Annals of the Rheumatic Diseases 1993; 52: 785-789

Occurrence of antiperinuclear, antikeratin, andanti-RA 33 antibodies in juvenile chronic arthritis

Cem Gabay, Anne-Marie Prieur, Olivier Meyer

AbstractObjectives-Antiperinuclear factor(APF), antikeratin antibodies (AKA), andanti-RA 33 antibodies are currentlyconsidered to be good markers for thediagnosis of adult rheumatoid arthritiswith or without rheumatoid factor (RF).The prevalence of these markers wasretrospectively reviewed in children withjuvenile chronic arthritis (JCA) todetermine whether they were associatedwith specific features.Methods-One hundred and twenty-fourpatients with JCA participated in thisstudy. Controls included 28 patients withjuvenile systemic lupus erythematosusand 21 healthy children. Antiperinuclearfactor and AKA were determined byindirect immunofluorescence on buccalmucosal cells and oesophagus sectionsrespectively. Anti-RA 33 antibodies weredetected using a Western blot techniqueon HeLa celi nuclear extract.Results-Antiperinuclear factor wasvirtually absent in all the tested subgroupsand anti-RA 33 antibodies were detectedonly in a subset of patients with RFpositive polyarticular onset. Antikeratinantibodies were found in 27% of allchildren withJCA and in 42% ofthose withRF negative polyarticular onset. Theseresults were statistically significantcompared with healthy controls, but thepresence of AKA was not specific to anypatient subgroup. Moreover, in contrastwith previous studies in adult RA, norelation was found between the presenceofAKA and disease severity or activity.Conclusion-These data suggest thatAPF, AKA, and anti-RA 33 antibodies arenot useful for the diagnosis orclassification ofJCA.

(Ann Rheum Dis 1993; 52: 785-789)

Detection of autoantibodies contributes to thediagnosis of many inflammatory rheumaticdiseases. Rheumatoid factor (RF) in rheu-matoid arthritis (RA) and antinuclear anti-bodies (ANA) in systemic lupus erythematosus(SLE) are included in the diagnostic criteria forthese disorders.' 2 Antiperinuclear factor(APF), antikeratin antibodies (AKA), and anti-RA 33 antibodies are currently considered tobe good markers for adult RA, with or withoutRF, but little information is available on theiroccurrence in juvenile chronic arthritis (JCA).

Antiperinuclear factors are IgG antibodiesdirected against keratohyaline granulessurrounding the nuclei of human buccalmucosal cells.3 I The exact biochemical natureof these antigens remains unknown, however.The presence ofAPF has been reported in theserum of 48-92% of patients with RA5-7 and of8.6-76% of seronegative patients with RA.5-11Although their occurrence in various otherrheumatic disorders10-12 and in patientsinfected with the Epstein-Barr virus has alsobeen reported,'3 the specificity of APF for RAhas been estimated at more than 900/o.7 9 11 14

Antikeratin antibodies are complementfixing IgG antibodies directed against thestratum corneum of rat oesophagus squamouscells.5 16 Evidence suggests that the keratinantigen shares some homology with the Cterminal portion of the heterogeneous nuclearRNP protein Al.7 These antibodies have beendetected in 36-69% of patients withRA5 "' 15 16 18-22 and in 5-47% of seronegativepatients with RA.5 11 16 20 23 Antikeratinantibodies were shown to be highly specific forRA when detected in synovial fluid.24 Theywere found in the serum of O-8% of healthycontrols and patients with other rheumaticdisorders. 5 19-21 23 24 Higher percentages werealso reported in subjects with progressivesystemic sclerosis22 25 and in patients withankylosing spondylitis.22More recently, anti-RA 33, an IgG antibody

directed against a nuclear protein antigen, witha molecular weight of 33 kilodaltons and whichseems be identical to the A2 protein of theheterogeneous nuclear ribonucleoprotein(hnRNP),26 was detected in approximately onethird of patients with RA27 and in 27-45% ofseronegative RA serum samples. "1 27

Juvenile chronic arthritis presents as a varietyof clinical patterns characterised by the absenceof the usual serological markers, with theexception ofRF positive JCA and pauciarticularonset JCA positive for ANA. The occurrence ofAPF in JCA serum samples has already beenreported.28 Serum APF was significantlyassociated with JCA only in patients with RFpositive and RF negative polyarticular onset,however. Furthermore, these data are atvariance with a study in which APF was nearlylacking in serum samples from patients withJCA.29 Antikeratin antibodies have beendetected in only one of 28 children with JCAincluded in a study of the occurrence of AKAin various rheumatic disorders.2' Antibodies toRA 33 have not been studied in JCA to date.The aim of this study was to determine the

occurrence of these three autoantibodies in a

Clinique deRhumatologie, HopitalBichat-ClaudeBernard, Paris, FranceC Gabay0 MeyerUnite fonctionnelile deRhumatologiepediatrique,Departement dePediatrie, H6pital des'Enfants-Malades',Paris, FranceA-M PrieurCorrespondence to:Dr C Gabay,Division de Rhumatologie,H6pital CantonalUniversitaire, CH-1211Geneva 14, Switzerland.

Accepted for publication15 July 1993

785

Gabay, Prieur, Meyer

large cohort of patients with JCA, across thespectrum of onset patterns, and to look forassociations between these antibodies andspecific clinical features. Furthermore, becauseAPF and AKA have been associated withincreased disease severity7 8 16 18 20 23 and extra-articular features'2 18 21 24 in patients with RA,we also sought to determine whether theseantibodies were ofuse for predicting the courseof articular manifestations or the developmentof selected extra-articular features in patientswith JCA.

Patients and methodsPATIENTS

Serum samples from 124 patients with JCAfulfilling the EULAR/WHO criteria"l wereanalysed retrospectively. Nearly all thechildren were followed up regularly at the'Enfants-Malades' Hospital in Paris by one ofus (AMP). Patients with JCA were classifiedaccording to the mode of onset: systemic in 24cases; early pauciarticular in 51 cases; RFnegative polyarticular in 24 cases, RF positivepolyarticular in seven cases; and juvenilespondylarthropathy in 18 cases. This lattersubgroup was defined according to criteriaproposed in 1990.3' Sex ratio, mean (SE) ageat onset, and mean (SE) disease duration at thetime of the study are shown in table 1 for thedifferent subtypes.The course of the articular disease,

occurrence of specific systemic features and ofacute or chronic uveitis, and treatment duringfollow up were abstracted from clinical charts.Among patients with pauciarticular onset JCA,those with persistent pauciarticular diseasewere defined as having a cumulative total offour or fewer affected joints over the entireperiod of follow up, whereas those whosubsequently converted to a polyarticularcourse were defined as having a cumulativetotal of more than five affected joints. Chronicuveitis was defined as the presence of typicalsigns on slit lamp examination (performed twoto three times a year in all patients withpauciarticular onset JCA during follow up).Physical examination with particular attentionto the presence of active inflammatory osteo-articular features was performed at the time ofserum collection. Erythrocyte sedimentationrate (ESR) was also measured in 109 of the 124patients.

Control patients included 21 childrenwithout rheumatic disease nor inflammatorydisorders and 28 children with SLE.

Table 1 Sex ratio, age at disease onset, and disease duration at the time of the study in thevarious subgroups ofjuvenile chronic arthritis (JCA). Age and disease duration are givenas arithmetic mean (SE)

Diagnosis Sex ratio Age at Disease(FIM) disease onset duration

(years) (years)

Systemic onsetJCA (n = 24) 12/12 4-96 (0 85) 4-84 (0 74)Pauciarticular onset JCA (n = 51) 43/8 3-14 (0 33) 4-21 (0 50)Polyarticular onset JCA, RF positive (n = 7) 6/1 8-36 (0-86) 7-49 (2 69)Polyarticular onset JCA, RF negative (n = 24) 17/7 7 59 (0-84) 5-93 (1-31)Juvenile spondylarthropathy (n = 18) 4/14 11-05 (0-57) 4-82 (1-03)

Abbreviations: F = female; M = male; RF = rheumatoid factor.

ANTIPERINUCLEAR ANTIBODY ASSAY

This assay was carried out as previouslydescribed by Youinou et al.9 Briefly, humanbuccal mucosal cells were scraped with awooden tongue depressor from the inner sideof both cheeks of a healthy volunteer. Theywere washed three times in phosphate bufferedsaline (PBS), pH 7 4, and resuspended in PBScontaining colimycin and sodium azide. Thecells were then transferred dropwise tomultispot slides, roughly 10 000 cells/well.After drying with a fan, the slides were used asthe substrate. Serum samples were initiallydiluted 1:40 and applied to the slide for 90minutes in a moist chamber. Slides were thenwashed three times with PBS, air dried, andincubated for 30 minutes with fluorescentantihuman IgG (Cappel Products, Flobio,Coubevoie, France) diluted to 1:80. The slideswere then washed twice with PBS and oncewith distilled water, air dried, mounted withglycerol, and studied with a fluorescencemicroscope. The fluorescence of perinuclearhomogeneous spheres in at least 10% ofexamined cells was considered to indicate apositive test for antiperinuclear antibodies. Ifpositive at this dilution, serum samples werefurther diluted at 1:80.

ANTIKERATIN ANTIBODY ASSAY

Unfixed cryostat sections (2 pLm thick) of thedistal third of rat oesophagus were used asantigen targets for the detection ofAKA usingan indirect immunofluorescent technique.Cryostat sections were transferred to slides anddried by silica gel granules (Prolabo, Paris,France). Slides were kept at -30°C until theassays (for a few weeks). Slides were againdried with silica gel before the assay. Patientserum samples diluted to 1:20 in PBS wereincubated on the slides for 30 minutes. Theslides were rinsed twice for 10 minutes in PBS,air dried, and incubated for 30 minutes with afluorescein labelled antihuman IgG (InstitutPasteur, Paris, France) diluted to 1:80. Slideswere then rinsed twice for 10 minutes in PBS,air dried, mounted with glycerol, and studiedwith a fluorescence microscope. Only serumsamples showing laminar fluorescent stainingof the superficial layer of the oesophagealepithelium were considered positive. If positiveat this dilution, serum samples were furtherdiluted until the last positive titre wasdetermined. The sensitivity and specificity ofthis assay for seropositive RA have been shownto be 63 and 96% respectively. Sensitivity forseronegative RA was 47-6%.6ll

ANTI-RA 33 ANTIBODY ASSAY

Nuclear extract was prepared as previouslydescribed27 and subjected to 12% sodiumdodecylsulphate polyacrylamide gel electro-phoresis (SDS-PAGE) according to theprocedure described by Laemmli.32 Immuno-logical detection of RA 33 antibodies wasperformed after transfer to nitrocellulose33 (onehour at 2-5 mA) using serum samples diluted25-fold in incubation buffer (PBS containing

786

Autoantibodies and juvenile chronic arthritis

3% non-fat dried milk with and without 0 01%Triton-X100). Nitrocellulose strips wereincubated with serum samples for 120 minuteswith constant shaking. Bound autoantibodieswere detected using phosphatase conjugatedgoat antihuman IgG antibodies (Cappelproducts) in PBS-3% non-fat dried milk-Triton 0 01%.

OTHER SEROLOGICAL ASSAYSAntinuclear antibodies were determined beforeand at the time of the study by indirectimmunofluorescence on rat liver sections. Thetest was considered as positive at a dilution of1:20. Rheumatoid factor was evaluated at leasttwice by the Singer Plotz latex test during thefollow up of all children with polyarticularonset JCA.

STATISTICAL ANALYSISFisher's exact test for two or five groups wasused to compare percentages. Statisticalanalysis was performed with Statxact (CytelSoftware, Cambridge, MA, USA). Quan-titative data (age of onset, disease duration,ESR values) were tested using unpaired t tests;two tailed p values <0.05 were consideredsignificant.

ResultsThe presence of AKA, anti-RA 33 antibodies,and APF in 124 children with different onsetpatterns ofJCA, 28 patients with juvenile SLE,and 21 healthy controls are shown in table 2.Antikeratin antibodies were detected in 27% of

Table 2 Presence of antiperinuclearfactors, antikeratin antibodies, and anti-RA 33antibodies in children with juvenile chronic arthritis (7CA), juvenile systemic lupuserythematosus and healthy controls. Results given are number positive (%o)Diagnosis No (%) No (oo) No (%o)

positive APF positive AKA positive anti-RA 33

Juvenile chronic arthritisSystemic onset (n = 24) 0 (0) 3 (13) 2 (8)Pauciarticular onset (n = 51) 0 (0) 12 (24) 1 (2)Polyarticular onset, seropositive (n = 7) 1 (14) 2 (29) 4 (57)*Polyarticular onset, seronegative (n = 24) 0 (0) 10 (42)t 2 (8)Juvenile spondylarthropathy (n = 18) 1 (6) 6 (33) 3 (17)Total (n = 124) 2 (1-6) 33 (27)t 14 (11)

Juvenile systemic lupus erythematosus (n = 28) 0 (0) 5 (18) 6 (21)Healthy controls (n =21) 0 (0) 1 (5) 0 (0)

Abbreviations: APF = antiperinuclear factor; AKA = antikeratin antibodies. Statistical analysis byFisher's exact test.*p<0.05 v controls; comparisons of percentages between JCA subgroups showed significantdifferences.tp<0 05 v controls; but comparisons of percentages between JCA subgroups showed nosignificant differences.tp<0-05 v controls.

Table 3 Comparison ofpatients with rheumatoidfactor negative polyarticular onsetjuvenile chronic arthritis positive (AKA+) and negative (AKA-) for antikeratinantibodies. p Values were not significant in any instance

AKA+ AKA-(n= 10) (n= 14)

Mean (SD) age at onset (years) 7-72 (1-52) 7-48 (0-97)Mean (SD) disease duration (years) 4-58 (1-7) 7-65 (1-9)Mean (SD) ESR* 34-2 (7 9) 25-6 (4 7)No (%/6) patients with active synovitis* 7/10 (70) 11/14 (79)No (%) patients receiving DMARDs* 3/10 (30) 5/14 (36)

Abbreviations: ESR = erythrocyte sedimentation rate; DMARDs = disease modifyingantirheumatic drugs. Statistical analysis by Fisher's exact test and unpaired t test.*At time of study.

all patients with JCA. This proportion wassignificantly greater than that in healthycontrols (p<0.05). Antikeratin antibodies werefound in 42% of patients with RF negativepolyarticular onset JCA. This proportion wassignificantly higher than that in healthycontrols. The proportions of AKA positivepatients were not significantly different in thefive JCA subgroups (5 x 2 tailed Fisher's exacttest, however). Anti-RA 33 antibodies werefound in 1 1% of all patients with JCA. Thisproportion was not significantly different fromthat in controls. The occurrence of theseantibodies varied significantly across the fiveJCA subgroups (p<0 001); the subset ofpatients with RF positive polyarticular onsetJCA had a significantly higher percentage ofanti-RA 33 positive serum samples thanhealthy controls (p<0 01). Specificity ofRA 33antibodies in healthy controls was 100% (95%confidence interval, 84 to 100%). Anti-RA 33antibodies were, however, also found in 21%of the children with SLE. Antiperinuclearantibodies were nearly always absent in thedifferent groups of subjects.To look for associations between AKA and

specific clinical outcomes, we comparedchildren positive and negative for AKA withchildren with RF negative polyarticular onsetJCA. Age at disease onset and disease durationwere identical in the two groups. As shown intable 3, no difference was found for diseaseactivity during the study, as defined by thenumber of patients with active synovitis andincreased ESR levels. The number of patientsreceiving disease modifying antirheumaticdrugs was identical in the two groups. Systemicand ocular manifestations were absent in thetwo groups. Functional outcome could not beassessed as none of these children had afunctional score higher than 2, as defined bythe Steinbrocker's functional index.34 Thepresence of AKA was not associated withspecific clinical features in early onsetpauciarticular JCA. Conversion to poly-articular disease and occurrence of chronicuveitis were not influenced by the status ofAKA.With the exception of one patient with

polyarticular seropositive onset, dilution titresof serum samples positive for AKA rangedfrom 1:20 to 1:160 in patients with JCA,whereas these values are generally higher inadult RA, reaching 1:1000 in some instances.Dilution titres found in our study werecompared with those in 14 RA serum samplespositive for AKA assessed at the same time.Dilution titres in the RA serum samplespositive for AKA were significantly higher thanthose found in our patients with JCA (data notshown). These results are in accordance withprevious reports showing that high dilutiontitres of AKA are pathognomonic for adultRA 16 20

Results of tests for ANA performed beforeand during the study in patients with JCA andin healthy controls are shown in table 4. Aspreviously reported,35 these antibodies werefound in significantly higher percentages of thepauciarticular and seropositive polyarticular

787

Gabay, Prieur, Meyer

Table 4 Presence of antinuclear antibodies (ANA) in children with juvenile arthritisbefore and during the study

Diagnosis No (Go) ANA positive

Before study During study

Juvenile chronic arthritisSystemic onset (n = 24) 1 (4) 0 (0)Pauciarticular onset (n = 51) 45 (88) 36 (71)*Polyarticular onset, RF positive (n = 7) 2 (29) 4 (57)*Polyarticular onset, RF negative (n = 24) 2 (8) 2 (8)Juvenile spondylarthropathy (n = 18) 0 (0) 1 (6)Total (n = 124) 49 (40) 43 (35)*

Healthy controls (n = 21) 0 (0)

*p<O O 1 v healthy controls by Fisher's exact test.

Table S Comparison between antinuclear antibodies (ANA) and antikeratin antibodies(AKA) or anti-RA 33 antibodies in serum samples ofpatients with juvenile chronicarthritis (JCA). Results given are numbers ofpositive (+) or negative (-) serum samples

Diagnosis ANA AKA Anti-RA 33

+ _ + _

Systemic onset (n = 24) + 0 0 0 03 21 2 22

Early onset pauciarticular (n = 51) + 8 28 0 364 11 1 14

Polyarticular onset, seronegative (n = 24) + 1 1 0 29 13 2 20

Polyarticular onset, seropositive (n = 7) + 1 3 2 21 2 2 1

Juvenile spondylarthropathy (n = 18) + 0 1 0 17 10 3 14

Total JCA (n = 124) + 10 33 2 41- 24 57 10 71

Statistical analysis by Fisher's exact test shows no correlations between ANA and AKA or anti-RA 33 in patients with JCA (p values ranged from 0-2 to 1).

onset subgroups than in controls (p<001 andp<00001 respectively). Antinuclear antibodystatus was independent of the results of testsfor other antibodies (table 5). Among the 45patients with pauciarticular onset JCA positivefor ANA, 41 (91%) had a homogeneousfluorescent pattern. Antinuclear antibodieswere present in all the patients with chronicuveitis and in 79% of those without ocularcomplications (p<006). Native anti-DNAantibodies were not detected in any of the JCAserum samples.

DiscussionConsistent with a previous report,29 APF wasrarely detected in the different JCA subgroupsincluded in this study. This result is, however,at variance with that of Nesher et al,28 whofound these antibodies in up to 37% ofchildren with polyarticular onset JCA. Thispercentage was even higher in RF positive JCA(83%).The numerous hypotheses which may

explain this difference include inadequatesensitivity of our assay; methodological differ-ences involving serum dilution and criteria forpositive results; and ethnic differences betweenstudy populations.The sensitivity and specificity of our assay

have been assessed in RA and in other adultrheumatic disorders. The occurrence of APFwas found in 56% of RF positive (73 serumsamples) and 33% of RF negative (21 serumsamples) patients with RA, with a specificity of93% (unpublished data). These results are inaccordance with those published elsewhere. Amajor cause of variability in APF results is thevariation of substrate antigen between different

donors.36 This problem probably had noinfluence in our assay as we used a singledonor.Nesher et al found that APF was significantly

detected only in undiluted serum or at adilution of 1:5.28 As in previous studies,9 serumsamples were diluted to 1:40 and 1:80 in ourassay. To determine whether this differencemight explain the variation in APF results, wetested 30 APF negative JCA serum samples atdilutions ranging from 1:2 to 1:20.Antiperinuclear antibodies were detected atdilution 1:2 in only two patients, who belongedto different JCA subgroups (data not shown).Most centres have their own criteria for

positive APF results. We and others9 considerthat the test is positive when 10% or more ofthe cells show the typical perinuclear fluor-escence. This threshold of positivity has beenshown to increase specificity to 97% for thediagnosis of RA, without appreciably reducingsensitivity.9 For others, one or a few typicalcells are sufficient for a positive result. Previousstudies in JCA have not provided details on thispoint.

It has been shown that the sensitivity of theAPF test for the diagnosis of adult RA can beincreased by pretreating cells with thedetergent Triton-X1 00.3 This procedurereduces specificity, however. We therefore didnot use Triton-X100 in our assay. Further-more, as found by Von Essen et al,29 Triton-X100 did not increase the sensitivity ofAPF forthe diagnosis ofJCA.Nesher et al reported that the prevalence of

APF varies across populations.38 Our patientswere ethnically diverse and, consequently, it isunlikely that our results were due to a selectionof a particular ethnic group.The presence ofAKA was significant only in

a subset of patients with RF negativepolyarticular onset JCA compared with healthycontrols. The sensitivity ofAKA was, however,rather poor, not exceeding 50%. In addition,AKA could not be considered specific for thissubset of JCA as the percentages of serumsamples positive for AKA did not differsignificantly across JCA subgroups. In contrastwith previous reports in adult RA,16 18 20 21 23AKA status was not related to disease activityor severity. The value of our result is, however,limited by the fact that our patients were notfollowed up longitudinally.Anti-RA 33 antibodies have been previously

reported as strongly specific for the diagnosisof RA.27 In one study, however, anti-RA 33antibodies were also detected in patients withother rheumatic disorders, including in 60% ofpatients with mixed connective tissue diseaseand 25% of those with SLE (unpublisheddata). Our findings in children with SLE arethus in accordance with results in adults. Therate of anti-RA 33 positivity was significantonly in patients with RF positive polyarticularonset JCA, a fact which limits the value of thistest. This finding requires confirmationbecause of the small number of patients in thissubgroup.

In conclusion, our results show that APF,AKA, and anti-RA 33 antibodies do not

788

Autoantibodies and juvenile chronic arthritis

contribute to the diagnosis or classification ofJCA, which still relies mainly on clinicalfindings and on the presence of RF and ANA.These data are at variance with those found inadult RA serum samples, suggesting that, withthe exception of a subset of patients with RFpositive polyarticular onset JCA, JCA is aheterogeneous entity different from adult RFpositive RA.

This study was supported by a grant from a French researchassociation against rheumatism (ARP). We thank Dr A Elias,MD (Centre de transfusion, H6pital Bichat, Paris) and Mrs THaim (Laboratoire d'Immuno-rhumatologie, Faculte deMedecine Xavier Bichat-Paris 7) for their technical assistance.We are grateful to Miss B Mermillod (Centre d'informatiquehospitaliere, H6pital Cantonal Universitaire, Geneva) for herstatistical advice. Dr C Gabay was in receipt of a fellowship fromthe University Hospital of Geneva, Switzerland.

1 Arnett F C, Edworthy S M, Bloch D A, et al. The AmericanRheumatism Association 1987 revised criteria for theclassification of rheumatoid arthritis. Arthritis Rheum1987; 31: 315-24.

2 Tan E M, Cohen A S, Fries J F, et al. The 1982 revisedcriteria for the classification of systemic lupuserythematosus. Arthritis Rheum 1982; 25: 1271-7.

3 Nienhuis R L F, Mandema E. A new serum factor inpatients with rheumatoid arthritis: the antiperinuclearfactor. Ann Rheum Dis 1964; 23: 302-5.

4 Veys E M, De Keyser F, De Vlam K, Verbruggen G. Theantiperinuclear factor. Clin Exp Rheumatol 1990; 8:429-31.

5 Johnson G D, Carvalho A, Holborow E J, Goddard D H,Russell G. Antiperinuclear factor and antikeratinantibodies in rheumatoid arthritis. Ann Rheum Dis 1981;40: 263-6.

6 Inez R J, Sondag-Tschroots M, Aaij C, Smit J W, FeltkampT E W. The antiperinuclear factor. 1. The diagnosticsignificance of the antiperinuclear factor for rheumatoidarthritis. Ann Rheum Dis 1979; 38: 248-51.

7 Janssens X, Veys E M, Verbruggen G, Declercq L. Thediagnostic significance of the antiperinuclear factor forrheumatoid arthritis. Jf Rheumatol 1988; 15: 1346-50.

8 Westgeest A A A, Boerbooms A M Th, Jongmans M,Vandenbroucke J P, Vierwinden G, Van de Putte L B A.Antiperinuclear factor: indicator of more severe disease inseronegative rheumatoid arthritis. J Rheumatol 1987; 14:893-7.

9 Youinou P, Le Goff P, Dumay A, Lelong A, Fauquert P,Jouquan J. The antiperinuclear factor. I. Clinical andserologic associations. Clin Exp Rheumatol 1990; 8:259-64.

10 Vivino F B, Maul G G. Histologic and electron microscopiccharacterization of the antiperinuclear factor antigen.Arthritis Rheum 1990; 33: 960-9.

11 Meyer 0, Tauxe F, Fabregas M, Haim Th, Kahn M F.Antiperinuclear factor (APF), antikeratin (AKA), andanti-RA 33 antibodies in rheumatoid arthritis (RA)[abstract]. Clin Rheumatol 1992; 11: 130.

12 Youinou P, Pennec Y, Le Goff P, Ferec C, Morrow W J W,Le Menn G. Antiperinuclear factor in Sjogren's syndromein the presence or absence of rheumatoid arthritis. ClinExp Rheumatol 1984; 2: 5-9.

13 Westgeest A A A, Van Loon A M, Van der Logt J T M, Vande Putte L B A, Boerbooms A M Th. Antiperinuclearfactor, a rheumatoid arthritis specific autoantibody: itsrelation to Epstein-Barr virus. J Rheumatol 1989; 16:626-30.

14 Feltkamp T E W, Berthelot J M, Boerbooms A M T, et al.Interlaboratory variability of the antiperinuclear factor(APF) test for rheumatoid arthritis. Clin Exp Rheumatol1993;11: 57-9.

15 Young B J J, Mallya R K, Leslie R D G, Clark C J M,Hamblin T J. Anti-keratin antibodies in rheumatoidarthritis. BM3r 1979; 2: 97-9.

16 Vincent C, Serre G, Lapeyre F, et al. High diagnostic valuein rheumatoid arthritis of antibodies to the stratumcomeum of rat oesophagus epithelium, so called'antikeratin antibodies'. Ann Rheum Dis 1989; 48:712-22.

17 Montecucco C, Caporali R, Negri C, et al. Antibodies frompatients with rheumatoid arthritis and systemic lupuserythematosus recognize different epitopes of a singleheterogeneous nuclear RNP core protein. Possible role ofcrossreacting antikeratin antibodies. Arthritis Rheum1990; 33: 180-6.

18 Hajiroussou V J, Skingle J, Gillet A P, Webley M.Significance of antikeratin antibodies in rheumatoidarthritis.JRheumatol 1985; 12: 57-9.

19 Kataaha P K, Mortazavi-Milani S M, Russell G, HolborowE J. Anti-intermediate filament antibodies, antikeratinantibody, and antiperinuclear factor in rheumatoidarthritis and infectious mononucleosis. Ann Rheum Dis1985; 44: 446-9.

20 Kirstein H, Mathiesen F K. Antikeratin antibodies inrheumatoid arthritis. Scand J Rheumatol 1987; 16:331-7.

21 Ordeig J, Guardia J. Diagnostic value of antikeratinantibodies in rheumatoid arthritis. J Rheumatol 1984; 11:602-4.

22 Quisimorio F P Jr, Kaufman R L, Beardmore T, MonganE S. Reactivity of serum antibodies to the keratin layerof rat esophagus in patients with rheumatoid arthritis.Arthritis Rheum 1983; 26: 494-9.

23 Paimela L, Gripenberg M, Kurki P, Leirisalo-Repo M.Antikeratin antibodies: diagnostic and pronostic markersfor early rheumatoid arthritis. Ann Rheum Dis 1992; 51:743-6.

24 Youinou P, Le Goff P, Colaco C B, et al. Antikeratinantibodies in serum and synovial fluid show specificity forrheumatoid arthritis in a study of connective tissuediseases. Ann Rheum Dis 1985; 44: 450-4.

25 Scott D L, Delamere J P, Jones L J, Walton K W.Significance of laminar antikeratin antibodies to ratoesophagus in rheumatoid arthritis. Ann Rheum Dis 1981;40: 267-71.

26 Steiner G, Hartmuth K, Skriner K, et al. RA 33 isindistinguishable from the A2 protein of theheterogeneous nuclear ribonucleoprotein [abstract]. ClinRheumatol 1992; 11: 134.

27 Hassfeld W, Steiner G, Hartmuth K, et al. Demonstrationof a new antinuclear antibody (anti-RA 33) that is highlyspecific for rheumatoid arthritis. Arthritis Rheum 1989; 32:1515-20.

28 Nesher G, Moore T L, Grisanti M W, El-Najdawi E,Osborn T G. Antiperinuclear factor in juvenilerheumatoid arthritis. Ann Rheum Dis 1992; 51: 350-2.

29 Von Essen R, Ylijoki H, Savolainen A, Haapasaari J, OksalaU-M. Antiperinuclear factor in chronic juvenile arthritis.Ann Rheum Dis 1992; 51: 1264.

30 Ansell B M. Heberden oratio: chronic arthritis in childhood.Ann Rheum Dis 1978; 37: 107-20.

31 Prieur A M, Listrat V, Dougados M, Amor B. Evaluationof the ESSG and the Amor criteria for juvenilespondylarthropathies USA). Study of 310 consecutivechildren referred to one pediatric rheumatology center.Arthritis Rheum 1990; 33: D159.

32 Laemmli U K. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 1979;227: 680-5.

33 Towbin H, Staehelin T, Gordon J. Electrophoretic transferof proteins from polyacrylamide gels to nitrocellulosesheets. Proc NatlAcad Sci USA 1979; 76: 4350-4.

34 Steinbrocker 0, Traeger C H, Batterman R C. Therapeuticcriteria in rheumatoid arthritis. _JAMA 1949; 140:659-62.

35 Moore T L, Osbom T G, Weiss T D, et al. Autoantibodiesin juvenile arthritis. Semin Arthritis Rheum 1984; 13:329-36.

36 Youinou P, Seigneurin J M, Le Goff P, Dumay A, VicariotM, Lelong A. The antiperinuclear factor. II. Variability ofthe perinuclear antigen. Clin Exp Rheumatol 1990; 8:265-9.

37 Hoet R M A, Boerbooms A M Th, Arends M, Ruiter D J,Van Venrooij W J. Antiperinuclear factor, a markerautoantibody for rheumatoid arthritis: colocalisation ofthe perinuclear factor and profilaggrin. Ann Rheum Dis1991;50:611-8.

38 Nesher G, Moore T L. Antiperinuclear factor in chronicjuvenile arthritis. Ann Rheum Dis 1992; 51: 1264.

789