Correspondence: Naoki Watanabe, Division of Laboratory Diagnosis, Sapporo Medical University Hospital, Department of Clinical Laboratory Medicine, School of Medicine, Sapporo Medical University, South-1, West-16, Chuo-ku, Sapporo 060-8543, Japan.
Acquired aplastic anaemia (aAA) is recognized as an autoimmune disorder; however, the autoantigens and target cells involved remain elusive. Expression of autoantibodies and their target cells were examined using the haematopoietic cell line K562 and bone marrow stromal cell line hTS-5; 43·5% and 21·7% of aAA expressed autoantibody against K562 and hTS-5 cells, respectively. The autoantigens were identified by serological identification of antigens through recombinant cDNA expression cloning. This study indicates that haematopoietic cells are the targets of immune abnormality in aAA. These autoantibodies may be utilized to distinguish patients associated with immune abnormality from bone marrow failure syndrome.
Aplastic anaemia (AA) is a syndrome characterized by peripheral cytopenia with hypoplastic bone marrow. Immunosuppressive therapy is an effective first line treatment for acquired AA (aAA) given that 50%–70% of patients respond to the therapy (Risitano, 2010). Non-responding, severely affected patients receive haematopoietic stem cell transplantation to reconstitute the haematopoietic cells as well as the immune system.
It is widely accepted that cellular immunity plays a critical role in the pathogenesis of aAA (Young et al, 2006; Risitano, 2010); the role of humoral immunity in aAA remains elusive. Immunological diseases are classified into four types according to the nature of the reactions involved (Coombs & Lachmann, 1975). In type IV hypersensitivity, self-reacting T-cells mediate the reaction to develop the disease, nevertheless, autoantibodies are produced against autoantigens. Despite the accumulating data showing that immunological abnormalities exist in patients with aAA, the identity of the target cells and molecules involved remains unknown. Analysing autoantibody response is a valuable way to identify autoantigens. In this study, using serological identification of antigens by recombinant cDNA expression cloning (SEREX), we identified autoantibodies that are expressed in patients with aAA accompanied by immune abnormality.
The methods are described in detail in the Supporting Information (Data S1).
aAA serum samples react with K562 cells
In order to investigate the target of the immune abnormality that occurs in aAA, we first verified the expression of autoantibodies in the sera obtained from the 23 patients with aAA by fluorescence-activated cell sorting (FACS) analysis using K562, a haematopoietic cell line, and hTS-5, a bone marrow stromal cell line (Fig 1A). K562 cells showed higher positivity (43·5%, 10/23) than hTS-5 cells (21·7%, 5/23, P =0·029), suggesting that autoantibodies expressed in aAA serum samples preferentially target haematopoietic cells rather than stromal cells.
Identification of SEREX antigens in aAA
To identify the autoantigens involved in aAA, a cDNA library of K562 was created, subjected to SEREX and eight SEREX antigens were identified: CLIC1, SLIRP, HSPB11, NHP2L1, SLC50A1, RPL41, RPS27 and SNRPF.
Autoantibodies against SEREX-defined antigens are expressed in sera from patients with aAA
Enzyme-linked immunosorbent assay was performed to confirm that the sera of aAA patients expresses autoantibodies against the SEREX-defined antigens. The titres of autoantibodies against CLIC1, HSPB11 and RPS27 were higher in patients with aAA than in normal volunteers. The P values for CLIC1, HSPB11 and RPS27 were 6·41 × 10−3, 4·48 × 10−4 and 2·77 × 10−4, respectively. The positivity of IgG-type autoantibodies against CLIC1, HSPB11 and RPS27 were 32·1% (9/28), 39·3% (11/28) and 50·0% (14/28), respectively (Table 1).
Compared with normal individuals, the titres of IgG-type autoantibodies against NHP2L1, SLC50A1, RPL41, SNRPF and SLIRP were not elevated in patients with aAA. The positivity of IgG-type autoantibodies against NHP2L1, SLC50A1, RPL41, SNRPF and SLIRP were 7·1% (2/28), 21·4% (6/28), 21·4% (6/28), 21·4% (6/28) and 7·1% (2/28) respectively, in aAA serum samples.
SEREX-defined antigens are expressed in K562 cells and CD34-positive haematopoietic cells
Reverse transcription polymerase chain reaction was performed to confirm that the SEREX-defined antigens are expressed in haematopoietic cells. All the SEREX-defined antigens were expressed in K562 cells (Fig 1B). When the expression levels were compared with those in hTS-5 cells, most of the antigens showed higher levels in K562 cells. We then verified whether these antigens were expressed in the haematopoietic cells of normal volunteers. CLIC1, NHP2L1, RPL41, RPS27, SLC50A1 and SNRPF were expressed in bone marrow mononuclear cells (BMMNCs); most of the genes were expressed at higher levels in CD34-positive cells compared with CD34-negative cells (Fig 1B).
IgG titres of CLIC1 AND RPS27 correlate positively with response to immunosuppressive therapy
The titres of IgM-type autoantibodies did not show significant differences with the values observed in normal volunteers. From the expression patterns of the autoantibodies, aAA patients could be classified into four groups: (i) IgM- and IgG-positive, (ii) IgM-positive and IgG-negative, (iii) IgM-negative and IgG-positive, and (iv) IgM- and IgG-negative. In this study, 18 out of the 28 patients received immunosuppressive therapy. At least one IgG-type autoantibody against CLIC1, HSPB11 or RPS27 was expressed in 12 of 13 responders, whereas all non-responders did not express any IgG-type autoantibody.
Furthermore, the titres of IgG-type autoantibodies against CLIC1 or RPS27 were higher in responders (partial and complete responders) than in non-responders (Fig 1C). Moreover, higher titres of IgG-type autoantibodies against CLIC1, HSPB11 or RPS27 were associated with higher response rates to immunosuppressive treatment (Fig 1C). The correlation coefficients obtained for CLIC1, HSPB11 and RPS27 were 0·835 (P =0·018), 0·647 (P =0·0037) and 0·960 (P =2·92 × 10−10), respectively. These results indicate that higher titres of IgG-type autoantibodies at presentation may predict a favourable response to immunosuppressive therapy in patients with aAA.
Accumulating evidences have shown that the immune abnormality observed in aAA targets haematopoietic cells. However, many studies have indicated the involvement of stromal cell dysfunctions in the pathogenesis of aAA (Chen, 2005). Most of the autoantigens found in this study are preferentially expressed in K562 and CD34-positive BMMNCs (Fig 1B). Moreover, the autoantibodies targeted K562 haematopoietic cells rather than hTS-5 stromal cells (Fig 1A), indicating that haematopoietic cells are the targets of the immune abnormality in aAA.
The autoantibodies identified in this study may provide two major clinical benefits. The first is that they may be potential serological markers to differentiate aAA from myelodysplastic syndrome (MDS). The expression of autoantibodies was more frequently observed in aAA samples (Table 1). Among the SEREX-defined autoantibodies, expression of anti-RPS27 IgG was higher in aAA patients (50·0%, 14/28) than MDS patients (19·0%, 4/21; P = 6·95 × 10−5). The four patients who expressed autoantibodies in the MDS group were the same individuals (Table 1). These patients may have complicated immune abnormality with MDS.
Another advantage of identifying these autoantibodies is that they may help to predict the outcome of immunosuppressive therapy, given that the response rate to immunosuppressive therapy positively correlated with the titres of IgG-type autoantibodies (Fig 1C). Of the 12 patients who expressed an IgG-type autoantibody (anti-CLIC1, HSPB11 or RPS27), all responded to immunosuppressive therapy, whereas only 16·7% (1/6) responded to the therapy when no IgG-type autoantibody was expressed (P =2·02 × 10−4). A longer interval between diagnosis and transplantation and preceding immunosuppressive therapy is related to a poorer outcome of bone marrow transplantation in patients with aAA (Ades et al, 2004; Kobayashi et al, 2006). Measuring the titre of the autoantibodies at presentation may assist in deciding the course of treatment, and therefore, improve the outcome of patients who require bone marrow transplantation.
Formerly, anti-moesin, diazepam-binding inhibitor-related protein 1, kinectin, postmeiotic segregation increased 1 and HNRNPK antibodies were reported to be expressed in aAA (Hirano et al, 2003, 2005; Feng et al, 2004; Takamatsu et al, 2007; Qi et al, 2010). The common characteristic of the autoantigens found in this study with the ones previously identified is that HNRNPK and SLIRP, RPL41, RPS27 and SNRPF are proteins engaged in RNA biogenesis. The ribosomal protein family has been linked to haematological disorders in bone marrow failure syndrome (Dokal & Vulliamy, 2010; McGowan et al, 2011). Thus, the open reading frame and 5′-untranslated region of the RPL41 and RPS27 mRNAs in half of the aAA patients enrolled in this study were sequenced but no mutation was found.
This study left a couple of open questions. The first is whether the expression of autoantibodies found is related to the pathogenesis of aAA or a result of an immune response after ineffective haematopoiesis. The results showing that most of the patients with MDS did not express autoantibodies (Table 1) suggest that their production is not a result of ineffective haematopoiesis. The expression of IgG-type antibodies correlated with a favourable outcome for immunosuppressive therapy (Fig 1C), supporting the possibility that the expression of autoantibodies is linked to the pathogenesis of aAA.
The second question is whether these autoantibodies are involved in the inhibition of haematopoiesis. We monitored the titres of autoantibodies in several patients with aAA, including three patients who achieved a complete haematological response after immunosuppressive therapy. All of the patients showed a gradual decrease in autoantibody titres; however, the decrease did not coincide with the recovery of blood counts, indicating that the autoantibodies are not the sole cause of aAA. The expression of IgG- but not IgM-type autoantibodies correlated with a favourable outcome after immunosuppressive therapy. In general, IgG production results from class switch recombination, which depends on antigen-specific helper T cells (Shapiro-Shelef & Calame, 2005); intracellular antigens are processed and presented on major histocompatibility complex (MHC) class II molecules (Gannage & Munz, 2009). These observations represent the existence of immunological subsets of cells, including not only plasma cells but also helper T cells targeting these autoantigens; these T cells may play an essential role in the pathogenesis of aAA (Figure S1).
In conclusion, this study identified candidate autoantigens of aAA, which are preferentially expressed in haematopoietic cells. The autoantibodies may be powerful clinical tools to distinguish patients associated with immune abnormality from bone marrow failure syndrome.
The authors would like to thank Dr Kobune (Sapporo Medical University) for providing the hTS-5 cell line. This study was supported by the Grants-in-Aid from Scientific Research in Japan (Grant No. 23590681, KK and NW). KK and NW were supported by the Charitable Trust Laboratory Medicine Research Foundation of Japan.
NW is the principal investigator and takes primary responsibility for this study. KK wrote the paper. MG, KK, YT, TK, and MT performed the laboratory work. KK, DK, and NW coordinated the research.
Maki Goto, Kageaki Kuribayashi and Naoki Watanabe are applying for a Japanese patent regarding this study.