Autoantibody to c-Mpl (thrombopoietin receptor) in systemic lupus erythematosus: Relationship to thrombocytopenia with megakaryocytic hypoplasia

Authors


Abstract

Objective

To examine the prevalence, clinical associations, and pathogenic role of autoantibodies to c-Mpl, the thrombopoietin (TPO) receptor, in patients with systemic lupus erythematosus (SLE).

Methods

Sera from 69 SLE patients, 84 patients with idiopathic thrombocytopenic purpura (ITP), and 60 healthy individuals were screened for anti–c-Mpl antibodies by enzyme-linked immunosorbent assay using recombinant c-Mpl as an antigen. Clinical findings, autoantibody profiles, and serum TPO levels were compared between SLE patients with and without anti–c-Mpl antibodies. A pathogenic role for the anti–c-Mpl antibody was evaluated by examining its inhibitory effect on TPO-dependent cell proliferation and megakaryocyte colony formation.

Results

Serum anti–c-Mpl antibody was detected in 8 SLE patients (11.6%) and 7 ITP patients (8.3%), but in none of the healthy controls. Anti–c-Mpl antibody was associated with thrombocytopenia (P = 0.0002) and a decrease in bone marrow megakaryocytes (P = 0.02) in SLE patients. Serum TPO levels in thrombocytopenic SLE patients with anti–c-Mpl antibodies were significantly elevated compared with levels in those without the antibodies (P = 0.007). IgG fractions purified from anti–c-Mpl antibody–positive sera bound to c-Mpl expressed on the cell surface and inhibited TPO-dependent cell proliferation and megakaryocyte colony formation.

Conclusion

Autoantibody to c-Mpl is present in a subset of SLE patients with thrombocytopenia and megakaryocytic hypoplasia. It is likely that the impaired thrombopoiesis in these patients is mediated by the anti–c-Mpl antibody, which functionally blocks an interaction between TPO and c-Mpl.

Thrombocytopenia is one of the major complications in patients with systemic lupus erythematosus (SLE), with a prevalence ranging from 7% to 30% (1). Severe thrombocytopenia with an apparent bleeding tendency is uncommon in SLE patients, but can be fatal due to bleeding into the brain or gastrointestinal tract. In fact, the presence of thrombocytopenia is an independent risk factor for a worse prognosis in patients with SLE (2).

A variety of pathogenic processes are reported to be involved in thrombocytopenia in SLE. The most common process is increased platelet destruction mediated by antiplatelet autoantibodies, which is analogous to the mechanism in patients with idiopathic thrombocytopenic purpura (ITP) (3). Other abnormalities associated with thrombocytopenia include thrombotic thrombocytopenic purpura, disseminated intravascular coagulation, hemophagocytic syndrome, and antiphospholipid syndrome (3). All of these conditions are associated with increased platelet consumption in the peripheral circulation, with normal or increased bone marrow megakaryocytes, and platelet production is not primarily affected. However, there are several case reports of patients with SLE complicated by amegakaryocytic thrombocytopenia (AMT) (4–7), which is associated with impaired thrombopoiesis due to a selective absence of megakaryocytes in the bone marrow (8). Although AMT is a rare disorder in SLE patients, a decreased proportion of megakaryocytes is occasionally detected in the bone marrow of SLE patients with thrombocytopenia. T cell–mediated and antibody-mediated suppression of megakaryocyte colony formation have both been reported as possible mechanisms for AMT (6, 7), but the underlying processes of megakaryocytic hypoplasia in SLE patients are largely unknown.

Thrombopoietin (TPO) has recently been identified as a megakaryocyte colony-stimulating factor and megakaryocyte potentiator (9). TPO binds to its receptor, c-Mpl, on the surface of hematopoietic stem cells and megakaryocytes and induces their proliferation and maturation (10). Recently, defective c-Mpl expression due to c-Mpl gene mutations was shown to be the major cause of congenital AMT (11). Because the presence of autoantibodies to various self proteins is one of the immunologic features in patients with SLE (12), an autoantibody that interferes with the interaction between TPO and c-Mpl might be present in thrombocytopenic SLE patients, especially those with megakaryocytic hypoplasia including AMT. To test this hypothesis, we developed an enzyme-linked immunosorbent assay (ELISA) system to detect serum anti–c-Mpl antibodies and used it to screen sera from SLE patients with and without thrombocytopenia. Clinical characteristics associated with the presence of anti–c-Mpl antibodies in SLE patients and the pathogenic roles of these antibodies in megakaryocytic hypoplasia were also examined in this study.

PATIENTS AND METHODS

Patients and controls

Sixty-nine consecutive patients with SLE (6 men and 63 women) and 84 patients with chronic ITP (23 men and 61 women) who were followed up at Keio University Hospital were examined in this study. All SLE patients satisfied the American College of Rheumatology (ACR) classification criteria for SLE (13). The criteria for the diagnosis of chronic ITP were as follows: 1) thrombocytopenia (≤50,000 platelets/mm3), 2) normal or increased megakaryocytes in the bone marrow smears, without morphologic evidence of dysplasia, and 3) disease duration >6 months (14). To assure that patients with the idiopathic form of TP were selected, clinical and laboratory examinations were carefully carried out to exclude those with other primary clinical disorders that can give rise to all or some of the conditions seen in ITP (i.e., SLE, autoimmune thyroid diseases, human immunodeficiency virus–related immunologic thrombocytopenia, lymphoma, chronic lymphocytic leukemia, hypersplenism, and drug-induced thrombocytopenia) (14). All ITP patients had a platelet count of <100,000/mm3 at the time of serum sampling. The mean ± SD age at examination was 42.9 ± 12.3 years (range 18–69 years) for the SLE patients and 45.8 ± 15.7 years (range 17–76 years) for the ITP patients.

Sixty healthy individuals (26 men and 34 women) with normal platelet counts were used as a control group. All blood samples were obtained after the patients and control subjects had given their written informed consent, as approved by the Keio University Institutional Review Board.

Clinical features of the SLE patients

Demographic and clinical features were evaluated for each SLE patient at the time of serum collection. Thirty-five clinical and laboratory findings were recorded; these were individual items included in the ACR criteria (13) and the SLE Disease Activity Index (SLEDAI) (15). Thrombocytopenia was defined as a platelet count of <100,000/mm3 (13). The SLEDAI score was calculated and used to evaluate disease activity in SLE patients.

Serologic analysis

Anti–double-stranded DNA (anti-dsDNA) antibody was detected by the Farr assay. Anti-Sm, anti–U1 RNP, anti-SSA/Ro, and anti-SSB/La antibodies were identified using an RNA immunoprecipitation assay with unlabeled HeLa cell extracts (16). Anti–β2-glycoprotein I (anti-β2GPI) antibody was detected with an ELISA kit (Yamasa, Choshi, Japan), in which cardiolipin-coated plates were incubated with purified human β2GPI as a cofactor. Serum TPO levels were measured using an ELISA kit (Quantikine; R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. In some experiments, IgG was purified from sera by affinity chromatography using a HiTrap Protein G column (Amersham Pharmacia Biotech, Uppsala, Sweden) according to the manufacturer's protocol. IgG fractions were dialyzed against phosphate buffered saline (PBS) and sterilized by passage through 0.22 μm–pore syringe filters.

Bone marrow evaluation

Bone marrow aspiration was performed in 12 SLE patients with thrombocytopenia and in all 84 ITP patients. The majority of the patients underwent bone marrow biopsy at the same time. The proportion of megakaryocytes in bone marrow nucleated cells was evaluated on bone marrow cell smears. At least 1,000 nucleated cells were counted in each sample. A megakaryocyte proportion of ≤0.2% was regarded as “decreased,” and >1.0% was regarded as “increased.” Bone marrow samples apparently containing peripheral blood cells were excluded from the analysis.

Antigen preparations

A recombinant protein encoding the entire extracellular domain of human c-Mpl (amino acid residues 1–466) was provided by Kirin Brewery (Takasaki, Japan). Recombinant c-Mpl has been expressed in HEK cells and is thought to have native posttranslational modification and conformation based on its capacity to bind to TPO. Platelet surface glycoprotein IIb-IIIa (GPIIb-IIIa), a major target recognized by antiplatelet autoantibodies in ITP patients (17), was affinity purified from outdated platelet concentrates as described elsewhere (18).

Mouse cell line transfected with human c-Mpl

The mouse myeloid leukemia cell line FDC/P2 with and without transfected full-length human c-Mpl complementary DNA (cDNA) was provided by Kirin Brewery (19). The wild-type FDC/P2 cell line requires mouse interleukin-3 (IL-3) to proliferate, while the FDC/P2 line transfected with human c-Mpl cDNA (FDC/P2-c-Mpl) is able to proliferate in response to either mouse IL-3 or human TPO. These cell lines were maintained in Iscove's modified Dulbecco's medium containing 10% fetal bovine serum, 2 mML-glutamine, 50 units/ml penicillin, and 50 μg/ml streptomycin in the presence of recombinant mouse IL-3 (1 ng/ml) at 37°C in a humidified atmosphere of 5% CO2.

Detection of platelet-associated anti–GPIIb-IIIa antibodies

IgG antibodies to GPIIb-IIIa that bound to platelet surfaces were measured by ELISA according to the published method (20). Immunoglobulins eluted from 5 × 107 platelets were used as the primary antibody in an ELISA. All samples were tested in duplicate, and antibody units were calculated from the optical density at 450 nm (OD450 nm), using a standard curve obtained from serial concentrations of a monoclonal antibody to GPIIb-IIIa (clone HPL1; Harlan Sera Laboratory, Leicester, UK). The cutoff value was determined as the mean plus 3SD in 20 samples from healthy individuals (3.3 units).

ELISA for detecting anti–c-Mpl antibodies

An ELISA for detecting anti–c-Mpl antibodies was established as described previously (21). Briefly, polyvinyl 96-well microtitration plates (Immulon 2; Dynatech Laboratories, Chantilly, VA) were coated with recombinant c-Mpl at 1 μg/ml diluted in PBS, at 4°C for 12 hours. The remaining free binding sites were blocked with 1% bovine serum albumin (BSA) in PBS for 1 hour at room temperature. Wells were incubated with serum samples diluted 1:100 in ELISA buffer (PBS containing 0.1% BSA and 0.05% Tween 20) at room temperature for 2 hours, and subsequently incubated with peroxidase-conjugated goat anti-human IgG (ICN/Cappel, Aurora, OH) diluted 1:5,000 in ELISA buffer at room temperature for 1 hour. The antibody binding was visualized by incubation with tetramethylbenzidine (1 mg/ml) in phosphate–citrate buffer containing DMSO. After the reaction was stopped by the addition of 1M sulfuric acid, the OD450 nm was read with an automatic plate reader (Bio-Rad Laboratories, Hercules, CA). All incubations were followed by 3 washes with ELISA buffer. Samples were tested in duplicate, and antibody units were calculated from the OD450 nm results, using a standard curve obtained from serial concentrations of a rabbit polyclonal antibody to human c-Mpl (Kirin Brewery). The cutoff value was considered to be the mean plus 3SD of 20 healthy control sera (18.0 units).

The specificity of the anti–c-Mpl antibody reactivity was confirmed by an ELISA competition assay (21). Briefly, sera positive for anti–c-Mpl antibody were preincubated with various competitors at room temperature for 1 hour before being added to antigen-coated ELISA wells. Recombinant c-Mpl and platelet-derived human GPIIb-IIIa were used as competitors at concentrations of 1–40 μg/ml. Serial dilutions of cultured FDC/P2 and FDC/P2-c-Mpl cells were also used as competitors (102–106 cells).

Immunoblots

Recombinant c-Mpl was fractionated on sodium dodecyl sulfate–10% polyacrylamide gels and transferred onto nitrocellulose membranes. The membranes were incubated with a 1:100 dilution of patient sera or a 1:1,000 dilution of rabbit polyclonal anti–c-Mpl antibodies. The membranes were subsequently incubated with alkaline phosphatase–conjugated goat anti-human or anti-rabbit IgG (ICN/Cappel), and the reactivities were visualized by development with 4-nitroblue tetrazolium chloride/BCIP.

Flow cytometric analysis

Wild-type FDC/P2 and FDC/P2-c-Mpl cells were incubated with IgG isolated from patient and control sera (500 μg/ml) or rabbit polyclonal anti–c-Mpl antibody (50 μg/ml), followed by incubation with fluorescein-5-isothiocyanate–conjugated goat anti-human or anti-rabbit IgG F(ab′)2 fragment (ICN/Cappel). Cell staining was analyzed on a FACSCalibur flow cytometer (BD PharMingen, San Diego, CA) using the CellQuest software (BD PharMingen).

Cell proliferation assay

Proliferation of FDC/P2-c-Mpl cells in response to mouse IL-3 or human TPO was measured as described previously (19). Exponentially growing FDC/P2-c-Mpl cells in the presence of mouse IL-3 were washed 3 times and resuspended in complete medium without mouse IL-3 at 2.5 × 104 cells/ml. The cell suspension (100 μl) was seeded into each well of 96-well flat-bottomed tissue culture plates and cultured in the presence or absence of serial concentrations of IgG fractions or rabbit anti–c-Mpl antibody for 1 hour. Mouse IL-3 (1 ng/ml) or human recombinant TPO (5 ng/ml; Kirin Brewery) was then added to the cultures, and the cells were further incubated for 72 hours. After a final 16-hour incubation with 0.5 μCi/well of 3H-thymidine, the cells were harvested and the 3H-thymidine incorporation was determined in a TopCount microplate scintillation counter (Packard, Meriden, CT). All cultures were carried out in quadruplicate, and the results were expressed as the mean of the 4 values. In some experiments, before its use in the cell proliferation assay, the complete medium containing the IgG fractions was preincubated twice in sterile polyvinyl 96-well plates that were previously incubated with recombinant c-Mpl at 0–32 μg/ml and then with PBS containing 3% BSA. In some instances, the results were expressed as the percentage of inhibition, which was calculated as the difference between the counts per minute incorporated in the cultures with and without IgG fraction at 80 μg/ml divided by the cpm incorporated in the culture without IgG fraction.

Megakaryocyte colony assay

The effects of IgG fractions on the growth of megakaryocyte colonies were examined by soft agar culture assay as described previously (22, 23). Briefly, the CD34+ cell–enriched fraction prepared from normal human bone marrow mononuclear cells using an immunomagnetic separation system (Mini-MACS; Miltenyi Biotec, Bergisch Gladbach, Germany) was incubated with or without IgG fractions (50 μg/ml or 250 μg/ml) or rabbit anti–c-Mpl antibody (5 μg/ml) for 1 hour and then cultured in serum-free soft agar for 14 days in the presence of human TPO (0.5 ng/ml). It has been shown that only megakaryocyte colonies are able to grow in this assay condition (22).

Statistical analysis

All comparisons for statistical significance between 2 patient groups were performed using the chi-square test or Student's t-test. The correlation coefficient was determined using a single regression model.

RESULTS

Prevalence of anti–c-Mpl antibodies in SLE patients

Serum samples from 69 SLE patients, 84 ITP patients, and 60 healthy individuals were analyzed by ELISA using recombinant c-Mpl as an antigen (Figure 1). Anti–c-Mpl antibody was detected in 8 SLE patients (11.6%) and 7 ITP patients (8.3%), but in none of the healthy controls. The prevalence of anti–c-Mpl antibodies was similar in SLE and ITP patients, but when the SLE patients were divided into 2 groups based on the presence or absence of thrombocytopenia, anti–c-Mpl antibody was more frequent in the SLE patients with thrombocytopenia than in those without thrombocytopenia (38.9% versus 2.0%; P = 0.0002) or those with ITP (P = 0.002).

Figure 1.

Serum anti–c-Mpl antibodies determined by enzyme-linked immunosorbent assay in 69 patients with systemic lupus erythematosus (SLE), 84 patients with idiopathic thrombocytopenic purpura (ITP), and 60 healthy controls. SLE patients were divided into 2 groups according to the presence or absence of thrombocytopenia (TP). The broken line denotes the positive cutoff point (18.0 units).

To examine the specificity of anti–c-Mpl antibody binding in the ELISA, we conducted a competitive ELISA in which serum samples were preincubated with serial amounts of competitors, including recombinant c-Mpl, GPIIb-IIIa, FDC/P2 cells, and FDC/P2-c-Mpl cells. Representative results from an SLE serum are shown in Figure 2. The anti–c-Mpl antibody reactivity was inhibited in a dose-dependent manner by preincubation of patient serum with c-Mpl, but preincubation with GPIIb-IIIa had no effect. Moreover, preincubation of serum with FDC/P2-c-Mpl cells, but not with wild-type FDC/P2 cells, resulted in loss of anti–c-Mpl antibody reactivity. Principally concordant results were obtained from 4 additional anti–c-Mpl antibody–positive serum samples, including samples from 2 SLE and 2 ITP patients.

Figure 2.

Competition assay to examine the specificity of anti–c-Mpl antibody reactivity detected by enzyme-linked immunosorbent assay in a representative serum sample from a patient with systemic lupus erythematosus. Serial amounts of recombinant c-Mpl or platelet surface glycoprotein IIb-IIIa (GPIIb-IIIa) (A) and FDC/P2 cells or FDC/P2 cells expressing human c-Mpl (FDC/P2-c-Mpl cells) (B) were used as competitors. OD450 = optical density at 450 nm.

Fifteen serum samples from SLE and ITP patients that were positive for anti–c-Mpl antibody by ELISA were further examined by immunoblotting using recombinant c-Mpl as an antigen. A representative result in SLE sera is shown in Figure 3, and the antibody binding to c-Mpl on the membrane was detected in 2 of 5 sera positive for anti–c-Mpl antibody by ELISA. Overall, only 7 of 15 anti–c-Mpl–positive sera, including 3 SLE and 4 ITP sera, reacted with c-Mpl in immunoblots. There was no statistically significant difference in the mean ± SD anti–c-Mpl antibody levels determined by ELISA between the 7 samples that bound to c-Mpl in immunoblots and the 8 that did not (43.0 ± 37.6 antibody units versus 29.6 ± 22.3 antibody units). None of 80 serum samples that were negative for anti–c-Mpl antibody by ELISA (from 30 SLE patients, 30 ITP patients, and 20 healthy controls) reacted with c-Mpl in immunoblots.

Figure 3.

Immunoblots using recombinant c-Mpl as an antigen in sera from patients with systemic lupus erythematosus (SLE). Recombinant c-Mpl was fractionated on sodium dodecyl sulfate–10% polyacrylamide gels, transferred onto nitrocellulose membranes, and then probed with SLE sera that were positive or negative for anti–c-Mpl antibodies by enzyme-linked immunosorbent assay (ELISA). Lane 1, Rabbit anti–c-Mpl polyclonal antibodies; lanes 2–6, anti–c-Mpl–positive SLE sera; lanes 7–11, anti–c-Mpl–negative SLE sera.

Clinical characteristics of SLE patients with anti–c-Mpl antibody

Demographic and clinical findings as well as coexisting autoantibodies were compared between SLE patients with and without anti–c-Mpl antibody; selected findings are shown in Table 1. There were no differences in the frequencies of clinical characteristics except for thrombocytopenia, which was more frequently found in patients with anti–c-Mpl antibody than in those without (P = 0.0002). The SLEDAI score was the same for these 2 patient groups, indicating that the anti–c-Mpl antibody was not associated with disease activity. The frequencies of other autoantibodies, including anti-dsDNA, anti-Sm, anti–U1 RNP, anti-SSA/Ro, anti-SSB/La, and anti-β2GPI, were also similar. In contrast, platelet-associated anti–GPIIb-IIIa antibody frequently coexisted with the anti–c-Mpl antibody (P = 0.03). Platelet-associated anti–GPIIb-IIIa antibody was detected in 10 of 18 SLE patients with thrombocytopenia, but in none of 34 patients without thrombocytopenia (P = 0.00001). Moreover, because the frequency of platelet-associated anti–GPIIb-IIIa antibody was similar between the anti–c-Mpl–positive and –negative SLE patients with thrombocytopenia (57% versus 45%), it appeared to be primarily associated with thrombocytopenia rather than with the anti–c-Mpl antibodies.

Table 1. Selected clinical and serologic characteristics of SLE patients with or without serum anti–c-Mpl antibody*
Clinical or laboratory findingAnti–c-Mpl positive (n = 8)Anti–c-Mpl negative (n = 61)P
  • *

    SLE = systemic lupus erythematosus; NS = not significant; SLEDAI = SLE Disease Activity Index; anti-dsDNA = anti–double-stranded DNA; anti-β2GPI = anti–β2-glycoprotein I; anti–GPIIb-IIIa = anti–platelet surface glycoprotein IIb-IIIa.

No. of men/no. of women1/75/56NS
Age at examination, mean ± SD years44.9 ± 7.642.6 ± 12.9NS
Disease duration, mean ± SD years5.8 ± 2.46.2 ± 3.7NS
SLEDAI score, mean ± SD5.6 ± 6.54.1 ± 4.0NS
Renal disorder, %2528NS
Neurologic disorder, %135NS
Leukopenia (<4,000 leukocytes/mm3), %3836NS
Thrombocytopenia (<100,000 platelets/mm3), %88180.0002
Hypocomplementemia, %3828NS
Anti-dsDNA antibody, %3833NS
Anti-Sm antibody, %2520NS
Anti-β2GPI antibody, %1316NS
Platelet-associated anti–GPIIb-IIIa antibody, % (frequency)57 (4/7)13 (6/45)0.03

Anti–c-Mpl antibody and the proportion of megakaryocytes in bone marrow.

Of 12 SLE patients with thrombocytopenia, 5 had an isolated decrease in megakaryocytes, and 1 of these 5 patients showed an absence of megakaryocytes, compatible with the diagnosis of AMT. The decrease in megakaryocytes in these patients was confirmed by bone marrow biopsies. As shown in Table 2, an absence of or decrease in megakaryocytes was found in all 4 anti–c-Mpl–positive patients, but in only 1 of 8 anti–c-Mpl–negative patients (P = 0.02).

Table 2. Proportion of bone marrow megakaryocytes in SLE patients with thrombocytopenia in the presence or absence of serum anti–c-Mpl antibody*
Bone marrow megakaryocytesAnti–c-Mpl positive (n = 4)Anti–c-Mpl negative (n = 8)
  • *

    SLE = systemic lupus erythematosus.

Absent10
Decreased31
Normal05
Increased02

Correlations between anti–c-Mpl antibody and serum TPO levels

Serum TPO levels were measured in SLE patients, ITP patients, and healthy controls with or without thrombocytopenia and anti–c-Mpl antibodies (Figure 4). Sera with high TPO levels from 2 patients with aplastic anemia were also analyzed as a control (24). The mean ± SD TPO level in 26 healthy controls was 83.9 ± 18.7 pg/ml. TPO levels were significantly higher in thrombocytopenic SLE patients with anti–c-Mpl antibodies than in those without (P = 0.007) or in nonthrombocytopenic SLE patients (P < 0.001) or in healthy controls (P < 0.001). It was noteworthy that the TPO levels in some anti–c-Mpl–positive SLE patients exceeded those in patients with aplastic anemia. An SLE patient with AMT showed the highest TPO level, which was >20 times higher than the mean level in healthy controls. ITP patients with anti–c-Mpl antibodies also had higher TPO levels compared with ITP patients without the antibodies and compared with healthy controls (P = 0.004 and P < 0.001, respectively). However, the TPO levels in anti–c-Mpl–negative SLE patients with and without thrombocytopenia were greater than those in healthy controls (P = 0.002 and P = 0.001, respectively).

Figure 4.

Serum thrombopoietin (TPO) levels in SLE and ITP patients in the presence or absence of TP and anti–c-Mpl antibody, and in healthy controls. Sera with high TPO levels from 2 patients with aplastic anemia were also analyzed as a control. TPO levels between individual groups were compared using Student's t-test. Solid bars indicate mean TPO levels. Double bar indicates threshold for the single outlier, an SLE patient with amegakaryocytic TP. See Figure 1 for other definitions.

Binding of the anti–c-Mpl antibodies to c-Mpl expressed on cell surfaces

To further examine effects of the anti–c-Mpl antibodies on the functional interaction between TPO and c-Mpl, IgG fractions were purified from 6 anti–c-Mpl–positive sera (from 3 SLE and 3 ITP patients) and 6 anti–c-Mpl–negative sera (from 3 SLE patients, 2 ITP patients, and 1 healthy control) and used in the following experiments. The binding of anti–c-Mpl antibodies in patient sera to c-Mpl molecules expressed on the cell surface was examined by flow cytometry using FDC/P2-c-Mpl cells. As shown in Figure 5, rabbit anti–c-Mpl antibodies bound to FDC/P2-c-Mpl cells, but not to wild-type FDC/P2 cells. IgG purified from a representative anti–c-Mpl–positive SLE serum (serum FY) at a concentration of 500 μg/ml, comparable to one-thirtieth of the serum IgG concentration, also bound to the surface of FDC/P2-c-Mpl cells. In contrast, no binding was detected when the cells were incubated with the IgG from SLE patients or healthy individuals who were negative for anti–c-Mpl antibodies. All IgG preparations obtained from 6 anti–c-Mpl–positive sera stained FDC/P2-c-Mpl cells (see Table 3), but positive staining was detected in none of 6 anti–c-Mpl–negative sera.

Figure 5.

Binding of anti–c-Mpl antibodies to c-Mpl expressed on the cell surface. Wild-type FDC/P2 cells and FDC/P2 cells expressing human c-Mpl (FDC/P2-c-Mpl cells) were incubated with rabbit anti–c-Mpl antibodies (50 μg/ml) or IgG purified from 1) anti–c-Mpl–positive serum FY from a patient with systemic lupus erythematosus (SLE), 2) anti–c-Mpl–negative serum from an SLE patient, or 3) anti–c-Mpl–negative serum from a healthy control (500 μg/ml each), followed by incubation with fluorescein-5-isothiocyanate–conjugated goat anti-human or anti-rabbit IgG F(ab′)2 fragment. Cell staining was analyzed by flow cytometry and is shown as shaded histograms. Open histograms represent controls stained with the secondary antibody alone. Results shown are representative of 3 experiments.

Table 3. Inhibitory effect of the IgG fraction on TPO-induced cell proliferation and its associations with laboratory findings in SLE and ITP patients seropositive for anti–c-Mpl antibody*
Patient, serumNo. of platelets/mm3% megakaryocytes in bone marrowAnti–c-Mpl antibody, unitsc-Mpl binding in immunoblotsSerum TPO level, pg/mlBinding to FDC/P2-c-Mpl cells by flow cytometry% inhibition of TPO-induced cell proliferation
  • *

    TPO = thrombopoietin; SLE = systemic lupus erythematosus; ITP = idiopathic thrombocytopenic purpura; FDC/P2-c-Mpl cells = FDC/P2 cells expressing human c-Mpl; ND = not determined.

  • Calculated as the difference between the cpm incorporated in the cultures with and without IgG fraction at 80 μg/ml divided by the cpm incorporated in the culture without IgG fraction.

  • From a patient with amegakaryocytic thrombocytopenia.

SLE       
 MT18,0000.0819.4+1,387+80
 FY21,000084.1>2,000+84
 MA49,000ND122.2+909+78
ITP       
 AF17,0000.4029.0+58+54
 SK37,0000.2922.6243+72
 ST52,0000.2459.2+448+81

Interference with the TPO–c-Mpl interaction by anti–c-Mpl antibody

The TPO-induced proliferation of FDC/P2-c-Mpl cells was evaluated in the presence of IgG purified from serum samples that were positive or negative for anti–c-Mpl antibody (Figure 6A). Rabbit anti–c-Mpl antibodies inhibited the TPO-induced cell proliferation in a dose-dependent manner. IgG from representative anti–c-Mpl–positive sera from an SLE patient (serum MA) and an ITP patient (serum ST) also inhibited the cell proliferation. In fact, all 6 serum samples that were positive for anti–c-Mpl antibodies showed a similar inhibitory effect on the TPO-induced cell proliferation (Table 3). The percentage of inhibition of the cell proliferation was independent of anti–c-Mpl antibody levels by ELISA or the binding to c-Mpl in immunoblots, although there was a trend toward a positive correlation between percent inhibition of TPO-induced cell proliferation and serum TPO levels (r = 0.72, P = 0.11). In contrast, none of the IgG from the 6 anti–c-Mpl–negative serum samples showed any suppressive effect. However, none of the IgG fractions, including those from 6 anti–c-Mpl–positive sera, suppressed the proliferation of FDC/P2-c-Mpl cells induced by mouse IL-3. An excess amount of human TPO did not reverse the inhibition of cell proliferation by IgG from anti–c-Mpl–positive sera (data not shown).

Figure 6.

Inhibition of thrombopoietin (TPO)–induced proliferation of FDC/P2 cells expressing human c-Mpl (FDC/P2-c-Mpl cells) by IgG fractions obtained from SLE and ITP patients with and without anti–c-Mpl antibodies. A, FDC/P2-c-Mpl cells were cultured for 3 days with human TPO (5 ng/ml) in the presence or absence of serial concentrations of rabbit anti–c-Mpl polyclonal antibodies or IgG purified from 1) serum MA from an anti–c-Mpl–positive SLE patient, 2) serum ST from an anti–c-Mpl–positive ITP patient, 3) serum from an anti–c-Mpl–negative SLE patient, or 4) serum from a healthy control. After the cells were harvested, 3H-thymidine incorporation was determined by liquid scintillation counting. Shown are the results of 1 of 3 experiments with similar findings. B, FDC/P2-c-Mpl cells were cultured with human TPO (5 ng/ml) in the presence of IgG purified from 1) serum MA from an anti–c-Mpl–positive SLE patient or 2) serum ST from an anti–c-Mpl–positive ITP patient (80 μg/ml each). Each IgG fraction was pretreated with serial concentrations of c-Mpl immobilized on polyvinyl plates. Results shown are representative of 2 experiments. See Figure 1 for other definitions.

To confirm that the inhibitory effect was mediated through the anti–c-Mpl antibody in the IgG preparations, each fraction was pretreated twice with recombinant c-Mpl immobilized on polyvinyl plates and subsequently added to the cultures of FDC/P2-c-Mpl cells in the presence of human TPO (Figure 6B). An inhibitory effect of the IgG fractions in sera from 2 anti–c-Mpl–positive patients was cancelled when the IgG fractions were preincubated with immobilized c-Mpl.

We further examined the effect of IgG fractions obtained from anti–c-Mpl–positive sera on megakaryocyte colony formation (Figure 7). The IgG fractions from anti–c-Mpl–positive SLE serum MA and ITP serum ST reduced the number of megakaryocyte colonies stimulated with TPO, but this inhibitory effect was not observed in the IgG fraction from a healthy control.

Figure 7.

Inhibition of megakaryocyte colony formation by IgG fractions obtained from SLE and ITP patients with anti–c-Mpl antibodies. Bone marrow CD34+ cells were cultured for 14 days in serum-free soft agar with human thrombopoietin (TPO; 0.5 ng/ml) in the presence or absence of rabbit anti–c-Mpl polyclonal antibodies (5 μg/ml) as well as IgG purified from 1) serum MA from an anti–c-Mpl–positive SLE patient, 2) serum ST from an anti–c-Mpl–positive ITP patient, or 3) serum from a normal healthy control (NHC) (50 μg/ml or 250 μg/ml each). The cell aggregates were counted under a microscope. Values are the mean and SD and are representative of 2 experiments. See Figure 1 for other definitions.

Clinical course of an SLE patient with AMT and anti–c-Mpl antibody

Table 4 shows the platelet count and the serum anti–c-Mpl antibody and TPO levels over time in serum FY from an SLE patient who had AMT. In September 1997, the patient was admitted to our hospital because of a disease flare involving rash, neurologic disorder, and thrombocytopenia. A bone marrow examination revealed the absence of megakaryocytes with normal myeloid and erythroid series. The serum was anti–c-Mpl antibody positive at that time, but stored serum obtained in May 1993 was negative. After corticosteroid pulse therapy combined with intravenous cyclophosphamide, the platelet count increased and the serum became negative for the anti–c-Mpl antibody. The serum TPO levels were positively correlated with the anti–c-Mpl antibody levels and negatively correlated with the platelet count during the disease course. Platelet-associated anti–GPIIb-IIIa antibodies were not detected at any of the time points. The clinical course of this particular patient further supported the association between impaired thrombopoiesis and anti–c-Mpl antibodies.

Table 4. Serial measurement of platelet counts, serum anti–c-Mpl antibody levels, and serum TPO levels in an SLE patient with amegakaryocytic thrombocytopenia*
 May 1993October 1997February 1998March 1999
  • *

    TPO = thrombopoietin; SLE = systemic lupus erythematosus; ND = not determined; anti–GPIIb-IIIa = anti–platelet surface glycoprotein IIb-IIIa.

  • Normal <18.0 units.

  • Normal <3.3 units.

No. of platelets/mm3131,00021,000103,000128,000
Anti–c-Mpl antibody, units10.184.111.49.8
Serum TPO level, pg/ml499>2,000626ND
Platelet-associated anti–GPIIb-IIIa antibody, unitsND2.41.31.9

DISCUSSION

This is the first report describing the presence of autoantibodies to c-Mpl in patients with SLE. Anti–c-Mpl antibody was associated with thrombocytopenia, especially thrombocytopenia with megakaryocytic hypoplasia, but the frequencies of other clinical and laboratory findings as well as the disease activity were the same for SLE patients with and without the anti–c-Mpl antibody. Anti–c-Mpl antibodies in patients' sera interfered with TPO function by blocking its ligation to c-Mpl expressed on the transfected mouse leukemia cell lines as well as on human hematopoietic stem cells in vitro, indicating a pathogenic role of this autoantibody in megakaryocytic hypoplasia. It is also possible that the anti–c-Mpl antibody mediates a cytotoxic activity toward megakaryocytes through the Fc receptor–dependent and/or the complement-dependent pathway in the bone marrow. The involvement of the anti–c-Mpl antibody in impaired thrombopoiesis was further supported by the clinical course of an SLE patient with AMT, in whom the platelet counts were negatively correlated with the serum anti–c-Mpl antibody levels.

A variety of self proteins, mostly nuclear and cytoplasmic proteins, are targeted by autoantibody responses in SLE patients, but functional impairment through direct binding to the target autoantigenic structure has not been proven for the majority of SLE-related autoantibodies (12). Therefore, the anti–c-Mpl antibody is a rare example of SLE-related autoantibodies that alter the normal function of the target molecules.

Nearly half of the serum samples that reacted with c-Mpl by ELISA did not bind to c-Mpl in its denatured form in immunoblots. This discordant result cannot be explained simply by the difference in sensitivity between these two detection methods, because the anti–c-Mpl antibody levels measured by ELISA were similar for serum samples that were positive by immunoblotting and those that were negative. Furthermore, the anti–c-Mpl antibodies in the patients' sera were shown to bind to the c-Mpl molecule expressed on cell surfaces and to inhibit the TPO-induced cell proliferation, independent of their anti–c-Mpl antibody reactivity in immunoblots (see Table 3). These findings strongly suggest that pathogenic effect of the anti–c-Mpl antibodies is dependent on the antibody recognition of conformational epitopes expressed on the extracellular domain of c-Mpl.

Our results indicate that impaired thrombopoiesis induced by the anti–c-Mpl antibody is one of the mechanisms of thrombocytopenia in SLE patients. AMT, an extreme condition of impaired megakaryopoiesis, is not a distinct clinical entity, but rather, is a syndrome caused by a variety of pathogenic mechanisms, such as an intrinsic defect of the megakaryocyte progenitor cells as well as the suppression of megakaryocytic differentiation (8). There are many reports describing the efficacy of corticosteroids and immunosuppressive drugs for acquired AMT with and without an association with SLE (5, 7, 8, 25), suggesting that the immune process is a dominant mechanism for acquired AMT unrelated to exposure to drugs or toxic materials. In this regard, the presence of antibodies that suppress megakaryocyte colony formation in acquired AMT was included in several case reports (6, 8, 26). Anti–c-Mpl antibody was presumed to be involved in the pathogenic process in some of those patients. However, several groups have reported individual cases of acquired AMT that was probably due to suppression of megakaryocytes or their progenitor cells mediated by bone marrow accessory cells, including E-rosette–positive T cells (7), macrophages (27), and CD8+ T cells (28). Nevertheless, it is also possible that antibodies secreted by bone marrow B cells in vitro are involved in the cell-mediated suppression of megakaryocyte colony formation in their culture systems. Other possible mechanisms include autoantibodies that neutralize TPO, which have been reported in a patient with acquired AMT unrelated to SLE (23).

SLE patients who were positive for serum anti–c-Mpl antibodies had markedly elevated TPO levels in their circulation. This is in accordance with a recent report showing increased serum TPO levels in patients with acquired AMT, one of whom had concomitant SLE (29). TPO is constitutively produced mainly in the liver, and its serum concentration was initially thought to be regulated by its absorption by c-Mpl molecules expressed on the surfaces of circulating platelets (30). However, TPO levels are markedly elevated in patients with aplastic anemia and those with AMT, in whom bone marrow megakaryocytes are nearly absent, while only a slight increase in TPO levels is seen in ITP patients (24, 31). These findings led to an alternative theory that megakaryocyte counts, rather than platelet counts, primarily control the serum TPO levels (31). Therefore, the elevated TPO concentration in sera from SLE patients with anti–c-Mpl antibodies may simply reflect a decrease in the number of bone marrow megakaryocytes. Alternatively, the anti–c-Mpl antibody may compete with TPO for the binding site on c-Mpl. The present study also showed increased serum TPO levels in SLE patients lacking anti–c-Mpl antibodies compared with healthy individuals independent of thrombocytopenia, but the precise reason for the elevated TPO levels in these patients remains to be clarified.

Anti–c-Mpl antibodies were detected not only in SLE patients, but also in ITP patients who had a normal or increased number of megakaryocytes in their bone marrow smears. This finding appears to be inconsistent with the association between anti–c-Mpl antibodies and megakaryocytic hypoplasia in SLE patients, since it has been shown that thrombocytopenia in ITP patients results from an accelerated destruction of platelets in the peripheral circulation, mainly in reticuloendothelial systems (32).

One of the possible explanations for this inconsistent finding is contamination of AMT patients in our ITP patients, because evaluation of megakaryocyte mass using bone marrow smears is not accurate in certain cases, and some of our ITP patients with anti–c-Mpl antibody showed extremely high serum TPO levels. Alternatively, several kinetic studies that were carefully performed using autologous platelets revealed the existence of ineffective thrombopoiesis in a significant proportion of ITP patients (33, 34). Moreover, the presence of antibodies specifically bound to megakaryocyte membranes was reported in ITP patients (35). This finding implies that the impaired thrombopoiesis in ITP patients could be partly accounted for by the presence of antibodies reactive with megakaryocyte-related antigens, such as c-Mpl. In fact, Hoffman and colleagues described a patient with ITP who later developed AMT associated with antibodies that suppressed megakaryocyte colony formation (26).

Four of 7 SLE patients with anti–c-Mpl antibodies in our study also had platelet-associated anti–GPIIb-IIIa antibodies, which have been shown to be specific to ITP (36) and which have also been detected in SLE patients with thrombocytopenia (37). Because GPIIb-IIIa is known to be the main target of pathogenic antiplatelet autoantibodies in ITP patients (17), impaired thrombopoiesis may not be the sole mechanism for thrombocytopenia in patients who have both anti–c-Mpl and anti–GPIIb-IIIa antibodies. A small amount of c-Mpl is expressed on the surface of circulating platelets as well as on megakaryocyte surfaces (38), whereas GPIIb-IIIa is expressed in abundance on circulating platelets and to a lesser extent on megakaryocytes. Therefore, both anti–c-Mpl and anti–GPIIb-IIIa antibodies are regarded as antibodies to megakaryocyte/platelet lineage cells. Interestingly, their biologic effects on thrombopoiesis are entirely different and dependent on the expression levels of target antigens on circulating platelets versus bone marrow megakaryocytes. Further studies are necessary to clarify the mechanism by which autoantibody responses to multiple molecules expressed on megakaryocyte/platelet lineage cells are induced in a subset of SLE patients.

Acknowledgements

We thank Mutsuko Ishida for assisting in the RNA immunoprecipitation assay and Tetsuya Hagiwara for assisting in the megakaryocyte colony assay.

Ancillary