCH1 HSAD027624 0.45 0.0012 4.55E-06 3.24E-10 SIGLEC1 HSAD004329 0.46 1.00E-04 7.00E-07 4.11E-14 HIF1A HSAD005171 0.46 0 0 0 STEAP4 HSAD005462 0.46 0.0109 2.97E-05 1.73E-08 CD55 HSAD014113 0.46 0 0 0 ARID5B HSAD010849 0.46 5.00E-04 2.44E-06 1.72E-11 EIF2AK2 HSAD005033 0.46 0.0055 1.70E-05 5.16E-09 MCTP2 HSAD016444 0.46 0 0 0 PSME2 HSAD023053 0.46 0 0 0 IFIH1 HSAD032686 0.47 0 0 0 SAMD9 HSAD018365 0.47 2.00E-04 1.17E-06 2.67E-13 BHLHE40 HSAD017316 0.47 0 0 0 PARP9 HSAD040549 0.47 0.0012 4.55E-06 2.34E-10 CLEC4D HSAD021580 0.48 0 0 0 LITAF HSAD036816 0.48 2.00E-04 1.17E-06 7.49E-13 BHLHE40 HSAD014030 0.48 0 0 0 XPO6 HSAD027979 0.48 0 0 0 PARP14 HSAD006958 0.49 0 0 0 AIM2 HSAD006308 0.49 0 0 4.44E-16 TNFSF13B HSAD024048 0.49 0.0077 2.27E-05 8.39E-09 HERC6 HSAD002112 0.50 2.00E-04 1.17E-06 1.36E-13 ADAM8 HSAD000791 0.50 0 0 0 RTP4 HSAD028846 0.50 6.00E-04 2.86E-06 2.45E-11 CR1 HSAD028964 0.50 0.0299 6.89E-05 1.22E-07 HLA-DQA2 HSAD012893 0.50 0 0 0 PSME2 HSAD033376 0.50 9.00E-04 3.72E-06 6.26E-11 TMEM123 HSAD006500 0.50 1.00E-04 7.00E-07 2.08E-14 LMNB1 HSAD027252 0.50 1.00E-04 7.00E-07 3.51E-14 PARP14 HSAD025174 0.51 0.001 3.96E-06 1.40E-10 PRKCA HSAD027541 0.51 1.00E-04 7.00E-07 7.24E-14 FAM129A HSAD008970 0.51 0 0 0 SHISA5 HSAD035943 0.51 0 0 0 BLVRA HSAD006107 0.51 0.0265 6.28E-05 1.00E-07 OAS3 HSAD009123 0.51 8.00E-04 3.57E-06 3.02E-11 HIF1A HSAD000397 0.51 0.0061 1.86E-05 6.13E-09 HLA-DQA1 HSAD022618 0.51 0 0 0 SAMD9 HSAD013624 0.51 1.00E-04 7.00E-07 1.09E-14 HEG1 HSAD005742 0.51 0.0105 2.88E-05 1.61E-08 SPATS2L HSAD022754 0.51 0.0029 1.01E-05 1.13E-09 GALNT3 HSAD013737 0.51 0 0 0 ADARB1 HSAD021327 0.51 0 0 0 APOL6 HSAD001671 0.51 0 0 0 EMR1 HSAD006103 0.51 0 0 0 STEAP4 HSAD028177 0.52 0.001 3.96E-06 1.71E-10 CD55 HSAD001906 0.52 0 0 0 TNFSF10 HSAD035237 0.52 0.0035 1.19E-05 1.70E-09 VNN2 HSAD010850 0.52 0.001 3.96E-06 1.56E-10 EMR3 HSAD033775 0.52 4.00E-04 1.97E-06 2.71E-12 MT2A HSAD015631 0.53 0 0 0 C5orf56

120 HSAD020140 0.53 9.00E-04 3.72E-06 6.68E-11 EIF2AK2 HSAD009573 0.53 0 0 0 RABGAP1L HSAD024002 0.53 0 0 0 TNNT1 HSAD002463 0.53 2.00E-04 1.17E-06 1.88E-13 TOR1B HSAD012413 0.53 2.00E-04 1.17E-06 1.02E-12 MT2A HSAD021381 0.53 1.00E-04 7.00E-07 1.73E-14 AL109948.1 HSAD003083 0.53 0.0053 1.69E-05 4.25E-09 HERC6 HSAD031024 0.53 0.0013 4.87E-06 4.07E-10 LIMK2 HSAD035276 0.54 2.00E-04 1.17E-06 4.00E-13 DHX58 HSAD039650 0.54 2.00E-04 1.17E-06 2.64E-13 RABGAP1L HSAD033385 0.54 8.00E-04 3.57E-06 3.12E-11 HSAD007509 0.54 0 0 3.33E-16 CLEC5A HSAD012248 0.54 1.00E-04 7.00E-07 2.04E-14 MT2A HSAD040207 0.54 0 0 0 AKT3 HSAD006073 0.54 0.0041 1.36E-05 2.15E-09 HSAD028174 0.54 1.00E-04 7.00E-07 5.11E-14 CDKN2D HSAD034534 0.54 0 0 0 TTC7B HSAD029340 0.54 0 0 7.77E-16 TAP1 HSAD038971 0.55 0.001 3.96E-06 1.75E-10 IFITM3 HSAD008946 0.55 9.00E-04 3.72E-06 8.50E-11 SOD2 HSAD035600 0.55 0 0 0 SCO2 HSAD025893 0.55 0.0246 5.88E-05 8.78E-08 NEURL HSAD038758 0.56 0 0 0 GBP3 HSAD027784 0.56 0 0 6.66E-16 NA HSAD034571 0.56 0 0 0 AC122108.1 HSAD040287 0.56 0.0105 2.88E-05 1.58E-08 AC025280.1 HSAD022378 0.57 0.0027 9.55E-06 9.43E-10 C5orf32 HSAD026711 0.57 4.00E-04 1.97E-06 4.99E-12 RP11-379B18.2 HSAD012259 0.57 0 0 0 CHMP5 HSAD013181 0.57 0 0 1.78E-15 IL1RN HSAD027318 0.57 0.0017 6.27E-06 5.67E-10 RAB37 HSAD006598 0.57 4.00E-04 1.97E-06 3.38E-12 XRN1 HSAD036555 0.57 0.0064 1.94E-05 6.54E-09 TSPAN2 HSAD001075 0.57 4.00E-04 1.97E-06 5.98E-12 CD44 HSAD027910 0.57 4.00E-04 1.97E-06 3.02E-12 TRANK1 HSAD022465 0.57 0 0 1.11E-16 SP110 HSAD008462 0.58 0 0 0 DTX3L HSAD034011 0.58 0.0055 1.70E-05 4.72E-09 RNF213 HSAD005780 0.58 2.00E-04 1.17E-06 3.38E-13 HSAD000751 0.58 8.00E-04 3.57E-06 4.54E-11 PARP12 HSAD030133 0.58 0.0064 1.94E-05 6.53E-09 KLHDC7B HSAD002639 0.58 0.0184 4.59E-05 5.04E-08 CRIP1 HSAD032209 0.58 4.00E-04 1.97E-06 8.38E-12 PXK HSAD027992 0.58 4.00E-04 1.97E-06 4.22E-12 IFIT5 HSAD018698 0.58 9.00E-04 3.72E-06 7.41E-11 RP1-193M11.1 HSAD002298 0.58 4.00E-04 1.97E-06 1.40E-12 PIM3 HSAD020632 0.58 0.0419 9.03E-05 2.24E-07 NT5C3 HSAD024138 0.59 0 0 0 GBP2 HSAD037903 0.59 0.0041 1.36E-05 2.11E-09 SLITRK2 HSAD019382 0.59 0.0028 9.82E-06 1.02E-09 SFMBT2 HSAD009776 0.59 0 0 0 GCH1 HSAD040734 0.59 0 0 3.33E-16 JAK3 HSAD018674 0.59 0 0 1.44E-15 PML HSAD006367 0.59 4.00E-04 1.97E-06 1.54E-11 MOV10

121 HSAD012490 0.59 2.00E-04 1.17E-06 1.12E-13 ST6GAL1 HSAD036741 0.59 9.00E-04 3.72E-06 7.20E-11 CD36 HSAD015931 0.59 2.00E-04 1.17E-06 6.75E-13 MEI1 HSAD037351 0.59 0.0052 1.67E-05 4.06E-09 CR1L HSAD019671 0.59 0 0 0 TMTC1 HSAD006271 0.59 9.00E-04 3.72E-06 7.58E-11 ARID5B HSAD001782 0.59 0.0034 1.16E-05 1.61E-09 CCR5 HSAD030269 0.60 8.00E-04 3.57E-06 4.77E-11 OPTN HSAD000473 0.60 0 0 0 NFE2L3 HSAD024840 0.60 0.0411 8.92E-05 2.14E-07 AQP9 HSAD022451 0.60 4.00E-04 1.97E-06 7.23E-12 OLFM1 HSAD018604 0.60 4.00E-04 1.97E-06 1.17E-11 RAB11FIP4 HSAD017288 0.60 0.0087 2.51E-05 1.07E-08 IFI6 HSAD007997 0.60 4.00E-04 1.97E-06 2.28E-12 SAMD9 HSAD036699 0.60 1.00E-04 7.00E-07 3.39E-14 MPZL2 HSAD013332 0.60 0 0 0 NMI HSAD009022 0.61 0.0018 6.59E-06 5.82E-10 RELB HSAD004139 0.61 0 0 2.22E-16 HPSE HSAD003523 0.61 0 0 0 HSAD002621 0.61 9.00E-04 3.72E-06 8.88E-11 FAM117A HSAD038844 0.61 1.00E-04 7.00E-07 7.05E-14 GCH1 HSAD001850 0.61 0.0286 6.67E-05 1.14E-07 GK2 HSAD015168 0.61 2.00E-04 1.17E-06 9.98E-14 FAM8A1 HSAD002183 0.61 0 0 2.22E-16 HSH2D HSAD032464 0.61 0.02 4.96E-05 5.61E-08 NA HSAD018492 0.61 0.0341 7.64E-05 1.58E-07 RP4-781L3.1 HSAD005744 0.61 0.0354 7.85E-05 1.64E-07 USP41 HSAD001886 0.61 0 0 1.11E-16 DDX60L HSAD040883 0.62 0 0 2.22E-16 HPSE HSAD006096 0.62 0.0071 2.13E-05 7.21E-09 NID1 HSAD025236 0.62 0 0 0 FAM129A HSAD039791 0.62 0.0055 1.70E-05 4.56E-09 NRP1 HSAD006863 0.62 2.00E-04 1.17E-06 1.52E-13 FYN HSAD026283 0.62 2.00E-04 1.17E-06 1.06E-13 CD36 HSAD012613 0.62 0.037 8.10E-05 1.79E-07 RNF213 HSAD038212 0.62 0 0 0 OXCT1 HSAD010351 0.62 0.0048 1.56E-05 2.99E-09 NUAK2 HSAD026667 0.62 0.0098 2.73E-05 1.40E-08 MTMR11 HSAD030252 0.63 0 0 0 TRIM22 HSAD030766 0.63 4.00E-04 1.97E-06 2.73E-12 HSAD014942 0.63 2.00E-04 1.17E-06 2.26E-13 TMEM8B HSAD017456 0.63 0 0 0 CELF2 HSAD008470 0.63 1.00E-04 7.00E-07 1.44E-14 PCGF5 HSAD011841 0.63 8.00E-04 3.57E-06 3.53E-11 APOL6 HSAD014168 0.63 0 0 0 VRK2 HSAD016417 0.63 2.00E-04 1.17E-06 1.27E-13 ADAR HSAD012593 0.64 0.0011 4.29E-06 2.13E-10 RP5-903G2.1 HSAD024068 0.64 4.00E-04 1.97E-06 3.61E-12 GBP3 HSAD028075 0.64 0 0 5.55E-16 DTX3L HSAD004165 0.64 0.0028 9.82E-06 1.04E-09 SYNE1 HSAD029308 0.64 0.0055 1.70E-05 4.66E-09 BATF3 HSAD026746 0.64 2.00E-04 1.17E-06 1.94E-13 AFG3L1 HSAD019654 0.64 1.00E-04 7.00E-07 6.38E-14 PSME1 HSAD005067 0.64 0 0 1.67E-15 CD99

122 HSAD000814 0.64 0.0298 6.89E-05 1.21E-07 POU2AF1 HSAD011252 0.64 0.0031 1.07E-05 1.35E-09 APCDD1 HSAD011296 0.65 0 0 1.11E-16 IL23A HSAD039716 0.65 0.0475 0.000101905 2.67E-07 SMPDL3A HSAD024275 0.65 4.00E-04 1.97E-06 2.44E-12 AP001011.3 HSAD020366 0.65 0.001 3.96E-06 1.74E-10 HIVEP2 HSAD032335 0.65 0.0215 5.21E-05 6.69E-08 IFI35 HSAD035752 0.65 2.00E-04 1.17E-06 7.65E-13 AIG1 HSAD017818 0.65 0.0039 1.31E-05 1.98E-09 ATP10A HSAD002661 0.65 0.0106 2.90E-05 1.66E-08 GK3P DNAJC25- HSAD004891 0.65 0.0095 2.68E-05 1.28E-08 GNG10 HSAD031601 0.65 0.0206 5.05E-05 6.05E-08 STMN4 HSAD034856 0.65 0 0 0 PER2 HSAD004039 0.65 4.00E-04 1.97E-06 6.17E-12 PCGF5 HSAD010689 0.65 0.0096 2.70E-05 1.34E-08 MCTP2 HSAD021176 0.65 2.00E-04 1.17E-06 7.65E-13 RP11-397P13.6 HSAD032325 0.65 0 0 5.55E-16 IFI16 HSAD009591 0.66 0.0362 7.96E-05 1.72E-07 CD55 HSAD024184 0.66 0 0 1.11E-16 TCTEX1D2 HSAD011242 0.66 4.00E-04 1.97E-06 1.53E-12 TMTC1 HSAD031208 0.66 0.001 3.96E-06 1.52E-10 CDS2 HSAD008180 0.66 0.0223 5.35E-05 7.19E-08 KLHL5 HSAD021603 0.66 0.0345 7.70E-05 1.60E-07 EIF4E3 HSAD010377 0.66 0.0269 6.35E-05 1.07E-07 ITM2B HSAD012106 0.66 4.00E-04 1.97E-06 9.54E-12 P2RY14 HSAD030349 0.66 0 0 2.22E-16 RP11-397P13.6 HSAD014646 0.66 0.0012 4.55E-06 2.80E-10 FAM110B HSAD033171 0.66 0.011 2.99E-05 1.93E-08 PHF11 HSAD011878 0.67 9.00E-04 3.72E-06 8.86E-11 ITGAX HSAD000788 0.67 4.00E-04 1.97E-06 5.46E-12 PTPRC HSAD012913 0.67 0.0091 2.59E-05 1.19E-08 EMR2 HSAD005741 0.67 0.0339 7.62E-05 1.56E-07 GPR137B HSAD003625 0.67 0.0055 1.70E-05 4.75E-09 PDE4B HSAD025970 0.67 0.0026 9.22E-06 9.01E-10 WIPI1 HSAD020021 0.67 9.00E-04 3.72E-06 8.33E-11 HSAD036078 0.67 4.00E-04 1.97E-06 1.57E-11 ATXN1 HSAD016978 0.67 8.00E-04 3.57E-06 4.67E-11 SCO2 HSAD035199 0.67 0.0215 5.21E-05 6.71E-08 FNDC3B HSAD031593 0.67 0.0011 4.29E-06 1.87E-10 RP11-198M15.1 HSAD031130 0.67 0.0021 7.58E-06 6.93E-10 CLMN HSAD004772 0.67 0.0041 1.36E-05 2.40E-09 MTHFD2 HSAD007445 0.67 0.0041 1.36E-05 2.22E-09 IFIT5 HSAD037754 0.67 4.00E-04 1.97E-06 2.73E-12 IFI16 HSAD011135 0.67 0.0138 3.59E-05 2.93E-08 PPA1 HSAD024304 0.68 0.005 1.61E-05 3.75E-09 PSPH HSAD006796 0.68 0.0022 7.90E-06 7.39E-10 LGALS9B HSAD014133 0.68 0.0083 2.41E-05 9.73E-09 CASP7 HSAD033962 0.68 8.00E-04 3.57E-06 4.06E-11 HSAD037504 0.68 0.0075 2.22E-05 8.04E-09 LYSMD2 HSAD027056 0.68 0 0 1.33E-15 ACER3 HSAD003330 0.68 0.0154 3.93E-05 3.61E-08 KIAA1958 HSAD002960 0.68 0.0489 0.000104288 2.79E-07 SYTL3 HSAD012770 0.68 0.015 3.86E-05 3.29E-08 TRIM5 HSAD028781 0.68 0.0196 4.86E-05 5.44E-08 AC074093.1

123 HSAD026751 0.68 0.0016 5.98E-06 4.77E-10 MTHFS HSAD020740 0.68 9.00E-04 3.72E-06 9.87E-11 CRY1 HSAD029671 0.68 0.0413 8.94E-05 2.17E-07 ITGAX HSAD036422 0.68 6.00E-04 2.86E-06 1.99E-11 AC003682.1 HSAD009361 0.69 0.0044 1.44E-05 2.72E-09 DNAJA1 HSAD032356 0.69 0.0112 3.03E-05 1.98E-08 SLAMF8 HSAD028538 0.69 0.0302 6.96E-05 1.26E-07 UBE2L6 HSAD006842 0.69 2.00E-04 1.17E-06 1.09E-12 RBM43 HSAD035335 0.69 4.00E-04 1.97E-06 1.57E-11 FAM92A1 HSAD023689 0.69 0.0021 7.58E-06 6.94E-10 APOBEC3G HSAD038414 0.69 2.00E-04 1.17E-06 1.85E-13 UBE2Z HSAD027676 0.69 4.00E-04 1.97E-06 1.03E-11 SPSB1 HSAD001923 0.69 9.00E-04 3.72E-06 9.22E-11 CD52 HSAD034215 0.69 0 0 1.11E-16 PTGS1 HSAD004326 0.69 0.0116 3.10E-05 2.12E-08 AC091046.1 HSAD036457 0.69 0 0 1.11E-16 FYB HSAD003881 0.69 0.0351 7.82E-05 1.62E-07 EVL HSAD033564 0.69 0.015 3.86E-05 3.34E-08 CD226 HSAD034103 0.69 0.0061 1.86E-05 5.86E-09 ZMIZ1 HSAD040777 0.69 0.022 5.30E-05 7.10E-08 SYTL3 HSAD025755 0.69 0.0074 2.21E-05 7.85E-09 HSAD036821 0.69 1.00E-04 7.00E-07 7.18E-14 ATG3 HSAD004862 0.69 0.0114 3.06E-05 2.04E-08 TNFRSF10C HSAD017420 0.69 0.0187 4.66E-05 5.09E-08 WSB1 HSAD011153 0.70 0.0055 1.70E-05 4.95E-09 TDRD7 HSAD004323 0.70 0.0263 6.25E-05 9.94E-08 HSAD008004 0.70 0.001 3.96E-06 1.53E-10 STAT2 HSAD039531 0.70 0.0053 1.69E-05 4.18E-09 MCTP2 HSAD004534 0.70 0.0055 1.70E-05 4.93E-09 GTPBP1 HSAD031020 0.70 0.0105 2.88E-05 1.64E-08 TBC1D8 HSAD029289 0.70 0 0 2.89E-15 TRIM22 HSAD025050 0.70 0.0121 3.20E-05 2.27E-08 AC104297.1 HSAD032221 0.70 4.00E-04 1.97E-06 3.28E-12 TUBB3 HSAD020107 0.70 0.027 6.35E-05 1.08E-07 KMO HSAD014434 0.70 0.0242 5.80E-05 8.36E-08 MXD3 HSAD038174 0.70 8.00E-04 3.57E-06 3.86E-11 HSAD026761 0.70 0.0031 1.07E-05 1.34E-09 IL2RG HSAD020893 0.70 0.0023 8.22E-06 8.06E-10 ACAT2 HSAD002134 0.70 0.0378 8.27E-05 1.84E-07 H3F3C HSAD029911 0.70 8.00E-04 3.57E-06 4.07E-11 HSAD019712 0.70 0.0017 6.27E-06 5.62E-10 RP11-274B18.2 HSAD017870 0.70 0.0177 4.42E-05 4.63E-08 LYZ HSAD008147 0.71 0.0155 3.95E-05 3.73E-08 APOL6 HSAD024756 0.71 0.001 3.96E-06 1.53E-10 CCDC109B HSAD001282 0.71 0.0012 4.55E-06 3.62E-10 AC046176.3 HSAD036307 0.71 0.0022 7.90E-06 7.56E-10 USP53 HSAD034562 0.71 0.0035 1.19E-05 1.65E-09 ALDH6A1 HSAD035021 0.71 1.00E-04 7.00E-07 2.65E-14 C12orf23 HSAD003415 0.71 0.0168 4.23E-05 4.38E-08 HSAD000207 0.71 0.0075 2.22E-05 8.27E-09 MYO10 HSAD030768 0.71 9.00E-04 3.72E-06 1.34E-10 FCGR2A HSAD031556 0.71 0.0012 4.55E-06 3.70E-10 VWA5A HSAD012199 0.71 0.0028 9.82E-06 9.59E-10 PSME1 HSAD031009 0.71 0.0038 1.28E-05 1.90E-09 ASAP2

124 HSAD023635 0.72 6.00E-04 2.86E-06 1.76E-11 FAM8A1 HSAD023779 0.72 0.0177 4.42E-05 4.59E-08 CYSLTR1 HSAD031429 0.72 0.0241 5.80E-05 8.28E-08 C16orf57 HSAD015629 0.72 0.024 5.73E-05 8.06E-08 PHF21A HSAD004313 0.72 6.00E-04 2.86E-06 2.07E-11 NRG1 HSAD030873 0.72 0.0101 2.80E-05 1.52E-08 HSAD035928 0.72 0.0012 4.55E-06 3.05E-10 SLC22A4 HSAD004079 0.72 0.0043 1.42E-05 2.49E-09 SEC14L1 HSAD017458 0.72 0.0012 4.55E-06 2.48E-10 CD47 HSAD002491 0.72 0.0206 5.05E-05 6.07E-08 HSAD021753 0.72 0.0358 7.92E-05 1.68E-07 RP11-351O1.2 HSAD030442 0.72 2.00E-04 1.17E-06 6.16E-13 PTPRC HSAD017478 0.72 2.00E-04 1.17E-06 8.12E-13 RAB24 HSAD003314 0.72 0.037 8.11E-05 1.79E-07 BNIP2 HSAD032538 0.72 0.0012 4.55E-06 2.81E-10 KIAA0226 HSAD026893 0.72 0.0023 8.22E-06 7.77E-10 USP6NL HSAD005030 0.72 9.00E-04 3.72E-06 8.92E-11 KLHL6 HSAD020896 0.72 0.0106 2.90E-05 1.67E-08 PSMB9 HSAD022162 0.72 0.0055 1.70E-05 4.94E-09 RGS10 HSAD034385 0.72 0.0105 2.88E-05 1.64E-08 HSAD026696 0.72 0.0153 3.92E-05 3.47E-08 CNIH HSAD018399 0.72 0.0038 1.28E-05 1.85E-09 CD99 HSAD022934 0.72 0.001 3.96E-06 1.73E-10 R3HDM2 HSAD001213 0.73 0.0017 6.27E-06 5.12E-10 ADAMTSL3 HSAD000513 0.73 0.0177 4.42E-05 4.60E-08 CASP8 HSAD033667 0.73 0.0223 5.35E-05 7.25E-08 TESK1 HSAD029999 0.73 0.0378 8.27E-05 1.84E-07 LIMS1 HSAD033597 0.73 0.0339 7.62E-05 1.55E-07 C20orf111 HSAD031560 0.73 0.0034 1.16E-05 1.62E-09 C6orf192 HSAD005164 0.73 0.0265 6.28E-05 1.00E-07 SIK3 HSAD007573 0.73 0.036 7.95E-05 1.69E-07 DEK HSAD033874 0.73 0.0033 1.14E-05 1.53E-09 PLAGL2 HSAD031067 0.73 9.00E-04 3.72E-06 7.43E-11 STK24 HSAD038061 0.73 9.00E-04 3.72E-06 1.05E-10 NT5C2 HSAD010107 0.73 0.0173 4.35E-05 4.52E-08 TSR1 HSAD020771 0.73 0.0075 2.22E-05 8.17E-09 AC011479.1 HSAD023787 0.73 6.00E-04 2.86E-06 2.12E-11 SIK3 HSAD018879 0.73 0.0093 2.64E-05 1.26E-08 STK24 HSAD034876 0.73 5.00E-04 2.44E-06 1.70E-11 TRAM1 HSAD018496 0.73 0.0282 6.59E-05 1.14E-07 GBP4 HSAD006338 0.74 0.0049 1.59E-05 3.47E-09 GNB4 HSAD033864 0.74 0.0203 5.02E-05 5.75E-08 DDR1 HSAD014231 0.74 0.0115 3.08E-05 2.09E-08 HSAD019177 0.74 0.0021 7.58E-06 7.00E-10 C1GALT1 HSAD022817 0.74 0.0098 2.73E-05 1.44E-08 IL3RA HSAD026476 0.74 0.0142 3.68E-05 3.05E-08 MAK HSAD003142 0.74 4.00E-04 1.97E-06 7.20E-12 DTNBP1 HSAD016949 0.74 9.00E-04 3.72E-06 6.17E-11 ACAT2 HSAD010366 0.74 0.0134 3.49E-05 2.81E-08 ANKRD13A HSAD015614 0.74 0.0034 1.16E-05 1.55E-09 BRP44L HSAD006431 0.74 0.0228 5.45E-05 7.53E-08 CD48 HSAD028447 0.74 0.031 7.09E-05 1.31E-07 C7orf60 HSAD024878 0.74 0.0104 2.88E-05 1.55E-08 RASSF5 HSAD022253 0.74 0.0041 1.36E-05 2.44E-09 FYB

125 HSAD036483 0.74 0.0088 2.53E-05 1.12E-08 SEL1L3 HSAD015515 0.74 0.0129 3.38E-05 2.74E-08 SERPINB9 HSAD020304 0.75 0.0068 2.05E-05 6.82E-09 HSAD028492 0.75 0.0354 7.85E-05 1.63E-07 EIF4E3 HSAD027488 0.75 0.0017 6.27E-06 5.06E-10 DHTKD1 HSAD011304 0.75 0.0313 7.15E-05 1.34E-07 CXCL9 HSAD002403 0.75 0.0198 4.90E-05 5.48E-08 ACER3 HSAD003446 0.75 0.0259 6.17E-05 9.66E-08 NR3C1 HSAD036644 1.50 0.023 5.49E-05 7.64E-08 C17orf55 HSAD021723 1.51 0 0 5.55E-16 MEST HSAD014652 1.52 4.00E-04 1.97E-06 1.39E-12 CABC1 HSAD032521 1.52 0 0 7.77E-16 HSAD004849 1.52 0.0079 2.31E-05 8.88E-09 TMEM39A HSAD033316 1.53 2.00E-04 1.17E-06 3.43E-13 RNASE6 HSAD032413 1.53 1.00E-04 7.00E-07 3.53E-14 IL13RA1 HSAD022136 1.53 1.00E-04 7.00E-07 1.90E-14 GEMIN4 HSAD030804 1.53 9.00E-04 3.72E-06 7.19E-11 CCND2 HSAD000735 1.54 0 0 0 AC023055.2 HSAD002941 1.54 0.0097 2.73E-05 1.34E-08 APEX1 HSAD026533 1.54 0.0163 4.12E-05 4.05E-08 CLEC3B HSAD011310 1.55 0 0 3.11E-15 RP11-474P12.3 HSAD017233 1.55 0.0037 1.26E-05 1.80E-09 RBP7 HSAD034125 1.55 0 0 0 C13orf31 HSAD021036 1.56 0 0 0 RAB13 HSAD013694 1.56 0.001 3.96E-06 1.68E-10 DIP2B HSAD040767 1.56 4.00E-04 1.97E-06 9.42E-12 IL13RA1 HSAD000437 1.57 2.00E-04 1.17E-06 5.19E-13 OLIG1 HSAD028345 1.57 2.00E-04 1.17E-06 6.66E-13 UACA HSAD031977 1.57 0.0073 2.18E-05 7.67E-09 AC093627.10 HSAD000682 1.57 0.0265 6.29E-05 1.01E-07 HSAD037001 1.57 4.00E-04 1.97E-06 7.94E-12 ADAMTS2 HSAD020029 1.58 0 0 0 GPX1 HSAD008255 1.58 6.00E-04 2.86E-06 1.98E-11 UACA HSAD032939 1.60 8.00E-04 3.57E-06 2.96E-11 ARMC4 HSAD037246 1.61 8.00E-04 3.57E-06 4.44E-11 ANG HSAD007715 1.61 0 0 3.11E-15 MYO7A HSAD007456 1.62 1.00E-04 7.00E-07 2.60E-14 APEX1 HSAD027857 1.62 2.00E-04 1.17E-06 7.78E-13 MYO1E HSAD021730 1.63 0.001 3.96E-06 1.42E-10 MAT2A HSAD012197 1.64 0.0055 1.70E-05 4.94E-09 GAMT HSAD034689 1.65 0 0 0 IL13RA1 HSAD003586 1.65 1.00E-04 7.00E-07 1.55E-14 LDHB HSAD007399 1.65 1.00E-04 7.00E-07 4.61E-14 SEZ6L HSAD011298 1.66 8.00E-04 3.57E-06 4.84E-11 MMP19 HSAD034409 1.70 2.00E-04 1.17E-06 1.14E-13 MFGE8 HSAD014347 1.71 0.0299 6.90E-05 1.23E-07 AC138472.5 HSAD021080 1.71 1.00E-04 7.00E-07 7.15E-14 MAT2A HSAD040907 1.71 4.00E-04 1.97E-06 4.17E-12 HSAD023416 1.72 0 0 0 P2RY12 HSAD024518 1.73 0 0 0 AMY1C HSAD018696 1.73 0 0 0 STS HSAD008117 1.75 4.00E-04 1.97E-06 9.44E-12 MAF HSAD004298 1.75 0 0 0 AHR HSAD018227 1.75 0 0 0 RNASE4

126 HSAD015603 1.76 2.00E-04 1.17E-06 1.88E-13 METTL7A HSAD039692 1.78 0 0 0 C20orf27 HSAD009482 1.87 0.0055 1.70E-05 4.63E-09 SESN1 HSAD033733 1.89 0 0 0 PLD3 HSAD023553 1.89 0.0013 4.87E-06 3.79E-10 KCNMA1 HSAD039868 1.89 0 0 3.33E-16 NFIA HSAD019236 1.89 0 0 0 METTL7A HSAD009107 1.90 0 0 0 FOXH1 HSAD010650 1.93 0.0041 1.36E-05 2.40E-09 TMOD1 HSAD039652 1.94 0 0 0 CPM HSAD020751 1.94 0 0 0 TOP1MT HSAD001512 1.95 0 0 0 IL18 HSAD008490 2.02 0.0413 8.94E-05 2.19E-07 TXNIP HSAD030136 2.02 0 0 0 MAF HSAD008798 2.03 0 0 0 CPM HSAD001704 2.05 0.015 3.86E-05 3.25E-08 SESN1 HSAD039180 2.06 0 0 0 OLFML2A HSAD028960 2.09 0 0 0 LIPA HSAD011921 2.14 0.0061 1.86E-05 6.25E-09 OLAH HSAD029622 2.28 0 0 0 LHFPL2 HSAD014418 2.30 0.0077 2.27E-05 8.36E-09 C5orf30 HSAD002662 2.36 0 0 0 MARVELD1 HSAD034232 2.37 0 0 0 GPER HSAD021281 2.43 0 0 0 RNASE1 HSAD005114 2.47 0 0 0 SHMT1 HSAD007358 2.64 0.0277 6.50E-05 1.10E-07 FN1 HSAD018237 2.64 0.0031 1.07E-05 1.36E-09 ADORA3 HSAD014973 2.70 0 0 0 GPR34 HSAD005548 2.70 0 0 1.11E-16 ADAMTS2 HSAD005982 2.75 0 0 0 FLT3 HSAD021104 2.95 0 0 0 TCN2 HSAD008218 3.03 0 0 0 HTRA1 HSAD014496 3.17 2.00E-04 1.17E-06 1.95E-13 AMPH HSAD004899 3.22 0 0 0 ALOX15B HSAD028845 3.31 0 0 4.44E-16 TMEM45A HSAD037876 3.45 0 0 0 ADAMTS2 HSAD007885 3.83 0 0 0 AMPH HSAD017642 3.97 0 0 0 TPST1 HSAD036439 5.16 0 0 0 RNASE1 HSAD009260 5.22 0 0 0 LPL HSAD023017 11.14 0 0 0 VSIG4 HSAD010189 11.55 0 0 0 IL1R2 HSAD021559 12.25 0 0 0 IL1R2

Supplementary Table 4: Differentially expressed genes between T24 and T0 of CD14+monocytes of non-Responders, after FWER multiple testing correction and a cut-off of 1.5 up- or down-regulation

Supplementary Table 5: reporterId Fold change / Ratio pval.FWER pval.BH pval.ttest geneSymbol HSAD028102 4.43 0 0 0 S100A12 HSAD019991 4.41 0.0012 0.000122129 2.23488E-13 PROK2 HSAD001600 3.83 0 0 0 MMP8 HSAD031645 3.17 0.0053 0.00029422 1.64944E-11 PLBD1

127 HSAD022035 2.53 0.006 0.000305323 2.82655E-11 PADI4 HSAD040348 2.17 0.0022 0.000167928 1.62448E-12 RETN HSAD020442 2.11 0.0477 0.001204194 9.77894E-09 CACNA1I HSAD038429 2.08 0.0016 0.000150313 6.20282E-13 HSAD029694 2.05 0.001 0.000111027 1.9873E-14 S100A9 HSAD019276 2.01 0.0247 0.00087933 1.80499E-09 HSAD027135 1.95 0.0029 0.000208338 3.98659E-12 S100A8 HSAD009997 1.92 0.0263 0.000894514 2.11344E-09 SNX29P2 HSAD015498 1.78 0.0001 1.7447E-05 0 ANXA3 HSAD026532 1.75 0.0049 0.000284968 1.40505E-11 PPAPDC2 HSAD025911 1.68 0.0001 1.7447E-05 0 NEAT1 HSAD031521 1.66 0.0022 0.000167928 1.801E-12 HSAD009404 1.66 0.0166 0.000679853 5.09453E-10 CIRBP HSAD022281 1.65 0.0214 0.000828189 1.17901E-09 JAKMIP1 HSAD017701 1.64 0.0005 7.63308E-05 1.22125E-15 HSAD006198 1.62 0.0049 0.000284968 1.33594E-11 RP11-617D20.1 HSAD004617 1.61 0.0299 0.000951982 2.98489E-09 CBLL1 HSAD040511 1.60 0.0266 0.000894514 2.17754E-09 SLC4A7 HSAD027523 1.60 0.0355 0.001052638 4.54625E-09 CXCR6 HSAD038632 1.60 0.001 0.000111027 3.9857E-14 TAGAP HSAD036039 1.59 0.008 0.000389874 6.05924E-11 HSAD000103 1.56 0.0315 0.000980087 3.27101E-09 CD97 HSAD000422 1.56 0.0124 0.000522208 2.19037E-10 C11orf21 HSAD026890 1.54 0.0285 0.000932039 2.73417E-09 C17orf48 HSAD037810 1.51 0.0039 0.000264613 7.14728E-12 KIAA0240 HSAD023882 1.51 0.0431 0.001204194 7.07909E-09 HSAD024193 0.73 0.0472 0.001204194 9.51043E-09 PCOLCE2 HSAD019026 0.70 0.0481 0.001204194 1.01369E-08 SEH1L HSAD006611 0.69 0.0328 0.000997885 3.71098E-09 NR4A2 HSAD018278 0.69 0.044 0.001204194 7.9804E-09 FAM26F HSAD008462 0.66 0.0056 0.000297358 1.88636E-11 DTX3L HSAD033136 0.65 0.024 0.00087933 1.56634E-09 HNRNPA1 HSAD018123 0.65 0.0242 0.00087933 1.6346E-09 LAMP3 HSAD017655 0.61 0.0482 0.001204194 1.02648E-08 JUNB HSAD026350 0.59 0.0083 0.000389874 7.27431E-11 LDLR HSAD016267 0.55 0.0001 1.7447E-05 0 AIF1 HSAD022317 0.55 0.0022 0.000167928 2.03182E-12 NDFIP2 HSAD008962 0.53 0.0008 0.000108559 3.44169E-15 SECTM1 HSAD005876 0.47 0.0001 1.7447E-05 0 MPO HSAD038430 0.45 0.0364 0.00105372 4.80123E-09 STAT1 HSAD033365 0.43 0.0049 0.000284968 1.36057E-11 JUNB HSAD032558 0.29 0.0001 1.7447E-05 0 SOCS3

Supplementary Table 5: Differentially expressed genes between T24 and T0 of CD4+T-cells of GC-Responders, after FWER multiple testing correction and a cut-off of 1.5 up- or down-regulation

128 Supplementary Table 6: reporterId Fold change / Ratio pval.FWER pval.BH pval.ttest geneSymbol HSAD001600 4.94 0 0 0 MMP8 HSAD028102 4.10 6.00E-04 9.52E-06 1.17E-11 S100A12 HSAD031645 4.07 3.00E-04 5.68E-06 5.72E-13 PLBD1 HSAD027135 3.70 0.0206 0.000118408 5.49E-08 S100A8 HSAD029694 3.68 0.0284 0.000153923 9.48E-08 S100A9 HSAD034977 3.31 7.00E-04 1.01E-05 1.57E-11 OLFM4 HSAD028713 3.14 0.0404 0.000203603 1.83E-07 CLEC4E HSAD001676 2.98 0.0145 9.41E-05 2.94E-08 MMP9 HSAD015498 2.98 0.0037 3.28E-05 1.65E-09 ANXA3 HSAD008296 2.89 0.0037 3.28E-05 1.61E-09 LCN2 HSAD032151 2.81 0.0166 0.000100474 3.74E-08 LTF HSAD030998 2.56 0 0 4.44E-16 TCN1 HSAD015473 2.46 0.0034 3.17E-05 1.34E-09 OLFM4 HSAD022035 2.34 0.0018 1.94E-05 2.85E-10 PADI4 HSAD003135 2.27 0.0319 0.000170129 1.18E-07 DDIT4 HSAD040348 2.21 7.00E-04 1.01E-05 1.54E-11 RETN HSAD020877 2.21 0.0037 3.28E-05 1.70E-09 RGS18 HSAD010686 2.15 0.0021 2.24E-05 4.41E-10 HSAD001704 2.09 0 0 0 SESN1 HSAD009482 2.01 0 0 0 SESN1 HSAD029229 2.00 0.0128 8.46E-05 2.29E-08 CRISP3 HSAD000362 1.96 0.0105 7.21E-05 1.54E-08 DHRS9 HSAD020442 1.95 0 0 0 CACNA1I HSAD017701 1.88 0.0037 3.28E-05 1.66E-09 HSAD007381 1.85 0 0 6.77E-15 HSAD008490 1.80 1.00E-04 2.27E-06 4.97E-14 TXNIP HSAD007077 1.79 0 0 0 VIPR1 HSAD015409 1.75 0.0033 3.13E-05 1.08E-09 MYO1B HSAD032190 1.73 0.0278 0.000153923 8.68E-08 CLK1 HSAD022120 1.70 0 0 0 FAM13AOS HSAD038762 1.68 0.0161 9.89E-05 3.48E-08 RP11-356I2.4 HSAD029511 1.67 8.00E-04 1.05E-05 3.24E-11 TPM2 HSAD033664 1.67 6.00E-04 9.52E-06 7.92E-12 C14orf132 HSAD037492 1.65 0.0161 9.89E-05 3.48E-08 HIST1H2AC HSAD016421 1.64 0.0051 4.20E-05 3.26E-09 TSC22D3 HSAD011960 1.64 0.0496 0.000235562 2.61E-07 PGLYRP1 HSAD031299 1.64 0 0 0 HSAD020937 1.64 6.00E-04 9.52E-06 8.34E-12 RP11-420G6.4 HSAD000422 1.62 0.0018 1.94E-05 2.56E-10 C11orf21 HSAD036164 1.61 0 0 6.66E-16 ANKRD55 HSAD001969 1.59 7.00E-04 1.01E-05 1.31E-11 RP4-717I23.3 HSAD036720 1.59 8.00E-04 1.05E-05 2.07E-11 NLRP1 HSAD037967 1.58 0 0 1.11E-16 KCNMB4 HSAD024518 1.56 0.0012 1.51E-05 5.52E-11 AMY1C HSAD035118 1.55 0 0 0 BTG1 HSAD014470 1.54 7.00E-04 1.01E-05 1.37E-11 NRBP2 HSAD026562 1.53 0 0 7.11E-15 RP4-717I23.3 HSAD004525 1.53 4.00E-04 7.05E-06 3.57E-12 ADHFE1 HSAD026967 1.52 0.0039 3.39E-05 1.91E-09 GOLGA8G HSAD013882 1.52 0.0071 5.33E-05 6.94E-09 GPR37 HSAD034122 1.51 0 0 8.88E-16 OGT HSAD022494 1.51 3.00E-04 5.68E-06 7.63E-13 YPEL3

129 HSAD025126 1.50 0.0033 3.13E-05 1.04E-09 KRT72 HSAD012253 1.50 0.0059 4.68E-05 4.67E-09 PBX4 HSAD020996 1.50 0.0101 7.08E-05 1.44E-08 AC019322.4 HSAD025440 1.50 0.0024 2.48E-05 5.45E-10 AC100756.4 HSAD010563 0.75 0.0013 1.51E-05 7.51E-11 C3orf26 HSAD040621 0.75 0.0319 0.000170129 1.18E-07 DYRK2 HSAD025403 0.75 0.0449 0.000221087 2.20E-07 MRPL24 HSAD028523 0.75 0.0013 1.51E-05 9.44E-11 EDA HSAD004062 0.75 0.0045 3.81E-05 2.53E-09 TAPBP HSAD038266 0.75 0.0066 5.14E-05 6.05E-09 MAP2 HSAD013075 0.75 0.0033 3.13E-05 1.18E-09 STAT3 HSAD017544 0.75 0.0317 0.000170129 1.16E-07 FBXO36 HSAD016267 0.74 0.0353 0.000183087 1.36E-07 AIF1 HSAD018282 0.74 0.0033 3.13E-05 1.08E-09 IRF9 HSAD030193 0.74 5.00E-04 8.58E-06 7.27E-12 BYSL HSAD025342 0.74 8.00E-04 1.05E-05 2.20E-11 METTL1 HSAD019026 0.74 8.00E-04 1.05E-05 2.48E-11 SEH1L HSAD034413 0.74 0.0084 6.09E-05 9.20E-09 NOLC1 HSAD004287 0.74 2.00E-04 4.30E-06 4.64E-13 BOLA3 HSAD018021 0.74 0.0036 3.28E-05 1.51E-09 CLUAP1 HSAD033732 0.74 0.0012 1.51E-05 7.03E-11 PARP11 HSAD030204 0.74 0.0051 4.20E-05 3.33E-09 BST2 HSAD006796 0.73 0.0076 5.61E-05 8.08E-09 LGALS9B HSAD019882 0.73 6.00E-04 9.52E-06 1.01E-11 IFITM3 HSAD014023 0.73 4.00E-04 7.05E-06 1.45E-12 BOLA3 HSAD037947 0.73 4.00E-04 7.05E-06 1.41E-12 LZTS1 HSAD029678 0.73 0.0159 9.89E-05 3.29E-08 OAS2 HSAD024011 0.73 0.0125 8.31E-05 2.22E-08 TMSL4 HSAD009214 0.73 0.0013 1.51E-05 1.24E-10 GPATCH4 HSAD029376 0.73 0.0466 0.000225891 2.35E-07 CALR HSAD017741 0.73 0.0356 0.000183871 1.41E-07 DDX21 HSAD036890 0.73 0.0083 6.09E-05 8.82E-09 NOP16 HSAD034700 0.73 0.0059 4.68E-05 4.36E-09 LTB HSAD004965 0.72 0.0017 1.91E-05 1.57E-10 IFITM8P HSAD007622 0.72 1.00E-04 2.27E-06 1.16E-13 SLC29A2 HSAD009942 0.72 0.016 9.89E-05 3.40E-08 FAM100B HSAD031237 0.72 0.0184 0.000108227 4.52E-08 HSAD005208 0.72 0.0284 0.000153923 9.36E-08 C10orf2 HSAD023053 0.71 2.00E-04 4.30E-06 4.52E-13 IFIH1 HSAD025301 0.71 0.0017 1.91E-05 1.88E-10 HSAD002298 0.71 0.0039 3.39E-05 1.82E-09 PIM3 HSAD022025 0.71 0.0459 0.000223332 2.27E-07 ST6GALNAC1 HSAD019757 0.71 0.0119 7.95E-05 1.89E-08 C6orf105 HSAD028075 0.71 3.00E-04 5.68E-06 1.09E-12 DTX3L HSAD010054 0.71 6.00E-04 9.52E-06 8.69E-12 GBP1 HSAD026668 0.71 0.0024 2.48E-05 6.14E-10 NEFL HSAD015961 0.70 0.0202 0.000116654 5.41E-08 PPIL1 HSAD019654 0.70 1.00E-04 2.27E-06 1.67E-13 PSME1 HSAD035143 0.70 0.0067 5.19E-05 6.30E-09 OASL HSAD017111 0.70 0.0238 0.000134185 6.82E-08 RP11-381O7.3 HSAD015771 0.69 1.00E-04 2.27E-06 5.78E-14 DYRK2 HSAD022520 0.69 0.0032 3.12E-05 9.06E-10 MX2 HSAD016074 0.69 0.0014 1.62E-05 1.38E-10 LAP3 HSAD012072 0.69 0.0092 6.60E-05 1.07E-08 ST8SIA1

130 HSAD023366 0.69 1.00E-04 2.27E-06 1.51E-13 GBP1 HSAD001671 0.69 0.0373 0.000191675 1.50E-07 EMR1 HSAD030737 0.69 0.0142 9.31E-05 2.77E-08 PTMA HSAD001026 0.68 0.0027 2.70E-05 7.51E-10 MX2 HSAD001023 0.68 1.00E-04 2.27E-06 1.24E-14 PIM2 HSAD017539 0.68 0.0252 0.000142334 7.48E-08 MTHFD1L HSAD010795 0.68 0 0 5.88E-15 TMSB4X HSAD012199 0.67 3.00E-04 5.68E-06 1.02E-12 PSME1 HSAD010499 0.67 0 0 0 SNTB1 HSAD010226 0.67 3.00E-04 5.68E-06 7.99E-13 SUSD4 HSAD013847 0.67 0 0 1.22E-15 SH3BP5 HSAD025239 0.66 0 0 2.22E-16 SAMD9L HSAD007295 0.66 0.0013 1.51E-05 1.13E-10 CHI3L2 HSAD029257 0.66 0.0027 2.70E-05 7.43E-10 HSAD038971 0.65 0 0 0 IFITM3 HSAD010959 0.65 0 0 0 RNF144A HSAD031199 0.64 0 0 3.77E-15 PLAC8 HSAD040734 0.64 3.00E-04 5.68E-06 7.23E-13 JAK3 HSAD018475 0.64 0 0 0 PLAC8 HSAD004464 0.64 0 0 1.55E-15 CNP HSAD026206 0.63 0.0069 5.25E-05 6.55E-09 IFI44 HSAD023609 0.62 0 0 0 DNTT HSAD010294 0.62 0.0349 0.000183087 1.33E-07 HSAD026289 0.61 0.0141 9.29E-05 2.62E-08 EPSTI1 HSAD011522 0.61 0 0 0 EPSTI1 HSAD040287 0.61 0 0 0 AC025280.1 HSAD022456 0.61 2.00E-04 4.30E-06 3.75E-13 IRF7 HSAD016444 0.60 0 0 0 PSME2 HSAD012893 0.60 0 0 0 PSME2 HSAD011031 0.60 0 0 0 NDFIP2 HSAD020140 0.59 0.0013 1.51E-05 8.88E-11 EIF2AK2 HSAD024618 0.59 0 0 0 IFITM4P HSAD023134 0.59 1.00E-04 2.27E-06 5.50E-14 SUSD4 HSAD026350 0.58 0 0 2.22E-16 LDLR HSAD029659 0.57 0 0 0 SPON1 HSAD007835 0.57 0.017 0.000101915 4.02E-08 ISG15 HSAD038088 0.56 0.0191 0.000111806 4.92E-08 XAF1 HSAD022317 0.55 0 0 0 NDFIP2 HSAD009022 0.55 0 0 0 RELB HSAD029454 0.54 0 0 0 DDX60 HSAD017515 0.53 0 0 8.88E-16 OAS3 HSAD007491 0.53 0 0 0 MX1 HSAD010770 0.53 0 0 0 IFITM2 HSAD034204 0.53 0.0108 7.37E-05 1.62E-08 CST3 HSAD010299 0.53 1.00E-04 2.27E-06 1.38E-13 OAS1 HSAD024773 0.53 0.0029 2.85E-05 8.36E-10 IFIT1P1 HSAD008462 0.52 0 0 0 DTX3L HSAD017316 0.51 0 0 0 PARP9 HSAD013968 0.51 0 0 0 DDX60 HSAD026985 0.49 7.00E-04 1.01E-05 1.54E-11 USP18 HSAD017273 0.48 0.0034 3.17E-05 1.25E-09 IFIT1 HSAD031902 0.47 0 0 1.22E-15 XAF1 HSAD002849 0.45 8.00E-04 1.05E-05 2.95E-11 CMPK2 HSAD014767 0.45 7.00E-04 1.01E-05 1.49E-11 XAF1

131 HSAD028888 0.45 0 0 0 IFI44L HSAD018123 0.43 0 0 3.33E-16 LAMP3 HSAD001824 0.42 0 0 0 STAT1 HSAD038430 0.40 0 0 0 STAT1 HSAD037946 0.38 0 0 0 RSAD2 HSAD033860 0.33 0.0018 1.94E-05 2.53E-10 IFIT3 HSAD032558 0.26 8.00E-04 1.05E-05 2.68E-11 SOCS3

Supplementary Table 6: Differentially expressed genes between T24 and T0 of CD4+T-cells of Non-Responders, after FWER multiple testing correction and a cut-off of 1.5 up- or down-regulation

Supplementary Figure 1:

Supplementary Figure 1. Common pathway involvement of IFI44L, RSAD2, STAT1, IRF7, ISG15 and IFIH1 (=MDA5). A KEGG pathway search clustered ISG15, IRF7 and MDA5 significantly in the RIG-1 (Type I IFN/response to virus) pathway (light green boxes and red stars). Using GeneRIF annotations, STAT1, RSAD2 and IFI44L were directly linked to the pathway as being type I IFN response genes (Light blue boxes, red stars).

Supplementary text: cells and is shown in supplemental Figure An alternative method to find differences 2A. in the working mechanism of GC in GC- qPCR validation was performed Responders and non-Responders is the (supplemental Figure 2C) and confirmed direct comparison of differentially significantly increased expression of expressed genes between patient groups ERAP2 in GC-Responders (Fold change after 24 hours of GC-therapy at T24. (FC) GC-Responders vs. non-Responders: Genes with p<0.05 using the false 4.9, p<0.01) and CEP350, a centrosomal discovery rate correction (FDR) were protein (FC: 1.8; p=0.05). A trend was considered statistically significantly seen for DICER, a ribonuclease involved different. in RNA interference (FC: 1.5; p= 0.06). This approach yielded a list of 15 probes, The differential expression of LNPEP, representing 14 different genes in CD14+ MYCBP2 and LIPA in monocytes seen in

132 the microarray results could not be was confirmed for ERAP2 (FC GC- confirmed by qPCR (not shown). Responders vs. non-Responders: 7.5, The comparison of differentially expressed p<0.01). Whereas decreased expression for genes in CD4+cells between GC- cyclin B2 (CCNB2; ratio: 0.49; p< 0.01) Responders and non-Responders at T24 was highly significant, decreased yielded a list of 20 probes, representing 16 expression of TMP4 in GC-Responders different genes, as shown in supplemental was marginal (ratio: 0.62; p=0.059). Only a Figure 2B. qPCR validation is shown in trend was seen towards lower expression supplemental Figure 2D. Significantly of ADA in GC-Responders (ratio: 0.7; increased expression in GC-Responders p=0.08).

Supplementary Figure 2A:

133

Figure legend Supplementary Figure 2: Differentially expressed genes in CD14+monocytes and CD4+T-cells from Glucocorticoid-Responders versus non-Responders at T24: Heat map of gene expression in CD14+monocytes (supp. Figure 2A) and CD4+T-cells (supp. Figure 2B) of individual patients. Transcript levels are expressed as median centered ratios across all samples. Scale bar (log2 ratio): increased (red), decreased (green) or identical (black) ratio. Validation by qPCR for selected genes is shown for CD14+monocytes (supp. Figure 2C) and CD4+T-cells (supp. Figure 2D). CT (CT gene of interest – CT GAPDH) levels are given as quantification of gene expression with higher CT corresponding to a lower gene expression. P<0.05 (using unpaired Students T-test) was considered statistically significant. *: p<0.05; **: p<0.01.

134

Chapter 9

Summary and General Discussion

135

Summary subsets and assessing their phenotype after rounds of stimulation and periods of rest. Whereas the CCR7-/CD27- T-cells retained Memory T-cell subsets in health and their phenotype, the other subsets matured disease along the proposed maturational pathway. Measurement of the telomere length further substantiated this differentiation pathway, as Immunological memory is a key feature of this marker of cell senescence significantly our immune system and relies on a functional decreased steadily from naïve  T-cell memory compartment. If this system is CCR7+/CD27+  CCR7-/CD27+  CCR7- unhinged, several pathological conditions can /CD27-. The decrease in telomere length was ensue, including immunodeficiencies, paralleled by a decrease in inducible allergies, and autoimmune diseases. It is thus telomerase activity, the enzyme responsible of eminent importance to decipher the for counteracting telomere length attrition, maturational steps of T-cells in order to from naïve to CCR7+ to CCR7- subsets. influence an aberrant differentiation leading Also the proliferative capacity, a further to a specific disease state. The characteristic of cell maturation, which characterization of the maturational pathway declines as cells undergo senescence, of CD4+ memory T-cells in health and (auto- decreased along the suggested pathway. immune) disease was the scope of the first Interestingly, the CCR7- subsets were prone part of my thesis and is described in to undergo apoptosis upon stimulation, which Chapters 2 and 3. was hardly seen in naïve cells or CCR7+ In Chapter 2 we expanded the knowledge of cells. Examining the function of theses the developmental stages of CD4 T-cells by subsets in terms of cytokine secretion, we employing two cell surface markers, namely found a clear pattern in the CD4+ T-cell CCR7 and CD27. CCR7 is a chemokine subsets. The signature cytokines for Th1 and receptor, which orchestrates the homing of Th2 cells were increased in the CCR7- cells to secondary lymphoid organs. Its /CD27+ and CCR7-/DC27- subset expression divides memory T-cells into respectively. This is consistent with the CCR7+ central memory (TCM) and CCR7- conclusion that there is an association of effector memory (TEM). (1) TCM are thought “effector” cytokine secretion with the loss of to recirculate to secondary lymphoid organs CCR7 expression. However, the TEM using CCR7 , whereas TEM do not migrate to population, as defined by CCR7- only, was lymphoid organs, but instead migrate to not homogeneous, and the subdivision into inflammatory sites and exert their function CCR7-/CD27+ and CCR7-/CD27- cells by cytokine secretion. (figure 1, revealed different cytokine secretion patterns. introduction). By combining CCR7 with As maturation stages also differ in terms of CD27, we could further specify this trafficking potential, the chemokine suggested differential pathway as depicted in expression of the defined subsets was figure 2. According to our findings, CD4 T- examined and revealed a decrease in CXCR4 cells mature from naïve to the CCR7+/ and CXCR5 expression in the CCR7- CD27+ stage (accounting for 69.8±2.1% of subsets. As both these chemokine receptors CD4+CD45RO+ T-cells), further to the serve to attract T-cells into lymphoid organs, CCR7-/CD27+ (21.5±1.8% of their decreased expression in the terminally CD4+CD45RO+ T-cells) before they also differentiated subsets is compatible with our lose surface expression of CD27 and become proposed model. (2,3) Altogether, the data CCR7-/CD27- (5.6±0.6% of presented in Chapter 2 lead to a more CD4+CD45RO+ T-cells). We could confirm precise delineation of the maturational stages this pathway by culturing these different of CD4+ T-cells forming the basis of

136 evaluation of a potential aberrant maturation cells are known. (7) However, not much in diseases In SLE, the imbalance between thought had been given to the overall different functional T-cell subsets had been functioning of T-cells in SLE, reflected by the focus of a considerable amount of their differentiation status. Therefore, in studies, showing alternately a skewing to Chapter 3 using knowledge of Chapter 2 Th1, Th2 or Th17 cell subsets (4-6); we examined the maturational pathway in furthermore, intrinsic abnormalities of T- SLE-patients. .

Figure 1: The maturational pathway of CD4+ T-cells. After encounter with their antigen, naive T-cells become memory T-cells, which is marked by the replacement of CD45RA by CD45RO on the cell surface. The cell surface markers CCR7 and CD27 are gradually lost, as the cells mature through additional stimulation. The cells are depicted with their preferential secreted cytokine.

The pathway delineated in healthy controls controls (allergy- or RA-patients). Secondly, (HC; Figure 1) was also seen in SLE patients, examining the cell cycle status of the which was evident from cell culture and cell different subsets, a higher percentage of cycle experiments, measurement of telomere CCR7- cells from SLE patients were in the length and telomerase as described for G2/M phase ex vivo, and were more prone to healthy controls in Chapter 2. However, undergo apoptosis after stimulation. The some crucial differences were seen in SLE latter was mirrored by a significantly reduced patients. Firstly, the distribution of the proliferative capacity and survival of CCR7- subsets differed significantly between SLE /CD27- T-cells derived from SLE patients. patients and HC or disease controls. A clear Thirdly, the mean inducible telomerase increase of CCR7- subsets was observed in activity was significantly decreased in the SLE. (CCR7+/CD27+: SLE: 54.1±2.8% , SLE naive CD4+ and CCR7+/CD27+ HC: 69.8±2.1%); CCR7-/CD27+ cells: SLE: memory T cells compared with those from 29.1±2.0%, HC: 21.5±1.8%; CCR7-/CD27-: normal donors (p<0.02), whereas the SLE 13.9±2.3%, HC:5.6±0.6%; p<0.00003; telomere length between subsets of SLE p<0.007 and p<0.002, respectively). This patients and HC did not differ. This reduced difference did not seem to be influenced by telomerase activity was not due to a different disease activity or medication and appears proportion of recent thymic emigrants in the SLE specific, as it was not seen in disease CCR7+ T-cell departments, as assessed by

137 CD31 expression. Lastly, the examination of especially the humoral response to H1 chemokine receptors showed a decreased correlates with SLE disease activity (9) and expression of CXCR4 in all SLE derived appears to be more specific for SLE than Abs subsets, and an increase of CXCR5 in the to core histones.(9, 10) Histones were CCR7-/CD27- T-cells of active SLE patients identified as the major T-cell antigens (Ags) compared to HC. Thus, in Chapter 3, we in SLE and histone-specific T-cell clones verified in SLE-patients the maturational augmented the production of anti-dsDNA in pathway proposed in Chapter 2 for healthy mice and men. (11,12) However, details of individuals and could identify several the histone-specific cellular response in differences in distribution, cell survival and human SLE, such as described for the proliferation capacity as well as trafficking humoral response are lacking. In particular, it potential. As these differences were not is not clear which histone constitutes the dependent on disease activity or medication major auto-Ag in T cell activation. use, our findings cannot be directly translated into clinical practice as e.g. an alternative Therefore, in Chapter 4 we examined the T- marker for disease activity. However, the cell reactivity to the different core histones findings suggest an increased turnover of (H2A, H2B, H3 and H4) and H1. We cells, implying immunological activity investigated the proliferative and cytokine despite clinically quiescent disease. This response of peripheral blood mononuclear suggests the need for immunosuppression cells (PBMCs) upon incubation with single even in times of remission. histones in cells from healthy controls and from SLE patients. The most apparent difference between these two populations Auto-reactive T-cells in SLE and was seen in the reactivity against H1, which RA was significantly higher in SLE patients in frequency and degree, measured by the In addition to an altered composition of the Stimulation index (SI). (frequency of positive T-cell compartment, the auto-reactivity of the response: HC: 11%, SLE patients: 44%; T cell population is considered pivotal in the p<0.01; mean SI: HC: 1.5 ± 0.4, SLE pathogenesis of auto-immune disease. The patients: 2.3±1.5, p<0.01). Also H2A elicited second part of my thesis was therefore a T-cell response which was significantly devoted to identifying targets and properties more outspoken in SLE patients (measured of antigen-specific T-cells against prominent by SI), but not significantly more frequent. auto-antigen in SLE (Chapter 4 and 5) and All other histones showed a similar response in RA (Chapter 6). concerning frequency and degree in HC and SLE patients. It is noteworthy that only in In Chapter 4 and Chapter 5 T-cell auto- SLE and not in HC the cellular response to reactivity in SLE is examined. Antibodies to H1 generally correlated with the response dsDNA are a key serologic characteristic of elicited by the other histones. The SLE and can target different chromatin proliferative responses were accompanied by components, either the complete nucleosome production of IFN and TNFin both or parts thereof. In Chapter 4 histones are examined populations, suggesting a Th1 the studied auto-antigens, followed in response to histones. Histone H1 induced the Chapter 5 by evaluating cellular reactivity to highest TNF response, which was generally hnRNP-A2, a component of the spliceosome. more robust than IFN secretion. The same The latter was also subject of study for RA pattern of cytokine secretion could be seen in patients as is described in Chapter 6. 7 H1-specific T-cell clones (TCC) derived from one SLE patient and 2 cloned from HC. Notably, the antibodies against histones are All TCC were CD4+CD8-TCR+/+, among the earliest to be detected (8); showed a Th1 cytokine pattern and did not

138 cross-react with other histones. Only H1 A difference between SLE patients and HC showed a correlation between cellular and in frequency and magnitude of the cellular humoral reactivity. Overall, a positive reactivity to hnRNP-A2 was found. Whereas serologic response to histones was seen 51% 66% of SLE patients displayed a positive T- of SLE sera and 7% of sera from HCs and cell response in PBMC cultures with more was predominantly directed to H1, H3 and than 50% of patients reacting with an SI>4, H4 (prevalence 51%, 51% and 49%, only 24% of HC did show a positive reaction respectively). Furthermore, most of the anti- and only 5% had an SI>4. The mean SI of histone Abs were of the IgG2 type. The SLE patients amounted to 6.7±2.3 and was potential to induce an antibody response by much lower in the control group, (mean SI H1-specific T-cells was tested in primary 2.3±0.2; p<0.00002). There was no cultures of PBMC of SLE- and RA-patients correlation with humoral reactivity. Tetanus as well as HC. IgG baseline levels were toxoid (TT) was used as positive control similar in all cultures and moderate but not recall antigen and elicited similar reactions in significant increases were observed upon patients and controls. However, whereas in unspecific stimulation or in the presence of HC TT induced a significantly higher the recall antigen tetanus toxoid (TT). Upon response than hnRNP2, this was not the case exposure to histone H1, IgG production in SLE patients. The examination of increased only in SLE patient derived hnRNPA2 specific T-cell clones concerning cultures (p=0.02 as compared to HC phenotype, cytokine production, and cultures). Also, stimulation with H1 resulted specificity showed similarities in phenotype in a significant increase of anti-histone Ab between TCC raised from SLE patients and levels from baseline (p<0.04) only in SLE HC, with approximately 2/3 being CD4+. In patients; these were significantly higher than contrast, the expression of the co-stimulatory in H1-stimulated HC cultures (p=0.02). molecule CD28 differed between TCC Altogether, we could demonstrate that H1 is derived from HC or SLE patients. Whereas the major cellular auto-antigen in SLE all examined TCC of HC were CD28+, the patients underlying the humoral response to majority of SLE-TCC lacked CD28 chromatin and that these H1-reactive T-cells expression. The CD8+CD28- TCC from SLE belong to the Th1 cell subtype. patients furthermore differed in cytokine secretion when compared to the CD4+ But histones are not the only antigens CD28- from SLE patients as well as to the recognized by SLE patients. In 20-30% of CD8+CD28+ TCC from HC; the SLE SLE patients auto-Ab to hnRNP-A2, a derived CD8+CD28- produce hardly any component of the spliceosome, can be IFN, but large amounts of IL-10. Generally, detected. (13) The function of hnRNP-A2 all generated TCC did not produce IL-4, and includes regulation of alternative splicing, most secreted IFN, indicating a Th1 transport of mRNA, and regulation of response to hnRNPA2. As the CD8+CD28- translation. (14) Only few reports about the SLE TCC had a distinct cytokine profile, a T-cell response to spliceosomal antigens are possibly unique, regulatory effector function available, and one of them characterized was assessed and compared to CD4+CD28+ hnRNP-A2 specific CD4+ T-cell clones TCC from SLE patients. Unexpectedly, all (TCCs) in mixed connective tissue disease supernatants induced a proliferative response patient and SLE patients, attributing a in the stimulated effector CD4 T-cells, in possible pathogenic role to these T-cells in contrast to incubation with purified IL-10. SLE. (15) However, the possible differences Thus, as for the cellular reactivity to histones in auto-antigen specific cellular reactivity (Chapter 4), we saw a distinct difference in between patients and HC had not been cellular reactivity against hnRNP-A2 in SLE investigated thus far and was the objective of patients compared to HC and could identify the study described in Chapter 5. them as predominantly being Th1 cells.

139 Additionally, we identified a possibly new citrullinated peptides. To learn more about subset of not-suppressive, IL-10 secreting the T-cell reactivity possibly underlying CD8+CD28- T-cells against hnRNP-A2, ACPA production, we investigated the which was seen only in SLE patients. cellular response of primary cultures to Filaggrin and 2 known citrullinated Filaggrin As auto-Ab to hnRNP-A2 can be detected in (cFil) B-cell epitopes in their original as well about one-third of RA-patients and are a as in citrullinated form. We did not find any marker of early disease (16) the cellular substantial difference between HC and RA- reactivity to hn-RNP-A2 in RA-patients was patients. Also no difference in pattern investigated in Chapter 6. We found the recognition between citrullinated or non- majority of RA-patients reacting (58%) on a citrullinated antigens could be observed. cellular level to this auto-antigen in contrast However, the observation that cFil almost to 20% of the control populations. As seen in always induced a weaker primary response SLE patients, also RA-patients displayed a suggested that citrullination indeed alters T- significantly more vigorous reactivity with cell recognition of antigens. hnRNP-A2 than controls, which was, again, Altogether, in Chapter 6, we described a not correlated with the humoral response. vigorous pro-inflammatory T-cell reactivity (RA patients: mean SI: 4 ±3.5; HC: mean SI: against hnRNP-A2 in RA-patients and an 1.5 ± 1.1) The cytokine secretion pattern seen absent cellular reactivity to (citrullinated) in RA and HC was a Th1 response, and only Filaggrin, suggesting a pathological differed in terms of a higher IL-2 production involvement of hnRNP-A2 specific T-cells in by RA-patients. Synovial expression of the the pathogenesis of RA, which could not be antigen hnRNP-A2 as well as synovial T-cell demonstrated for (citrullinated) Filaggrin. reactivity were measured as well; the In Chapters 4-6 we could identify and predominant localization of hnRNP-A2 was characterize the underlying cellular reactivity the lining and sublining layers of RA- to important auto-antigens of SLE and RA. synovium. HnRNP-A2 was mainly expressed We showed that histone H1 is the most by CD68-positive macrophages, fibroblast- important target of cellular reactivity in SLE like synoviocytes, and endothelial cells and underlying the pathognomonic antibody was scarcely observed in osteoarthritis reaction towards dsDNA. Furthermore we synovial tissue as a non-inflammatory demonstrated the importance of hnRNP-A2 disease control. The positive T-cell response as cellular auto-antigen in SLE and RA. It is was also seen in RA-synovial fluid samples, noteworthy that the vast majority of T-cells which was generally higher than the one seen displayed a Th1 type, irrespective of antigen in peripheral blood, indicating a preferential or disease. However, we also discovered an localization to the joint of these specific T- SLE specific T-cell subset with a distinct cells. Also, T-cell clones (TCC) against phenotype reacting to hnRNP-A2. hnRNP-A2 were Th1 cells, marked by high Altogether, our data substantiate the role of IFNand absent IL-secretion. In contrast to T-cells in these two auto-immune diseases. hnRNP-A2 TCC from SLE patients, phenotypical evaluation of TCC showed almost always CD4+ T-cells in RA, whereas T-cells in RA and the response to HC also generated CD8+ TCC. Generally, Glucocorticoids the IFN production was higher in TCC from RA-patients, although blood from one HC CD4+ T-cell are not only important in the gave rise to three CD4+CD8+ TCC, which disease process of RA, but are also major produced high amounts of IFN. targets of the classical immunosuppressant Additionally we sought to characterize the drug used in this disease, glucocorticoids underlying T-cell reactivity to other (GC). Although in general very effective, important RA-associated auto-Ags, namely approximately one third of the patients does

140 not react appropriately to GC (17). This altered expression of 19 individual genes in prompted us to investigate the difference in CD4+ T-cells (48 in monocytes) exclusively gene expression profile of CD4+ T-cells and in GC-Responders, 104 genes (253 in monocytes between GC-Responders and monocytes) in Non-Responders exclusively, Non-Responders by using cDNA microarrays and 18 overlapping genes (104 in monocytes) and qPCR validation, as presented in in GC-Responders as well as Non- Chapter 7 and Chapter 8. The direct Responders. This is less than the expected comparison of gene expression profiles 10-20% genes regulated by GC, which is due between GC-Responders and Non- to the strict statistical correction applied and Responders before pulse therapy yielded a the additional cut-off point of 1.5 fold change number of genes in T-cells as well as used in our study. monocytes, which have not been implicated Pathway analysis revealed only one distinct in GC action before. The most striking pathway in CD4+ of GC-Responders, difference was the highly significant >4 fold involving calcium ion binding, which turned increased expression of ERAP2 in GC- out to be overlapping in GC-Responders and Responders before therapy in CD4+ T-cells Non-Responders. In Non-Responders genes and monocytes when compared to Non- were clustered into the pathways: response to Responders. (CD4+ T-cells: 4.9-fold, p<0.02; virus, IL-22 soluble Receptor signaling monocytes: 6.6-fold, p<0.007). In pathway, and immune response and showed a monocytes, also TRIM33 expression was highly and very significantly enriched IFN- significantly higher in GC-Responders (1.4- signature. The genes changed by GC in fold, p< 0.04). In CD4+T-cells increased monocytes were distributed over the expression of FAM26F (4.4-fold, p<0.006) pathways immune response, defense and LST1 (2.5-fold, p<0.009) was detected. response, and negative regulation of defense Interestingly, the differential gene expression response, whereas genes from Non- of ERAP2 was sustained 24 hours after Responders were clustered in immune therapy. Furthermore, the correlation with response, response to virus, defense disease activity after therapy proved response, response to wound healing, and significant for ERAP2 in monocytes (p<0.05) inflammatory response. and LST1 and FAM26F in CD4+cells (both: Again, a dominance of IFN related genes p<0.02), underlining their potential were seen in the Non-Responders. predictive value. As smoking has been The observed stronger suppression of IFN- implicated in GC-resistance in asthma, the related genes in Non-Responders seen in the association of the differentially expressed microarray analysis was verified in CD4+ T- genes and smoking - measured by cigarette cells by qPCR. pack years -was examined. It demonstrated a Additionally, several transcription-factors significant correlation between the were found in both cell types from Non- expression of LST1 and FAM26F in Responders, which are able to influence CD4+cells. In this chapter we could thus multiple genes from the generated gene lists identify several potential predictors of GC- described above. In CD4+ cells, LEF1 and response in RA-patients, the most important SP1 were found, whereas in monocytes, the of which seems to be ERAP2. Future studies transcription-factors were LEF1, NFAT-1, have to confirm the usefulness of these IRF-1 and NFkB1. Several of these predictors in clinical practice. transcription factors could be correlated in To identify possible differences in the terms of gene expression values with their working mechanisms of GC related to the target genes, by which (statistically) a clinical response, the change in gene distinct transcription-factor profile could be expression pattern before and 24 hours after discerned. This indicates that a specific pulse therapy was investigated as described transcription-factor network might actually in Chapter 8. GC action resulted in the orchestrate the processes underlying Non-

141 Respondership. Altogether, in Chapter 8 we counting by FACS beads, and thymidine identified a difference in GC-working in incorporation. Together, these data are most CD4+ T-cells between GC-Responders and consistent with the conclusion that Non-Responders, which is mirrored by a proliferating TCM eventually give rise to TEM differential down-regulation of the IFN with diminished in vivo and in vitro signature. proliferative potential. The decreased proliferative capacity found in our study in the CCR7- subsets was paralleled by an General Discussion: increased propensity to undergo apoptosis upon stimulation. Recently, this aspect was

examined and the underlying mechanisms

were discovered to be Fas-dependent. (20) Memory T-cell subsets in health and CCR7-/CD27- T-cells were more prone to disease (Part 1) Fas-induced apoptosis and therefore also to RICD (restimulation induced cell death) The proposed maturational pathway of elicited by TCR stimulation. This effect CD4+T-cells from naïve  CCR7+/CD27+ seemed to be depend on Fas-signaling, as the  CCR7-/CD27+  CCR7-/CD27- allows CCR7+ positive subsets had decreased death for a more detailed view of human CD4 inducing signalling complex (DISC) activity. memory T cell maturation. The data in this Furthermore, in contrast to the CRR7+ cells, thesis confirmed the suspected micro-domains of lipid rafts of the terminally developmental relationship between TCM and differentiated cells were shown to contain TEM as previously suggested (1) and Fas. Lipid rafts are glycosphingolipid and specified different subpopulations in the TEM cholesterol enriched, low-density lipoprotein subset. In a more recent study, our data were fractions of the plasma membrane. Pre- confirmed by authors delineating the localization of Fas to lipid rafts makes T-cells maturation by the use of CD45RA, CCR7, more sensitive to apoptosis after TCR CD27, and CD28. (18) The authors stimulation (21) and may thus be a correlated phenotypic expression with contributing factor for the increased cytokine secretion and found the subset inclination to undergo apoptosis in the CD45RA-/CCR7-/CD27+/CD28+ to CCR7- T-cell subsets. predominantly include Th0 and Th1 cells. Th2 cells were distributed to the CD45RA- The increase of CCR7- subsets in SLE /CCR7-/CD27-CD28+ and CD45RA-/CCR7- patients is noteworthy, as it cannot be /CD27-/CD28- cells, in line with our explained by mere transient down-regulation findings. of these receptors after in vivo stimulation, as A previous report had found a higher was proven by a lack of re-expression also turnover of CCR7- cells, which is in line with after a resting period. (22,23) This suggests our finding of shortened telomeres in these an accumulation of more differentiated cells subsets. However, in contrast to the authors in SLE. Additionally, significant deficiencies we found a diminished proliferative capacity in cell survival and proliferation of the ex vivo of the CCR7- subsets. (19) The CCR7-/CD27+ and CCR7-/CD27- subsets discrepancy might be due to the difference compared with their counterparts from between ex vivo and in vitro measurements healthy controls were apparent. Whether this (in our study) as compared to the in vivo is due to in vivo stimulation or due to the situation where possible other constraints on known dys-regulation of cell death in SLE the CCR7+ subsets are present. Importantly, (24,25) remains unclear. However, the the in vivo experiments did not examine the confinement of enhanced apoptosis to the potential proliferative capacity as we CCR7- subsets suggest an abnormality examined by cell cycle analysis, cell resulting from in vivo differentiation and not

142 a characteristic of all SLE memory T cells. cycling naive T cells. Moreover, there was no Likewise, if the underlying cause for the increase in “antigen-experienced” accumulation of differentiated memory T- peripherally expanded naïve CD4 T cells, cells in SLE, is an intrinsic problem leading characterized by the lack of CD31 to a lower threshold of activation (7) or a expression, (34) in the SLE patients greater exposure to auto-antigens remains to examined. An alternative explanation could be determined. be a genetic abnormality in telomerase As mentioned above, the higher rate of expression causing reduced functionality of apoptosis in CCR7- subsets in HC was partly this enzyme. In this context. a lupus found to be caused by decreased pre- susceptibility locus has been found in the localization of Fas in lipid rafts of these genetic region of hTERT, the enzymatically subsets. (21) In SLE, accelerated dynamics active component of telomerase. (35) Also, a and altered composition of lipid rafts have different cytokine milieu in patients with been shown, possibly contributing to SLE might alter inducible telomerase amplification of TCR mediated signals in activity, as an increase in serum TNF levels SLE T-cells. (26) However, no investigation is characteristic of SLE, (36) and TNF is of Fas localization in lipid rafts of SLE known to regulate telomerase. (37) patients has been conducted so far, but an increased expression of Fas has been noted in Altogether the data discovered in these CD4+ memory T-cells. (27) From the studies warrant further investigations into combined data, increased localization in the these defined T-cell subsets; ongoing CCR7- subsets of SLE patients is expected, examination of their gene expression profile and may delineate a new target for by microarray analysis is expected to reveal therapeutic approaches. more details on the maturational pathway of Noteworthy is the decreased inducible T-cells and on the differences between these telomerase expression in naïve and CCR7+ subsets derived from SLE compared to HC. cells of SLE patients. Several studies have been undertaken in lupus patients, and come to different conclusions about telomere Auto-reactive T-cells in SLE and length in SLE. Our data show no difference RA (Part 2) in telomere length between age matched HC and SLE patient using Flow-FISH, as does a (T-cell) auto-reactivity in SLE: later report using the same technique, (28) Autoimmune responses to chromatin play a suggesting a difference in technique as pivotal role in the pathogenesis of SLE and possible reason for different results seen in the pathognomonic anti-dsDNA Ab are earlier studies. (29-31) The decreased known to be crucial. However, they may not inducible telomerase found in our study be the first target of the auto-Ab response, seems to contradict other reports, (29-32) but could be preceded by Abs to other however, the reported telomerase activity in chromatin components, particularly histones. our study is after additional in vitro (8) Anti-histone Abs, particularly Abs to stimulation. Furthermore, the other reports histone H1, not only correlate with SLE did not stratify telomerase activity according disease activity, but also seem to be more to T-cell subsets, complicating a direct specific for SLE than Abs to core histones. comparison. Our findings of reduced (10) As these auto-Ab are class-switched, the inducible telomerase activity in naïve and underlying T-cell auto-reactivity has been the CCR7+ T-cell subsets could be due to an in focus of some studies and identified histones vivo stimulation of these cells, leading to a as the major T-cell antigens in murine and reported down-regulation upon re-stimulation human SLE. In a murine lupus model the by our in vitro experiments. (33) However, auto-epitopes of lupus nephritis-inducing this is unlikely, since there was no increase in CD4+ T-cells: H2B10–33, H416–39 H471–94 and

143 H385–102 and H122–42 were identified. (38) attributed in SLE, this issue is certainly These peptides could accelerate lupus worthwhile examining in the future. (42, 43). nephritis in vivo by generating Th1 cells, A point of interest is the high amount of which induced the production of pathogenic TNF production in SLE T-cell clones and anti-nuclear auto-Ab. In a next step, these PBMCs. Since TNF correlates with SLE epitopes were tested in human lupus T-cell disease activity (44) and its therapeutical lines and PBMC and the epitopes from the blockade results in clinical benefit, including core histones were confirmed as human lupus amelioration of lupus nephritis, (45) it is auto-antigens. (39) In a different approach tempting to speculate that histone specific T- (12) it was shown that T-cell clones raised cells, particularly those directed to H1, could against a mixture of histones could augment be important stimulators of inflammation in antibody response against dsDNA in SLE. autologous B-cells, underlining the importance of T-cell reactivity against (core) It is interesting that, despite the relatively histones. In our study we found the most high prevalence of auto-Abs to H3 and H4 in pronounced T-cell reactivity to histones in sera of patients with SLE, cellular reactivity PBMC to be directed to H1, which has not to these histones was not different from been reported earlier. In mice the epitopes healthy controls. Moreover, TNF secretion derived from H1 were the most potent elicited by H3 and H4 in the primary culture inducers of nephritis; (38) additionally, it has assays was considerably lower than been shown in humans that anti-H1-Ab are TNFsecretion induced by H1. Thus, it is superior to anti-core histone Ab in terms of possible that the autoimmune response to H3 lupus specificity and disease activity and H4 may have less pathogenic correlation. (9) These findings suggest a consequences, or might even exert protective prominent role of anti-H1 T-cell reactivity in functions. This assumption is supported by human lupus as seen in our study. observations made in experimental lupus in Interestingly, cellular reactivity to H1 which peptides derived from H3 or H4, in correlated significantly with cellular contrast to H1-specific peptides, suppressed responses to histones H2A, H2B and H4 or delayed lupus symptoms when applied to suggesting that H1-specific T-cells may at lupus-prone mice. (46,47) least partially drive immune responses to In line with previous observations, histones other histones, presumably due to the H1, H3 and H4 appeared to be major mechanism of intermolecular epitope antigens of the humoral response in SLE, spreading as has been described for while Abs to H2A and H2B were less spliceosomal auto-antigens in SLE. (40) frequently found.(9) However, concerning The auto-reactive T-cells against histones in these latter two histones, H2A–H2B dimers our study all showed a Th1 like pattern, with may rather form the preferred target of high amounts of TNF and (to a lesser autoAbs, especially in drug-induced LE (48), extent) IFNcombined with a lack of IL-4 an issue that was not addressed in our study. production. This is in line with previous Also possible modifications of histones due reports, (12,39) and adds to the body of to apoptotic changes, such as acetylation of evidence supporting the role of Th1 cells in H2B (49,50) have not been investigated by SLE, which is further substantiated by the us; likewise examining possible reactivity to predominance of IgG2 subtype of the anti- modified histones derived from neutrophil histone Abs. However, we cannot rule out the extracellular traps (NETs) seems desirable. possibility of the T-cell clones being in fact NETs, consisting of chromatin laced with Th17 cells, as it has been reported that Th17 histones and other bactericidal proteins, are cells are able to produce TNF and low extruded by neutrophils and have recently amounts of IFNConsidering the been implicated in SLE pathogenesis. (51,52) increasingly important role Th17 cells are Thus, although the cellular response seems to

144 be predominantly directed against H1, we Chapter 4 and previous reports, (12, 15) we cannot exclude that modification of the observed a relatively high percentage of histone-proteins as seen in apoptosis or CD8+ TCCs in SLE patients. Of particular NETosis may also elicit a substantial T-cell interest, most of these clones lacked CD28 response, which should be investigated in expression and produced neither IFNγ nor future studies. IL-4, but did produce IL-10. This indicates a special pathological role of this T-cell subset Although the auto-reactivity to chromatin because this phenotype appeared to be components is crucial in SLE, also Ab to restricted to SLE patient derived TCCs and other auto-antigens are seen in SLE, such as was not found in HC. Ribonucleoproteins (RNP; Ref 53) possibly contributing to lupus pathogenesis. Therefore The relatively high proportion of hnRNP-A2 we also examined the underlying T-cell specific CD8+ TCCs might mirror the reactivity against hnRNPA2, auto-Ab to common understanding of the importance of which can be seen in 20% of SLE patients. CD8+ T-cells in SLE. An increased (13) CD8:CD4 ratio in SLE patients (54-56) has The data obtained showed the existence of a been reported; this seems to be correlated pronounced cellular response to hnRNP-A2 with disease activity, including nephritis in the majority of SLE patients, which was (57,58) and determines SLE prognosis. (59) far more vigorous than in HC. In contrast, the However, there are only anecdotal reports response to tetanus toxoid was similar in about auto-reactive CD8+ TCCs in SLE (53, patients and controls, demonstrating the 60) and their role is not entirely clear and specificity of the immune response towards may be quite complex indeed. hnRNP-A2. Moreover, the response of SLE The rather low IFNγ secretion of patient patients to hnRNP-A2 was of similar derived CD8+ TCCs might be due to magnitude, or even slightly higher, than the inherent abnormalities of CD8+ (suppressor) response to the classic recall antigen. It is T cell function in SLE, as has been shown. interesting that the robustness of cellular (61) On the other hand, these TCCs secreted response to hnRNP-A2 in terms of considerable amounts of IL-10 and, in proliferation seemed considerably higher contrast to the TCCs derived from healthy than the one seen against H1, suggesting a controls, were CD28 negative. They could possibly underestimated role of auto- thus be regulatory CD8+ cells (62-64) reactivity to RNPs in SLE. Nevertheless, as exerting their function in a cell contact the experiments were not carried out independent manner via secretion of IL-10. simultaneously, we cannot exclude the As such they could constitute an effort of differences in patients and materials (batch counter-regulation within the disturbed effect) as contributing factor. However, immune system of SLE patients. Although whereas the cellular response to H1 was there are no reports over an increased correlated with the presence of the respective frequency of such cells in lupus, (65,66) auto-Abs, the serologic reactivity to increased survival and telomerase activity of hnRNPA2 was far less frequent than the this subset has been described. (31) cellular response. Although this might be due Unexpectedly though, the supernatants of the to fluctuations in antibody response hnRNP-A2 specific CD8+CD28- TCCs (unpublished observation) it is also thinkable derived from SLE patients did not inhibit that hnRNP-A2 may predominantly be a T- proliferation of CD4+ T cells, but instead cell stimulus. enhanced anti-CD3/anti-CD28 induced stimulation, similar to supernatants derived Although the cellular reactivity to hnRNP-A2 from CD4+CD28+ TCCs. As IL-10 alone appeared to be primarily a Th1 response had an anti-proliferative effect, the increased similar to the response to histones in content of IL-10 in the supernatants of

145 CD8+CD28- TCCs was obviously described for B-cell epitopes, (71) counteracted by other yet unknown differences in the T-cell response were stimulatory components of the supernatant. apparent, upon characterization of T-cell Of note, the lack of CD28 expression is also clones. RA patients raised an almost a hallmark of senescent T cells, which are exclusive Th1, thus CD4+ mediated known to be increased in various response, which was diminished by MHCII autoimmune diseases and chronic blockade. This is in contrast to the above inflammatory conditions and may show a mentioned hnRNP-A2 induced CD8+ T-cells rather aggressive phenotype. (67) Therefore, seen in lupus, and reflects the different it is conceivable that the CD8+CD28- TCC importance ascribed hitherto to CD8+ cells in may be derived from the senescent T-cell these two diseases. However, in both SLE pool of SLE patients. and RA, in contrast to HC, hnRNP-A2 Elevated IL-10 serum levels have been reactivity exceeded the one against tetanus reported in SLE patients and correlated with toxoid, suggesting an Ag-driven expansion in clinical and serological disease activity, both diseases. Although the observed IFN especially anti-dsDNA antibody titres. (68) secretion had implicated a pro-inflammatory Several cell potential of these cells, the current view of types have been implicated in the production the importance of Th17 cells in RA- of IL-10 (69) and augmented IL-10 secretion pathogenesis and the inhibitory potential of has been linked to auto-Ab production in a IFN on such Th17 cells could offer a model where PBMC from patients with SLE different, maybe even protective, role to were hnRNP-A2 specific T-cells in RA. (72). This transferred into mice with severe combined could also explain the reported association of immunodeficiency. (70) So, altogether, the hnRNP-A2 auto-Abs in patients with a CD8+CD28- T cell subpopulation might milder disease course. (73) However, in a enhance the inflammatory process in SLE recent study, immunodominant T-cell patients by stimulation of B cells via IL-10. peptides of hnRNP-A2 were identified, Unfortunately, because of the short life-span eliciting an IFN response in RA patients. of the clones, additional functional assays The presence of these specific T-cells were could not be performed and remain the significantly associated with active disease in subject of future investigations. RA patients and with bone erosion, In conclusion, our data identify a pronounced suggesting involvement of hnRNP-A2 T cell response against H1 and hnRNP-A2 in specific cellular autoimmune responses in the majority of SLE patients compared to a RA pathogenesis.(74) scarcer and lower response in healthy An interesting finding of our study is the subjects. Furthermore we could characterize discovery of three CD4+CD8+ double the response in more detail and found positive (DP) hnRNP-A2 specific TCC, all increased TNF production in TCC against derived from a healthy control. These DP T- H1. Moreover, we discovered a distinct cells constitute a small population in the CD8+CD28- IL-10 producing T-cell type in peripheral blood of healthy individuals, SLE patients. Both these cell types contribute increasing with age. (75) Furthermore, these to lupus pathogenesis in a different way and cells may be increased in patients with might have therapeutical implications. autoimmune diseases, viral infections (76-78) and auto-immune diseases. (79-81) Although (T-cell) auto-reactivity in RA: their precise function has not been The primary cellular response to hnRNP-A2 elucidated, it seems that the DP T-cells in was also assessed in RA and was similar to peripheral blood or target tissues are mature the one seen in SLE, with the majority of T-cells that may contribute to auto-immune patients reacting with a strong proliferation pathology. However, to our knowledge, this and IFN production. However, as had been is the only report of auto-reactive DP T-cell

146 clones in a healthy donor (who has not important targets of GC, we investigated the developed RA in the following 10 years). difference in gene expression profiling in We also assessed the T-cell response to these cells to identify 1) potential predictors citrullinated Filaggrin, which is a known for the clinical response to GC (Chapter 7) target of ACPA. We did not find any and 2) potential differences in GC-action substantial difference between healthy mechanism between GC-Responders and controls and RA-patients or preferential Non-Responders (Chapter 8). We found recognition of the citrullinated peptides over interesting answers to both question. The non-citrullinated ones. This indicates that most striking difference between GC- filaggrin does not constitute a T-cell antigen. Responders and Non-Responders was the The characterization of T-cell response significantly higher expression of ERAP2 underlying ACPA has remained a difficult (endoplasmic reticulum aminopeptidase 2) in task. However, several recent papers GC-Responders before, and also 24 hours identified T-cell reactivity to citrullinated after start of pulse therapy in both cell types. peptides, and found a proliferative response ERAP2 is suggested to be a gene duplication accompanied by IL-17 secretion for of ERAP1 (endoplasmic reticulum citrullinated aggrecan. (82) Also citrullinated aminopeptidase 1) and both genes have been vimentin peptides were shown to induce identified as susceptibility loci for cytokine production, though no data on the autoimmune diseases (reviewed in 85). As proliferation were given. (83) However, as peptidases, both enzymes cause the removal these studies were carried out using PBMC, of amino acids from N-termini of peptides or the exact nature of the specific T-cell proteins. This function is used, among others, remained elusive. In this respect, only in a in the trimming of peptides to optimal length recent paper the proliferative response as for their binding into the MHC class I well as the cytokine response upon groove, crucially affecting antigen- incubation with citrullinated and native presentation. (86) Interestingly, ERAP2 vimentin, collagen, fibrinogen, and aggrecan preferably cleaves basic aminoacids as was measured in combination with Arginine or Lysine. Consequently, targets assessment of the T-cell phenotype. (84) Of for deimination are made accessible which all these auto-antigens, citrullinated aggrecan can result in citrulline residues. Considering was the most immunogenic in terms of the importance of citrullinated peptides and cytokine secretion, but did not induce a the thus far incompletely understood T-cell proliferative response. The reactive T-cells reactivity against them, a role herein for were all CD4 positive, and in RA-patients ERAP2 is imaginable, although the comprised CD28+ and CD28- cells. The fact mechanism is not yet understood. that despite intense research efforts, the An additional function of ERAP1 is the cellular response to ACPA is still proteolytic cleavage of cytokine receptors incompletely characterized, as e.g. in terms TNFRI, IL-6R, and IL-1RII, (85,87) which of the discrepant proliferative and cytokine in their soluble state act as decoy for their secreting capacities induced by citrullinated respective inflammatory cytokines. As antigens, suggests that ACPA constitute TNFand IFN induce the expression of more of a B-cell antigen than a T-cell both peptidases, this suggests ERAP1 to be stimulus. the center of an inflammatory negative feedback loop. Considering that ERAP2 has originated as duplicate of ERAP1 and T-cells and monocytes in RA and the displays close functional similarity and response to Glucocorticoids (Part 3) homology, cytokine receptor shedding by ERAP2 seems possible. However, such a As CD4+ T cells and monocytes are not function has not been reported in the only important in RA-pathogenesis, but also literature so far. When we tried to correlate

147 ERAP2 expression with serum levels of the clinical response to Rituximab (91,92) an soluble receptors TNFRI, IL6Rand IL1RII, increase after Rituximab treatment, was we could not find such an association. correlated with a good clinical response. (93) Nevertheless, taking into account the The opposite relation was seen for the multiple factors influencing soluble receptors association of TNF blockade and the in serum, the lack of this correlation does not interferon signature. Here, good responders preclude an influence of ERAP2 in receptor to anti-TNF-therapy showed a high interferon shedding on a cellular level and should be signature at baseline. (94) These results addressed in future studies. reflect the complex interplay of TNF and Another gene of interest that we found to be type I interferons in general and in RA in increased in GC-Responders was LST1, the specific. The original view of a mutual leucocyte-specific transcript 1, which negative regulation originated from the regulates T-cell proliferation and constitutes observation that TNF blockade induced an a susceptibility locus in RA. (88) Increased increased interferon signature (95) clearly is expression of LST1 has been reported in oversimplified. blood and synovial tissue from RA-patients in association with a higher degree of As GC regulate interferon gene expression, synovial inflammation. (89) In our study we the down-regulation of interferon response found a 2.5 fold increased expression of genes seen in both cell types was expected. LST1 in CD4-T-cells from GC-Responders However, the more prominent down- compared to Non-Responders and, of note, a regulation in Non-Responders, which was correlation with smoking. This newly found confirmed in qPCR experiments in CD4+ connection between LST1, GC- cells was surprising. Due to the interplay of responsiveness and smoking is worthwhile interferons and TNF it cannot be excluded exploring in the future. that this enhanced down-regulation in Non- Our second aim was to identify differences in Responders is in fact caused by known GC-action between GC-Responders and effects of GC on TNF production and Non-Responders. This was determined by signalling. Indeed, in our transcription factor comparing the changes in gene expression analysis, both IRF-1 as well as NFkB1 were induced by GC pulse therapy between GC- found to be regulated by GC in Non- Responders and Non-Responders. We Responders making a distinction between a observed a stronger down-regulation of IFN direct effect of GC on IFN and an indirect via related genes in both cell types of Non- TNF difficult. However, the indirect Responders. influence on the interferon signature via TNF Interferons are pleiotropic cytokines, which seems less likely, as an up-regulation of IFN not only confer an antiviral state, but are also signature denominated a worse response after involved in lymphoid differentiation, TNF-blockade, (96) whereas a stronger homeostasis, tolerance, and memory. Both down-regulation of interferon related genes type I (IFNand type II interferons was seen in Non-Responders to GC. This (IFN) are known to contribute to RA- may have clinical implications, e.g. a pathogenesis. The presence of an “interferon possibly favourable combination treatment of signature”, as observed in approximately TNF-blockade with GC. half of RA-patients, has only recently been acknowledged in RA-pathogenesis. (90) In patients with asthma, NFkB1 had an 81% Apart from involvement in RA-pathogenesis, accuracy in predicting the GC-response. (97) the interferon signature has also been In our microarray results, NFkB1 was implicated in therapy response to biologicals, identified as a transcription factor of although yielding variable results. Whereas importance in monocytes from Non- an increased baseline expression of IFN Responders, but the actual expression values related genes seems to be predictive of worse of NFkB1 were not significantly changed.

148 Interestingly, expression in CD4+cells was References down-regulated (ratio 0.29) in Non- Responders, but slightly up-regulated in GC- 1) Sallusto F, Lenig D, Forster R, Lipp Responders (FC: 1.26), although both not M, Lanzavecchia A. 1999. Two subsets statistically significant. The relevance of this of memory T lymphocytes with distinct finding is not clear, especially as it did not homing and effector functions. withstand (FWER) statistical correction. Nature;401:708-12. Another noteworthy finding was a significant 2) Scimone ML, Felbinger TW, Mazo TB, down-regulated RELB in monocytes and Stein JV, von Andrian UH, and CD4-Tcells of Non-Responders exclusively, Weninger W. 2004. CXCL12 mediates which implies a differential role of NFkB in CCR7-independent homing of central GC-response also in RA-patients. memory cells, but not naive T cells in In infant acute lymphoblastic leukemia peripheral lymph nodes. J Exp (ALL) cells a causative relation of increased Med;199:1113–20. S100A8 and S100A9 –expression leading to 3) Breitfeld D, Ohl L, Kremmer E, a hampered Ca-flux to the mitochondria, Ellwart J, Sallusto F, Lipp M, and which is necessary for GC-induced Forster R. 2000. Follicular B helper T apoptosis, was descried. In our study, we did cells express CXC chemokine receptor not see a differential gene induction of 5, localize to B cell follicles, and calgranulin between Non-Responders and support immunoglobulin production. J. GC-Responders in CD4+T-cells. (98) Exp Med;192: 1545–52. The discrepancy between the findings of 4) M. Funauchi, S. Ikoma, H. Enomoto, these mentioned studies and ours suggest a and A. Horiuchi. 1998. Decreased Th1- disease-specific GC-resistance pattern in RA- like and increased Th2-like cells in patients, which is further substantiated by the systemic lupus erythematosus. Scand J pre-treatment increase of ERAP2 reported in Rheum; 27: 219–24. GC-responding RA-patients by us. 5) Akahoshi M, Nakashima H, Tanaka Y, Kohsaka T, Nagano S, Ohgami E, Altogether, we have identified a possible Arinobu Y, Yamaoka K, Niiro H, novel marker of GC-response in RA, Shinozaki M, Hirakata H, Horiuchi T, ERAP2. As our study was carried out in a Otsuka T, Niho Y. 1999. Th1/Th2 limited number of patients, the verification of balance of peripheral T helper cells in our finding in a larger cohort is of great systemic lupus erythematosus. Arthritis importance. However, the functional aspects Rheum;42:1644-8. of ERAP2 in RA harbour potentially new 6) Shah K, Lee WW, Lee SH, Kim SH, insight in pathological mechanisms and are Kang SW, Craft J, Kang I.2010. worthwhile exploring further. The Dysregulated balance of Th17 and Th1 characterized difference in working cells in systemic lupus erythematosus. mechanism of GC in GC-Responders and Arthritis Res Ther;12:R53. Non-Responders associated with a more 7) Crispin JC, Liossis SN, Kis-Toth K, prominent suppression of the interferon Lieberman LA, Kyttaris VC, Juang YT, signature in Non-Responders is of particular Tsokos GC. 2010. Pathogenesis of interest and awaits to be verified in a larger human systemic lupus erythematosus: cohort as well. A special interest should be recent advances. Trends Mol given to examine the differential down- Med;16:47-57 regulation of interferon related genes also in 8) Burlingame RW, Rubin RL, Balderas diseases not known for their interferon RS, Theofilopoulos AN. 1993. Genesis signature. and evolution of antichromatin autoantibodies in murine lupus implicates T-dependent immunization

149 with self antigen. J Clin Invest; disease patients assist in autoantibody 91:1687–96. production. Arthritis Rheum;50:2216- 9) Schett G, Smolen J, Zimmermann C, 22. Hiesberger H, Hoefler E, Fournel S, 16) Hassfeld W, Steiner G, Graninger W, Muller S, Rubin RL, Steiner G. 2002. Witzmann G, Schweitzer H, and The autoimmune response to chromatin Smolen JS. 1993. Autoantibody to the antigens in systemic lupus nuclear antigen RA33: a marker for erythematosus: autoantibodies against early rheumatoid arthritis. Br J histone H1 are a highly specific marker Rheumatol;32:155-59. for SLE associated with increased 17) van Schaardenburg D, Valkema R, disease activity. Lupus;11:704–15. Dijkmans BA, Papapoulos S, 10) Schett G, Rubin RL, Steiner G, Zwinderman AH, Han KH, Pauwels Hiesberger H, Muller S, Smolen J. EK, Breedveld FC. 1995. Prednisone 2000. The lupus erythematosus cell treatment of elderly-onset rheumatoid phenomenon: comparative analysis of arthritis. Disease activity and bone antichromatin antibody specificity in mass in comparison with chloroquine lupus erythematosus cell-positive and - treatment. Arthritis Rheum;38:334-42. negative sera. Arthritis Rheum;43:420– 18) Okada R, Kondo T, Matsuki F, Takata 28. H, Takiguchi M. 2008. Phenotypic 11) Kaliyaperumal A, Michaels MA, Datta classification of human CD4+ T cell SK. 2002. Naturally processed subsets and their differentiation. Int chromatin peptides reveal a major Immunol; 20:1189-99 autoepitope that primes pathogenic T 19) Macallan DC, Wallace D, Zhang Y, de and B cells of lupus. J Lara C, Worth A, Ghattas H, Griffin G, Immunol;168:2530–37. Beverley PCL, and Tough DF. 2004. 12) Voll RE, Roth EA, Girkontaite I, Fehr Rapid turnover of effector memory H, Herrmann M, Lorenz HM, Kalden CD4 T cells. J Exp Med;200: 255–60. JR. 1997. Histonespecific Th0 and Th1 20) Ramaswamy M, Cruz AC, Cleland SY, clones derived from systemic lupus Deng M, Price S, Rao VK, Siegel RM. erythematosus patients induce double- 2011. Specific elimination of effector stranded DNA antibody production. memory CD4+ T cells due to enhanced Arthritis Rheum;40:2162–71. Fas signaling complex formation and 13) Hassfeld W, Steiner G, Studnicka- association with lipid raft Benke A, Skriner K, Graninger W, microdomains. Cell Death Fischer I, Smolen JS.1995. Differ;18:712-20. Autoimmune response to the 21) Muppidi JR, Siegel RM. 2004. Ligand- spliceosome: an immunologic link independent redistribution of Fas between rheumatoid arthritis, mixed (CD95) into lipid rafts mediates connective tissue disease, and systemic clonotypic T cell death. Nat lupus erythematosus. Arthritis Immunol;5: 182–89. Rheum;38:777-85 22) Hintzen RQ, Jong J, Lens SM, van Lier 14) Krecic AM, Swanson MS.1999. RA. 1994.CD27: marker and mediator HnRNP complexes: composition, of T-cell activation. Immunol structure, and function. Curr Opin Cell Today;15:307–11. Biol;11:363-71. 23) Langenkamp A, Nagata K, Murphy K, 15) Greidinger EL, Gazitt T, Jaimes KF, Wu L, Lanzavecchia A, Sallusto F. Hoffman RW.2004. Human T cell 2003. Kinetics and expression patterns clones specific for heterogeneous of chemokine receptors in human nuclear ribonucleoprotein A2 CD4+ T lymphocytes primed by autoantigen from connective tissue

150 myeloid or plasmacytoid dendritic 31) Honda M, Mengesha E, Albano S, cells. Eur J Immunol;33:474–82. Nichols WS, Wallace DJ, Metzger A, 24) Emlen W, Niebur JA, Adera R. 1994. Klinenberg JR, Linker-Israeli M.2001. Accelerated in vitro apoptosis of Telomere shortening and decreased lymphocytes from patients with replicative potential, contrasted by systemic lupus erythematosus. J continued proliferation of telomerase- Immunol;152:3685–92. positive CD8+CD28(lo) T cells in 25) Silvestris S, Grinello D, Tucci M, patients with systemic lupus Cafforio P, Dammaco F. 2003. erythematosus. Clin Immunol;99:211- Enhancement of T cell apoptosis 21. correlates with increased serum levels 32) Klapper W, Moosig F, Sotnikova A, of soluble Fas (CD95-Apo-I) in active Qian W, Schroder JO, Parwaresch R. lupus. Lupus;12:8–14. 2004. Telomerase activity in B and T 26) Krishnan S, Nambiar MP, Warke VG, lymphocytes of patients with systemic Fisher CU, Mitchell J, Delaney N, lupus erythematosus. Ann Rheum Tsokos GC.2004. Alterations in lipid Dis;63:1681–3. raft composition and dynamics 33) Weng NP, Palmer LD, Levine BL, contribute to abnormal T cell responses Lane HC, June CH, Hodes RJ. in systemic lupus erythematosus. J. 1997.Tales of tails: regulation of Immunol;172:7821–31 telomere length and telomerase activity 27) Amasaki Y, Kobayashi S, Takeda T, during lymphocyte development, Ogura N, Jodo S, Nakabayashi T, differentiation, activation, and aging. Tsutsumi A, Fujisaku A, Koike T. Immunol Rev;160:43–54. 1995. Up-regulated expression of Fas 34) Kimmig S, Przybylski GK, Schmidt antigen (CD95) by peripheral naive and CA, Laurisch K, Möwes B, Radbruch memory T cell subsets in patients with A, Thiel A. 2002. Two subsets of naive systemic lupus erythematosus (SLE): a T helper cells with a distinct T cell possible mechanism for lymphopenia. receptor excision circle content in Clin Exp Immunol;22:245–50. human adult peripheral blood. J Exp 28) Beier F, Balabanov S, Amberger CC, Med;195:789–94. Hartmann U, Manger K, Dietz K, 35) Nath SK, Namjou B, Garriott CP, Kötter I, Brummendorf TH.. Telomere Frank S, Joslin PA, Kilpatrick J, Kelly length analysis in monocytes and JA, Harley JB. 2004. Linkage analysis lymphocytes from patients with of SLE susceptibility: confirmation of systemic lupus erythematosus using SLER1 at 5p15.3. Genes multi-color flow-FISH Lupus, 16: 955– Immun;5:209–14. 62 36) Aringer M, Stummvoll GH, Steiner G, 29) Kurosaka D, Yasuda J, Yoshida K, Köller M, Steiner CW, Höfler E, Yoneda A, Yasuda C, Kingetsu I, Hiesberger H, Smolen JS, Graninger Toyokawa Y, Yokoyama T, Saito S, WB. Serum interleukin-15 is elevated Yamada A. 2006.Abnormal telomerase in systemic lupus erythematosus. activity and telomere length in T and B Rheumatology (Oxford) 2001;40:876– cells from patients with systemic lupus 81. erythematosus. J Rheumatol;33:1102–7 37) Akiyama M, Yamada O, Hideshima T, 30) Kurosaka D, Yasuda J, Yoshida K, Yanagisawa T, Yokoi K, Fujisawa K, Yokoyama T, Ozawa Y, Obayashi Y, Eto Y, Yamada H, Anderson KC. Kingetsu I, Saito S, Yamada A. 2003. 2004. TNFinduces rapid activation Telomerase activity and telomere and nuclear translocation of telomerase length of peripheral blood mononuclear in human lymphocytes. Biochem cells in SLE patients. Lupus;12:591–9. Biophys Res Commun;316:528–32.

151 38) Kaliyaperumal A, Mohan C, Wu W label study. Arthritis Rheum;50:3161– and SK Datta. 1996. Nucleosomal 9. peptide epitopes for nephritis-inducing 46) Kang HK, Chiang MY, Liu M, T helper cells of murine lupus. J Exp Ecklund D, Datta SK. 2011. The Med;183:2459–69. histone peptide H4 71-94 alone is more 39) Lu L, Kaliyaperumal A, Boumpas DT effective than a cocktail of peptide & Datta SK. 1999. Major peptide epitopes in controlling lupus: autoepitopes for nucleosome-specific T immunoregulatory mechanisms. J Clin cells of human lupus. J Clin Invest;104: Immunol;31:379-94. 345–55. 47) Kaliyaperumal A, Michaels MA, Datta 40) Monneaux F, Muller S. 2001. Key SK. 1999.Antigen-specific therapy of sequences involved in the spreading of murine lupus nephritis using the systemic autoimmune response to nucleosomal peptides: tolerance spliceosomal proteins. Scand J spreading impairs pathogenic function Immunol;54:45–54. of autoimmune T and B cells. J 41) van Hamburg JP, Asmawidjaja PS, Immunol;162:5775–83. Davelaar N, Mus AM, Colin EM, 48) Rubin RL, Bell SA, Burlingame RW. Hazes JM, Dolhain RJ, Lubberts E. 1992. Autoantibodies associated with 2011. Th17 cells, but not Th1 cells, lupus induced by diverse drugs target a from patients with early rheumatoid similar epitope in the (H2A-H2B)- arthritis are potent inducers of matrix DNA complex. J Clin Invest;90:165– metalloproteinases and 73. proinflammatory cytokines upon 49) van Bavel CC, Dieker J, Muller S, synovial fibroblast interaction, Briand JP, Monestier M, Berden JH, including autocrine interleukin-17A van der Vlag J. 2009. Apoptosis- production. Arthritis Rheum;63:73-83. associated acetylation on histone H2B 42) Shah K, Lee WW, Lee SH, Kim SH, is an epitope for lupus autoantibodies. Kang SW, Craft J, Kang I.2010. Mol Immunol;47:511–6. Dysregulated balance of Th17 and Th1 50) Dieker JW, Fransen JH, van Bavel CC, cells in systemic lupus erythematosus. Briand JP, Jacobs CW, Muller S, Arthritis Res Ther 2010;12:R53. Berden JH, van der Vlag J. 2007. 43) Crispín JC, Oukka M, Bayliss G, Apoptosis-induced acetylation of Cohen RA, Van Beek CA, Stillman IE, histones is pathogenic in systemic Kyttaris VC, Juang YT, Tsokos GC. lupus erythematosus. Arthritis 2008. Expanded double negative T Rheum;56:1921–33. cells in patients with systemic lupus 51) Garcia-Romo GS, Caielli S, Vega B, erythematosus produce IL-17 and Connolly J, Allantaz F, Xu Z, Punaro infiltrate the kidneys. J Immunol; M, Baisch J, Guiducci C, Coffman RL, 181:8761-6. Barrat FJ, Banchereau J, Pascual V. 44) Aringer M, Feierl E, Steiner G, 2011. Netting neutrophils are major Stummvoll GH, Höfler E, Steiner CW, inducers of type I IFN production in Radda I, Smolen JS, Graninger WB. pediatric systemic lupus erythematosus. 2002. Increased bioactive TNF in Sci Transl Med;3:73ra20 human systemic lupus erythematosus: 52) Lande R, Ganguly D, Facchinetti V, associations with cell death. Frasca L, Conrad C, Gregorio J, Meller Lupus;11:102–8. S, Chamilos G, Sebasigari R, Riccieri 45) Aringer M, Graninger WB, Steiner G, V, Bassett R, Amuro H, Fukuhara S, Smolen JS. 2004. Safety and efficacy Ito T, Liu YJ, Gilliet M. 2011. of tumor necrosis factor a blockade in Neutrophils activate plasmacytoid systemic lupus erythematosus: an open- dendritic cells by releasing self-DNA-

152 peptide complexes in systemic lupus predicts prognosis in autoimmune erythematosus. Sci Transl disease. Nat Med;16:586–91. Med;3:73ra19. 60) Contin-Bordes C, Lazaro E, Richez C, 53) Holyst MM, Hill DL, Hoch SO, Jacquemin C, Caubet O, Douchet I, Hoffmann RW. 1997. Analysis of T Viallard JF, Moreau JF, Pellegrin JL, and B cell responses against U small Blanco P. 2011.Expansion of myelin nuclear ribonucleoprotein 70-kD, B autoreactive CD8+ T lymphocytes in and D polypeptides among patients patients with neuropsychiatric systemic with systemic lupus erythematosus and lupus erythematosus. Ann Rheum mixed connective tissue disease. Dis;70:868-71 Arthritis Rheum;40:1493-503. 61) Filaci G, Bacilieri S, Fravega M, 54) Smolen JS, Chused TM, Leiserson Monetti M, Contini P, Ghio M, Setti WN, Reeves JP, Alling D, Steinberg M, Puppo F, Indiveri F.2001. AD.1982. Heterogeneity of Impairment of CD8+ T suppressor cell immunoregulatory T-cell subsets in function in patients with active systemic lupus erythematosus. Am J Systemic Lupus Erythmatosus. J Med;72:783-90. Immunol;166:6452-7. 55) Bakke AC, Kirkland PA, Kitridou RC, 62) Filaci G, Fravega M, Negrini S, Quismorio FP Jr, Rea T, Ehresmann Procopio F, Fenoglio D, Rizzi M, GR, Horwitz AD.1983. T lymphocyte Brenci S, Contini P, Olive D, Ghio M, subsets in systemic lupus Setti M, Accolla RS, Puppo F, Indiveri erythematosus. Arthritis F. 2004.Nonantigen specific CD8+ T Rheum;26:745-50. suppressor lymphocytes originate from 56) Maeda N, Sekigawa I, Matsumoto M, CD8+CD28- T-cells and inhibit both Hashimoto H, Hirose S. 1999. T-cell proliferation and CTL Relationship between CD4+/CD8+ T function.Hum Immunol;65:142-56. cell ratio and T cell activation in 63) Colovai AI, Liu Z, Ciubotariu R, systemic lupus erythematosus. Scand J Lederman S, Cortesini R, Suciu-Foca Rheumatol;28:166-70 N.2000. Induction of xenoreactive 57) Blanco P, Pitard V, Viallard JF, et al. CD4-T cell anergy by suppressor 2005. Increase in activated CD8+ T CD8+CD28- Tcells. Transplantation; lymphocytes expressing perforin and 69:1304-10. granzyme B correlates with disease 64) Najafian N, Chitnis T, Salama AD, Zhu activity in patients with systemic lupus B, Benou C, Yuan X, Clarkson MR, erythematosus. Arthritis Rheum;52: Sayegh MH, Khoury SJ.2003. 201–11. Regulatory functions of CD8+CD28- T 58) Couzi L, Merville, Deminière C, cells in an auoimmune disease mode. J Moreau JF, Combe C, Pellegrin JL, Clin Invest;112:1037-48. Viallard JF, Blanco P.2007. 65) Tulunay A, Yavuz S, Direskeneli H, Predominance of CD8+ T lymphocytes Eksioglu-Demiralp E. 2008. among periglomerular infiltrating cells CD8+CD28-, suppressive T cells in and link to the prognosis of class III systemic lupus erythematosus. and class IV lupus nephritis. Arthritis Lupus;17:630-7. Rheum;56:2362–70. 66) Alvarado-Sánchez B, Hernández- 59) McKinney EF, Lyons PA, Carr EJ, Castro B, Portales-Pérez D, Baranda L, Hollis JL, Jayne DR, Willcocks LC, Layseca-Espinosa E, Abud-Mendoza Koukoulaki M, Brazma A, Jovanovic C, Cubillas-Tejeda AC, González- V, Kemeny DM, Pollard AJ, Macary Amaro R. 2006. Regulatory T cells in PA, Chaudhry AN, Smith KG. 2010. A patients with systemic lupus CD8+ T cell transcription signature erythematosus. J Autoimmun;27:110-8.

153 67) Vallejo AN. 2005.CD28 extinction in diagnostic and prognostic tool for human T cells: altered functions and rheumatoid arthritis. Ann Rheum Dis; the program of T-cell senescence. 64:1731-6. Immunol Rev;205:158-69. 74) Trembleau S, Hoffmann M, Meyer B, 68) Houssiau FA, Lefebvre C, Vanden Nell V, Radner H, Zauner W, Hammer Berghe M, Lambert M, Devogelaer JP, J, Aichinger G, Fischer G, Smolen J, Renauld C.1995.Serum IL-10 titres in Steiner G.2010. Immunodominant T- systemic lupus erythematosus reflect cell epitopes of hnRNP-A2 associated disease activity. Lupus;4:393-5. with disease activity in patients with 69) Llorente L, Richaud-Patin Y, Fior R, rheumatoid arthritis. Eur J Alcocer-Varela J, Wijdenes J, Fourrier Immunol;40:1795-808. BM, Galanaud P, Emilie D.1994. In 75) Ghia P, Prato G, Stella S, Scielzo C, vivo production of interleukin-10 by Geuna M, Caligaris-Cappio non-T cells in rheumatoid arthritis, F.2007.Age-dependent accumulation of Sjogren's syndrome, and systemic monoclonal CD4+CD8+ double lupus erythematosus. Arthritis positive T lymphocytes in the Rheum;37:1647-56. peripheral blood of the elderly. Br J 70) Llorente L, Zou W, Levy Y, Richaud- Haematol;139:780–90. Patin Y, Wijdenes J, Alcocer-Varela J, 76) Martinez-Gallo M, Puy C, Ruiz- Morel-Fourrier B, Brouet JC, Alarcon- Hernandez R, Rodriguez-Arias JM, Segovia D, Galanaud P, Emilie Bofill M, Nomdedeu JF, Cigudosa JC, D.1995. Role of interleukin-10 in the B Rodriguez-Sanchez JL, de la Calle- lymphocyte hyperactivity and Martin O. 2008. Severe and recurrent autoantibody production of human episodes of bronchiolitis obliterans systemic lupus erythematosus. J Exp organising pneumonia associated with Med;181:839-44. indolent CD4+CD8+ T-cell leukaemia. 71) Skriner K, Sommergruber K, Tremmel Eur Respir J;31:1368–72. WH, Fischer V, Barta A, Smolen JS 77) Parel Y, Aurrand-Lions M, Scheja A, and Steiner G.1997. Anti-A2/RA33 Dayer JM, Roosnek E, Chizzolini autoantibodies are directed to the RNA C.2007. Presence of CD4+CD8+ binding region of the A2 protein of the double-positive T cells with very high heterogeneous nuclear interleukin-4 production potential in Ribonucleoprotein complex. lesional skin of patients with systemic Differential epitope recognition in sclerosis. Arthritis Rheum; 56: 3459– rheumatoid arthritis, systemic lupus 67. erythematosus, and mixed connective 78) Parel Y, Chizzolini C.2004. tissue disease. J. Clin. Invest;100:127– CD4+CD8+ double positive (DP) T 35. cells in health and disease. Autoimmun 72) Hirota K, Hashimoto M, Yoshitomi H, Rev; 3: 215–20. Tanaka S, Nomura T, Yamaguchi T, 79) Chizzolini C, Parel Y, De Luca C, Iwakura Y, Sakaguchi N, Sakaguchi Tyndall A, Akesson A, Scheja A, S.2007. T cell self-reactivity forms a Dayer JM.2003.Systemic sclerosis Th2 cytokine milieu for spontaneous cells inhibit collagen production by development of IL-17+ Th cells that dermal fibroblasts via membrane- cause autoimmune arthritis. J Exp associated TNF-alpha. Arthitis Med.;204:41-47. Rheum;48:2593–604. 73) Nell VP, Machold KP, Stamm TA, 80) De Maria A, Malnati M, Moretta A, Eberl G, Heinzl H, Uffmann M, Pende D, Bottino C, Casorati G, Smolen JS, Steiner G.2005. Cottafava F, Melioli G, Mingari MC, Autoantibody profiling as early Migone N, Romagnani S, Moretta L

154 1987. CD3q4-8- WT31-(T cell receptor in the endoplasmic reticulum. Nat gammaq) cells and other unusual Immunol;6:689-97. phenotypes are frequently detected 87) Cui X, Rouhani FN, Hawari F, Levine among spontaneously interleukin 2- SJ. 2003.Shedding of the type II IL-1 responsive T lymphocytes present in decoy receptor requires a the joint fluid in juvenile rheumatoid multifunctional aminopeptidase, arthritis. A clonal analysis. Eur J aminopeptidase regulator of TNF Immuno ;17:1815–19. receptor type 1 shedding. J 81) Mizutani H, Katagiri S, Uejima K, Immunol;171:6814-9. Ohnishi M, Tamaki T, Kanayama Y, 88) Kilding R, Iles MM, Timms JM, Tsubakio T, Kurata Y, Yonezawa T, Worthington J, Wilson AG. Tarui S.1985.T-cell abnormalities in 2004.Additional genetic susceptibility patients with idiopathic for rheumatoid arthritis telomeric of the thrombocytopenic purpura: the DRB1 locus. Arthritis Rheum.;50:763- presence of OKT4q8q cells. Scand J 9. Haematol;35:233 –9. 89) Mulcahy H, O'Rourke KP, Adams C, 82) Von Delwig A, Locke J, Robinson JH, Molloy MG, O'Gara F. 2006.LST1 and Ng W. 2010. Response of Th17 cells to NCR3 expression in autoimmune a citrullinated arthritogenic aggrecan inflammation and in response to IFN- peptide in patients with rheumatoid gamma, LPS and microbial infection. arthritis. Arthritis Rheum;62:143–9 Immunogenetics;57:893-903. 83) Snir O, Rieck M, Gebe JA, Yue BB, 90) van der Pouw Kraan TC, Wijbrandts Rawlings CA, Nepom G, Malmström CA, van Baarsen LG, Voskuyl AE, V, Buckner JH.2011.Identification and Rustenburg F, Baggen JM, Ibrahim functional characterization of T cells SM, Fero M, Dijkmans BA, Tak PP, reactive to citrullinated-vimentin in Verweij CL. 2007.Rheumatoid arthritis HLA-DRB1*0401 humanized mice subtypes identified by genomic and RA patients. Arthritis profiling of peripheral blood cells: Rheum;63:2873-83. assignment of a type I interferon 84) Law SC, Street S, Yu CH, Capini C, signature in a subpopulation of Ramnoruth S, Nel HJ, van Gorp E, patients. Ann Rheum Dis;66:1008–14. Hyde C, Lau K, Pahau H, Purcell AW, 91) Raterman HG, Vosslamber S, de Thomas R. 2012.T-cell autoreactivity Ridder S, Nurmohamed MT, Lems to citrullinated autoantigenic peptides WF, Boers M, van de Wiel M, in rheumatoid arthritis patients carrying Dijkmans BA, Verweij CL, Voskuyl HLA-DRB1 shared epitope alleles. AE. 2012.The interferon type I Arthritis Res Ther;17;14:R118. signature towards prediction of non- 85) Fierabracci A, Milillo A, Locatelli F, response to rituximab in rheumatoid Fruci D. 2012.The putative role of arthritis patients. Arthritis Res endoplasmic reticulum Ther;14:R95. aminopeptidases in autoimmunity: 92) Thurlings RM, Boumans M, Tekstra J, insights from genomic-wide van Roon JA, Vos K, van Westing association studies. Autoimmun DM, van Baarsen LG, Bos C, Kirou Rev;12:281-8. KA, Gerlag DM, Crow MK, Bijlsma 86) Saveanu L, Carroll O, Lindo V, Del JW, Verweij CL, Tak PP. Val M, Lopez D, Lepelletier Y, Greer 2010.Relationship between the type I F, Schomburg L, Fruci D, Niedermann interferon signature and the response to G, van Endert PM.2005. Concerted rituximab in rheumatoid arthritis peptide trimming by human ERAP1 patients. Arthritis Rheum;62:3607-14. and ERAP2 aminopeptidase complexes

155 93) Vosslamber S, Raterman HG, van der BA, Tak PP, Verweij CL. 2010. Pouw Kraan TC, Schreurs MW, von Regulation of IFN response gene Blomberg BM, Nurmohamed MT, activity during infliximab treatment in Lems WF, Dijkmans BA, Voskuyl AE, rheumatoid arthritis is associated with Verweij CL. 2011. Pharmacological clinical response to treatment. Arthritis induction of interferon type I activity Res Ther;12:R11 following treatment with rituximab 97) Hakonarson H, Bjornsdottir US, Halapi determines clinical response in E, Bradfield J, Zink F, Mouy M, rheumatoid arthritis.Ann Rheum Helgadottir H, Gudmundsdottir AS, Dis;70:1153-9. Andrason H, Adalsteinsdottir AE, 94) Mavragani CP, La DT, Stohl W, Crow Kristjansson K, Birkisson I, Arnason T, MK. 2010. Association of the response Andresdottir M, Gislason D, Gislason to tumor necrosis factor antagonists T, Gulcher JR, Stefansson K. 2005. with plasma type I interferon activity Profiling of genes expressed in and interferon-beta/alpha ratios in peripheral blood mononuclear cells rheumatoid arthritis patients: a post hoc predicts glucocorticoid sensitivity in analysis of a predominantly Hispanic asthma patients. Proc Natl Acad Sci U cohort. Arthritis Rheum;62:392-401. S A;102:14789-94. 95) Palucka AK, Blanck JP, Bennett L, 98) Spijkers-Hagelstein JA, Schneider P, Pascual V, Banchereau J. 2005.Cross- Hulleman E, de Boer J, Williams O, regulation of TNF and IFN-alpha in Pieters R, Stam RW. Elevated autoimmune diseases. Proc Natl Acad S100A8/S100A9 expression causes Sci U S A.;102:3372-7. glucocorticoid resistance in MLL- 96) van Baarsen LG, Wijbrandts CA, rearranged infant acute lymphoblastic Rustenburg F, Cantaert T, van der leukemia. Leukemia. 2012;26:1255-65. Pouw Kraan TC, Baeten DL, Dijkmans

156

Nederlandse samenvatting

157 Deze thesis beschrijft verschillende aspecten van T-cellen in het afweersysteem van gezonde mensen en patiënten met reumatologische ziekten, met name Systemische lupus erythematosus (SLE) en Reumatoide artritis (RA). Het eerste deel betreft de ontwikkeling van T-cellen. Het tweede deel onderzoekt T-cellen, die ziekte-specifieke antigenen (doel-eiwitten van T-cellen en antistoffen) herkennen, zoals histonen in SLE and hnRNP-A2 in SLE en RA. Het derde deel laat zien welke veranderingen in T-cellen door de therapie met glucocorticoiden (bijvoorbeeld Prednison) veroorzaakt worden in RA-patiënten.

1) Geheugen T-cellen in ziekte en geheugen T-cellen (TEM) te verdelen. Van gezondheid TCM wordt aangenomen dat zij terugkeren naar de secundaire lymfoïde organen met T-cellen zijn onontbeerlijk voor het behulp van CCR7, terwijl TEM niet naar opbouwen van ons immunologisch geheugen, lymfoïde organen maar naar en zonder dit geheugen kan ons ontstekingsgebieden migreren en daar hun immuunsysteem niet goed functioneren. Als functie uitoefenen door cytokines uit te dit systeem wordt verstoord kunnen stoten (Figuur 1, introduction). Door CCR7 verschillende soorten ziekten ontstaan, van en CD27 (een ander eiwit, dat op de immuundeficiënties tot allergieën of auto- oppervlakte van sommige immuuncellen, immuunziekten. Het ontrafelen van de waaronder T-cellen, gevonden wordt) te verschillende stappen in het rijpingsproces combineren konden we dit van T-cellen is daarom van groot belang om ontwikkelingsproces preciezer specificeren de afwijkingen daarin te kunnen beïnvloeden, (Figuur volgende pagina). voordat deze tot specifieke ziekten kunnen Onze bevindingen laten zien dat CD4+ T- leiden. Het karakteriseren van de cellen rijpen van naïeve cellen naar het verschillende stappen in het rijpingsproces CCR7+CD27+ stadium, en daarna naar het van CD4+ T-cellen, de zogenoemde T CCR7-CD27+ stadium voordat zij hun helper-cellen (Th), is het onderwerp van het oppervlakte-expressie van CD27 verliezen en eerste deel van dit proefschrift. T helper- CCR7-CD27- worden. We konden dit cellen bereiken hun effect vooral door het ontwikkelingspad bevestigen door deze uitstoten van immunologisch werkzame verschillende subsets op te kweken en hun oplosbare eiwitten (cytokines). Met de fenotype te bepalen na verschillende cycli verschillen in samenstelling van deze van stimulatie en rust. Waar de CCR7-CD27- eiwitten kunnen de Th worden ingedeeld in T-cellen hun fenotype behielden, Th1, Th2 en Th17 –cellen. De resultaten van ontwikkelden de andere subsets zich via het het onderzoek naar het rijpingsproces in de bovenbeschreven ontwikkelingspad. CD4+ T-cellen in gezondheid en (auto- Bepaling van de telomeerlengte, een marker immuun) ziekten vinden hun neerslag in van veroudering, die afneemt naarmate de cel Hoofdstukken 2 en 3. ouder wordt, leverde verdere onderbouwing In Hoofdstuk 2 hebben we de kennis en voor deze hypothese, aangezien de inzicht over de ontwikkelingsstadia van telomeerlengte gestaag afnam van naïeve CD4+ T-cellen verder uitgebreid door cellen → CCR7+CD27+ → CCR7-CD27+ gebruik te maken van 2 cel-oppervlakte → CCR7-CD27-. Deze werd vergezeld door eiwitten: CCR7 en CD27. CCR7 is een een afname in induceerbare telomerase chemokine receptor, die het gericht opzoeken activiteit. Telomerase is het enzym dat (homing) van de secundaire lymfoïde verantwoordelijk is voor het tegengaan van organen door immuuncellen reguleert. De telomeerlengte verlies. Ook de proliferatieve expressie van CCR7 maakt het mogelijk capaciteit (de capaciteit om zich te delen) is geheugen T-cellen in CCR7+ centrale een kenmerk van celrijping die afneemt als cellen verouderen. Zoals verwacht, nam deze geheugen cellen (TCM) en CCR7- effector

158 af naarmate cellen het ontwikkelingspad CCR7 expressie. Aan de andere kant was de verder volgden. Interessant was, dat CCR7- TEM populatie, gedefinieerd als CCR7-, niet T-cellen geneigd waren apoptose homogeen, en de onderverdeling in CCR7- (geprogrammeerde celdood) te ondergaan /CD27+ en CCR7-/CD27- cellen toonde nadat ze gestimuleerd werden, hetgeen in verschillende patronen van cytokine-uitstoot. naïeve of CCR7+ cellen nauwelijks werd Alle data gepresenteerd leiden in Hoofdstuk gezien. Onderzoek naar de functie van de 2 leiden tot een preciezere beschrijving van verschillende cel-subsets met betrekking tot de rijpingsstadia van CD4+ T-cellen. In een hun cytokine productie liet een duidelijk later verschenen onderzoek (21) werd ons verschil tussen de enkele T-cel subsets zien. voorgesteld model bevestigd en een De voor Th1 en Th2 specifieke cytokines mogelijke verklaring voor de verhoogde waren verhoogd in respectievelijk de CCR7- neiging tot apoptose in CCR7- CD4+ T- CD27+ en de CCR7-CD27- subsets. Dit past cellen gevonden, die afhankelijk is van de bij de conclusie dat de uitstoot van “effector” verdeling van Fas (de “receptor des doods”) cytokines geassocieerd is met het verlies van in de celmembranen van deze cel-subsets.

Figuur stelt het door ons vastgestelde rijpingsproces van CD4+ T-cellen voor. Na blootstelling aan een antigeen worden de naïeve T-cellen geheugen T-cellen, hetgeen wordt gemarkeerd door de vervanging van CD45RA door CD45RO op hun celoppervlak. De oppervlakte markers CCR7 en CD27 verdwijnen langzaam naarmate de cellen door verdere stimulatie rijpen. De cellen zijn tezamen met hun preferentieel uitgescheiden cytokines afgebeeld.

In SLE is de disbalans tussen functionele T- Het ontwikkelingspad dat in gezonde cel subsets het onderwerp van een groot controles (HC, healthy control, Figuur) aantal studies, die afwisselend een gezien werd, was ook in SLE-patiënten verschuiving naar Th1, Th2 of Th17 T-cel aanwezig, hetgeen met experimenten analoog subsets laten zien. (4-6) Daarnaast zijn er aan de proeven in Hoofdstuk 2, kon worden intrinsieke abnormaliteiten van T-cellen in bevestigd. Er waren echter cruciale SLE bekend. (7) Er is echter weinig aandacht verschillen bij SLE-patiënten. Ten eerste was voor het algemeen functioneren van T-cellen er een duidelijke toename van de CCR7- T- in SLE geweest, zoals dat in hun cel subsets in SLE patiënten vergeleken met differentiatie status wordt gereflecteerd. HC of ziekte-controles. Dit verschil leek niet Daarom hebben we in Hoofdstuk 3, beïnvloed door ziekte-activiteit of medicatie, voortbouwend op Hoofdstuk 2, de maar SLE-specifiek te zijn. Ten tweede rijpingspaden van T-cellen in SLE bekeken. waren de CCR7- cellen van SLE-patiënten

159 meer geneigd tot apoptose na stimulatie dan essentieel voor het ontstaan van auto- dezelfde T-cel-subsets van gezonde mensen. immuun ziekten beschouwd. Daarom is het Dit uitte zich ook in een significant 2e deel van mijn proefschrift gewijd aan het verminderde proliferatieve capaciteit en identificeren van de antigenen en functionele overleving van CCR7-CD27- T-cellen eigenschappen van antigeen-specifieke T- afkomstig van SLE-patiënten. Ten derde was cellen, gericht tegen belangrijke auto- de gemiddelde induceerbare telomerase antigenen in SLE (Hoofdstuk 4 en 5) en RA activiteit in de SLE naïeve CD4+ en (Hoofdstuk 6). CCR7+CD27+ geheugen T-cellen in In Hoofdstuk 4 en Hoofdstuk 5 wordt de vergelijking tot die van gezonde donoren (p< auto-reactiviteit van T-cellen in SLE 0.02) verlaagd, terwijl de telomeerlengte onderzocht. De aanwezigheid van auto- tussen de verschillende subsets van SLE- antilichamen (tegen bv dsDNA) zijn een patiënten en HC niet verschilde. Op deze belangrijk kenmerk van SLE. Deze kunnen wijze konden we het rijpingspad zoals we dat tegen verschillende celkern-componenten in Hoofdstuk 2 voor gezonde mensen gericht zijn, soms tegen volledige voorgesteld hebben, in Hoofdstuk 3 nucleosomen (complex van DNA en verifiëren voor SLE-patiënten en daarbij geassocieerde eiwit, ook wel histon meerdere verschillen in verdeling, cel genoemd), maar ook tegen delen daarvan. In overleving en proliferatie capaciteit Hoofdstuk 4 zijn de histonen de onderzochte aantonen. auto-antigenen, vervolgens wordt in De grotere percentage aan CCR7- T-cellen in Hoofdstuk 5 de cellulaire activiteit tegen SLE is interessant, en suggereert een hnRNP-A2, een component van het opeenhoping van meer gerijpte cellen in het spliceosome (zie onder) beschreven. Deze bloed van SLE-patiënten. Of dit veroorzaakt laatste is ook het onderwerp van studie bij wordt door een intrinsieke lagere drempel RA-patienten, beschreven in Hoofdstuk 6. van SLE T-cellen om uit te rijpen, of door Opvallend genoeg zijn antilichamen tegen een frequentere stimulatie door vaker histonen een van de vroegst gedecteerde aanwezige auto-antigen in SLE, moet nog auto-antilichamen (8) in SLE; vooral de worden bepaald. humorale (= gemedieerd door antistoffen) Omdat de door ons gevonden verschillen niet respons tegen H1 correleert met SLE ziekte- afhankelijk waren van ziekte-activiteit of activiteit (9) en lijkt meer SLE-specifiek dan medicatie-gebruik, kunnen onze resultaten antilichamen tegen core-histonen (= H2A, niet direct vertaald worden naar de klinische H2B, H3, H4). (9,10) Histonen zijn ook praktijk, bijvoorbeeld als alternatieve marker geïdentificeerd als de belangrijkste T-cel voor ziekte-activiteit. Daarentegen tonen antigenen in SLE, echter ontbreken de details onze bevindingen een versnelde rijping van van deze histon-specifieke cellulaire respons. cellen in SLE, hetgeen immunologische Het is vooral onduidelijk welke histonen activiteit suggereert, ondanks een klinisch voor de T-cel activatie belangrijk zijn. rustige ziekte. Dit kan wijzen op de noodzaak Daarom onderzochten we in Hoofdstuk 4 de van immunosuppressieve behandeling zelfs T-cel activiteit tegen de verschillende core- tijdens periodes van remissie. histonen (H2A, H2B, H3 en H4) en tegen H1. We onderzochten het verschil tussen gezonde controles en SLE patiënten in 2) Auto-reactieve T-cellen in SLE en proliferatie en cytokine uitstoot van RA. mononucleaire cellen in het perifere bloed (PBMC, voor een groot deel T-cellen), na Naast de veranderde samenstelling van het T- incubatie met enkelvoudige histonen. Het cel compartiment wordt ook de auto- duidelijkste verschil werd gezien in de reactiviteit, het herkennen van lichaamseigen reactiviteit tegen H1, die duidelijk verhoogd doel-eiwitten, van de T-cel populatie als was in SLE-patiënten. Dit geldt zowel voor

160 de frequentie van voorkomen als in van Th1 cellen en het door hen heftigheid, uitgedrukt in de Stimulatie Index geproduceerde TNF in SLE. Het (SI). Ook H2A wekte een T-cel respons op therapeutisch effect van TNF-blokkade in die meer uitgesproken was in SLE patiënten, SLE-patiënten, dat in kleine studies tot maar die niet significant vaker optrad. Alle klinische verbetering leidde, zou dus andere histonen lieten een vergelijkbare tenminste deels op de bestrijding van deze respons in zowel frequentie als heftigheid H1 specifieke T-cellen kunnen berusten. zien bij HC en SLE-patiënten. Het is Nu zijn histonen niet de enige auto-antigenen opmerkelijk dat alleen in SLE-patiënten en die bij SLE herkend worden. In 20-30% van niet in HC de cellulaire respons tegen H1 in SLE-patiënten kunnen auto-antilichamen het algemeen correleerde met de respons tegen hnRNP-A2, een component van het opgewekt door de andere histonen. De spliceosome, worden herkend. (13) De proliferatieve respons werd vergezeld door functie van hnRNP-A2 bestaat onder andere productie van de cytokines IFNγ en TNFα bij uit het reguleren van verschillende stappen in zowel patiënten als ook bij gezonde het proces tussen genetische codering en controles, hetgeen een Th1 respons tegen uiteindelijk eiwitproductie, zoals alternatieve histonen suggereert. Histon H1 induceerde de splicing, transport van mRNA, en het hoogste TNFα respons. Het zelfde patroon reguleren van translatie. Er zijn slechts een van cytokine secretie werd gezien voor alle paar beschrijvingen van de T-cel respons H1-specifieke T cel klonen (TCC) afkomstig tegen spliceosomale antigenen beschikbaar. van 1 SLE-patiënt en van HC. Alle TCC Een daarvan beschreef hnRNP-A2 specifieke waren CD4+ T-cellen, produceerden een Th1 CD4+ T-cel klonen (TCCs) in patiënten met cytokine patroon en reageerden exclusief met bindweefselziekten, waaronder SLE, en één specifieke histon, en niet met alle andere schreef een mogelijk pathogene rol toe aan histonen. Alleen H1 toonde een correlatie deze T-cellen in SLE. De mogelijke tussen cellulaire en antistof-reactiviteit. De verschillen in de hnRNP-A2 specifieke T- capaciteit van H1-specifieke T-cellen om een cellen tussen patiënten en gezonde controles (specifieke) antilichaam respons te induceren (HC) zijn voorheen niet onderzocht en waren werd getest in SLE- en RA- patiënten en in onderwerp van studie in Hoofdstuk 5. gezonde controles. Na blootstelling aan H1 We vonden een verschil in zowel frequentie nam de totale antistof productie alleen toe in van voorkomen als in grootte van de kweken afkomstig van SLE-patiënten cellulaire respons tegen hnRNP-A2 tussen (p=0.02 in vergelijking met HC kweken). SLE-patiënten en HC. Waar 66% van de Daarnaast resulteerde H1 stimulatie alleen bij SLE-patiënten een positieve T cel respons SLE-patiënten in een significante toename lieten zien (waarbij meer dan 50% met een van specifieke anti-histon antilichaam Stimulatie index SI>4 reageerden), liet niveaus ten opzichte van de uitgangswaarden slechts 24% van de gezonden een positieve (p<0.04); deze niveaus waren dan ook reactie zien (slechts 5% met een SI>4). De significant hoger dan in met H1 gemiddelde SI van SLE-patiënten was met gestimuleerde HC kweken (p=0.02). 6.7 ± 2.3 veel hoger dan in de controle groep Alles tezamen konden we hiermee aantonen (gemiddelde SI 2.3 ± 0.2; p < 0.0002). Er dat H1 het belangrijkste cellulaire auto- was geen correlatie met de hnRNP-A2 antigeen is in SLE-patiënten en ten grondslag specifieke antistof respons. Verder onderzoek ligt aan de antistof-respons tegen chromatine van hnRNP-2A specifieke T-cel klonen en celkern antigenen. Bovendien werd (TCC) naar fenotype, cytokine productie en vastgesteld dat deze H1-reactieve T-cellen tot specificiteit liet overeenkomsten zien in TCC het Th1 cel subtype behoren. Dit onderstreept afkomstig van respectievelijk HC en SLE aan de ene kant de relevantie van H1 bij het patiënten, met ongeveer 2/3 CD4+ T-cellen. ontstaan van de ziekte. Aan de andere kant Maar waar alle onderzochte TCC van HC het bieden onze bevindingen steun voor de rol co-stimulatoire eiwit CD28 droegen (CD28+

161 waren), ontbrak CD28 expressie bij de Omdat auto-antilichamen tegen hnRNP-A2 meerderheid van SLE-TCC. De cellen, die in ongeveer een derde van RA-patiënten niet CD4+ waren, droegen CD8 aan hun gevonden worden en deze een marker voor oppervlakte, en waren dus CD8+ , vroege ziekte zijn (16) onderzochten we in zogenoemde cytotoxische T-cellen. Ook in Hoofdstuk 6 de cellulaire activiteit tegen deze cellen was de CD28 expressie hnRNP-A2 in RA-patiënten. We vonden dat verschillend tussen HC en SLE patiënten. De een meerderheid van RA patiënten (58%) op CD8+CD28- TCC afkomstig van SLE cellulair niveau reageerden met dit auto- patiënten verschilden daarnaast van zowel antigeen, in tegenstelling tot slechts 20% van CD4+CD28- cellen van SLE patiënten als de gezonden controles. Net als bij de SLE- van CD8+CD28+ TCC van HC in secretie patiënten lieten ook RA patiënten een van cytokines en produceerden nauwelijks significant heftiger reactie tegen hnRNP-A2 IFNγ, maar wel grote hoeveelheden IL-10. In zien dan de controles, en deze respons was het algemeen produceerden de TCC geen IL- opnieuw niet geassocieerd met de antistof- 4, en de meeste scheidden IFNγ uit, wijzend respons. Het patroon van cytokine uitstoting op een Th1 respons tegen hnRNP-A2. Daar dat we zowel in RA als in HC zagen, was een de CD8+CD28- SLE TCC een uitgesproken Th1 respons, en onderscheidde zich bij RA- cytokine profiel (veel IL-10) hadden, patiënten alleen door een hogere IL-2 onderzochten we de mogelijkheid van een productie. Een positieve T-cel respons zagen unieke, regulerende functie van deze cellen wij ook in monsters van RA synoviaal- op andere immuun cellen. De cellen hadden vloeistof, die over het algemeen hoger was echter een stimulerend effect op proliferatie dan de respons in het perifeer bloed, hetgeen van andere immuuncellen (CD4+ T-cellen), een voorkeurslocalisatie in de gewrichten vergelijkbaar met het effect van de CD4+ voor deze specifieke T-cellen impliceert. CD28+ T-cellen, waar dit effect verwacht Daarnaast waren T-cel klonen (TCC) tegen werd, en in contrast met het effect van hnRNP-A2 wederom Th1-cellen. In contrast gepurificeerd IL-10. met de hnRNP-A2 TCC van SLE patiënten, Zo zagen we, net als bij de cellulaire liet fenotypische evaluatie van TCC vrijwel activiteit tegen histonen, een duidelijk altijd CD4+ T-cellen in RA zien, terwijl HC onderscheid in cellulaire activiteit tegen ook CD8+ TCC aanmaakten. Over het hnRNP-A2 bij SLE-patiënten vergeleken met algemeen was de IFNγ productie hoger in HC en konden we deze als vooral Th1-cellen TCC van RA-patiënten, hoewel uit het bloed identificeren. Daarnaast identificeerden we van één HC 3 CD4+CD8+ TCC ontstonden, een mogelijke nieuwe subset van niet- die grote hoeveelheden IFNγ produceerden. suppressieve, IL-10 uitscheidende Ook stelden wij ons ten doel de CD8+CD28- T-cellen tegen hnRNP-A2, die onderliggende T-cel reactiviteit tegen andere alleen gezien werd bij SLE patiënten. So wie belangrijke RA- geassocieerde auto- so was het hoge percentage CD8+ T-cel antigenen, te weten gecitrullineerde peptiden, klonen in SLE opmerkelijk, en ondersteunt in kaart te brengen. Antistoffen tegen eerdere rapportages van een disbalans tussen gecitrullineerde peptiden (ACPA) zijn CD4+ en CD8+ T-cellen in SLE, ten gunste specifiek voor RA. Een van de antigenen van CD8+ T-cellen. Daar de meeste CD8+ voor de ACPA is gecitrullineerd filaggrine. TCC van SLE echter ook CD28- waren, lijkt Om meer te leren over de T-cel respons die dit een eigen, in SLE vaker voorkomende, mogelijk aan de ACPA ten grondslag ligt, subpopulatie te zijn. Nader onderzoek naar onderzochten we de cellulaire respons van deze cel populatie zou opheldering kunnen primaire kweken tegen gecitrullineerd brengen, of hun functie in het lichaam een filaggrine (cFil) en filaggrin in zijn originele, ontstekings-remmende of juist een niet gecitrullineerde vorm. We vonden geen onstekings-bevorderende is. substantieel verschil tussen HC en RA patiënten in herkenning van gecitrullineerde

162 of niet-gecitrullineerde antigenen. Omdat worden en welke genen onderdrukt worden. filaggrine niet het enige eiwit is, waar ACPA Omdat het testen van alle genen tegelijk veel op reageren, suggereren onze bevindingen, (statistische) ruis kan veroorzaken, hebben dat niet gecitrullineerd filaggrine, maar een we onze bevindingen door qPCR (waarmee ander gecitrullineerd eiwit de cellulaire slechts de activiteit van enkele genen getest reactiviteit induceert, en dat die dan zelf weer worden) bevestigd, zoals beschreven in de antistof (ACPA) productie aanzet. Hoofdstuk 7 en Hoofdstuk 8. Inderdaad zijn desbetreffend zowel De directe vergelijking van genexpressie gecitrullineerd aggrecan als ook profielen tussen GC-gevoelige en GC- gecitrullineerd vimentine als antigenen voor ongevoelige patiënten voorafgaand aan een T-cellen in RA beschreven. (82,83) stootkuur met Glucocorticoiden leverde een In Hoofdstukken 4-6 konden we de aantal genen in zowel T-cellen als onderliggende cellulaire reactiviteit tegen monocyten op, waarvan voorheen geen rol in belangrijke auto-antigenen in SLE en RA de werking van GC bekend was. identificeren en karakteriseren. We lieten Het meest opvallende verschil was de zien dat histon H1 het belangrijkste doelwit hogelijk significante, meer dan 4-voudige is van de cellulaire respons die ten grondslag toegenomen expressie van ERAP2 bij GC- ligt aan de pathognomonische antilichaam gevoelige patiënten in vergelijking met niet- reactie tegen dsDNA. Daarnaast toonden we gevoelige patiënten in zowel CD4+ T-cellen aan hoe belangrijk hnRNP-A2 is als cellulair als monocyten voor behandeling. In auto-antigeen in SLE en RA. Het is monocyten was ook een ander gen, TRIM33, opmerkelijk dat de grote meerderheid van T- significant meer tot expressie gebracht in cellen van het Th1 type was, onafhankelijk GC-gevoelige patiënten en in CD4+ T-cellen van antigeen of ziekte. Daarnaast ontdekten gold dit voor de genen FAM26F en LST1. we een SLE specifieke T-cel subset met een Interessant genoeg hield het verschil in afwijkend fenotype die reageerde met genexpressie van ERAP2 24 uur na hnRNP-A2. Alles bij elkaar geven onze data behandeling aan. Daarnaast correleerde de verder gewicht aan de rol van T-cellen in genexpressie van ERAP2 voor therapiestart deze 2 auto-immuunziekten. met ziekteactiviteit na behandeling, hetgeen de predictieve waarde van een verhoogde expressie voor een goed GC-respons verder 3) T-cellen in RA en de reactie op onderstreept. Daarmee konden we in dit Glucocorticoiden. hoofdstuk meerdere mogelijke voorspellers van GC-gevoeligheid identificeren, waarvan CD4+ T-cellen zijn niet alleen belangrijk in ERAP2 de meest belangrijke leek te zijn. De de ontwikkeling van RA, maar zijn ook een bekende functie van ERAP2 is het op de belangrijk doelwit van het klassieke juiste maat knippen van eiwitstukken voor de immunosuppressieve medicijn bij deze presentatie aan (CD8+) T-cellen ziekte, glucocorticoiden (GC). Hoewel GC (antigenpresentatie). Het zou echter ook nog meestal zeer effectief zijn, reageert ongeveer zo kunnen zijn, dat ERAP2. verschillende een derde van de patiënten onvoldoende op receptoren van ontstekings-cytokinen van het behandeling. Dit leidde ons ertoe te cek-oppervlakte zou kunnen knippen. Dit is onderzoeken welke verschillen er in een bekende functie van ERAP1, het gen, genexpressie van CD4+ T-cellen en waarvan ERPA2 een duplicatie is. Daarmee monocyten bestaan tussen GC-gevoelige en zou ERAP2 ontstekingen kunnen reguleren. GC-ongevoelige patiënten. Om verschillen in Verder werd ERAP2 al eerder in sommige genexpressie aan te tonen gebruikten we auto-immuunziekten gevonden als cDNA microarrays. Hiermee kunnen alle predisponerende factor. In de toekomst zijn genen in een cel tegelijk bekeken worden, en verdere studies noodzakelijk om het nut van het wordt nagegaan, welke genen geactiveerd al deze voorspellers in de daagse klinische

163 praktijk te bevestigen. Ook de precieze rol autoimmunziekten aangetoond, waaronder van ERAP2 in RA en in de respons op RA en SLE. We konden de bij de micro- therapie moet nog worden bepaald. arrays gevonden sterkere onderdrukking van Om mogelijke verschillen in het IFN-genen bij de niet-gevoelige patiënten werkingsmechanisme van GC in relatie met bevestigen in CD4+ cellen middels qPCR. de klinische respons te identificeren, hebben Daarnaast hebben we meerdere we de verandering in genexpressie patronen transcriptiefactoren gevonden in beide 24 uur na de stootkuur (ten opzichte van voor celtypes van GC-ongevoelige patiënten, die de stootkuur) onderzocht bij de GC- in staat zijn meerdere genen in de gevoelige en bij de GC-ongevoelige bovenbeschreven genenlijsten te patiënten met RA. Dit is beschreven in beïnvloeden. In CD4+ cellen vonden we Hoofdstuk 8. Behandeling met GC leidde bij LEF1 en SP1, terwijl in monocyten de uitsluitend de GC-gevoelige patiënten tot de transcriptiefactoren LEF1, NFAT-1, IRF-1 veranderde expressie van 19 individuele en NFκB1 naar boven kwamen. Dit is een genen in CD4+ T-cellen (48 in monocyten), indicatie dat de processen die ten grondslag en bij uitsluitend ongevoelige patiënten tot liggen aan het niet gevoelig zijn voor GC veranderde expressie van 104 genen (253 in door een specifiek transcriptie-factor-netwerk monocyten). Voor 18 genen veranderde de bestuurd kunnen worden. expressie zowel in de gevoelige als in de Samenvattend hebben we in Hoofdstuk 8 een ongevoelige RA-patiënten in CD4+T-cellen verschil in GC-effecten in CD4+ T-cellen (104 genen in monocyten). De volgende stap van GC-gevoelige en ongevoelige patiënten was, om in deze grote hoeveelheid genen een kunnen onderscheiden, hetgeen wordt patroon te herkennen. Dit bracht een weerspiegeld in een verschillende mate van belangrijk verschil tussen GC-gevoelige en onderdrukking van het IFN-patroon. Ook in GC-ongevoelige patiënten te voorschijn. In andere therapieën (bv Infliximab of de microarray analyse van zowel CD4+ T- Rituximab) werd een associatie van het cellen als ook monocyten van GC- therapie effect en de IFN signatuur gezien; ongevoelige RA- patiënten was de expressie echter, de gepubliceerde data zijn hier niet van genen, die door interferon alpha (IFN) eensluidend. In de toekomst is het belangrijk gereguleerd wordt, significant meer om deze effecten precies te onderzoeken, om veranderd dan in die van GC-gevoelige RA- een ideale, op maat gesneden therapie (of patiënten. IFNiseen cytokine, dat normaal combinatie van therapieën) voor de RA- gesproken bij virusinfecties door cellen patiënt te kunnen maken. geproduceerd wordt, maar de laatste jaren is . een verhoogd IFN activiteit (“IFN signatuur”) ook in meerdere

164

Dankwoord / Acknowledgements

165

Dankwoord:

Dit proefschrift is een reflectie van een decennium onderzoek in twee continenten, drie verschillende instituten, landen en systemen. Er zijn dan ook velen om te bedanken. Nu alles af is overheerst het gevoel wetenschap en de wetenschappelijke wereld van verschillende zijden te hebben kunnen zien en het goede van elk systeem te kunnen behouden. Daarmee is de weg ook langer en soms in omgekeerde volgorde geworden (van Habilitation naar promotie, hoewel de tweede eigenlijk juist aan de eerste voorafgaat en de eerste hogere eisen stelt) Gelukkig is in het leven de weg het doel ;-)

Allereerst wil ik de Oostenrijkse, Amerikaanse en Nederlandse patiënten en vrijwilligers bedanken voor het steeds weer afstaan van tijd en lichaamssappen. Zonder hen was ik ook na 10 jaar nog niet verder dan de inleiding van dit proefschrift.

Prof. dr. Bijlsma, beste Hans, zonder jou was dit proefschrift in deze vorm er niet gekomen, of ik daar onverdeeld dankbaar voor moet zijn weet ik nog steeds niet. Zonder gekheid is je rustige steun, ook bij momenten van Midden-Europese emotie van grote waarde geweest, zowel puur wetenschappelijk, als organisatorisch-politiek. Je begrip en ondersteuning tijdens een periode van groot persoonlijk verlies heb ik nodig gehad en als zeer bijzonder ervaren. Ook hiervoor mijn warme dank.

Prof. dr. Lafeber, beste Floris, vanaf het eerste moment heb ik bij jou het gevoel gehad dat je me als wetenschapper volledig vertrouwde en me de ruimte gaf het onderzoek naar mijn eigen ideeën in te richten, mocht ik dingen willen bespreken of vragen stond je altijd klaar. De moeite die je hebt genomen mij in mijn carrière in een voor mij vreemd land te ondersteunen en aan te moedigen apprecieer ik zeer.

Sehr geehrter professor Smolen, lieber Joschi, eine postdoc-Kollegin im NIH hat mich den Spruch “what would Jesus do” gelehrt, nach dem sie ihr Leben und ihre Wissenschaft ausrichtete. In der klinischen Reumatologie und in der wissenschaftlichen Immunologie kann man sich ein schlechteres Leitmotiv als “what would Joschi do” auswaehlen. Auch wenn unsere Wege dann doch sehr verschieden verlaufen (sind), so kann ich heute mit Stolz sagen, dass wir beide die einzigen österreichischen Reumatologen mit einem Doktorat einer holländischen Universität sind……..herzlichen, aufrichtigen Dank fuer Deine vielen großen und kleinen Gesten der Unterstützung.

Dear prof. Lipsky, dear Peter, it’s in your lab that I’ve come into my own as a scientist. The freedom to pursue my own ideas combined with helpful and critical guidance (your dogma of “MORE SEQUENCES!” to prove a finding still echoes in my brain) is something I’m tremendously grateful for. Your approach to science and your ability to see the ironic side of science and life in general have been a great example for me. Above all, I am thankful for your concern for my development and wellbeing also beyond the scientific aspects.

Sehr geehrter prof. Kraft, ohne die Chance, die ersten wissenschaftlichen Gehversuche, in Ihrem Labor unternehmen zu können, wären die T-zellen nie zu so einem Fokus meiner wissenschaftlichen Arbeit geworden. Vielen Dank für diese Gelegenheit! Mein Dank gilt natürlich auch den Menschen, die mich bei diesen ersten Schritten begleitet haben. Die kritischen Fragen von Barbara Bohle, Ursula Wiedermann und Beatrice Schmid (meist auch noch mit Lösungsvorschlägen gespickt) zusammen mit dem handwerklichen know-how von Renate und Ute haben mir ein sehr nützliches Rüstzeug gegeben. Die lebendige und schnelle

167 Art von Christof Ebner, seine Gruppe zu leiten wird mir ebenso in Erinnerung bleiben wie der Humor, mit dem alles gewürzt wurde.

Einmal auf der Rheuma, haben die T-zellen erst dann richtig wachsen wollen, als Daniela Müllegger (und spater Bettina Zweifler) sich um sie gekümmert hat. Für Eure Hingabe an unser Projekt bin ich sehr dankbar; Ihr beide habt ein bewundernswertes Durchhaltevermögen, das sich auch auf privater Ebene zeigt (und bezahlt macht!)….. Sehr geehrter Prof. Streiner, lieber Günter, ohne Deine sanfte Lenkung wär RA33 noch immer kein T-zell antigen…..vielen Dank für die Geduld und die liebenswerte Art, Deinen Mitarbeitern ihren Freiraum zu lassen… Von den vielen, mir lieben KollegInnen auf der Rheuma im AKH sind es zwei, denen ich besonders dankbar bin. Lieber Martin, Prof. Aringer, Deine ruhige Art, auch schwierige Themen direkt an zu sprechen und immer für Deine Freunde und Kollegen mit Rat und (vor allem) Tat bereit zu stehen, ist unschätzbar wertvoll. Tja, und Stups, ohne Deinen Schmäh, hätten wir so manches Semester nicht überstanden……Deine Kombination von Humor und Präzision gehört unter Denkmalschutz…..

Dear Margarida, Klaus, Nancy, Gary, Rachel, Xavier, AnneMarie and Melissa, without you the time at the NIH would have been scientifically interesting, but certainly not so memorable. I cherish the late night shifts at the lab (or Kramer books) as well as scientific (or political) discussions. What a treat!

Mijn lieve klinische Utrechtse collega’s, de overgang van Wenen naar Nederland was een cultuurschok, die door jullie steun en vriendschap veel makkelijker te overwinnen was: Aike - vanaf het eerste moment heb ik me bij jou gewaardeerd en als gelijke gevoeld, ook toen ik als habilitierte Oostenrijks interniste weer als assistente in Utrecht begon. Mijn diepe dank voor deze houding. Janneke - jij bent eenvoudig een van de meest collegiale arts die ik in alle ziekenhuizen en continenten ben tegen gekomen, dank voor jouw steun en hulp! Evelien - niet alleen draait de afdeling en poli op jouw organisatietalent, daarnaast weet je soms beter dan ik wanneer nog meer erbij doen gewoon niet meer past. Hans J - wat vond ik je soms een zeur toen ik net begon, hoe waardeer ik je precisie en heldere analytische blik nu. Ben jij veranderd of ben ik het? Helen - ik bewonder je inzet en drive, en ben ervan overtuigd dat we samen ook in de toekomst iets heel moois van de Utrechtse klinische immunologie gaan maken. Leen - je ongelofelijk analytische denkvermogen en de droge humor maakten besprekingen en gewone koffiepauzes erg vermakelijk. Jammer, dat dit nu alleen maar buiten het UMC gebeurt! Willemijn – je bent een echte people’s person….je denkt mee, en probeert te helpen waar mogelijk, dank hiervoor!

Geachte Prof. Derksen, beste Ron- ik heb veel van je geleerd, klinisch en wetenschappelijk, maar ook hoe kliniek en wetenschap te organiseren en positioneren. Daarnaast zijn besprekingen met jou altijd leuk en vrolijk.

Geachte Prof. Radstake, beste Tim- met je inzicht, overzicht en overredingskracht heb je in korte tijd je stempel op de Utrechtse Immunologie gedrukt, ik verheug me op onze verdere vruchtbare samenwerking bij het lupus onderzoek en daarbuiten.

168 Geachte Prof. van Laar, beste Jaap, hoewel de laatste maanden een leuke en interessante tijd onder jouw leiding voorspellen, heb je me de grootste beloning en uitdaging al gegeven, toen je die Nederlandse jonge student mee naar de NIH nam, die me uiteindelijk naar Utrecht bracht. Dit wordt moeilijk om te toppen!

Remco – het is zo stil, sindsdien je weg bent. En nu moeten Jocea en ik de koekjes zelf kopen! Je was een erg gezellige, en heel gedreven collega, en ik mis je als mens en als SLE- sparring partner!

Marieke – je bent mijn eerste collega hier geweest, en hebt mijn dagelijkse frustraties over het feit, dat geneeskunde in Oostenrijk blijkbaar het tegenovergestelde van geneeskunde in NL is, opgevangen. Je deed dit met jouw droge humor, waarvan ik als vriendin sindsdien ook vaak mocht genieten.

Collega’s uit het lab, Sandra- When you came to our lab you were a great student, meticulous, hard-working, willing to learn and think for yourself, you’re clearly on your way to become a great scientist. Jasper- zonder jouw rust, vakkennis en geduld weet ik niet hoe de laatste hoofdstukken tot stand gekomen zouden zijn. Hartelijk dank! Marian- niet alleen was je bij de micro-arrays ongelofelijk precies en georganiseerd, ook bij het onvermijdelijk trouble-shooten was je altijd voor vragen beschikbaar en had je eigenlijk ook altijd het antwoord op de problemen. Ook jou van harte bedankt! Arno, Kim, Katja, Marion, Dorien, veel dank voor jullie inzet en vakkennis, en natuurlijk voor jullie geduld met de buitenlandse medicus met al haar buitenissige ideeën en plannen. AnneKarien, Marjolein (Sijvers), Annemiek en Stefan veel dank voor jullie geduld voor mij bij het opzetten van mijn nieuw uitdaging — het lupus cohort. En dat, hoewel ik mijn hoofd eigenlijk altijd elders (bv dit boekje) had! Dank! Jonas, wat kan je mooie plaatjes maken…en dan ben je ook nog een uitstekende immunologische wetenschapper. Wat leuk, dat ik van beide feiten mag profiteren ;-) Joel, ik waardeer je rustige manier, je vakkennis en ook je humor (de ironische glimlach) waarmee we weer eens een (immunologisch) obstakel op weg naar de METC of publicatie opzij schuiven….leuk, om de SLE lijn met jou te trekken. Marieke (Vianen), veel dank voor al jouw precies werk bij het corrigeren en lay-outen van dit proefschrift! Wat een verschil!

Lieve Jocea en Martijn, jullie heb ik veel te danken, jullie bijdrage aan dit boekje is vooral het behouden van mijn “sanity” in de laaste 7 jaar. Jocea, jouw directe manier en je empathie maken dat je een goede arts bent, maar ook een echte vriend. En jouw fantasie maakt, dat het gewoon erg gezellig met je is….Martijn, door jouw schuld wonen we in Bunnik, en willen we hier ook niet meer weg. Jij (en Lieke) staan altijd klaar voor ons, met name als ik weer aan dit boekje moet werken…..het is af! We kunnen weer gewoon koffie drinken!

Dear close friends in Austria, the NL, the rest of Europe and the US, thank you all for the hours of (telephone) psychotherapy or late night drinks, or your frequent trips to Utrecht / Washington to keep me company and keep me happy!

Lieve schoonfamilie, lieve Neline en Anton, jullie steun is op de ene kant heel materieel: het tijdsgeschenk, dat jullie ons geven, als de kinderen weer een hele weekend, of avond bij jullie doorbrengen (tot hun genot!). Van de andere kant, heel immaterieel: de steun en dagelijkse waardering, en “positive feedback”, waarmee je dan weer de volgende obstakel bestrijdt.

169 Coen en Anouk, jullie hulp – ondanks eigen drukte ook nog onze 2 kinderen over te nemen – gaf en geeft ons de mogelijkheid, de boekjes af te maken. En dan is het ook nog gezellig!!!

Liebe Eltern, die Habil hat nicht gereicht, PhD mußte es auch noch sein. Natürlich wäre dies ohne Euren Beitrag nicht möglich gewesen: Genetik, Erziehung, das Bereitstellen von verschiedensten Möglichkeiten,…..aber auch die liebevolle Stütze in kritischen Momenten. Allen voran steht allerdings das tradierte Credo, zu versuchen, das Beste zu geben, beruflich und als Mensch. Schön, daß wir ab jetzt mehr Zeit haben, das Beste von einander einfach nur zu genießen.

Clairezl, Schwesterherz, ohne Dich geht sowieso niks. Wir sind gemeinsam zur hebrew day school und durch die Patho/Bürgerliches Recht Prüfung, durch Beatles und Celentano, durch Venedig und Washington gegangen. Wer weiß wo uns der Weg noch überall hin führt!

Lieve Jan, hoe zeer we bij elkaar horen, merk ik nu, dat we zo vaak apart zijn. De mooiste dagen van mijn leven beginnen en eindigen naast jou, und ich bin für jeden einzelnen davon dankbar. Lieber Philip, liebe Charlotte, ohne Euer Kichern ist selbst die tollste Publikation nichts wert. Was bin ich froh, dieses Kichern jeden Tag so oft hören zu können.

170 List of publications

171

Publications:

Original papers:

L. Bauer, C. Ebner, R. Hirschwehr, B. Wüthrich, C. Pichler, R. Fritsch, O. Scheiner and D. Kraft: IgE-crossreactivity between birch pollen, mugwort pollen and celery is due to at least three distinct crossreacing allergens. Immunoblot investigationof the birch-mugwort-celery syndrome. Clin. Exp. Allergy; 1996;26:1161-70 R. Fritsch, H. Ebner, D. Kraft and C. Ebner: Food allergy to pumpkin seed. Characterization of allergens. Allergy; 1997; 52: 335-7 Y. Cahen, R. Fritsch, B. Wüthrich. Food allergy with monovalent sensitivity to poultry meat. Clin. Exp. Allergy; 1998; 28:1026-930 U. Wiedermann, B. Schmid, R. Fritsch, L. Bauer, H. Renz, D. Kraft and C. Ebner: Effects of adjuvants on the immune response to allergens in a murine model of allergen inhalation: cholera toxin induces a Th1 like response to Betv1, the major birch pollen allergen. Clin. Exp. Immunol.; 1998; 111: 144-51 R. Fritsch, B. Bohle, U. Vollmann, U. Wiedermann, B. Jahn-.Schmid, M. Krebitz, H. Breiteneder, D. Kraft and C. Ebner: Bet v 1, the major birch pollen allergen, and Mal d 1, the major apple allergen, crossreact on the level of allergen-specific T helper cells. J Allergy Clin Imunonol; 1998; 102:679-86. E. Jensen-Jarolim, N. Reider, R. Fritsch and H. Breiteneder: Fatal outcome of anaphylaxis to camomile-containing enema during labor: a case study: J Allergy Clin Immunol; 1998; 102(6Pt2):104. R. Fritsch, D. Eselböck, B. Jahn-Schmid, C. Scheinecker, B. Bohle, K. Skriner , J. Neumüller, J. Smolen and G. Steiner. Characterization of auto-reactive T cells to the autoantigens hnRNP-A2/RA33 and Filaggrin in Patients with Rheumatoid Arthritis and Controls. J Immunol; 2002; 169(2):1068-76. R.D. Fritsch, X. Shen, G.P. Sims, K.S. Hathcock, R.J. Hodes, P.E. Lipsky. Stepwise differentiation of CD4 memory T cells defined by expression of CCR7 and CD27.J Immunol; 2005;175(10):6489-97. R. D. Fritsch*, X. Shen*, G. G. Illei, C. H. Yarboro, C. Prussin, K. S. Hathcock, R. J. Hodes, P. E. Lipsky. Abnormal Differentiation of Memory T Cells in Systemic Lupus Erythematosus. Arthritis and Rheum; 2006; 54(7):2184-97 *Authors contributed equally R. Fritsch-Stork, D. Eselböck, B. Jahn-Schmid, J. Neumüller, B. Bohle, K. Skriner, J. Smolen, and G. Steiner: Characterization of RA33 (hnRNP-A2/B1) auto-reactive T cells in SLE-Patients. Arthritis Research & Therapy 2006, 8:R118 Souto-Carneiro MM, Fritsch R, N Sepulveda, MJ Lagareiro, N Morgado, NS Longo, PE Lipsky. The NF-κB canonical pathway is involved in the control of the exonucleolytic processing of coding-ends during V(D)J recombination. J Immunol. 2008;15;180(2):1040-9. Stummvoll GH*, Fritsch RD*, Meyer B, Hoefler E, Aringer M, Smolen JS, Steiner G. Characterization of cellular and humoral autoimmune responses to histone H1 and core histones in human systemic lupus erythematosus. *Authors contributed equally. Ann Rheum Dis. 2009;68(1):110-6.

173 Souto-Carneiro MM, Mahadevan V, Takada K, Fritsch-Stork R, Nanki T, Brown M, Fleisher TA, Wilson M, Goldbach-Mansky R, Lipsky PE. Alterations in peripheral blood memory B cells in patients with active rheumatoid arthritis are dependent on the action of tumour necrosis factor. Arthritis Res Ther. 2009;11(3):R84. Epub 2009 Jun 5. Fritsch-Stork RD, Leguit RJ, Derksen RH. Rapidly fatal HTLV-1-associated T-cell leukemia/lymphoma in a patient with SLE. Nat Rev Rheumatol. 2009;5(5):283-7. Van Avondt K, Fritsch-Stork R, Derksen RHWM and L Meyaard. SIRL-1 suppresses the release of neutrophil extracellular traps in systemic lupus erythematosus. Accepted. Plos one Fritsch-Stork RD, MD; Cardoso SCS, MSc; Groot-Koerkamp MJA, BSc; Broen J, MD, PhD; Concepcion AN, BSc; Giovannone B, PhD; Ploner A, PhD; Lafeber FPJ, PhD; Bijlsma JWJ, MD, PhD. Potential predictors of Glucocorticoid-response in Rheumatoid arthritis. A role for endoplasmic reticulum aminopeptidase 2 (ERAP2) and leucocyte-specific transcript 1 (LST1) ? Paper submitted Fritsch-Stork RD, MD; Broen J, MD, PhD; Cardoso SCS, MSc; Groot-Koerkamp MJA; BSc; v Roon J, PhD; K van der Wurff-Jacobs, BSc; Radstake TRDJ; MD, PhD; Lafeber FPJ, PhD; Bijlsma JWJ, MD, PhD. Glucocorticoid-responsiveness correlates with an Interferon signature in Rheumatoid Arthritis Patients. Paper submitted Cardoso SCS, MSc; MD; Broen J, MD, PhD; Groot-Koerkamp MJA; BSc; Lafeber FPJ, PhD; Bijlsma JWJ, MD, PhD; Fritsch-Stork RD. The gene expression changes mediated by oral Glucocorticoids (GC) in GC-responding Rheumatoid Arthritis Patients . Paper in preparation * these autors contributed equally

Reviews/Commentary:

R. Fritsch, C. Ebner and D. Kraft: Allergenic Crossreactivities: Pollens and Vegetable Foods. In: Allergic crossreactions, Clinical Reviews in Allergy and Immunology ; 1997; 15:397-404 Hueber AJ, Distler JH, Fritsch-Stork RD. International rheumatology networking: The 2008 American College of Rheumatology/European League Against Rheumatism exchange program. Arthritis Rheum. 2009; 60(7):1881-3. RD Fritsch-Stork and PO Fritsch. Dermatologist versus rheumatologist in the management of lupus patients. Lupus. 2010, 19(9):1153-5. Schoneveld JL, Fritsch-Stork RD, Bijlsma JW. Nongenomic glucocorticoid signaling: new targets for immunosuppressive therapy? Arthritis Rheum. 2011 Dec;63(12):3665-7. RK Luijten, R Fritsch-Stork, JWJ Bijlsma, RHWM Derksen. The use of glucocorticoids in Systemic Lupus Erythematosus; after 60 years still more an art than science"; Autoimmun Rev. 2013 Mar;12(5):617-28. Others: JWG Jacobs en RDE Fritsch-Stork. Knuckle pads. Medisch Contact 2010, 65(19):870 A v Kessel en RDE Fritsch-Stork: Jacob Churg, Lotte Strauss en het syndroom van Churg- Strauss; Nederlands tijdschrijft voor Reumatologie 2012, 4

174 RDE Fritsch-Stork en A v Kessel: Eponiem: Emanuel Labman, Benjamin Sacks en Libman- Sacks endocarditis; Nederlands Tijdschrijft voor Reumatologie 2013, 2

Abstracts/oral Presentations/Posters:

R. Fritsch, H. Ebner and D. Kraft: Food allergy to pumpkin seed-Characterization of allergens. EAACI (European Academy of Allergology and Clinical Immunology), June 1997, Rhodes, Greece. R. Fritsch, B. Bohle, U. Vollmann, U. Wiedermann, B. Jahn-.Schmid, M. Krebitz, H. Breiteneder, D. Kraft and C. Ebner: Bet v 1, the major birch pollen allergen, and Mal d 1, the major apple allergen, crossreact on the level of allergen-specific T helper cells. ÖGAI (Österreichische Gesellschaft für Allergologie und Immunologie), October 1997, Innsbruck, Austria R. Fritsch, B. Bohle, U. Vollmann, U. Wiedermann, B. Jahn-Schmid, M. Krebitz, H. Breiteneder, D. Kraft and C. Ebner: Birch pollen and apple share T cell epitopes. ÖGAI, October 1998, Graz, Austria. R. Fritsch, D. Eselböck, B. Jahn-Schmid, C. Scheinecker, B. Bohle, K. Skriner, J. Smolen and G. Steiner: Characterization of T cell reaction to the autoantien RA33 (hnRNP-A2/B1) in SLE and RA patients. EWWR (European Workshop on Rheumatology Research), March 2000, Oxford, GB and EULAR (European League against Rheumatism), June 2000, Nice, France. R. Fritsch, D. Eselböck, B. Jahn-Schmid, C. Scheinecker, B. Bohle, K. Skriner, J. Smolen and G. Steiner: Characterization of auto-reactive T cells in peripheral blood of patients with RA and SLE. EFIS (European Federation of Immunological Societies), September 2000, Poznan, Poland. R. Fritsch, D. Eselböck, B. Jahn-Schmid, C. Scheinecker, B. Bohle, K. Skriner, J. Smolen, J. Neumüller and G. Steiner: Characterization of auto-reactive T cells to the autoantigens hnRNP-A2/RA33 and Filaggrin in Patients with Rheumatoid Arthritis and Controls. EWRR March 2001,Vienna, Austria. R. Fritsch, D. Eselböck, B. Jahn-Schmid, J. Neumüller, B. Bohle, K. Skriner, J. Smolen, and G. Steiner: Characterization of RA33 (hnRNP-A2/B1) auto-reactive T cells in SLE-Patients EWRR, March 2001, Vienna, Austria and 6th International Lupus Conference, March 2001, Barcelona, Spain. R. Fritsch, D. Eselböck, B. Jahn-Schmid, J. Neumüller, B. Bohle, K. Skriner, J. Smolen, and G. Steiner: Characterization of RA33 (hnRNP-A2/B1) auto-reactive T cells in SLE-Patients; EULAR, June 2001, Prague, Czech Republic. R. Fritsch, D. Eselböck, B. Jahn-Schmid, B. Bohle, K. Skriner, J. Smolen, and G. Steiner: Characterization of auto-reactive T cells in SLE-Patients. ACR (American College of Rheumatology), November 2001, San Francisco, USA. R. Fritsch: Myalgie oder Myositis; Tiroler Ärztetage 2001, October 2001, Mayerhofen, Austria.

175 Fritsch R, Olson D, Illei G, Yarboro C, and Lipsky PE: Characterization of differences in memory T cell subsets of SLE patients and Healthy Controls, ACR, October 2003, Orlando, USA, and FASEB, 2004, Washington DC, USA and FOCIS, July 2004, Montreal, Canada. Souto-Carneiro MM, Fritsch R, Yanovski J, Lipsky PE. Abnormalities in V(D)J Rearrangements in Patients with X-Linked Hyper-IgM Ectodermal Dysplasia Syndrome Indicate a Requirement for the Ikkgamma Dependent Signaling Pathway in V(D)J recombination. FASEB, 2004, Washington DC, USA. G.H. Stummvoll, R.D. Fritsch, B. Meyer, M. Aringer, J.S. Smolen, G. Steiner. Autoreactive T cells to Histone H1 and Core Histones in SLE patients. EULAR, June 2005, Vienna, Austria Yuko Shirota, Cheryl Yarboro, Gary Sims, Ruth Fritsch, Rachel Ettinger, Xavier Valencia, Randy Fischer, Amrie Grammer, Tuyet-Hang Pham, Peter E. Lipsky, Gabor G. Illei. The Impact of In Vivo Anti IL-6 Receptor Blockade on Circulating T and B Cell Subsets in Patients with Systemic Lupus Erythematosus. ACR, October 2005, San Diego, USA. R. Fritsch-Stork , X. Shen, G. G. Illei, C. H. Yarboro, C. Prussin, K. S. Hathcock, R. J. Hodes, P. E. Lipsky: T cells display distinct abnormalities in memory cell maturation. EULAR, June 2006, Amsterdam, The Netherlands. R. Fritsch-Stork. Difficult SLE cases, EULAR, June 2012, Berlin, Germany. R. Fritsch-Stork. Het antifosfolipidensyndroom. NVLM vakbeurs, march 2013, Den Haag, The Netherlands. R. Fritsch-Stork. Antiphospholipidsyndrom (APLAS): Ak, Diagnostik, was tun? Reumatologische Wintertagung 2013, Leogang, Austria. March 2013 R. Fritsch-Stork, Cardoso S, Broen J, Groot-Koerkamp M, Concepcion A, Lafeber F and Johannes W.J. Bijlsma. Distinction Between Glucocorticoid-Responders and Non-Responders In Rheumatoid Arthritis: A New Role For Endoplasmic Reticulum Aminopeptidase 2. ACR, October 2013, San Diego, USA. RK Luijten, R Fritsch-Stork, JWJ Bijlsma, RHWM Derksen. The use of glucocorticoids in Systemic Lupus Erythematosus; after 60 years still more an art than science"; NVVI (Dutch Society of Immunologists), december 2013, Noordwijkerhout, The Netherlands

176

Curriculum vitae

177

Curriculum vitae

Ruth Fritsch-Stork was born on 21 August 1968 in Vienna and attended the lycee francais de Vienne before graduating with distinction from the Innsbrucker Gymnasium (highschool). She studied medicine at the Medical University of Vienna and after her graduation worked as a research fellow in the Department of Pathophysiology and Allergy-research of the Medical University of Vienna under supervision of prof. dr. Dietrich Kraft. From 1997 to 2007 she followed her residency training in internal medicine at the General Hospital of Vienna (the University Hospital of the Vienna) in the Department of Internal Medicine III/Rheumatology under prof. dr. Josef Smolen. She interrupted her residency for several years to concentrate on research at NIAMS (National Institute for Arthritis, Musculoskeletal and Skin diseases, Bethesda, MD, USA) under prof. dr. Peter Lipsky. The research conducted during her clinical training together with some work from the NIH culminated in her “Habilitation” at the Medical University of Vienna in 2007. Part of the studies performed at the NIH and in Vienna were also included in this thesis. From 2007 to 2011 she received further training in rheumatology at the University Medical Center Utrecht (UMCU) under prof. dr. Hans Bijlsma. Since 2011 she has been working as internist-rheumatologist in the UMCU. In 2012 she also concluded her subspecialty training in clinical immunology and was registered as internist-clinical immunologist in the Netherlands. During her clinical work at the UMCU she carried out basic research activities, which are included as the last chapters in this thesis. Ruth Fritsch is married to Jan Stork and proud mother of Philip and Charlotte.

Ruth Fritsch-Stork werd 21 augustus 1968 in Wenen geboren en ging aan het lycee francais de Vienne en later het Innsbrucker Gymnasium naar school. Zij studeerde geneeskunde aan de Medizinische Universität Wien. Na het artsexamen werkte zij als onderzoeker op de afdeling Pathophysiologie und Allergieforschung aldaar, onder begeleiding van prof. dr. Dietrich Kraft. Van 1997 tot 2007 werkte zij als arts-assistent in opleiding tot internist aan de Medizinische Universität Wien, afdeling Innere Medizin III/Rheumatologie, met als opleider prof. dr. Josef Smolen. Zij onderbrak haar opleiding om enkele jaren onder prof. dr. Peter Lipsky onderzoek te doen aan het National Institute for Arthritis, Musculosceletal and Skin diseases. Ook tijdens haar opleiding tot internist deed zij onderzoek, hetgeen in 2007 leidde tot een succesvolle verdediging van haar Habilitation aan de Medizinische Universität Wien. Een deel van het toen verrichte onderzoek is in dit proefschrift opgenomen. Van 2007 tot 2011 specialiseerde zij zich in het Universitair Medisch Centrum Utrecht verder tot reumatoloog, haar opleider daar was prof. dr. Hans Bijlsma. Vanaf 2011 werkt zij als internist-reumatoloog in het UMCU. In 2012 werd zij ook als internist-klinisch immunoloog geregistreerd. Ook naast haar klinische werkzaamheden in het UMCU bleef zij wetenschappelijk onderzoek doen, leidend tot het derde deel van dit proefschrift.

Ruth Fritsch is getrouwd met Jan Stork en trotse moeder van Philip en Charlotte.

179