Autoimmune rheumatic diseases days held in Athens in June 2004

Autoimmunity Reviews 3 Suppl. 1(2004) S7–S73

1568-9972/04/$ - see front matter Elsevier Science B.V.. Published by European Society of Cardiology All rights reserved.PII: S1568-9972Ž04.00043-6

Mini reviews

PATHOGENETIC MECHANISMS IN AUTOIMMUNE DISEASES

1.Type 1 Diabetes as a model forautoantibodies as predictors ofautoimmune diseases

Abner Louis Notkins, M.D.,(Experimental MedicineSection, Oral Infection and Immunity Branch, NationalInstitute of Dental and Craniofacial Research, NationalInstitutes of Health, Bethesda, Maryland, USA)

Autoantibodies as predictors of type 1 diabetesEvidence that type 1 diabetes is an autoimmune diseaseor at least has a major autoimmune component was firstrecognized in the mid 1970s by the demonstration thatsera from type 1 patients stained by immunofluorescencenormal pancreatic islets. These autoantibodies becameknown as islet cell antibodies(ICA) and the two majorautoantigens with which ICA reacted were identified inthe 1990s as an isoform of glutamic acid decarboxylase(GAD 65) and IA-2 (1). GAD is involved in the con-version of glutamic acid to gamma-aminobutyric acid, amajor neuroinhibitory transmitter. IA-2 is a member ofthe transmembrane protein tyrosine phosphatase familylocated in dense core secretory vesicles and is involvedin insulin secretion(2). By using recombinant GAD andIA-2 in radioimmune-precipitation tests, 70% to 90% ofnewly diagnosed patients with type 1 diabetes werefound to have autoantibodies to one or the other or bothof these proteins. Further studies showed that these auto-antibodies appeared months or years before the devel-opment of clinical disease. This was demonstrated byboth prospective and retrospective studies in whichserum samples were collected over a 5 to 10 year periodand then tested for the presence of autoantibodies whenclinical diabetes developed. Three different at risk pop-ulations were tested: identical twins, first-degree rela-tives of patients with type 1 diabetes and a general(school children) population. The results from thesethree populations showed that autoantibodies to IA-2,GAD and insulin(an autoantibody not detected by the

ICA test) could be used as predictive markers and thatthe number of autoantibodies that were positive seemedto be more important than the titer of a single antibody.Estimates based on first degree relatives showed that thelikelihood of developing type 1 diabetes within fiveyears was approximately 10% in the presence of oneautoantibody, 50% in the presence of two autoantibod-ies, and 60% to 80% in the presence of three autoanti-bodies. Thus, autoantibodies have turned out to be gooddiagnostic markers to distinguish autoimmune type 1diabetes from non-autoimmune type 2 diabetes and goodpredictive markers to identify individuals at high risk ofdeveloping type 1 diabetes. The value of these predictiveantibodies is that it is now possible to rapidly screenlarge populations to identify high-risk subjects for entryinto therapeutic intervention trials long before there istotal loss of beta cell function and the onset of clinicaldisease. Autoantibodies(primarily to GAD 65) also arefound in 5% to 10% of adults with diabetes suggestingthat these individuals either were misdiagnosed and inreality have type 1 diabetes or have a combination oftype 1 and type 2 diabetes. Autoimmune diabetes inadults has been referred to as LADA or type 1.5 dia-betes. Since there are at least 16 million people in theUnited States with type 2 diabetes this would representa near doubling of the number of people with autoim-mune diabetes.The studies on type 1 diabetes, which were performedon tens-of-thousands of subjects and in many differentlaboratories around the world, suggest that autoantibod-ies also might serve as predictive markers for at leastsome of the other 30 to 40 autoimmune diseases. Auto-immune diseases, taken as a group, are the third leadingcause of morbidity and mortality after heart disease andcancer. Many of the autoimmune diseases are thought tobe complex genetic diseases(i.e. involving more thanone gene). Despite extensive efforts, thus far, very fewof these genes have been identified. Since many of theautoimmune diseases are chronic in nature, autoantibod-ies also might appear in these diseases long before the

S8 Autoimmunity Reviews 3 Suppl. 1 (2004) S7–S73

development of clinical symptoms. The appearance ofautoantibodies would strongly suggest that the diseaseprocess is already underway, whereas the presence ofspecific genes might simply indicate that the potnetialfor a particular disease exists, not that the disease willdevelop. Thus, autoantibodies may prove to be betterpredictive markers than the presence of high-risk genes,even if such genes are identified.

Principles of autoantibodies as predictors

Although a large number of different autoantigens havebeen identified, linked to specific autoimmne diseasesand used as diagnostic markers, with the exception oftype 1 diabetes, there have been very few large-scalestudies on autoantibodies as predictors of these otherautoimmune diseases. The studies on type 1 diabetespoint to several critical issues that must be taken intoconsideration when searching for and evaluating the pre-dictive value of autoantibodies(3). Lessons learnedfrom type 1 diabetes include the importance of: 1) theperformance characteristics and validation of the assay;2) the use of recombinant autoantigens and the confor-mation of the autoantigen; 3) the value of screening formore than one autoantibody; 4) the need for autoanti-body screening at different times during the course ofthe disease and at different ages; and 5) the sensitivity,specificity and predictive value of the autoantibodies.Autoantibodies with positive predictive value might beuseful as markers for disease classification, for initiatingearly treatment before the onset of clinical disease, forprognosis of disease course and severity and for avoid-ing potential triggering agents(e.g. glutens in celiacdisease).

Status of autoantibodies as predictors in other auto-immune diseases

Although still at a very early stage, recent findings havedemonstrated the predictive value of autoantibodies inseveral diseases. For example, autoantibodies to syn-thetic citrullinated peptide(4) appear years before thedevelopment of clinical rheumatoid arthritis and highautoantibody titers may be of value in predicting pro-gression to erosive disease. Autoantibodies to nuclearantigens also appear years before the development ofsystemic lupus erythematosus(SLE). In a recent studythese autoantibodies were found in 80% of the subjectswho went on to develop SLE and some of these auto-antibodies were present up to 9.4 years before diagnosis(5). In Addison’s disease autoantibodies to 21-hydrox-ylase are widely used to differentiate autoimmune fromnon-autoimmune adrenal insufficiency. At the time ofdiagnosis 80 to 90% of patients with Addison’s disease

have autoantibodies to 21-hydroxylase, but because ofthe rarity of the disease(i.e. approximately one case per8000 people) predictive studies are difficult to perform.However, in poly endocrine disease, 0.5 to 5.0% of thepatients with Hashimoto’s thyroditis or type 1 diabeteshave autoantibodies to 21-hydroxylase. In children withpolyendocrine disease, autoantibodies to 21-hydroxylaseare highly predictive with up to 90% going on to devel-op Addison’s disease(6). In contrast, in adults withpolyendocrine disease, autoantibodies to 21-hydroxylaseare less predictive with only 20% developing Addison’sdisease. In the case of multiple sclerosis(MS), autoan-tibodies to myelin basic protein or myelin oligodentro-cyte glycoprotein have not been shown thus far to bepredictive of the disease, but there is new evidence thatthese autoantibodies may be predictive of a relapse.Approximately 95% of patients with both of these auto-antibodies had a relapse within a mean of 7.5 monthsafter their first demyelinating attack as compared to only23% of autoantibody negative subjects who had arelapse within a mean of 45 months(7). Autoantibodiesto desmoglein 1 and desmoglein 3 have been used asdiagnostic markers for pemphigus. This disease is rela-tively rare so predictive studies have been difficult toperform. However, an endemic form of pemphigus fol-iaceus, known as fogo selvagem, is found in about 4.0%of the population(1200) in Limao Verde, Brazil. Pro-spective studies on this population showed that autoan-tibodies to the C-terminal domain of desmoglein 1(EC5) appeared months to years before the onset of clin-ical disease whereas autoantibodies to the N-terminaldomain of desmoglein 1(EC1 and EC2) appeared at theonset of clinical disease suggesting that epitope spreadplays a role in the pathogenesis of this disease(8,9).The diagnostic and predictive value of autoantibodies intwenty different autoimmune diseases has recently beensummarized(10).In conclusion, once an autoantibody is identified as avalid predictive marker it is likely to become part of thephysician’s laboratory armament. As the list of predic-tive autoantibodies increases, vigorous attempts will bemade to develop high-throughput screening proceduresto test simultaneously for dozens of autoantibodies atlow cost. If this becomes a reality, autoantibody screen-ing is very likely to become a routine part of the annualmedical examination, especially for those diseases inwhich therapeutic intervention is available.

References

1. Notkins A.L. Immunologic and genetic factors in type 1 dia-betes. J. Biol. Chem. 2002 15;277:43545–8.

S9Autoimmunity Reviews 3 Suppl. 1 (2004) S7–S73

2. Saeki K., Zhu M., Kubosaki A., Xie J., Lan M.S. Notkins A.L.Targeted disruption of the protein tyrosine phosphatase-likemolecule 1A-2 results in alterations in glucose tolerance testsand insulin secretion. Diabetes. 2002;51:1842–50.

3. Leslie D., Lipsky P., Notkins A.L. Autoantibodies as predictorsof disease. J. Clin. Invest. 2001;108:1417–22.

4. Rantapaa-Dahlqvist S., de Jong B.A., Berglin E., Hallmans G.,Wadell G., Stenlund H., Sundun U., van Venrooij W.J. Anti-bodies against cyclic citrullinated peptide and IgA rheumatoidfactor predict the development of rheumatoid arthritis. Arthri-tis Rheum. 2003;48:2741–9.

5. Arbuckle M.R., McClain M.T., Rubertone M.V., Scofield R.H.,Dennis G.J., James J.A., Harley J.B. Development of autoan-tibodies before the clinical onset of systemic lupus erythe-matosus. N. Engl. J. Med. 2003;349:1536–33.

6. Betterle C., Volpato M., Rees Smith B., Furmaniak J., ChenS., Zanchetta R., Greggio N.A., Pedini B., Boscaro M., Pre-sotto F. II. Adrenal cortex and steroid 21-hydroxylase auto-antibodies in children with organ-specific autoimmunediseases: markers of high progression to clinical Addison’sdisease. J. Clin. Endocrinol. Metab. 1997;82:939–42.

7. Berger T., Rubner P., Schauter F., Egg R., Ulmer H., MayringerI., Dilitz E., Deisenhammer F., Reindl M. Antymyelin anti-bodies as a predictor of clinically definite multiple sclerosisafrer a first demyelinating event. N. Engl. J. Med.2003;349:139–45.

8. Warren S.J., Lin M.S., Giudice G.J., Hoffmann R.G., Hans-Filho G., Aoki V., Rivitti E.A., Santos V., Diaz L.A. The prev-alence of antibodies against desmoglein 1 in endemicpemphigus foliaceus in Brazil. Cooperative Group on FogoSelvagem Research. N. Engl. J. Med. 2000;343:23–30.

9. Li N., Aoki V., Hans-Filho G., Rivitti E.A., Diaz L.A. The roleof intramolecular epitope spreading in the pathogenesis ofendemic pemphigus foliaceus(fogo selvagem). J. Exp. Med.2003;197:1501–10.

10. Notkins A.L., Lernmark A., Leslie D.(Editors). Autoantibod-ies as Diagnostic and Predictive Markers of Autoimmune Dis-eases. Autoimmunity. 2004(in press).

2.Apoptosis in systemic autoimmunity

J. Kalden,(Institute and Poliklinik fur Klinische Immunologie undRheumatologie, Krankenhausstrabe 12, 3560-8520,Erlangen, Germany).Email: [email protected]

Apoptosis or programmed cell death is a normal physi-ological process occurring in senescent and damagedcells. Apoptosis, which has become one of the major

areas of research today was first described in 1885 byFlemming. The term of apoptosis was coined by Kerrand Wyllie in 1972. Four different faces can be definedduring an apoptotic process: Initiation, Effector, Degra-dation, Engulfment.The fourth face, the receptor-mediated uptake of apop-totic bodies by phagocytes is of special interest for thepathogenesis of SLE.Autoimmune disease situations have been described inwhich too much or too little of apoptosis might beinvolved in pathogenic events. An example in which toolittle cell deaths leads to autoimmune phenomena is theMRL-lpr mouse strain which develops a SLE-like syn-drome including a massive lymphadenopathy caused bya mutation in the Fas-receptor gene. The correspondinghuman disease is theCanale-Smith-Syndrome or auto-immune lymphoproliferative syndrome(ALPS). Thishuman disease is also characterized by a significant lym-phadenopathy and appearance of autoantibodies.Various reports suggested defects in the regulation ofapoptosis of different cell types within the inflamed syn-ovium as implicated in the pathogenesis ofrheumatoidarthritis (RA). Studies have shown an over-expressionof anti-apoptotic genes like Bola, BolXL and surviving.Therefore, the hypothesis has been forwarded that theseanti-apoptotic genes might contribute to the synovialhyperplasia. In addition, somatic mutations of the tumorsuppressor gene p53 might be responsible for too littlecell deaths within the inflamed synovium with conse-quences for perpetuating chronic inflammatory response.Finally, fibroblast-like cells have been demonstrated tobe sensitive to Fas-induced cell deaths by anti-Fas-anti-bodies, however, only a small proportion of Fibroblastsactually undergo spontaneous apoptosis in RA jointspecimens.A systemic autoimmune disease which is characterizedby a defective clearance of apoptotic cell bodies isSLE.In SLE, as originally described in our laboratory byHerrmann et al., peripheral moncyte derived macro-phages have a defect for the uptake of apoptotic cellmaterial. This leads to a persistant increase in the serumof SLE patients of oligonucleosomes. Oligonucleosomescontain components, which can be potential autoanti-genes for SLE. To verify the inefficient clearance ofapoptotic cells in vivo, lymph node biopsies from SLEpatients and from controls were histologically analyzed.In germinal centers from a subgroup of patients withSLE and active disease, the numbers of so-called tingi-ble body macrophages(TBM) were significantlyreduced and multiple non-enfulged apoptotic cells weredetected. In addition, Tunel-positive apoptotic materialwas observed not internalized but attached to the surface


of follicular dendritic cells, were it maybe presented toautoreactive B cells. In contrast, control-groups revealedapoptotic nuclei only in association with TBM.Since 1980, it has been known that macrophages ofpatients with SLE have an impaired phagocytic activityagainst yeast and bacteria. Based on our findings, thefollowing hypothesis is suggested for the induction ofautoantibodies in a subgroup of patients with SLE.Apoptotic B cells in the GC are not adequately clearedin the early phase of apoptosis. Therefore, apoptosis pro-gresses and the cells enter secondary necrosis. In thisstate, activation of complement may result in depositionof C3b on disintegrated apoptotic cells. The apoptoticcell-derived nuclear fragments bind to CR2yCD21 onFDC. Because of the lost membrane integrity, autoreac-tive B cells gain access to potential intracellular autoan-tigens, including dsDNA, which can now provide ashort-term survival signal. An important initial controlmechanism of B cell tolerance is circumvented by theretention of autoantigens on the surfaces of FDC. Underthese conditions, B cell tolerance relies mainly on thepresence of a functional T cell tolerance to nuclearautoantigens. However, histone-specific T cells, able toprovide in vitro help for dsDNA specific B cells, havebeen detected in patients with SLEw5,6x. In addition,the presentation to T cells of apoptotic cell-derived mod-ified autoantigens also challenges the T cell tolerance.If autoreactive T cells are present in the mantle zone ofthe lymph node they may provide a long-term survivalsignal for autoreactive B cells. The latter may furtherdifferentiate into plasma cells producing those nuclearautoantibodies that represent the hallmark of SLE.That a defective clearance might play a central role inthe pathogenesis of SLE and SLE-like syndromes is fur-ther supported by recently described animal models:knock-out mice for C1q, for SAP or DNA’s I.

3.Untying the gordian knot of autoimmunityand b cell lineage lymphomas

Herbert C. Morse III, Laurence Morel, Derry Roopenian,Peter Lipsky, Warren Leonard, Wendy Davidson,(NIAID, NIAMS and NHLBI, NIH; University of Flor-ida; Jackson Laboratory; American Red Cross, 5640Fishers Lane, Rockville, MD 20852).Email: [email protected]

An association between autoimmunity and the devel-opment of B cell lineage non-Hodgkin lymphomas, firstestablished for patients with Sjogren’s syndromeylym-¨phoepithelial sialadenitis(SSyLESA), has been known

for more than 30 years(1). The lymphomas of patientswith SSyLESA are most often low-grade extranodalmarginal zone B cell lymphomas(MZL) of mucosa-associated lymphoid tissue(MALT ) type, as are thyroidlymphomas of patients with Hashimoto’s thyroiditis.Lymphomas of patients with systemic lupus erythema-tosus(SLE) cover a spectrum of diagnoses includingsmall lymphocytic, follicular, diffuse large B cell(DLBCL), multiple myeloma (MM), and MALT.Patients with rheumatoid arthritis develop DLBCL at anincreased frequency that correlates with the level of dis-ease activity and with highly active disease may occurat a 25-fold higher frequency. Finally, lymphomas inpatients with autoimmune lymphoproliferative syndrome(ALPS) due to mutations in the FASyFASL signalingpathway are mostly follicular.Studies of mice from certain inbred strains and geneticcrosses have also revealed an association between auto-immunity and B cell lymphomas. NZB mice developsplenic MZL, a previously unrecognized subset ofmouse B cell lymphomas. In addition to the prototypesensitive strain, BALByc, NZB and (NZBxBALB yc)F mice are the only mice sensitive to pristane induc-1

tion of plasmacytomas(PCT), a neoplasm of matureplasma cells. More recently, mice with mutations inFas(lpr) andFasl (gld) were reported to develop neoplasmsof immature plasma cells termed plasmacytoid lympho-mas(2).Both genetic and environmental factors are thought tocontribute to the genesis of autoimmune disorders. Map-ping studies have defined chromosomal intervals in bothhumans and mice containing genes that predispose toautoimmunity. Some of these loci are known to relaxnormally stringent barriers to T and B cell responses toautoantigens. In one model, chronic ‘self’ antigen-drivenstimulation of normally silent autoreactive B cellsthrough multiple rounds of division could enhance thechances for genetic errors contributing to transformation.The critical replicative drive could be either B cellautonomous or exogenous. Our studies of autoimmunemice demonstrate that some strains have a high inci-dence of B cell lymphomas. Further, they suggest genet-ic determinants, both B cell-intrinsic and extrinsic, aloneor in combination, are central to the pathogenesis ofthese neoplasms.(NZBxNZW)F mice (ByW) develop profound auto-1

immunity associated with production of multiple auto-antibodies and lethal immune complex glomerulo-nephritis. Several strains of mice derived from ByWcrosses carry three loci –Sle1, Sle2 and Sle3 – thatmediate full-blown SLE when recombined on the non-autoimmune genetic background of C57BLy6 (B6)mice


(3). These loci are also carried by two additional ByW-derived strains, NZM2410 and TAN that are strikinglydifferent for their signs of autoimmunity. NZM2410mice exhibit high levels of autoantibodies and developdisease with features and severity comparable to that ofByW mice or the triple congenic B6 strain, B6.Sle1,Sle2, Sle3 (B6.TC). In contrast, TAN mice exhibit onlylow levels of autoantibodies and develop a mild, non-lethal glomerulonephritis like that of strain NZW. Stud-ies of the three ByW-derived strains for histopathologyof spleens and lymph nodes at approximately 1 year ofage showed that over half the TAN mice had clonalsplenic MZL while about 25% of NZM mice had clonallymphomas classified as follicular or DLBCL. Remark-ably, B6.TC mice had no clonal B cell lymphoprolifer-ative disorders at this age. This suggests that loss oftolerance to self antigens occasioned by the presence ofSle1, Sle2, Sle3 will provide the chronic B cell stimu-lation required for autoantibody production and muta-genic drive responsible for B cell transformation only inthe context of other genetic elements yet to be defined.Crosses between TAN and NZM2410 and B6.TC micewill provide opportunities to map these determinants.While the antigenic specificity of immunoglobulins pro-duced by lymphomas of the ByW-derived strains is notknown to be anti-self, the reactivity of secreted immu-noglobulins from the plasmacytoid lymphomas ofBALB yc.gld mice for nuclear antigens including DNAis well established. The fact that the immunoglobulinproducts of these tumors are class switched and aremutated in variable region sequences attests to their pas-sage through the germinal center and their origins fromself-reactive specificities and suggests that chronic stim-ulation by self antigens is critical to their genesis. In thisregard, it is most likely that these tumors derive fromthe expanded populations of autoreactive memory-likeB cells found in aginggld mice.BXSB-mice bearing theYaa mutation on the Y chro-mosome exhibit the only male dominant form of system-ic autoimmunity. Interestingly, disease of similar severitydevelops in genetic crosses with ByW but with not non-autoimmune strains of mice. While identification of theYaa mutation itself has not been achieved, insights intoits mode of action were suggested by real time PCRscreens of gene expression patterns over the time courseof disease. The results demonstrated that there weremarked increases in expression of the cytokine IL-21 inassociation with disease progression. These changes maywell be BXSB-Yaa-specific as IL-21 levels in spleens ofaged mice bearing thelpr mutation do not differ fromcontrols. In normal humoral immune responses, IL-21acts as a master regulator of life and death decisions in

B cells, driving naıve B cells to apoptosis while directing¨the maturation of antigen-experienced, T-dependent Bcells to memory and plasma cells(4). The possibilitythat the spleens of aging BXSB-Yaa mice might harborclonal populations of B cells, like those of BALByc-gldygld, is under investigation.The underpinnings of the relations between autoimmun-ity and B cell neoplasia are thus rapidly becoming moreapproachable in mice and will soon gain pace inhumans. Efforts in mouse models to define the non-neo-plastic origins of clonal malignancies, to elucidate thecytokine milieu conducive to expansion and mutationand to pinpoint the critical inherited polymorphisms per-missive for both diseases hold great promise for better-ing the care and treatment of those with autoimmunedisorders.

1. Ioannidis J.P., Vassiliou V.A., Moutsopoulos H.M. Long-termrisk of mortality and lymphoproliferative disease and predictiveclassification of primary Sjogren’s syndrome. Arthritis Rheum.2002;46:741–7.

2. Davidson W.F., Giese T., Fredrickson T.N. Spontaneous devel-opment of plasmacytoid tumors in mice with defective Fas-Fasligand interactions. J Exp Med. 1998;187:1825–38.

3. Morel L., Croker B.P., Blenman K.R., Mohan C., Huang G.,Gilkeson G., Wakeland E.K. Genetic reconstitution of systemiclupus erythematosus immunopathology with polycongenicmurine strains. Proc Natl Acad Sci USA. 2000;97:6670–5.

4. Ozaki et al.,(submitted for publication).

4.Diagnostic and pathogenic implications ofthe heterogeneity of antiendothelial cellantibodies

Christophe Dugue, Yves Renaudineau, Christophe Jam-ín, Pierre Youinou,(EA 2216 ‘Immunologie et Pathologie’, Laboratory ofImmunology, Brest University Medical School Hospital,BP824, F29609 Brest Cedex, France).Email: [email protected]

Despite differences in the pathophysiology of immune-mediated diseases, anti-endothelial cell antibodies(AECA) are commonly identified in a vast array of con-ditions associated with vascular injury. Not only do theseinclude systemic lupus eythematosus(SLE), Wegener’sgranulomatosis and rheumatoid arthritis, but an increas-ing number of disease states are reported to be associ-ated with AECAs. Their levels parallel the disease


activity, so that such autoantibodies are believed to bepathogenic. In contrast, the antigen specificities ofAECAs remain an unresolved enigma. Some have, how-ever, been pinpointed, such as heat shock protein 60,most notably in SLE.

Keywords: endothelial cell, antiendothelial cellautoantibody

Since IgG reacting with endothelial cells(EC) wereidentified on kidney biopsy specimens from patientswith systemic lupus erythematosus(SLE), a good dealof controversy has been sparked over the diagnostic andpathogenic implications of these anti-EC-antibodies(AECA). The impressive diversity of settings encoun-tered suggests that they constitute an extremely hetero-geneous family of autoantibodies. Recent data provideinsight into conceptualization of the mechanisms bywhich the EC function may be influenced. Stimulatingpredictions derive from these findings, and justify inten-sive studies currently in progress to recognize their tar-get antigens(Ag).

Disease associations

AECAs have frequently been reported in a vast array ofautoimmune states, of which the sole common denom-inator is the presence of an immune-mediated inflam-mation of the vessel walls. The quintessentialrepresentative of such conditions is SLE, where AECAsexist in 15 to 85% of the casesw1x. They have subse-quently been described in rheumatoid arthritis(RA),when associated with vasculitis, in idiopathic inflam-matory myopathies, such as polymyositis(PM) and der-matomyositis, and in systemic sclerosis(SS),particularly those at risk of developing vascular crises.AECAs can also been found in patients with diabetesmellitus, coronary artery disease following cardiac trans-plantation, uremia treated with hemodialysis, hyperpro-lactinemia, anterior uveitis and borderline hypertension.Whether the AECA titer increases are predictive of clin-ical conditions worsen warrants address in follow-upstudies, but shows promise as a marker for ongoingmyocardial infarction. It is noteworthy that AECAs aresignificantly associated with antiphospholipid(PL) anti-bodies in a number of connective tissue diseasesw2x.The interest for further analysis of infection-inducedAECAs has also been recently revived by the report thatMycobacterium leprae, cytomegalovirus and denguevirus colonize ECs. They contribute to the pathophy-siology of the related vasculitidis, and could encouragethe development of autoimmune complications.

Although of unknown significance, AECAs are a com-mon finding in hepatitis C, even more in the presenceof mixed cryoglobulinemia. They have also been foundin under half of the sera from infective endocarditis andover half of those from leprosy. In the latter infection,they were found across the whole spectrum of the dis-ease, but preferentially associated with its multibacillarythan its paucibacillary variants.

Pathogenic effects

Although the relevance of AECAs has long remaineduncertain, an experimental model of systemic vasculitishas, based on autoantibody idiotype, provided evidencesuggesting that they are pathogenicw3x. In support ofthis view, AECAs are associated with vasculitis in RA,lung sclerosis in PM and renal failure in SLE. In thisrespect, it is interesting that the autoantibody levels fluc-tuate with disease activity in patients with SLE, Wegenergranulomatosis(WG) and Kawasaki syndrome.ECs can be activated by such autoantibodies. AECA-containing sera may thus raise the density of adhesionmolecules, and promote the synthesis of cytokines.Hence, speculation about the mechanism of vascularconditions associated with AECAs has focused on endo-thelial expression of adhesion molecules. They have alsothe capacity to augment the production of tissue factor,and thereby to favour coagulation. Another appealingproperty of AECAs is to encourage ECs to undergoapoptosis, particularly in SSw1x and leprosy. Presumably,activation does not constitute a prerequisite for AECA-mediated apoptosis of ECs. There is, however, everylikelihood that AECAs act as opsonins to facilitateremoval of injured cells.Curiously, AECAs are associated with anti-PLs antibod-ies directed at complexes of anionic PL-binding pro-teins, such asb2-glycoprotein I. Apoptosis-inducingAECAs are responsible for PL reaching the surface ofthe cell. Their defective clearance by macrophages maycontinuously challenge T cell tolerance and lead to anti-PL antibody production.

Antigen specificity

Taken together, impaired clearance and possible abnor-mal processing of apoptotic ECs, through unveiling ofcryptic epitopes from these proteins, may drive previ-ously ignorant self-reactive T cells to provide help toself-specific B cells, resulting in the production of path-ogenic autoantibodies. The heterogeneity of the diseasepresentation reflects indeed the variety of AECAs. Cer-tain Western blot patterns have been claimed to be spe-


cific for RA, SLE or SS, but these results vary from onestudy to another.A variety of autoantibodies may even bind to ECsthrough electric charges, whereas their contribution tothe EC reactivity is considered negligible. These includeanti-double-stranded DNA, anti-PL, and anti-heparansulfate antibodies. They may also react with the extra-cellular matrix or the glycosaminoglycan part of extra-cellular receptors. Most investigators agree that AECAsand antineutrophil cytoplasmic antibodies are two dis-tinct antibody populations. Both may have EC-bindingproperties, though they are not overlapping Ag targets.Identification of genuine Ag deserves further study. Thetwo most promising approaches are EC complementaryDNA expression libraries and two-dimensional electro-phoreses(2 DE) of EC lysates. Those from patients withleprosy appear to react with calreticulin, vimentin, tub-ulin, and heat-shock protein(HSP)-70. Distinct speci-ficities have been identified in SLE, such as AKAP 350,actin gamma 1, NADH deshydrogenase, ADH andmethyltransferase. Furthermore, using 2-DE, a panel ofeight Ag were shown to be involved in SLE. This doesnot imply that all SLE sera are positive for all Ag, butthat these specificities are found together exclusively inSLE. Among them, HSP-60 reactivity would be relatedto vasculitis as suggested by its presence in arterial dis-eases, including WG,Periarteritis Nodosa and Behcet¸disease, but not in various nonorgan-specific autoim-mune diseases, e.g. SS, PM and RA.

Given unacceptable discrepancies between the diseasesin a given group of investigators, and between thegroups of investigators in one and a single disease, thereremains a crucial need for a reliable assay. Nonetheless,the forthcoming identification of a handful of Ag couldpermit the development of ELISAs specific for diseasesassociated with specific AECAs. For the time being, itis consequent to conclude that most of the findings raisethe issue of the physiological role of natural AECAs inthe removal of unwanted damaged ECs, thereby pre-venting autoimmune vasculitis.

References

1. Bordron A., Revelen R., Youinou P. Anti-endothelial cell auto-´ ántibodies and systemic disease. Isr. Med. Assoc. J. 2000;2:544–9.

2. Meroni P.L., Raschi E., Testoni C., Tincani A., Balestrieri G.,Youinou P. Antiphospholipidyendothelial cell interaction in thepathogenesis of the antiphospholipid syndrome. In: AshersonR.A., Cervera R., Piette J.C., Shoenfeld Y., editors. The anti-phospholipid syndrome II: autoimmune thrombosis. Amster-dam, Elsevier Science, 2002:79–89.

3. Damianovich M., Gilburd B., George J., et al. Pathogenic roleof anti-endothelial cell antibodies in vasculitis: an idiotypicexperimental model. J. Immunol. 1996; 156:6946–51.


RHEUMATOID ARTHRITIS: PATHOGENETIC ASPECTS

5.Cell-cell interactions and tissue damage inrheumatoid arthritis

J.M. Dayer and D. Burger,(Service of Immunology and Allergy, University Hos-pital, 24 Rue Micheli du Crest, CH-1211 Geneva 14,Switzerland).Email: [email protected]

In many chronic inflammatory diseases such as rheu-matoid arthritis(RA), inflammation is characterized bythe migration into the target tissue of T and B lympho-cytes, dendritic cells, neutrophils, mast cells and mac-rophages(Mf). This process is associated with theproliferation of invading and resident cells(i.e. synovio-cytes and endothelial cells) leading to destruction andremodeling of the extracellular matrix. The destructionof the organic phase by proteases, mainly metallopro-teinases(MMP), is accompanied by the resorption ofthe inorganic phase of bone, mainly due to the action ofthe receptor activator of NFkB and its ligand(RANK-RANKL) and prostanoids. The expression of proteasesand their inhibitors is regulated by various stimuli,including soluble factors(i.e. cytokines, hormones),contact with extracellular matrix components and directcellular interactions(1). The activity of pro-inflamma-tory cytokines(i.e. TNF, IL-1) and MMPs is counter-balanced by numerous mechanisms of which cytokineinhibitors-IL-1 receptor antagonist(IL-1Ra), IL-1 solu-ble receptor(IL-1sRII), TNF soluble receptor(TNFsR)-and tissue inhibitor of MMP(TIMP). It is generallyacknowledged that the imbalance between cytokines andtheir respective inhibitors is responsible for the persist-ence of chronic inflammation and maybe even necessaryfor its initiation. As demonstrated by human clinical tri-als, there is now considerable evidence that cytokinessuch as TNF and IL-1 in addition to other cytokines(i.e.IL-6, IL-15, IL-17, IL-18) are involved in RA patho-genesis(Fig. 1).

T cell-monocyte interactions

Although infiltration of T lymphocytes into the targettissue precedes joint damage-suggesting their pathogeniceffect – Mf, the main producers of IL-1 and TNF, arealso present at an early stage in the lesion, and interac-tions are likely to occur between T lymphocytes and

Mf. As summarized in Fig. 1, our studies strongly arguethat direct cellular contact with stimulated T cells is amajor pathway for the production of IL-1 and TNF inMf (1). Indeed, contact-mediated activation of Mf bystimulated T cells is as potent as optimal doses of LPSin inducing IL-1b and TNF production in Mf and ishighly relevant to the pathogenesis and maintenance ofchronic destructive, inflammatory reactions in RA.Most T cell types including T cell clones, freshly iso-lated T lymphocytes and T cell lines such as HUT-78cells induce IL-1 and TNF in Mf. Various stimuliinduce T lymphocytes to activate monocytes by directcellular contact, including PHAyPMA, cross-linking ofCD3 with or without cross-linking of the co-stimulatorymolecule CD28, antigen-recognition on antigen-specificT cell clones, and cytokines. Furthermore, depending onT cell type and T cell stimulus, cellular contact withstimulated T lymphocytes can induce different patternsof products in Mf. This suggests that multiple ligandsand counter-ligands are involved in the contact-mediatedactivation of Mf, which are differentially induced in Tcells depending on the stimulus. Besides, upon contactwith stimulated T cells the balance between IL-1b andIL-1Ra production in Mf was ruled by SeryThr phos-phatase(s) and contact-activated THP-1 cells expressedmembrane-associated protease(s) neutralizing TNFactivity both by degrading the latter cytokine and bycleaving its receptors at the cell surface. Thus the trig-gering of these intra- and extra-cellular processes bydirect contact with stimulated T lymphocytes may reg-ulate the pro-inflammatory cytokines and their inhibi-tors, and the balance of their production in Mf dictatesin part the outcome of the inflammatory process. Alongthese lines we demonstrated that PI3 kinase(PI3K) wasan important signalling effector regulating the produc-tion of pro- and anti-inflammatory cytokines in humanMf. Indeed, PI3K was involved in the repression of IL-1b and the induction of IL-1Ra in isolated human bloodmonocytes upon contact with stimulated T cells. ThusPI3K represents a key effector that might be dysregu-lated in pathological conditions.The identity of the molecules on the T cell surface, thatare involved in contact-mediated signaling of Mf acti-vation as well as their counter-ligands remain to be iden-tified. However, several ligands and counterligands havebeen involved in the contact-mediated activation of Mf

(i.e. LFA-1yICAM-1, CD2yLFA3, CD40yCD40L).These studies suggest that some known surface mole-


cules are involved in T cell signaling of Mf. However,inhibitors (e.g. antibodies) of these molecules fail toabolish monocyte activation altogether, suggesting thatthe factor(s) required for T cell signaling of human Mfby direct contact remain(s) to be identified. Further-more, none of these molecules is known to interact withhigh-density lipoproteins(HDL) or their major proteincomponent apolipoprotein A-I(apo A-I), which wereshown to inhibit both TNF and IL-1b production in Mf

activated by cellular contact with stimulated T cells(3).

T cell–fibroblast interactions

Cellular contact induces an imbalance between MMP-1and TIMP-1 production by dermal fibroblasts(4). Sincedirect cell-cell contact with stimulated T lymphocytesinduced an imbalance between the production of MMPsand TIMP-1 by synoviocytes in vitro, it may, in analogy,favor tissue destruction in vivo. In addition to MMP-1and TIMP-1, direct cell-cell contact with stimulated Tlymphocytes induces PGE production on human dermal2

fibroblasts and synoviocytes(4) that also contribute totissue destruction by favoring bone resorption. The Tcell surface molecules involved in MMP induction infibroblasts are mainly membrane-associated IL-1 andTNF. Similarly, membrane-associated cytokines areinvolved in the inhibition of deposition of the majorextracellular matrix components such as types I and IIIcollagen. Indeed, direct contact with either plasma mem-branes or fixed, stimulated T cells markedly inhibitedthe synthesis of types I and III collagen in dermal fibro-blasts and synoviocytes, whether untreated or treatedwith transforming growth factor-b (TGFb). This inhi-bition was associated with a marked decrease in steady-state levels of pro-a I and III collagen mRNAs, whichwas due to a diminished transcription rate. This inhibi-tion of extracellular matrix production mediated by Tcell contact was partially due to additive effects of Tcell membrane- associated IFN-g, TNF, and IL-1a (5).Thus, direct contact with stimulated T cells favors extra-cellular matrix catabolism by enhancing MMP produc-tion while diminishing collagen synthesis in fibroblastsand synoviocytes. Interestingly, similar membrane-asso-ciated cytokines were involved in both these processes.Conclusions: Direct cell–cell contact with stimulated Tlymphocytes is one of the principal pathways triggeringactivation of Mf in the absence of infectious agents,which suggests that it is an important mechanism inchronic inflammatory diseases of autoimmune etiologyincluding RA. Many more investigations are needed toidentify the surface the surface molecules-ligands andcounter-ligands-involved in this process. This may pro-

vide the basis for the development of novel agents inter-fering with the inflammatory response induced by cell–cell contact and leading to tissue destruction in chronicinflammatory diseases.

References

1. Burger D., Dayer J.M. The role of human T lymphocyte-mon-ocyte contact in inflammation and tissue destruction. ArthritisRes. 2002; 4(suppl. 3):S169–S176.

2. Burger D. Cell contact-mediated signaling of monocytes bystimulated T cells: a major pathway for cytokine induction. Eur.Cytokine Netw. 2000; 11:346–353.

3. Hyka N., Dayer J.M., Modoux C., Kohno T., Edwards C.K.,III, Roux-Lombard P. et al. Apolipoprotein A-I inhibits the pro-duction of interleukin-1beta and tumor necrosis factor-alpha byblocking contact-mediated activation of monocytes by T lym-phocytes. Blood 2001; 97:2381–2389.

4. Burger D., Rezzonico R., Li J.M., Modoux C., Pierce R.A.,Welgus H.G. et al. Imbalance between interstitial collagenaseand tissue inhibitor of metalloproteinases 1 in synoviocytes andfibroblasts upon direct contact with stimulated T lymphocytes:involvement of membrane-associated cytokines. ArthritisRheum. 1998; 41:1748–1759.

5. Rezzonico R., Burger D., Dayer J.M. Direct contact between Tlymphocytes and human dermal fibroblasts or synoviocytesdown-regulates types I and III collagen production via cell-associated cytokines. J. Biol. Chem. 1998; 273:18 720–18 728.

Fig. 1: Scheme of the activation cascade from Tlymphocytes(T) to monocyte-macrophages(Mf)and fibroblastsysynoviocytes(FyS) that leads totissue destruction in RA. Activated T trigger MfL

to produce pro-inflammatory cytokines that in turninduce the production of matrix-destructive metal-loproteinases (MMPs) and prostaglandin E ,2

(PGE ), the latter products being involved in car-2

tilage destruction and bone resorption. These pro-cesses are controlled by pro-inflammatory factors(IL-15, IL-2, IL-18, IL-17) and anti-inflammatory


factors (IL-4, IL-10, GM-CSF, IFNb). Further-more, naturally occurring inhibitors(IL-1sRII, IL-1Ra, TNFsRI, TNFsRII) inhibits the activity ofIL-1 and TNF-a, which production is blocked byHDLyapo A-I and decreased by exogenous anti-bodies to CD69 andb2-integrins (CD11b)CD11c)CD11a).

6.BiP: a new biologic immunomodulator forthe treatment of rheumatoid arthritis

Gabriel Panayi and Valerie Corrigall,(Academic Department of Rheumatology, King’s Col-lege London, Guy’s Hospital, London SE1 9RT, UK).Email: [email protected]

Introduction

Stress proteins are induced by a number of factors thatinclude heat, hypoxia, low glucose, and reactive oxygenspecies. During periods of stress, stress protein transcrip-tion is increased. Stress proteins protect the cell from theharmful effects of the induced stress including apoptoticcell death. This is the first function assigned to stressproteins. More recently, it has been realised that stressproteins may be released into the extracellular environ-ment where they acquire distinct and novel functions.They act as ‘‘chaperokines’’ broadcasting stress in tis-sues and activating the innate immune system. Stressproteins are able to activate the innate immune systembecause they interact with specific cell receptors onmonocytes and a variety of other cells.Glucose regulated proteins 78(GRP78) or BiP is amember of the HSP70 family of stress proteins. It is notinduced by heat but by the other stresses mentionedabove. It is an endoplasmic reticulum(ER) protein andits function is to allow correct folding of nascent poly-peptides as they emerge from the ER. We have dem-onstrated that it can be found in the synovial fluid ofpatients with rheumatoid arthritis and this raises theintriguing possibility as to its extracellular as distinctfrom its intracellular functions.

The extracellular functions of BiP

We have described a series of extracellular functions ofBiP both in human as well as in animal experimentalmodels.

Effect of BiP on human T cell and monocyte function

● When human monocytes are cultured with BiP theystimulate a large secretion of interleukin 10, up to 5ngyml within 48 h, and this secretion can persist forup to 5 days. We have evidence that this stimulationof interleukin 10 is via a putative BiP receptor butthe nature of this receptor is presently unknown.

● BiP is able to induce the proliferation of synovial T-cells predominantly from patients with rheumatoidarthritis. The degree of proliferation is low and nointerferon gamma is produced. However the respond-ing cells produce interleukin 10. We have been ableto clone CD8 positive BiP specific T-cells from theperipheral blood of normal human subjects. Thesecells are able to secrete interleukin 10,y4 andy5.Their regulatory function is presently under investi-gation.

● BiP is able to inhibit the development of human den-dritic cells. When human monocytes are cultured inthe presence of GM-CSF and interleukin 4 theyundergo differentiation into dendritic cells. At the endof 5 days, they have undergone maturation intoimmature dendritic cells. We have shown that BiP butnot another recombinant protein, such as theb-glu-curonidase, is able to inhibit the development ofimmature dendritic cells from their precursors.

We postulate that the combined effect of these twoextracellular functions of BiP will be to switch from aTH1 immune response to a TH2 immune response withanti-inflammatory consequences. This can only be ascer-tained by animal experiments.

Effect of BiP in mouse experimental systems

● BiP is not able to induced arthritis when given alongwith Freund’s complete adjuvant to several strains ofmice and rats. However, when BiP is given intrave-nously or subcutaneously to DBAy1 or HLA-DR1 transgenic mice it is able to inhibit theqyq

development of collagen-induce arthritis(CIA).There is a reduction in the antibodies to type 2 col-lagen and the isotype of the antibodies made suggestthat there has been a switch of the immune responsefrom TH1 to TH2.

● Subsequent experiments have shown that intravenousor subcutaneous BiP can treat ongoing CIA both inDBAy1 and in HLA-DR1 transgenic mice. Theqyq

nature of the anti-collagen antibody response suggests


that there has been a switch from a TH1 to a TH2immune response here also. This is an exciting resultas it suggests that BiP may be useful for the treatmentof patients with rheumatoid arthritis.

Conclusion

BiP although a member of the HSP70 family, unlikeother members of that family, has anti-inflammatory andimmune regulatory properties in its extracellular form.Two conclusions may be drawn from these observations.The first is that within the stress protein universe thereare pro- and anti-inflammatory molecules that are impor-tant, in the former case, for initiating innate immunityand hence modulating adaptive immunity and, in the lat-ter case, of down regulating the innate immune responseand alter the subsequent nature of the adaptive immuneresponse. The second important conclusion is that BiPmay be a novel immunomodulatory drug being able tocontrol immune reactions, such as those taking place inthe rheumatoid joint, not by inhibiting pro-inflammatorycytokine activity but by modulating both the inflam-matory response itself as well as the nature of the adap-tive immune response taking place within the joint.

References

1. Corrigall V.M., Panayi G.S. Autoantigens and immune path-ways in rheumatoid arthritis. Crit. Rev. Immunol. 2002;22:281–93. Review.

2. Panayi G.S., Corrigall V.M., Pitzalis C. Pathogenesis of rheu-matoid arthritis. The role of T cells and other beasts. Rheum.Dis. Clin. North Am. 2001;27:317–34. Review.

3. Bodman-Smith M.D., Corrigall V.M., Kemeny D.M., PanayiG.S. BiP, a putative autoantigen in rheumatoid arthritis, stimu-lates IL-10-producing CD8-positive T cells from normal indi-viduals. Rheumatology(Oxford). 2003;42:637–44.

4. Corrigall V.M., Bodman-Smith M.D., Fife M.S., Canas B.,Myers L.K., Wooley P., Soh C., Staines N.A., Pappin D.J., BerloS.E., van Eden W., van Der Zee R., Lanchbury J.S., Panayi G.S.The human endoplasmic reticulum molecular chaperone BiP isan autoantigen for rheumatoid arthritis and prevents the induc-tion of experimental arthritis. J. Immunol. 2001;166:1492–8.

5. Corrigall V.M., Bodman-Smith M.D., Brunst M., Cornell H. andPanayi. G.S. Inhibition of antigen-presenting cell function andstimulation of human peripheral blood mononuclear cells toexpress an anti-inflammatory cytokine profile by the stress pro-tein, BiP. Arthritis Rheum. 2004;50:1164–71.

7.Anti-CCP antibodies: the new rheumatoidfactor in the serology of rheumatoidarthritis

Walther J. van Venrooij, Erik R. Vossenaar and AlbertJ.W. Zendman,(Department of Biochemistry 161, University of Nij-megen, Nijmegen, The Netherlands).Email: [email protected]

Introduction

Since almost 60 years, the serology of rheumatoid arthri-tis (RA) depends primarily on the testing of the rheu-matoid factor(RF). However, RF is not very specificfor RA since it is merely a marker of inflammation andtherefore present in many inflammatory diseases andeven in healthy controls, especially in the elderly. As aconsequence there is a need for other, more specificserological tests. Such a test should be very specific forRA, present in the majority of RA patients and usefulfor the detection of RA in the very beginning of thedisease. Ideally, it should also have some prognosticabilities, that is it should indicate whether erosive dis-ease is developing or not. In this short review, we willpresent evidence that the anti-CCP(cyclic citrullinatedpeptide) antibody system is such a specific and usefulmarker for(early) RA.

The CCP2 system is specific and sensitive

RA sera show a remarkable variety in the reactivity pat-tern towards different citrulline-containing peptides,indicating that the amino acids flanking the citrullineresidue are important for the antigenicity of the epitopew1x. To increase the low sensitivity of these linear cit-rulline-containing peptides, they were modified to adapta more stringent structure in which the citrulline moietyis optimally exposed for antibody bindingw2x. With asingle filaggrin-derived cyclic citrullinated peptide(CCP), antibodies could be detected in about 68% ofRA sera with a very high specificity(98%) w2x. Thiscyclic peptide was used as the antigenic substrate in theCCP1 test, which had only a moderate sensitivity of 50–70% w3x. To improve the sensitivity, a dedicated libraryof citrulline-containing peptides was screened for addi-tional antigenic peptides. This work finally culminatedin the present CCP2 test that is reactive with more than80% of the RA sera while maintaining the superior spec-ificity of 98% (reviewed inw4–6x.


Clinical use of the CCP2 test

A serological marker for disease is useful for the clini-cian when the antibodies directed to it are present in themajority of patients and are highly specific for the dis-ease. The data given in Table 1 illustrate that the CCP2system is highly specific for RA with a sensitivity com-parable to the RF(see alsow4–6x). However, a goodmarker antibody should also be present early in the dis-ease and able to predict disease outcome.When a random population of a rheumatology clinic istested for anti-CCP, about 2–5% of the patients test pos-itive but do not suffer from RA(our unpublished obser-vations). Two recent studies have provided clearevidence that such supposedly ‘false’ positives might bein the early phase of developing RA. Rantapaa-Dahlqv-¨ ïst and collaboratorsw7x analyzed blood samples from83 blood donors that later developed RA. Anti-CCP2antibody could be detected in some patients many years(25% on average 1.5 to 9 years before onset of the firstsymptoms of the disease) before the first clinical symp-toms. The percentage of CCP2 positivity increasedsharply in the last 1.5 years before the first diseasesymptoms. Anti-CCP2 positivity was 52% in the lattergroup. More than 70% of the patients were anti-CCP2positive at their first visit to the clinicw7x. Nielen andcoworkersw8x measured anti-CCP1 and IgM-RF levelsin serial blood samples of 72 blood donors that laterdeveloped RA. Anti-CCP positivity could be observedup to 14 years before the first clinical symptoms and41% of the patients were CCP positive at presentationto the clinician. For IgM-RF the corresponding para-meters were 10 years and 28% positivity. Thus, anti-CCP detected more positive subjects and longer beforethe start of the complaints compared to IgM-RFw8x. Vit-tecoq and coworkers studying 314 early arthritis patientsconfirmed these results. At one year follow-up 90% ofthe patients that were CCP positive at baseline wereclassified as established RA patientsw9x. The conclusionfrom these studies is that anti-CCP antibodies are presentearly in disease, and that their presence is able to accu-rately predict the development of RA.There are also several studies indicating that both theanti-CCP1 and anti-CCP2 systems are able to predicterosive disease(seew5, 6x). However, the obvious ten-dency that anti-CCP antibody is present preferentially inpatients with erosive disease cannot exclude the fact thatthis may not be the case in an individual patient. Howcan we be sure that(erosive) RA is developing in aCCP-positive individual without serious complaints?The answer to this question certainly needs additionalstudy but has recently been approached by Berglin and

coworkersw10x. In the study of Rantapaa–Dahlqvistw7x¨ ¨the Odds Ratio(OR) for anti-CCP2 to predict devel-opment of RA was about 25, a high number when com-pared to the ORs of RF(about 2) and shared epitope(SE, about 2). The presence of 2 or 3 of these para-meters in an individual patient drastically increases thechance that indeed RA is developing. Berglin andcoworkersw10x, using the patient material from the Ran-tapaa-Dahlqvist studyw7x, found that the combination of¨ ¨RF and SE gave an OR of about 40, and that the com-bination of SE and anti-CCP2 positivity resulted in anelevated OR for predicting future development of RA ofalmost 67. These ORs are even higher than could beexpected upon multiplification of the single parameterORs. These data not only show that these parametersmay act in a synergetic way, but also indicate that acombination of serological and genetic factors in thefuture will be able to predict the development of(ero-sive) RA in an individual with high accuracy. A futuregenetic and serological RA passport could thus be ofconsiderable help in the decision whether treatmentshould be applied or not.

Table 1. Sensitivity and specificity of the anti-CCP2 andIgM-RF antibody systems(taken fromw5x).

Rheumatoid Arthritis CCP2 IgM-RF

n pos (%) n pos (%)

1117 865 (77) 1119 827 (74)

Healthy individuals 431 1 (0) 335 38 (11)Osteoarthritis 81 7 (9) 79 14 (18)Juvenile rheumatoid arthritis 21 6 (29) 21 8 (38)Various connective tissue 814 44(5) 573 217 (38)diseases

Various inflammatory diseases 225 2 (1) 160 14 (9)Various arthropathies 297 13 (4) 72 5 (7)Various viral infections 188 2 (1) 171 43 (25)Various bacterial infections 147 2 (1) 179 21 (12)Various parasitic infections 93 2 (2) 99 22 (22)Total non-RA 2297 79 (3) 1689 382 (23)Specificity 97% 77%

References

1. Schellekens G.A., de Jong B.A., van den Hoogen F.H., van de

Putte L.B., van Venrooij W.J. Citrulline is an essential con-

stituent of antigenic determinants recognized by rheumatoid

arthritis-specific autoantibodies.J. Clin. Invest. 1998,

101:273–281.


2. Schellekens G.A., Visser H., de Jong B.A., van den Hoogen

F.H., Hazes J.M., Breedveld F.C., van Venrooij W.J. The diag-

nostic properties of rheumatoid arthritis antibodies recognizing

a cyclic citrullinated peptide.Arthritis Rheum. 2000, 43:155–

163.

3. van Boekel M.A., Vossenaar E.R., van den Hoogen F.H., van

Venrooij W.J. Autoantibody systems in rheumatoid arthritis:

specificity, sensitivity and diagnostic value.Arthritis Res.

2002, 4:87–93.

4. van Venrooij W.J., Hazes J.M., Visser H. Anticitrullinated pro-

teinypeptide antibody and its role in the diagnosis and prog-

nosis of early rheumatoid arthritis.Neth. J. Med. 2002,

60:383–388.

5. Vossenaar E.R., van Venrooij W.J. Anti-CCP antibodies, a spe-

cific marker for(early) rheumatoid arthritis.Clin. Appl. Immu-

nol. Rev., in press.

6. Zendman A.J.W., Vossenaar E.R., van Venrooij W.J.: Auto-

antibodies to citrullinated(poly)peptides: a key diagnostic and

prognostic marker for rheumatoid arthritis.Autoimmunity, in

press.

7. Rantapaa-Dahlqvist S., de Jong B.A., Berglin E., Hallmans G.,¨ ¨

Wadell G., Stenlund H., Sundin U., van Venrooij W.J. Anti-

bodies against citrullinated peptide and IgA rheumatoid factor

predict the development of rheumatoid arthritis.Arthritis

Rheum. 2003, 48:2741–2749.

8. Nielen M.M.J., van Schaardenburg D., Reesink H.W.R., van

de Stadt R.J., van der Horst-Bruinsma I.E., de Koning

M.H.M., Habibuw M.R., Vandenbroucke J.P., Dijkmans B.A.

Specific autoantibodies precede the symptoms of rheumatoid

arthritis: a study of serial measurements in blood donors.

Arthritis Rheum. 2004, 50:380–386.

9. Vittecoq O., Incaurgarat B., Jouen-Beades F., Legoedec J.,

Letourneur O., Rolland D., Gervasi G., Menard J.F., Gayet A.,

Fardellone P., Daragon A., Jolivet M., Le Loet X., Tron F.

Autoantibodies recognizing citrullinated rat filaggrin in an

ELISA using citrullinated and non-citrullinated recombinant

proteins as antigens are highly diagnostic for rheumatoid

arthritis. Clin. Exp. Immunol. 2004, 135:173–180.

10. Berglin E., Padyukov L., Hallmans G., van Venrooij W.J., Kla-

reskog L., Rantapaa-Dahlqvist S. Presence of the shared epi-¨ ¨

tope genes increase the predictive value of antibodies against

cyclic citrullinated peptide(CCP) for rheumatoid arthritiswabstractx. Arthritis Rheum. 2003, 48:S678.


RHEUMATOID ARTHRITIS: CLINICAL PICTURE AND THERAPEUTIC CONSIDERATIONS

8.Epidemiology of rheumatoid arthritis

A.A. Drosos,(Rheumatology Clinic, Department of Internal Medi-cine, Medical School, University of Ioannina, 45110,Ioannina, Greece).Email address: [email protected]

Keywords: Rheumatoid arthritis; Epidemiology; Sharedepitope; Rheumatoid factor; Age, Gender

Take–home messages

● RA in Greece and in the Mediterranean area has aregionally determined genetic and clinical profile.

● In Mediterranean countries RA seems to be mild,with less EAM and less radiological damage.

● It seems that age at disease onset and gender do notinfluence disease expression and outcome.

● Factors that may contribute to this different clinicalprofile are environmental(sun exposure, ultravioletradiation) and dietary like olive oil, fish consumptionand the Mediterranean diet.

Introduction

Rheumatoid arthritis(RA) is a chronic inflammatorydisease affecting the synovial membrane and causes sub-stantial morbidity and mortalityw1x. The prevalence ofthe disease approaches between 0.5 and 1% in the adultpopulation in most western countriesw2x. The etiologyof RA is obscure, however previous studies implicateboth genetic and environmental factorsw3x. AlthoughRA is often a disease of middle–aged women, its inci-dence increases through the seventh decade of lifew4x.Changes over time in the incidence and in the age andsex distribution of RA might implicate specific environ-mental factors and suggest new approaches whenenquiring to the etiology. Furthermore the severity andoutcome of RA have been shown to be influenced by avariety of factors such as gender, age at disease onset,shared epitope(SE), the presence of rheumatoid factor(RF) and othersw5, 6x.

Epidemiological approach of RA in Greece

Considerable clinical heterogeneity exists amongpatients with RA and it has been proposed that instead

of a simple disease, RA may be better classified as asyndromew7x. The importance of genetic factors hasbeen shown by studies of HLA associations with RA.The strongest association has been reported with theHLA–DRb1 alleles called the SEw8x. The different dis-ease phenotype reported among RA patients may be dueto genetic differences and to unknown environmentalfactors among populations. In early 1980, we havereported that RA in Greece differs clinically and serol-ogically grounds from patients of northern Europeancountriesw9x. This observation prompted us to set up acomparative clinical study to investigate the similaritiesand differences in RA in Greek and British hospital pop-ulations as the basis for a subsequent molecular analysisof HLA association in these two populations. We foundthat British patients had more severe articular diseaseand had more frequently EAM and more severe radio-logical damagew10x. In a subsequent DNA analysis, wefound that 57% of the Greek patients lack the putativeHLA–DRb motif, which suggests that considerableimmunogenetic heterogeneity underlies disease suscep-tibility in this population w11x. Our results were thenconfirmed by other investigators in the Mediterraneanarea who reported that RA in Mediterranean countriesis milder, with less EAM and less radiological damagew12,13x.The above observations prompted us to investigate theprevalence and incidence of RA in the Ioannina districtof northwest Greece, during the period of 1987–1995.A total of 428 cases of RA were identified during thestudy period. We found that the total prevalence for wasmen 2.05 and for women 4.75 casesy1000 inhabitants,and the total womenymen ratio was 2.33. Annual inci-dence rates fluctuated between 0.15 and 0.36y1000inhabitants. These findings suggested a low prevalenceand low incidence of RA in northwest Greecew14x. Ina subsequent study we examined the strength of theassociation between SE and RA susceptibility, articulardisease severity and EAM in Mediterranean populations.We found that SE alleles were significantly more com-mon in RA patients than in controls(odds ratiowORx2.5; 95% confidence intervalwCIx 1.4–4.3). Larsenradiologic score was predicted by SE and disease dura-tion. However, SE did not increase the risk of any EAM.The meta–analysis showed a pooled OR of 3.7(95%CI, 2.6–5.2) for susceptibility to RA conferred by SE(OR 3.4 vs. 3.9 in Greek vs. non-Greek populations).These results suggest that SE determine articular


destruction without increasing the risk of EAM. Theimmunogenetic association of RA susceptibility is con-sistent, but their strength may depend on the SE preva-lence in different ethnic groupsw15x.To evaluate whether RF influence disease severity andoutcome, we examined 417 patients with early RA(dis-ease duration less than one year) during the period1981–1999 using a database. This study showed that RFseems to be a predictor of more severe disease activityw16x. In a subsequent study, we investigated if age atdisease onset comprises a separate parameter for diseaseexpression, prognosis and outcome in early RA patients.For this reason, 438 patients(time period 1981–2001)were analyzed according to age at disease onset(youngpatients aged less than 60 years at disease onset vs eld-erly patients aged more than 60 years at disease onset).The results of this study showed that age at disease onsetdoes not influence the clinical course and outcome ofearly RA w17x.Finally, we investigated whether gender comprises a sep-arate parameter for disease expression, prognosis andoutcome in early RA patients. Thus, 438 patients wereanalyzed according to gender during the period 1991–2000. We studied 312 women and 126 men with earlyRA. The female to male ratio was 2.5:1. There were nodifferences between gender in the general symptoms,clinical, laboratory and radiological findings. It seemsthat sex does not influence disease expression, severityand outcomew18x.

Comments

The results of the above studies reinforce the notion thatRA in Greece has a regionally determined genetic andclinical profile w19–21x. Other factors that may contrib-ute to this different clinical profile are dietary factorssuch as olive oil and fish consumption or even the Med-iterranean diet could offer a protective effect of diseasedevelopment. In addition, sun exposure and ultravioletradiation may have immunosuppressive propertiesw22,23x. Finally, the milder climate conditions in the Medi-terranean area may also contribute to different environ-mental factors from those of the US and NorthernEuropean countries. However, the role of these factorsremains uncertain. Further study is needed to investigatethe possible role of environmental factors in the epide-miology of RA in the area.

References

1. Wolfe F., Mitchell D.M., Sibley J.T., et al. The mortality ofrheumatoid arthritis. Arthritis Rheum. 1994;37:481–94.

2. Hochberg M.C., Spector T.D. Epidemiology of rheumatoidarthritis: update. Epidemiol. Rev. 1990;12:247–52.

3. Seldin M.F., Amos C.I., Ward R., Gregersen P.K. The geneticsrevolution and the assault on rheumatoid arthritis. ArthritisRheum. 1999;42:1071–9.

4. Gabriel S.E., Crowson C.S., O’Fallon W.M. The epidemiologyof rheumatoid arthritis in Rochester, Minnesota, 1955–1985.Arthritis Rheum. 1999;42:415–20.

5. van de Heijde D.M., van Riel P.L., van Rijswijk M.H., van dePutte L.B. Influence of prognostic features on the final out-come in rheumatoid arthritis: a review of the literature. Semin.Arthritis Rheum. 1988;17:284–92.

6. Jantti J.K., Kaarela K., Luukkainen R.K., Kautiainen H.J. Pre-diction of 20-year outcome at onset of seropositive rheumatoidarthritis. Clin. Exp. Rheumatol. 2000;18:387–90.

7. Weyand C.M., Goronzy J.J. Association of MHC and rheu-matoid arthritis. HLA polymorphisms in phenotypic variantsof rheumatoid arthritis. Arthritis Res. 2000;2:212–6.

8. Gregersen P.K., Silver J., Winchester R.J. The shared epitopehypothesis. An approach to understanding the moleculargenetics of susceptibility to rheumatoid arthritis. ArthritisRheum. 1987;30:1205–13.

9. Moutsopoulos H.M., Giotaki H., Maddison P.J., MavridisA.C., Drosos A.A., Skopouli F.N. Antibodies to cellular anti-gens in Greek patients with autoimmune rheumatic diseases:anti–Ro(SSA) antibody a possible marker of penicillamine–D intolerance. Ann. Rheum. Dis. 1984;43:285–7.

10. Drosos A.A., Lanchbury J.S., Panayi G.S., MoutsopoulosH.M. Rheumatoid arthritis in Greek and British patients. Acomparative clinical, radiologic, and serologic study. ArthritisRheum. 1992;35:745–8.

11. Boki K.A., Panayi G.S., Vaughan R.W., Drosos A.A., Mout-sopoulos H.M., Lanchbury J.S. HLA class II sequence poly-morphisms and susceptibility to rheumatoid arthritis in Greeks.The HLA–DR beta shared–epitope hypothesis accounts forthe disease in only a minority of Greek patients. ArthritisRheum. 1992;35:749–55.

12. Ronda E., Rouiz M.T., Pascual E., Gibson T. Differencesbetween Spanish and British patients in the severity of rheu-matoid arthritis: comments on the article by Drosos et al.Arthritis Rheum. 1994;37:147–8.

13. Benazet J., Reviron D., Mercier P., Roux H., Roudier J. HLA–DRB1 alleles associated with rheumatoid arthritis in southernFrance. Absence of extraarticular disease despite expression ofthe shared epitope. J. Rheumatol. 1995;22:607–10.

14. Drosos A.A., Alamanos I., Voulgari P.V., et al. Epidemiologyof adult rheumatoid arthritis in northwest Greece 1987–1995.J. Rheumatol. 1997;24:2129–33.

15. Ioannidis J.P.A., Tarassi K., Papadopoulos I.A., et al. Sharedepitopes and rheumatoid arthritis: disease associations inGreece and meta–analysis of Mediterranean European popu-lations. Semin. Arthritis Rheum. 2002;31:361–70.


16. Papadopoulos I.A., Katsimbri P., Katsaraki A., TemekonidisT., Georgiadis A., Drosos A.A. Clinical course and outcomeof early rheumatoid arthritis. Rheumatol. Int. 2001;20:205–10.

17. Papadopoulos I.A., Katsimbri P., Alamanos Y., Voulgari P.V.,Drosos A.A. Early rheumatoid arthritis patients: relationshipof age. Rheumatol. Int. 2003;23:70–4.

18. Voulgari P.V., Papadopoulos I.A., Alamanos Y., Katsaraki A.,Drosos A.A. Early rheumatoid arthritis: Does gender influencedisease expression? Clin. Exp. Rheumatol. 2004(in press).

19. Drosos A.A., Moutsopoulos H.M. Rheumatoid arthritis inGreece: clinical, serological and genetic considerations. Clin.Exp. Rheumatol. 1995;13(suppl 12):S7–S12.

20. Andrianakos A., Trontzas P., Christoyannis F., et al. Prevalenceof rheumatic diseases in Greece: a cross-sectional populationbased epidemiological study. The ESORDIG Study. J. Rheu-matol. 2003;30:1589–1601.

21. Gorman J.D., Lum R.F., Chen J.J., Suarez–Almazor M.E.,Thomson G., Criswell L.A. Impact of shared epitope genotypeand ethnicity on erosive disease: a meta–analysis of 3,240rheumatoid arthritis patients. Arthritis Rheum. 2004;50:400–412.

22. Skoldstam L., Hagfors L., Johansson G. An experimentalstudy of a Mediterranean diet intervention for patients withrheumatoid arthritis. Ann. Rheum. Dis. 2003;3:208–214.

23. Ponsonby A.L., McMichael A., van der Mei I. Ultraviolet radi-ation and autoimmune disease: insights from epidemiologicalresearch. Toxicology 2002;181–182:71–78.

9.Multiple faces of rheumatoid arthritis:diagnostic and therapeutic algorithms

F. Breedveld,(University Hospital, Department of Rheumatology, P.O.Box 9600, 2300 RC Leiden, The Netherlands).Email: [email protected]

Aim: To evaluate the one-year clinical and radiologicaloutcomes of four treatment strategies for early RA, withtreatment adjustments based on disease activity scores(DAS ): (1) sequential monotherapy starting with44

methotrexate(MTX) up to 25 mgyweek, next treatmentstep sulphasalazin(SSA) 2000 mg, followed by leflu-nomide 20 mg;(2) step-up therapy from MTX, next stepadd SSA, then add hydroxychloroquine 400 mg;(3)step-down therapy from MTXqSSAqprednisone 60mg tapered to 7.5 mg, and(4) treatment with MTX and

infliximab 3 mgykg, dose increased or reduced to nildepending on DAS .44

Methods: The BeSt trial is a multicenter, randomized,single blinded trial in 508 patients with newly(-2 yearscomplaints) diagnosed RA(ACR 1987 criteria, at inclu-sionG6y66 swollen joints andG6y68 tender joints, andESRG28 mmyh or VAS global healthG20 mm) pre-viously not treated with DMARDs. Adjustment in treat-ment for each treatment strategy was dictated by threemonthly calculations of the DAS , with the goal to44

achieve a DASF2.4. DAS and health assessment44 44

questionnaires(HAQ) were obtained by blinded asses-sors. Sharpyvan der Heijde scores(SHS) were per-formed by two independent blinded physicians.Outcomes were calculated in an’intention to treat’-analysis.Results: At baseline, there were no significant differenc-es in patient characteristics between the groups. Afterone-year follow-up, patients in groups 3 and 4 had alower DAS , lower HAQ and a lower SHS than patients44

in groups 1 and 2(see table). Patients in groups 3 and4 improved earlier than patients in groups 1 and 2: at 3months 55% of patients in group 3 and 47% in group 4had a DAS F2.4, compared to 17% and 19% of44

patients in groups 1 and 2, respectively(P-0.001). TheHAQ also improved more and earlier in groups 3 and 4than in 1 and 2. There was no significant difference inthe number of drop-outs or in the number of seriousadverse events(SAE’s) between the groups. Numeri-cally, SAE’s occurred more often in group 3 than in theother groups.Conclusion: In early RA patients treated by a protocolwith three monthly treatment adjustments aiming at aDAS F2.4, initial treatment with combination therapy44

and initial treatment with infliximab plus MTX, resultsin a significantly better and faster clinical response aswell as in significantly less radiological damage thansequential monotherapy or step-up therapy.Outcomes after one-year follow-up

Group 1 Group 2 Group 3 Group 4 OverallMono- step-up combi- infliximab P-valuetherapy therapy nationns125 ns122 ns133 ns128

DAS F2.4 (% of pts)44 53 65 71 74 0.004DAS -1.6 (% of pts)44 28 30 34 36 0.614HAQ, mean change y0.7 y0.7 y0.9 y0.9 0.040Median SHS-progression 2.0 2.5 1.0 0.5 -0.001No SHS-progression 27 29 37 46 0.007(% of pts)

Discontinuation 5 6 6 2 0.494-1 year(no.of pts)

No. of serious 3 5 13 6 0.160adverse events


10.Autoimmunity and rheumatoid arthritis*

J.S. Smolen, S. Hayer, G. Schett, K. Redlich, M. Arin-ger, G. Kollias, E. Wagner, G. Steiner,(Medical University Vienna and Fleming Institute,Athens).Email: [email protected]

Rheumatoid arthritis(RA) is (i) an inflammatory jointdisease,(ii) of destructive nature; and(iii ) characterizedby autoimmune phenomena. One of the open questionsin its pathogenesis relates to the links between thesecharacteristics: which role does autoimmunity play inthe context of the inflammatory and destructive events?Rheumatoid factor(RF) is the first autoantibody everdescribed in a rheumatic disease. In the course of thepast decades it was shown that RF-positivity is relatedto disease severity and thus constitutes a bad prognosticmarker, that the presence of RF is associated with HLA-DR4, that RF can precede the evolution of clinical RAby many years, that RFq synovial fluids are commonlyhypocomplementemic suggesting the presence and aproinflammatory role of immune complexes, and thatthis autoantibody fluctuates with disease activity. How-ever, it is rarely found in experimental arthritis.Another important group of autoantibodies are thoseagainst citrullinated proteins, including a cyclic citrulli-nated peptide(CCP), which like RF are associated withdisease severity and also can precede disease. Again,these autoantibodies are not usually found in animalmodels of arthritis. Finally, an autoantibody against anuclear protein, hnRNP-A2, anti-RA33, can also befound in RA; while this autoantibody may not be a

marker of severe disease, it can also precede clinical RA,is accompanied by significant T-cell reactivity to theautoantigen which is even more frequent than thehumoral response, and can be found in a variety ofexperimental arthritides. Moreover, the antigen is over-expressed in the RA synovial membrane.Tumor necrosis factor(TNF) is a central proinflamma-tory cytokine. TNF-blockade is an effective therapy forRA. TNF-transgenic mice develop severe destructivearthritis. Interestingly, these mice do not express RF oranti-CCP, but bear anti-RA33 autoantibodies and theirsynovial membrane overexpresses hnRNP-A2. The auto-antibodies become apparent with the development ofarthritis, but do not occur or occur with highly decreasedfrequency with therapeutic interference with or blockadeof inflammation and destruction. When immune com-plexes are incubated with macrophages in vitro, the lat-ter secrete increased amounts of TNF. Together, thesefindings are compatible with at least an enhancing roleof autoimmunity in arthritis.Recent evidence suggests that targeting B-cells is aneffective therapy for RA: anti-CD20 treatment leads tosimilar efficacy as anti-TNF and to a decrease of RF-levels. This finding from clinical trials further supportsthe involvement of autoantibodies in the pathogenesis ofRA. However, neither TNF-blockade nor anti-B-celltherapy appear sufficient to induce full, persistent remis-sions. It will have to be seen if combined application ofTNF- and B-cell targeting may lead to better and sus-tained effects.

*Dedicated to Harry, a highly distinguished scientist and along-term dear friend since the days at NIH, with best wisheson the occasion of his birthday. Josef


TARGETING THE IMMUNE SYSTEM

11.Modeling the function of tumor necrosisfactor in immune pathophysiology

G. Kollias,(Biomedical Sciences Research Centre ‘AlexanderFleming’, Vari-Athens, Greece).E-mail: [email protected]

TNF is produced in response to infection or immuno-logical injury and effects multiple responses, that extendbeyond its well characterized proinflammatory proper-ties to include divert signals for cellular differentiation,proliferation and death. Part of the complexity of TNF-mediated responses may be related to the apparent dif-ferential bioactivities of its soluble and transmembraneforms(1) and the differential functioning of its two TNFreceptors. The in vivo significance of these pathwaysremains elusive and poorly defined. With reference todisease pathogenesis, especially in autoimmunity, therole of TNF is equally unpredictable. There is now clearevidence that aberrations in TNF production in vivo maybe either pathogenic or protective. For example, tem-poral and spatial deregulation of TNF production intransgenic, non-autoimmune-prone mice, promotesp55TNF-R-dependent pathologies which mimic thedevelopment of site-specific or multi-organ inflamma-tion (2), as well as specific human diseases of autoim-mune nature such as rheumatoid arthritis(3), multiplesclerosis(4–6) and inflammatory bowel disease(7).However, in contrast to the well documented enhancingeffect of chronic inflammation on autoimmune reactiv-ity, there is now good evidence to suggest that TNF-induced inflammation may in fact inhibit thedevelopment of autoimmunity, as was recently demon-strated by the prevention of autoimmune diabetes intransgenic NOD mice over-expressing TNF in theirislets, or even more recently by the exacerbated auto-immunity and clinical scores of experimental autoim-mune encephalomyelitis (EAE) developing inTNF-deficient mice(8).Deregulated TNF production, be it low or high, char-acterizes many autoimmune syndromes which areaccompanied by tissue damage processes. Consistentevidence supports a dualistic role for TNF in these con-ditions. On the one hand, TNF exerts deleterious tissue

damaging effects mainly through its pro-inflammatoryactivities, but on the other hand this same moleculeexerts beneficial activities by dampening aggressiveautoimmune responses. Therefore, blocking TNF inautoimmune-prone chronic inflammatory disease maylead to unpredictable outcomes, depending on timingand duration of treatment. Indeed, blockade of TNF inrecent clinical trials of rheumatoid arthritis or inflam-matory bowel disease, although so far impressively ben-eficial for the majority of patients, it has also led in somecases to a significant incidence of drug induced anti-dsDNA production or even in manifestations of neu-roinflammatory disease. Moreover, anti-TNF treatmentof multiple sclerosis patients has led almost exclusivelyto immune activation and disease exacerbation. Theapparent heterogeneity of receptor usage by TNF inautoimmune suppression versus inflammatory tissuedamage in murine disease models, may provide cluesfor alternative ‘anti-TNF’ therapeutic strategies. Forexample, either the p55TNFR or the p75TNFR appearssufficient to suppress autoimmune reactivity in bothorgan-specific (EAE) and systemic autoimmunity(SLE). In contrast, at least for EAE, the p55TNFRappears to be required for the damaging proinflamma-tory activities of TNF. It is therefore conceivable thatblocking the p55TNFR instead of TNF in organ-specificautoimmunity, may prove advantageous since it mayinhibit the deleterious proinflammatory and tissue dam-aging activities of TNF without compromising its immu-nosuppressive properties. A similar line of thought maybe followed to support blockade of the p75TNFR in sys-temic autoimmunity, since p75TNFR function appearsredundant in the suppression of autoimmune reactivitywhile at the same time there are indications that it maybe required for tissue damage in lupus. These hypotheses(especially the role of the p75TNFR in lupus) shouldcertainly be confirmed in more direct experiments usingappropriate model systems. Further insight into the pre-cise mechanisms mediating the autoimmune suppressiveand tissue damaging effects of TNF in chronic immu-nopathology, and the identification of the respective roleof the two TNF receptors in these phenomena, shouldlead to more effective TNFyTNF-R modulating strate-gies in combating autoimmune conditions in humans.A first step in the regulation of TNF function occurs atthe biosynthetic level. Deregulated production of TNFin transgenic models variably leads to the development


of pathologies resembling rheumatoid arthritis(RA),multiple sclerosis(MS), inflammatory bowel disease(IBD) or multi-organ inflammation(9). Targeted dele-tion of AU-rich elements(ARE) residing in the 3’UTRof TNF mRNA (TNFDARE mice) results in alteredtemporal and spatial patterns of TNF expression and tothe development of chronic joint and gut immunopath-ology (7). Absence of the ARE from the TNF mRNAleads to a loss of anti-inflammatory regulation of TNFmRNA translation, including TNF-blocking signalsinstigating from the IL-10 receptor and interfering withMAPKySAPK modules regulating TNF translation(10).The inability of MAPKySAPK-mediated signals tointerfere with TNF production in the TNFDARE mousemodel, provide an opportunity to dissect their independ-ent role in the transduction of TNF signals responsiblefor pathophysiology.Analysis of the cellular targets of TNF, which areresponsible for the development of IBD pathology in theTNF DARE model, show that TNF receptor signalsindependently restricted to either hemopoietic or non-hemopoietic cells can drive IBD with equal potency(11). This finding provides a first evidence for redun-dant cellular pathways of disease induction, foundimmediately downstream of the activity of a single cyto-kine. Future research should aim at delineating the iden-tity of the specific effector signals operating in thedifferent cellular pathways of TNF-mediated disease, asa means to achieve more selective molecular therapiesfor these diseases in humans(12, 13).

References

1. Grell, M., Douni, E., Wajant, H., Lohden, M., Maxeiner, B.,Georgopoulos, S., Lesslauer, W., Kollias, G., Pfizenmaier, K.and Scheurich P. The transmembrane form of tumor necrosisfactor is the prime activating ligand of the 80 kDa tumornecrosis factor receptor. Cell(1995) 83, 793–802.

2. Probert, L., Keffer, J., Corbella, P., Cazlaris, H., Patsavoudi,E., Stephens, S., Kaslaris, E., Kioussis, D. and Kollias, G.Wasting, ischemia, and lymphoid abnormalities in miceexpressing T cell-targeted human tumor necrosis factor trans-genes. J. Immunol.(1993) 151, 1894–1906.

3. Keffer, J., Probert, L., Georgopoulos, S., Cazlaris, H., Kaslaris,E., Kioussis D. and Kollias, G. Transgenic mice expressinghuman tumour necrosis factor: a predictive genetic model ofarthritis. EMBO J.(1991) 10, 4025–4031.

4. Probert L., Akassoglou K., Pasparakis M., Kontogeorgos G.,Kollias G. Spontaneous inflammatory demyelinating diseasein transgenic mice showing central nervous system-specificexpression of tumor necrosis factor alpha. Proc. Natl. Acad.Sci. USA (1995) 92, 11 294–11 298.

5. Akassoglou K., Probert L., Kontogeorgos G., Kollias G. Astro-cyte-specific but not neuron-specific transmembrane TNFtriggers inflammation and degeneration in the centralnervous system of transgenic mice. J. Immunol.(1997) 158,438–445.

6. Akassoglou, K., Bauer, J., Kassiotis, G., Pasparakis, M.,Lassmann, H., Kollias, G. and Probert, L. Oligodendrocyteapoptosis and primary demyelination induced by localTNFyp55TNF receptor signaling in the central nervous systemof transgenic mice: models for multiple sclerosis with primaryoligodendrogliopathy. Am. J. Pathol. (1998) 153,801–813.

7. Kontoyiannis D., Pasparakis M., Pizarro T., Cominelli F., Kol-lias G. Impaired onyoff regulation of TNF biosynthesis inmice lacking TNF AU-rich elements: implications for joint andgut associated immunopathologies. Immunity(1999) 10, 387–398.

8. Kassiotis G. and Kollias, G. Uncoupling the proinflammatoryfrom the immunosuppressive properties of tumor necrosis fac-tor (TNF) at the p55 TNF receptor level: implications for path-ogenesis and therapy of autoimmune demyelination. J. Exp.Med. (2000) 193, 427–434.

9. Kollias G. Modeling multi-organ failure, rheumatoid arthritis,multiple sclerosis and inflammatory bowel disease by engi-neering defects in TNFyTNF-R biosynthesis and function.Immunol. Rev.(1999) 169, 175–194.

10. Kontoyiannis D., Kotlyarov A., Carballo E., Alexopoulou L.,Blackshear P.J., Gaestel M., Davis R., Flavell R. and KolliasG. Interleukin-10 targets p38 MAPK to modulate ARE-dependent TNF mRNA translation and limit intestinal pathol-ogy. EMBO J.(2001) 20, 3760–3770.

11. Kontoyiannis D., Boulougouris G., Manoloukos M.,Armaka M., Apostolaki M., Pizzaro T., Kotlyarov A., ForsterI., Flavell R., Gaestel M., Tsichlis P., Cominelli F. and KolliasG. Genetic dissection of the cellular pathways and signalingmechanisms in modeled tumor necrosis factor-inducedCrohn’s-like inflammatory bowel disease. J. Exp. Med.(2002)196, 1563–74.

12. Kollias G. and Kontoyiannis D. Role of TNFyTNFR in auto-immunity: specific TNF receptor blockade may be advanta-geous to anti-TNF treatments. Cytokine Growth Factor Rev.(2002) 13:315–321.

13. Kassiotis G. and Kollias G. TNF and receptors in organ-spe-cific autoimmune disease: multi-layered functioning mirroredin animal models. J. Clin. Invest.(2001) 107: 1507–8.


12.Angiogenesis in inflammation

E. Bagli , A. Xagorari , A. Papapetropoulos , C.1 2 2

Murphy , and T. Fotsis ,1 1

( Laboratory of Biological Chemistry, Medical School,1

University of Ioannina, 45110 Ioannina, Greece,George P. Livanos Laboratory, Evangelismos Hospital,2

Department of Critical Care and Pulmonary Services,University of Athens, Athens, Greece).Email: [email protected]

Alterations in the function of endothelial cells are inte-gral part of many inflammatory processes, irrespectiveof the initial pathogenetic mechanism. Several cytokinescontribute in regulating endothelial cell responses, suchas adhesion properties and the generation of new vessels,during inflammatory processes. TNFa is a cytokine thathas been reported to regulate the expression of adhesionmolecules in the surface of endothelial cells, induceangiogenesis, and participate in the pathogenesis of sev-eral inflammation-associated diseases. Indeed, inhibitionof TNFa is considered to be a rational approach fordeveloping novel therapeutic approaches for the controlof these diseases. In this context, we have investigatedtyrosine kinase inhibitors with regard to their ability toalter endothelial cell responses induced by TNFa. Wehave particularly tested a set of flavonoids, which wehave previously shown to inhibit proliferation of endo-thelial cells.We have, initially, screened several flavonoids withregard to their ability to inhibit the expression of adhe-sion molecules, such as E-selectin, ICAM-1 and VCAM-1, in the surface endothelia cells. Indeed, one of them,luteolin, could inhibit TNFa-induced expression of allthe three adhesion molecules. Next we have investigatedthe mechanism of action of luteolin in inhibiting expres-sion of adhesion molecules in endothelial cells byTNFa. As activation of the NF-kB is instrumental forthe expression of E-selectin, ICAM-1 and VCAM-1, wehave investigated the effect of luteolin on reporter assaysencompassing various parts of the E-selectin promoter.Indeed, luteolin inhibited TNFa-induced transcriptionalactivation of the E-selectin promoter. Moreover, the NF-kB binding sequences were important for transcriptionalactivation of the E-selectin promoter by TNFa. Indeed,overexpression of p65 caused a dramatic increase in theactivation of E-selectin promoter constructs, an effectthat was inhibited by luteolin. However, luteolin did notaffect neither phosphorylation of IkBa or IkBb kinasesnor translocation of NF-kB to the nucleus.

These results suggested that luteolin was probably tar-geting the transactivation activity of NFkB. On this con-text, luteolin did not inhibit binding of NF-kB to itscognate DNA sequence as evidenced by electrophoreticmobility shift assays(EMSA). Also, luteolin did notinhibit other signaling cascades, such as that of p38MAPK, that emanate of TNFa and could influenceTNFa-induced transcriptional activation via phospho-rylation of TBP. Taken together these results suggestedthat luteolin did not affect the assembly of the NF-kBcomplex on the E-selectin promoter and that luteolincould affect the phosphorylation pattern of p65 render-ing it transcriptionally inactive. An important phospho-rylation of p65 occurs in serine 529 as consequence ofthe activity of casein kinase II(CKII). Using a specifickinase assay, we have found that luteolin is an inhibitorof casein kinase II offering a possible molecular targetfor the activity of luteolin in modulating TNFa-inducedexpression of adhesion molecules in endothelial cells.

13.Role of Tec family kinases in respiratoryinflammation

P. Sideras J. Forssell , Ch. Eriksson , K. Rydell , M.1 2 3 4

Malm-Erjefalt , Per Olof Eriksson , and J.S. Erjefalt ,5 3 4¨ ¨( Center of Immunology and Transplantations, Founda-1

tion for Biomedical Research Academy of Athens, Sor-anou tou Efesiou 4, 11527 Athens, Greece; Institute for2

Medical Biosciences, Umea University, Umea, Sweden,˚ ˚AstraZeneca R&D, Lund, Sweden, Department of3 4

Physiological Sciences, Department of Clinical Phar-5

macology, Lund University, Lund, Sweden).*These authors contributed equally to this work.Email: [email protected]

Allergic asthma and rhinitis are common inflammatoryairway diseases with increasing prevalences. The1

response to allergens is generally divided into an acute,largely mast cell-derived response, and a late phaseychronic inflammation promoted by T cells. The acute2,3

reaction is initiated by the release of stored mast cellmediators such as histamine, which induce an immediateplasma extravasation, hypersecretion, and broncho-constriction. Antigen recognition by IgE on the cell4–6

surface triggers mast cell degranulation. Aggregation ofsurface-bound FceRI-IgE complexes leads to activationof non-receptor tyrosine kinases of the Src, Syk and Tecfamily as well as lipid kinases including phosphatidyl-inositol-3-kinase(PI3K), the combined action of whichorchestrate the eventual release of mast cellmediators.7


Development of late phase inflammation(e.g. eosino-philia, goblet cell hyperplasia, and airway hyperrespon-siveness) requires differentiation of CD4 T cells intoq

Th cells and release of Th cytokines, e.g. IL-4, IL-5,2 2

and IL-13. Mast cells may also participate in the8,9

development of a late phase inflammation by the secre-tion of inflammatory mediators and cytokines such asIL-4, possibly in part by driving uncommitted Th cellstowards a Th phenotype.10–12

2

Polarization of CD4 cells and development of theq

Th phenotype is influenced by several factors, includ-2

ing cytokines, signals from the TCR and costimulatorymolecules. Early events induced by triggering thesereceptors, as in FcRI mediated activation of mast cells,include among others activation of the Src-, Syk-, andTec-family of tyrosine kinases. Recent findings haverevealed roles for Tec kinases in Th differentiation andItk was shown to play a critical role in the establishmentof a Th phenotype.13–15

2

Btk and Itk, two members of the Tec family kinases,have previously been implicated in the activation ofmast cells in vitro and the development of Th2 cells invivo, respectively . Since both these processes could16–17

potentially influence the unfolding of allergic airwayinflammatory responses in vivo, we have used an in vivomodel of allergic airway inflammation to compare acuteand late phase allergic reactions in wild type mice(C57BL6) and corresponding responses in mice carry-ing targeted Btk or Itk alleles on the same genet-yyy yyy

ic background.Btk mice showed minor protection against allergicyyy

symptoms when challenged with allergen. In sharp con-trast, both acute and late phase allergic responses wereessentially extinct in Itk mice. Therefore, Itk,yyy

beyond its known function to regulate signaling in Tcells, also function as a critical regulator in mast cells.Thus, while having normal IgE and IgG responses,1

allergen challenged Itk mice had severely impairedyyy

mast cell degranulation and acute plasma extravasation.The degranulation defect was confirmed in passivelysensitized mice, using anti-DNP IgE antibodies, and bydirect challenge with the mast cell secretagogue c48y80.Moreover, late phase associated inflammatory changes,including eosinophilia, leukocyte infiltration and Th2cytokine production in the lungs were eliminated inItk mice. Collectively, these findings suggest a crit-yyy

ical role for Itk in mast cell degranulation in vivo andidentify Itk as a ‘master regulator’ of both acute and latephase allergic reactions at least in the airways.From the findings listed above it is concluded that Itkis a very attractive target for anti-allergic therapy that

could potentially target both acute and late phase allergicreactions.

References

1. Tattersfield A.E., Knox A.J., Britton J.R. & Hall I.P. Asthma.Lancet (2002) 360, 1313–1322.

2. Busse W.W., Calhoun W.F. & Sedgwick J.D. Mechanism ofairway inflammation in asthma.Am. Rev. Respir. Dis. (1993)147, S20–4.

3. Herrick C.A. & Bottomly K. To respond or not to respond: Tcells in allergic asthma.Nat. Rev. Immunol. (2003) 3, 405–12.

4. Bochner B.S. & Lichtenstein L.M. Anaphylaxis.N. Engl. J.Med. (1991) 324, 1785–1790.

5. Metcalfe D.D., Baram D. & Mekori, Y.A. Mast cells.Physiol.Rev. (1997) 77, 1033–1079.

6. Persson C.G.A. Centennial notions of asthma as an eosino-philic, desquamative, exudative, and steroid-sensitive disease.Lancet (1997) 350, 1021–1024.

7. Kawakami T. & Galli S.J. Regulation of mast-cell and basophilfunction and survival by IgE.Nat. Rev. Immunol. (2002) 2,773–86.

8. Foster P.S. et al. Elemental signals regulating eosinophil accu-mulation in the lung.Immunol. Rev. (2001) 179, 173–181.

9. Wills-Karp M. & Chiaramonte M. Interleukin-13 in asthma.Curr. Opin. Pulm. Med. (2003) 9, 21–27.

10. Marone G., Galli S.J. & Kitamura Y. Probing the roles of mastcells and basophils in natural and acquired immunity, physi-ology and disease.Trends Immunol. (2002) 23, 425–427.

11. Bradding P. The role of the mast cell in asthma: a reassess-ment.Curr. Opin. Allergy Clin. Immunol. (2003) 3, 45–50.

12. Wills-Karp M. Immunologic basis of antigen-induced airwayhyperresponsiveness.Annu. Rev. Immunol. (1999) 17, 255–81.

13. Schaeffer E.M. et al. Requirement for Tec kinases Rlk and Itkin T cell receptor signaling and immunity.Science (1999) 284,638–41.

14. Fowell D.J. et al. Impaired NFATc translocation and failure ofTh2 development in Itk-deficient CD4q T cells. Immunity(1999) 11, 399–409.

15. Schaeffer E.M. et al. Mutation of Tec family kinases alters Thelper cell differentiation.Nat. Immunol. (2001).

16. Khan W.N. et al. Defective B cell development and functionin Btk-deficient mice.Immunity (1995) 3, 283–99.

17. Liao X.C. & Littman D.R. Altered T cell receptor signalingand disrupted T cell development in mice lacking Itk.Immu-nity. (1995) 3, 757–769.


14.Molecular analysis of signal transductionpathways

G. Panayotou,(Biomedical Sciences Research Center ‘AlexanderFleming’, Vari, Greece).Email: [email protected]

Cells respond to signals from their environment throughspecific cell surface receptors, which elicit cascades ofbiochemical events culminating in specific changes ingene expression patterns. A large number of these recep-tors, especially those for growth factors, cytokines, hor-mones, antigens and adhesion molecules, belong to thereceptor tyrosine kinase family or signal though asso-ciated tyrosine kinase subunits. Upon activation, thereceptors themselves become phosphorylated on tyrosineresidues, which then act as recruiting signals for a largenumber of intracellular enzymes or adaptors, leading toan intricate pattern of macromolecular complex forma-tion and enzymatic regulation(Schlessinger, 2000; Paw-son, 2004).The interaction of extracellular stimuli with their recep-tors is a process characterized by exquisite specificity.Indeed highly varied structures have evolved to allowreceptors to select accurately their respective ligands. Asimilar level of high specificity is also observed for theculminating events in the nucleus, which involve pro-tein–protein as well as protein–DNA interactions andresult in highly coordinated patterns of gene expression.However, the intervening cytoplasmic interactions thattransmit the signal do not appear to follow the same rule.Their inherent specificity, as measured with isolatedcomponents, is not high enough to fully explain a linearflow of high-specificity information from the cell sur-face to the nucleus. For example, SH2 domains distin-guish between different tyrosine-phosphorylation sitesbased on the context of the aminoacids surrounding thephosphotyrosine, however the measured difference inaffinity between ‘specific’ and ‘non-specific’ sites canbe as low as one order of magnitude. The same is truefor other signaling interactions, such as those involvingSH3 or Pleckstrin Homology domains. Moreover, sig-naling enzymes, for example kinases or phosphatases,often also display redundancy towards their target sites.While the relative low specificity of many intracellularprocesses confers plasticity in the ability of a cell torespond in a dynamic fashion to a changing environ-ment, cells still have to choose particular fates and elicitvery specific events. It is obvious therefore that thetranslation of diffuse, low-specificity signals to highly

specific physiological responses must involve other fac-tors. While negative selection of non-favorable interac-tions within an organism has been reported(Zarrinparet al., 2003), it is generally agreed that many more pro-cesses are in operation: Spatial segregation throughmodular interactions with diverse membrane phosphoi-nositides or with cytoskeletal or vesicular components,as well as trafficking between organelles; temporal seg-regation, where a given signal may be allowed to pro-ceed for variable times giving rise to distinct outcomes;participation in macromolecular complexes, such asthose with proteins that have regulatory or scaffoldingroles and which could be involved in presenting ormasking of substrates or in allosteric regulation; cell-type specific expression of substrates or regulators, etc.In order to study these complex processes, traditionalapproaches that focus on isolated components are notsufficient. It is becoming apparent that global approach-es to analyze the complex networks of cellular signalingare required. The modern methodologies of systemsbiology are beginning to be applied to this research area.Analyses at the genome, proteome and interactome levelare particularly important. Global changes in geneexpression and protein function are helping to revealnew targets of specific signaling pathways and the modeof their regulation. Protein arrays are revolutionizing ourability to study simultaneously in a cellular context largefamilies of distinct enzymes, such as kinases. Already,basic maps of the entire network of protein interactionsin specific organisms have been made available, forexample in yeast, Drosophila and C. elegans(Giot et al.,2003; Li et al., 2003). It is expected that these will beextended to higher organisms and, perhaps more impor-tantly, they will be refined to include changes that occurupon specific perturbations of a given cell state. It willbe important to determine the interactions of entire setsof proteins that participate in a signaling pathway andhow changes in these interactions relate to specific stim-uli and distinct physiological outcomes. A first indica-tion of this approach has been presented for the NF-kBpathway (Bouwmeester et al., 2004). The pace ofimprovements in high-throughput technologies and inbioinformatics is expected to accelerate these discover-ies and offer us a better picture of the complex machin-ery that regulates cellular responses to extracellularstimuli.

References

1. Bouwmeester T., et al. A physical and functional map of thehuman TNF-ayNF-kB signal transduction pathway. Nat. CellBiol. (2004) 6, 97–105.


2. Giot L., et al. A protein interaction map of Drosophila melan-ogaster. Science(2003) 302, 1727–36.

3. Li S., et al. A map of the interactome network of the metazoanC. elegans. Science(2003) 303, 540–543.

4. Pawson T. Specificity in signal transduction: From phosphoty-rosine-SH2 domain interactions to complex cellular systems.Cell (2004) 116, 191–203.

5. Schlessinger J. Cell signaling by receptor tyrosine kinases. Cell(2000) 103, 211–225.

6. Zarrinpar A., et al. Optimization of specificity in a cellular pro-tein interaction network by negative selection. Nature(2003)426, 676–680.


THE IMMUNE COMPONENT OF ATHEROSCLEROSIS AND STRESS

15.Heat shock proteins and stress inatherosclerosis

G. Wick, M. Knoflach, M. Kind, B. Henderson and D.Bernhard,(Institute of Pathophysiology, Medical University ofInnsbruck, Austria, Rennweg 10, A-6020 Innsbruck).Email: [email protected]

It is now accepted by the experimental and clinical com-munity that atherosclerosis has a definitive and impor-tant immunologic-inflammatory facet (1). Ourlaboratory is interested in the very early stages of thisdisease with special emphasis on immunologic effectormechanisms that may be operative at that stage. We haveshown that humoral and cellular immunity to a stressprotein, heat shock protein 60(HSP60), is the first path-ogenetic event that starts an inflammatory process in thearterial intima even before the occurrence of fatty streaksthat have for a long time been considered as the firsthallmark of atherosclerosis.Our ‘autoimmune hypothesis’ of atherogenesis postu-lates an important role for pre-existent humoral and cel-lular immunity against atherogenic epitopes of HSP60.This molecule is phylogenetically highly conserved withstrong sequence homology from bacteria to humans.Thus, HSP60 of various bacterial species show an over97% homology on the DNA and protein levels and anover 50% homology still exists between bacterial andhuman HSP60. Therefore, every normal subject pos-sesses protective immunity against bacterial and parasit-ic HSP60 as well asbona fide autoimmunity intendedto remove chemically altered autologous HSP60 andstressed cells expressing such molecules as ‘danger sig-nals’ on their surface. HSP60 plays an important role inthe folding and intracellular transport of proteins. How-ever, they also act as chaperones by associating withother cellular proteins under stress conditions protectingthese from denaturation. Interestingly, we and other lab-oratories have shown that HSP60, in principle a mito-chondrial protein, under stress not only translocates tothe cytosol but also appears on the cell surface. In addi-tion, bacterial and autologous HSP60 bind passively tothe surface of cells via Toll-like receptors(TLR-2y6,TLR-4yCD14).

Classical atherosclerosis risk factors such as high bloodpressure, smoking, diabetes, chemically altered lipopro-teins, etc., first act as stress factors for endothelial cells(ECs) entailing the expression of HSP60. We haveshown that arterial ECs are significantly more sensitiveto the HSP60-inducing effect of such risk factors, mostprobably due to the fact that they are pre-stressed by thelifelong higher arterial as compared to venous bloodpressure. Furthermore, the same stress factors, includingproinflammatory cytokines(e.g. TNF-a), oxygen radi-cals, oxidized LDL and bacterial toxins simultaneouslyinduce the expression of HSP60 and adhesion moleculesthus providing the prerequisites for the interaction ofHSP60-reactive T cells with endothelial target cells. Thisis especially true for the arterial branching areas whereECs are subjected to turbulent rather than laminar shearstress, i.e. those that are also known as classical predi-lection sites for the development of atherosclerosis. Wehave studied this phenomenon applying an experimentalarterial-venous bypass technique developed in our lab-oratory where the murine common carotid artery isreplaced by the jugular vein of a histocompatible donor.When such venous conduits are subjected to arterialblood pressure, they behave like arteries and restenose.We have shown that the first event in this restenosisprocess is the massive expression of HS60 by the ECsof the venous conduit followed by mononuclear intimacell infiltration, intimal hyperplasia and extracellularmatrix deposition finally resulting in complete oblitera-tion. Incidentally, we have also used this approach toclarify the so far unanswered question, if low frequency(i.e. household) electromagnetic induction(‘electros-tress’) represents a pro-atherogenic risk factor. Ourresults clearly exclude this possibility since 50 Hz 700Micro-Tesla electromagnetic inductions have no HSP60-inducing effect in this bypass model(2, 3), we haveidentified both linear(4) and, more importantly, confor-mational pro-atherogenic B cell epitopes that can beused for new diagnostic test systems and may also haveconsiderable therapeutic potential.Following the lead of our experimental results, we havealso embarked on several large clinical studies regardingthe possible effect of anti-HSP60 immunity for thedevelopment of atherosclerosis. Thus, since 1990 we arefollowing a cohort of originally nearly 1000 volunteersin a prospective atherosclerosis prevention study, the so-called ‘Bruneck-Study’. In this cohort, we have firstshown a significant correlation of the presence of son-ographically demonstrable atherosclerotic lesions in the


carotid artery with the presence of bacterial-humancross-reactive anti-HSP60 antibodies.This proved not only to be a new parameter for morbid-ity, but – as evident from follow up studies – also mor-tality from cardiovascular disease.The importance of bacterial-human humoral and cellularanti-HSP60 crossreactivity is also underlined by theobservation that the infectious load of an individual sig-nificantly correlates with an increased odds ratio todevelop atherosclerosis.In the year 2000, we supplemented these analyses ofhumoral anti-HSP60 immunity with investigations onHSP60 reactivity of peripheral blood T cells as well asin vivo skin test with material rich in mycobacterialHSP65(Mantoux-test). In contrast to the clearcut cor-relation of sonographically evident atherosclerosis withhumoral anti-HSP60 immunity, no such correlationemerged with respect to both, in vitro and in vivo T cellreactivity. More recently, we then repeated this type ofstudy with a smaller cohort of young(17–18 year old)male volunteers where we measured the arterial intima-media thickness(IMT) at eight different sites and ana-lysed both, humoral antibodies and peripheral T cellreactivity against human and bacterial HSP60(5). Sur-prisingly, we found that 28% of these clinical healthyyoung men already had a pathologically increased IMTat at least one of the eight measured sites. Furthermore,this feature showed a significant correlation with T cellreactivity against bacterial and-even more pronounced-human HSP60, statistically followed by a correlationwith the titer of anti-HSP60 antibodies. Thus, cellulartogether with humoral immunity against HSP60 seemsto be one of the first independent risk factors for thedevelopment of the earliest stages of atherosclerosis. Ofcourse, also in this latter group, the most important riskfactor was smoking. However, according to our concept,smoking should be a potent inducer of endothelialHSP60 expression and also bring about the release andbiochemical modification of copious amounts of auto-logous HSP60. We are now in the process of studyingthe effect of this most important atherosclerosis risk fac-tor in conjunction with our autoimmune hypothesis ofatherogenesis. In a first series of experiments, we haveshown that the constituents of cigarette smoke lead tonecrosis rather than apoptosis of human ECs in vitro(6).As expected, this process could not be blocked by broad-band caspase inhibitors.In summary, we have identified humoral and cellularimmunity to HSP60 as the most important early patho-

genetic event in atherogenesis that is independent fromother classical laboratory parameters. We also do notdeny the role of classical risk factors in the pathogenesisof the disease, but rather assign a new role to them inthe earliest stages, i.e. their effect as stress factors. Thisnew concept thus encompasses the former classical con-cepts of atherogenesis, i.e. the ‘‘response to injury’’hypothesis and the ‘altered lipoprotein’ hypothesis andopens new avenues for the prevention, diagnosis andtherapy of this killer number one.

Acknowledgements

This work was supported by the Austrian ResearchFund (P14741), the Austrian Ministry of Defense andthe Vereinigung the Elektrizitatswerke Osterreichs¨¨(VEO).¨

References

1. G. Wick, M. Knoflach and Q. Xu. Autoimmune and inflam-matory mechanisms in atherosclerosis. Annu. Rev. Immunol.22: 11.1–11.43(2004).

2. B.R. Henderson, G. Pfister, G. Bock, M. Kind and G. Wick.¨

Expression levels of heat shock protein 60 in human endothelialcells in vitro are unaffected by exposure to 50 Hz magneticfields. Cell Stress Chaperones(2003) 8: 172–182.

3. B.R. Henderson, A. Tagwerker, G. Pfister, G. Bock, H. Ulmer,¨

H. Dietrich and G.WICK. Progression of arterio-venous bypassrestenosis in mice exposed to a 50 Hz magnetic field. CellStress Chaperones(2003) 8: 373–380.

4. H. Perschinka, M. Mayr, G. Millonig, C. Mayerl, R. Van derzee, S.G. Morrison, R.P. Morrison, Q. Xu and G. Wick. Cross-reactive B-cell epitopes of microbial and human heat shock pro-tein 60y65 in atherosclerosis. Arterioscl. Thromb. Vasc. Biol.(2003) 23y6: 1060–1065.

5. M. Knoflach, S. Kiechl, M. Kind, M. Said, R. Sief, M. Gisinger,R. Van der zee, H. Gaston, E. Jarosch, J. Willeit and G. Wick.Cardiovascular risk factors and atherosclerosis in young males– ARMY Study (Atherosclerosis risk factors in male young-sters). Circulation(2003) 108: 1064–1069.

6. D. Bernhard, G. Pfister, C.W. Huck, M. Kind, W. Salvenmoser,G. Bonn and G. Wick. Disruption of vascular endothelial home-ostasis by tobacco smoke: impact on atherosclerosis. FASEB J.(2003) 17: 2302–2304.


16.The infectious etiology of theantiphospholipid syndrome (APS)

Y. Shoenfeld Y and M. Blank,1,2 1

( Department of Medicine ‘B’ and The Center for Auto-1

immune Diseases, Sheba Medical Center, Tel-Hashomer,and Sackler Faculty of Medicine, Tel-Aviv University,Israel; Incumbent of the Laura Schwarz-Kipp Chair for2

Research of Autoimmune Diseases, Tel-Aviv University,Israel).Email: [email protected]

Autoimmune diseases have a multifactorial etiology,entailing genetic hormonal, immunologic, and environ-mental factors. Among the environmental agents bacte-ria or viruses are prominent as inducers of autoimmunediseases. They can do it by a variety of mechanisms(1).For example, proteins of certain infectious agents canact as polyclonal activators on unique lymphocyte sub-sets. Viruses can preferentially infectydestroy a partic-ular T cell subset, leading to an imbalance in theimmune response. In other instances, infectious agentscan up-regulate Th1 cytokines, thereby increasingexpression of selected molecules such as MHC glyco-proteins, as well as activation of costimulatory mole-cules. Several microbial agents have been found toencode superantigens that can preferentially activatesubset(s) of T cells. Microbes can also direct the releaseof cytokines and chemokines, which may have roles ofgrowth, differentiation, or chemotactic factors for dif-ferent Th populations and also may regulate the expres-sion of MHC class I and class II molecules(1).Epstein-Barr virus(EBV)-infected autoreactive B cellsproliferate and become latently infected memory B cells.Since these B cells express virus-encoded anti-apoptoticmolecules, they are resistant to apoptosis that occursduring normal B cell homeostasisw2x. The EBV-infectedautoreactive B cells lodge in organs where their targetantigen is expressed, and act there as APCs and as asource for the monoclonally expanded B cells. Recently,it was reported that EBV infected autoreactive B cells,presenting their cognate and self-antigens to activatedautoreactive T cells, traffick through the target organ andcause local pathology(2).The classical ‘‘Hughes Syndrome’’-antiphospholipidsyndrome(APS) is characterized by the presence of cir-culating anti-phospholipid antibodies(anti-PL Abs)which bind target molecules mainlyvia ß2GPI(beta-2-glycoprotein-I), andyor lupus anticoagulant, associatedwith recurrent pregnancy loss, thromboembolic phenom-ena, and thrombocytopenia(3). The panoply of mani-

festations suggest that APS is a systemic autoimmunedisease(4), associated not exclusively with coagulationabnormalities or pregnancy failure but with manydiverse clinical manifestations such as endocarditis,CNS, skin, adrenal, and other organ involvements(4).The common denominator of all the systemic featuresin APS is the association with the presence of patho-genic polyclonal anti-PL Abs directed mainly to cardi-olipin,phosphatidylserine or phospho-ethanolamine viab2GPI. The human ß2GPI molecule is a heavily gly-cosylated membrane-adhesion glycoprotein(326aa),presented in healthy and APS patients’ plasma.b2GPIexhibits properties of anticoagulant and has a role inclearance of apoptotic cells from the circulation in vitro.Antibodies tob2GPI itself have a pathogenic activityentailing endothelial and platelet cell activations in-vitroand induction of experimental APS in-vivo(5).The healthy immune system is tolerant to the moleculesof which the body is composed of. However, one canfind that among the major antigens recognized during awide variety of bacterial, viral and parasitic diseases,many belong to conserved protein families, sharingextensive sequence identity or conformational fits, withhost’s molecules, namely molecular mimicry. Antigenicsimilarity of either molecules’ linear amino-acidsequences or their conformational structure betweenantigens of infectious agents and host tissues might trig-ger an immune response against the shared determinant.As a result, the tolerance to autoantigens breaks down,and the pathogen-specific immune response that is gen-erated, cross-react with host structures to cause tissuedamage and disease. A role for molecular mimicry inthe pathogenesis of autoimmune diseases has recentlybeen shown in APS(6–8).Accumulating evidence point to an association betweenthe presence of circulating aPL and mainly anti-b2GPI,with infectious agents including bacteria, viruses andparasites, which in several cases involves overt APSmanifestations, while in others lead to catastrophic APS(9). Experimental APS models proved the molecularmimicry betweenb2GPI related synthetic peptides andstructures within bacteria, viruses(e.g. CMV), and tet-anus toxoid, and confirmed their caustative role inexperimental APS(6–8). Previously, we have identifiedand characterized severalb2GPI related synthetic pep-tides. These peptides were fished from a hexapeptidephage display library employing human anti-b2GPImonoclonal Abs derived from APS patients. The pep-tides had neutralizing biological activity in-vitro and in-vivo (5).In the protein data base we found homologies betweenour peptides and common bacteria, viruses, yeast and


tetanus-toxin(Fig. 1). In order to prove the involvementof molecular mimicry mechanism between the patho-gens andb2GPI molecule in APS pathogenesis, weimmunized naıve mice with microbial pathogens, which¨shared structural homology with one of the hexapeptide(TLRVYK ). IgG anti-TLRVYK were found in theimmunized mice and were affinity purified on aTLRVYK-column and than passively infused i.v. intoanother set of naive mice at day 0 of pregnancy. APS-clinical parameters were evaluated in the infused miceon day 15 of pregnancy. Following immunization, var-th

ious levels of mouse anti-b2GPI Abs were observed, thehighest being detected in those mice immunized withhaemophilus influenzae, neisseria gonorrhoeae or teta-nus toxoid. Mice infused with affinity purified anti-b2GPIypeptide relate Abs had developed thrombo-cytopenia, prolonged aPTT and raised percentage offetal loss, findings characteristic OF experimental APS(6). Our results established a pathogenic molecularmimicry mechanism in experimental APS, which in con-cert with the presence of circulating anti-peptide Abs inAPS patients’ sera(10), propose an infectious origin foranti-b2GPI or aPL. An infection with one of theb2GPIcross- reactive infecting agents in a susceptible person(i.e. specific HLA) may lead to an overt APS.Interestingly, also the anti-b2GPI Abs located on thevalves of patients with APS having endocarditis, bindalso the TLRVYK. These results point to the possibleinfectious origin of the ‘non-infectious Libman-Sacksendocarditis’. Since APS resemble rheumatic fever(RF)in its presentations(valve deformations and chorea) itis not surprising that cross-reactions were found by usbetween the streptococcal M-protein peptides and ourb2GPI (e.g. bacterial) peptides.Furthermore, anti-Saccharomyces cervisiae antibodis(ASCA) were detected in 33% patients with APS incomparison to 7% of ‘healthy’ individuals. Anti-b2GPI,affinity purified from ASCA positive sera, bound spe-cifically the phosphopeptidomannan(PPM) compoundof the yeast. The crossreactivity was also demonstratedby competition assays(95–98%). The PPM inhibiteddifferentially the anti-b2GPI binding tob2GPI.All the above studies point to an infectious etiology forAPS. The interaction of the specific peptides with theprone MHC molecules has to be delineated.

References

1. Shoenfeld Y., Rose N.(Eds) Infections and Autoimmunity;Elsevier Publication, Amsterdam, The Netherlands, 2004; pp1–650.

2. Pende M.P. Infection of autoreactive B lymphocytes with EBV,causing chronic autoimmune diseases.Trends. Immunol. 2003;24,584–588.

3. Hughes G.R.V., Harris E.N., Gharavi. A.E. The anti-cardioli-pin syndrome. J. Rheumatol. 1986; 13:486–9.

4. Shoenfeld Y. Systemic antiphospholipid syndrome. Lupus2003;497–498.

5. Blank M., Shoenfeld Y., Cabilli S., Heldman Y., Fridkin M.,Katchalski-Katzir E. Prevention of experimental antiphospho-lipid syndrome and endothelial cell activation by syntheticpeptides, Proc. Natl. Acd. Sci. 96,1999; 5164–5168.

6. Blank M., Krause I., Fridkin M., Keller N., Kopolovic J.,Goldberg I., Tobar A., Shoenfeld Y. Bacterial induction ofautoantibodies to beta2-glyco-protein-I accounts for the infec-tious etiology of antiphospholipid syndrome. J. Clin. Invest.2002; 109:797–804.

7. Shoenfeld Y. Etiology and pathogenetic mechanisms of theanti-phospholipid syndrome unraveled. Trends Immunol.2003; 24:2–4.

8. Gharavi A.E., Pierangeli S.S., Espinola R.G., Liu X., Colden-Stanfield M., Harris. E.N. Antiphospholipid antibodiesinduced in mice by immunization with a cytomegalovirus-derived peptide cause thrombosis and activation of endothelialcells in vivo. Arthritis Rheum. 2002; 46:545–552.

9. Zandman-Goddard G., Blank M., Shoenfeld Y. Antiphospho-lipid antibodies and Infections. In: The Antiphospholipid Syn-drome II. Autoimmune Thrombosis. Asherson R.A. CerveraR., Shoenfeld Y., Piette J.-C.(Eds) Elsevier Publication,Amsterdam, The Netherlands, 2002; pp 343–358.

10. Shoenfeld Y., Krause I., Kvapil F., et al. The correlation ofantibodies againstb2-glycoprotein-I and variousb2-glycopro-tein-I peptides with clinical manifestations in the antiphospho-lipid syndrome. J. Clin. Immunol. 2003; 23:377–83.


Figure 1: Bacteria or viruses having cross-reactingpeptides to b2GPI infect susceptible subjects andinduce anti- bacterialyviruses antibodies that reactwith similar peptides on the b2GPI molecules andlead to overt manifestations for APS.

Take-home messages:

● Bacteria and viruses(i.e. EBV) have crossreacting peptides to B2GPI.

● APS is induced in a susceptible(i.e MHCgenes) subject exposed to these bacteria orviruses.

● Different manifestation of APS may be due todifferent peptides inducing anti- Pl antibodies.

● Libman-sacks endocarditis(in APS) may bedues also to an infectious cause.

● Saccharomyces cervisiae may be one of theimportant inducing infectious agents of APS

● Rheumatic fever and APS share common clini-cal manifestations and most probably also com-mon infecting agents-cross reaction between Mprotein peptides of streptococcus and b2GPIpeptides of APS and other infecting agents.

17.Systemic inflammation and well-being

G.P. Chrousos,(Athens University Medical School, Athens, Greece).Email: [email protected]

Like the stress response, the inflammatory reaction ofan individual is crucial for survival of the self and spe-cies. Also like the stress response, inflammation is meantto be tailored to the stimulus and time-limited. A fullyfledged systemic inflammatory reaction consists of acti-vation of immune and immune accessory cells and resul-tant stimulation of four major programs:(1) the acutephase reaction,(2) the sickness syndrome,(3) the painprogram, mediated by the afferent sensory and autonom-ic systems, and(4) the stress program, mediated by thestress system, i.e. the hypothalamic-pituitary-adrenal(HPA) axis and the locus ceruleus- norepinephrineysym-pathetic system. The main effector substances of the sys-temic inflammatory response are the inflammatorycytokines, such as TNFalpha, IL-1 and IL-6, chemoki-nes, such as IL-8, and other mediators of inflammation;the acute phase reactants, mostly of hepatic origin, such

as C-reactive protein(CRP), cell adhesion molecules,fibrinogen and plasminogen activator inhibitor 1; theeffectors of the sensory afferent system, such as sub-stance P; and, of the stress system, namely hypothalamicCRH and vasopressin, cortisol, the catecholamines nor-epinephrine and epinephrine, and peripheral neuronalCRH.Be it an inflammatory focus with spillover of inflam-matory effector molecules into the systemic circulationor a truly generalized, systemic inflammatory reaction,the programs that are activated during inflammationhave both synergistic and antagonistic actions. Forinstance, the inflammatory cytokines stimulate CRP pro-duction by the liver and this effect is potentiated byglucocorticoids, which however also inhibit the secretionof inflammatory cytokines, bringing inflammation to aclose. The sickness syndrome consists of anorexiaynau-sea, fatigue andyor depressed affect, somnolence, hyper-algesia, sleep disturbances, elevated temperature and anincreased metabolic rate, all manifestations suppressedby glucocorticoids. Yet, peripheral neuronal CRH acti-vated by stress or the inflammatory reaction, and sub-stance P activated by the inflammatory reactionpotentiate inflammation. In fact, through the formermechanism stress may trigger andyor exacerbate aninflammatory condition such as asthma or rheumatoidarthritis.Chronic systemic inflammation, depending upon itsdegree, varies from asymptomatic to mildly, to severelysymptomatic. Regardless of the presence of overt symp-tomatology of sickness syndrome manifestations, chron-ic elevations of circulating inflammatory cytokines andyor activation of the stress system result in a combinationof immune and metabolic disturbances, including endo-thelial inflammation andyor a Thelper1 to Thelper 2switch, osteoporosis, hypercoagulability of the blood,dyslipidemia, insulin resistance, carbohydrate intoler-ance andyor diabetes type 2, and hypertension. The non-immune manifestations constitute the visceral fatsyndrome which deteriorates with time in patients withchronic inflammation andyor stress; this represents anexacerbation of a phenomenon that occurs with advanc-ing age in both men and women. These immune andmetabolic changes increase all cause mortality, primarilycardiovascular due to atherosclerosis, but also cancer-and infection- related; they also cause significant mor-bidity, potentially including clinically significant osteo-porosis.Chronic or intermittent but frequent inflammation dueto presence of inflammatory foci, such as those in aller-gic rhinitis, bronchial asthma, periodontitis,Helicobac-ter pylori infection or multiple sclerosis, may be


responsible for varying degrees and patterns of sicknesssyndrome manifestations and may be associated with thechronic immune, metabolic and cardiovascular compli-cations of inflammation mentioned above.Interestingly, adipose tissue secretes large amounts ofTNFalpha and IL-6 in a neurologically, hormonally andmetabolically regulated fashion. The plasma levels ofthese cytokines are proportional to the body mass index(BMI) and are further elevated in patients with visceralobesity. The secretion of inflammatory cytokines has acircadian pattern, with elevations in the evening and inthe early morning hours. This pattern is maintained inpatients with inflammatory diseases and in obese sub-jects, albeit at a higher level, is affected by the qualityof sleep, and correlates with manifestations of the sick-ness syndrome. In obesity, the hypercytokinemia is asso-ciated frequently with some manifestations of thesickness syndrome, such as fatigue and somnolence, andof the other programs that may be activated during theinflammatory reaction. Thus, obesity and, especially thevisceral type, can be considered as a chronic inflam-matory state, with many of the behavioral, immune, met-abolic and cardiovascular sequelae of such a state.

References

Chrousos G.P. The Hypothalamic-Pituitary-Adrenal Axis andImmune-Mediated Inflammation. NE. J. Med.(1995)332:1351–1362.Chrousos G.P. 1997 Hans Selye Memorial Lecture: Stressors,Stress and Neuroendocrine Integration of the Adaptive Response.Ann. NY Acad. Sci.(1998)851:311–335Chrousos G.P., Torpy D., Gold P.W. Interactions Between theHypothalamic-Pituitary- Adrenal Axis and the Female Reproduc-tive System: Clinical Implications. Ann. Intern. Med.(1998)129:229–240Clauw D.J., Chrousos G.P. Chronic Pain and Fatigue Syndromes:Overlapping Clinical and Neuroendocrine Features and PotentialPathogenic Mechanisms. NeuroImmunoModulation(1997)4:134–153.Crofford L.J., Kalogeras K.T., Mastorakos G., Magiakou M-A.,Wells J., Kanik K.S., Gold P.W., Chrousos G.P., Wilder R.L. Cir-cadian Relationships Between Interleukin(IL)-6 and Hypothalam-

ic-Pituitary-Adrenal Axis Hormones: Failure of IL-6 to CauseSustained Hypercortisolism in Patients with Early Untreated Rheu-matoid Arthritis. J. Clin. Endocrinol. Metab.(1997)82:1279–1283.Elenkov I.J., Chrousos G.P. Stress Hormones, Th1yTh2-patterns,ProyAnti- Inflammatory Cytokines and Susceptibility to Disease.Trends Endo Metab.(1999)10: 359–368Franchimont D., Kino T., Galon J, Meduri G.U., Chrousos G.P.Glucocorticoids and Inflammation Revisited. NIH Clinical StaffConference. NeuroImmunoModulation(2003)10:247–260Orban Z., Remaley A.T., Sampson M., Trajanoski Z., ChrousosG.P. The Differential Effect of Food Intake and -Adrenergic Stim-ulation of Adipose-Derived Hormones and Cytokines in Man. J.Clin. Endocrinol. Metab.(1999)84:2126–2133Papanicolaou D.A., Tsigos C., Oldfield E.H., Chrousos G.P. AcuteGlucocorticoid Deficiency is Associated with Plasma Elevation ofInterleukin-6: Does the Latter Participate in the Symptomatologyof the Steroid Withdrawal Syndrome and Adrenal Insufficiency?J. Clin. Endocrinol. Metab.(1996)81:2303–2306.Papanicolaou D.A., Wilder R.L., Manolagas S.C., Chrousos G.P.The Pathophysiologic Roles of Interleukin-6 in Humans. Ann.Intern. Med.(1998)128:127–137.Pillemer S.R., Bradley L.A., Crofford L.J., Moldofsky H., Chrou-sos G.P. The Neuroscience and Endocrinology of Fibromyalgia.Arthritis & Rheumat.(1997)40:1928–1939.Reincke M. Allolio B., Arlt W., Heppner C., Petcke F., MbulamberiD., Siekman L., Vollmer D., Winkleman W., Chrousos G.P. Impair-ment of Adrenocortical Function Associated with Increased PlasmaTumor Necrosis Factor-alpha and Interleukin-6 Concentrations inAfrican Trypanosomiasis. NeuroImmunoModulation(1994)1:14–22.Vgontzas A.N., Papanicolaou D.A., Bixler E.O., Kales A., TysonK., Chrousos G.P. Elevation of Plasma Cytokines in Disorders ofExcessive Daytime Sleepiness: Role of Sleep Disturbance andObesity. J. Clin. Endocrinol. Metab.(1997)82:1313–1316.Vgontzas A.N., Papanicolaou D.A., Bixler E.O., Lotsikas A.,Zachman K., Kales A., Prolo P., Wong M., Licinio J., Gold P.W.,Hermida R.C., Mastorakos G., Chrousos G.P. Circadian Interleu-kin-6 Secretion and Quality and Depth of Sleep, J. Clin. Endocri-nol. Metab.,(1999)84:2603–2607Vgontzas A.N., Papanicolaou D.A., Bixler E.O., Hopper K., Lot-sikas A., Kales A., Chrousos G.P. Resistance, and Hypercytoki-nemia, J. Clin. Endocrinol. Metab.,(2000)85:1151–1158


INFLAMMATORY MYOPATHIES

18.The place of autoimmunity in myositis

P. Plotz,(Arthritis and Rheumatism Branch, National Institute ofArthritis and Musculoskeletal and Skin Diseases,National Institutes of Health, Clinical Center 9N244,9000 Rockville Pike, Bethesda, MD 20892-1820).Email: [email protected]

Autoimmunity, it sometimes seems to me, is often usedas the apparently respectable and unassailable explana-tion for what we do not understand. The discovery auto-antibodies in disease more than half a century agoappeared to break wide open the understanding of manypreviously mysterious diseases, especially inflammatorydiseases. It is hard now to recall that the current classi-fication of many diseases grew up hand in hand withthe progressive refinement of techniques for definingautoantibody specificity. Investigators sought and foundautoantibodies with convenient pathological propertiesfor the occasion in all sorts of diseases. The great major-ity of these have disappeared, but what didn’t disappearwas the conviction that any disease with chronic inflam-mation on biopsy is considered likely to be autoimmunein origin. Supporting this bias has been the associationof particular HLA genotypes with some of thesediseases.In fact, the majority of autoantibodies associated withinflammatory diseases-being directed at ubiquitous intra-cellular antigens-cannot yet persuasively be tied to dis-eases pathogenesis. The instinct has been to conclude,therefore, that specific cellular immunity-target tissuespecific T cells of various kinds-must be responsible forthe pathology and the clinical disease. Much excellentscience coupled to excellent clinical observation seemsto support this bias, but it is reasonable to observe thatthe term autoimmune disease most often just means adisease of chronic inflammation that often waxes andwanes in the absence of a recognizable cause, accom-panied-in at least some cases-by the presence of auto-antibodies. In extremely few cases have T cells ofrelevant specificity been identified, and in almost no cas-es can they be shown to be the cause of the disease.

Furthermore, it is now evident that many genetic manip-ulations of the murine or even human immune systemlead to the same relatively limited repertoire of autoan-tibody specificities, and it would not surprise one tolearn one day that the same will turn out to be true forT lymphocyte specificities in the lymphocytic infiltra-tions that accompany these manipulations.This matters today, because disease classification that isbased upon the uni-dimensional consolidation of patientsinto clinically- or immunologically- or pathologically-or even genetically-defined groups is increasingly unsat-isfactory. In each person, all of these elements are at playin unique and very complex ways that defy full under-standing.Where we stand today with muscle disease is emblem-atic. The elements of genetic disease, of inflammation,including lymphocytic attack without identifiable cause,of muscle cell degeneration and apoptosis, of myoblastand probably other stem cell repair, and of humoral andcellular autoimmunity can be found in patients withmany diseases in our current classification schemes.They are not all always there. But consider these exam-ples or paradoxes: inflammation can be a transient oreven a long-lasting accompaniment of several musclediseases unquestionably initiated by any of several dys-trophic genetic mutations; severe muscle cell damage insome genetic metabolic myopathies is almost neveraccompanied by inflammation; autoantibodies may befound in idiopathic myopathy in which lymphocyticinflammation is sparse or absent; in one group of chronicmyopathies-inclusion body myopathyymyositis-thedefining pathological change can be found in an unques-tionably genetic disorder or in a sporadic phenocopywith ‘‘autoimmune’’ genetic markers and lacking otherrecognized genetic abnormalities. Knowing that some orall of these elements are likely to be part of all diseaseprocesses in the muscle(or elsewhere) must give onepause. This is not the moment for rigidity. This a timeto be flexible and ecumenical in our approach. Andmuch difficult research lies ahead.The career of Professor Haralampos Moutsopoulos,beautifully illustrates the value of an open mind, andthis wonderful celebration illustrates the accomplish-ment of his open heart.


19.Update on the molecular pathogenesis ofinflammatory myopathies

M.C. Dalakas,(Neuromuscular Diseases Section, National Institute ofNeurological Disorders and Stroke, National Institutes ofHealth, Bethesda, MD. NINDS, NIH, Building 10,Room 4N248, 10 Center Dr. MSC 1382, Bethesda, MD20892-1382.).Email: [email protected]

Classification and diagnosis

The inflammatory myopathies comprise three major anddistinct subsets: dermatomyositis(DM), polymyositis(PM), and sporadic inclusion-body myositis(s-IBM). PM as a stand-alone entity is rare in adults1–5( )

and very rare in children; DM occurs in all ages; IBMis the commonest inflammatory myopathy above the ageof 50. Dermatomyositis is a distinct disease that affectsmuscle and skin. Polymyositis(PM) remains a diagnosisof exclusion; it is best defined as an acquired myopathyof subacute onset that occurrs in patients who do nothave family history, exposure to myotoxic drugs or tox-ins, another disease caused by endocrine, metabolic orneurogenic causes, IBM or a dystrophy. IBM developsslowly (over months or years) and affects distal andproximal muscle groups in a characterisitic distribu-tion. 1–5( )

The clinical diagnosis of PM, DM and IBM is confirmedby the presence of:(a) elevated serum muscle enzymes(from 5 to 50-fold); (b) myopathic motor unit actionpotentials on electromyography; and c) typical changeson the muscle biopsy. In PM and IBM the inflam-6( )

mation is primary, a term coined to indicate that endo-mysial T cell infiltrates surround individual, healthymuscle fibers that eventually invade; most importantly,the MHC-1 antigen is unbiquitously expressed on thesarcolemma even in fibers not invaded by CD8qcells. The presence of the ‘CD8yMHC-1’ lesion is1–3,7( )

specific for PM and IBM and essential to excludemyopathies with secondary, non-specific inflammation,hence the need to be in the diagnostic criteria. In2,3,8( )

DM, the endomysial inflammation is predominantly per-ivascular or in the interfascicular septae and around,rather than within, the fascicles. The muscles fibers8( )

undergo degeneration and phagocytosis, often in groupsinvolving a portion of a muscle in a wedgelike shape orat the periphery of the fascicle, resulting in perifascicularatrophy. In IBM, in addition to the intense inflammationand the ‘CD8yMHC-1 lesion’, there are rimmed vacu-

oles, abnormal mitochondria and congophilic amyloiddeposits that immunoreact for various amyloid-relatedproteins such asbAPP, tau, ubiquitin, chymothrypsin orprion.

Immunopathogenesis

The autoimmune origin of these disorders is supportedby their association with other systemic autoimmune,viral, or connective tissue diseases; the presence of var-ious autoantibodies against ribonucleoproteins involvedin protein synthesis(anti-synthetases) or translationaltransport(anti-signal-recognition particles); their asso-ciation with histocompatibility genes; the evidence of Tcell-mediated myocytotoxicity or complement-mediatedmicroangiopathy; and their response to immunothera-pies. However, the specific muscle or capillary targetantigens have not been identified, and the agents initi-ating self-sensitization are still unknown.In DM, the endomysial infiltrates have a higher thannormal percentage of B cells, CD4q)CD8q cellsespecially in the proximity to B cells and macrophages,and a relative absence of lymphocytic invasion of non-necrotic muscle fibers. The immune process is mediatedby the complement C5b-9 membranolytic attack com-plex directed against microvascular antigens, resulting innecrosis of the endothelial cells, reduced number ofendomysial capillaries, ischemia, muscle-fiber destruc-tion often resembling microinfarcts, and inflammation.Complement activation occurs early in the disease, andtriggers the release of proinflammatory cytokines andchemokines which upregulate VCAM-1 and ICAM-1 onthe endothelial cells and facilitate the transmigration ofactivated lymphoid cells to the perimysial and endo-mysial spaces.In PM and IBM there is evidence of an antigen-directedcytotoxicity mediated by CD8q cytotoxic T cells. Theendomysial CD8q cells, along with macrophages, sur-round, invade and eventually destroy healthy, nonne-crotic, muscle fibers that aberrantly express class I MHCmolecules (‘CD8yMHC-I complex’). Several of theautoinvasive CD8q T cells of specific T cell Receptor(TCR) families are clonally expanded and their CDR3region, the antigen-binding region of the TCR, has con-served aminoacid sequences, suggesting of an antigen-driven T cell response. In contrast, the non-9–13( )

invasive, bystander T cells are clonally diverse. Clonalrestriction of the same Vb families, most frequentlyVb3, Vb5.1, Vb6.7 Vb13, persits even when sequentialmuscle biopsies performed over a 2 year period. In10( )

PM and IBM, in spite of the Fas antigen expression onthe muscle fibers and the Fas-L on the autoinvasive


CD8q cells, the muscle fibers do not undergo apoptosis,but necrosis in a Fas-independent pathway mediated byperforin released from the autoinvasive sensitized Tcells. The B7-family of costimulatory molecules, BB-1and ICOS(Inducible Costimulator), and their ligandsCD28yCTLA-4 and LICOS are upregulated on the mus-cle fibers and the autoinvasive CD8q cells. Cytokines(IL , TNF-alpha and interferon-gamma), chemo-1,2,4,5,6

kines MCP-1(CCL2), Mig (CXCL9) and their recep-tors, matrix metalloproteinsases(MMP-2 and MMP-9),adhesion molecules on leukocytes(L-selectin and inte-grins LFA-1, VLA-4) and their respective ligands onendothelial cells(GlyCAM-1, ICAM-1, VCAM-1) areupregulated and facilitate adhesion and transmigrationof activated T cells through the endothelial cell wall.In IBM, in addition to the autoimmune mechanismsdescribed above, there is also a concomitant degenera-tive process exemplified by the deposition within thevacuolated fibers of beta amyloid along with Alzhei-mer’s disease-like amyloid related proteins, such asamyloid precursor protein(APP), chymotrypsin, apoli-poprotein E and phosphorylated tau. The vacuolesincrease with disease chronicity and they are almostalways present in muscle fibers not invaded by T cells;in contrast, the fibers invaded by the cytotoxic CD8qT cells are never vacuolated. These degenerative fea-tures are probably secondary, caused by the effects ofaging, disease chronicity and upregulated cytokinesreleased by the abundant macrophages and activatedendomysial T cells. In IBM, like in Alzheimer disease,there is an intriguing link between b-APP and inflam-matory cytokines, especially IL-1, TNF-a and MMP-2,which are upregulated on the amyloid deposits. The14( )

excess of IL-1b may be derived not only from the abun-dant endomysial macrophages and T cells but also bythe b-APP, which is a known enhancer of IL-1b pro-duction; in turn, IL-1b upregulatesb-APP andb-APPgene expression and closes the proposed loop:IL1b°b-APP°IL1b°inflammation.14( )

Treatment

Because the specific target antigens in DM, PM andIBM are unknown, the immunosuppressive therapies arenot selectively targeting the autoreactive T cells or thecomplement-mediated process on the intramuscularblood vessels. Instead, they are inducing a non-selectiveimmunosuppression or immunomodulation. Agents15,16( )

used in the treatment of PM and DM include: Cortico-steroids, Azathioprine, Methotrexate, Cyclophosphami-de, Chlorambucil, Cyclosporine, and IntravenousImmunoglobulin(IVIg). The preferred drugs in the hier-

archy of treatment, used alone or in combination, dependon patient’s age, degree of disability, tolerance, experi-ence with the drug, patient’s general health and cost.Progress in molecular immunology promises a more tar-get-oriented immunotherapy. Ongoing trials are current-ly utilizing monoclonal antibodies against:(a) signaltransduction in T lymphocytes targeting CD52, costi-mulary molecules or IL2R, resulting in inhibition of Tcell activation or T cell depletion;(b) immunomodulat-ing cytokines;(c) complement C and(d) Adhesion5

molecules and receptors.15,16( )

References

1. Dalakas M.C. Polymyositis, Dermatomyositis and Inclusion-Body Myositis. N. Engl. J. Med. 1991;325:1487–1498.

2. Dalakas M.C., Hohlfeld R. Polymyositis and Dermatomyositis.Lancet 2003; 362:971–982.

3. Dalakas M.C. Polymyositis, Dermatomyositis and InclusionBody Myositis. In: Harrison’s Principles of Internal Medicine(16th edition). Braunwald E., Fauci A.S., Kasper D.L., HauserS.L., Longo D.L., Jameson J.L.(eds). McGraw-Hill, NewYork, NY. In press.

4. Mastaglia F.L., Phillips B.A. Idiopathic inflammatory myopa-thies: epidemiology, classification and diagnostic criteria.Rheum Dis Clin N Am 2002; 28:23–41.

5. Griggs R.C., Askanas V., Di Mauro S., et al. Inclusion bodymyositis and myopathies. Ann. Neurol. 1995:38:705–713.

6. Dalakas M.C. Muscle biopsy findings in inflammatory myopa-thies. Rheum. Dis. Clin. N. Am. 2002; 28:779–98.

7. Hohlfeld R., Engel A.G. The immunobiology of muscle.Immunol. Today 1994;15:269–274.

8. Dalakas M.C., Hohlfeld R. Diagnostic criteria for Polymyositisand Dermatomyositis. Lancet 2003;362:1762–1763.

9. Bender A., Ernst N., Iglesias A., Dornmair K., Wekeria H.,Hohlfeid R. T Cell Receptor repertoire in polymyositis: clonalexpansion of autoaggresive CD8 T Cells. J. Exp. Med.1995;181:1863–1868.

10. Amemiya K., Granger R.P., Dalakas M.C. Clonal restrictionof T-cell receptor expression by infiltrating lymphocytes inInclusion Body Myositis persists over time: studies in repeatedmuscle biopsies. Brain 2000;123:2030–2039.

11. Murata K., Dalakas M.C. Expression of the costimulatory mol-ecule BB-1, the ligands CTLA-4 and CD28 and their mRNAin inflammatory myopathies. Am. J. Pathol. 1999;155:453–460.

12. Benveniste O., Cherin P., Maisonobe T., et al. Severe pertur-bations of the blood T cell repertoire in Polymyositis, but notdermatomyositis patients. J. Immunol. 2001;167:3521–3529.

13. Hofbauer M., Wiesner S., Babbe H., et al. Clonal tracking ofautoagressive T cells in polymyositis by combining laser


microdissection, single-cell PCR and CDR3 spectratype anal-ysis. Proc. Natl. Acad. Sci. USA 2003;100:4090–4095.

14. Dalakas M.C. Molecular Immunology and Genetics of Inflam-matory Muscle Diseases. Arch. Neurol. 1998;55:1509–1512.

15. Hohlfeld R., Dalakas M.C. Basic Principles of Immunotherapyin Neurological Diseases. Sem. Neurol. 2003;23:121–132.

16. Dalakas M.C. Therapeutic Approaches in Patients with Inflam-matory Myopathies. Sem. Neurol. 2003;23:199–206.


SYSTEMIC LUPUS ERYTHEMATOSUS

20.Pathogenic aspects of antiphospholipidantibodies

S.A. Krillis,(Immunology, Allergy & Infectious Diseases St. GeorgeHospital, 2 South Street, Kogarah NSW 2217,Australia).Email: [email protected]

Introduction

The term antiphospholipid syndrome(APS) refers topatients with a combination of autoantibodies and a his-tory of thrombosis, recurrent miscarriage and other clin-ical events consistent with the syndrome(1). Most ofthe clinical events in APS are thrombotic in nature. Theautoantibodies in this condition are thought to play acritical role in the pathogenesis of thrombosis and recur-rent miscarriage. The major evidence comes from pas-sive transfer of the antiphospholipid antibodies(aPA) torodents, which produce features of APS and increase inclot formation in a model of venous injury. Contrary tothe name of the syndrome, phospholipid itself is not thetarget for clinically significant antibodies but antibodiesfrom patients with APS bindb GPI, a major plasma2

protein that binds to anionic phospholipids(2).Although there have been a number of potential mech-anisms of thrombosis in the APS, no unifying hypothesishas been proposed. This brief review will outline somecurrent exciting findings in this area.

b -Glycoprotein I2

b -Glycoprotein is a glycoprotein of 50 kDa that cir-2

culates in plasma at approximately 4mM. It has 326amino acids that consist of repeated sequences in a formtypical of the complement control protein(CCP) mod-ule. Individual modules are also known as short consen-sus repeats(SCR), a key feature of which is disulphidebridges joining the 1st to 3rd and 2nd to 4th cysteineresidues.b GPI has affinity for negative charged mac-2

romolecules such as anionic phospholipids and proteo-glycans. The 5th domain is critical for phospholipid andheparin binding and is highly conserved.b GPI is the2

primary target antigen recognised by autoantibodies inpatients with the antiphospholipid syndrome. Binding ofautoantibodies tob GPI is now generally accepted as an2

important feature of APS and a number of studies haveshown there is a significant correlation between throm-botic manifestations and the presence of anti-b GPI anti-2

bodies. Although the physiological function ofb GPI in2

normal individuals remains to be elucidated, plasmafrom b GPI knockout mice exhibits impaired thrombin2

generation in vitro(3). b GPI has multiple pro-coagu-2

lant and anti-coagulant effects in vitro but the in vivofunction of this molecule has not been elucidated. ManyaPA are intrinsically of low affinity and bivalent bindingis an absolute requirement to bindb GPI (4). It has2

recently been demonstrated that the major immunodom-inant epitope onb GPI is localised on domain I(5).2

Anti-b GPI antibodies from patients with APS can pos-2

sess lupus anticoagulant activity, which is detected bytheir ability to prolong in vitro phospholipid dependentcoagulation assays.b GPI forms bivalent complexes2

when it interacts with anti-b GPI antibodies, which2

leads to an increase in affinity ofb GPI for negatively2

charged phospholipids. Chimeric dimers ofb GPI have2

been demonstrated to induce increased binding to phos-pholipids and lupus anticoagulant activity in the absenceof antiphospholipid antibodies. Furthermore dimericb2GPI increased platelet adhesion to collagen in an invitro system, which mimicked the effect on plateletswhen whole blood was spiked with patient derived poly-clonal anti-b GPI antibodies(6). It is likely that auto-2

antibodies tob GPI form complexes withb GPI able to2 2

bind to activated platelet surface receptors. After bindingdimeric b GPI reacts with platelet receptor and results2

in sensitisation of platelets to low concentrations ofthrombin which are able to respond to collagen whichthey adhere to. Dimericb GPI has been demonstrated2

to interact with ApoER2 on platelets stimulated withthrombin in an in vitro flow system perfused over col-lagen or fibronectin(6). In this system dimericb GPI2

constructs and anti-b GPI antibodies induced platelet2

adhesion, thrombus formation and thromboxane A2 pro-duction(6).

Apolipoprotein E receptor 2 (ApoER2)

ApoER2 is a member of the LDL receptor family andmembers of this family have specific protein modulesof extracellular ligand binding repeats, EGF precursorhomolog repeats and the cytoplasmic region containingmotifs for endocytosis and signal transduction(7).ApoER2 is involved in receptor mediated signal trans-duction and the gene encodes for a number of differ-


entially spliced transcripts. These involve variations inthe ligand binding in intracellular domain of the receptorand are species and tissue specific. Platelets, megakar-yocytic cell lines and endothelial cells express the majortranscript lacking ligand binding repeats 4–6 due to theabsence of exon 5 but contain the full cytoplasmic tail.Ligand interactions with all members of this family canbe antagonised by the receptor associated protein(RAP)a unique ligand frequently used as a tool in the study ofligand receptor interactions within this family. Theincreased platelet adhesion observed with dimericb GPI2

(6) disappeared after preincubation with RAP. Mono-meric b GPI did not induce platelet adhesion or throm-2

boxane A2 production in this system due to the fact itcould not bind efficiently or crosslink ApoER2. We haverecently identified a peptide region on ApoER2, whichspecifically interacts withb GPI using phage display2

technology(unpublished observations). It is proposedthat the interaction ofb GPI with ApoER2 on platelets2

and endothelial cells crosslinked with anti-b GPI anti-2

bodies activates the platelets and endothelial cells.

b -Glycoprotein I, plasmin cleavage and platelet2

activation

The 5th domain ofb GPI contains a surface exposed2

loop which is susceptible to proteolytic cleavage. Wehave previously reported thatb2GPI is proteolyticallyclipped in the 5th domain abolishing binding to anionicphospholipids. This cleavage is generated in vitro byplasmin and at low efficiency by factor Xa and in vivoin pathological states of increased fibrinolysis.b2GPIwill bind to many negatively charged macromoleculesincluding heparin glycosaminoglycan. Heparin is regu-larly used in the treatment of APS patients both in theacute therapy and in prophylaxis. We have demonstratedthat b2GPI binds heparin on the 5th domain and thatheparin potentiates the plasmin mediated inactivation ofb GPI (8). Thus heparin may exert its beneficial effects2

in APS patients by two distinct mechanisms includinginactivation ofb GPI. It is likely that anti-b GPI-b GPI2 2 2

complexes bind to activated platelets via negativelycharged phospholipids that become exposed after acti-vation of the platelets. After binding dimericb GPI2

interacts with ApoER2 which results in sensitisation ofthe platelets that make them more prone to respond tocollagen to which they adhere. After interaction dimer-isation of the ApoER2 receptor could lead to thrombox-ane formation by receptor mediated signalling. Plasmingenerated at the site of thrombus formation in the pres-ence of heparan sulphate proteoglycans which are pres-

ent on the endothelial cell surface would potentiate theproteolytical cleavage ofb GPI in the 5th domain abol-2

ishing its binding to anionic phospholipids and thus pre-venting anti-b GPI-b GPI complexes binding to2 2

ApoER2. Cleavage ofb GPI in b GPI–anti-b GPI2 2 2

complexes associated with the platelet ApoER2 wouldinduce dissociation of the complex from ApoER2(pro-posed model see Fig. 1).

b -Glycoprotein I plasmin cleavage and FXI2

activation

Activation of factor XI by thrombin in vivo plays a crit-ical role in coagulation by providing an important pos-itive feedback mechanism for additional thrombingeneration. Factor XI is activated in vitro by thrombinor factor XIIa in the presence of dextran sulphate. Wehave recently reported thatb2-Glycoprotein I binds FXIin vitro and inhibits its activation to FXIa by thrombinand FXIIa. Inhibition of FXI activation occurred withlower concentrations ofb2GPI than found in humanplasma. Proteolytic clipping ofb GPI by plasmin abol-2

ished its inhibition of FXI activation by thrombin(9).This provides a novel mechanism of regulation wherebyphysiological concentrations ofb2GPI may attenuatethrombin generation in vivo by inhibition of FXI acti-vation. Plasmin cleavage ofb GPI provides a negative2

feedback that counteracts its inhibition of FXI activa-tion. As b GPI is the dominant autoantigen in patients2

with antiphospholipid syndrome disregulation of thispathway by autoantibodies may be an important patho-physiological mechanism for thrombosis in patients withthe antiphospholipid syndrome.


References:

1. Kandiah D.A., Sali A., Sheng Y., Victoria E.J., Marquis D.M.,

Coutts S.M., Krilis S.A. Current insights into the ‘‘antiphos-

pholipid’’ syndrome: clinical, immunological, and molecular

aspects. Adv. Immunol. 1998;70:507–63.

2. McNeil H.P., Simpson R.J., Chesterman C.N., Krilis S.A. Anti-

phospholipid antibodies are directed against a complex antigen

that includes a lipid-binding inhibitor of coagulation: beta 2-

glycoprotein I(apolipoprotein H). Proc. Natl. Acad. Sci. USA

1990;87(11):4120–4.

3. Sheng Y., Reddel S.W., Herzog H., Wang Y.X., Brighton T.,

France M.P., Robertson S.A., Krilis S.A. Impaired thrombin

generation in beta 2-glycoprotein I null mice. J. Biol. Chem.

2001 27;276:13817–21.

4. Sheng Y., Kandiah D.A., Krilis S.A. Anti-B2-glycoprotein I

autoantibodies from patients with the antiphospholipid syn-

drome bind to B2-glycoprotein I with low affinity. Dimerization

of B2-glycoprotein I induces a significant increase in anti-B2-

glycoprotein I antibody affinity. J. Immunol. 161: 2038–2043,

1998.

5. Iverson G.M., Reddel S., Victoria E.J., Cockerill K.A., Wang

Y.X., Marti-Renom M.A., Sali A., Marquis D.M., Krilis S.A.,

Linnik M.D. Use of single point mutations in domain I of beta

2-glycoprotein I to determine fine antigenic specificity of anti-

phospholipid autoantibodies. J. Immunol. 2002;169:7097–103.

6. Lutters B.C., Derksen R.H., Tekelenburg W.L., Lenting P.J.,

Arnout J., de Groot P.G. Dimers of beta 2-glycoprotein I

increase platelet deposition to collagen via interaction with

phospholipids and the apolipoprotein E receptor 29. J. Biol.

Chem. 2003;278:33 831–8.

7. Schneider W.J., Nimpf J. LDL receptor relatives at the crossroad

of endocytosis and signaling. Cell Mol. Life Sci. 2003;60:892–

903.

8. Guerin J., Sheng Y., Reddel S., Iverson G.M., Chapman M.G.,

Krilis S.A. Heparin inhibits the binding of beta 2-glycoprotein

I to phospholipids and promotes the plasmin-mediated inacti-

vation of this blood protein. Elucidation of the consequences of

the two biological events in patients with the anti-phospholipid

syndrome. J. Biol. Chem. 2002;2774:2644–9.

9. Tong Shi G., Michael Iverson, Jian C., Qi Keith A. Cockerill,

Matthew D. Linnik, Pamela Konecny, Steven A Krilis.b-

Glycoprotein I binds Factor XI and inhibits its activation by2

thrombin and Factor XIIa. Loss of Inhibition by Clippedb-

Glycoprotein I. Proc. Natl. Acad. Sci. USA.(in press 2004).2

21.Treatment of severe systemic lupuserythematosus: a work in progress

D. Boumpas and P. Sidiropoulos,(Departments of Internal Medicine and Rheumatology,Clinical Immunology and Allergy, University of Crete,P.O. Box 2208, Heraklion GR-71003, Greece).Email: [email protected]

The management of severe lupus has fascinated physi-cians across a wide range of specialties of internal med-icine and requires superb skills in general internalmedicine as well expertise in the use of immunosup-pressive therapy. Although severe lupus is traditionallyconsidered as involvement of major organs(kidneys,lungs, blood, heart, intestine and others) severe involve-ment of other organs such as the skin and the jointsrepresents an equally important challenge. Moreover,major improvements in survival of patients with severedisease have stimulated an increased interest in mini-mizing drug related toxicity and improving quality oflife and compliance of therapy.For practical purposes the treatment of SLE and lupusnephritis has been divided into induction and mainte-nance phases. Recommendations for both induction andmaintenance treatment of lupus nephritis are a subjectof continuous debate.For mild proliferative nephritis without adverse risk fac-tors (especially high risk histology with crescents, fibri-noid necrosis or significant interstitial fibrosis andtubular atrophy, or renal insufficiency) corticosteroidsalone or in combination with azathioprine or mycophen-olate mofetil may be adequate. Failure to achieve remis-sion of disease within three to six months shouldprovoke discussions about more intensive immunosup-pressive therapy.For patients with moderate to severe proliferative dis-ease controlled trials have shown that pulse cyclophos-phamide is the treatment of choice. Cyclophosphamideis given in monthly intravenous pulses for at least sixconsecutive months. Not surprisingly, exacerbations oflupus nephritis were reduced when intravenous cyclo-phosphamide was given every three months for an addi-tional two years. Long-term follow-up of patientsparticipating in one of these controlled trials suggest thatcombining pulse cyclophosphamide with pulse methyl-prednisolone increases efficacy but not toxicity. Pulsecyclophosphamide is associated with an increased riskfor herpes zoster infections on the short-term and withsustained amenorrhea or azoospermia on the long-term.


Because of concerns about the toxicity of intravenouspulse cyclophosphamide there is intense interest inexploring alternative treatments initially for maintenanceand more recently for induction. Following encouragingresults with azathioprine as a maintenance therapy forANCA positive vasculitides, studies are in progress inEurope to further explore the role of azathioprine main-tenance therapy in lupus nephritis. Other studies haveexplored the role of mycophenolate mofetil(MMF).In a controlled study in patients with diffuse proliferativelupus nephritis, Chan et al from Hong-Kong comparedprednisolone and mycophenolate mofetil to prednisoloneand cyclophosphamide. Both treatment arms were fol-lowed by prednisolone and azathioprine. The authorsconcluded that treatment with mycophenolate mofetil isas effective as and has fewer side effects than sequentialtreatment with cyclophosphamide and azathioprine(1).However, in this study follow up was short, patients hadrelatively mild disease and patients with high-risk fac-tors were not included. Longer follow-up demonstratedthat relapses were twice as common in patients treatedwith MMF. Another controlled study from USA com-paring MMF to a 6 month course of pulse cyclophos-phamide for induction therapy was presented by EMGinzler et al in 2003(2). The investigators concludedthat MMF is better than pulse cyclophosphamide for theinduction of remission of lupus nephritis. However thestudy was not powered to demonstrate superiority, fol-low-up was short and there was an unusually high-num-ber of withdrawals from the cyclophosphamide group.More recently, Contreras and colleagues(3) reported theresults of a prospective controlled trial comparing threemaintenance regimens: quarterly intravenous injectionsof cyclophosphamide(pulse cyclophosphamide), oralMMF, and oral azathioprine. Randomization took placeafter patients had received four to six monthly intrave-nous doses of cyclophosphamide given as inductiontherapy. Remission of nephritis occurred in 83 percentof patients during the pulse-cyclophosphamide inductionphase. The proportions of patients who met the criteriafor remission(some degree of reduction in proteinuriaand stable or improved serum creatinine levels) wereevenly distributed among the three maintenance-therapygroups. Overall survival among patients was signifi-cantly higher with azathioprine than with cyclophos-phamide as maintenance therapy. Although the authorsconcluded that after induction therapy with cyclophos-phamide, both azathioprine and mycophenolate mofetilare superior for maintenance therapy to cyclophospham-ide, several limitations of the study suggest a more cau-tious interpretation of these results(3)

A major limitation of these studies of lupus nephritisusing MMF is the short follow-up. Nonetheless thesestudies suggest that MMF is a useful addition to thearsenal of treatments for patients with moderately severelupus nephritis or lupus nephritis refractory to cyclo-phosphamide treatment. The available data also suggestthat both azathioprine and mycophenolate mofetil aregood options for maintenance therapy in patients withproliferative lupus nephritis. Along the same lines, arecent Cochrane review of the studies until January 31,2003 concluded that ‘‘until future RCTs of newer agentsare completed, the current use of cyclophosphamidecombined with steroids remains the best option to pre-serve renal function in proliferative lupus nephritis. Thesmaller effective dose and duration of therapy should beused to minimize gonadal toxicity without compromis-ing efficacy’ (4). Small uncontrolled studies or caseseries have also demonstrated efficacy of pulse cyclo-phosphamide in other types of severe lupus includingneuropsychiatric, hematological, gastrointestinal, pul-monary and dermatologic disease.Renal relapses after an initial response are emerging asa significant problem in proliferative lupus nephritis.Nephritic relapses(defined as active urine sediment withvarious degrees of proteinuria) affect adversely theprognosis of renal disease especially when are associatedwith a decrease in renal function. Reinstitution of immu-nosuppressive therapy is essential in these cases.Other investigations explore the therapeutic potential ofhigh-dose cyclophosphamide in combination with nucle-oside analogs, or biologic response modifiers. High-dosecyclophosphamide or combinations of low-doses withfludarabine may result in profound bone marrow andimmune suppression. Autologous stem cell therapy isalso under evaluation for severe, life threatening lupus.More recently rituximab, a chimeric monoclonal anti-body directed against the CD20 antigen, has been inves-tigated in patients with disease refractory toconventional therapy. A study involving anti-CD40Lwas prematurely terminated because of thromboticevents. Blocking of B Lymphocyte Stimulator(TALL-1, zTNF4, BAFF, THANK) or IFN-g has been shownto ameliorate lupus activity and tissue injury, and maybe tested in humans. Therapy directed at proinflamma-tory cytokines such as TNF-a or IL-1 are also underconsideration. Combinations of cyclophosphamide withbiologic response modifiers have shown encouragingresults in preclinical animal studies. Proteasome inhibi-tors that decrease NF-kB DNA binding activity by pre-venting degradation of IkB, are currently underevaluation in multiple myeloma and may provide an


additional modality to deplete antibody producing plas-ma cells and decrease inflammation in lupus.

References

1. Chan T.M., Li F.K., Tang C.S.D., et al. Efficacy of mycophen-olate mofetil in patients with diffuse proliferative lupus nephri-tis. N. Engl. J. Med. 2000;343:1156–1162.

2. Ginzler E.M., Aranow C., Buyon J., et al. A multicenter studyof mycophenolate mofetil(MMF) vs intravenous cyclophos-phamide(IVC) as induction therapy for severe lupus nephritis(LN): preliminary results. ACR 2003, Abstract 1690.

3. Contreras G., Pardo V., Leclercq B., et al. Sequential therapiesfor proliferative lupus nephritis. N. Engl. J. Med. 2004;350:971–980.

4. Balow J.E., Austin H.A. 3rd. Maintenance therapy for lupusnephritis--something old, something new. N. Engl. J. Med.2004;350:1044–6.

5. Flanc R.S., Roberts M.A., Strippoli G.F.M., Chadban S.J., KerrP.G., Atkins R.C. Treatment for lupus nephritis(CochraneReview). In: The Cochrane library, Issue 1, 2004. Chichester,UK: John Wiley and Sons, Ltd.


SYSTEMIC VASCULITIS

22.The role of bacterial infections for theinitiation and exacerbation of systemicvasculitis

C.G.M. Kallenberg,(Department of Clinical Immunology, University Hos-pital Groningen, The Netherlands).Email: [email protected]

Introduction

Vasculitis can be classified into primary or idiopathicvasculitides and vasculitides secondary to other diseaseprocesses. Within the latter category infections are fre-quently the underlying condition. In some cases directinvasion of a microbial agent into the vessel wall resultsin vasculitis but in most cases immune complexes areinvolved in the pathogenesis of the vasculitic process.These complexes, consisting of microbial antigens andtheir cognate antibodies, can either be formed in situ ordeposited from the circulation resulting in, generallysmall-vessel, vasculitis. One of the classical examples isinfective endocarditis, which is frequently associatedwith vasculitis in skin, kidneys and other organs.Do infections play a role in the primary vasculitides?They certainly do in various forms of primary vasculitisalthough the precise mechanisms involved have not yetbeen fully elucidated. Nevertheless, the associationbetween polyarteritis nodosa and Hepatitis B Virusinfection has been observed for a long time, and, morerecently, mixed essential cryoglobulinemia has beenlinked to Hepatitis C Virus infection. In Kawasaki dis-ease, a form of systemic vasculitis of unknown etiologythat primarily affects infants and young children, a rolefor superantigens derived fromStaphylococcus aureushas been postulated. Superantigens are proteins that bindto class II MHC molecules on antigen presenting cellsand interact simultaneously with specific Vb segmentsof the T-cell receptor(TCR). As such, they are able tostimulate all T-cells that utilize a particular group ofTCR Vb segments, in an antigen-independent way. Var-ious superantigens are also able to stimulate B-cellsexpressing particular variable regions of the heavy chainwithout antigen being present. In Kawasaki disease,S.aureus strains have been isolated expressing varioussuperantigens, in particular the toxic-shock-syndrometoxin-1 (TSST-1) superantigen, and analysis of the Vb

repertoire on the TCR of circulating T-cells showed T-cell expansion compatible with superantigen driven T-cell proliferation w1x. The data presently availablesuggest that, at least in a subpopulation of patients withKawasaki disease,S. aureus derived superantigens playa role in the pathogenesis of the disease.In this review I will focus on the role of infection in theinitiation and exacerbation of Wegener’s Granulo-matosis.

Wegener’s Granulomatosis (WG) and Staphylococ-cus aureus

WG is characterized by granulomatous necrotizinginflammation of the upper and lower respiratory tract inconjunction with systemic vasculitis and necrotizingcrescentic glomerulonephritis. Antineutrophil cytoplas-mic antibodies(ANCA) directed to proteinase 3(PR3-ANCA) are a hallmark of the disease. The diseasefrequently follows an indolent course for weeks tomonths or even years, characterized by ongoing inflam-mation of the upper airways, before the full-blown man-ifestations are apparent. This may suggest that infectiousmicro-organisms are involved in its pathogenesis.Indeed, infections with several micro-organisms wereassociated with relapse of WGw2x.In 1994, Stegeman et al.w3x described a close associa-tion between chronic nasal carriage ofS. aureus and theoccurrence of relapses in patients with WG. First, 36 outof 57 patients with WG were chronic nasal carriers ofS. aureus defined as the presence ofG75% of nasalcultures positive forS. aureus during long-time follow-up. Secondly, relapses almost exclusively occurred inpatients who were carrier ofS. aureus (relative risk forrelapse of 7.16). Otherwise, an association was observedbetween the persistence of ANCA after induction ofremission and the occurrence of relapses: 22 out of 33patients who were intermittently or persistently positivefor ANCA during follow-up developed a relapse asopposed to only one of 21 patients who were persistentlynegative for ANCA during follow-up. In a second study,Stegeman et al. demonstrated that maintenance treatmentwith co-trimoxazole was able to reduce the number ofrelapses in WG patients by 60%w4x.What do we conclude from the foregoing data? First,carriage ofS. aureus appears to create an environmentin which relapsing disease occurs. Secondly, there seems


to be a relationship betweenS.aureus carriage and per-sistence of ANCA.

Pathogenic role of S. aureus in Wegener’sGranulomatosis?

Several studies have been undertaken to elucidate a pos-sible pathogenic role ofS. aureus in WG.The first set of studies started from the observation thatcationic proteins fromS. aureus can bind, due to chargeinteractions, to glomerular structures. We hypothesizedthat, in the presence of specific antibodies, focal glom-erulonephritis could ensue; in the concomitant presenceof ANCA mild glomerulonephritis could be turned intosevere necrotizing glomerulonephritis as we had dem-onstrated before in a model of anti-glomerular basementmembrane diseasew5x. Indeed, we observed thatS.aureus derived acid phosphatase, a cationic protein,could bind, in vitro, to endothelial cellsw6x. Moreover,antibodies to this antigen were present in patients withWG, significantly more than in healthy controls, and theantibodies were able to bind to endothelial cell-boundacid phosphatasew6,7x. Furthermore, the antigen wasdetected in 3 of 19 renal biopsies from patients with WGbut in none of 24 biopsies from control patients. Asmentioned, we hypothesized that vasculitic lesions inWG could start from focal immune deposits with minorinflammation and progress to severe vasculitis withoutimmune deposits due to the neutrophil-activating capac-ity of ANCA. Therefore, we studied biopsies from earlyskin lesions in patients with WG, and we observedimmune deposits in the vessel walls in 4 of 11 biopsiestaken at initial presentation and 4 of 21 biopsies takenat the onset of relapse, whereas 9 renal biopsies in thepatients showed pauci-immune glomerulonephritis irre-spective of the presence(ns5) or absence(ns4) ofimmune deposits in the skin biopsyw8x. Taken together,all of these data suggest that(cationic) antigens fromS.aureus are involved in the initiation and relapses of WG.Secondly, we further evaluated the possible role ofS.aureus derived superantigens in WG. In a longitudinalcohort study we confirmed the relation betweenS.aureus carriage and the occurrence of relapses, as pre-viously found. More importantly, the risk for relapse wasmodulated according to the presence and type of super-antigen fromS. aureus, with toxic shock syndrome tox-in-1 (TSST-1) being associated with a higher risk forrelapse(RR 14.5; 95% CI 2.4–85.4). Next, we assessedwhether activated B- and T-lymphocytes(activated, pos-sibly due toS. aureus derived superantigen stimulation)were present in the peripheral blood of patients withWG. Indeed, we observed increased percentages of acti-

vated B- and T-lymphocytes in WG, with B-cell acti-vation being related to active disease and T-cellactivation persisting during remissionw9x. However,expansions of T-cells were not associated with carriageof specific staphylococcal superantigens in WGw10x.Very recently, Voswinkel et al. analyzed the B-lympho-cyte repertoire in biopsies from early lesions in the upperrespiratory tract in patients with WGw12x. Interestingly,they observed somatic mutation of these B-cells sug-gesting(local) antigenic stimulation. Furthermore, anal-ysis of the heavy chain repertoire of these B-cellssuggested usage of a restricted number of V -genes, andH

V -gene usage was compatible with superantigen stim-H

ulation by S. aureus derived superantigensw11x. Thus,the data presently available suggest that superantigensfrom S. aureus could be involved in the initial stage ofWG by (locally) activating a restricted polyclonalresponse possibly including a specific autoimmuneresponsew12x. An hypothetical scheme of the supposedpathogenesis is given in fig. 1.

Conclusion

Infections are underlying many cases of secondary vas-culitis. In the primary vasculitides microbial agentsseems operative as well. In Wegener’s Granulomatosiscarriage ofS. aureus is associated with disease activa-tion. Various in vivo and in vitro findings are supportivefor the role of S. aureus in the pathogenesis of thisdisease.

References

1. Leung D.Y.M., Meissner H.C., Schlievert P.M. The etiologyand pathogenesis of Kawasaki disease – how close are we toan answer. Curr. Opin. Infect. Dis. 1997;10:226–232.

2. Pinching A.J., Rees A.J., Pussell B.A. et al. Relapses in Wege-ner’s Granulomatosis: the role of infection. Br. Med. J.1980;281:836–838.

3. Stegeman C.A., Cohen Tervaert J.W., Manson W.L., SluiterW.J., de Jong P.E., Kallenberg C.G.M. Association of chronicnasal carriage ofStaphylococcus aureus and higher relapserates in Wegener’s Granulomatosis. Ann. Int. Med.1994;120:12–17.

4. Stegeman C.A., Cohen Tervaert J.W., de Jong P.E., KallenbergC.G.M. Trimethoprim-sulfamethoxazole for the prevention ofrelapses of Wegener’s Granulomatosis. N. Engl. J. Med.1996;335:16–20.

5. Heeringa P., Brouwer E., Klok P.A., Huitema M.G., van denBorn J., Weening J.J., Kallenberg C.G.M. Autoantibodies to


myeloperoxidase aggravate mild anti-glomerular-basement-membrane-mediated glomerular injury in the rat. Am. J. Path-ol. 1996;149:1695–1706.

6. Brons R.H., Bakker H.I., van Wijk R.T., et al. Staphylococcalacid phosphatase binds to endothelial cells via charge inter-action; a pathogenic role in Wegener’s granulomatosis? Clin.Exp. Immunol. 2000;119:566–573.

7. Brons R.H., Kallenberg C.G.M., Cohen Tervaert J.W. AreANCA-associated vasculitides pauci-immune? Rheum. Dis.Clin. North Am. 2001;27:833–848.

8. Brons R.H., de Jong M.C.J.M., de Boer N.K., Stegeman C.A.,Kallenberg C.G.M., Cohen Tervaert J.W. Detection of immunedeposits in skin lesions of patients with Wegener’s Granulom-atosis. Ann. Rheum. Dis. 2001;60:1097–1102.

9. Popa E.R., Stegeman C.A., Bos N.A., Kallenberg C.G.M.,Cohen Tervaert J.W. Differential B- and T-cell activation inWegener’s Granulomatosis. J. Allergy Clin. Immunol.1999;103:885–894.

10. Popa E.R., Stegeman C.A., Bos N.A., Kallenberg C.G.M.,Cohen Tervaert J.W. Staphylococcal superantigens and T-cellexpansions in Wegener’s Granulomatosis. Clin. Exp. Immunol.2003;132:496–504.

11. Voswinkel J., Kramer J., Muller A., et al. B lymphocytes infil-¨ ¨trating Wegener’s granuloma: the immunoglobulin V geneH

repertoire from granulomatous tissues displays an antigen-driven maturation and suggests a microbial trigger. ArthritisRes. 2004;6:S1,24.

12. Popa E., Stegeman C.A., Kallenberg C.G.M., Cohen TervaertJ.W. Staphylococcus aureus and Wegener’s granulomatosis.Arthritis Res. 2002;4:77–79.

Figure 1: Hypothesized pathophysiological role ofS. aureus in Wegener’s Granulomatosis.S. aureus infection can elicit an activation of theinnate immune system via the inflammatory path-

way. Activated monocytes produce proinflamma-tory cytokines such as IL-1 and TNF-a, resultingin upregulation of adhesion molecules on endothe-lial cells and upregulation of adhesion moleculesand expression of proteinase 3(PR3) on the cellsurface of PMNs. As a consequence, PMNs adhereto the vessel wall. Phagocytes engulf staphylococcithat are killed. During this process, degranulationof phagocytes occurs resulting in the release ofPR3. Furthermore, penetration of staphylococcalsuperantigens(SAgs) may result in the activationof monocytes and of PR3-specific B cells to anenhanced production of ANCA when these B cellsare, in the presence of PR3, ‘bridged’ to SAg-reac-tive helper T cells. Subsequently, ANCA stimulatesPMNs that adhere to endothelial cells which resultsin local production of reactive oxigen species(ROS) and release of the proteolytic enzymes thatdamage the vessel wall. Finally, SAgs may bridgeautospecific T cells to other antigen presentingcells than B cells, leading to autoreactive T-cellactivation, as an intermediate step towards granu-loma formation.

23.Immunopathology and new therapeuticconsiderations in ANCA-associatedvasculitides

W.L. Gross,(Department of Rheumatology, University Hospital ofSchleswig-Holstein, Campus Luebeck, and Rheumakli-nik Bad Bramstedt, Ratzeburger Allee 160, 23538 Lue-beck, Germany)Email: [email protected]

From the first identification of antineutrophil cytoplas-mic antibodies(ANCA) 20 years ago till to date, grow-ing evidence suggests that ANCA are not only avaluable diagnostic tool in entities now termed the‘ANCA-associated vasculitides’(Wegener’s Granulom-atosis (WG), Microscopic Polyangiitis(MPA) andChurg-Strauss Syndrome(CSS), but are also pathogenicby activating neutrophils what finally leads to endothe-lial cell damage and vasculitis. First evidence for a pos-sible role of ANCA in vasculitis arose in 1982, whenthe presence of a ‘antineutrophil autoantibody’ with nec-rotizing glomerulonephritis was observed. One year lat-er, Hall et al. found the same autoantibodies in four


patients with unclassified vasculitis. In 1985, a group ofdutch and danish researchers were the first to show astrong association of a distinct fluorescence type of anti-neutrophil antibodies(ACPA; now: cANCA) and amore distinct form of vasculitis(WG), a finding whichwas confirmed by us few month later. Then it was foundthat certain types of small vessel vasculitides—the socalled pauci-immune vasculitides—were associated witha different fluorescence pattern, a cytoplasmic pattern(C-ANCA) in WG and a perinuclear pattern(P-ANCA)in MPA. Soon after this first period of immunodiagnos-tic progress, the target antigen of C-ANCA in WG, Pro-teinase 3(PR3, synonym ‘‘Wegener’s autoantigen’’)was detected by us, while the Chapel-Hill-group iden-tified myeloperoxidase(MPO) as the main target anti-gen of P-ANCA in MPA. Till to date, many studies haveconfirmed that testing both C-ANCAplus PR3-ANCAand P-ANCA plus MPO-ANCA is the major step inproving the clinical diagnosis of one of the ANCA asso-ciated vasculitides. In addition, the initial impressionthat ANCA titers correlate with disease activity has beenconfirmed at least partially. Thus, from a clinical pointof view there is a strong belief in the pathogenic capa-bility of PR3- or MPO-ANCA. Therefore, over the lastdecade the ANCA scientific community pushed forwardthe observations that ANCA participate in the pathogen-esis of the associated diseases. Among the number ofhypotheses how ANCA may cause vasculitis, the mostaccepted (‘ANCA-Cytokine-Sequence-Theory’) isbased on the observations that the cytokine-inducedexpression of granule proteins(i.e., PR3) on the surfaceof neutrophils and monocytes allows ANCA an inter-action with surface antigens. This interaction results inan activation of neutrophils and interaction with theendothelium resulting in its damage. Neutrophils play animportant role in the pathogenesis of pauci-immune vas-culitis (i.e. those without immune deposits in situ in con-trast to the immune-complex induced forms): theypredominate at the site of tissue injury(necrotizing vas-culitis and granuloma) and they are the main target cellassociated with the ANCA antigens. How neutrophilsbecome activated by ANCA, in particular, which signaltransduction pathways are used, is still a matter of inves-tigation. ANCA signaling is most likely a consolidationof signals produced by both ANCA-F(ab‘)2 and ANCA-Fc engagement. Furthermore, ANCA-induced signalingcan synergize with arachidonic acid- and with tumor-necrosis factor- a(TNF-a)-signaling pathways.Increased interest in ANCA as a signaling molecule hasbeen fueled by findings that neutrophils respond to thephysical cues of ANCA by up-regulating transcriptionof genes, such as IL-1b, IL-8, cyclooxygenase 2 and

differentiation-dependent gene 2. The consequence ofneutrophil activation for endothelium and tissue is aninflammatory process that becomes dysregulated. Incontrast to regular uptake of apoptotic cells by macro-phages, which is noninflammatory, interaction of ANCAwith apoptotic neutrophils expressing PR3 enhancesclearance by macrophages and induces a proinflamma-tory response with release of IL-1, IL-8 and TNF-a.These observations may help to explain the amplifica-tion and perpetuation of inflammation associated withsevere necrotizing vascular injury. Furthermore, a recentstudy showed the development of ANCA, albeit ofundefined specificity, in rats after multiple injections ofapoptotic neutrophils. So, theoretically, accumulation ofapoptotic neutrophils may boost the PR3-specific auto-immune response.Very recently, Xiao et al., for the first time providedevidence that ANCA can in fact be pathogenic in vivoand that this involves the activation of neutrophils. Theytransferred autoantibodies directed against murine MPOinto recombinase-activating gene-2-deficient(Rag-2)knockout mice which lack functional T and B andobserved the development of vasculitic lesions whichreassemble those seen in MPA, thus fulfilling Witeb-sky’s postulate for autoimmunity. In summary, we nowknow that a passive transfer of MPO-ANCA is sufficientto induce disease, but it remains to be discovered howthe production of ANCA is triggered. Furthermore, asatisfactory animal model for PR3-ANCA induced vas-culitis has yet to be developed.Today, these findings are of major interest with regardto new treatment modalities. The recent observationfrom a controlled trial that the removal of antibodies byplasmapheresis is superior to conventional immunosup-pressive treatment with methylprednisolone furtheralienates the importance of ANCA in the developmentof vasculitis. In addition, the knowledge about the keyrole of cytokines in the pathogenesis of vasculitides hasopened avenues for therapeutic interventions(e.g. TNF-a blocking agents). Finally, the effectiveness of anti-CD20 antibodies in patients with refractory WG pointstowards a possible role of B-cells in WG which warrantsfurther investigation.Thus, after 20 years of extensive research we have learn-ed that activation of neutrophils by ANCA is central forthe pathogenesis of the ANCA-associated vasculitides.In the future, a better understanding of the underlyingmechanisms that lead to the generation and productionof ANCA may open new ways for treatment of theseformerly incurable diseases.


24.Clinical pictures of and therapeuticstrategies for systemic vasculitides

L. Guillevin,(Service de Medecine Interne, Hopital Cochin, Assis-´ ˆtance Publique–Hopitaux de Paris, Universite Paris V,ˆ ´27, rue du Faubourg Saint-Jacques, 75679 Paris Cedex14, France.).E-mail: [email protected]

Introduction

Systemic vasculitides can affect every organ, causingsome clinical manifestations that are severe. Vasculitistreatment of should be adapted to the severity of theclinical involvement. The combination of corticosteroids(CS) and cyclophosphamide(CYC) has been widelyprescribed for polyarteritis nodosa(PAN), microscopicpolyangiitis (MPA), Churg–Strauss syndrome(CSS)and Wegener’s granulomatosis(WG). The regimenshould be chosen after careful analysis of classification,predictable outcome, etiology, pathogenesis, severity,age and other involvement. We review herein the ther-apeutic strategies now being applied to treat severe sys-temic vasculitides.

Factors influencing treatment choice

The vasculitic process may affect blood vessels of allsizes from the aorta to capillaries. The American Collegeof Rheumatology has defined different classification cri-teria for distinct vasculitides(1–3) but the classificationthat has been adopted by the scientific community todifferentiate PAN from MPA, is the Chapel Hill Nomen-clature(4). The outcome of vasculitides varies from onedisease to another and the relapse rate also varies, from5% for hepatitis B virus-related PAN(HBV-PAN) (5)to 23.4% for CSS(6), 34.1% for MPA (7) and morethan 50% for WG(8). Treatment duration should prob-ably be decided, at least in part, according to the risk ofrelapse. For WG, prolonged therapy does not seem ableto lower the relapse rate below 50%(8), although treat-ments lasting less than 18 months are almost alwaysassociated with a high relapse rate, ranging from 50 to100%.The etiologies of a few cases of vasculitides have beenidentified, with infections, viral or bacterial, proven ordiscussed. HBV is considered to be the etiological agentof a minority of patients with non-PAN systemic vas-culitides (-1%), although it has been consideredresponsible for more than one-third of the patients devel-oping PAN (5). Hepatitis C virus(HCV) is now con-

sidered to be the etiological agent of more than 90% ofpatients with mixed cryoglobulinemia(9).

TREATMENTS

Treatment of HBV- and other virus-related PAN

HBV-related PAN

Based on the efficacy of antiviral agents in chronic hep-atitis and of plasma exchanges(PE) in PAN, we com-bined both therapies to treat HBV-PAN. The rationale ofthe therapeutic sequence was as follows: initial CS, torapidly control the most severe life-threatening manifes-tations of PAN which are common during the first weeksof the disease; their abrupt stoppage to enhance immu-nological clearance of HBV-infected hepatocytes andfavor HBe antigen to anti-HBe antibody seroconversion;and PE to control the course of PAN.In HBV-PAN, the combination of antiviral agents(vidar-abine, interferon-alpha 2b(IFNa2b) and, more recently,lamivudine) and PE gave excellent overall therapeuticresults and should be preferred to conventional regi-mens, which are more risky in this setting(as they facil-itate virus replications and subsequent relapse) (5).

Mixed cryoglobulinemia

HCV-related cryoglobulinemia is asymptomatic in themajority of patients but persists for decades and the dis-ease duration could be a factor associated with theoccurrence of clinical symptoms(10). In asymptomaticpatients, no argument supports initiating treatment andclose monitoring could be sufficient. In patients withmoderate symptoms(arthralgias, purpura, sensoryperipheral neuropathy, for instance), IFNa2b or a com-bination of IFNa2b and ribavirin should be tried(11,12). In patients with severe clinical symptoms of mixedcryoglobulinemia, PE could be useful, especially to cureleg ulcers and to help reverse other manifestations.

Treatment choice according to disease severity

Initial therapeutic choice

It now seems reasonable to adapt the first-line therapyto the severity of the vasculitis and not to propose sys-tematically a standard treatment. To help the clinicianchoose the most effective regimen and to avoid over-treatment, we have established(13), a five-factor score(FFS), which has significant prognostic value, andwhose parameters, defined as follows, were responsiblefor higher mortality: proteinuria)1 gyday, renal insuf-ficiency (creatininemia )140 mmolyl), cardiomyo-


pathy, gastrointestinal manifestations and centralnervous system involvement. For FFS of 0, 1 andG2,the respective 5-year mortality rates were: 12%, 26%and 46%. In a study on 278 patients(14) with PAN,MPA or CSS, we demonstrated that combining CYC andCS was beneficial for patients with an FFSG2. Thepatients who died from severe vasculitides had moreoften been treated with CS than with the combination ofCS and CYC. Other criteria, like the Birmingham vas-culitis activity score(BVAS) (15), could also be usedto determine the intensity of treatment to be prescribedand are being evaluated in the prospective trials pro-posed by the European Vasculitis(EUVAS) group.

Indications of steroids and cyclophosphamide inpolyarteritis nodosa and Churg–Strauss syndrome

CS and immunosuppressive drugs, especially CYC, havetransformed the prognoses of these vasculitides. CSalone have been able to increase the 5-year survival ratefor untreated patients from 10% to about 55% in themid-to-late 1970s. Survival was further prolonged byadding immunosuppressants, either azathioprine or CYCto the treatment regimen.

Cyclophosphamide

In PAN and CSS, when CYC is indicated in patientswith factors of poor prognosis(FFSG1), an IV pulseshould be preferred to oral administration. The IV routeobtains a more rapid clinical response than oral CYC,which is important in patients with active disease. Treat-ment duration with CS and CYC should not exceed 1year.

Treatment of Wegener’s granulomatosis

Because CS and CYC treatments are prescribed to allWG patients, we will concentrate on cytotoxic agents.To date, no consensus has been reached concerning CYCtreatment, even though its indication is universallyaccepted. Oral administration of 2 mgykgyd is pre-scribed(8). The dose should be adapted according tothe therapeutic response, the occurrence of side effects,renal function and age. Treatment should last at least 18months but other immunosuppressants, like azathioprine,can be prescribed as a maintenance therapy.Pulse CYC has also been assessed. CYC was given eve-ry 3 to 4 weeks at a dose of 0.5 to 0.7 gym . Remission2

rates were comparable to those obtained with the oralroute. Nevertheless, the number of relapses was highafter stopping treatment(16).

Treatment of microscopic polyangiitis

We now recommend treating MPA like WG, based onthe presence of putative common pathogenetic mecha-nisms and the preliminary results of ongoing trials.

MISCELLANEOUS TREATMENTS

Plasma exchanges

PE can be a useful tool, as second-line treatment, in non-HBV-related PAN refractory to conventional therapy. Inpatients with crescentic glomerulonephritis responsiblefor severe renal insufficiency(creatininemia)500mmolyl), Pusey and coworkers have shown that PE canrapidly improve renal function and enable patients tostop dialysis(unpublished data).

Intravenous immunoglobulins

IVIg have essentially been used in WG and MPA. Themajority of the patients appeared to improve, with sus-tained benefit and reduced requirement for immunosup-pression(17).

Other immunosuppressive agents; new drugs

Azathioprine is commonly used as maintenance therapyand seems to be effective and well tolerated(18).Methotrexate has also been proposed for maintenanceand relapsing patients(19). The initial dose of 0.3 mgykg is delivered(intramuscular injection or orally) onceweekly. Although its efficacy is inferior to that of CYC,good results have been obtained(19).Mycophenolate mofetil and leflunomide are being eval-uated for maintenance therapy, or in patients whorelapsed or were refractory to the combination of CSand CYC. The initial results have been promising butcannot yet be extrapolated to general use. Recently, anti-tumor necrosis factor-alpha(TNFa) antibodies(inflixi-mab) (20) or TNFa-receptor analogues(e.g. etanercept)have been proposed to treat systemic vasculitides.

References

1. Lightfoot R.W., Jr., Michel B.A., Bloch D.A., Hunder G.G.,Zvaifler N.J., McShane D.J., et al. The American College ofRheumatology 1990 criteria for the classification of polyarter-itis nodosa. Arthritis Rheum. 1990;33:1088–93.

2. Masi A.T., Hunder G.G., Lie J.T., Michel B.A., Bloch D.A.,Arend W.P., et al. The American College of Rheumatology1990 criteria for the classification of Churg–Strauss syndrome(allergic granulomatosis and angiitis). Arthritis Rheum.1990;33:1094–100.


3. Leavitt R.Y., Fauci A.S., Bloch D.A., Michel B.A., HunderG.G., Arend W.P., et al. The American College of Rheuma-tology 1990 criteria for the classification of Wegener’s gran-ulomatosis. Arthritis Rheum. 1990;33:1101–7.

4. Jennette J.C., Falk R.J., Andrassy K., Bacon P.A., Churg J.,Gross W.L., et al. Nomenclature of systemic vasculitides. Pro-posal of an international consensus conference. ArthritisRheum. 1994;37:187–192.

5. Guillevin L., Lhote F., Cohen P., Sauvaget F., Jarrousse B.,Lortholary O., et al. Polyarteritis nodosa related to hepatitis Bvirus. A prospective study with long-term observation of 41patients. Medicine(Baltimore) 1995;74:238–53.

6. Guillevin L., Cohen P., Gayraud M., Lhote F., Jarrousse B.,Casassus P. Churg–Strauss syndrome. Clinical study and long-term follow-up of 96 patients. Medicine(Baltimore)1999;78:26–37.

7. Guillevin L., Durand Gasselin B., Cevallos R., Gayraud M.,Lhote F., Callard P., et al. Microscopic polyangiitis: clinicaland laboratory findings in eighty-five patients. ArthritisRheum. 1999;42:421–30.

8. Hoffman G.S., Kerr G.S., Leavitt R.Y., Hallahan C.W., Lebov-ics R.S., Travis W.D., et al. Wegener granulomatosis: an anal-ysis of 158 patients. Ann. Intern. Med. 1992;116:488–98.

9. Agnello V., Chung R.T., Kaplan L.M. A role for hepatitis Cvirus infection in type II cryoglobulinemia. N. Engl. J. Med.1992;327:1490–5.

10. Rieu V., Cohen P., Andre M., Mouthon L., Godmer P., Jar-´

rousse B., et al. Characteristics and outcome of 49 patientswith symptomatic cryoglobulinaemia. Rheumatology(Oxford)2002;41:290–300.

11. Ferri C., Marzo E., Longombardo G., Lombardini F., La CivitaL., Vanacore R., et al. Interferon-alpha in mixed cryoglobuli-nemia patients: a randomized, crossover-controlled trial. Blood1993;81:1132–6.

12. Cacoub P., Lidove O., Hausfater P., Maisonobe T., Thibault V.,Charlotte Fea. Antiviral treatment and outcome in patientswith hepatitis C virus systemic vasculitis. Arthritis Rheum.2001;44: S56,21.

13. Guillevin L., Lhote F., Gayraud M., Cohen P., Jarrousse B.,Lortholary O., et al. Prognostic factors in polyarteritis nodosaand Churg–Strauss syndrome. A prospective study in 342patients. Medicine(Baltimore) 1996;75:17–28.

14. Gayraud M., Guillevin L., le Toumelin P., Cohen P., Lhote F.,Casassus P., et al. Long-term followup of polyarteritis nodosa,microscopic polyangiitis, and Churg-Strauss syndrome: anal-ysis of four prospective trials including 278 patients. FrenchVasculitis Study Group. Arthritis Rheum. 2001;44:666–75.

15. Luqmani R.A., Bacon P.A., Moots R.J., Janssen B.A., Pall A.,Emery P., et al. Birmingham Vasculitis Activity Score(BVAS)in systemic necrotizing vasculitis. Quat. J. Med. 1994;87:671–8.

16. Guillevin L., Cordier J., Lhote F., Cohen P., Jarrousse B., Roy-er I., et al. A prospective, multicenter, randomized trial com-paring steroids and pulse cyclophosphamide versus steroidsand oral cyclophosphamide in the treatment of generalizedWegener’s granulomatosis. Arthritis Rheum. 1997;40:2187–98.

17. Jayne D. Intravenous immunoglobulins in the therapy of sys-temic vasculitis. Tranfus Sci. 1992;13:317–24.

18. Jayne D., Rasmussen N., Andrassy K., Bacon P., et al. A ran-domized trial of maintenance therapy for vasculitis-associatedwith antineutrophil cytoplasmic antibodies. N. Engl. J. Med.2003;349:36–44.

19. Langford C., Talar-Williams C., Barron K., Sneller K. A stagedapproach to the treatment of Wegener’s granulomatosis: induc-tion of remission with glucocorticoids and daily cyclophos-phamide switching to methotrexate for remisison maintenance.Arthritis Rheum. 1999;42:2666–73.

20. Bartolucci P., Ramanoelina J., Cohen P., Godmer P., Le HelloC., Guillevin L. Efficacy of anti-TNF antibody infliximabagainst refractory systemix vasculitides. An open pilot studyon 10 patients. Rheumatology(in press) 2002.

25.Systemic small vessel vasculitis:pathophysiological aspects ofantineutrophil cytoplasm antibodies(ANCA). A review on recent findings

A. Wiik,(Department of Autoimmunology, Statens Serum Insti-tut, Artillerivej 5, DK-2300 Copenhagen S, Denmark.).E-mail: [email protected]

Certain systemic small vessel vasculitides involvingvenules, capillaries and often arterioles with necrotizingvessel wall lesions are regularly associated with produc-tion of specific autoantibodies to neutrophil and mono-cyte granule constituents(ANCA), particularly directedto proteinase 3(PR3) or myeloperoxidase(MPO).These diseases are known as Wegener9s granulomatosis(WG), microscopic polyangiitis(MPA) and Churg-Strauss syndrome(CSS) as well as limited forms ofthese conditions.Immunoinflammatory events leading to evolution ofsuch conditions are thought to involve activation of neu-trophils and endothelial cells by pro-inflammatory cyto-kines, ANCA and upregulated adhesion moleculeswhich cause neutrophils to adhere in a fixed way to thevessel walls. This results in the induction of respiratoryburst with release of toxic oxygen radicals and activelysosomal enzymes onto the vessel wall and subsequent-


ly structural damage. Autoantibodies to endothelial cellsmay cooperate with ANCA in orchestrating and propa-gating this abnormal inflammatory cascade attack on thevessel walls.PR3-ANCA of the IgG class are produced in most WGpatients during active phases of the disease. Though thesubclass distribution of the IgG ANCA may tend to shifttowards more IgG3 production in the active phase westill lack good explanations why IgG PR3-ANCAderived from active phase WG patients seem to have aspecial potential to activate neutrophils by attaching tosurface exposed PR3 by their Fab2 part and to Fcg-receptors (FcgR2a) on neutrophils, thereby causingthem to become hyperactivated. These PR3-ANCA havealso been suggested to be capable of interfering withnormal complexation of PR3 witha1-antitrypsin, thenatural serum inhibitor of several serine proteases. Thismay leave enzyme activity uninhibited and further pro-mote tissue damage. The imbalance between enzymeand inhibitor is further worsened by inactivation ofa1-antitrypsin by oxygen radical damage to this protein. Onthe other hand, PR3-ANCA taken from active WGpatients has also been shown to form complexes withthe active enzyme thereby inhibiting the enzyme activity,however, this inhibition may not be long-lasting sinceIgG molecules are themselves a substrate for enzymaticdegradation by PR3.Another fascinating property of PR3-ANCA taken fromactive WG patients is its immunogenicity in animals.Thus, active immunization of mice with minute amountsof human PR3-ANCA has been shown to cause produc-tion of an anti-idiotypic network response where the ani-mals start producing PR3-ANCA after several months,reflecting an anti-anti-idiotype response(1). In this par-ticular model only a modest mononuclear cell infiltrationwas produced and a full pathophysiological potentialwas not elicited.Very recent studies have further elucidated how an anti-idiotype response can lead to production of PR3-ANCA(2). Pendergraft et al. detected how autoimmunity toPR3 can be triggered through ‘autoantigen complemen-tarity’. This scenario involves immune response to anantigen having an aminoacid sequence that is comple-mentary (anti-sense aminoacid sequence) to humanPR3. The particular peptide studied was the areabetween aminoacids 105 and 201. This peptide wasfound to bind with high selectivity to the natural 105–201 peptide of PR3. Once an immune response towardsthe complementary peptide arises the resulting antibodycarrying an idiotope will further trigger production ofanti-idiotypic antibody which exactly mimics PR3-ANCA directed to peptide 105–201. This was proven

by immunization of mice with the complementary pep-tide which elicited antibody production to this peptideas well as to the sense 105–201 peptide. The two anti-body populations were shown to be independent, butthey were able to interact with each other as could beexpected from idiotypic pairs.When patients with PR3-ANCA-associated vasculitiswere studied sera from 7 out of 34 were found to har-bour antibodies to the complementary peptide alongwith PR3-ANCA as shown by ELISA and immunopre-cipitation techniques. These populations were affinitypurified and again behaved as an idiotypic pair with nocross-reactivity.The complementarity peptide-like molecules might bederived from endogenous or exogenous proteins. Whencirculating leucocytes from PR3-ANCA vasculitispatients were studied by RT-PCR 10 out of 22 werepositive showing that such molecules had been tran-scribed. None of the healthy controls, the ANCA-nega-tive vasculitis patients nor patients with systemic lupuserythematosus were positive. Interestingly, whensearched for in databases complementarity peptide-likesequences were found in several microbial agents andamong themStaphylococcus aureus andEntamoeba his-tolytica. These two microbes have been associated withproduction of PR3-ANCA during active infection.Chronic nasal carriage of staphylococci additionally hasbeen shown to increase the risk of getting relapses inWG (3).Most animal models of vasculitis have been too artificialto give insight into the pathogenic role of the ANCAthemselves. Recent studies by Xiao et al. utilized MPOknockout mice which were immunized with mouseMPO (4). Spleen cells from these mice were injectedinto recombinase-activating-gene-2-deficient(Rag2yyy) mice that lack functional T- and B-cells, and this ledto development of severe necrotizing vasculitis and cres-centic glomerulonephritis. Both Rag2yyy and wildtype C57BLy6J mice developed focal necrotizing andcrescentic glomerulonephritis with few immune depositsupon intravenous injection of anti-MPO IgG alone,while normal mouse IgG did not produce lesions. Thesedata support the assumption that MPO-ANCA carrytheir own pathophysiological potential if produced insufficient amounts.Earlier data from several groups have shown that PR3-specific T-cells are present in WG patients and can reactwith proliferation and cytokine production.There is, however, little agreement regarding MHCgenes conferring risk for vasculitis or ANCA produc-tion. MPO-specific T-cells are more difficult to studysince MPO is toxic to lymphocytes.


Conclusions

Though many investigations have pointed to very com-plex pathogenetic pathways leading to small vessel vas-culitis, there is emerging evidence that both PR3-ANCAand MPO-ANCA have important roles in the pathophy-siology of these diseases. Exogenous antigens are prob-ably important for the induction of PR3-ANCAresponses.

1. Tomer Y., Gilburd B., Blank M., Lider O., Hershkovitz R., Fish-man P., Zigelman R., Meroni P.L., Wiik A., Shoenfeld Y. Char-acterization of biologically active antineutrophil cytoplasmicantibodies induced in mice. Pathogenic role in experimentalvasculitis. Arthritis Rheum. 38: 1375–1381,1995.

2. Pendergraft W.F., Preston G.A., Shah R.R., Tropsha A., CarterC.W., Jennette J.C., Falk R.J. Autoimmunity is triggered bycPR3 (105–201), a protein complementary to human autoan-tigen proteinase-3. Nature Med. 10: 72–79,2003.

3. Stegeman C.A., Tervaert J.W., Sluiter W.J., Manson W.L., de-Jong P.E., Kallenberg C.G. Association of chronic nasal carriageof Staphylococcus aureus and higher relapse rates in Wegener9sgranulomatosis. Ann. Intern. Med. 120:12–17,1994.

4. Xiao H., Heeringa P., Hu P., Liu Z., Zhao M., Aratani Y., MaedaN., Falk R.J., Jennette J.C. Antineutrophil cytoplasmic autoan-tibodies specific for myeloperoxidase cause glomerulonephritisand vasculitis in mice. J. Clin. Invest. 110:955–963,2002.

26.The lumps and bumps of Behcet’s¸syndrome

H. Yazici,(Division of Rheumatology, Dept. of Medecine, Cerrah-pasa Medical Faculty, University of Istanbul, AksarayIstanbul, 34303 Turkey)Email: [email protected]

The nodular lesions of BS are of basically two differentpathologies: a. e. nodosum like lesions and b. superficialthrombophlebitis(STM). More recently we are becom-ing aware that the demography, clinical differentiationand the clinical associations of these lesions are ratherinformative. Several years ago we have shown(1) thatthe e. nodosum – like lesions are, in fact, e. nodosum –like in that, contarary to e. nodosum, idiopathic or asso-ciated with other diseases, contain areas of vasculitis.The main histology of the STM lesions on the otherhand, is, as expected, a thrombosed vein. The clinicaldifferention between the two could be difficult by nakedeye. In a recent study by C. Mat and colleagues(2)among 55 BS patients with nodular lesions and 19 none

BS patients with nodular lesion due to other causes,using blinded histology as the golden diagnostic tool anddermal ultrasound(US) – where the e. nodosum likelesions have a hyper and a the STM lesions a hypo echo-ic pattern-as an aid to clinical diagnosis, we haveobserved that the sensitivity and the specificity of thenaked eye to diagnose these lesions were between 70and 95% and the use of US slightly improved our diag-nostic accuracy. It was however interesting to note thatall but one of the 17 biopsy proven STM were amongthe males while the MyF ratio among those patients withEN like lesions was 1y5. We were aware for many yearsthat EN like lesions of BS were more common amongthe females. We now learn that STM is more commonthis time among the males. This information further tiesin with our recent observations that a. STM and deepvein thrombosis(DVT) clustered together in a prospec-tive factor analysis among 272 patients with BS(3) andb. among 88 patients with central nervous systeminvolvement 65% of 17 patients with CNS disease in theform of dural sinus thrombi(DST) had DVT while thefrequency of DVT was 19% in the remaining patientswith parenchymal CNS disease(4). Similar to what wasobserved in STM, 16y17 of the patients with DST weremales.Thus the thrombophilia of BS, unlike the more commonforms of thrombophilia, is mainly a burden for the maleand has a spectrum of STP™DVT™DST.The third important lump or bump in BS is the papu-lopustular lesion. These lesions are usually indistin-guishable from ordinary acne with the main exceptionthat they may occur also in sites uncommon for ordinaryacne like arms and legs. The conventional wisdom wassuch that these lesions were sterile. We have recentlyformally studied the microbiology of these lesions(5)and showed this was not the case.Staphylococcus aureus was grown in 58% of pustules inBS patients, while it was grown in 29% of pustules fromacne vulgaris patients. Pustules on unusual acne sitessuch as arms and legs seemed to be mainly responsiblefor this difference in thatS. aureus was grown in 86%of pustules on unusual acne sites.Prevotella sp. bacteriawhich grow on secondarily infected skin lesions such aspsoriasis and eczema rather than in acne vulgaris weregrown in 17 of 70(24%) BS pustules, whereas it wasnot grown in any of the 37 pustular acne vulgarislesions. These bacteria usually grow on secondarilyinfected skin lesions such as psoriasis and eczema andit is early to say that what we have grown have patho-genetic importance. On the other hand these observa-tions make us think that search for a possible infectiousetiology for at least some of the manifestations of BS


has yet to be exhausted(6). This is particularly true forthe arthritis of BS which we have also recently shownto be associated with the papulopustular lesions in twodifferent studies(4, 6).Unlike the potentially blinding eye involvement and thesinister pulmonary artery disease there is nothing threat-ening or glorious about the skin lumps and bumps inBS. Yet, we have hopes they might be of considerablehelp in solving its enigma.

References

1. Demirkesen C., Tuzuner N., Mat C., Senocak M., BuyukbabaniN., Tuzun Y., Yazici H. Clinicopathologic evaluation of nodularcutaneous lesions of Behcet syndrome. Am. J. Clin. Pathol.,2001;116:341–6.

2. Mat C. et al. manuscript in preparation.3. Tunc R., Keyman E., Melikoglu M., Fresko J., Yazici H. Target

organ associations in Turkish patients with Behcet’s disease: across-scrtional study by exploratory factor analysis. J. Rheu-matol., 2002;29:2393–6.

4. Tunc R. et al. Ann. Rheum. Dis., in print.5. Hatemi G. et al. Ann. Rheum. Dis., in print.6. Diri E., Mat C., Hamuryudan V., Yurdakul S., Yazici H. Papu-

lopustular skin lesions are seen more frequently in patients withBehcet syndrome who have arthritis: a controlled and maskedstudy. Ann. Rheum. Dis., 2001;60:1074–6.

27.Clinical presentation and outcome ofsystemic sclerosis

P.G. Vlachoyiannopoulos,(Department of Pathophysiology, Medical School,National University of Athens, 75 Mikras Asias str,11527, Athens, Greece).Email: [email protected]

Systemic sclerosis(SSc) is a multi-organ disease withconsiderable accompanying morbidity. Several clinicalmanifestations and laboratory parameters have been pro-posed as candidate predictors of mortality in previousstudies.There are also considerable differences across studies ineligibility criteria, definitions of manifestations and out-comes. The accumulated literature raises questions onwhether the mortality impact and predictors thereof inSSc are similar across diverse settings or whether thereis genuine heterogeneity in different countries, ethnici-ties and clinical practices.We present the results of the International Meta-analysisof Mortality Impact of Systemic Sclerosis(IMMISS)

which is an international collaborative meta-analysis ofindividual-level data. We examined the consistency ofsurvival outcomes and predictors thereof in SSc indiverse settings using common definitions for diagnosisand organ involvement across a number of medicalcenters.

Objectives

The meta-analysis had two primary objectives:(1) toestimate the standardized mortality ratio for patientswith SSc across different cohorts, and(2) to evaluateand validate predictors that consistently affect mortalityacross studies. Follow-up started at the time of first visitto each participating center with the diagnosis of SSc.

Individual patient level database

Patients were divided into incident and prevalent cases,depending on whether enrollment(first cohort visit)occurred within 6 months of the first physician diagnosisof SSc (incident case) or diagnosis had preceded thefirst visit by )6 months(prevalent case). The reportedmain analyses address only incident cases, unless spec-ified otherwise.

Standardized mortality ratios

For each cohort, the standardized mortality ratio(SMR)was estimated by comparing the number of observeddeaths with the number of expected deaths adjusting forage and sex, according to country-specific life tables forthe calendar years of follow-up in each cohort. Confi-dence intervals were calculated using the inverse of theobserved number of deaths as the variance of the naturallogarithm of SMR. The method of Rothman and Boicegave very similar results. Cohort-specific SMRs weretested for heterogeneity based on general variancemodels.

RESULTS

Meta-analysis database

Seven cohorts(2 from medical centers in the USA, 4from Europe and one from Japan) contributed data on1645 incident SSc cases followed for 11 521 person-years(578 deaths)Common reported causes of death were renal(ns30),cardiac (ns42) and pulmonary involvement(ns35)and pulmonary arterial hypertension(ns41).

Standardized mortality ratios

All cohorts showed significantly increased SMR esti-mates. However, there was very large between-cohortheterogeneity in these estimates, with a range between1.5 and 7.2(P-0.001).


Predictors of mortality

After adjusting for age, sex, and year of enrollment, anti-topoisomerase I antibodies, as well as renal, cardiac andpulmonary involvement, independently increased therisk of death. Anticentromere antibodies, anti-U3RNPantibodies, and esophageal involvement did not adverse-ly affect survival. Analyses including all 3311 subjects(incident and prevalent cases) showed that diffuse skininvolvement at enrollment had an independent adverseeffect on mortality(hazard ratio 1.2, 95% CI, 1.0–1.4,Ps0.03).In all cohorts renal, cardiac and pulmonary involvementwere important predictors, but there was significantbetween-cohort heterogeneity. Esophageal involvementwas not a prominent risk factor for mortality in anycohort.Patients with one serious organ involvement were morelikely to also have another serious organ involvementduring follow-up (correlation coefficients 0.20, for theco-occurrence of cardiac-renal involvement, 0.17 forcardiac-pulmonary involvement, and 0.15 renal-pulmo-nary involvement,P-0.001 for all). Diffuse cutaneousinvolvement at onset was correlated with developmentof renal (rs0.23) and cardiac involvement during fol-low-up (rs0.10), and was also related to anti-topoiso-merase I antibodies(rs0.20, P-0.001 for all).Anti-topoisomerase I antibodies were correlated withpulmonary(rs0.28) and, less so, cardiac involvement(rs0.11,P-0.001 for both).

Discussion and conclusions

SSc clearly increases the risk of death compared withthe general population, with SMRs ranging between 1.5and 7.2 across different cohorts. The development ofrenal, pulmonary and cardiac involvement are importantindependent adverse predictors and the presence of anti-topoisomerase I antibodies increases the risk by an addi-tional 1.3-fold. Patients with one type of major organinvolvement were more likely to have another majororgan system involved as well.Our analysis has some limitations. We could not haveassembled all the study teams working in the SSc field.Non-participating centers may have used different cri-teria and definitions. However, the meta-analysis includ-

ed a very large sample and provides a paradigm forbuilding meta-analyses of individual patient data for pre-dictive factors. Second, we did not use all of the varia-bles that have been considered potential predictivefactors in past studies, in particular laboratory parame-ters such as erythrocyte sedimentation rate, proteinuria,hemoglobin and serum protein levels. We decided not tofocus on non-specific laboratory markers, since they areprobably secondary to the underlying major organ man-ifestations. Third, we could not collect detailed infor-mation on the time of diagnosis of pulmonary arterialhypertension in the available patients. Pulmonary arterialhypertension is also an important cause of mortality inSSc, while renal disease is probably no longer the lead-ing cause of death in the last decade. Fourth, we did notcollect any treatment data. Finally, one cohort contrib-uted the majority of the patients. Unavoidably, theresults in this cohort heavily influence the overall sum-mary results.Our work suggests that for relatively uncommon, butclinically important, diseases such as SSc, it is possibleto accumulate large-scale evidence by collating stan-dardized information from diverse teams. The use ofstandardized individual-level information may helpbypass the problems encountered in meta-analyses ofprognostic factors.We can conclude that SSc confers ahigh mortality risk, but there is considerable heteroge-neity across settings. Internal organ involvement andanti-topoisomerase I antibodies are important determi-nants of mortality.

References

1. Scussel-Lonzetti L., Joyal F., Raynauld J.P. et al. Predictingmortality in systemic sclerosis: analysis of a cohort of 309French Canadian patients with emphasis on features at diagnosisas predictive factors for survival. Medicine(Baltimore).2002;81:154–67.

2. Ferri C., Valentini G., Cozzi F. et al. Systemic sclerosis: dem-ographic, clinical, and serologic features and survival in 1,012Italian patients. Medicine(Baltimore). 2002;81:139–53.

3. Vlachoyiannopoulos P.G., Dafni U.G., Pakas I. et al. Systemicscleroderma in Greece: low mortality and strong linkage withHLA-DRB1*1104 allele. Ann. Rheum. Dis. 2000;59:359–367.


EPIDEMIOLOGIC CONSIDERATIONS

28.The value of outcome measures inautoimmune rheumatic diseases

M. Mosca, C. Baldini, S. Bombardieri,(Rheumatology Unit, Department of Internal Medicine,University of Pisa, Italy).Email: [email protected]

Autoimmune diseases are very complex conditions char-acterized by a variable clinical picture, both betweenpatients as well as in the same patient, by a very largerange of clinical and serological manifestations, by arelapsing-remitting course, and whose clinical manifes-tations can be attributed to active disease as well as toirreversible damage induced by the disease itself or bytherapy. Furthermore these conditions may greatly affectpatients quality of life.These aspects have made necessary the definition of uni-versally accepted measurement items to assess patientsand to improve scientific communication in observation-al trials but more importantly to assess the effectivenessof new drugs in randomised clinical trials.In this view in 1992 the ‘Outcome Measures in Rheu-matology’ (OMERACT) was established with the pur-pose to select outcome measures for clinical trials invarious autoimmune conditions such as rheumatoidarthritis, systemic lupus erythematosus, ankylosingspondylitis, osteoporosis, osteoarthritis and many others.Along with the need for sensitive and specific classifi-cative criteria for the definition of the different condi-tions, discussion emphasized the need to include variousinstruments to assess: disease activity, disease damage,health related quality of life, and drugs toxicities.In Table 1 we report the results so far obtained in somediseases.

Classificative Activity Damage Quality Respondercriteria of life index

Rheumatoid q q q q qarthritis

Systemic lupus q q q q yerythematosus

Systemic sclerosisq q y q ySjogren Syndromeq y y q y

Rheumatoid arthritis represents the better defineddisease with the definition of a set of disease activ-

ity measures for assessing outcome. In 1995 a def-inition of improvement in RA trials was proposedas 20% improvement in tender and swollen jointcounts and 20% improvement in 3 of 5 remainingACR core set measures: patient and physicianglobal assessments, pain, disability and an acute-phase reactant. The responder index, the ACR 20,is now broadly used both in routine clinical prac-tice as well as in randomized trials.Important results have also been achieved in sys-temic lupus erythematosus, and disease activityindices have been developed and validated, as wellas a damage index, furthermore the SF36 is nowwidely used to assess patients quality of life in thisdisease. However, a responder index is still inunder definition, in fact recently the ACR hasorganized a consensus building process to developa priori response criteria for SLE Activity Meas-ures and preliminary results will be soon available.Taking into consideration the good results obtainedin RA, certainly the development of a ResponderIndex in many conditions represent an unvaluablegoal, since such index could represent a furtherimprovement for scientific communication and forthe performance of trials. Indeed many of the con-cepts used in the assessment of SLE and RApatients, are now used for the definition of out-come measures on other diseases such as systemicsclerosis, Sjogren’s syndrome and idiopathicïnflammatory myopathies.

References

1. Felson D.T., Anderson J.J., Boers M., Bombardier C., Furst D.,

Goldsmith C., Katz L.M., Lightfoot R. Jr., Paulus H., Strand V.,

et al. American College of Rheumatology. Preliminary defini-

tion of improvement in rheumatoid arthritis. Arthritis Rheum.

1995;38:727–35.

2. Gladman D., Ginzler E., Goldsmith C., et al. The development

and initial validation of the Systemic Lupus International Col-

laborating ClinicsyAmerican College of Rheumatology damage

index for systemic lupus erythematosus. Arthritis Rheum.

1996;39:363–9.

3. Strand V., Gladman D., Isenberg D., Petri M., Smolen J., Tug-

well P. Endpoints: consensus recommendations from OMER-


ACT IV. Outcome Measures in Rheumatology. Lupus.

2000;9:322–7.

4. Strand V., Gladman D., Isenberg D., Petri M., Smolen J., Tug-

well P. Outcome measures to be used in clinical trials in sys-

temic lupus erythematosus. J. Rheumatol. 1999;26:490–7.

5. Ward M.M., Marx A.S., Barry N.N. Comparison of the validity

and sensitivity to change of 5 activity indices in systemic lupus

erythematosus. J. Rheumatol. 2000;27:664–70.

6. Isenberg D.A., Allen E., Farewell V. et al. International Myositis

and Clinical Studies Group(IMACS). International consensus

outcome measures for patients with idiopathic inflammatory

myopathies. Development and initial validation of myositis

activity and damage indices in patients with adult onset disease.

Rheumatology(Oxford). 2004;43:49–54.

7. Bowman S.J. Collaborative research into outcome measures in

Sjogren’s syndrome. Update on disease assessment. Scand. J.

Rheumatol. 2002;(116):23–7.

29.The value of meta-analysis in rheumatologyresearch

John P.A. Ioannidis,(Department of Hygiene and Epidemiology, Universityof Ioannina School of Medicine, Ioannina, 45110,Greece and Institute for Clinical Research and HealthPolicy Studies, Tufts-New England Medical Center,Tufts University School of Medicine, Boston, USA).Email: [email protected]

Meta-analysis, the quantitative synthesis of data fromdiverse studies on the same question, has been increas-ingly used across medical disciplines, including rheu-matology. An evaluation of 34 meta-analyses publishedon rheumatology-related topics in 2002–2003 showsthat the majority of them still pertain to randomized con-trolled trials, but meta-analyses of non-randomized evi-dence are also conducted. About two-thirds of themeta-analyses pertain to rheumatoid arthritis and oste-oarthritis, but the method has also been applied to sev-eral other rheumatologic diseases. Two-thirds pertain totherapeutic interventions(split almost equally betweendrugs and other therapies), but understanding geneticand other epidemiological associations is also anincreasingly common objective. Most meta-analyses to-date have been based on data from the published liter-ature. Nevertheless, there are already examples availableof meta-analyses based on collaborative consortia andon projects involving individual-level data. Meta-anal-

ysis provides a prime opportunity for improving thepower to answer important research questions and toidentify and possibly explain bias and genuine hetero-geneity in the results of studies addressing the samequestion. Moreover, a comprehensive perspectivetowards research may help improve future studies, andmay lead to the better appreciation of the global natureof the research endeavor in rheumatology, as well as inother related fields.Meta-analysis is a set of quantitative, systematic meth-ods that synthesize information from diverse studies onthe same question. Meta-analysis aims to provide sum-mary estimates, to quantify the extent of diversity(het-erogeneity) between the results of various studies, toidentify bias in the results of each study as well as inthe overall synthesis of all studies, and, to identify rea-sons for the presence of heterogeneity and biasw1x. Therelevant applications are continuously expanding andmeta-analysis is already accepted as the highest level inthe hierarchy of medical evidencew2x. Rheumatologyhas seen a rapid expansion of the meta-analytic appli-cations during the last few years.Table 1 describes the characteristics of 34 meta-analyseson rheumatology-related topics published in 2002–2003.The Cochrane Database of Systematic Reviews, a prod-uct of the international Cochrane Collaborationw3x, is aprime source of meta-analyses across medical disci-plines. Several meta-analyses have also appeared in allmajor rheumatology journals as well as in general med-ical journals. Most of the evaluated meta-analyses tar-geted randomized controlled trials, since this type ofstudy has been traditionally been considered the mostreliable design for addressing the effectiveness of med-ical interventions. However, there is an increasing lit-erature also on meta-analyses of non-randomized dataand a few projects that combine both randomized andnon-randomized evidence.Approximately two-thirds of the meta-analyses in thefield pertain to rheumatoid arthritis and osteoarthritis,but a wide variety of other rheumatologic diseases hasalso been targeted, including many that are not partic-ularly common, e.g. systemic lupus erythematosus andlupus nephritis or antiphospholipid syndrome. A meta-analytic approach is very attractive in uncommon, butimportant, autoimmune diseases, since single studies areunlikely to have sufficient sample size to answer anyresearch question with certainty.Two-thirds of the rheumatology-related meta-analysesaddress therapeutic interventions. These include mostlymedications, but also psychological, surgical, manage-ment, exercise, and other therapeutic options. Meta-anal-ysis has been applied also successfully in the field to


decipher genetic associations and linkage and other epi-demiological associations. Meta-analyses of prognosisand prevalence are less common. Studies of diseaseprevalence and prognostic factors often tend to give con-troversial results and use very divergent approaches anddefinitions that need some formal reconciliation. Anoth-er important application that was not represented in thisanalyzed sample is meta-analysis of diagnostic test per-formancew4x. It is important to understand the sensitiv-ity and specificity of new diagnostic markers for variousdiseases and subgroups thereof. Most autoimmune dis-eases have convoluted diagnostic criteria. The diagnosticliterature is furthermore hampered by small sample sizesthat leave wide uncertainty about the diagnostic per-formance of a proposed test or marker. For example, weare currently conducting a meta-analysis of the diagnos-tic performance of anti-riboprotein P antibodies for dis-criminating patients at risk for specific neuropsychiatricmanifestations of systemic lupus erythematosus; 14teams of investigators worldwide have agreed to providedata on their patients and other interested investigatorsare welcome to join the project.Most meta-analyses use group-level data that havealready been published in the literature. An importantlimitation is that the quality of the data may be sub-optimal w5x. Nevertheless, a meta-analysis offers a primeopportunity to study rigorously the quality of each rel-evant study and to examine whether the results correlatewith any important quality parameters across investiga-tionsw6x. Second, not all studies on a question of interestmay be published. Small studies with non-significantresults may remain unpublished or may be publishedlater than studies of similar size and quality that findstatistically significant results. This could lead to pub-lication bias and time lag biasw7x. Nevertheless, diag-nostics have been developed that provide hints onwhether such biases are operating. Occasionally the ear-ly results about the efficacy of a new treatment or theimportance of a research finding may be exaggeratedand appropriate caution is required until promising find-ings are replicated consistently.Meta-analyses may also be conducted as part of largecollaborative efforts involving all investigators workingon a specific field. Such collaborative meta-analysesmay also use detailed data from individual patients.They may also employ more standardized definitions forthe collection of data across participating studies. Whilesuch approaches are more demanding in time and mon-ey, they are worth pursuing, when resources can be met.For example, in a recent meta-analysis of individual-level dataw8x, we collected data on the outcome of 3311patients with systemic sclerosis followed up for almost

20 000 person-years, using consistent definitions fororgan system manifestations and auto-antibodies across8 participating study teams. Group data from publishedstudies would be too incompatible to combine for thepurpose of understanding prognostic factors for systemicsclerosis.Finally, meta-analysis is usually performed with datathat are already available, but one may also collect pro-spective data form diverse teams. Prospective meta-anal-ysis resembles a multicenter clinical study. However, itallows greater flexibility in that the background studydesign of each participating team is retained, and theanalytical approach respects the potential differencesacross study teams rather than simply summing up num-bers as in a multicenter study. For example, we recentlyset up GENOMOS, a consortium of European univer-sities that has been genotyping in a standardized fashiona large number of gene polymorphisms on approximate-ly 20 000 subjects enrolled across 8 major Europeancohort studies with data on osteoporosis, fractures andother bone diseases. Collaborative meta-analyses havealso been performed for the genetics of systemic lupuserythematosus and antiphospholipid syndromew9, 10x.Still there is much room for further growth of theseapproaches and for the set up of prospective meta-anal-yses in this field.Meta-analysis provides a global perspective on researchquestions. The emphasis is on seeing the available evi-dence in its totality and to understand for each researchquestion whether the available data are conclusive, het-erogeneous, or insufficient. Thus, meta-analysis mayalso help determine the need for conducting furtherresearch and can specify what kind of research is war-ranted and how extensive data are needed for futureinvestigations. Finally, fostering a collaborative spiritcan improve the quality and efficiency of researchefforts worldwide.

References

1. Lau J., Ioannidis J.P., Schmid C.H. Summing up evidence: oneanswer is not always enough. Lancet 1998;351:123–7.

2. Olkin I. Meta-analysis: reconciling the results of independentstudies. Stat. Med. 1995;14:457–72.

3. Clarke M., Langhorne P. Revisiting the Cochrane Collabora-tion. Meeting the challenge of Archie Cochrane--and facingup to some new ones. BMJ 2001;323:821.

4. Irwig L., Macaskill P., Glasziou P., et al. Meta-analytic meth-ods for diagnostic test accuracy. J. Clin. Epidemiol.1995;48:119–30

5. Karassa F.B., Tatsioni A., Ioannidis J.P.A. Design, quality andbias in randomized controlled trials of systemic lupus erythe-matosus. J. Rheumatol. 2003;30:979–84.


6. Balk E.M., Bonis P.A., Moskowitz H., et al. Correlation ofquality measures with estimates of treatment effect in meta-analyses of randomized controlled trials. JAMA2002;287:2973–82.

7. Ioannidis J.P.A. Effect of the statistical significance of resultson the time to completion and publication of randomized effi-cacy trials. JAMA 1998;279:281–6.

8. Ioannidis J.P.A, Vlachoyiannopoulos P.G., Haidich A.-B., etal. Mortality in systemic sclerosis: an international meta-anal-ysis of individual patient data. Am. J. Med.(in press).

9. Karassa F.B., Trikalinos T.A., Ioannidis J.P.A. TheFcgammaRIIIA-F158 allele is a risk factor for the develop-ment of lupus nephritis: a meta-analysis. Kidney International2003;63:1475–82.

10. Karassa F.B., Bijl M., Davies K.A., et al. The role of theFcgammaRIIA polymorphism in the antiphospholipid syn-drome: an international meta-analysis. Arthritis Rheum.2003;48:1930–8.

Table. Formal meta-analyses on rheumatology-relatedtopics published in 2002–2003 in MEDLINE-indexed journals*

Characteristic N

Year2002 202003 14JournalArthritis and Rheumatism 6Cochrane Database of Systematic Reviews 6Journal of Rheumatology 4

Rheumatology 2Other(16 diverse journals) 16Type of studies includedRandomized trials(direct comparisons) 20Non-randomized studies 11Both 2Randomized trials(indirect comparisons) 1Type of disease(s)Rheumatoid arthritis 15Osteoarthritis 7Rheumatoid arthritis and osteoarthritis 1Systemic lupus erythematosus 3Juvenile arthritis 3Other 5Type of research questionTreatmentDrug 14Psychological 2Other 8Genetic risk factorsAssociation 4Linkage 1Other associations and risk factors 3Prevalence 1Prognosis 1

*MEDLINE searches used an array of terms for the mostcommon rheumatic diseases(exact search strategy availableupon request). The sample retained only those evaluationsthat had strict systematic inclusionyexclusion criteria forincluding investigations in the analysis and quantitativesynthesis had been used for the relevant data.


SJORGEN’S SYNDROME AUTOIMMUNE EPITHELITIS:¨

30.Do clinical studies add valuableinformation to a syndrome described 70years ago?

F.N. Skopouli,(Department of Dietetics and Nutrition Science, Haro-kopio University of Athens, Greece).Email: [email protected]

In 1930 Hennik Sjogren(1) clearly demonstrated withhis classical study, that the syndrome, which later wasnamed after him, is a systemic disease, since sicca man-ifestations from eyes were co-expressed, in the sameindividuals, with polyarthritis. Thirty years later thegroup of NIH headed by Bunim(2) presented thediverse clinical picture of the syndrome. Further moreTalal and Bunim(3) showed that Sjogren’s syndromecan evolve to lymphoid neoplasia. At the same timeShearn and collaborators(4) in the West Cost of USApresented in detail the renal tubular defects due to theextension of infiltrating activated lymphocytes in therenal inerstitium. Ten years later Moutsopoulos and hiscollaborators(5, 6) at NIH on the basis of clinical, sero-logic and genetic studies indicated that the syndromeshould be divided in primary when it appears alonewithout clinical or serologic manifestations of otherautoimmune disorders and secondary when it manifestsitself in patients with rheumatoid arthritis. At that timeMoutsopoulos described also that immune complexglomerulonephritis can develop in patients with PrimarySjogren’s syndrome(7).Our group in Greece, in the last twenty years, guidedand inspired by Professor Moutsopoulos developed aspecific interest on epidemiological, clinical and patho-phyciological studies of the syndrome.From our clinical studies in patients with primary Sjo-gren’s syndrome it was indicated that subclinical andclinical manifestations from the respiratory tract and theliver had histopathological background consisting ofround cell infiltrates mainly around bronchi and chol-langia respectively(8, 9). This pathological picture wasvery similar to that described previously in the patho-logical lesions of exocrine(lacrimal or salivary) glandsand the renal interstitium.The above clinical and pathological observations indi-cated that the main cell of the affected organs, regardingif there are exocrine glands or parenchymal organs is the

epithelial cell. This was further substantiated withimmunopathological and functional studies, whichrevealed that these cells appear intrinsically activatedand play a significant role in the initiation and perpet-uation of the autoimmune process(10). These obser-vations prompted us to rename primary Sjogren’ssyndrome as ‘autoimmune epithelitis’(11).Attendingand following-up carefully a large cohort of primarySjogren’s syndrome patients we identified that

1. The renal, liver or lung involvement in the sense ofepithelitis, despite the fact that are not frequent, arepresent at the time of diagnosis of the disorder togeth-er with the manifestations of the exocrine glands orsome times may precede the sicca findings. This pro-cess after many years of follow-up does not seem toprogress, remains rather stable and does not lead tofailure of the affected organs. In addition the auto-antibody response(rheumatoid factor and a-Ro anda-La antibodies) is also present from the beginningof the disease and does not change during follow-up(12).

2. A subgroup of primary Sjogren’s syndrome patients(about 20%) presents at the time of diagnosis a spe-cific immune system activation, manifesting withmixed monoclonal cryoglobulins, which probablyutilize complement and lead to low levels of C4 com-ponent. This immune system activation does not seemto correlate with the degree of extension of epithelitisin the parenchymal organs. These patients may pres-ent withyor develop later additional clinical manifes-tations of small vessel vasculitis of the skin, likepalpable purpura or cutaneous ulcers and may devel-op, after many years, glomerulonephritis(12, 13).Among these patients around 3% at five years and5% at ten years of disease duration, develop overtlymphoprolipherative disease(14). This lymphopro-liferative disorder has the characteristics of an antigendriven lymphoma, described as a marginal zone typelymphoma(15). Mortality rate is greater in this groupof patients(12, 14, 15).

In conclusion the majority of primary Sjogren’s syn-drome patients manifest a complete clinical and labor-atory picture at the time of diagnosis, which does notchange significantly during follow-up. Thus, it is notnecessary to follow this group of patients closely andrepeated testing for auto-antibodies appears not valuable.


These patients do not need aggressive therapeuticintervention.A subgroup of patients, which clinically present fromthe time of diagnosis with cutaneous small vessel vas-culitis andyor have low serum C4 levels need closemonitoring for the possibility of developing glomerulo-nephritis andyor lymphoma.This specific immune response, in this subgroup ofpatients, leads to an increased morbidity and mortalityand can be predicted many years in advance. Thus, it isworth it to develop collaborative efforts in order to dis-cover therapeutic interventions, which can be appliedearly in the disease course.These clinical observations change our notion on follow-ing and treating primary Sjogren’s syndrome patients aswell as on disease prognostic factors for increased mor-bidity and mortality.In addition, the above clinical observations improvedour notion on the physical history of the disease, andenhanced our concept of epithelitis as the main patho-physiologic process of the disease, which, in somepatients, presents with a more aggressive immuneresponse leading to lymphoprolipheration.

References

1. Sjogren H. Keratoconjunctivitis sicca. Hugiea 1930; 82:829.2. Bloch K.J., Buchanan W.W., Wohl M.J., Bunim J.J. Sjogren’s

syndrome: aclinical, pathological and serological study of six-ty-two cases. Medicine 1965; 44:187–231.

3. Talal N., Bunim J.J. The development of malignant lymphomain the course of Sjogren’s syndrome. Am. J. Med. 1964;36:529–540.

4. Shearn M.A., Tu W. Nephrogenic diabetes insibidus and otherdefects of renal tubular function in Sjogren’s syndrome. Am.J. Med. 1965; 39:312–318.

5. Moutsopoulos H.M., Webber B.L., Vlagopoulos T.P., ChusedT.M., Decker J.L. Differences in the clinical manifestations ofsicca syndrome in the presence and absence of rheumatoidarthritis. Am. J. Med. 1979; 66:733–736.

6. Moutsopoulos H.M., Mann D.L., Johnson A.H., Chused T.M.Genetic differences between primary and secondary sicca syn-drome. N. Engl. J. Med. 1979; 301:761–763.

7. Moutsopoulos H.M., Balow J.E., Cawley T.J., Stahl N.I.,Antonovych T.T., Chused T.M. Immune complex glomerulo-nephritis in sicca syndrome. Am. J. Med. 1978; 64:955–960.

8. Skopouli F.N., Barbatis C., Moutsopoulos H.M.. Liverinvolvement in primary Sjogren’s syndrome. Br. J. Rheum. 33:745–748, 1994.

9. Papiris S.A., Maniati M.A., Constantopoulos S.H., RoussosCh., Moutsopoulos H.M., Skopouli F.N. Lung involvement in

primary Sjogren’s syndrome is mainly related to small airwaydisease. Ann. Rheum. Dis. 58: 61–4, 1999.

10. Tzioufas A.G., Moutsopoulos H.M. Sjogren’s syndrome. In:Rheumatology Eds: Hochberg et al. 3d Edition 2003 p: 1431–1443.

11. Moutsopoulos H.M. Sjogren’s syndrome: autoimmune epi-thelitis. Clin. Immunol. Immunopathol. 1994; 72:161–165.

12. Skopouli F.N., Dafni U., Ioannidis J.P., Moutsopoulos H.M.Clinical evolution, and morbidity and mortality of primarySjogren’s syndrome. Semin. Arthritis. Rheum. 2000; 29: 296–304.

13. Goules A., Masouridi S., Tzioufas A.G., Ioannidis J.P., Sko-pouli F.N., Moutsopoulos HM. Clinically significant and biop-sy-documented renal involvement in primary Sjogrensyndrome. Medicine(Baltimore). 2000; 79: 241–9.

14. Ioannidis J.P.A., Vassiliou V.A., Moutsopoulos H.M. Longterm risk of mortality and lymphoprolipherative disease andpredictive classification of primary Sjogren’s syndrome.Arthritis Rheum. 2002;46: 741–747

15. Voulgarelis M., Dafni U.G., Isenberg D.A., MoutsopoulosH.M. Malignant lymphoma in primary Sjogren’s syndrome: amulticenter, retrospective, clinical study by the European Con-certed Action on Sjogren’s Syndrome. Arthritis Rheum. 1999;42: 1765–72.

31.The central role of epithelial cells inSjogren’s syndrome or autoimmuneëpithelitis

E. Kapsogeorgou and M.N. Manoussakis,(Department of Pathophysiology, Medical School,National University of Athens, 75 Mikras Asias str11527, Athens, Greece).E-mail: [email protected]

Autoimmune diseases are thought to result from theaberrant activation of the immune system in associationwith defects in the failsafe immunoregulatory mecha-nisms that normally maintain immune tolerance.Although the pathogenesis of these disorders is believedto involve the concerted action of genetic, hormonal andenvironmental factors and typical immunologicresponses, the precise mechanisms that operate remainunclear. Autoaggressive immune reactions most certain-ly involve typical presentation of autoantigen(s) to Tcells by suitably equipped ‘antigen-presenting’ cells(APC) and the subsequent cross-talk of T cells with Blymphocytes for the production of specific autoantibod-ies. The presentation of antigens in the context of MHCmolecules that are expressed in the surface of APC is


not sufficient for the effective activation of T lympho-cytes, but requires assessory signals provided by ‘nonantigen-specific’ receptors that mediate T cell costimu-lation, adhesion and expansion of immune responses.‘Ectopic’ expression of such ‘immunoregulatory’ mol-ecules by non-immune tissues is considered to have amajor role in both the regulation of tolerance, as well asthe development of inflammation and autoimmune reac-tions (1).During the recent years, we have been studying the roleof epithelial cells in the pathogenesis of Sjogren’s syn-¨drome (SS), aiming to better understand the involve-ment of epithelial tissues in the development ofautoimmune responses. SS is a chronic autoimmune dis-order that is characterized by the dysfunction anddestruction of exocrine glands, associated with T and Blymphocytic infiltrates(2). The widespread involvementof epithelial tissues in SS patients had lead to the ren-aming of the syndrome as autoimmune epithelitis(3).SS is relatively common(prevalence 0.5–1%) andaffects mainly middle-age women. It displays a broadspectrum of clinical manifestations that range fromorgan-specific exocrinopathy to systemic disorder andthe development of lymphoma and thus, it represents anexcellent model for the study of chronic organ-specificand systemic autoimmune diseases. Immunopathologicanalyses of the lymphoepithelial lesions in the affectedtissues had revealed the ‘‘activated’’ phenotype of theadjacent epithelial tissues, as this was illustrated by theaberrant expression of various activation proteins,including oncogenes, several immunoreactivity mole-cules and autoantigens(2).During the recent years we sought to address whetherepithelial cells are innocent bystanders or active playersthat participate in the initiation and the perpetuation ofSS lesions. For this purpose we had established a repro-ducible in-vitro system for the long-term cultivation andstudy of salivary gland epithelial cells(SGEC) obtainedfrom patients with SS and disease controls(4). The par-allel phenotypic analysis of salivary gland epithelialcells in biopsy specimens(in-situ) and in cultured epi-thelial cells(in-vitro) had revealed that SGECs derivedfrom SS patients aberrantly express molecules implicat-ed in lymphoid recruitment and activation. SGEC fromSS patients appear to be capable to attract immune cellsin the inflammatory lesion. These cells express variousmolecules involved in lymphoid cell homing and theamplification of epithelial-immune cells interactions,including proinflammatory cytokines lymphoattractantchemokines, apoptosis and adhesion molecules(5–8,11, 13). Importantly, SS-SGECs appear suitablyequipped for the presentation of antigenic peptides and

the transmittance of activation signals to T cells, as sug-gested by the expression of MHC class I and class IImolecules and functional B7 costimulatory molecules(8). B7 costimulatory molecules represent the mostextensively studied costimulatory pathway. B7 signalingis considered critical for the activation of naıve T cells.¨Furthermore, B7 molecules play major role in the reg-ulation of immune responses by the transmission of T-cell activation or inhibitory signals through theirinteraction with CD28 or CTLA4 receptors, respectively,on T cells (9). Notably, B7.2 molecules expressed bySGECs were found to display unique binding propertiesdenoted by the functional interaction with CD28 recep-tor and reduced binding to the negative-regulator ofimmune responses CTLA4(10). In addition, SGECsexpress functional CD40 molecules(11). The interac-tion of CD40 on APC with its receptor CD40-ligand(CD40L) is thought to be implicated in the expansionof inflammatory responses by the transmittance of directactivating signals to T cells or by the upregulation ofthe expression of the costimulatory and adhesion mole-cules on APC(12). These findings suggest a critical roleof epithelial cells in the regulation of local immuneresponses in salivary glands, possibly by acting as non-classical antigen-presenting cells.Cultured SGEC lines derived from SS patients displayelevated expression of all the abovementioned immune-rectivity molecules that is stable even after four monthsof cultivation. This fact suggests that the ‘‘activationphenotype’’ of SGEC that is observed in-situ in the sal-ivary gland tissues of SS patients is likely due to intrin-sic activation processes that operate in these epithelialcells, rather than to the influence of local micro-envi-ronmental factors(e.g. inflammatory cytokines). Fur-thermore, our data suggest that the inflammatory milieupresent in the lymphoepithelial lesions of SS patientsmay drive the injuries of epithelia by apoptotic celldeath, as illustrated by the susceptibility of SGEC toFas-mediated apoptosis following treatment with the T-cell derived cytokine interferon-g (13). The apoptoticepithelial cell death possibly results to the release ofautoantigens that are captured by APC and presented tothe immune cells, resulting thus to the perpetuation andthe expansion of SS immune responses.

References

1. Abbas A.K., Lightman A.H., Pober J.S. 1994. Self-toleranceand autoimmunity. Cellular and molecular immunology(2ndedition) W.B. Saunders Company; Philadelphia p. 377.


2. Manoussakis M.N., Moutsopoulos H.M. 2000. Sjogren’s syn-¨

drome: autoimmune epithelitis. Bailliere’s Clinical Rheuma-tology 14:73.

3. Moutsopoulos H.M. Sjogren’s syndrome: Autoimmune epi-¨

thelitis. Clin Immunol Immunopathol. 1994; 72:162.4. Dimitriou I.D., Kapsogeorgou E.K., Moutsopoulos H.M.,

Manoussakis M.N. Establishment of a convenient system forthe long-term culture and study of non-neoplastic human sal-ivary gland epithelial cells. Eur. J. Oral. Sci. 2002; 110:21–30.

5. Xanthou G., Polihronis M., Tzioufas A.G., Paikos S., SiderasP., Moutsopoulos H.M. ‘Lymphoid’ chemokine messengerRNA expression by epithelial cells in the chronic inflamma-tory lesion of the salivary glands of Sjogren’s syndromepatients: possible participation in lymphoid structure forma-tion. Arthritis Rheum. 2001; 44:408–18.

6. Skopouli F.N., Moutsopoulos H.M. Cytokines in Sjogren’ssyndrome. Ann. Med. Interne. 1995; 146:219–22.

7. Kapsogeorgou E.K., Dimitriou I.D., Abu Helu R.F., Moutso-poulos H.M., Manoussakis M.N. Activation of epithelial andmyoepithelial cells in the salivary glands of patients with Sjo-gren’s syndrome: high expression of intercellular adhesionmolecule-1(ICAM.1) in biopsy specimens and cultured cells.Clin. Exp. Immunol. 2001; 124:126–33.

8. Manoussakis M.N., Dimitriou I.D., Kapsogeorgou E.K., Xan-thou G., Paikos S., Polichronis M., Moutsopoulos H.M.Expression of B7 costimulatory molecules by salivary glandepithelial cells in patients with Sjogren’s syndrome. Arthritis¨

Rheum., 1999; 42:229–39.9. Carreno B.M., Collins M. The B7 family of ligands and its

receptors: new pathways for costimulation and inhibition ofimmune responses. Annu. Rev. Immunol., 2002; 20:29–53.

10. Kapsogeorgou E.K., Moutsopoulos H.M., Manoussakis M.N.Functional expression of a costimulatory B7.2(CD86) proteinon human salivary gland epithelial cells that interacts withCD28 receptor, but has reduced binding to CTLA4. J. Immu-nol., 2001; 166: 3107–13.

11. Dimitriou I.D., Kapsogeorgou E.K., Moutsopoulos H.M.,Manoussakis M.N. CD40 on salivary gland epithelial cells:high constitutive expression by cultured cells from Sjogren’ssyndrome patients indicating their intrinsic activation. Clin.Exp. Immunol., 2002; 127:386–92.

12. Grewal I.S., Flavell R.A. CD40 and CD154 in cell-mediatedimmunity. Annu. Rev. Immunol. 1998; 16:111–35.

13. Abu Helu R.F., Dimitriou I.D., Kapsogeorgou E.K., Moutso-poulos H.M., Manoussakis M.N. Induction of salivary glandepithelial cell injury in Sjogren’s syndrome: in vitro assess-ment of T cell-derived cytokines and Fas protein expression.J. Autoimmun. 2001; 17:141–53.

32.Origin, antigenic specificity and regulationof autoantibodies targeting RoyLa RNP

J.G. Routsias and A.G. Tzioufas,(Department of Pathophysiology, School of Medicine,National University of Athens, 11527, Athens, Greece).Email: [email protected]

Antibodies recognizing the RoyLa RNP particle arecommonly found in a high proportion of sera frompatients with systemic lupus erythematosus or Sjogren’ssyndrome. Although, the mechanism by which theseautoantibodies arise is not known, their autoantigenictargets have been studied extensively. The RoyLa ribo-nucleoprotein complex(RNP) is formed by the noncov-alent association of La and Ro60 autoantigens with asmall cytoplasmic RNA(hYRNA) w1, 2x. Ro52 autoan-tigen is also transiently associated with RoyLa RNPw2x.Additional components of the complex have beenrecently identified as the proteins calreticulinw3x andnucleolin w4x.Epitope mapping with synthetic peptides, in our labor-atory, revealed the precise antigenic regions of Ro60kDin 169–190 and 211–232 parts of the antigenw5x. Oneof them, the 169–190 epitope, was found to share con-formational and antigenic similarity with HLADR3b-chain. The homologous regions in these two proteins(HLA-DR3 b-chain and Ro60KD) were found to sharesimilar molecular conformation(as defined by circulardichroism and molecular modeling), as well as commonantigenic featuresw6x. This finding is particularly inter-esting since the autoimmune response directed towardsRoyssa and Layssb autoantigen is highly associated withthis particular HLA class II alloantigen. Thus, autoan-tibodies reacting with such exposed regions of the majorhistocompatibility complex(MHC)-II are potentiallycapable to activate B cells or macrophages throughdimerization and cross-linking of these molecules.Although Ro60kd epitopes were identified as smallpeptidic moieties(22 aa in length) with rather limitedreactivity against patient sera, their recognition by auto-antibodies is conformation-dependent and their antige-nicity is dramatically enhanced upon interaction with themolecular chaperone calreticulinw7x. Using complexesof highly purified human calreticulin with the linear epi-topes of Ro60Kd, it was found that almost all positiveanti-Ro60Kd sera bound strongly onto the newly formedconformation of the epitopesw7x. When calreticulin orthe linear epitopes of Ro60Kd were tested individuallywith the same sera, the prevalence of positive reactionswas much lower. In addition, sera from pSS or SLE


patients without anti-RoySSA antibodies did not reactwith the calreticulin–linear epitope complexes ofRo60Kd w7x. These observations suggest conformation-dependent enhancement of antigenicity of the Ro60Kdepitopes upon interaction with the chaperone protein cal-reticulin and such kind of complexes can potentially beused as substrates for the efficient detection ofautoantibodies.Recent studies in our laboratory have been also focusedon the zinc finger motif of Ro60Kd protein. The zincfingers are secondary structure elements, responsible forprotein–DNA and protein–protein interactions. Theycan also hold putative conformational B-cell epitopes,since their structure is affected by zinc binding andredox conditions. Using synthetic peptide analogues cor-responding to(i) to the zinc finger motif of Ro60Kd,spanning the region 301–327aa(Zif-1), (ii) a truncatedform of the zinc finger motif, without the intermediateloop (310–319aa) of the molecule(Zif-2), and(iii ) theintermediate loop of the zinc finger motif(Zif-3). It wasfound that the majority of anti-RoySSA and LaySSBpositive sera from patients with pSS bound in the full-length peptide, in the absence of zinc ions. In contrast,the native form of the zinc finger domain, in the pres-ence of zinc ions, could bind to Ro52Kd, but not toautoantibodiesw8x. Thus, different conformations of thezinc finger domain of Ro60kD, were employed in inter-action with Ro52kD polypeptide and pSS autoanti-bodies.B-cell epitope mapping of LaySSB was performed alsoin our laboratory using 20-mer synthetic peptides over-lapping by eight amino acids covering the wholesequence of the protein. Peptides highly antigenicwere those spanning the sequences: HKAFKGSI147 154

(located within RRM motif: 113 – 182aa), NGNL-291

QLRNKEVT , VTWEVLEGEVEKEALKKI and302 301 318

GSGKGKVQFQGKKTKF w9x. The peptide-based349 364

ELISA assays, with the above described epitopes, pre-sented sensitivities ranging from 78 to 90% and speci-ficities from 69 to 94%. The most sensitive and specificpeptide 349GSGKGKVQFQGKKTKF364()90% sen-sitivity and specificity) was synthesized in attachmentwith a tetramer sequential oligopeptide carrier SOC4 andused for immunoassay development. Ninety percent ofanti-La positive sera were reactive with both the syn-thetic peptide 349–364aa and the recombinant La pro-tein w10x. Thus, this epitope analogue exhibitedcomparable with the recombinant LaySSB value for thedetection of anti-LaySSB antibodies. Clinical aspects ofantibodies to linear B-cell epitopes of LaySSB in pSSwere also studied by our groupw11x. It was found thatautoantibodies to the LaySSB epitope, p349–364aa,

were significantly positively associated with longer dis-ease duration, recurrent or permanent parotid glandenlargement, and a higher proportion of non-exocrinemanifestations, compared to patients without autoanti-bodiesw11x.Anti-idiotypic antibodies, reactive with idiotypes ofautoantibodies, are capable of regulating the autoim-mune responsew12x. The same antibodies may alsointerfere in autoantibody detection by competing withantigen for binding in the same paratopic site(antigeninhibitable or Ab2h anti-idiotypic antibodies, accordingto Jerne’s classification). In order to derive peptidescapable of neutralizing anti-idiotypic antibodies, wehave taken advantage of the antisenseycomplementarypeptide approachw13x. This approach is based on themolecular recognition theory. According to this theory,the translation of two complementary mRNAs producesa pair of peptides with inverted hydrophobicity profilesthat leads, under certain conditions, to strong interactionbetween these two(sense and antisense) peptidesw14x.Interestingly, these peptides have the ability to generateinteracting pairs of idiotypic and anti-idiotypic antibod-ies upon their application in animal immunizationsw15x.In this regard, we prepared complementary peptides cor-responding to major epitopes of LaySSB (289–308aaand 349–364aa) w13x. These peptides reacted with a sig-nificant proportion of patient sera with anti-La specific-ity. From these patients sera, anti-complementaryepitope and anti-epitope antibodies were purified anddigested with pepsin in order to produce F(ab) frag-2

ments. The antibodies against epitopes found to specif-ically interact with the F(ab) fragments of antibodies2

recognizing complementary epitopes and vice versa,suggesting their idiotype–anti-idiotype relation. Inhibi-tion experiments demonstrated that anti-idiotypic anti-bodies compete with the antigen for the binding site(paratope) of antibodies against LaySSB epitopes. Itwas also found that immunizations with either pep orcpep led to the appearance of antibodies against theimmunogen peptide by day 31 which subsequently wasfollowed by antibody production to its complementarypeptide by day 55. Using the complementary epitopesas inhibitors of the anti-idiotypic antibodies, we wereable to recover the hidden anti-LaySSB reactivity inpatient seraw13x. This methodology was applied in 44anti-La(y), anti-RoyANA (q) sera from patients withSLE and Sjogren’s syndrome. Ninety-four percent ofSjogren’s syndrome sera and 80% of SLE sera werefound negative for anti-pep 349–364 antibodies in ELI-SA prior to the treatment. After the heatqcomplemen-tary epitope treatment, all SS and SLE sera becamepositive for anti-epitope 349–364 antibodies, while none


of the normal sera exhibited a positive reaction. Thus,virtually all anti-RoyANA (q) sera possess also hiddenanti-LaySSB antibodies that can be unmasked by treat-ment with the complementary epitopew13x.

Acknowledgment

This work was supported by a grant from Greek Sec-retariat of Research and technology PEND No.O1ED164.

References

1. Ben Chetrit E., Chan E.K., Sullivan K.F., Tan E.M. A 52-kDprotein is a novel component of the SS-AyRo antigenic par-ticle. J. Exp. Med. 1988; 167:1560–71.

2. Slobbe R.L., Pluk W., van Venrooij W.J., Pruijn G.J. Ro ribo-nucleoprotein assembly in vitro. Identification of RNA-proteinand protein-protein interactions. J. Mol. Biol. 1992; 227:361–6.

3. Cheng S.T., Nguyen T.Q., Yang Y.S., Capra J.D., SontheimerR.D. Calreticulin binds hYRNA and the 52-kDa polypeptidecomponent of the RoySS-A ribonucleoprotein autoantigen. J.Immunol. 1996; 156:4484–91.

4. Fouraux M.A., Bouvet P., Verkaart S., van Venrooij W.J.,Pruijn G.J. Nucleolin associates with a subset of the humanRo ribonucleoprotein complexes. J. Mol. Biol. 2002; 320:475–88.

5. Routsias J.G., Tzioufas A.G., Sakarellos-Daitsiotis M., Saka-rellos C., Moutsopoulos H.M. Epitope mapping of the RoySSA60KD autoantigen reveals disease-specific antibody-binding profiles. Eur. J. Clin. Invest. 1996; 26:514–21.

6. Routsias J.G., Sakarellos-Daitsiotis M., Tsikaris V., SakarellosC., Moutsopoulos H.M., Tzioufas A.G. Structural, molecularand immunological properties of linear B-cell epitopes ofRo60KD autoantigen. Scand. J. Immunol. 1998; 47:280–7.

7. Staikou E., Routsias J.G., Makri A.et al. Calreticulin bindspreferentially with B-cell linear epitopes of Ro60kD autoan-tigen, enhancing the recognition by anti-Ro60kD autoantibod-ies. Clin. Exp. Immunol. 2003;in press.

8. Routsias J.G., Makri A., Sakarellos C.et al. Zinc fingerdomain of Ro60kD autoantigen is essential for binding ofRo52kD and autoantibodies. Arthritis Res. Ther. 2003; 5(Suppl 1):22.

9. Tzioufas A.G., Yiannaki E., Sakarellos-Daitsiotis M., RoutsiasJ.G., Sakarellos C., Moutsopoulos H.M. Fine specificity ofautoantibodies to LaySSB: epitope mapping, and characteri-zation. Clin. Exp. Immunol. 1997; 108:191–8.

10. Yiannaki E.E., Tzioufas A.G., Bachmann M.et al. The valueof synthetic linear epitope analogues of LaySSB for the detec-tion of autoantibodies to LaySSB; specificity, sensitivity andcomparison of methods. Clin. Exp. Immunol. 1998; 112:152–8.

11. Tzioufas A.G., Wassmuth R., Dafni U.G.et al. Clinical, immu-nological, and immunogenetic aspects of autoantibody pro-duction against RoySSA, LaySSB and their linear epitopes inprimary Sjogren’s syndrome(pSS): a European multicentrestudy. Ann. Rheum. Dis. 2002; 61:398–404.

12. Kohler H., Kaveri S., Kieber-Emmons T., Morrow W.J., MullerS., Raychaudhuri S. Idiotypic networks and nature of molec-ular mimicry: an overview. Methods Enzymol. 1989; 178:3–35.

13. Routsias J.G., Touloupi E., Dotsika E.et al. Unmasking theanti-LaySSB response in sera from patients with Sjogren’ssyndrome by specific blocking of anti-idiotypic antibodies toLaySSB antigenic determinants. Mol. Med. 2002; 8:293–305.

14. Blalock J.E., Bost K.L. Binding of peptides that are specifiedby complementary RNAs. Biochem. J. 1986; 234:679–83.

15. Routsias J.G., Dotsika E., Touloupi E.et al. Idiotype-anti-idi-otype circuit in non-autoimmune mice after immunization withthe epitope and complementary epitope 289–308aa of LaySSB: implications for the maintenance and perpetuation of theanti-LaySSB response. J. Autoimmun. 2003; 21:17–26.

33.B cell monoclonal proliferation in Sjogren’ssyndrome

Michalis Voulgarelis,(Department of Pathophysiology, Medical School,National University of Athens, 75 Mikras Asias str,11527, Athens, Greece).Email: [email protected]

Over the past years numerous studies in humans linkedseveral different autoimmune diseases with malignantlymphoproliferation. Adequate studies establish strongassociations between B cell lymphomas and Sjogren’ssyndrome(SS) (1), and autoimmune thyroiditis(2).There are weaker associations between B cell lympho-mas and systemic lupus erythematosus and rheumatoidarthritis.Among all autoimmune diseases, SS(autoimmune epi-thelitis) (3) best illustrates the autoimmunity-lympho-proliferation-lymphoma sequence. The SS-associatedlymphoproliferation ranges from an increased frequencyof mixed monoclonal cryoglobulinemia, increased levelsof circulating CD5-positive B cells, circulating mono-clonal immunoglobulin, to an increased frequency ofmalignant non Hodgkin’s lymphomas(NHL) (4–6). Anotable histological feature in a parotid gland is the lym-phoid follicle-like structures with germinal centers thatsimulate the architecture of peripheral lymphoid nodeswhere B-lymphocytes are often oligoclonal(7) with a


risk of progression to B cell lymphoma. The transitionof reactive lymphoepithelial sialadenitis(LESA) frombenign monoclonality to monoclonal lymphoma is gen-erally considered to represent a multistep process. Thepresumed chronic antigen stimuli in SS, virus or autoan-tigen (the rheumatoid factor autoantigen) is uncertain(8), however the limited repertoire of VH gene segmentsin SS-associated lymphomas, implicates that LESA-associated clones may bind the same or similar antigensand are selected for clonal expansion(9). In addition,oncogenic events like inactivation of tumor suppressiongenes andyor activation of proto-ongogenes(10, 11) andsomatic hypermutation in germinal centers affectinggenes such as Fas receptors are usually required tomalignant transformation.Our department reported a study(11) in which p53 andp21 protein expression as well as DNA sequence anal-ysis of the p53 gene were performed on tissue samplesobtained from patients with SS in order to examine thepossible involvement of these cell cycle check pointgenes in the pathophysiology of SS. Immunohistochem-istry and Western blot analysis were used to study p53and p21 protein expression in minor salivary gland(MSG) biopsies from seven patients with SS and fivecontrols. In addition, sequence analysis of the p53 genewas performed on DNA samples obtained from MSGbiopsies from the same SS patients and from fourpatients with SS and NHL. This study revealed increasedprotein expression of p53 and p21 in MSG biopsies fromSS patients compared with controls, while sequenceanalysis showed that the p53 gene was wild type. Fur-thermore, sequence analysis of the p53 gene frompatients with SS and NHL revealed two novel mutationsin exon 5 of the p53 gene. Three patients had a G-to-Tsubstitution of codon 155(TGGyTGG), while onepatient had a G-to-C substitution at codon 127(GGAyGGA). The novel mutations of the p53 gene implicatethe deregulation of this tumor suppressor gene as a pos-sible mechanism of lymphoma development in SSpatients.Several investigators have attempted to establish predic-tive clinical and serological factors(palpable purpura,low C4 and mixed monoclonal cryoglobulinemia) forlymphoma development. Patients with these factors con-stitute a separate subgroup that should be monitored andmanaged closely than the other SS patients(22).The prevalence of NHL in SS patients is 4.3% and usu-ally develops later during the illness. The majority oflymphomas in SS are low-grade lymphoma of mucosa-associated lymphoid tissue(MALT ) type. A study con-structed by our department(11) and the members of theEuropean Concerted Action for SS presented a retro-

spective study of the clinical course and evolution ofmalignant NHL in 33 patients with SS followed in nineEuropean Medical Centers. The NHLs were primarilysituated in the marginal zone(48.5%). Manifestationswere mostly extranodal(78.8%) and most often identi-fied in the salivary glands(54.6%). Lymphadenopathy(65.6%), skin vasculitis (33.3%), peripheral nerveinvolvement(24.2%), low-grade fever(25.0%), anemia(48.1%), lymphopenia(78.6%), and cryoglobulinemia(50%) were observed significantly more frequently thanin the general SS population. Patients with high andintermediate grade malignancy have significantly worsesurvival (Ps0.041). Presence of B-symptoms and alarge tumor diameter()7 cm) are additional independ-ent risk factors for death. The novel observations of thisstudy are the type of the developing NHL, the overallsurvival of these patients and the role of skin vasculitis,peripheral nerve involvement, anemia, and lymphopeniaas important clinical predictors for the lymphoma devel-opment. Some previously reported results on extranodalmanifestations were also confirmed.In conclusion, SS as a clinical paradigm of autoimmun-ity-associated lymphoma is the best tool to dissect themultiple components of the autoimmunity and lym-phomatogenesis with the goal to our understanding bothof the processes.

References

1. Kassan S.S., Thomas T.L., Moutsopoulos H.M., Hoover R.,Kimberly R.P., Budman D.R., et al. Increased risk of lympho-ma in sicca syndrome. Ann. Inter. Med. 1978; 89:888–92.

2. Ansell S.M., Grant C.S., Habermann T.M. Primary thyroidlymphoma. Semin. Oncol. 1999;26:316–23.

3. Tapinos N.I., Polihronis M., Tzioufas A.G., MoutsopoulosH.M. Sjogren’s syndrome. Autoimmune epithelitis. Adv. Exp.Med. Biol. 1999;455:127–34.

4. Tzioufas A.G., Boumba D.S., Skopouli F.N., MoutsopoulosH.M. Mixed monoclonal cryoglobulinemia and monoclonalrheumatoid factor cross-reactive idiotypes as predictive factorsfor the development of lymphoma in primary Sjogren’s syn-drome. Arthritis Rheum. 1996;39:767–72.

5. Royer B., Cazals-Hatem D., Sibilia J., Agbalika F., CayuelaJ.-M., Soussi T., Maloisel F., Claurel J.P., Brouet J.C., Mar-riette X. Lymphomas in patients with Sjogren’s syndrome aremarginal zone B-cell neoplasms, arise in diverse extranodaland nolal sites and are not associated with viruses. Blood1997;90:766–75.

6. Voulgarelis M., Dafni U.G., Isenberg D.A., MoutsopoulosH.M. Malignant lymphoma in primary Sjogren’s syndrome.Arthritis Rheum. 1999;42:1765–72.


7. Moutsopoulos H.M., Tzioufas A.G., Bai M.K., PapadopoulosN.M., Papadimitriou C.S. Association of serum IgMk mono-clonicity in patients with Sjogren’s syndrome with anincreased proportion ofk positive plasma cells infiltrating thelabial minor salivary glands. Ann. Rheum. Disease1990;49:929–31.

8. Martin T., Weber J.C., Levallois H., Labouret N., Soley A.,Koenig S., et al. Salivary gland lymphomas in patients withSjogren’s syndrome may frequent develop from rheumatoidfactors B cells. Arthritis Rheum. 2000;43:908–16.

9. Miklos J.A., Swerdlow S.H., Bahler D.W. Salivary glandmucosa-associated lymphoid tissue lymphoma immunoglobu-lin VH genes show frequent use of V1-69 with distinctiveCDR3 features. Blood 2000;95:3878–3884.

10. Pisa E.K., Pisa P., Kang H.I., Fox R.I. High frequency oft(14;18) translocation in salivary gland lymphomas from Sjo-gren’s syndrome patients. J. Exp. Med. 1991; 174:1245–50.

11. Tapinos N.I., Polihronis M., Moutsopoulos H.M. Lymphomadevelopment in Sjogren’s syndrome. Novel p53 mutations.Arthritis Rheum. 1999;42:1466–72.

12. Skopouli F.N., Dafni U., Ioannidis J.P.A., Moutsopoulos H.M.Clinical evolution, and morbidity and mortality of primarySjogren’s syndrome. Semin. Arthritis Rheum. 2000;29:296–304.

34.Lymphoproliferative disorders in patientswith Sjogren’s syndrome¨

Susumu Sugai, Yasufumi Masaki and Lingli Dong,(Hematology & Immunology, Internal Medicine, Kana-zawa Medical University.)Email: [email protected]

SS is a chronic organ-specific autoimmune disease char-acterized by lymphocytic infiltration into the salivaryand lacrimal glands, resulting in keratoconjunctivitis sic-ca and xerostomia(1). About half of SS patients devel-op systemic disorders, and about 5% develop malignantlymphomas (2). During disease progression, manyorgans develop lesions with lymphocytic infiltration. Wereport here the associations between SS and lymphopro-liferative disorders.

1. Three stages of SSA long-term (over 10 year) follow up study of 31patients with primary SS showed that they could bedivided into two groups. One(15 patients, or 48%)experienced no changes in symptoms or laboratory dataduring the course of the disease, while the other.(16patients, or 52%) experienced changes in laboratory par-

ameters or sustained further lymphocytic organ damage,including interstitial pneumonia, renal tubular acidosis,neuropathy, vasculitis, andyor lymphomas or other typesof cancer. Six of the 23 SS patients who died(26%)had developed malignant lymphoma, with 5 of these 6patients dying within 5 years of SS onset. Six otherpatients(26%) developed other forms of cancer.SS can be divided into 3 stages(3). In Stage I(about45% of cases), patients have only sicca syndrome anddo not experience systemic involvement or changes inlaboratory parameters. In Stage II(about 50% of cases),patients experience lymphocytic organ damage, whichmay involve the pulmonary, renal, hepatic, hematologic,andyor dermatologic systems. Finally, in Stage III(about5% of cases), patients develop malignant lymphomas.Patients in Stages II and III may have some as yetunknown triggering factors that lead to further pro-gression.

2. Monoclonal gammopathy and monoclonal rheuma-toid factorOf the 23 SS sera with monoclonal serum immunoglob-ulins, 5 contained IgG, 9 contained IgA, 7 containedIgM, and 2 contained both IgG and IgM. Eight of these23 monoclonal immunoglobulins had rheumatoid factoractivity. Using monoclonal anti-idiotypic antibodiesagainst these rheumatoid factors, we detected a highincidence of rheumatoid factor idiotypes in the sera andlymphoma cells of other SS patients(4). ELISAsrevealed that a particular idiotype, SF18y2, was presentin 18% of SS sera, as well as being on the surface of Bcells in some SS patients. One myeloma cell line(ILKM3 ) established in our laboratory showed cross-reactivity with this idiotype, and the light-chain gene ofthis cell line was 96% homologous to the Vg germlinegene.These results suggested that the SF18y2 idiotype wasderived from the Vg germline gene or a similar gene.We therefore assayed the frequency of Vg gene usagein SS patients. When we amplified this gene fromperipheral blood mononuclear cell DNA, we found thatband intensity was much stronger in SS patients than incontrols, indicative of more frequent usage of the Vggene.Using a combinatorial phage display method and bonemarrow cells from an SS patient, we generated 8 humanmonoclonal anti-anti-idiotypic antibody clones againstmouse monoclonal anti-SF18y2 idiotypic antibody.Molecular analysis revealed that the VH3 gene was usedin all 8 clones and the Vk3 gene in 7(5). These resultsindicate that the idiotype network of this rheumatoid fac-tor is operative in SS patients, and that use of the Vg


gene, in conjunction with the VH3 and Vk3 genes, isimportant for the emergence of monoclonal B cells inSS patients.3. Benign lymphoproliferative disorders and malignantlymphomaWe observed 3 types of lymphoproliferative disorders inour SS patients: polyclonal lymphoproliferation, mono-clonal lymphoproliferation, and MALT lymphoma orhigh-grade malignant lymphoma.PCR amplification of the IgVH-CDR3 region in DNAfrom labial salivary biopsies resulted in oligo- or mon-oclonal bands in the DNA of 7 of 50 patients(6).Thirteen of our SS patients had Mikulicz’s disease.While 8 of these patients had features typical of SS, theother 5 did not have sicca complex. Biopsies of theenlarged glands in the first set of 8 patients showed that5 had marked lymphocytic infiltration with lymphoepi-thelial lesions. Two of these 5 patients had monoclonalDNA bands in the IgVH-CDR3 region, suggesting mon-oclonal lymphoproliferation. The other three, who hadno monoclonal characteristics, were thought to havepolyclonal lymphoproliferation. PCR amplification ofthe IgVH-CDR3 region and sequencing of the resultingbands showed progression from polyclonal or oligoclon-al to monoclonal lymphoproliferation. These patientsalso have the possibility of further transition to malig-nant lymphoma.Of our 25 SS patients with lymphoma, three were malesand 22 were females(Table 1). Twenty had primary SSand 5 had secondary SS. Thirteen patients had extran-odal types of lymphoma, including glandular lympho-mas, whereas 12 had nodal lymphomas. One hadHodgkin’s disease, whereas the other 24 had non-Hodg-kin’s lymphomas. Four had T cell type tumors(angioim-munoblastic lymphadenopathy with dysproteinemia,AILD-like T cell), and 21 had B cell type. Eight hadMALT lymphomas. Assays of the lymphoma biopsies of12 patients showed that only 2, one with a B cell andone with a T cell lymphoma, had Epstein-Barr virus.The incidence of autoantibodies in these patients waslower than in most SS patients.The importance of the CD40L–CD40 interaction hasbeen shown in our experiments in NZByW F1 mice, agood animal model for human systemic lupus erythe-matosus and SS. Treatment of mice with anti-CD40Lantibody prior to the onset of disease almost completelyinhibited disease development(7). These findings sug-gest close interactions among epithelial cells, T cells andB cells through cell surface molecules such as CD40,CD40L, CD80 and CD28. These interactions may allowthese cells to live cooperatively in lymphoepitheliallesions and not undergo apoptosis.

4. Disease progression in SSThe transition from polyclonal lymphoproliferation tomonoclonal lymphoproliferation to MALT lymphomaand finally to high-grade malignant lymphoma is con-sidered to be a multi-step process. Antigenic stimulationof B cells and oncogenic events may be important indisease progression. One of the candidate antigens maybe IgG, which stimulates rheumatoid factor clones.Important contributory factors in this progression are:

(1) Stimulation of specific B cells through surface Igbinding of exogenous or autoantigens.

(2) B cell activation through CD40L-CD40 binding.(3) Formation of lymphoepithelial lesions, enabling B

cells to become oligoclonal and then monoclonal.(4) Abnormal genetic events, including trisomy 3, p15y

16 hypermutation andt(1;14) translocation, further-ing the emergence of MALT lymphoma.

(5) Mutation of the p53 gene and deletion of the p16gene, allowing MALT lymphoma to progress tohigh-grade malignant lymphoma.

References

1. Moutsopoulos H.M., Chused T.M., Mann D.L., Klippel J.H.,Fauci A.S., Frank M.M., Lawley T.J., Hamburger M.I.: Sjo-¨gren’s syndrome(Sicca syndrome): current issues. Ann. Intern.Med. 1980; 92: 212–226.

2. Ramos-Casals M., Font J., Garcia-Carrasco M., Brito M.P.,Rosas J., Calvo-Alen J., Pallares L., Cervera R., Ingelmo M.Primary Sjogren syndrome: hematologic patterns of diseaseexpression. Medicine(Baltimore) 2002; 81:281–292.

3. Sugai S., Saito I., Masaki Y., Takeshita S., Shimizu S., Tachi-bana J., Miyasaka N. Rearrangement of the rheumatoid factor-related germline gene Vg and bcl-2 expression inlymphoproliferative disorders in patients with Sjogren’s syn-¨drome. Clin. Immunol. Immunopathol. 1994; 72:181–186.

4. Sugai S., Shimizu S., Tachibana J., Imaoka S., Konda S. A highincidence of rheumatoid factor idiotypes in monoclonal proteinsin the serum and in lymphoma cells in patients with Sjogren’s¨syndrome. J. Autoimmun. 1989; 2: 471–476.

5. Kim C.G., Masaki Y. Epitope cloning of anti-idiotypic antibodyA-SF18y2 reaction to immunoglobulin in Sjogren’s syndrome¨patients. J. Kanazawa Med. Univ. 2002; 27: 130–136(in Jap-anese, abstract in English)

6. Jordan R.C.K., Masaki Y., Takeshita S., Speight P.M., Sugai S.High prevalence of B-cell monoclonality in labial gland biop-sies of Japanese Sjogren’s syndrome patients. Int. J. Hematol.¨1996; 64: 47–52.

7. Cui G., Sugai S. Suppression of autoimmune diseases by block-ade of the CD40–CD40L interaction. J. Kanazawa Med. Univ.1999; 24: 137–144(in Japanese, abstract in English)


Table 1. Malignant Lymphomas Associated with Sjo-¨gren’s Syndrome

Total Extranodal Nodal

No. of cases 25 13 12Sex F22 M3 F11 M1 F10 M2

SS Duration(Y) 0–26 0–11 0–26P or S P20 S5 P12 S(PSS) 1 P8 S(RA) 4ML NHL24 HD1 NHL13 NHL11 HD1T or B cell T4 B20 B12 T1 B8 T3MALT type 8 7 1EBV q2 y10 q2 y6 q0 y4Laboratory dataANA 16y25 (64%) 6y13 (46.2%) 10y12 (83.3%)RF 10y25 (40%) 5y13 (38.5%) 5y12 (41.7%)A-RoySS-A 9y22 (40.9%) 6y13 (46.2%) 3y9 (33.3%)A-LaySS-B 3y22 (13.6%) 2y13 (15.4%) 1y9 (11.1%)

Abbreviations: SS, Sjogren’s syndrome; F, female; M, male; P, pri-¨

mary SS; S, secondary SS; PSS, progressive systemic sclerosis;

RA, rheumatoid arthritis; ML, malignant lymphoma; NHL, non-

Hodgkin’s lymphoma; HD, Hodgkin’s disease; MALT, mucosa-

associated lymphoid tissue; EBV, Epstein-Barr virus; ANA,

anti-nuclear antibody; RF, rheumatoid factor; A-RoySS-A, anti-

RoySS-A antibody; A-LaySS-B, anti-LaySS-B antibody.

35.Possible involvement of Epstein-Barr virusand its regulatory gene in rheumatoidsynovitis

Shigemasa Sawada and Masami Takei,(Department of Medicine, University School of Medi-cine,Tokyo, Japan.)Email: [email protected]

Epstein Barr virus(EBV), a human herpes virus, sub-sequent to primary infection is capable of remaininglatent in host lymphocytes. This virus is well known tobe associated with infectious mononucleosis, AfricanBurkitt’s lymphoma and nasopharyngeal carcinoma(1).As EBV plays an important role in the progressive pro-liferation of cells, EBV might be associated with theproliferating synovial cells in the patients with rheu-matoid arthritis(RA) since patients with RA have anti-bodies have antibodies that react with an antigen in thenucleus of EBV-transformed B cells called RA-associ-ated nuclear antigen(RANA) (2). Furthermore, thenumber of infected peripheral lymphocytes in RApatients tends to be increased compared to normal indi-viduals and RA patients show impaired ability to gen-erate EBV-specific cytotoxic T lymphocytes(3). The

association between RA and EBV has been suggested,based on the molecular mimicry hypothesis that is sup-ported by the similarity between RANA and EBNA-1(4). The antibody to EBNA-1 reacts with a 62 kDa pro-tein in the synovium of patients with RA(5), and ahomology in aminoacid sequences exists betweengp110, which is a component of the EBV capsid protein,and HLA-DR4 (6). The most important question how-ever, is whether EBV is present or not in the synovialcells from RA patients. Fox et al demonstrated thatimmunofluorescence and Southern blot assays were notsensitive enough for detection of EBV in synovial cellsof RA patients(7). However, there are several reportsfrom various sources including our department that havedemonstrated that parts of the EBV genome were ampli-fied by the polymerase chain reaction method(PCR)from synovial cells of RA patients(8–11), indicatingthat EBV exists in the RA synovial tissue. Nonetheless,PCR detection cannot specify the cell population in thesynovial tissue in which Epstein-Barr virus is present.We investigated the presence of EBV by detecting EBV-encoded small RNAs(EBER) in synovial cells with insitu hybridization and the expression of CD21 moleculesor latent membrane protein(LMP-1) and EBNA-2 withimmunostaning. EBER was detected in synovial cellsand lymphocytes from RA patients(23.5%) but not incontrol synovial tissue(from osteoarthritis and psoria-sis). In some cases, EBER were localized in synoviallining cells that were located at the apex of villus pro-liferating lesions. Furthermore, LMP-1 was also detectedin synovial cells. However CD19, CD21 molecules andEBNA-2 were not detected. The incidence of EBV-pos-itive cells in synovial cells was higher in heavily infil-trated tissues compared to the ones that were moderatelyinfiltrated. Recent studies have also demonstrated thatEBER or EBV DNA can be detected in synovial cellsfrom RA patients(9). Takeda et al and Blaschke et alalso demonstrated that EBV DNA or EBER-1 wasdetected in the synovial cells from patients with RA(10,11). In contrast Niedobitek et al did not detect EBVproducts in RA samples(12). However, Brousset dem-onstrated small, EBV-infected bystander lymphocytes bydetecting EBV-encoded RNA with in situ hybridisation(13). Niedobiteks, failure to detect EBV in synovial tis-sue may be atributed to BZLF1(ZEBRA) protein,which is strongly related to the lytic cycle of virus(13).It is well known that LMP-1 is involved in the malignanttransformation of fibroblasts, possibly due to the acti-vation of the bcl-2 gene, inhibiting apoptosis pathways.The aberrant proliferation of rheumatoid synovial cellsmay be due to IMP-1.


The entry of EBV to synovial cells is not clear becauseCD21, receptor for EBV, was negative on the synovialmembrane. It might be that CD21 molecule may beexpressed at the time of infection and subsequentlydown-regulated and be a requisite for the EBV infectionof synovial cells. Cell fusion with EBV-infected lym-phocytes has been suggested as a possible route of viralentry into non-lymphoid cells(14). Entry of EBV intosynovial cells may be due to cell fusion of EBV-positiveB lymphocytes. Recently, it has proposed that presynov-ial stem cells for synoviocyte is recruited into jointsfrom the bone marrow.It may be thought that patients with rheumatoid arthritishave EBV infected pre-synovial stem cells in their mar-row and those cells move to joint synovial membrane.In a recent study, in an attempt to locate the causativegene of X-linked lymphoproliferative syndrome(XLP),different groups have independently identified in XLPfamilies mutations in a novel gene(15–16). Its productis referred to as singnaling lymphocytic-activation mol-ecule(SLAM)-associated protein(SAP) or Src homolp-gy 2 domain-containing protein(SH2D1A). The fourtyrosine-based motifs in the cytoplasmic domain ofSLAM is known as CDw150(17), may be involved inthe recruitment of Src homology 2(SH-2)-domain-con-taining protein tyrosine phosphatase-2(SHP-2) as wellas the association between SLAM and SAP, whichmeans that SAP competes with SHP-2 for binding tophosphorylated SLAM(15). As to the SLAMySAPpathway, it has been proposed that SAP functions as aregulator of the signal transduction pathway initiated by2B4, which is primarily expressed on a subset of CD8qT lymphocytes and NK cells.(18,19). 2B4 is a surfacemolecule involved in the activation of NK cell-mediatedcytotoxicity. The malfunction of 2B4 molecules may bedirectly involved in the inability of cytotoxic lympho-cytes to kill EBV-infected cells(20)Indeed, the defective SAP protein caused by the mutatedSAP gene plays a crucial role in the pathogenesis of theinherited immunodeficiency XLP.We investigated the involvement of the SAP gene inpatients with RA. Using a quantitative real-time PCR,the expression level of SAP transcripts in peripheral Tlymphocytes or leukocytes was examined. The SAPtranscripts level in peripheral leukocytes of RA patientswas significantly lower than that of normal individuals,that of inactive systemic lupus erythematosus patients,and that of with chronic renal diseases patients(21). Thedecreased of SAP transcripts level in patients with RAwas also observed in peripheral CD2 T lymphocytescompared with normal individuals. Furthermore, wefound that the nucleotide sequence of the SAP cDNAs

was not shown any mutations or deletions in their cod-ing region compared with wild-type SAP cDNA. AsAltered 2B4 function by the decreased SAP or SH2D1Aproteins may contribute to fail the suppressive controlto EBV infective cells by cytotoxic T lymphocytes orNK cells, because that the mechanisms to sustaining theelimination of EBV-infected cells might be the same asthose of the elimination in patients with XLP. CD48 isthe ligand of 2B4 and the expression on EBV-trans-formed B cells is 10-fold greater than that on EBV-neg-ative B cells (22). The up-regulation of CD48 onEBV-transformed B may act as a signal to specificallyactivate NK cells via 2B4 and induce lysis of trans-formed cels. Impaired 2B4ySAP pathway may contrib-ute to fail to eliminate EBV-infected cells by cytotoxicT lymphocytes or NK cells in patients with RA.The role of the decreased SAP transcripts in patientswith RA is unclear. Genomic polymorphism at promoteror enhancer regions may exist in patients with RA. Thispossible pathogenesis should be investigated in thefuture.

References

1. Klein G. 1973. The Epstein-Barr virus. In Kaplan A.S. ed.,The Herpes Viruses, p. 521. Academic Press, New York.

2. Alspaugh M.A., Jensen F.C., Rabin H., Tan E.M. Lymphocytetransformed by Epstein-Barr virus. Induction of nuclear anti-gen reactive with antibody in rheumatoid arthritis. J. Exp.Med. 147: 1018–27, 1978.

3. Tosato G., Steinberg A.D., Yarchoan R., Heilman C.A., PikeS.E., De-Seau V., Blaese R.M. Abnormally elevated frequencyof Epstein-Barr virus-infected B cells in the blood of patientswith rheumatoid arthritis. J. Clin. Invest. 73: 1789–95, 1984.

4. Venables P.J., Pawlowski T., Mumford P.A., Brown C., Craw-ford D.H., Maini R.N. Reaction of antibodies to rheumatoidarthritis nuclear antigen with a synthetic peptide correspondingto part of Epstein-Barr nuclear antigen. Ann Rheum. Dis. 47:270–9, 1988.

5. Fox R., Sportsman R., Rhodes G., Luka J., Pearson G.,Vaughan J. Rheumatoid arthritis synovial membrane containsa 62 000-molecular weight protein that shares an antigenic epi-tope with the Epstein-Barr virus encoded associated nuclearantigen. J. Clin. Invest. 77: 1539–47, 1986.

6. Roudier J., Rhodes G., Petersen J., Vaughan J.H., Carson D.A.The Epstein-Barr virus glycoprotein gp110, a molecular linkbetween HLA DR4, HLADR1 and rheumatoid arthritis. Scand.J. Immunol. 27: 367–71, 1988.

7. Fox R.I., Chilton T., Rhodes G., Vaughan J.H. Lack of reac-tivity of rheumatoid arthritis synovial membrane DNA withcloned Epstein-Barr virus DNA probes. J. Immunol. 137: 498–501, 1986.


8. Newkirk M.M., Watanabe D.K.H., Leclerc J., Lambert N., Shi-roky J.B. Detection of cytomegalovirus, Epstein-Barr virus andHerpes virus-6 in patients with rheumatoid arthritis with orwithout Sjogren’s syndrome. Br. J. Rheum. 33: 317–22, 1994.

9. Takei M., Mitamura K., Fujiwara S., Horie T., Tyu J., OsakaS., Yoshiono S., Sawada S. Detection of Epstein-Barr virusencoded small RNA 1 and latent membrane protein 1 in syn-ovial linig cells from rheumatoid arthritis patients. Int. Immu-nol. 9: 739–43, 1997.

10. Takeda T., Mizugaki Y., Matsubara L., Imai S., Koike T., Tak-ada K. Lytic Epstein-Barr virus infection in the synovial tissueof patients with rheumatoid arthritis. Arthritis Rheum. 43:1218–25, 2000.

11. Blaschke S., Schwarz G., Monke D., Binder L., Muller G.,Reuss-Borst M. Epstein-Barr virus infection in peripheralblood mononuclear cells, synovial fluid cells and synovialmembranes of patients with rheumatoid arthritis. J. Rheumatol.27:866–73, 2000.

12. Niedobitek G., Lisner R., Swoboda B., Rooney N., FassbenderH.G., Kirchner T., Aigner T., Herbst H. Lack of evidence foran involvement of Epstein-Barr virus infection of synovialmembranes in the pathogenesis of rheumatoid arthritis. Arthri-tis Rheum. 43: 151–4, 2000.

13. Brousset P. Lack of Epstein-Barr virus infection of synovialmembranes in patients with rheumatoid arthritis: comment onthe article by Niedobitek et al. Arthritis Rheum. 43: 2614,2000.

14. Timens W., Boes A., Vos H., Poppema S. Tissue distributionof D3dyEBV-receptor: CD21 monoclonal antibodies reactivewith a variety of epithelial cells, medullary thymocytes andperipheral T cells. Histochemistry 95: 605–11, 1991.

15. Sayos J., Wu C., Morra M., Wang N., Zhang X., Allen D., etal. The X-linked lymphoproliferative-disease gene productSAP regulates signals induced through the co-receptor SLAM.Nature 395:462–9, 1998.

16. Coffey A.J., Brooksbank R.A., Brandau O., Oohashi T., How-ell G.R., Bye J.M., et al. Host response to EBV infection inX-linkedlymphoproliferative disease results from mutaions inan SH-2-domain encoding gene. Nat. Genet. 20: 129–35,1998.

17. Cocks, B.G., Chang C.C., Carballido J.M., Yssel H., de VriesJ.E., Aversa G. A novel receptor involved in T-cell activation.Nature 376: 260–3, 1995.

18. Mathew P.A., Garni-Wargner B.A., Landd K., Takashima A.,Stoneman E., Bennet M., et al. Cloning and characterizationof the 2B4 gene encoding a molecule associated with non-MHC-restricted killing mediated by activated natural killercells and T cells. Immunol. 151: 5328–37, 1993.

19. Boles K.S., Nakajima H., Colonna M., Chuang S.S., SteppS.E., Bennett M., et al. Molecular characterization of a novelhuman natural killer cell receptor homologous to mouse 2B4.Tissue Antigen 54: 27–34, 1999.

20. Tangye S.G., Phillips J.H., Lanier L.L., Nichols K.E. Cuttingedge: Functional requirement for SAP in 2B4-mediated acti-vation of human natural killer cells as revealed by the X-linkedlymphoproliferative syndrome. J. Immunol. 165:2932–6,2000.

21. Takei M., Ishiwata T., Mitamura K., Fujiwara S., Sasaki K.,Nishi T., et al. Decreased expression of signaling lymphocytic-activation molecule-associated protein(SAP) transcripts in Tcells from patients with rheumatoid arthritis. Int. Immunol. 13:559–68, 2001.

22. Thorley-Lawson D.A., Schooley R.T., Bhan A.K., Nadler L.M.Epstein-Barr virus superinduces a new human B cell differ-entiation agtigen(B-LAST 1) expressed on transformed lym-phoblasts. Cell 30: 415–36, 1982.

36.Autoimmune diseases: role ofcoxsackieviruses in their pathogenesis

Liakos D.A., Triantafyllopoulou A., Kapsogeorgou E.K.and Moutsopoulos H.M.,(Department of Pathophysiology, Medical School,National University of Athens, 75 Mikras Asias str,11527, Athens, Greece).Email: [email protected]

Autoimmune epithelitis or Sjogren’s syndrome(SS) (1)ïs a chronic autoimmune disease that is characterized bythe dysfunction and destruction of exocrine glands, as aresult of focal lymphocytic infiltrates. The syndrome isas common as rheumatoid arthritis and mainly affectswomen in their forth and fifth decade of life. The clinicalspectrum of the disease is wide and ranges from focalexocrinopathy, to parenchymal organ involvement andneoplasia in the form of B cell lymphoma. The diseasecan be seen as an entity alone(primary SS) or in asso-ciation with other rheumatic autoimmune diseases(sec-ondary SS). In addition Sjogren’s syndrome is¨characterized by the presence of autoantibodies againstcellular antigens. Furthermore, the affected tissue(minorsalivary gland) is easily accessible, making biopsies rel-atively easy to perform. All these above characteristicsmake Sjogren’s syndrome a prototype autoimmune dis-ëase for the study of autoimmune responses as well asneoplasia.As indicated by findings in minor salivary gland epithe-lial cells (SGEC) from Sjogren’s syndrome patients, epi-¨thelium plays an active role in the initiation andperpetuation of autoimmune processes. The expressionof proinflammatory cytokines, lymphoattractant che-mokines and adhesion molecules in addition to the


expression of MHC class II molecules and functional co-stimulatory molecules by SS epithelial cells(2–5), sug-gests that they may act as antigen presenting cells.Furthermore, long term cultures of SGECs from Sjo-¨gren’s syndrome patients display an activated phenotypeindicating the operation of intrinsic activation factors.Based on these findings, we proposed for the first timein international literature the term ‘‘autoimmune epi-thelitis’’, a term that has been accepted by the scientificcommunity(6).The implication of viruses in the development of Sjo-¨gren’s syndrome has long been suspected. There is noclear evidence so far that proves viral induction of Sjo-¨gren’s syndrome. However, there is data that links virus-es and the disease. It has been shown that Sjogren’s¨syndrome patients express EBV-associated antigens intheir salivary glands and also display an increased con-tent of EBV-DNA in their saliva, but these data havebeen challenged by other groups. Another link betweenviruses and the disease comes from the fact that chroniclymphocytic sialadenitis that is linked to viral infectionis histologically very similar to Sjogren’s syndrome(7).¨Retroviruses can also cause sialadenitis in patientsinfected with human immunodeficiency virus(HIV) (8)and human T-lymphotropic virus I(HTLV-I ) (9).Although there is great similarity between sialadenitisand Sjogren’s syndrome, the two entities are different¨since there are differences in severity of tissue damagepresence of autoantibodies.In order to identify genes that may contribute to primarySjogren’s syndrome pathogenesis the differential display¨protocol was applied to minor salivary gland RNA sam-ples of a patient with primary Sjogren’s syndrome andä healthy control individual. After sequencing of severaldifferentially expressed genes, a fragment homologousto coxsackievirus RNA expressed exclusively in the dis-eased sample was identified.Coxsackieviruses belong to the large viral family ofpicornaviruses. Coxsackieviruses, are small non-envel-oped RNA viruses that are classified on the basis of theirantigenic response and are divided into two major sero-type groups, A and B. Coxsackievirus infection givesrise to a variety of human diseases. Herpangina, acutehemorrhagic conjuctivitis and hand-foot-and-mouth dis-ease are all caused by type A coxsackieviruses. Type Bcoxsackieviruses are known to cause myocarditis, peri-carditis and meningoencephalitis. Both types also giverise to aseptic meningitis, respiratory and undifferentiat-ed febrile illnesses and hepatitis. The virion of coxsack-ieviruses consists of a capsid shell of 60 subunits,arranged in pentamers, each being composed of fourproteins (VP1-VP4). The genome of the virus is an

RNA molecule of about 7.4 kb that codes for a singlepolyprotein that is cleaved to produce the various pro-teins that are required for virion structure and replica-tion. Coxsackie viruses gain entrance into cells bybinding to two different receptors, coxsackie-adenovirusreceptor(CAR) and decay-accelerating factor(DAF).In order to further examine the presence of coxsackie-viral RNA in Sjogren’s syndrome patient samples weäpplied an RT-PCR protocol designed to amplify aregion that is conserved among enteroviruses. Ourresults show that coxsackieviral RNA is present in tis-sues and cultured SGEC from SS patients but not innormals. To exclude the possibility of a generalizedinfection, we also examined peripheral blood lympho-cytes collected at the same time point of the biopsy.Coxsackieviral RNA was not detected, indicating theabsence of systemic viral infection that could affectpatient tissue samples. These findings indicate for thefirst time in human disease tissues the presence ofcoxsackieviral sequences, finding which suggests thatthese viruses may participate in the pathogenesis of Sjo-¨gren’s syndrome.Based on the above we propose the following workingmodel on the role of coxsackieviruses in the pathogen-esis of Sjogren’s syndrome: In genetically predisposedïndividuals the virus remains latent following infection.After stimulation through hormonal, environmental andstress related factors the virus becomes active and par-ticipates in the stimulation of the epithelial cells. Theactivated state of the epithelium leads to the autoimmuneresponse and tissue damage that characterizes Sjogren’s¨syndrome.

Bibliography

1. Manoussakis M.N., Moutsopoulos H.M. Sjogren’s syndrome:current concepts. Adv. Intern. Med. 2001;47:191–217.

2. Skopouli F.N., Fox P.C., Galanopoulou V., Atkinson J.C., JaffeE.S., Moutsopolous H.M. T cell subpopulations in the labialminor salivary gland histopathologic lesion of Sjogren’s syn-drome. J. Rheumatol. 1991;18:210–4.

3. Manoussakis M.N., Dimitriou I.D., Kapsogeorgou E.K., Xan-thou G., Paikos S., Polihronis M., Moutsopoulos H.M. Expres-sion of B7 costimulatory molecules by salivary gland epithelialcells in patients with Sjogren’s syndrome. Arthritis Rheum.1999;42:229–39.

4. Dimitriou I.D., Kapsogeorgou E.K., Moutsopoulos H.M., Man-oussakis M.N. CD40 on salivary gland epithelial cells: highconstitutive expression by cultured cells from Sjogren’s syn-drome patients indicating their intrinsic activation. Clin. Exp.Immunol. 2002;127:386–92.


5. Boumba D., Skopouli F.N., Moutsopoulos H.M. CytokinemRNA expression in the labial salivary gland tissues frompatients with primary Sjogren’s syndrome. Br. J. Rheumatol.1995;34:326–33.

6. Moutsopoulos H.M. Sjogren’s syndrome: autoimmune epithel-itis. Clin. Immunol. Immunopathol. 1994;72:162–5.

7. Mariette X., Zerbib M., Jaccard A., Schenmetzler C., Danon F.,Clauvel J.P. Hepatitis C virus and Sjogren’s syndrome. ArthritisRheum. 1993;36:280–1.

8. Kordossis T., Paikos S., Aroni K., Kitsanta P., DimitrakopoulosA., Kavouklis E., Alevizou V., Kyriaki P., Skopouli F.N., Mout-sopoulos H.M. Prevalence of Sjogren’s-like syndrome in acohort of HIV-1-positive patients: descriptive pathology andimmunopathology. Br. J. Rheumatol. 1998;37:691–5.

9. Terada K., Katamine S., Eguchi K., Moriuchi R., Kita M., Shi-mada H., Yamashita I., Iwata K., Tsuji Y., Nagataki S., et al.Prevalence of serum and salivary antibodies to HTLV-1 in Sjo-gren’s syndrome. Lancet 1994;344:1116–9.

Documents

Autoimmune rheumatic diseases days held in Athens in June 2004