Autoantibodies specific for apoptotic U1-70K are superior serological markers for mixed connective tissue disease

Patients suffering from autoimmune diseases are characterized by the presence of autoantibodies directed to a wide range of autoantigens. Mixed connective tissue disease (MCTD) is a relatively rare systemic autoimmune disease and includes a group of patients with overlapping clinical symptoms of systemic lupus erythematosus (SLE), systemic sclerosis (SSc), rheumatoid arthritis and polymyositis/dermatomyositis. Sharp and colleagues were the first to describe MCTD as a distinct rheumatic disease [1], but whether MCTD can be regarded as a distinct disorder has been a subject of discussion [2]. A characteristic serological feature that distinguishes MCTD patients from patients with other connective tissue diseases is high levels of autoantibodies directed against the U1 small nuclear ribonucleoprotein (snRNP) particle [1, 3]. The U1 snRNP is a highly conserved RNA–protein complex, located in the nucleus, where it is involved in the processing of pre-mRNA [4, 5]. It consists of the U1 snRNA molecule and several proteins: the U1A, U1C and U1-70K (70K) proteins are components specific for the U1 snRNP, whereas the seven Sm proteins (B/B', D1, D2, D3, E, F and G) are shared with other U snRNPs [6]. Most U1 snRNP components are autoantigenic in MCTD and SLE. Autoantibodies directed against U1A, U1C, 70K and the U1 snRNA molecule are mainly found in MCTD patients, whereas autoantibodies targeting Sm-D, Sm-B/B' and the E.F.G complex are more specifically associated with SLE [7, 8].

The mechanisms through which such autoantigens, generally highly conserved and ubiquitously expressed molecules, escape tolerance and are recognized by the immune system as non-self remain unclear, but it is proposed that cell death is important in the initiation of autoimmune responses [9, 10]. Recently, secondary necrosis has also been put forward as a source of proteolytically modified autoantigens [11], but the modifications that occur on autoantigens during apoptosis were studied most extensively. Apoptotic modifications on autoantigens include specific cleavage by caspases or granzyme B, (hyper)phosphorylation, dephosphorylation, citrullination, methylation and transglutaminase cross-linking [10, 12, 13], and it is thought that these modifications might be seen by the immune system as novel 'cryptic' epitopes. It is believed that these novel epitopes induce the primary immune response, and that secondary immune responses and epitope spreading result in autoantibodies that are directed against unmodified regions of the autoantigens and antigens that are associated with the initially modified autoantigen [9].

One of the apoptotic modifications occurring on the U1 snRNP is the cleavage of 70K at residue 341 by caspase-3 [14, 15]. Antibodies against 70K are in general the first autoantibodies to appear in anti-U1 snRNP (often referred to as anti-RNP) positive patients, indicating that 70K is important as an initial autoantigen [16]. The molecular and immunological characteristics of the major apoptotic isoform of 70K, a 40 kDa cleavage product that remains associated with the U1 snRNP complex [17], and its role in the triggering of the primary and possibly secondary autoimmune response, are therefore intriguing.

Recently it was shown that sera of some anti-U1 snRNP positive patients contain antibodies that specifically bind to the apoptotic form of 70K, which displays an epitope that is not present on the intact form [18, 19]. This epitope is dependent on the region between amino acids 180 and 205, partly overlapping with the RNA-binding domain and overlapping with the most common T cell epitope [20]. In this study we analyzed a cohort of MCTD and control patients for the presence of autoantibodies against intact and apoptotic 70K. Moreover, we longitudinally analyzed sera from another group of MCTD patients. Our results show that, early in disease, autoantibodies directed against the apoptotic form of 70K (70K^apop) are more strongly represented than autoantibodies against the intact form. Longitudinal studies also show that autoantibodies against 70K^apop are not significantly correlated with disease flares.

In this study, patient sera were analyzed for the presence of autoantibodies against 70K and its apoptotic product (70K^apop), on western blots containing extracts of non-apoptotic and apoptotic Jurkat cells. Two positive controls for the detection of 70K and 70K^apop were included in each experiment: anti-70K mouse monoclonal antibody 2.73 [23], which displays higher reactivities with 70K than with 70K^apop, and serum from MCTD patient B16, which reacts with both 70K and 70K^apop. The position of 70K^apop on western blots was confirmed by a recombinant monoclonal antibody recognizing both 70K and 70K^apop (Fig. 1a) [24]. The results show that in these apoptotic cells 70K is converted almost completely into 70K^apop. Besides positive controls for 70K and 70K^apop, mouse monoclonal antibody ANA125 directed against Sm-B/B' (Fig 1a), and anti-U1-A/U2-B" mouse monoclonal antibody 9A9 (not shown) were also used. To be able to detect autoantibody reactivities to the intact 70K and its apoptotic 40K fragment simultaneously and to facilitate a direct comparison of these reactivities, a mixture of apoptotic and non-apoptotic cell extracts was used to prepare western blots. An additional advantage of this approach was that differences between blots could be excluded, thereby allowing a more accurate comparison of reactivities with 70K and 70K^apop in a single patient serum. Serum antibody reactivities against 70K and 70K^apop were scored ranging from 0 to 5. Figure 1b shows a western blot containing such a mixture of non-apoptotic and apoptotic Jurkat cell extracts, probed with a serial dilution of serum from MCTD patient B16. It can be seen that the signals for 70K and 70K^apop increase when the serum is applied at a lower dilution, indicating that the western blot assay can be used for semi-quantitative interpretation.

Greidinger and colleagues recently showed that antibodies against the 70K protein in RNP-positive patients are often accompanied by antibodies directed against the apoptotic cleavage product of this autoantigen and that the B cell epitopes recognized on the apoptotic product are antigenically different from those contained in the intact form of the 70K protein [18, 19]. This study is the first to confirm and extend these findings and strongly suggests that the reactivity of a patient serum with anti-70K antibodies depends on the presence of antibodies against epitopes shared by 70K and 70K^apop, and on the presence of antibodies against epitopes exclusively present on 70K^apop. The major apoptosis-specific epitope on 70K has been shown to be located in the region containing the RNA-binding domain, and its formation depends on amino acids 180–205, overlapping with the most common T cell epitope [20]. Other apoptosis-specific epitopes on 70K have not yet been described. In our recent studies, monoclonal recombinant human antibodies against 70K were isolated from phage display libraries derived from SLE patients, and several of these monoclonal antibodies preferentially recognized 70K^apop on a western blot and in immunoprecipitation experiments ([24], and D Hof, unpublished results). It is believed that the apoptotic cleavage of 70K leads to the exposure of a neo-epitope that, if presented to the immune system, triggers the autoimmune response. Greidinger and colleagues showed that a mutant consisting of the amino-terminal 205 amino acids, was indeed able to induce an anti-70K^apop antibody response in mice, with subsequent epitope spreading. Interestingly, some of the immunized mice developed pulmonary lesions comparable to lesions found in lungs of MCTD patients. This finding supports the hypothesis that apoptosis-specific epitopes, and antibodies directed against them, might have a pathological role in the triggering and maintenance of the human autoimmune response to 70K [19].

In our study, a minority of MCTD sera (4%) contained autoantibodies exclusively reacting with intact 70K. We suggest that these sera derive from patients in a relatively late disease phase and primarily contain antibodies resulting from expanded epitope spreading. Most epitopes recognized by these sera might therefore be dependent on the carboxy-terminal part of the protein, which is cleaved off during apoptosis and is not present on 70K^apop. Patients that tested negative in our western blot experiments might either have low levels of anti-70K antibodies or might not produce such antibodies at all. Instead, other components of the U1 snRNP, such as the U1 RNA molecule, U1A or U1C, might be targeted by these sera and might explain their anti-U1 snRNP reactivity.

We show here that most U1 snRNP-positive patient sera preferentially recognize 70K^apop, which is most probably explained by the presence of antibodies targeting an apoptotic 70K epitope. These results are in line with reports by Greidinger and colleagues [18, 19], who found that about 50% of their RNP-positive sera contained 70K^apop autoantibodies.

How disease flares are induced is not completely understood. Correlations between serum levels of certain autoantibodies and disease activity have been reported for MCTD and SLE [21, 22], but it can be disputed whether these antibodies contribute to the disease flares or are merely epiphenomena. Our data show that antibodies against 70K^apop are not significantly correlated with disease activity, suggesting that there is no important role for 70K^apop in the initiation of disease flares. However, it is possible that the variations in antibody levels against the apoptosis-specific epitope are masked by the presence of antibodies against other epitopes on 70K and U1-70K^apop. Furthermore, a polyspecific secondary antibody was used to detect bound serum antibodies, and as a consequence variations in isotype-specific antibody levels might have remained undetected.

It has been reported that the first autoantibodies to appear in anti-RNP-positive patients are generally antibodies against 70K [16, 27]. Our results suggest that 70K^apop drives the primary autoimmune response to 70K, because in several patients antibodies against an epitope associated with 70K^apop preceded the appearance of reactivity with intact 70K. The fact that the first serum samples from relatively few patients exclusively contain anti-70K^apop antibodies might be due to the stage of disease development at which the patient enters the rheumatological clinic. It is likely that the first symptoms, later followed by the diagnosis of the disease, had been established years before the start of the longitudinal study. Moreover, it is possible that autoantibodies, especially those generated by the primary immune response, were already present before the manifestation of clinical symptoms and that subsequent epitope spreading might have occurred before the patient entered the rheumatological clinic. For example, anti-cyclic citrullinated peptide autoantibody is a very specific marker for rheumatoid arthritis, and such antibodies can be detected in patients up to 10 years before the occurrence of the first clinical symptoms [28, 29]. In our opinion this might explain why a relative enrichment of anti-70K^apop antibodies could not be detected in the early sera of all patients.

During apoptosis, the U1 snRNP complex is modified in several ways. In addition to cleavage of 70K, U1 snRNA and the Sm-F protein are cleaved, and phosphorylated serine–arginine proteins associate with the complex [30]. Apoptotic modifications of the U1A and U1C proteins have not yet been described. 70K can be cleaved by caspase-3 and granzyme B, and it can be oxidatively fragmented in the presence of metals, resulting in products of 40, 60 and 55 kDa, respectively. Correlations between the recognition of specific 70K fragments and disease manifestations are interesting. For example, patients suffering from Raynaud's phenomenon preferentially recognize the oxidatively modified 55 kDa fragment of 70K [31]. The findings that early MCTD sera are enriched for antibodies against the 40 kDa apoptotic fragment (70K^apop) and that most sera show a higher reactivity with this fragment suggest that caspase-3 cleaved 70K has a role in breaking tolerance in these patients. Although granzyme B is postulated to have a role in breaking tolerance [32] to 70K, it is unknown whether specific patient groups preferentially recognize the 60 kDa cleavage product generated by granzyme B, which would be interesting to study in more detail.

Analysis of a group of MCTD patient sera by western blotting showed that the majority of patient sera recognized 70K^apop more efficiently than the intact form of the 70K protein. The fact that the presence of these antibodies in most patients precedes the occurrence of other anti-70K antibodies suggests that 70K^apop is particularly important for the early detection of this disease in patients.