British Journal of Rheumatology 1996;35:534-541

PREDOMINANCE OF IgM ANTI-U1RNP ANTIBODIES IN PATIENTSWITH SYSTEMIC LUPUS ERYTHEMATOSUS

P. G. VLACHOYIANNOPOULOS, A. GUIALIS,* A. G. TZIOUFAS andH. M. MOUTSOPOULOS

Department of Pathophysiology, University of Athens, Medical School, 75 Mikras Asias Street, Athens 115 27 and^Institute of Biological Research and Biotechnology, The National Hellenic Research Foundation, Vassileos

Constantinou Avenue, Athens 116 35, Greece

SUMMARY

Anti-UlRNP antibodies occur in patients with mixed connective tissue disease (MCTD), systemic sclerosis (SSc), systemic lupuserythematosus (SLE) and other ill-defined connective tissue diseases. To associate the isotypes of anti-UlRNP antibodies withthe diagnosis of the disease, namely SLE or MCTD, sequential sera of patients positive for anti-UlRNP antibodies bycounterimmunoelectrophorcsis (CIE) (32 with SLE, 35 with MCTD) were tested for IgG and IgM anti-UlRNP antibodies byenzyme-linked immunosorbent assay (ELISA) using affinity-purified UlsnRNP complexes. Results from ELISA were confirmedby RNA precipitation. IgG RNA precipitation of HeLa cellular extracts was performed using the bulk of the IgG fractionremoved from each serum after binding to protein A-Sepharose beads. IgM RNA precipitation was carried out on the IgMfraction of the serum bound to protein A-Scpharose-rabbit anti-human IgM immune complexes. RNAs were electrophoresedin 10.5% acrylamide-7 M urea gels and detected with the silver stain. ELISA showed that all sera were positive to IgGanti-UlRNP, while 12 of the 35 MCTD and 21 of the 32 SLE patients possessed IgM anti-UlRNP (P < 0.025). IgManti-UlRNP reactivity was found during the follow-up in 20% of 44 sera from 17 MCTD patients and 68% of 112 sera from23 SLE patients (P < 0.0001). IgG from all the sera precipitated UlRNPs. Eight of the MCTD sera also precipitated U2RNPsand 14 of the SLE sera U2 and/or U4/U6, U5 RNPs. IgM from MCTD sera did not precipitate URNPs, while IgM from SLEsera precipitated predominantly UlRNPs. These data suggest that IgM anti-UlRNP antibodies occur predominantly in patientswith SLE. The occurrence of IgG anti-UlRNP without IgM is more frequent in MCTD.

KEYWORDS: Anti-UlRNP, Anti-Sm, Autoantibodies, Isotypes, SLE, MCTD.

THE term 'mixed connective tissue disease' (MCTD) [1]was introduced to describe patients with a combinationof features, such as non-erosive arthritis [indicative ofsystemic lupus erythematosus (SLE)] [2], Raynauds'phenomenon, puffy hands, sclerodactyly, oesophagealhypomotility [indicative of systemic sclerosis (SSc)] [3]and muscle involvement (indicative of myositis [4]).The sera of these patients recognized a nuclearantigen extracted in isotonic buffers [5] which wasfound to be a physical complex of two antigens:one sensitive to ribonuclease and trypsin, designated'nuclear RNA-protein' (nRNP) and the other resistant,designated Sm [6]. The sera of MCTD patients reactedexclusively against the nRNP antigen and, therefore,the anti-nRNP antibodies were considered as a distinctserological feature associated with MCTD [7]. Progressin molecular biology revealed that both nRNP and Smare complexes of small nuclear RNAs with proteins,in other words small nuclear ribonucleoproteins(snRNP), and that anti-nRNP antibodies recognizeonly components unique to UlsnRNP, while anti-Smantibodies recognize common proteins to Ul, U2,U4/U6, U5 snRNPs [8-10]. The UlsnRNP specificproteins are a 70 kDa polypeptide as well aspolypeptides A (34 kDa) and C (22 kDa) which in

Submitted 28 June 1995; revised version accepted 6 December1995.

Correspondence to: H. M. Moutsopoulos, Department ofPathophysiology, University of Athens Medical School, 75 MikrasAsias Street, Athens 115 27, Greece.

complex with U1RNP form the U1RNP antigen.The Sm proteins recognized by anti-Sm antibodiescorrespond to B' (29 kDa), B (28 kDa), D (16 kDa) andoccasionally to the peptides E (12 kDa), F (11 kDa)and G (9 kDa) [10-12]. In terms of small linearepitopes, a considerable cross-reactivity exists betweenpeptides B'/B, A and C of the Sm and U1RNP antigens[13]. Antibodies to the 70 kDa polypeptide are mainlyof the IgG class, while antibodies to the A polypeptideare of the IgM class [14]. IgG anti-70 kDa and IgManti-B'/B have been associated with arthralgias andRaynaud's phenomenon [15]. In a recent report, IgManti-A and anti-D polypeptides were associated withIgM anti-ds DNA in the sera of patients with SLE [16].However, the previous reports were based on studiesusing Western blotting and therefore the polypeptideswere denatured. In this report, an effort wasundertaken to detect the isotypes of antibodies againstsnRNPs by using the polypeptides of the Sm andU1RNP antigens in their natural configuration incomplexes with the respective RNAs, and to associatethese isotypes with the disease status.

MATERIALS AND METHODS

Clinical diagnoses and serum samplesSequential sera from consecutive patients with

autoimmune connective tissue diseases, positive foranti-UlRNP antibodies either alone or associated withother precipitating antibodies by counterimmuno-electrophoresis (CIE) [17], were tested for the presence

© 1996 British Society for Rheumatology534

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

VLACHOYIANNOPOULOS ET AL.\ IgM ANTI-UlsnRNP in SLE 535

of IgG and IgM anti-UlRNP antibodies by a sensitiveand specific enzyme-linked immunosorbent assay(ELISA) [18]. The IgG and IgM reactivities of the seraagainst Ul and Sm-snRNPs were further characterizedby RNA immunoprecipitation [19]. Sera from 32patients with SLE and 35 with MCTD were tested. Thediagnosis of SLE was established according to the1982 revised criteria of the American RheumatismAssociation for the diagnosis of SLE [2]. The diagnosisof MCTD was established when the patients presentedanti-UlRNP antibodies by CIE plus three of thefollowing four clinical manifestations: oedema of thehands, synovitis, myositis, Raynaud's phenomenonand acrosclerosis [20].

Affinity purification of UIRNP from calfthymus extractSera positive for antibodies to UIRNP by CIE [17]

were further characterized by RNA immunoprecipita-tion, using HeLa cellular extracts [19]. A serum withreactivity against only UIRNP by RNA precipitationrecognizing the 70 kDa specific UIRNP antigen byimmunoblotting [21] was chosen for the preparation ofan immunoaffinity column for UIRNP purification(Fig. 1). A protein A-Sepharose column (Pharmacia)was used to separate IgG from the monospecificanti-UlRNP serum, according to standard protocols[22]. Briefly, 80 mg of the above IgG were coupledto cyanogen bromide-activated Sepharose 4B(Pharmacia) at 5 mg IgG/ml of gel according to the

manufacturer's instructions. The remaining active siteswere blocked by 0.2 M glycine. Purification of UIRNPwas performed as previously described [18]. Briefly,antigen was extracted by homogenizing 100 g of calfthymus in 300 ml of 0.15 M phosphate-buffered saline(PBS) (pH 7.2) and partially purified by ammoniumsulphate fractionation. The 30-60% ammoniumsulphate fraction of the extract was passed through thecolumn and after extensive washing with PBS (pH 7.2)containing 0.5 M NaCl, the antigen bound to thecolumn was eluted with the 3 M guanidine-HCl(pH 7.2) and dialysed against PBS (pH 7.2).

ELISAA solid-phase ELISA was developed to detect the

antibody class (IgG and IgM) with anti-UlRNPspecificity [18]. Antigen [20 /ig/ml in PBS (pH 7.2)] wascoated onto 96 well polystyrene cuvettes pre-coatedwith poly-L-lysine (Sigma, Poole, Dorset) (50 /ig/ml inPBS). After washing with PBS, the non-specific bindingsites were blocked with 10% bovine serum in PBS at37°C. One hundred microlitres of sera diluted in 10%bovine serum in PBS (1:1000 in the case of IgG and1:100 in the case of IgM) were added in duplicates andincubated for 2 h after washing, antibody-boundantigen was detected with alkaline phosphatase-conjugated anti-human y or n chain specific antisera(SERA-LAB, Sussex). Finally, the plates were washed,substrate solution (p-nitrophenyl phosphate, 1 mg/ml)

IS

I

P-SS —rjnat

- U 2

—U4

!-U5

oI

3

I

- 7 O k

- L a

A. B.

Fio. I.—The serum GE used for the preparation of the immunoaffinity column for UIRNP purification was tested by RNA precipitation (A)as well as by immunoblot (B). It was shown that this serum precipitates U1RNA, but not U2, U4, U5 and U6 RNAs. The same is shownwith the prototype anti-UlRNP serum. Furthermore, when the same sera were tested by immunoblotting using nuclear HeLa cell extract (B),they demonstrated reactivity only to the 70 kDa peptide of the UIRNP antigen.

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

536 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 6

in diethanolamine buffer (pH 9.8) was added to eachwell and the absorbance was read at 405 nm. Thewashing buffer was PBS containing 0.05% Tween(PBS-Tween). Incubations took place at 37°C. Foreach ELISA (IgG and IgM), a serum with a strongreactivity, as detected in preliminary experiments, waschosen for the standard curve. The prototype serum forthe IgM ELISA was negative for rheumatoid factor(RF) activity. Furthermore, sera positive for RF weretested after absorption of RF by incubating the serawith human heat-aggregated IgG; when IgM ELISAwas repeated, IgM anti-UlRNP reactivity was reducedby only 0-10% of the original values. Sera wereconsidered positive when the optical density was higherthan the mean value of normals (n = 80) plus 3 S.D.

ImmunoblottingThis method was used to test the antigenicity of

our antigen preparation. Affinity-purified Ul-snRNPcomplexes, expected to contain Sm/UIRNP antigen,were run on a 12% polyacrylamide slab gel [23] and theproteins were transferred to nitrocellulose membranesusing Western blotting [21]. Strips cut from thenitrocellulose reacted with 1:100 diluted serum and thebinding bands were labelled by horseradish peroxidase(HRPO) anti-human IgG (Sigma) in a dilution 1:1000to detect IgG anti-Sm/UlRNP reactivity. Five per centnon-fat milk in Tris-buffered saline (TBS)-Tween 20was used as a blocking agent, as well as for the dilutionof the sera. Incubation of the nitrocellulose stripsof the sera was performed for 4 h and incubationwith anti-human IgG-HRPO conjugates for 1.5 h.A 4-chloro-l-naphthol (Sigma) substrate was finallyadded to detect binding of the conjugate.

RNA precipitation assays using IgG, IgM autoantibodyfractions

To discriminate between IgG and IgM autoantibodyreactivity in patients' sera, the sensitive and accurateRNA precipitation assay was applied to Ig fractionshighly enriched for either IgG or IgM antibodyclasses. This was accomplished by making use of thedifferential binding ability of the different Ig classes forprotein A molecules. As is well established, human IgGmolecules (with the exception of the IgG3 subclass)bind strongly to protein A, in contrast to the weakbinding exhibited by IgM and IgA antibody classes[22].

Initially, the bulk of the IgG molecules were removedfrom the serum by incubating 10 /il of serum with 3 mgprotein A-Sepharose beads (Pharmacia, Sweden) inNET-2 buffer [150 mM NaCl, 10 HIM Tris-HCl(pH 7.5), 0.05% NP-40] for 1.5 h with gentle rocking at4°C. Following a brief spin and saving of the proteinA-IgG complexes in the pellet (fraction a), thesupernatant of the serum was applied to a second cycleof incubation to largely deplete the remaining IgGmolecules (fraction b). The supernatant of the secondcycle was then allowed to react, in a third cycle ofincubation, with the complex of protein A-rabbitanti-human IgM, prepared by pre-incubating 3 mg

of protein A with 10 fi\ of rabbit anti-human IgM(fraction c) (13.8mg/ml; Dako Co., Denmark). Thebeads of protein A/antibody complexes from the threecycles of incubation were washed three times in NET-2buffer and applied to RNA precipitation assays usingHeLa total cellular extracts as previously described[19]. An aliquot of each immunoprecipitate tested inRNA precipitation was saved for testing the presenceof IgG or IgM on a Western blot using rabbitanti-human IgM alkaline phosphatase-conjugatedantisera. This experiment revealed that fraction (c)contained only IgM, but not IgG (data not shown). Allimmune pellets obtained were then phenol extractedand the purified RNA molecules were resolved on 7%urea-10% polyacrylamide gels. RNA on the gel wasvisualized by silver staining.

Statistical analysisContingency tables were used where indicated.

RESULTSDiagnoses, serological findings and evolution of patients

There were 32 patients with SLE and 35 withMCTD. The majority of the SLE patients fulfilled fourof the American Rheumatism Association (ARA) SLEcriteria (22 patients), seven fulfilled five, two patientssix and one patient eight of the criteria. The patientswith SLE developed full-blown disease (four or moreof the ARA criteria) 1-15 yr after the onset of the firstsymptom compatible with the disease. None of theMCTD patients fulfilled criteria for SLE. Six of the 32patients also fulfilled criteria for MCTD; three ofthem presented synovitis, oedema of the hands andRaynaud's phenomenon, two presented synovitis,myositis and Raynaud's phenomenon, and one patientsynovitis, oedema of the hands and myositis.Autoantibodies to Ro (SSA), La (SSB) and Smdetected by CIE, RF and anti-dsDNA antibodies inboth groups are presented in Table I.

IgG and IgM isotypes of anti-UlRNP detected byELISA

Since purified UlsnRNP fraction contained boththe UlsnRNP (70 kDa, A and C polypeptides) aswell as the Sm antigens (B/B' and D polypeptides),which are common to all UsnRNPs (Ul, U2, U4/U6,U5), ELISA cannot discriminate between antibodiesof anti-UlRNP and anti-Sm specificities. To avoidconfusion, we shall refer to the affinity-purified

TABLE IAutoantibodies in patients with MCTD and SLE

Autoantibodies in the oldestavailable serum samplesAnti-DNAAnti-RoAnti-LaAnti-SmRF

MCTD(n = 35)n (V.)2(6)3(9)0(0)2(6)

11 (31)

SLE(n -= 32)n (%)

18 (tt)10(31)2(6)6(19)

12(37)

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

VLACHOYIANNOPOULOS ET AL.\ IgM ANTI-UlsnRNP in SLE 537

UlsnRNP fraction applied in ELISA assays as thesnRNP antigen. Nonetheless, since, as shown inTable I, only 6% of MCTD sera and 19% of SLE seraalso contained anti-Sm antibodies, we assumed thatELISA data reflected by and large the specificityagainst UlsnRNP components. ELISA was used toscreen the sera of the patients for IgG and IgManti-Ul snRNP reactivity against affinity-purifiedUlsnRNP components. In a first experiment, the oldestserum samples from each patient that were available inour laboratory were tested. Twelve out of the 35MCTD patients were positive for IgM anti-Ul snRNPantibodies vs 21 of the 32 SLE patients (P < 0.025)(Fig. 2). On the contrary, all the patients were positivefor IgG anti-Ul snRNP antibodies.

Sera from patients with SLE and MCTD which hadbeen collected during their follow-ups were withdrawnfor testing with an order of one sample every 9 monthsduring their regular visits to the out-patient clinic. Atotal of 44 sera from 17 MCTD patients and 112 serafrom 23 SLE patients were collected. The IgG and IgMreactivity to UlsnRNP antigen was detected in thesesera by ELISA. Nine of the 44 MCTD sera (20%) vs77 of 112 SLE sera (68%) were positive for IgMantibodies to U1RNP (P < 0.0001). In fact, temporaryincreases in IgM anti-Ul snRNP reactivity above theupper limit of normals occurred in four patients withMCTD. On the contrary, a persistent increase in IgM

anti-Ul snRNP reactivity above control levels occurredin 18 of the 23 SLE patients. Figures 3 and 4 representthe IgM anti-Ul RNP activity in a group of eight SLEand MCTD patients. These patients were collected asbeing representative of each group in terms of thefollowing: (a) the SLE patients did not presentMCTD related features other than synovitis, whichoccurred in three of them; the MCTD patientspresented a stable clinical feature for years non-evolving to SLE or scleroderma; (b) the patients ofboth groups had a long follow-up; (c) from the initialpatients' sera, those with highest IgM anti-Ul snRNPactivity were used; (d) sequential serum samples wereavailable in our serum bank.

As is shown, five of the eight MCTD patients areinitially negative for IgM anti-Ul snRNP antibodiesand tend to be negative thereafter, with temporaryincreases in the IgM anti-Ul snRNP titre above normalvalues. On the other hand, five sera of SLE patients arepersistently positive for IgM anti-Ul RNP antibodies,three were initially negative and became positive overa period of 2-3 yr.

The RF activity of the sera does not interfere with theirIgM anti-UlRNP activity

Fifteen of the 23 patients positive for RF werealso positive for IgM anti-Ul snRNP antibodies.In addition, 18 of 44 patients negative for RF were

. J

1.4

M

U

U

3 0.9

{ 04

0.7

06

0.4

04

0*

0,1

0

*

I

i

l•

1

••••i

JSSSBb

ItjM

JL»!*it

•

•

:

i

V

i—lo« laM

MCTD sera SLE sera NORMAL sera

FIG. 2.—The absorbance values of anti-Ul RNP sera of SLE and MCTD patients as well as normal human sera tested for IgG(G) or IgM(M)anti-UlRNP antibodies.

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

538 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 6

10000 =Abaorbance (405nm).

1000:

100:

18 27 36 45Time (months)

54

FIG. 3.—IgM anti-UlRNP reactivity of sequential sera from eight patients with SLE.

also positive for IgM anti-UlsnRNPs. By usingcontingency tables, no statistically significant associ-ation between the above variables was observed(X7 = 3.57, P > 0.05). To further strengthen the aboveobservation, we eliminated serum RF by absorptionwith human heat-aggregated IgG as RF activity wasnot further detectable and the ELISA was repeated.

The absorbance values were decreased from 0 to 10%of the initial values in all the sera.

Characterization of the IgM anti-UlRNP reactivity byRNA precipitation

Although ELISA helped us to show the presence, inhigh frequency, of IgM anti-snRNP reactivity in sera

600

500

Absorbance (405nm)

27 36Time (months)

45

FIG. 4.—IgM anti-UlRNP reactivity of sequential sera from eight patients with MCTD.

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

VLACHOYIANNOPOULOS ET AL.\ IgM ANTI-UlsnRNP in SLE 539

6

1T a b c a b c | a b e T a b c a b c a b c

i i i i i i i i i j i i i I I i | i i i

!

Fio. 5.—RNA precipitation assays using HeLa cellular extracts were applied to eight sera from patients suffering from MCTD (1-5) or SLE(6-8). Each of the sera has gone through two cycles of incubation with protein A-Sepharose beads (a and b) and, subsequently, through a thirdcycle using protein A-Sepharose-rabbit anti-human IgM immune complexes (c). Therefore, fragment c contains the IgM antibody eluted fromthe protein A-Sepharose-rabbit anti-human IgM matrix. T refers to the RNA content of an aliquot (1/5) of the HeLa extract used in eachimmunoprecipitation assay. Sera 1, 2, 4, 5 and 7 were negative in ELISA for IgM anti-UlRNP antibodies. Sera 3, 6 and 8 were positive inELISA for IgM anti-UlRNP antibodies and these sera also demonstrated IgM reactivity against the UlRNP particle in immunoprecipitationassay. However, serum 8 also demonstrates weak reactivity against U2 and U4 RNPs in the IgM immunoprecipitation assay.

of SLE when compared to MCTD patients, it couldnot provide information on the specific antigeniccomponent (Sm vs UlsnRNP) involved. Therefore,sera from SLE and MCTD were further characterizedregarding their IgM, IgG reactivity by RNA precip-itation of HeLa cellular extracts. For each serum, threeimmunoprecipitates were analysed with respect toUsnRNA content. Immunoprecipitates (a) and (b)referred to the snRNA species bound by the bulk andthe residual IgG class, respectively. On the other hand,immunoprecipitate (c) should contain snRNP speciesrecognized solely by IgM class bind to the proteinA-rabbit anti-human IgM complex. All the sera werepositive for IgG anti-UlRNPs, as was shown by IgGRNA immunoprecipitation. However, eight of theMCTD sera also precipitated U2RNPs, while 14 of theSLE sera also precipitated U2RNPs and eight of themin addition precipitated U4/U6, U5 RNPs. Figure 5indicates the reactivity of each of the above samples(obtained from five MCTD and three SLE sera) againstsnRNPs by RNA immunoprecipitation. As is shown,four out of the five MCTD sera have IgG anti-UlRNPspecificity in the absence of any detectable IgMreactivity. Serum no. 3 was the only one with a veryweak anti-Sm specificity. This serum also precipitatesUlRNP by IgM antibodies; while the rest of the seraare negative for IgM anti-UlRNPs. The resultsobtained by ELISA showed that of the above five serathis serum was weakly positive for IgM anti-snRNPreactivity. Two out of the three SLE sera shown inFig. 5 are positive for IgM anti-snRNPs, and these arethe sera which also have anti-Sm specificity, as is clearby the presence of U2, U4/U6, U5 RNPs which areprecipitated by their IgG antibodies. The one SLEserum (no. 7) with IgG anti-UlRNP alone does notcontain any detectable specificity of the IgM class.

Thus, anti-UlRNP reactivity alone in either MCTDor SLE is of the IgG isotype. On the contrary,anti-UlRNP antibodies when existing together with

anti-Sm can be of both the IgG and IgM isotype.Anti-Sm reactivity in SLE can also be of the IgMisotype. The above results are compatible with thoseobtained by the ELISA.

DISCUSSIONThe present report indicates that among patients

with precipitating antibodies to snRNP components,the patients with SLE possess anti-snRNP antibodiesof the IgM class, while the patients with MCTD possessanti-snRNP (mainly anti-UlRNP) antibodies withpredominance of the IgG class. The above informationwas based on results obtained by ELISA using purifiedUlRNP antigen as the antigen source. Since thepurified antigen also contained Sm antigenic deter-minants [6, 8, 12, 18], it was important to discriminatethe reactivity of the sera attributed to the anti-UlRNPand not to the anti-Sm antibodies which probably existin some of the sera [18,24]. RNA precipitation, inagreement with CIE, indicated that the majority ofboth MCTD and SLE sera precipitated only UlRNAs,and thus they contain antibodies to the UlRNPspecific components. This was not surprising sinceanti-Sm antibodies are rare in Greek [24] and otherEuropean populations [25] with SLE. Therefore, theresults obtained by ELISA indicated mainly anti-U1RNP reactivity of the sera tested. Although thisfinding was clear for IgG anti-UlRNP reactivity, theIgM anti-UlRNP reactivity needed substantial proof.To discriminate between IgG and IgM autoantibodyto snRNP in patients' sera, we applied the RNAprecipitation assay as an extremely sensitive way ofidentifying anti-UlRNP and anti-Sm specificities. TheRNA precipitation assay confirmed that antibodies ofthe IgM class detected by ELISA really existed, sincethese antibodies could precipitate snRNPs in theabsence of the IgG fraction. Furthermore, the IgMfraction of the anti-Sm sera recognized predominantlyUlRNP, although these sera, by their IgG fraction,

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

540 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 6

also recognized all UsnRNPs (U2, U4/U6, U5 RNPs).In fact, the IgM anti-UlRNP reactivity occurredmainly in the sera with anti-Sm reactivity.

The above results are in line with previous resultsindicating that the 70 kDa polypeptide of the UlRNPantigen is mainly recognized by IgG antibodies [14, 15],while the A polypeptide is mainly recognized by IgMantibodies [15, 16]. The IgM reactivity to peptides A ofthe U1RNP and D of the Sm antigens has also beenassociated with anti-dsDNA antibodies in SLE seranegative for precipitating anti-UlRNP or anti-Smantibodies [16]. A considerable cross-reactivity ofanti-A and anti-D with anti-dsDNA antibodies hasbeen described [26], and can explain the aboveassociation. These reports were based on Western blotexperiments, where the denaturation of antigenicpeptides is inevitable. Therefore, small linear epitopes,but not all the conformational epitopes, could berecognized by the sera tested. A considerablecross-reactivity between peptides B'/B and D of the Smantigen and A of the UlRNP antigen has beendescribed [13]. This cross-reactivity can be attributedto a proline-rich motif such as the octapeptidePPPGMRPP, which is repeated several times withminor alterations within peptides B'/B, A and C[12, 14, 27].

In the present report, an effort was made to answerthe question of whether the IgM anti-UlRNPreactivity of the sera can also be explained by IgMantibodies recognizing the native forms of the UlRNPrelated peptides. The ELISA was used as a screeningtest and revealed a considerable number of SLEpatients to be positive for IgM antibodies to UlsnRNPcomponents, in contrast to MCTD patients. The RNAprecipitation confirmed the IgM reactivity of the seraand showed that this reactivity is turned on againstmainly UlRNP. Therefore a continuous, antigen-driven, T- and B-cell activation occurs in SLE,in contrast to MCTD. In this regard, the IgGanti-UlRNP activity observed in MCTD can beconsidered as a sign of a previous and never renewedantigen-driven activation. The fact that the sera of SLEpatients also possess IgG anti-UlRNP antibodiesindicates that antibody class switching does occur inSLE. The IgM reactivity when it occurs, is turned onagainst mainly UlRNPs, even among sera with IgGanti-Sm snRNPs. This finding indicates a stricthierarchy in the evolution of autoimmune reactivity tosnRNPs. Thus, UlRNP seems to be the target of theinitial response and the Sm snRNPs the targets ofsubsequent responses. There is experimental evidencethat the B'/B peptides of the Sm antigen carry at leastone determinant which triggers immune responses thatin a later stage spread to other epitopes related tosnRNPs, as well as to dsDNA [28].

Although many mechanisms for the productionof certain autoantibodies remain unresolved, it isbecoming clear that intact UlRNP is a majorimmunogen for patients with SLE and MCTD, andthat SLE is a disease where a continuous activation ofthe immune system with snRNPs operates.

REFERENCES

1. Sharp G, Irwin ES, Tan EM, Gould RG, Holman JR.Mixed connective tissue disease—an apparently distinctrheumatic disease syndrome associated with a specificantibody to an extractable nuclear antigen. Am J Med1972;52:148-59.

2. Tan EM, Cohen AS, Fries JF et al. The 1982 revisedcriteria for the classification of systemic lupus eryth-ematosus. Arthritis Rheum 1982^5:1271-7.

3. Massi AT, Rodnan GP, Medsger TA, Altman RD,D'Agnelo WA, Fries JF. Subcommittee for SclerodermaCriteria of the American Rheumatism AssociationDiagnostic and Therapeutic Criteria Committee: prelimi-nary criteria for the classification of systemic sclerosis(scleroderma). Arthritis Rheum 1980;23:581-90.

4. Bohan A, Peter JB. Polymyositis and dermatomyositis.N Engl J Med 1975^92:344-7.

5. Holman HR. Partial purification and characterization ofan extractable nuclear antigen which reacts with SLEsera. Ann NY Acad Sci 1965;124:800-6.

6. Mattioli M, Reichlin M. Physical association of twonuclear antigens and mutual occurrence of theirantibodies: the relationship of the Sm and RNA protein(Mo) systems in SLE sera. J Immunol 1973;ll(hl318-24.

7. Sharp GC, Irvin W, May CM et al. Association ofantibodies to ribonucleoprotein and Sm antigens withmixed connective tissue disease, systemic lupus erythe-matosus and other rheumatic diseases. N Engl J Med1976;265:1149-54.

8. Lerner MR, Steitz JA. Antibodies to small nuclear RNAscomplexcd with proteins are produced by patients withsystemic lupus erythematosus. Proc Natl Acad Sci USA1979;76:5495-9.

9. Steitz A. Autoimmune probes for mammalian geneexpression. Harvey Led 1986;80:39-46.

10. Petterson I, Hinterberger M, Mimori T, Gottlieb E,Steitz JA. The structure of mammalian small nuclearribonucleoproteins: identification of multiple proteincomponents reactive with anti-(Ul) ribonucleoproteinand anti-Sm autoantibodies. / Biol Chem 1984;259:5907-14.

11. Habets WJ, Bcrden JHM, Hoch SO, Van Venrooij WJ.Further characterization and subcellular localization ofSm and Ul ribonucleoprotein antigens. Eur J Immunol1985;15:992-6.

12. Habets WJ, Sillekens PTG, Hoet MH et al. Small nuclearRNA associated proteins are immunologically related asrevealed by mapping of autoimmune reactive B cellepitopes. Proc Natl Acad Sci USA 1989;86:4674-8.

13. Ohosone Y, Mimori T, Fujii T et al. Autoantigenicepitopes of the B polypeptide of Sm small nuclear RNPparticles. Arthritis Rheum 1992^5:960-6.

14. Williams DG, Stocks MR, Charles PJ, Maini RN.Antibody class differences are detected in anti-nRNP andanti-Sm antibodies directed against distinct antigensubsets. Rheum Int 1986;6:189-94.

15. Lundberg I, Nyman U, Pettersson I, Hedfors E. Clinicalmanifestations and anti-(Ul)snRNP antibodies: aprospective study of 29 anti-RNP antibody positivepatients. Br J Rheumatol 1992̂ 31:811—6.

16. Zhang W, Reichlin M. IgM anti-A and D SnRNPproteins and IgM anti-dsDNA are closely associated inSLE sera. Clin Immunol Immunopathol 1995;74:70-5.

17. Kurata N, Tan EM. Identification of antibodies tonuclear antigens by counterimmunoclectrophoresis.Arthritis Rheum 1976;19:564.

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from

VLACHOYIANNOPOULOS ET AL.: IgM ANTl-UlsnRNP in SLE 541

18. Maddison PG, Skinner RP, Vlachoyiannopoulos PG,Brennand DM, Hough D. Antibodies to nRNP, Sm,Ro(SSa) and La(SSB) detected by ELISA: theirspecificity and inter-relations in connective tissue diseasesera. Clin Immunol Immunopathol 1985;62:337—41.

19. Guialis A, Dangli A, Sekeris CE. Distribution of snRNPcomplexes in rat liver nuclear extracts: Biochemicaland immunochemical analysis. Mol Cell Biochem1987;76:147.

20. Alarcon-Segovia D, Villarreal M. Qassification anddiagnostic criteria for mixed connective tissue diseases.In: Kasukawa R, Sharp GC, eds. Mixed connectivetissue diseases and anti-nuclear antibodies. Amsterdam:Elsevier, 1987:33-40.

21. Laemmli MK. Clearance of structural protein during theassembly of the bacteriophage T4. Nature 1970;227:680-5.

22. Hudson L, Hay RF. Practical immunology, 2nd edn.Oxford: Blackwell Scientific Publications, 1980:222.

23. Towbin H, Staehelin T, Gordon J. Electrophoretictransfer of proteins from polyacrylamide gels tonitrocellulose sheets. Procedures and applications.Proc Natl Acad Sci USA 1979;76:4350-4.

24. Vlachoyiannopoulos PG, Karassa FB, Karakostas KX,Drosos AA, Moutsopoulos HM. Systemic lupuserythematosus in Greece. Clinical features evolution andoutcome: a descriptive analysis of 292 patients. Lupus19?3;2:303-12.

25. Cervera R, Khamashta MA, Foint J et al. andthe European Working Party on Systemic LupusErythematosus. Clinical and immunologic patterns ofdisease expression in a cohort of 1000 patients. Medicine1993;72:113-24.

26. Reichlin M, Martin A, Taylor-Albert E et al. Lupusautoantibodies to native DNA CTOSS react with theA and D snRNA polypeptides. / Clin Invest 1994;93:443-9.

27. Rokeach L, Jannatipour M, Hoch S. Heterologousexpression and epitope mapping of a human smallnuclear ribonucleoprotein-associated Sm-B'/B auto-antigen. J Immunol 1990;144:1015-22.

28. James JA, Gress T, Scofield RH, Harley JB. Immuno-globulin epitope spreading and autoimmune disease afterpeptide immunization: Sm B/B'-derived PPPGMRPPand PPPGIIRGP induce spliceosome autoimmunity.J Exp Med 1995;181:453-61.

by guest on July 13, 2011rheum

atology.oxfordjournals.orgD

ownloaded from