Anti-Fc Gamma Receptor Autoantibodies from Patients with Sjögren's Syndrome do Not React with Native Receptor on Human Polymorphonuclear Leukocytes

Journal of Autoimmunity (1996) 9, 181–191

Anti-Fc Gamma Receptor Autoantibodies from Patients with Sjögren’s Syndrome do Not React with Native Receptor on Human Polymorphonuclear Leukocytes Armelle Lamour1, Rozenn Le Corre1, Claude Soubrane,2 David Khayat,2 and Pierre Youinou1 1

Laboratory of Immunology, Brest Medical School Hospital, Brest 2 Department of Oncology, La Pitié-Salpétrière Hospital, Paris, France

Received 24 October 1995 Accepted 8 December 1995 Key words Sjögren’s syndrome, Fcã receptor, anti-Fc ã receptor antibody

Sera from patients with primary Sjögren’s syndrome (pSS) have been examined for the presence of cell-free Fc-gamma receptor (FcãR) IIIb, which is expressed in polymorphonuclear leukocytes (PMN), and the production of related autoantibody. Sera from 66 patients with pSS were evaluated by an ELISA using recombinant human FcãRIIIb as the substrate and by flow cytometry. Cell-free FcãRIIIb was also detected by an ELISA. The fine specificity of autoantibodies was established by inhibition with a preparation of FcãRI plus FcãRII, and two ELISAs using FcãRI and FcãRII as the substrates respectively. Anti-FcãRIIIb activity was found in 30 patients (45%), but 25 of them did not react with autologous PMN, whereas they bound to FcãRIIIb eluted from the same PMN in ELISA and Western blotting. Autoantibodies from one serum recognized the three receptors, six with FcãRII in addition to FcãRIII, and three sera were specific for the latter receptor. None of these reacted with FcãRI- and FcãRII-carrying cells. Cell-free FcãRIIIb, but negligible amounts of FcãRIIIa, were detectable in the patient sera. The membrane expression of CD15, an early activation marker, was diminished, while that of three PMN late activation markers was markedly enhanced. Taken together, these results suggest that autoantibodies are produced following the shedding of FcãRIIIb upon PMN activation. A credible candidate for this activation is IgG-containing immune complexes. © 1996 Academic Press Limited

Introduction

as a phosphatidyl inositol (PI)-anchored protein in polymorphonuclear leukocytes (PMN). The NA1/ NA2 and NB1+/NB1− alloantigen systems of FcãRIII are restricted to PMN. The latter cells are important in inflammation and immune complex (IC) clearance. Their functional status is reflected by the reduction of early markers, such as CD15, and the enhancement of late activation markers [8], viz. CR1 (CD35), CR3 (CD11b) and LFA-1 (CD11a). FcãRIIIb can be shed upon activation [9], and has indeed been identified [10–12] in the serum of patients with RA, SS and systemic lupus erythematosus (SLE). Furthermore, the FcãRIIIb-positive PMN population has been shown to be functionally more active than its FcãRIIIb-negative counterpart [13]. Autoantibodies and IC remain appealing candidates for the activation of PMN. Autoantibodies directed against different classes of FcãR have been found in the sera of autoimmune mice [14] and patients [15] using truncated recombinant murine FcãRII as the antigen. In this study, we have detected autoantibodies in patients with SS, using a novel ELISA and recombinant human FcãRIIIb

3

Sjögren’s syndrome (SS) is an autoimmune exocrinopathy in which lymphocyte infiltration and autoantibody production are prominent features [1, 2]. This syndrome may occur as a primary disorder or in association with rheumatoid arthritis (RA) or any other connective tissue disease. Both conditions give rise to an extensive clinical spectrum, encompassing restricted exocrinopathy through systemic presentation. Fc-gamma receptor III (CD16), defective in paroxysmal nocturnal hemoglobinuria (PNH), constitutes one of three structurally distinct families of receptors for the Fc domain of IgG (FcãR). FcãRI and FcãRII are designated CD64 and CD32, respectively [3–7]. FcãRIII is present in two different isoforms: IIIa which exists as a conventional transmembrane receptor on NK and tissue macrophages, and IIIb which presents Correspondence to: Prof. P. Youinou, Laboratory of Immunology, Brest University Medical School Hospital, BP 824, F29 609 Brest Cedex, France 187 0896-8411/96/020181+11 $18.00/0

© 1996 Academic Press Limited

A. Lamour et al.

182

(rhu FcγRIIIb). Autoantibodies predominated in those patients with activated PMN and high levels of cellfree FcγRIIIb and cross-reacted with FcγRI and FcγRII. The possibility, therefore, arises that such autoantibodies may account for the defective FcγR-mediated clearance by the reticuloendothelial system described in patients with SS [16].

Materials and methods Patients and controls Sixty-six patients with primary SS were enrolled in the study, all fulfilling the European preliminary criteria for the classification of SS [17]. The diagnosis was confirmed by labial and/or sublingual salivary gland biopsy [18]. None of the patients met clinical or serological criteria for additional connective tissue disease. All but one were women and their ages ranged from 25 to 74 years (mean 46). None of the patients suffered from acute viral or bacterial infection at the time of study. Normal sera were obtained from 34 members of the clinical or laboratory staff (14 men and 20 women whose mean age was 37 years) for the anti-FcγRIIIb assays. Normal controls in the cell-free FcγRIIIb assay consisted of members of the laboratory staff and residents of a home for the elderly (there were 8 men and 16 women, and their mean age was 45 years). This study was approved by the Brest Medical School Institutional Review Board.

Production of rhuFcãRIIIb The production of rhuFcγRIIIb has been described in detail elsewhere[4]. Briefly, a CD16-encoding fulllength cDNA-containing plasmid was used as the template and two oligonucleotides prepared to amplify the segment corresponding to the extracellular part of the molecule (residues 17 to 211). A BamHI fragment was inserted into the vector M13 mp18 at the EcoRI site. One of the recombinant clones was selected to prepare a 608 bp fragment and inserted into the vector pET3a at the NdeI and BamHI sites and transfected into E. coli-C 600. Following culture growth, bacterial cells were lysed, and DNA obtained by 50 U/ml of benzonase. Washed inclusion bodies were dissolved and the solubilized material filtered on 8 µm millipore membrane and, after dialysis, through 0.2 µm filter. SDS-PAGE under reducing conditions showed a silver-stained major band with a m.w. of approximately 27 kDa. Automated NH2 terminal sequencing established the presence of the nine expected residues.

Elisa autoantibody tests Prior to testing, all the serum samples were extensively absorbed with pooled human umbilical vein endothelial cells to remove irrelevant antibodies. To prevent non-specific binding of IgG through its Fc portion in the IgG autoantibody test, rhuFcγRIIIb was denatured by heating at 95°C for 5 min and treatment

with 15% 2-ME which was subsequently removed by dialysis. Microtiter plates (Dynatech, Marnes-laCoquette, France) were coated with 100 µl of this preparation (1 µg/ml in citrate buffer, pH 3.5). After an overnight incubation at 4°C, the plates were washed four times with PBS containing 0.05% Tween (PBS-T) and blocked with PBS supplemented with 5% BSA for 1 h at 37°C. Following another four washes, serum samples diluted 1/100 in PBS-T were dispensed into the wells and left for 90 min at room temperature. After washing, bound antibodies were detected by flooding the wells with horseradishperoxidase (HRP)-conjugated F (ab′)2 anti-human IgG or IgM (Dakopatt, Copenhagen, Denmark). After development, the absorbance was read at 492 nm on a Titertek Multiskan microplate reader (Flow Laboratories, McLean, VA). MAb IgG1 (Immunotech, Marseille, France) and IgM (Becton Dickinson, Mountain View, CA) directed towards FcγRIIIb were used as positive controls in the IgG and IgM autoantibody tests, respectively, and revealed with HRPconjugated goat anti-mouse IgG or IgM antibodies (Dakopatt). Three dilutions were tested, each assay performed in triplicate and the results averaged. FcγRI and FcγRII were first co-purified and quantitated using the Micro-BCA Protein Assay (Pierce, Rockford, IL). Ten sera were selected on the basis of high levels of anti-rhuFcγRIIIb antibody and an aliquot was incubated with 0.1–1.0 µg/ml of the FcγRI plus FcγRII preparation. The percentage of inhibition in the anti-rhuFcγRIIIb ELISA was calculated in comparison with the result obtained with an untreated aliquot. Purified FcγRI and II were denatured as described above and used as capture agents in the anti-FcγRI and anti-FcγRII antibody tests. Plates were coated with 100 µl of either antigen (5 µl/ml in carbonatebicarbonate buffer, pH 9.6) for 2 h at 37°C and overnight at 4°C. They were then washed four times with PBS-T, blocked with PBS-BSA for 1 h at 37°C, washed another four times and flooded with serum samples diluted 1/100 in PBS-T for 2 h at room temperature. Both tests were conducted, as described in the anti-rhuFcγRIIIb ELISA. A HRP-conjugated F(ab′)2 anti-human Ig (Dakopatts) was used as revealing agent.

Cell preparation After informed consent was obtained, blood samples from patients and controls were drawn into EDTA tubes prepared in the laboratory. Volunteers from the laboratory staff and patients were phenotyped by J. Cartron (Kremlin-Bicêtre Hospital, Paris, France), using specific workshop antisera to identify NA1+/ NA2−/NAB1+, NA1−/NA2+/NB1+ and NA1+/ NA2+/NAB1− individuals [19]. Cell suspensions containing more than 90% PMN were obtained by Dextran T500 sedimentation (Pharmacia, Uppsala, Sweden), followed by Ficoll-hypaque density gradients centrifugation (Eurobio, Paris, France) and hypotonic lysis of residual erythrocytes. Mononuclear cells were also collected and used in one experiment. The

Expression of anti-FcγR antibodies in pSS patients

cell viability exceeded 98%, as determined by staining with ethidium bromide (Sigma Chemical Co, Saint Louis, MO). In some experiments, to enhance FcγRIIIb expression, control PMN were stimulated with zymosan-activated autologous plasma overnight at 37°C, and then washed three times in PBS. NK were obtained from an RA patient [20] presenting with a severe neutropenia and expanded populations of large granular lymphocytes (LGL). These were isolated by continuous Percoll (Pharmacia) density gradient centrifugation, as described by Timonen and Saksela [21]. The LGL fraction contained approximately 40% CD16-positive cells. The FcγRI- and FcγRII- expressing [22] monoblastlike U937 cells and the FcγRII-expressing [23] K562 cells were originally purchased from the American Type Culture Collection (Rockville, MD) and maintained in RPMI 1640 (Eurobio) supplemented with 10% heat-inactivated FCS (Labsystem, CergyPontoise, France), 100 U/ml penicillin 100 µg/ml streptomycin and 200 mmol/l L-glutamine (BioMérieux, Marcy-L’Etoile, France) at 37°C with 5% CO2. To increase the amount of surface FcγRI, U937 cells were treated with 250 IU/ml of IFN-γ (Genzyme, Cambridge, MA) for 18 h in the culture. The cells were harvested after chilling the flasks for 30 min at 4°C.

Flow cytometric detection of autoantibodies NA1+/NA2−/NB1+, NA1−/NA2+/NB1+ and NA1+/NA2+/NB1− donors and two anti-FcγRIIIB autoantibody-positive patients were selected. Their PMN were collected, purified, fixed for 2 h with cold 2% paraformaldehyde (PFA) to prevent non-specific binding via FcγR, and adjusted at 5×106 cells/ml in PBS supplemented with 2% BSA and 0.1% sodium azide (PBS-BSA-azide). In some experiments, the PMN were not fixed. NK included in the mononuclear cell population were used as a control. One hundred µl of this cell suspension was incubated with 50 µl of the two anti-FcγRIIIb autoantibody-positive sera diluted 1/10 in PBS for 30 min at 37°C. The serum from a normal donor served as a negative control and that of a patient with idiopathic neutropenia as a positive control. After three washes in PBS-BSA-azide, FITC-conjugated F(ab′)2 anti-human Fcγ or antihuman Fcµ (Amersham, Les Ulis, France) was added, together with biotin (B)-labelled anti-CD35 (CR1) MAb (Immunotech) to identify PMN or phyco-erythrin-conjugated anti-CD56 (NKH1) MAb (Immunotech) to identify NK. After a 30 min incubation at 37°C and three washes, allo-phyco-cyanin (APC)-conjugated streptavidin (Molecular Probes Inc, Eugene, OR) was added. Following another three washes, the cells were analysed in a flow cytometer equipped with 500 mW argon and helium-neon lasers (Coulter, Hialeah, FL).

Cell elution of FcγR FcγRIIIb were eluted from normal PMN. Cells from three selected donors were adjusted at 2×106 cells/ml in PBS and incubated with 0.1 IU/ml phospholipase C

183

(Sigma) for 2 h at 37°C, with constant stirring. The preparations were clarified by centrifugation and the supernatants concentrated. The levels of cell-free FcγRIIIb were determined by the assay described below, and were 0.38, 0.36 and 0.45 µg/ml, respectively. The eluted FcγRIIIb were then denatured and coated onto the plates. The above two anti-FcγRIIIbpositive sera were tested in the IgM-specific ELISA. An anti-FcγRIIIb IgM MAb served as a positive control and a normal serum as a negative control. FcγRIIIa and FcγRI plus FcγRII were extracted from LGL and U937 cells, respectively. Cells were washed three times in cold PBS. The resulting pellet was chilled in an ice-water bath and lysed with 0.5% Nonidet P-40 (Sigma) in the presence of protease inhibitors, i.e., 10 mM iodoacetamide, 0.3 IU/ml aprotinin, 1 µM pepstatin and 2 mM PMSF (all from Sigma, except iodoacetamide from Merck-Clevenot, Nogent, France). The lysates were clarified by centrifugation at 30,000×g for 40 min at 4°C.

SDS-PAGE and Western blotting RhuFcγRIIIb and eluted FcγRIIIb were electrophoresed on 1-mm thick 5 to 10% polyacrylamide gel [24]. The protein bands were transferred to 0.45 µm nitrocellulose filters at 15 V constant voltage for 12 h at room temperature (Bio-Rad, Ivry-sur-Seine, France). The filters were quenched with 5% nonfat dry milk in PBS-T for 1 h at room temperature and washed three times with PBS-T. FcγRIIIb was identified by incubating the nitrocellulose sheets with autoantibodypositive sera from patients, overnight at 4°C with gentle agitation. The monoclonal IgM anti-FcγRIII Leu11b (Becton Dickinson) served as a positive control, and the serum of a normal individual as a negative control. After three further washes, B-labelled F(ab′)2 anti-human IgM or anti-mouse IgM (Dakopatt) were added and developed with HRPconjugated streptavidin (Amersham). Autoantibody binding was revealed by the enhanced chemiluminescence technique (Amersham, Les Ulis, France) and a 5-second exposure to blue light sensitive autoradiography film.

Cell-free FcγRIIIb assay As previously described in detail [25], a sandwich ELISA was used with slight modifications. Briefly, plates were coated with 100 µl/ml of 30 µl/ml IgG1 3G8 kindly donated by J. C. Unkeless (Mount Sinai School of Medicine, New York). After an overnight incubation at 4°C, plates were washed four times with PBS-T, and 100 µl of serum diluted 1/10 in PBS were applied for 4 h at room temperature. The plates were washed again, and 100 µl of 0.8 µg/ml IgM antiLeu11b MAb was added. Following a 2 h incubation at room temperature, the plates were washed and developed with 100 µl of 1/2,500 dilution of HRPconjugated goat F(ab′)2 anti-mouse IgM (Dakopatts). After development, the absorbance was read at 492 nm. The standard curve was established by serial dilutions of rhuFcγRIIIb. Eluted FcγRIIIb served as a

A. Lamour et al.

184

positive control and the serum of a patient with HPN as a negative control. The description of two RA patients with high levels of soluble FcγRIIIa [26] urged us to evaluate the contribution of this receptor to the cell-free FcγRIII assayed using the above technique. Since we did not have access to MAb to discriminate FcγRIIIa from FcγRIIIb, specific ELISAs could not be set up. The NA1 and NA2 allotypes can, however, be distinguished from each other, but FcγRIIIa which is not polymorphic always types as NA2 [3]. Given that the m.w. of FcγRIIIb of the NA1 allotype and FcγRIIIa are overlapping [27] we selected three NA2 homozygous patients. Their sera were submitted to SDS-PAGE and blotted. The presence of cell-free receptors was then determined on the basis of their m.w. in the silverstained gels and revealed with anti-Leu11b MAb on the blot. LGL-eluted FcγRIIIa (45 kDa) and NA2 homozygous PMN-eluted FcγRIIIb (70 kDa) served as positive controls, while normal serum was used as a negative control.

Preparation of pure receptors The U937 lysates were sequentially applied to a Sepharose-4B (Pharmacia) and a Sepharose-BSA column to remove non-specifically binding debris. The flow-through was collected and FcγRI and FcγRII affinity purified with an IgG column made of proteinA-eluted IgG from Cohn fraction II (Sigma) coupled with cyanogen bromide-activated Sepharose beads. Such FcγRI plus FcγRII preparations were used in the inhibition experiments. FcγRI was then separated from FcγRII on an antiFcγRII column. Anti-CD32 MAb (kindly donated by M. Hirn, Immunotech) was pepsin-digested. The F(ab′)2 fragments were recovered by passing the preparation over a protein-G column and a G-100 column (both from Pharmacia). The purity of F(ab′)2 fragments was shown by SDS-PAGE, and the binding activity ascertained by testing the reagent with IFN-γ activated PMN in the flow cytometer and FITClabelled F(ab′)2 anti-mouse light chain antibody (Amersham) as a second layer. The anti-CD32 column was made by standard procedure. The mixture of FcγRI and FcγRII was run three times through this column. FcγRI was collected in the effluent and FcγRII eluted with 0.5 N CH3COOH and 0.05% NP40 in Tris, pH 8.6. Both FcγRI and FcγRII were homogenous, as judged by SDS-PAGE. Western blotting with antiCD32 and anti-CD64 MAb pooled together developed the corresponding bands.

anti-mouse Ig (Jackson, Avondale, PA). All these MAb were purchased from Immunotech. Patient PMN were adjusted at 107/ml in PBS-BSAazide, and 100 µl of the cell suspension incubated with 5 µl of each MAb for 30 min at 4°C in the dark. After two washes in PBS-BSA-azide, APC-streptavidin and FITC-(Fab′)2 anti-mouse antibody were added to B-labelled and purified MAb, respectively, and left for 30 min at 4°C. Following another three washes, the cells were PFA-fixed and passed through a nylonmesh filter before being enumerated in the flow cytometer. Appropriate settings of forward and orthogonal scatter gates were employed to examine 10,000 PMN per test. The measurement conditions were controlled by using unstained cells, isotypematched non-reactive FITC- or B-conjugated MAb. Antigen densities were indirectly ascertained by the mean fluorescence intensity (MFI) of the cells analysed. Values of MFI were obtained, but because we were aware of daily variations of MFI, the results were expressed as ratios of MFI of patients’ cells to those of controls.

Statistics All experiments were consistent in triplicate and variation between values was less than 10%. Data are presented as arithmetic mean and SEM. Because of the skewed distribution of results, non-parametric tests were applied. Significance was established by the chi-square test (with Yates’ correction when required), and the Mann-Whitney test. Correlations were calculated by the Spearman rank test.

Results ELISA detection of anti-FcγRIIIb antibodies The mean IgM and IgG autoantibody activities were 0.186±0.016 (P<0.001, compared with normal controls) and 0.153±0.014 (P<0.001), respectively, in patients, and 0.096±0.008 and 0.088±0.008 (mean± SEM), respectively, in normal controls (Figure 1). Assuming that the IgM cutoff level is 0.190 (mean of the normal controls +2 SD), autoantibodies were detected in 24 patients (36%). With regard to the IgG, if the cutoff level is 0.174, autoantibodies were present in 16 patients (24%). Elevated levels of at least one autoantibody isotype (IgM only in 14 patients, IgG only in 6 and IgM plus IgG in 10) were thus present in 45% of the patients with SS (Table 1). There was, however, a limited correlation between IgM and IgG anti-FcγRIIIb activity (r=0.24, P<0.05).

Flow cytometric cell surface analysis of PMN

Flow cytometric study of autoantibodies

The reagents used in this experiment were FITCconjugated anti-CD16 (FcγRIIIB), anti-CD11a (LFA-1) and anti-CD67 (PI-linked p100) MAb, B-labelled anti-CD35 MAb developed with APC-conjugated streptavidin, purified anti-CD15 (PMN early marker), anti-CD11b (CR3) and anti-CD48 (PI-linked gp41) MAb developed with FITC-conjugated F(ab′)2 goat

Twenty-five sera were found to be negative (Table 2) and five positive using control (5.4±0.5 and 63.2±1.9%) and autologous PMN (6.7±0.5 and 62.2±2.7%). The proportions of PMN from three controls as well as autologous PMN recognized by anti-FcγRIIIb autoantibody-positive sera were similar to the results obtained with the negative control,

Expression of anti-FcãR antibodies in pSS patients

IgM

IgG

0.3

0.2

0.1

ticular autoantibodies to the cells. Experiments were repeated with non-fixed PMN with exactly the same results. In contrast, when the FcγRIIIb were eluted from the PMN from the three normal donors (donors 1, 2 and 3 in Figure 3) and used to coat ELISA plates, autoantibodies from two prototype patients (patients 1 and 2 in Figure 3) did recognize their antigens. Similar results were obtained with seven sera. As previously established [22] and shown on the blots (Figure 4), the m.w. of eluted FcγRIIIb was 70 kDa whereas the non-glycosylated rhuFcγRIIIb migrated as two bands of smaller mass, between 20 and 30 kDa. The autoantibodies reacted with both of them, whilst the MAb Leu11b bound only to the lower m.w. band.

3

Cell-free Fcγ R III b (µg/ml)

Anti-Fcγ R III b activity OD (492 nm)

0.4

185

2

1

Binding specificity 0.0

0 P<0.001

P<0.001

P<0.001

Figure 1. Anti-Fc gamma receptor III b(FcγRIIIb) IgM and IgG autoantibodies are shown on the left panel (OD: bars denote arithmetic means). Plates were coated with heat-denatured recombinant human (rhu) FcγRIIIb. Serum samples diluted 1/100 were dispensed into the wells and left for 90 min at room temperature. Bound antibodies were detected by flooding the wells with horseradish peroxidaseconjugated F(ab′)2 anti-human IgG or IgM. MAb IgG1 and IgM directed to human FcγRIIIb were used as positive controls. Cell-free FcγRIII levels are given on the right panel. As described in detail (see ref. 24) the 3G8 MAb was used as a capture agent and the anti-Leu 11 b MAb as the revealing agent. The standard curve was constructed using serial dilutions of rhuFcγRIIIb (µg/ml). (●) patients (•) controls.

Table 1. Incidence of sera from patients with primary Sjögren’s syndrome that react with human Fc-gamma receptor IIIb Isotype IgM IgG IgM+IgG Total

Number of positive samples

Percent of positive samples

14 6 10 30

21 9 15 45

Plates were coated with denatured recombinant human Fc gamma receptor IIIb (FcγRIIIb). Serum samples diluted 1/100 were dispensed into the wells and left for 90 min at room temperature. Bound antibodies were detected by flooding the wells with horseradish peroxidase-conjugated F(ab′)2 anti-human IgG or IgM. Anti-FcγRIIIb, monoclonal IgG1 and IgM were used as positive controls.

irrespective of the NA and NB phenotype of the cells in 25 of the 30 sera previously found to be reactive in the ELISA, whereas those from the remaining five patients and one positive control with idiopathic neutropenia stained 58–68% of the PMN in the flow cytometric assay. Neither was there any binding to the FcγRIIIa-expressing NK. Representative examples are given in Figure 2. Enhancement of FcγRIIIb expression on control PMN with zymosan-activated autologous plasma did not influence the binding of these par-

The target of autoantibodies in 10 of the sera containing anti-FcγRIIIb IgG that did not react with the native receptor on PMN was further determined (Figure 5). The binding of autoantibodies from seven sera to the rhuFcγRIIIb coated onto the plate was significantly inhibited (30.1±7.3%) by overnight incubation at 4°C with a mixture of 0.1 µg/ml of denatured FcγRI and 0.1 µg/ml of denatured FcγRII. It was then established by using the FcγRI-specific assay and the FcγRIIspecific assay that autoantibodies from one serum recognized the three receptors (Table 3), from six sera reacted with FcγRII in addition to FcγRIIIb, and from the remaining three proved to be specific for FcγRIIIb. This was not unexpected, because autoantibodies from the latter sera did not bind the FcγRI-FcγRII preparation in the inhibition experiment. Interestingly, none of these ten sera bound to the FcγRI-FcγRIIexpressing U937 cells, or to the FcγRI-expressing K562 cells in the flow cytometric assay (data not shown).

Cell-free FcãRIIIb On average, there were 0.807±0.115 µg/ml cell-free FcγRIIIb in the serum of patients (Figure 1), compared with 0.174±0.021 µg/ml in that of the normals (P<0.001). There were correlations between the levels of cell-free FcγRIIIb and the titers of IgM (r=0.51, P<0.01) and IgG (r=0.28, P<0.02) anti-FcγRIIIb antibodies (Figure 6). To address the question as to whether the measurements of cell-free FcγRIIIb were independent of the FcγRIIIa concentrations, three NA2-homozygous patients were selected to be analysed further. As shown in a representative example (Figure 7, lane 4), their serum did not display any 45 kDa band that would have reflected the presence of cell-free FcγRIIIa. This implies that the amount of cell-free FcγRIIIa is negligible in the serum of patients with SS.

Expression of the membrane markers on the PMN The density of the CD16 molecules (Table 4), as determined by MFI, was reduced on the PMN of patients (0.84±0.07), i.e. the ratios of CD16 MFI of

A. Lamour et al.

186

Table 2. Flow cytometric detection of anti-polymorphonuclear leukocyte (PMN) autoantibodies Percentages of stained PMN (mean±SEM) Serum samples

Control PMN NA1+/NA2+/NB1−

Autologous PMN

63.2±1.9 5.4±0.5 58 2

62.2±2.7 6.7±0.5 65 4

Positive patients (n=5) Negative patients (n=25) Positive control (n=1) Negative control (n=1)

PMN of normal donors and patients were phenotyped using specific polyclonal anti-sera. NA1+/NA2−/ NB1+, NA1−/NA2+/NB1+ and NA1+/NA2+/NB1− donors and two anti-FcγRIIb autoantibody-positive sera were selected. One hundred µl of these cell suspensions were incubated with 50 µl of the two sera. The serum of a normal subject served as a negative control and that of a patient with idiopathic neutropenia as a positive control. FITC-conjugated F(ab′)2 anti-human Fcγ or anti-human Fcµ was added. Autoantibodies were sought on a flow cytometer.

1000

A

1

2

B

1

2

APC

APC

100 10 1 A 3 1

–1 1000

C

B 10 FITC

100

1

4 1000 –1 2

D

3 1

10 FITC

100

1

4 1000 2

APC

PE

100 10 1 E –1

3 1

10 FITC

100

4 1000 –1

H 3 1

10 FITC

100

4 1000

Figure 2. Staining patterns of human polymorphonuclear leukocytes (PMN) and NK with IgM anti-FcγRIIIB from a patient with Sjögren’s syndrome (A, C, D) and a positive control serum from a patient with idiopathic neutropenia (B), revealed by a second layer of FITC-labelled F(ab′)2 anti-human IgM. PMN were untreated (A, B) or stimulated with zymosan-activated plasma (C). Cells were double-stained with biotin-anti-CD35 revealed by allo-phyco-cyanin (APC)-conjugated streptavidin for PMN (A, B, D) and phyco-erythrin (PE)-conjugated anti-CD56 for NK (C).

patients’ cells to those of controls were lower than 1 in 85% of the patients. These ratios correlated inversely with the cell-free FcγRIIIb levels (r= −0.42, P<0.05). The PI-linkage of FcγRIIIb could not account for the FcγRIIIb shedding, since irrelevant PI-anchored markers, such as CD48 and CD67, were normally expressed at the membrane level. The membrane expression of the early PMN marker CD15, was reduced in a number of patients. In contrast, the densities of late activation markers, such as CD35, CD11b and CD11a, were markedly augmented. These differences in density can be clearly seen in representative examples (Figure 8).

Discussion In this study, we show that autoantibodies to FcγRIIIb are detectable, using rhuFcγRIIIb. These were of the IgM and/or the IgG classes in a number of patients. Autoantibodies from seven out of ten sera, crossreacted with FcγRII or with FcγRI as well as FcγRII, so that the majority were not specific for FcγRII. Our results confirm and extend those reported by Boros et al., who used truncated mouse FcγRII to search for antibodies in autoimmune mice [14] and patients [15]. Anti-PMN autoantibodies have also been detected in the serum of patients with juvenile neutropenia [28],

Expression of anti-FcγR antibodies in pSS patients

1.12

187

100

1.10

1.00

0.75

Patient 2

0.50

Positive control

80

Percent inhibition

OD (492 nm)

Patient 1

0.25 Negative control 0.00

Donor 1

Donor 2

Donor 3

1

2

3

4

5

40

20

Figure 3. IgM autoantibodies recognize FcγRIIIb eluted from NA1+/NA2−/NB1+ (donor 1), NA1−/NA2+/NB1+) (donor 2) and NA1+/NA2+/NB1− (donor 3) polymorphonuclear leukocytes. These were adjusted at 2×106 cells/ml and incubated with 0.1 IU/ml phospholipase C for 2 h at 37°C. The eluted FcγRIIIb were then denaturated and coated onto the plates. Two anti-FcγRIIIb autoantibody-positive sera (patients 1 and 2) were tested in the IgM-specific ELISA, as described in the legend of Figure 1. An anti-FcγRIIIb IgM MAb served as a positive sera. The monoclonal IgM anti-Fcγ RIII Leu11b served as the positive control and a normal serum as a negative control.

MW (kD)

60

6

200

0

1.0

0.5 Inhibitor (µg/ml)

0.2

0.1

Figure 5. Inhibition (percentages) of the binding of antirhuFcγRIIIb, following incubation of 10 strongly positive sera with serial dilutions of a combination of FcγRI and FcγRII. The latter receptors were eluted from U937 cells and affinity-purified by passing the preparation over a human IgG column.

Table 3. Binding specificity of 10 sera reacting with recombinant human (rhu) Fc-gamma receptor (FcγR)IIIb Class of FcγR

97.4

IIIb

II

I

+ + + +

+ + − −

+ − + −

Number of positive %

69 46

30

Total

1 6 0 3 10

(10) (60) (0) (30) (100)

21.5

Figure 4. IgM autoantibodies bind to recombinant human (rhu) FcγRIIIb (lane 2) and eluted FcγRIIIb (lane 5). A positive (lanes 1 and 4) and a negative (lanes 3 and 6) controls have been run in parallel. Rhu FcγRIIIb and eluted FcγRIIIb (see legend of Figure 3) were electrophoresed and transferred to 0.45 µm nitro-cellulose filters. FcγRIIIb was identified by incubating the sheet with autoantibodypositive sera. The monoclonal IgM anti-FcγRIII Leu11b served as the positive control and the serum of a normal individual as a negative control. Biotin-labelled F(ab′)2 antihuman or anti-mouse IgM were added and developed with horseradish peroxidase-conjugated streptavidin.

primary biliary cirrhosis [29] and infection with human immunodeficiency virus [30]. The latter observations do not of course imply that all these anti-PMN antibodies are directed towards FcγRIIIb.

rhuFcγRIIIb, FcγRI or FcγRII (both obtained by elution from U937 cells followed by sequential affinity chromatography) were used as capture agents in three separate ELISAs and the antibody binding revealed with horseradish peroxidase-conjugated goat anti-human Ig.

In contrast to the autoantibodies described by Boros et al. in patients with SLE, SS and systemic scleroderma [15], in 25 of our 30 ELISA-positive cases, the antibody was non-reactive with the cell-bound FcγRIIIb, irrespective of the NA and NB phenotype of the PMN donors. Neither did they bind to FcγRIIIacarrying NK nor to various cell lines carrying FcγRI and/or FcγRII. Intriguingly, the same autoantibodies bound to FcγRIIIb clipped off from the PMN of the donors and the patients whose cells had not been recognized in the flow cytometric experiment. This discrepancy suggests that epitopes are cryptic at the cell surface level, or that changes in FcγRIIIb occur

A. Lamour et al.

188

Cell-free Fcγ R III b (µg/ml)

3

2 IgM r = 0.51 P < 0.01

IgG r = 0.28 P < 0.02

1

0.2

0

0.4 0.6 0.0 0.2 Anti-Fcγ R III b autoantibodies = OD (492 nm)

0.4

0.6

Figure 6. Relationships of cell-free FcγRIIIb to the corresponding IgM and IgG autoantibodies. The former was assessed as described in the legend of Table 3, and the latter as described in that of Figure 1.

1

2

3

4

MW (kD) 200

Table 4. Mean fluorescence intensity (MFI) of the membrane markers on the polymorphonuclear leucocytes (PMN) PMN membrane markers

Number of patients

Mean± SEM

17 15 17 16 16 16 15

0.84±0.07 1.30±0.17 1.42±0.34 0.77±0.34 2.41±0.42 1.65±0.15 1.31±0.07

97.4 69

46

30

21.5

Figure 7. Western-blotting analysis of a selected pathological serum to demonstrate cell-free Fc-gamma receptor (FcγR). The serum of NA2-homozygous patients with high amount of cell-free FcγR was electrophoresed and blotted (lane 4). NA2 homozygous polymorphonuclear leukocyteeluted FcγRIIIb (lane 1) and large granular leukocyte-eluted FcγRIIIa (lane 2) served as positive controls. A normal serum was used as a negative control (lane 3). Cell-free FcγR were identified according to their m.w. and revealed with the anti-Leu 11b MAb.

after shedding. The released FcγRIIIb might be attached to a membrane fragment or molecule, or linked to its ligand. It is unlikely that epitopes become accessible through linkage to IgG or IgG-containing IC, inasmuch as eluted FcγRIIIb were obtained in serum-free medium and proved to be as efficient as the purified rhuFcγRIIIb in the ELISA.This has to be put together with the correlation between the titers of autoantibodies and the concentrations of cell-free FcγRIIIb. It is thus proposed that, beyond a certain threshold, these receptors self-aggregate through lectin-sugar residue associations to trigger the autoantibody production [31]. MAb raised against peptide

CD16 CD48 CD67 CD15 CD35 CD11b CD11a

Patient PMN were adjusted at 107/ml and 100 µl of the cell suspension incubated with 5 µl of each MAb for 30 min. Biotin-labelled anti-CD35 MAb was developed with allophyco-cyanin-conjugated streptavidin. Other purified MAb were developed with FITC-conjugated F(ab′)2 goat anti-mouse Ig. Antigen densities were indirectly ascertained by the MFI of the cells analyzed. The results were expressed as ratios of MFI of patients’ cells to those of controls.

fragments of FcγRIIIb [32] will be useful to further define the target of these autoantibodies. Such experiments are in progress in our laboratory. Since no specific MAb were available, we could not distinguish the relative contributions of FcγRIIIa and FcγRIIIb in the cell-free FcγR. However, the former receptor was not detectable on the blot from three selected patients. We may thus conclude that most, if not all, cell-free FcγR is derived from PMN, which is in line with the results obtained by Huizinga et al. [9]. Cell-free FcγRIIIb originates from release by PMN upon activation. There was indeed an inverse correlation between its quantity and the MFI of this receptor on the PMN membrane. Our negative control with PNH lacked FcγRIIIb on her PMN as well as her serum [33]. Schopf et al. have also described a selective FcγRIIIb defect in a single patient [34]. Additional support for this interpretation is that there is an increased cell-free FcγRIIIb concentration following PMN stimulation [35]. Our results suggest that

Expression of anti-FcγR antibodies in pSS patients

189

CD15

Count

Patient

Control C

CD11a

Control

Patient

C

0.1

1

10

100

FITC

Figure 8. Representative profiles of the polymorphonuclear leukocyte (PMN) markers in a patient and a control (CD15 which is a PMN early activation marker is at the top and CD11a which is a late activation marker at the bottom). Purified anti-CD15 MAb was developed with FITCconjugated F(ab′)2 goat anti-mouse Ig. Anti-CD11a (LFA-1) MAb was FITC-conjugated.

activated PMN may release FcγRIIIb. However, the PI-linkage in itself is not a sufficient condition, given that the MFI of two irrelevant PI-anchored proteins in PMN was normal. The high concentrations of cell-free FcγRIIIb at sites of inflammation, such as bronchoalveolar lavage fluids from patients with adult respiratory distress syndrome and synovial fluids from RA patients [36] support the concept that the quantity of cell-free FcγRIIIb are reflects of the activational state of PMN. There is also a report of soluble FcγRII [37]. The levels of autoantibody could conceivably been underestimated by cell-free FcγR interfering with the assays. Autoantibodies are unlikely to trigger PMN activation, as has been described for several murine antiFcγR MAb [38]. A credible candidate agent for this activation and the ensuing release of FcγRIIIb is IgGcontaining IC. Soluble aggregates activate PMN that have previously been primed, while the insoluble aggregates activate primed and unprimed PMN [39]. The following sequence could thus be operating: IC binding to FcγRIIIb on PMN, cell activation, shedding of FcγRIIIb, and self-aggregation of these receptors leading to autoantibody production. Some of the autoantibodies might be polyspecific, as those made by mouse hybridomas [40] or human CD5-positive B-cells [41]. Alternatively, FcγRIIIb shedding might result in a more efficient binding of IgG to the PMN

FcγRII, as suggested by Huizinga et al. [35]. It is relevant to this hypothesis that IC have been detected in patients with PNH [42] possibly as a result of the absence of FcγRIIIb on the PMN membrane. The lipid linkage of FcγRIIIb strongly argues that their release is influential in PMN function. Despite the increased risk from infections, such as pneumonia, oral candidiasis and bacterial conjunctivitis in SS, the evidence is conflicting [43, 44], as to whether PMN activity is defective in this disease. In fact, the sole FcγRIIIb function substantiated thus far is the release of granule proteins. Boros et al. have demonstrated that anti-mouse FcγRII mouse IgM MAb trigger PMN enzyme secretion via human FcγRIIIb [40], suggesting a possible role for similar antibodies in autoimmune inflammatory processes. Chronic neutropenia is uncommon in SS [45] and the causation of anti-PMN autoantibody has hitherto been documented in one patient [46]. Some of ours displayed mild neutropenia; cell-free FcγRIIIb was more often detectable in these patients than in the remainder. The interrelation between these two abnormalities is unclear. A systematic survey of SS patients has been undertaken in our laboratory and experiments, including the measurement of PMN life span, designed to delineate mechanisms. There is no agreement as to the cause of the reduced IC clearance in SS and related diseases. This has been shown to be associated with the HLA-B8/DR3 haplotype [16]. Steric hindrance of receptors by autoantibody cannot be offered as an explanation, since their target seems not to be accessible at the membrane level. Rather, binding of autoantibody and/or cell-free FcγRIIIb to circulating IC could prevent their capture and removal by the reticuloendothelial system. Study of FcγRIII-bearing macrophages in the liver and the red pulp of the spleen would be of obvious interest to address this issue.

Acknowledgements We gratefully acknowledge J. C. Unkeless for generously providing the 3G8 MAb, M. Colona for his gift of rhuFcγRIIIb and M. Hirn for kindly supplying the anti-CD32 MAb. Thanks are also due to J. Cartron for phenotyping the controls and to C. Jamin for technical help. The expert secretarial assistance of I. Guillou and A. Paul is appreciated.

References 1. Talal N., Moutsopoulos H.M. Kassan S.S. 1987. Sjögren’s syndrome, clinical and immunological aspects. Springer-Verlag, Berlin 2. Talal N. 1989. Sjögren’s syndrome. A model for understanding autoimmunity. Academic Press, New York 3. Fanger M.W., Graziano L., Shen L., Guyre P.M. 1989. FcγR in cytotoxicity exerted by mononuclear cells. Chem. Immunol. 47: 214–221 4. Fanger M.W., Shen L., Graziano R.F., Guyre P.M. 1989. Cytotoxicity mediated by human Fc receptors for IgG. Immunol. Today 9: 92–97

190

5. Unkeless J.C., Scigliano E., Freedman V.H. 1988. Structure and function of human and murine receptors for IgG. Annu. Rev. Immunol. 6: 251–281 6. Lin C.T., Shen Z., Boros P., Unkeless J.C. 1994. Fc receptor mediated signal transduction. J. Clin. Immunol. 14: 1–13 7. Fridman W.H., Bonnerot C., Dearon M., Amigorna S., Teillaud J.L., Sautès C. 1992. Structural bases of Fc gamma receptor functions. Immunol. Rev. 125: 49–76 8. Youinou P., Lamour A., Dumay A., Le Goff P. 1992. Neutrophil polymorphonuclear cell function in rheumatoid arthritis. In Rheumatoid arthritis, recent research advances. Smolen J.S., Kalden J.R., Maini R.N., eds. Springer-Verlag, Berlin, p. 150–163 9. Huizinga T.W.J., van der Schoot C.E., Jost C., Klaassen R., Kleijer M., von dem Borne A.E.G.Kr., Roos D., Tetteroo P.A. 1988. The PI-linked receptor FcRIII is released on stimulation of neutrophils. Nature 333: 667–669 10. Khayat D., Le Goff P., Soubrane C., Magadur G., Youinou P. 1990. Fc gamma receptor III shedding by polymorphonuclear cells in rheumatoid arthritis. Med. Sci. Res. 18: 893–894 11. Lamour A., Soubrane C., Ichen M., Pennec Y.L., Khayat D., Youinou P. 1993. Fc-gamma receptor III shedding by polymorphonuclear cells in primary Sjögren’s syndrome. Eur. J. Clin. Invest. 23: 97–101 12. Hutin P., Lamour A., Pennec Y.L., Soubrane C., Dien G., Khayat D., Youinou P. 1994. Cell-free Fc-gamma receptor III in sera from patients with systemic lupus erythematosus: correlation with clinical and biological features. Int. Arch. Allergy Immunol. 103: 23–27 13. Youinou P., Jouquan J., Muzellec Y., Le Goff P. 1987. Polymorphonuclear cell Fc-gamma receptors in rheumatoid arthritis. In Protides of the biological fluids. H. Peeters, ed. Pergamon Press, Oxford, pp. 251–254 14. Boros P., Chen J., Bona C., Unkeless J.C. 1990. Autoimmune mice make anti-Fc gamma receptor Ig. J. Exp. Med. 171: 1581–1595 15. Boros P., Muryoi T., Spiera H., Bona C., Unkeless J.C. 1993. Autoantibodies directed against different classes of Fc gamma R are found in sera of autoimmune patients. J. Immunol. 150: 2018–2024 16. Hamburger M.I., Haralampos M., Moutsopoulos H.M., Thomas J., Lawley T.J., Frank M.M. 1979. Sjögren’s syndrome: a defect in reticuloendothelial system Fc-receptor specific clearance. Ann. Intern. Med. 91: 534–538 17. Vitali C., Bombardieri S., Moutsopoulos H.M., Balestrieri G., Bencivelli W., Bernstein R., Coll J., De Vita S., Hatron P.Y., Hay F.C., Isenberg D.A., Janin A., Kalder J.R., Dater L., Konttinen Y.T., Maddison P., Maini R.N., Manthorpe R., Meyer O., Ostuni P., Pennec Y.L., Prause J.U., Sauvezie B., Schiodt M., Sciutto M., Shoenfeld Y., Skopouli F.N., Smolen J.S., Snaith M., Tincani A., Tishler M., Todesco S., Valesini G., Venables P.J.W., Wattiaux M.J., Youinou P. 1993. Diagnostic criteria for Sjögren’s syndrome: results of an European Multicenter prospective study. Arthritis Rheum. 36: 340–347 18. Pennec Y.L., Leroy J.P., Jouquan J., Lelong A., Katsikis P., Youinou P. 1990. Comparison of labial and sublingual salivary gland biopsies in the diagnosis of Sjögren’s syndrome. Ann. Rheum. Dis. 49: 37–39

A. Lamour et al.

19. Lucas G.F., Carrington P.A. 1990. Results of the first international granulocyte serology workshop. Vox. Sang. 59: 251–256 20. Loughran T.P., Starkebaum G., Kidd P., Neiman P. 1988. Clonal proliferation of large granular lymphocytes in the rheumatoid arthritis. Arthritis Rheum. 31: 31–39 21. Timonen T., Saksela E. 1980. Isolation of human NK cells by density gradient centrifugation. J. Immunol. Methods 36: 285–291 22. Looney R.J., Abraham G.N., Anderson C.L. 1986. Human monocytes and U937 cells bear two distinct Fc receptors for IgG. J. Immunol. 136: 1641–1647 23. Tetteroo P.A.T., van der Schoot C.E., Visser F.J., Bos M.J.E., von dem Borne A.E.G.Kr. 1987. Three different types of Fcγ receptors on human leukocytes defined by Workshop antibodies: FcγR low neutrophils, Fcγ low of K/NK lymphocytes and FcγRII. In Leukocyte typing III. White cell differentiation antigens. McMichael A.J., ed. University Press, Oxford, pp. 102–107 24. Laemmli U.K. 1970. Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227: 680–685 25. Khayat D., Geffrier C., Yoon S., Scigliano E., Soubrane C., Weil M., Unkeless J.C., Jacquillat C. 1987. Soluble circulating Fcγ receptors in human serum. A new ELISA assay for specific and quantitative detection. J. Immunol Methods 100: 235–241 26. de Haas M., Kleijer M., Michinson R.M., Roos D., von dem Borne A.E.G.Kr. 1994. Soluble FcγRIIIa is present in plasma and is derived from natural killer cells. J. Immunol. 152: 900–907 27. Ravetch J.V., Perussia B. 1989. Alternative membrane forms of Fc gamma RIII (CD16) on human natural killer cells and neutrophils. Cell type-specific expression of two genes that differ in single nucleotide substitutions. J. Exp. Med. 170: 481–497 28. Lalezari P., Khorshidi M., Petrosova M. 1986. Autoimmune neutropenia of infancy. J. Pediatr. 109: 764–769, 1986 29. Bux J., Robertz-Vaupel G.M., Glasmacher A., Dengler H.J., Mueller-Eckhardt C. 1991. Autoimmune neutropenia due to NA1 specific antibodies in primary biliary cirrhosis. Brit. J. Haematol. 77: 121–126 30. Klaassen R.J.L., Vlekke A.B.L., von dem Borne A.E.G.Kr. 1991. Neutrophil-bound immunoglobulin in HIV infection is of autoantibody nature. Brit. J. Haematol. 77: 403–409 31. Kurosaki T., Ravetch J.V. 1989. A single amino acid in the glycosyl phosphatidyl inositol attachment domain determines the membrane topology of FcγRIII. Nature 342: 805–806 32. Mouawad R., Visonneau S., Soubrane C., Hirel P.H., Autran B., Marceau-Suissa J., Weil M., Khayat D. 1992. Antibodies raised against peptide fragments of CD16 reacting with membrane and soluble form of CD16. I: Production and characterization. Hybridoma 11: 447–459 33. Selvaraj P., Rosse W.F., Silber R., Springer T.A. 1988. The major Fc receptor in blood has a phosphatidylinositol anchor and is deficient in paroxysmal noctural haemoglobinuria. Nature 333: 565–567 34. Schopf R.E., Rehder M., Laux B., Korting G.W. 1987. Functional Fc-receptor defect of polymorphonuclear

Expression of anti-FcγR antibodies in pSS patients

35.

36.

37.

38.

39.

40.

leukocytes in a patient with Sjögren’s syndrome. Klin. Wochenschr. 65: 342–344 Huizinga T.W.J., de Haas M., Kleijers M., Huijens J.N., Roos D., von dem Borne A.E.G.Kr. 1990. Soluble FcγRIII in human plasma originates from release by neutrophils. J. Clin. Invest. 86: 416–423 Fleit H.B., Kobasiuk C.D., Daly D., Furie F., Levy P.C., Webster R.O. 1992. A soluble form of FcγRIII is present in human serum and other body fluids and is elevated at sites of inflammation. Blood 79: 2721–2728 Ierino F.L.P., Miach P.J., Hogarth P.M., McKenzie I.F. 1992. Serum soluble Fc gamma RII levels in dialysis and renal transplant recipients. Transplant. Proc. 24: 2328–2330 Anegon I., Cuturi M.C., Trinchieri G., Perussia P. 1988. Interaction of Fc receptor (CD16) ligands induces transcription of interleukin 2 receptor (CD25) and lymphokine genes and expression of their products in human natural killer cells. J. Exp. Med. 167: 452–458 Robinson J.J., Watson F., Phelan M., Bucknall R.C., Edwards S.W. 1993. Activation of neutrophils by soluble and insoluble immunoglobulin aggregates from synovial fluid of patients with rheumatoid arthritis. Ann. Rheum. Dis. 52: 347–353 Boros P., Odin J.A., Muryoi T., Masur S.K., Bona C., Unkeless J.C. 1991. IgM anti-FcγR autoantibodies trigger neutrophil degranulation. J. Exp. Med. 173: 1473–1482

191

41. Youinou P., Buskila D., Mackenzie L.E., Shoenfeld Y., Lydyard P.M. 1993. CD5-positive B cells and disease. In Natural autoantibodies. Their physiological role and regulatory significance. Schoenfeld Y., Isenberg D.A., eds. CRC Press Inc, Boca Raton. pp. 143–165 42. Villaescusa R., Santos M.N., Garcia Y., Trujillo Y., Bernal B., Bernal J.M., Ballester J.M., Hernandez P. 1982. Circulating immune complexes in paroxysmal noctural hemoglobinuria. Acta Haemat. 68: 136–141 43. Pennec Y.L., Gentric A., Jouquan J., Youinou P. 1988. Polymorphonuclear cell functions in patients with primary Sjögren’s syndrome. Clin. Exp. Rheumatol. 6: 293–296 44. Gudbjörnsson B., Feltelius N., Hällgren R., Venge P. Neutrophil function in patients with primary Sjögren’s syndrome: relation to infection propesity. Ann. Rheum. Dis. 50: 685–690 45. Ramakrishna R., Chaudhuri K., Sturgess A., Manoharan A. 1992. Haematological manifestations of primary Sjögren’s syndrome: a clinicopathological study. Q.J.M. 84: 547–554 46. Yamato E., Fujioka Y., Masugi F., Nakamaru M., Tahara Y., Kurata Y., Ogihara T. 1990. Case report: autoimmune neutropenia with anti-neutrophil autoantibody associated with Sjögren’s syndrome. Am. J. Med. Sci. 300: 102–103.