Autoantibodies to DEK oncoprotein in systemic lupus erythematosus (SLE)

Abstract

Autoantibodies against the transcriptional DEK protein have been considered characteristic of the pauciarticular onset subtype of juvenile rheumatoid arthritis (JRA) associated with iridocyclitis in young girls. In this study we investigated the presence of anti-DEK autoantibodies in the sera of 288 patients with SLE using a recombinant DEK protein as autoantigenic target. Thirty sera (10·4%) were positive against DEK protein by immunoblotting. Patients with anti-DEK autoantibodies show a lower frequency of cutaneous manifestation, exhibit more frequently certain markers of a chronic inflammatory status like anaemia and positivity for C-reactive protein, as well as a higher frequency of anti-double-stranded DNA autoantibodies. In contrast to JRA patients positive for anti-DEK autoantibodies, no association with erosive arthritis nor iridocyclitis were found in SLE. In conclusion, our results show that 10·4% of SLE patients from our area show antibodies against DEK protein, although this feature did not clearly establish a clinical subset of the disease.

INTRODUCTION

DEK is an oncogene that was first identified in a fusion with the CAN nucleoporin protein in a specific subtype of acute myelogenous leukaemia with translocation (6;9), that involves breakpoints in gene dek on chromosome 6 and gene can on chromosome 9 and induces the formation of a dek–can fusion gene [1]. The function of DEK has recently been identified as a site-specific DNA binding protein that is involved in transcriptional regulation and signal transduction [2]. DEK has been shown to be an autoantigen, and anti-DEK autoantibodies have been found associated with pauciarticular onset juvenile rheumatoid arthritis (JRA) with iridocyclitis affecting young girls [3,4]. Autoantibodies to DEK have also been detected in a patient with SLE and sarcoidosis [5]. More recently, the coexistence of autoantibodies to transcriptional regulation proteins DEK and ALY has been found in the serum from a patient with SLE [6]. The aim of this study was to investigate the presence of anti-DEK autoantibodies in a group of patients with SLE and to analyse the clinical and biological features of anti-DEK⁺ patients in order to clarify whether there is a subset of SLE which is associated with anti-DEK autoantibodies.

PATIENTS AND METHODS

Two hundred and eighty-eight SLE patients fulfilling the ACR 1982 revised criteria for the classification of SLE [7] were included in this study. The group comprised 253 women and 35 men. The ages of the patients ranged from 11 to 88 years (mean 41·7 years), the mean evolution time was 11·9 years (range 1–57 years), and the mean age at disease onset was 31·2 years.

Antinuclear autoantibodies were detected by indirect immunofluorescence (IIF) using conventional techniques and rat substrate (kidney, liver, stomach) (Biosystems, Barcelona, Spain) and HEp2 cell line (Labodia, Yens, Switzerland). Titres over 1/40 were considered positive. Antibodies to double-stranded DNA (dsDNA) were detected by IIF with Crithidia luciliae (Biosystems). Antibodies against extractable nuclear antigens (ENA) were tested by counter-immunelectrophoresis and double immunodiffusion. Immunoblotting was performed with the recombinant DEK antigen obtained as reported elsewhere [6]. The prototype serum (PMV) which allowed the definition of antibodies against DEK, absorbed by bacterial proteins, was used as reference for anti-DEK reactivity. Electrophoresis of bacterial lysates derived from bacteria expressing isopropyl-β-d-thiogalactopyranoside (IPTG)-induced recombinant protein was performed in a 10% polyacrylamide gel according to Laemmli [8]. Immunoblotting was performed as described by Towbin et al. [9]. Sera were studied at a 1:500 dilution in 3% non-fat dry milk in Tris-buffered saline with 0·05% Tween 20 (TTBS). Previous tests showed that the working dilution of 1:500 gives the best discrimination. Bound antibodies were detected by incubating the nitrocellulose strips with biotinylated anti-human IgG immunoglobulins from rabbit (Dako, Glostrup, Denmark) diluted 1:1000 and alkaline phosphatase-conjugated streptavidin at 1:2000. Data were analysed by χ² test with Yates’ correction, Fischer’s exact test and Bonferroni correction when appropriate. Relative risks were calculated according to Haldane’s correction of Woolf’s test.

RESULTS

Thirty sera from 288 studied (10·4%) were positive against DEK recombinant protein by immunoblotting. Of these, 26/253 (10·27%) were from women and 4/35 (11·4%) from men. The age of the DEK⁺ subjects was 47·3 ± 14 years (mean ± s.d.) and of the DEK⁻ patients was 41·3 ± 15 years, this difference being statistically significant (P = 0·04). The age of onset of clinical manifestations was 33·4 ± 14 years for the DEK⁺ subjects, and 29·1 ± 14 years for the DEK⁻ patients (NS). The time of evolution was 13·8 ± 7·9 years for positive individuals and 12·3 ± 8 years for negative individuals (NS).

Clinical and biological data were obtained in a cumulative longitudinal follow up of the patients. In contrast to data on the juvenile chronic arthritis (JCA) [4], we did not find any correlation of DEK autoantibodies either with articular (erosive arthritis) or with ocular impairment (iridocyclitis), the symptoms of which are very infrequent in SLE. Tables 1 and 2 show the clinical and biological data and their correlation with the presence or absence of anti-DEK antibodies. The only significant differences were the cutaneous involvement, less frequent among DEK⁺ patients, and the positive correlation of these antibodies with the presence of chronic, non-haemolytic anaemia, increased levels of C-reactive protein (CRP) and presence of anti-native DNA antibodies. When the Bonferroni correction was applied these differences were no longer significant.

Table 1.

Clinical data of 288 SLE patients according to their anti-DEK status

	Total patients n = 288		DEK+ patients n = 30 (10·4%)		DEK− patients n = 258 (89·6%)
Manifestations	n	%	n	%	n	%	P	RR
Articular	244	87·1	28	93	216	83·7	N.S.
Erosive arthritis	15	5·2	2	6·6	13	5·03	N.S
Cutaneous	258	89·6	23	76·6	235	91	0·024*	−3·1
Malar erythema	203	70·4	19	63·3	184	71·3	N.S
Livedo reticularis	157	54·5	13	43	117	45·3	N.S
Raynaud’s syndrome	84	29·2	7	23·3	78	29·8	N.S
Sjögren’s syndrome	75	26	9	30	66	25·6	N.S
Renal	79	27·4	11	36·7	68	26·4	N.S
Neurological	91	31·6	8	26·7	83	32·2	N.S
Respiratory	89	30	13	43·3	76	29	N.S
Cardiac	100	34	12	40	88	34	N.S
Iridocyclitis	2	0·69	0	0	2	0·77	N.S

^*NS after Bonferroni correction.

Table 2.

Biological and immunological data of 288 patients with SLE according to their anti-DEK status

	Total SLE patients (n = 288)		DEK+ patients (n = 30)		DEK− patients (n = 258)
Manifestations	n	%	n	%	n	%	P	RR
Anaemia	167	78	23	76·7	144	55·8	0·032*	2·6
Haemolytic anaemia	28	9·7	5	16·7	23	8·9	NS
Leucopenia	190	66	24	80	166	64·3	NS
Lymphopenia	249	86·5	28	93·3	221	85·7	NS
Trombopenia	95	33	11	36·7	84	32·6	NS
Erythrocyte sedimentation rate	197	68·4	24	80	173	67·1	NS
Hypergammaglobulinaemia	121	42	15	50	106	41	NS
Presence of RF	82	28·5	5	16·7	77	29·8	NS
↑ of C-reactive protein	132	45·8	20	66·7	112	43·2	0·02*	2·6
Antinuclear antibodies	286	99·3	30	100	256	99·2	NS
Antibodies to native DNA	175	60·8	24	80	151	58	0·02*	2·8
Anti-SSA antibodies	90	31·2	8	26·67	82	31·8	NS
Anti-SSB antibodies	35	12·1	2	6·7	33	12·8	NS
Anti-Sm antibodies	24	8·3	2	6·7	22	8·53	NS
Anti-nRNP antibodies	52	18	4	13·3	48	18·6	NS
Syphilis serology	29	10	2	6·7	27	10·5	NS

^*NS after Boneroni correction.

DISCUSSION

The main finding of this study is that autoantibodies against DEK oncoprotein occur in 10·4% of patients with SLE. SLE patients positive for this marker are significantly older and display cutaneous manifestations less frequently than individuals negative for anti-DEK. In addition, the former more often exhibit certain markers for a chronic inflammatory state like anaemia and positivity for CRP, as well as a higher frequency of anti-dsDNA autoantibodies.

Studies of autoantibodies in autoimmune diseases have frequently shown the existence of ‘marker’ antibodies or strong correlations between the presence of certain autoantibody species and some pathological features. Anti-DEK autoantibodies were primarily described in strong association with JCA, particularly in young girls with pauciarticular onset and, especially, with iridocyclitis. The presence of antibodies against DEK protein in a patient diagnosed with SLE from our area [6] led us to study the frequency of this antibody in 288 SLE patients and to investigate the pathological features associated with the presence of this autoantibody. The low frequency of anti-DEK positivity found in previous studies where nuclear sonicates were used [4] necessitated a recombinant source of antigen in order to increase the sensitivity of the test and to allow larger population studies. In the present study a recombinant anti-DEK protein obtained in our laboratory was employed [6]. Since we applied a Western blot technique we cannot completely exclude the lack of recognition of native determinants if they exist. Furthermore, since our DEK protein lacks the first 101 N-terminal amino acids of the molecule, other epitopes may not be recognized. Previous studies performed comparing the reactivity of the purified 375-amino acid protein DEK and its truncated 35-kD antigenic derivative lacking 68–72 amino acids from the N-terminal end show that every serum reacting with DEK also recognizes its truncated derivative [4]. These data suggest that the main immunodominant region seems to be located at the C-terminal end and give support to our data. On the other hand, a study performed with recombinant full length antigen in one case of SLE associated with sarcoidosis [5] showed that the titre of serial serum samples of this patient with anti-native DEK antibodies tested by ELISA was comparable to the titre by immunoblot, suggesting that the major epitope(s) is relatively insensitive to denaturation. Taken together with these restrictions, our study identifies a minimal frequency of patients with anti-DEK autoantibodies among patients with SLE. Molecular analysis of the repertoire of antibodies in several autoimmune diseases has shown that autoantibodies to physically linked proteins tend to be generated in tandem as a ‘linked set’[10]. As the prototype serum employed by us to obtain the recombinant DEK antigen also contained anti-ALY antibodies, we looked at the DEK⁺ sera for possible association with anti-ALY activity. All anti-DEK⁺ sera studied were negative for anti-ALY. This preliminary result seems to indicate that the association of anti-DEK and anti-ALY antibodies is not a ‘tandem’ phenomenon as displayed by other autoantibodies like Sm and RNP.

In conclusion, autoantibodies against DEK oncoprotein occur in 10·4% of patients with SLE. SLE patients positive for this marker show a more active clinical form as seen through the non-specific chronic inflammatory signs. Moreover, although these observations need to be confirmed by other groups, anti-DEK⁺ SLE patients seem to be significantly older and less frequently exhibit cutaneous manifestations than individuals negative for anti-DEK. This autoantibody should be included in the still growing list of autoantibodies occurring in SLE patients.