Anti-platelet antibodies associated with the Canale–Smith syndrome bind to the same platelet glycoprotein complexes as those of idiopathic thrombocytopenic purpura patients

Abstract

The Canale–Smith syndrome (CSS) is an inherited disease characterized by massive lymphadenopathy, hepatosplenomegaly and systemic autoimmunity to erythrocytes and platelets. Idiopathic thrombocytopenic purpura (ITP) is an autoimmune disease in which approximately 60–80% of patients have anti-platelet antibodies directed against specific platelet glycoprotein complexes (GPCs) located on their membrane: GP IIb/IIIa, GPIb/IX, and GPIa/IIa. Almost all (95–100%) of the antibody-positive patients have antibodies directed against GPIIb/IIIa alone, or in combination with other glycoprotein targets. Our objective was to determine the specificities of the anti-platelet antibodies in CSS patients. The detection of anti-platelet antibodies was performed using a commercially available ELISA, the Pak-AUTO (GTI, Brookfield, WI), in which highly purified GPIIb/IIIa, GPIb/IX, and GPIa/IIa are immobilized on microtitre plates, incubated with serum or plasma, and subsequently developed with an antihuman polyclonal immunoglobulin. Of 14 CSS patients tested, 11 (79%) had anti-platelet antibodies in their serum directed toward at least one of the three major GPC, nine (82%) of which were against GPIIb/IIIa alone or in combination. Antibodies detected in the sera of ITP patients had similar specificities. No such antibodies were detected in samples from 25 consecutive normal controls. These results demonstrate that a genetically defined defect in lymphocyte apoptosis results in a humoral autoimmune response with anti-platelet specificities very similar to the common idiopathic form of autoimmune thrombocytopenia.

INTRODUCTION

The Canale–Smith Syndrome (CSS) [1,2], also known as the autoimmune lymphoproliferative syndrome (ALPS) [3], or the human lymphoproliferative syndrome (HLPS) [4], is an uncommon disease of childhood characterized by lymphadenopathy/ splenomegaly, autoimmune cytopenias (most frequently affecting the platelets and erythrocytes), and an increase (> 5%) in circulating double-negative T cells. The disease is most commonly associated with mutations in Fas (APO-1/CD95) leading to defective apoptosis of activated lymphocytes [2–6]. This, in turn, leads to progressive accumulation of activated lymphocytes and inappropriate survival of autoreactive cells.

The thrombocytopenia in CSS is similar to that observed in idiopathic thrombocytopenic purpura (ITP) in that it is antibody mediated, is characterized by remissions and exacerbations, and is generally responsive to the same modalities of therapy [2–4,7–10]. Up to 90% of patients with the chronic form of ITP have platelet-associated anti-platelet IgG antibodies directed against specific platelet glycoprotein complexes (GPCs), GPIIb/IIIa, GPIb/IX, and GPIa/IIa, located on the platelet membrane [7,9,11]. Free circulating autoantibodies to GPIIb/IIIa and/or anti-GPIb/IX are detected in 30–50% of ITP patients [7,9]. Anti-platelet antibodies can also be detected in normal individuals. However, these antibodies occur in low titre and are usually directed toward internal platelet proteins, such as intermediate filaments or cytoskeletal proteins [12,13].

So far, there has been little or no characterization of the autoantibodies produced by patients with CSS. To address whether the serum anti-platelet antibodies in CSS were similar to those observed in ITP and/or exaggerated the low-level autoantibodies observed in normal individuals, we evaluated their antigenic specificity in a cohort of CSS patients.

PATIENTS AND METHODS

Patients

CSS was diagnosed on the basis of a chronic disease of early childhood characterized by lymphoaccumulation and systemic autoimmunity [1]. All 14 of the patients included in this study have been described previously and had defective lymphocyte apoptosis caused by Fas mutations [2,5]. Their haematological and major clinical manifestations are listed in Table 1. Thrombocytopenia was defined as a blood platelet count <150 000/μl. Although some of these patients had previously been transfused, none had been pregnant, nor had any been treated with heparin or quinidine prior to the detection of the anti-platelet antibodies. Samples from 54 patients with ITP known to be positive for serum glycoprotein-specific antibodies and 25 consecutive normal individuals were also studied.

Table 1.

Clinical features of CSS patients

Patient	Sex	Age of onset	Clinical manifestations
P1	F	15 years	LAD/AIT/chronic hepatitis B and C Thyroid and breast adenomas/basal cell carcinomas
P2	M	7 months	LAD/AIHA/AIT (S)
P3	M	4 months	LAD/SM (S)/AIHA/AIT/neutropenia
P4	F	2·5 years	LAD/HSM/AIHA/AIT/aPL Reversible CNS disease/mGN
P5	M	Neonate	LAD/SM (S)/AIHA/AIT
P6	M	2 years	LAD/HSM/AIHA/AIT/aCL/urticarial rashes
P7	M	Neonate	LAD/HSM (S)/AIHA/AIT
P8	F	Neonate	SM (S)/severe AIHA, AIT, and neutropenia Rashes
P8 mother	F	Neonate	LAD/SM/AIHA
P9	F	Neonate	LAD/SM (S)/AIHA/neutropenia/lymphopenia Oral ulcerations/premature ovarian failure
P9 brother	M	Neonate	LAD/SM (S)/AIHA/Hodgkin’s disease
P10	M	Neonate	LAD/HSM (S)/AIHA/AIT/urticarial rashes
P10 mother	F	12 years	LAD/SM/AIT/Hodgkin’s disease
P11	M	4 years	LAD/SM/AIHA/AIT/Guillain–Barré syndrome

LAD, Lymphadenopathy; HSM, hepatosplenomegaly; AIHA, autoimmune haemolytic anaemia; aPL, antiphospholipid antibodies; CNS, central nervous system; mGN, membranous glomerulonephritis; SM, splenomegaly; (S), splenectomy; AIT,autoimmune thrombocytopenia; aCL, anticardiolipin antibodies.

Anti-platelet antibody detection

The detection of anti-platelet antibodies was performed using blinded samples and a commercially available ELISA, the Pak-AUTO (GTI, Brookfield, WI) that detects antibodies reactive with GPCs GPIIb/IIIa, Ib/IX, and Ia/IIa. Briefly, highly purified platelet GPCs GPIIb/IIIa, GPIb/IX, and GPIa/IIa from detergent-solubilized platelets are immobilized using mouse MoAbs specific for the relevant GPCs in wells of microtitre plates. Serum or plasma (diluted 1:4 in buffer containing mouse serum) or platelet eluates from patients or normal controls were incubated with the immobilized GPCs and subsequently washed. Bound anti-platelet antibodies were detected by ELISA using an alkaline phosphatase-linked anti-human immunoglobulin reagent (IgG, IgA, IgM). Control sera were used in each assay with a specific internal standard. A positive test was defined as twice the absorbance of the internal standard (1+). The results corresponding to values twice, three, or four times the value of the internal standard were scored as 2+, 3+ and 4+, respectively.

Western blot analysis was performed in six of the 11 CSS patients with anti-platelet antibodies (those with the highest antibody titres), according to methods described previously [14,15], using extracts of 10⁸ platelets per lane as a source of antigens. Sera were tested at a dilution of 1:50 and ECL reagents for IgM, IgG, and IgA detection.

RESULTS

Of the 14 CSS patients tested, 11 (79%) had serum or plasma anti-platelet antibodies directed toward at least one of the three major GPCs (GPIIb/IIIa, GPIb/IX, and GPIa/IIa). Nine of the 11 (82%) patients with autoantibodies recognized GPIIb/IIIa alone, or in combination, versus 83% of ITP patients’ sera. Eight (73%) and three (27%) of the anti-platelet antibody-positive CSS patients recognized GPIa/IIa and GPIb/IX, respectively, alone or in combination, versus 61% and 18% of ITP patients’ sera (Table 2 and Fig. 1). No such antibodies were detected in 25 consecutive normal controls. None of the CSS patients had anti-platelet antibodies reactive with HLA antigens suggestive of alloantibodies. To determine whether CSS patients had autoantibodies to other platelet antigens that were not detected in the ELISA, Western blot analysis of total platelet lysates was performed. All six of these patients had weak, but detectable, bands in the molecular weight regions corresponding to the platelet GPCs detected by the ELISA (not shown). Of the 11 CSS patients with detectable anti-platelet antibodies, 10 had autoimmune thrombocytopenia (ranging from 92 to 132 × 10⁹/l) (Tables 1 and 2). Thrombocytopenia was not observed in patients without anti-platelet antibodies. Titres of anti-platelet antibodies were not systematically tested over time in all CSS patients, but two patients (P9 and P10) who initially were negative upon testing remained negative in samples taken over 3–4 years, and one patient (P6) with anti-platelet antibodies remained positive despite recovery of the platelet count to normal.

Table 2.

Results of anti-platelet antibody detection in CSS patients(Pak-AUTO ELISA)

	Anti-platelet GPC specificity
CSS patient	GPIIb/IIIa	GPIa/IIa	GPIb/IX
P1	4+	2+	Negative
P2	1+	1+	Negative
P3	Negative	1+	Negative
P4	1+	1+	Negative
P5	1+	1+	Negative
P6	3+	2+	1+
P7	2+	2+	1+
P8	1+	1+	Negative
P8 mother	Negative	Negative	Negative
P9	1+	Negative	Negative
P9 brother	Negative	Negative	Negative
P10	Negative	Negative	1+
P10 mother	Negative	Negative	Negative
P11	1+	Negative	Negative

CSS, Canale–Smith syndrome; GPC, glycoprotein complex; 1+ to 4+, intensity of positive ELISA reaction (1+ = twice the value of the absorbance of the internal standard and the cut-off value for positivity; 2+, 3+ and 4+ correspond to twice, three times, and four times the cut-off value for positivity, respectively).

Fig. 1.

Frequency of anti-platelet autoantibodies in Canale–Smith Syndrome (CSS) and idiopathic thrombocytopenic purpura (ITP) patients by ELISA. ▪, CSS; □, ITP.

DISCUSSION

Immune-mediated thrombocytopenia is amongst the most frequent manifestations of autoimmunity in CSS [2–5, 10]. We observed that almost 80% of these patients had autoantibodies directed against one of the three major GPCs, GPIIb/IIIa, GPIb/IX, or GPIa/IIa, that are common targets of autoantibodies in patients with ITP [7,9]. Furthermore, most of these CSS patients, like patients with ITP [7,9,11,16], had antibodies to GPIIb/IIIa, either alone or in combination. Reactivity against GPIa/IIa and GPIb/IX was also similar in CSS compared with ITP. In view of the insensitivity of the Western blot method, we cannot exclude the possibility that other autoantibodies such as those against glycoprotein V were also present in these samples.

The above findings indicate that Fas deficiency leads to a humoral autoimmune response typical of the idiopathic variety of thrombocytopenia, raising the possibility that a defect in peripheral tolerance, possibly related to regulation of apoptosis, may also cause ITP. This is unlikely to be a Fas mutation, since most ITP patients do not develop prominent lymphadenopathy and/or splenomegaly, and a previous study failed to detect Fas defects in ITP patients, but did identify a subgroup of patients with resistance to ceramide-mediated apoptosis [17]. In humans, downstream effectors of apoptosis [18] or another apoptosis pathway could be involved in ITP, especially in those patients with associated haemolytic anaemia (Evans syndrome [19]).

Anti-platelet antibodies are occasionally observed in normal individuals [13] and may be induced by drugs, infections, transfusion of blood products, or pregnancy [7–9,11,15]. In normal subjects, the anti-platelet antibodies are low-titre and usually directed toward internal platelet proteins, such as intermediate filaments or cytoskeletal proteins, rather than surface GPCs, as in ITP [12,13]. Although certain drugs and viral infections have been shown to induce anti-GPC antibodies [13,20,21], they are usually transient and the initiating agent is apparent. The absence of any other identifiable cause, as well as the chronicity and antigenic specificity of the anti-platelet autoantibodies, indicates that immune thrombocytopenia in CSS is caused by a loss of tolerance to platelet surface antigens.

The Fas pathway of programmed cell death plays a pivotal role in the elimination of antigen-primed lymphocytes in the peripheral immune system [5,22]. Mutations in Fas or FasL in mice result in lymphoaccumulation and systemic autoimmunity in all strains of mice, although the clinical manifestations vary considerably, depending upon the strain background [23–29]. In some strains, such as MRL, mice produce anti-dsDNA autoantibodies and develop autoimmune disease very similar to systemic lupus erythematosus (SLE) with vasculitis and glomerulonephritis [24]. At the other end of the spectrum, C57Bl/6 mice produce anti-ssDNA and rheumatoid factors, but have minimal clinical expression of autoimmunity [25,26]. Despite the fact that CSS/ALPS occurs in an outbred population, each member of which has different MHC as well as different B and T-cell repertoires, the core clinical expression of disease (autoimmune haemolytic anaemia and thrombocytopenia) is identical [2–6,10] and the specificity of the humoral response appears to be similar.

The relatively restricted specificity of autoantibodies in CSS could be explained by several factors. Heterozygous Fas mutations in humans are usually dominant negative in action and profoundly impair lymphocyte apoptosis, leading to clinical expression early in life [2–6,10]. It is possible that peripheral tolerization to cell surface antigens, such as platelet GPCs, occurs in the neonatal period and early infancy, and that there is a stringent requirement for Fas in this process. Since the major source of antigens to which developing B cells would be exposed in the bone marrow are of haematopoietic origin, it is perhaps not surprising that autoimmune responses are directed against these multivalent cell surface antigens. In contrast, intracellular antigens that are targeted in SLE are oligovalent, and may require anergy, rather than apoptosis, as the major mechanism of B-cell tolerance induction [30,31].

In conclusion, knowledge that mutations in a defined molecular genetic pathway of apoptosis lead to a humoral anti-platelet response very similar to that observed in ITP provides insight into the mechanisms of tolerance toward multivalent cell membrane antigens. These findings also suggest that regulation of apoptosis requires a more thorough evaluation in idiopathic forms of autoimmune-mediated thrombocytopenia.