ASN.2004070609.Full.Pdf

JASN Express. Published on May 18, 2005 as doi: 10.1681/ASN.2004070609

Membrane Proteinase 3 Expression in Patients with Wegener’s Granulomatosis and in Human Hematopoietic Stem Cell–Derived Neutrophils

Adrian Schreiber,* Bjoern Otto,* Xinsheng Ju,† Martin Zenke,† Ursula Goebel,* Friedrich C. Luft,* and Ralph Kettritz* *HELIOS Klinikum-Berlin, Franz Volhard Clinic and Max Delbru¨ck Center for Molecular Medicine, Medical Faculty of the Charite´, Humboldt University of Berlin, Berlin; and †Helmholtz Institute of Biomedical Engineering, Department of Cell Biology, University Hospital of Aachen, Aachen, Germany

A large membrane proteinase 3 (mPR3)-positive neutrophil subset (mPR3high) is a risk for Wegener’s granulomatosis (WG). The relationship between mPR3 expression and clinical manifestations was investigated in 81 WG patients and mPR3 expression was studied in CD34؉ stem cell–derived human neutrophils. The mPR3high neutrophil percentage correlated with renal function, anemia, and albumin at the time of presentation. The mPR3high neutrophil percentage and renal failure severity correlated directly after 5 yr. For elucidating mechanisms that govern mPR3 expression, studies were conducted to determine whether the genetic information that governs mPR3 expression resides within the neutrophils, even without stimuli possibly related to disease. CD34؉ hematopoietic stem cells were differentiated to neutrophils, and their mPR3 expression was determined. A two-step amplification/differentiation protocol was used to differentiate human CD34؉ hematopoietic stem cells into neutrophils with G-CSF. The cells progressively expressed the neutrophil surface markers CD66b, CD35, and CD11b. The ferricytochrome C assay demonstrated a strong respiratory burst at day 14 in response to PMA but none at day 0. Intracellular PR3 was detectable from day 4 by Western blotting. An increasing percentage of a mPR3-positive neutrophil subset became detectable by flow cytometry, whereas a second subset remained negative, consistent with a bimodal expression. Finally, human PR3–anti-neutrophil cytoplasmic autoantibodies induced a stronger respiratory burst, compared with human control IgG in stem cell–derived neutrophils. Taken together, these studies underscore the clinical importance of the WG mPR3 phenotype. The surface mPR3 on resting cells is probably genetically determined rather than being dictated by external factors. J Am Soc Nephrol 16: ???–???, 2005. doi: 10.1681/ASN.2004070609

nti-neutrophil cytoplasmic autoantibodies (ANCA) mPR3high neutrophils is stable during a person’s lifetime and are found in patients with systemic small-vessel vas- does not change appreciably with activation status or the neu- A culitis (1–3). A pathogenic role of ANCA interacting trophil’s age (14–16). A high percentage of mPR3high neutro- with ANCA–antigen-containing neutrophils and monocytes phils was a risk factor for vasculitis, was associated with re- was suggested by numerous in vitro studies (4–10). Further lapse in a Dutch cohort of patients wit WG, and resulted in evidence for a central role of ANCA as a pathogenetic factor of stronger neutrophil activation by PR3 ANCA in vitro in earlier vasculitis was recently shown in an animal model of the disease studies (15–18). Thus, a detailed understanding of mechanisms (11). ANCA directed against proteinase 3 (PR3) are found in that control this dual mPR3 expression pattern is important. We Wegener’s granulomatosis (WG) (12,13). Intracellular PR3 is recently described a strong genetic influence on mPR3 expres- translocated from granules to the cell membrane upon neutro- sion in monozygotic and dizygotic twins, where a high within- phil activation. However, some PR3 is also expressed on cell pair correlation was found only in monozygotic twins (16). membranes of nonactivated resting neutrophils. It is interesting However, whether the genetic information that determines the that the total neutrophil population of a given individual can be percentage of mPR3high neutrophils resides in the cell or in the high divided into a membrane PR3-positive (mPR3 ) and a mem- extracellular host milieu is not known. The aim of this investi- low brane PR3-negative (mPR3 ) (14). Although the amount of gation was to explore the relationship between mPR3 pheno- mPR3 changes with activation, the individual percentage of type and clinical course in our WG patient cohort. In addition, we investigated whether hematopoietic stem cells could be differentiated into neutrophils and whether this differentiation Received July 28, 2004. Accepted March 17, 2005. would provide a feasible approach to study mPR3 expression

Published online ahead of print. Publication date available at www.jasn.org. and ANCA-induced neutrophil activation. Finally, we tested the hypothesis that the stem cells were innately equipped with Address correspondence to: Dr. Ralph Kettritz, Franz Volhard Clinic, Wiltberg- low high strasse 50, 13122 Berlin, Germany. Phone: ϩ49-30-9417-2202; Fax: ϩ49-30-9417- the information to generate a mPR3 and mPR3 neutro- 2206; E-mail: [email protected] phil subset.

Materials and Methods FCS, 2 mM l-glutamine, 0.1 mM 2-mercaptoethanol, 100 U/ml penicil- Patients and Control Subjects lin and streptomycin (Life Technologies-BRL), and 10 ng/ml G-CSF for ϫ 6 We included 81 PR3-ANCA–positive WG patients. At the time of 16d(1 10 cells/ml). Every 2 d, growth factors were added and cells ϫ 6 mPR3 phenotyping, 66 of the 81 patients had a positive ANCA test. The were maintained at 1 10 cells/ml cell density. After differentiation diagnosis was made on the basis of the criteria of the Chapel Hill into neutrophils, duplicate samples were used for all further assays. Consensus Conference (19) and the American College of Rheumatology Stem cell factor and thrombopoietin were from Amgen (Thousand (20). PR3-ANCA was assayed by indirect immunofluorescence on eth- Oaks, CA), and Flt3 ligand and G-CSF was from PeproTech (London, anol-fixed neutrophils and by PR3-specific enzyme-linked immunoad- UK). Hyper–IL-6 was produced in yeast as described previously (21). sorbent assay. Relapse was defined as a rapid rise in creatinine levels accompanied by urinary sediment activity, the detection of active vas- Determination of Surface Antigen Expression by Flow culitis, or glomerulonephritis; pulmonary hemorrhage or expanding Cytometry nodules; the observation of iritis or uveitis; or new mononeuritis mul- FACS was used as described previously to evaluate the expression of tiplex. Relapse occurred in 53% of the patients. A total of 154 healthy surface molecules on neutrophils (18). Briefly, cells were spun down at subjects were assayed for mPR3 expression and served as control. 200 ϫ g for 7 min at 4°C. Pellets were resuspended in HBSS without ϩ ϩ Ca2 /Mg2 before they were incubated with dilutions of the indicated Materials antibodies. In case the primary antibodies were not FITC conjugated, Recombinant TNF-␣ was obtained from Genzyme (Ru¨sselsheim, Ger- we used a secondary FITC-conjugated F(ab)2 fragment of goat anti- many). The monoclonal mouse antibody to PR3 was obtained from CLB mouse IgG. Flow cytometry was performed on the same day using a (Amsterdam, Netherlands), and FITC-conjugated F(ab)-fragment of FACScan (Becton Dickinson, Heidelberg, Germany), and 10,000 events goat anti-mouse IgG was from DAKO (Hamburg, Germany). Dextran per sample were collected. was purchased from Amersham Pharmacia (Amsterdam, Netherlands). HBSS, PBS, and trypan blue were from Seromed (Berlin, Germany). Preparation of Immunoglobulins Histopaque 1083 was obtained from Sigma-Aldrich (Deisenhofen, Ger- Human IgG was prepared from one patient with biopsy-proven WG many). The following antibodies were used: CD11b (FITC-conjugated) (PR3-ANCA) as well as from one healthy donor as described previ- and CD66b (FITC-conjugated) both from Immunotech (Krefeld, Ger- ously (18). Plasma samples were obtained from freshly drawn blood many) and CD35 (FITC-conjugated) from Cymbus Biotech (Hants, UK). and kept at Ϫ20°C. Plasma was filtered through a 0.2-␮m syringe filter Endotoxin-free reagents and plastic disposables were used in all exper- (Gelman Sciences, Ann Arbor, MI) and applied to a HiTrap protein G iments. affinity column (Pharmacia, Uppsala, Sweden). Bound IgG was eluted with 0.1 M glycine-HCl buffer (pH 2.75; elution buffer). After the Isolation of Human Neutrophils antibodies emerged, the pH was immediately adjusted to pH 7.0 using Neutrophils from healthy volunteers were isolated from heparinized 1 M Tris-HCl (pH 9.0). A mouse mAb to myeloperoxidase (MPO; whole blood by red blood cell sedimentation with dextran 1%, followed MPO-7, IgG1) and an isotype-matched control (IgG1) were purchased by Ficoll-Hypaque density gradient centrifugation and hypotonic from Dako (Hamburg, Germany). Before use, IgG preparations were ϫ erythrocyte lyses. Neutrophils were centrifuged (10 min at 1050 rpm) centrifuged at 10,000 g for 5 min to remove aggregates. ϩ and resuspended in HBSS with calcium and magnesium (HBSS2 ). The cell viability was detected by trypan blue exclusion and exceeded 99%. Western Blot Analysis of PR3 The neutrophil percentage in the suspension was Ͼ95% by Wright- Cells were lysed with 20 ␮l of ice-cold lysing buffer (20 mM Tris-HCl Giemsa staining. [pH 8.0] that contained 138 mM NaCl, 1% Triton X-100, 2 mM EDTA, 10% glycerol, 0.2 mM sodium orthovanadate, 1 mM PMSF, 10 ␮g/ml ␮ Differentiation of Human Hematopoietic Stem Cells into aprotinin, 10 g/ml leupeptin, 0.1 mM quercetin, and 5 mM Iodoacet- Neutrophils amide). Samples were kept on ice for 5 min, supernatant was recovered ϩ ϫ Granulocytes were obtained from CD34 stem cells of cord blood by centrifugation at 13,000 g for 5 min at 4°C, and protein concen- from two different donors and from peripheral blood of three different tration was measured by BCA protein assay (Pierce, Munich, Ger- healthy volunteers by using a two-step amplification/differentiation many). Loading buffer (250 mM Tris-HCL [pH 6.8] with 4% SDS, 20% ϩ protocol. CD34 cells were isolated using the MACS system (Miltenyi glycerol, and 0.01% bromphenol blue) was added, and samples were ␮ Biotec, Bergisch-Gladbach, Germany). In brief, mononuclear cells were heated for 5 min at 95°C. Each sample that contained 5 g of protein per recovered by Ficoll-Hypaque gradient centrifugation (density 1.077 lane was loaded on 15% SDS-polyacrylamide gel, electrophoresed, and g/ml) and incubated with FcR blocking reagent and hapten-conjugated transferred to nitrocellulose membranes. The membranes were blocked ϩ anti-CD34 antibody (15 min, 4°C). Cells were washed and reacted with in TBS-T nonfat dry milk 10% for 1 h and incubated overnight with microbeads conjugated to antihapten antibody (15 min, 4°C). CD34ϩ an antibody to PR3 (gift from Wieslab AB, Lund, Sweden; 1:2000 ϩ cells were obtained by two cycles of immunomagnetic bead selection, dilution) in TBS-T nonfat dry milk 1%. Membranes were washed and and selected cells were analyzed by flow cytometry. Purity was 83 to incubated with a secondary antibodies (horseradish peroxidase–labeled ϩ 90%. CD34 cells (0.3 to 0.5 ϫ 106 cells/ml) were cultured in StemSpan donkey anti-rabbit IgG [1:5000]; Amersham). Blot was developed by serum-free medium (StemCell Technologies Inc., Vancouver, BC, Can- incubation in a chemiluminescence substrate (ECL; Amersham) and ada) with 100 ng/ml stem cell factor, 50 ng/ml Flt3 ligand, 20 ng/ml exposed to x-ray film. thrombopoietin, and 10 ng/ml hyper–IL-6. Growth factors were added every 2 d, and cells were maintained at 1 ϫ 106 cells/ml cell density. Superoxide Generation Assay Cell numbers were determined at regular time intervals with an elec- Superoxide was measured using the assay of SOD-inhibitable reduc- tronic cell counter device (CASY1; Scha¨rfe Systems, Reutlingen, Ger- tion of ferricytochrome C as described by Pick and Mizel (22). Briefly, many). After 10 to 14 d of culture, progenitor cells were induced to neutrophils were pretreated with 5 ␮g/ml cytochalasin B for 15 min at differentiate into neutrophils in RPMI medium supplemented with 10% 4°C. Cells (0.75 ϫ 106) were primed with 2 ng/ml TNF-␣ for 15 min at J Am Soc Nephrol 16: ???–???, 2005 Membrane Proteinase 3 in Wegener’s Granulomatosis 3

37°C before human ANCA preparations were added. No priming was 81 PR3-ANCA–positive WG patients and 154 healthy control performed when cells were stimulated with PMA. The final concentra- subjects (Figure 1). The data confirm our earlier observations ␮ tions were 125 g/ml for purified IgG preparations and 25 ng/ml of obtained in a smaller cohort of 35 WG patients. They demon- PMA. All experiments were set up in duplicate. The samples were strate that the percentage of mPR3-positive neutrophils is incubated in 96-well plates at 37°C for up to 45 min, and the absorption skewed rightward in patients, compared with control subjects of samples with and without 300 U/ml SOD was scanned repetitively (P Ͻ 0.0001). Sixty-seven percent of the patients were analyzed at 550 nm using a Microplate Autoreader. The final ferricytochrome C concentration was 50 ␮M, and the final cell concentration was 0.75 ϫ during remission (Birmingham Vasculitis Activity Score 106/ml. [BVAS] 0), whereas 33% of the patients had active disease (mean BVAS 17). We analyzed 10 patients on two occasions, Cell Staining with Wright-Giemsa and Microscopy five during both active (mean BVAS 9.8) and inactive disease Cytocentrifuge preparations of cells were fixed in 100% methanol (BVAS 0) and five additional patients at two different time and stained using modified Wright-Giemsa stain. Cells were assessed points, both during inactive disease (BVAS 0). The percentage by light microscopy. of mPR3 expression was similar in any given patient regardless of the disease activity. The five patients with active disease had Statistical Analyses 77 Ϯ 18.3% (SD) of mPR3-positive neutrophils. When they were Results are given as mean Ϯ SD. Comparisons were made by t test or retested during remission, their mPR3 expression was still 77 Ϯ Mann-Whitney U test as appropriate. Nonparametric correlation 10.1%. The five patients in remission, who were tested twice, (Spearman) was used to test for a coherence of mPR3 expression with had 84 Ϯ 13.7% mPR3 expression at the first measurement and other known variables. Results are given as a correlation coefficient 84 Ϯ 14.4% at the second time point. In contrast to the percentage with their corresponding P values. Patient data were collected in a of mPR3 positive neutrophils, the amount of mPR3 expressed Filemaker database. showed a trend toward higher expression during active disease (mean fluorescence intensity for mPR3 in remission 222.4 Ϯ Results 87.7 versus 276.6 Ϯ 122 in active disease; n ϭ 5; NS). Percentage of mPR3high Neutrophil Subset and Patient Characteristics We then analyzed mPR3 expression in relation to clinical and The demographic characteristics at the time of presentation laboratory findings at presentation and during the course of the are listed in Table 1. We studied neutrophil mPR3 expression in disease in the cohort. Table 2 shows a selection of parameters that were studied at presentation and their correlation with the percentage of PR3high neutrophils. We found that WG patients who had a higher PR3high subset presented with higher creat- Table 1. Demographic patient characteristics at the time inine, lower creatinine clearance as estimated by the Cockroft of presentationa formula, higher C-reactive protein, lower albumin, and more severe anemia. In addition, 18 patients initially required renal Parameter Value (n;%orϮ SD) replacement therapy. These 18 individuals showed a significant higher mPR3high neutrophil percentage, compared with the 63 Gender patients without the need for dialysis (86.6 Ϯ 11.4 versus 69.3 Ϯ male 42 (52%) 22.7% mPR3high neutrophils; P Ͻ 0.001). Furthermore, we ob- female 39 (48%) served a strong trend toward higher BVAS (P ϭ 0.051). No such total 81 (100%) Mean age (yr) 51.9 (Ϯ14.1) Organ manifestation kidney 72 (88%) nose 54 (66%) joints 50 (61%) lung 49 (60%) eye 32 (40%) ear 24 (30%) skin 24 (30%) peripheral nerves 18 (22%) Mean BVAS 19.7 (Ϯ7.2) Follow-up (mo) 79 (Ϯ55) No. of relapses Figure 1. Distribution of the percentage of membrane proteinase No. of patients with relapse 44 (53%) 3–positive neutrophils (mPR3high) in a cohort of 154 healthy one relapse 15 (18%) German subjects compared with 81 patients with Wegener’s two relapses 17 (21%) granulomatosis (WG). Isolated neutrophils were stained with more than two relapses 12 (15%) an mAb against PR3, a secondary FITC-conjugated goat anti- Ⅺ median relapse-free survival (mo) 21.5 mouse antibody, and analyzed by flow cytometry. , percentage of mPR3high neutrophils from healthy individuals; f, per- aBVAS, Birmingham Vasculitis Activity Score. centage of mPR3high neutrophils from WG patients. 4 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

Table 2. Correlation of the percentage of mPR3high neutrophils with selected laboratory valuesa

Parameter Value Correlation Coefficient P Value

Hb 6.9 Ϯ 1.3 Ϫ0.305 0.006 Thrombocytes 384 Ϯ 162 0.037 0.74 Leukocytes 12.4 Ϯ 5.7 0.124 0.27 Creatinine 314 Ϯ 340 0.351 0.001 Cockroft GFR 37.3 Ϯ 29 0.399 0.000 C-reactive protein 110 Ϯ 219 0.256 0.021 Total protein 67.8 Ϯ 8.5 Ϫ0.161 0.151 Albumin 35.7 Ϯ 7.8 Ϫ0.397 Ͻ0.0001

aValues are mean Ϯ SD. mPR3high, membrane proteinase 3-positive.

correlation was observed, for example, for thrombocytes, leukocytes, and total protein concentration. We next assessed the percentage of mPR3high subset and follow-up parameters. Patients were followed for a mean of 79 Ϯ 55 mo. We observed relapsing disease in 44 of 81 WG patients. Fifteen patients experienced one relapse, 17 patients had two relapses, and 12 patients had more than two relapses. We found no correlation between relapse number and percentage of PR3high neutrophils. In addition, no significant correlation was found with regard to time intervals to relapse and to BVAS at relapse. We did find, however, a significant positive correlation between the percentage of PR3high neutrophils and creatinine up to 5 yr and a negative correlation with the calcu- lated GFR up to 6 yr of follow-up.

ϩ Differentiation of Human CD34 Hematopoietic Stem Cells into Neutrophils We had shown previously that the percentage of PR3high neutrophil subset is stable in a given individual and, using a twin approach, that the mPR3 expression phenotype is genetically determined (16). To gain further insight into the regula- tion of mPR3 expression, we explored the possibility that ϩ CD34 hematopoietic stem cells from umbilical cord could be differentiated into functional neutrophils and asked whether ϩ these stem cells were equipped with the information to gener- Figure 2. Differentiation of CD34 hematopoietic stem cells ϩ ate an mPR3low and mPR3high neutrophil subset. Human um- with G-CSF. Expanded CD34 cells at day 0 and after 4 and ϩ bilical cord CD34 stem cells were amplified followed by neu- 14 d of treatment with G-CSF were cytospun and stained with trophilic differentiation with G-CSF. Wright-Giemsa staining Wright-Giemsa. Microscopy indicates the progressive appear- demonstrated progressive neutrophilic differentiation after 4 d ance of cells with morphologic signs of neutrophilic matura- (Figure 2). At day 14, the majority of neutrophils had the typical tion. A typical of two independent experiments is depicted. ϫ ϫ size and nuclear morphology of mature neutrophils. Magnification, 40 in left images; 100 in right images. In addition to morphologic characteristics of neutrophils, ϩ G-CSF–differentiated CD34 stem cells progressively expressed the neutrophil marker CD66b. Figure 3 shows increas- 90% of the cells expressed these molecules after 14 d of neutro- ing expression of membrane CD66b with a small percentage of philic differentiation by G-CSF. CD66b-positive cells at day 0 and 87.9 Ϯ 1% at day 14 (n ϭ 2). Neutrophils from the myelocyte stage on respond to various Flow cytometry analysis revealed that G-CSF treatment re- stimuli with activation of the NADPH oxidase system. We sulted in increased expression of additional neutrophilic mol- tested the ability of neutrophils that were differentiated from ϩ ecules such as ␤2-integrins (CD11b) and CD35, a marker of CD34 cells to generate superoxide upon stimulation with ϩ secretory vesicles (Figure 4). Whereas a negligible percentage of PMA (Figure 5). CD34 cells that were incubated in buffer cells stained positive for the respective marker at day 0, nearly control alone did not generate superoxide at day 0 or at day 14 J Am Soc Nephrol 16: ???–???, 2005 Membrane Proteinase 3 in Wegener’s Granulomatosis 5

ϩ Figure 3. Membrane expression of CD66b on CD34 hematopoietic stem cells that were treated with G-CSF. Flow cytometry of cells that were stained with a FITC-conjugated mAb to CD66b or an isotype control was performed at the indicated time points. Treatment with G-CSF resulted in progressive expression of CD66b. A typical of two independent experiments is shown.

ϩ Figure 5. Respiratory burst activity in CD34 hematopoietic stem cells before and 14 d after treatment with G-CSF. Super- oxide was measured with the ferricytochrome C assay, and data from the 45-min time point are depicted. These experiments demonstrate that PMA triggered respiratory burst activity in G-CSF–differentiated neutrophils but not in undifferen- ϩ ϩ Figure 4. Membrane expression of CD11b and CD35 in CD34 tiated CD34 stem cells (n ϭ 2). Left, cells were treated without hematopoietic stem cells that were treated with G-CSF. Flow PMA; right, cells were treated with PMA. cytometry of cells that were stained with a FITC-conjugated specific antibodies or an isotype control was performed. The data demonstrate no CD11b and CD35 expression in undiffer- ϩ PR3 Expression in Neutrophils Differentiated from Human entiated CD34 cells. In contrast, nearly 90% of the cells became ϩ CD34 Hematopoietic Stem Cells positive after 14 d of treatment with G-CSF. A typical of two We next studied whether the information to generate an independent experiments is depicted. mPR3low and mPR3high subset resides already within the human stem cell. Assaying intracellular PR3 by Western blot of G-CSF treatment. In contrast, PMA did trigger a strong (Figure 6) and by intracellular flow cytometry (data not shown), respiratory burst in cells that were treated with G-CSF for 14 d we found only very small amounts of PR3 in undifferentiated ϩ (37 Ϯ 4 nmol) but had no effect on undifferentiated CD34 cells cells. In contrast, PR3 was clearly detectable in G-CSF–treated Ϯ Ϫ ϫ 5 ϩ (3.0 0 nmol O2 /7.5 10 cells per 45 min). Taken together, CD34 cells from day 4. We found no PR3 on the membrane of these data indicate that G-CSF induces fully functional neutro- undifferentiated hematopoietic stem cells (Figure 7). However, ϩ phil differentiation from CD34 cells. differentiation with G-CSF resulted in an mPR3high and an 6 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

Discussion The important findings in our study are that patients with a higher percentage of mPR3high neutrophils showed significantly higher creatinine and C-reactive protein values and lower albumin and GFR determinations at presentation. Along the same lines, patients who initially required dialysis had a Figure 6. Intracellular PR3 content was compared by immuno- significantly higher percentage of mPR3high neutrophils com- blotting. Jurkat cells (Ju) that do not express PR3 served as pared with those without severe renal failure. We also observed negative control and human neutrophils (PMN) as positive ϩ a strong trend toward a higher BVAS; however, this tendency control. CD34 stem cells were differentiated with G-CSF and did not quite achieve significance. In addition, analysis of cre- subjected to immunoblotting for PR3 after 0, 4, 10, and 14 d of atinine and GFR during follow-up revealed that the course of treatment. The experiments indicate that significant intracellu- ϩ renal function was worse in patients with a higher percentage lar amounts of PR3 were detected in CD34 stem cells after 4 d high of G-CSF treatment. A typical of two independent experiments of mPR3 neutrophils. This finding has not yet been re- is depicted. ported. These data extend those of an earlier report (15). Be- cause mPR3 expression seems pivotal and genetically deter- ϩ mined, we next differentiated human CD34 progenitor cells to neutrophils. We showed that this differentiation is possible, also in humans. We also found that the differentiated cells not only exhibit mPR3 but also display the same bimodal distribution as found in humans with or without disease. ANCA are pathogenic in animal models and activate in vitro cytokine-primed neutrophils that express ANCA antigens (4,5,8,10,11,23–26). Neutrophil priming mobilizes PR3, the WG autoantigen, from its intracellular stores, resulting in an increased membrane expression. However, resting neutrophils Figure 7. Membrane PR3 expression was studied by flow cy- ϩ also express PR3 on the outer cell membrane. On the basis of tometry. CD34 hematopoietic stem cells were stained with an this mPR3 expression pattern, neutrophils can be divided into mAb to PR3 before and after 4 and 14 d of G-CSF treatment. high low ϩ an mPR3 and an mPR3 subset. This bimodal expression The data demonstrate that differentiation of CD34 cells with G-CSF resulted in two distinct mPR3-expressing populations. pattern is stable in a given individual, does not change with cell An mPR3low and an mPR3high subset can be distinguished. activation, and is genetically determined (15–17). Other inves- Data from one representative experiment are shown. tigators and our own group demonstrated that a large percentage of mPR3high neutrophils provides a risk factor for WG (15–17). We investigated the correlation between a large mPR3high subset and the clinical and laboratory disease manifestation in a cohort of 81 WG patients. In addition, we exam- low mPR3 subset of cells. At day 14 of differentiation, 13% of ined the feasibility of generating human neutrophils from ϩ mPR3-positive cells in experiment 1 and at 38% in experiment CD34 hematopoietic stem cells and tested the hypothesis that 2 were detected. No further increase was observed up to day 16, these stem cells differentiate into an mPR3high and an mPR3low when the experiments were stopped. subset. Finally, the ability of stem cell–derived neutrophils to re- Recently, we presented data showing that mPR3high neutro- spond to PR3-ANCA was tested. Because of the limited cell phils respond to PR3-ANCA, but not to other stimuli, with number, we used in both experiments an IgG preparation of increased phosphatidylinositol 3 kinase/Akt signaling and re- one PR3-ANCA patient and one control subject. Stimulation of spiratory burst activation (18). This observation provides one of ␣ TNF- –primed hematopoietic progenitor cell–derived neutro- probably several explanations as to why a large percentage of ␮ Ϯ high phils with 125 g/ml PR3-ANCA resulted in 23 12 nmol mPR3 neutrophils are a risk factor for WG. Rarok et al. (15) Ϫ Ϯ O2 /45 min at day 14, whereas control IgG triggered only 14 analyzed 89 WG patients and found that a high percentage of Ϫ ϭ 14 nmol O2 (n 2). In the second experiment, we stimulated mPR3-expressing neutrophils was significantly associated with the differentiated neutrophils in addition to the human prepa- an increase in relapses. In their study, no association with ration with an activating mAb to MPO. A total of 10 ␮g/ml of specific clinical manifestations was detected. Our results extend Ϫ the mAb to MPO resulted in 34 nmol O2 /45 min compared the observations from the Dutch cohort. Our data are also in Ϫ ␣ with 10 nmol O2 in TNF- –primed cells that were incubated agreement with published results from other studies, showing with an isotype control. These very preliminary data suggest that the renal function at initial presentation is the best predic- that stem cell–derived neutrophils could respond to PR3- and tor for renal function at follow-up (27–29). However, in contrast MPO-ANCA. This observation supports the contention that to the Dutch study, no significant correlation with the relapse ϩ CD34 cell–derived neutrophils may be used for studying both risk was found in our cohort. One difference between our study mechanisms of membrane PR3 expression and ANCA-induced and the Rarok data is that we observed different mPR3 expres- activation. sion patterns. In the Dutch cohort, only 30% of the patients and J Am Soc Nephrol 16: ???–???, 2005 Membrane Proteinase 3 in Wegener’s Granulomatosis 7

ϩ a similar percentage of control subjects showed a bimodal mPR3 expression, we monitored mPR3 expression in CD34 mPR3 expression pattern. We do not know the explanation for progenitor cells, which has not been studied before in any stem this difference between the German and the Dutch cohorts. cell line. Our data clearly indicate that during neutrophil dif- However, in a study of French patients, Witko-Sarsat et al. (17) ferentiation from stem cells, mPR3 expression is increased until found an mPR3 expression pattern that was very similar to our a stable level is achieved. However, some cells persistently cohort. The clinical data obtained from our study support the remained mPR3 negative, resulting in a bimodal mPR3 pattern contention that the mPR3 expression pattern is correlated with similar to what is observed in neutrophils that are isolated from clinical parameters at presentation of WG and, moreover, af- the circulation of human donors. This result suggests that stem fects renal outcome. Increased interaction of ANCA with cells contain the genetic information to generate both mPR3low mPR3high neutrophils resulting in stronger cell activation, as and mPR3high neutrophil subsets. Targeting candidate mole- found in our previous in vitro experiments, may provide some cules that may be involved in membrane insertion, such as explanation for these important clinical findings (18). However, SNARE by si-RNA, may gain new insight into mechanisms that whereas increased amounts of PR3 on neutrophils in active control mPR3 expression and should be explored in future ANCA disease were shown, ANCA bound to neutrophils has studies. not yet been demonstrated in vivo. The latter state of affairs To understand better this bimodal expression type of mPR3, could have several explanations, including low-affinity nature we would have liked to compare membrane expression of the of the antibodies (30), technical difficulties to detect membrane- differentiated neutrophils with the mPR3 expression of the bound ANCA, clearing of ANCA-coated polymorphonuclear stem cell donor. Unfortunately, various restrictions made such leukocytes (PMN) from circulation, or the possibility that bind- a study impossible. Therefore, we tried a different approach ϩ ing of ANCA to PMN occurs only in the microcirculation after and differentiated neutrophils from CD34 hematopoietic stem neutrophils have been primed by inflammatory cytokines (31). cells from peripheral blood of three adult donors to compare The significance of a large mPR3high neutrophil percentage their mPR3 expression pattern with isolated mature neutro- subset for clinical manifestations and outcome in WG patients phils from the same person (data not shown). Although the underscores the need to understand better the mechanisms that differentiated neutrophils expressed CD66b, no bimodal mPR3 determine the mPR3 expression pattern. Because neutrophils expression pattern that is typical for mature neutrophils iso- have a very limited life span and are hard to transfect, studies lated from blood was observed. Possibly, adult hematopoietic on mechanisms that control mPR3 expression are difficult. We stem cells that are harvested from the peripheral blood have showed previously that genetic factors have a strong influence. properties that differ from more immature cord blood stem ϩ However, we still do not know whether the information for the cells. Earlier studies showed that CD34 cells from peripheral membrane expression lies in the differentiating hematopoietic blood and bone marrow differ significantly in the expression of stem cell itself or in the body’s environment. For that reason, several cell-surface markers (34). Furthermore, hematopoietic we tested the feasibility of alternative models to study mecha- stem cells that are harvested from the peripheral blood or bone nisms of mPR3 expression, such as the use of stem cell–derived marrow have a lower cloning efficiency and proliferative ca- neutrophils. Recently, Lieber et al. (32) demonstrated the in vitro pacity, compared with progenitor cells from cord blood (35). In production of neutrophils from mouse embryonic stem cells. our experiments, we found a much lower proliferation rate for These investigators generated functional murine neutrophils. the progenitor cells that were harvested from peripheral blood. Conceivably, this technique would have utility for studies with However, we found a faster differentiation in these cells (data ϩ genetically manipulated mice. We showed that human CD34 not shown). hematopoietic stem cells could also be expanded and differen- In summary, our data underscore the clinical importance of tiated into fully functional neutrophils in vitro. The resulting the mPR3 phenotype in WG patients. Creatinine and GFR G-CSF–differentiated cells show the typical neutrophil mor- during follow-up were worse in patients with a higher percent- phology, express characteristic surface molecules, and can be age of mPR3high neutrophils. We also demonstrated that hema- activated as shown by respiratory burst activity. CD66b expres- topoietic stem cells could differentiate in vitro into mPR3high sion begins at the pro-myelocytic stage, CD11b begins at the and mPR3low neutrophil subsets, indicating that hematopoietic myelocytic stage, whereas CD35 is detectable only in bands and stem cells contain genetic information resulting in a bimodal segmented granulocytes. Thus, the expression of the surface mPR3 expression pattern of mature neutrophils. This stem cell markers assayed in this study established that our differentia- model may provide a new opportunity to study mechanisms of ϩ tion protocol resulted in approximately 90% mature (CD35 ) mPR3 expression. neutrophils (33). This two-step system follows the natural path- way of neutrophil differentiation from hematopoietic stem cells Acknowledgment and provides an interesting new opportunity to study cell This study was in part supported by InnoRegio 03 i 4509B. differentiation, signaling pathways, and biologic functions in human neutrophils. Modern techniques such as si-RNA and other methods to manipulate gene expression that are inher- References ently difficult to apply to neutrophils may now become feasible 1. van der Woude FJ, Rasmussen N, Lobatto S, Wiik A, to address other questions in this cell type. Permin H, van Es LA, van der Giessen M, van der Hem Because we are interested in mechanisms that determine GK, The TH: Autoantibodies against neutrophils and 8 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

monocytes: Tool for diagnosis and marker for disease ac- related to relapse in PR3-ANCA-associated vasculitis. JAm tivity in Wegener’s granulomatosis. Lancet 1: 425–429, 1985 Soc Nephrol 13: 2232–2238, 2002 2. Falk RJ, Jennette JC: Anti-neutrophil cytoplasmic autoan- 16. Schreiber A, Busjahn A, Luft FC, Kettritz R: Membrane tibodies with specificity for myeloperoxidase in patients expression of proteinase 3 is genetically determined. JAm with systemic vasculitis and idiopathic necrotizing and Soc Nephrol 14: 68–75, 2003 crescentic glomerulonephritis. N Engl J Med 318: 1651– 17. Witko Sarsat V, Lesavre P, Lopez S, Bessou G, Hieblot C, 1657, 1988 Prum B, Noel LH, Guillevin L, Ravaud P, Sermet Gaudelus 3. Jennette JC, Falk RJ: Small-vessel vasculitis. N Engl J Med I, Timsit J, Grunfeld JP, Halbwachs Mecarelli L: A large 337: 1512–1523, 1997 subset of neutrophils expressing membrane proteinase 3 is 4. Falk RJ, Terrell RS, Charles LA, Jennette JC: Anti-neutro- a risk factor for vasculitis and rheumatoid arthritis. JAm phil cytoplasmic autoantibodies induce neutrophils to de- Soc Nephrol 10: 1224–1233, 1999 granulate and produce oxygen radicals in vitro. Proc Natl 18. Schreiber A, Luft FC, Kettritz R: Membrane proteinase 3 Acad Sci U S A 87: 4115–4119, 1990 expression and ANCA-induced neutrophil activation. Kid- 5. Kettritz R, Jennette JC, Falk RJ: Crosslinking of ANCA- ney Int 65: 2172–2183, 2004 antigens stimulates superoxide release by human neutro- 19. Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross phils. J Am Soc Nephrol 8: 386–394, 1997 WL, Hagen EC, Hoffman GS, Hunder GG, Kallenberg 6. Savage CO, Pottinger BE, Gaskin G, Pusey CD, Pearson JD: CGM, McCluskey RT, Sinico RA, Rees AJ, van Es LA, Autoantibodies developing to myeloperoxidase and pro- Waldherr R, Wiik A: Nomenclature of systemic vasculiti- teinase 3 in systemic vasculitis stimulate neutrophil cyto- des: The proposal of an international consensus confer- toxicity toward cultured endothelial cells. Am J Pathol 141: ence. Arthritis Rheum 37: 187–192, 1994 335–342, 1992 20. Leavitt RY, Fauci AS, Bloch DA, Michel BA, Hunder GG, 7. Harper L, Ren Y, Savill J, Adu D, Savage COS: Antineu- Arend WP, Calabrese LH, Fries JF, Lie JT, Lightfoot RW Jr, trophil cytoplasmic antibodies induce reactive oxygen-de- et al.: The American College of Rheumatology 1990 criteria pendent dysregulation of primed neutrophil apoptosis and for the classification of Wegener’s granulomatosis. Arthritis clearance by macrophages. Am J Pathol 157: 211–220, 2000 Rheum 33: 1101–1107, 1990 21. Fischer M, Goldschmitt J, Peschel C, Brakenhoff JP, Kallen 8. Porges AJ, Redecha PB, Kimberly WT, Csernok E, Gross KJ, Wollmer A, Grotzinger J, Rose-John S: I. A bioactive WL, Kimberly RP: Anti-neutrophil cytoplasmic antibodies designer cytokine for human hematopoietic progenitor cell engage and activate human neutrophils via Fc gamma expansion. Nat Biotechnol 15: 142–145, 1997 RIIa. J Immunol 153: 1271–1280, 1994 22. Pick E, Mizel D: Rapid microassays for the measurement of 9. Franssen CF, Huitema MG, Kobold AC, Oost Kort WW, superoxide and hydrogen peroxide production by macro- Limburg PC, Tiebosch A, Stegeman CA, Kallenberg CG, phages in culture using an automatic enzyme immunoas- Tervaert JW: In vitro neutrophil activation by antibodies to say reader. J Immunol Methods 46: 211–226, 1981 proteinase 3 and myeloperoxidase from patients with cres- 23. Mulder AH, Heeringa P, Brouwer E, Limburg PC, Kallen- centic glomerulonephritis. J Am Soc Nephrol 10: 1506–1515, berg CG: Activation of granulocytes by anti-neutrophil 1999 cytoplasmic antibodies (ANCA): A Fc gamma RII-depen- 10. Grimminger F, Hattar K, Papavassilis C, Temmesfeld B, dent process. Clin Exp Immunol 98: 270–278, 1994 Csernok E, Gross WL, Seeger W, Sibelius U: Neutrophil 24. Ben-Smith A, Dove SK, Martin A, Wakelam MJ, Savage activation by anti-proteinase 3 antibodies in Wegener’s CO: Antineutrophil cytoplasm autoantibodies from pa- granulomatosis: Role of exogenous arachidonic acid and tients with systemic vasculitis activate neutrophils through leukotriene B4 generation. J Exp Med 184: 1567–1572, 1996 distinct signaling cascades: Comparison with conventional 11. Xiao H, Heeringa P, Hu P, Liu Z, Zhao M, Aratani Y, Fc␥ receptor ligation. Blood 98: 1448–1455, 2001 Maeda N, Falk RJ, Jennette JC: Antineutrophil cytoplasmic 25. Kettritz R, Choi M, Butt W, Rane M, Rolle S, Luft FC, Klein autoantibodies specific for myeloperoxidase cause glomer- JB: Phosphatidylinositol 3-kinase controls antineutrophil ulonephritis and vasculitis in mice. J Clin Invest 110: 955– cytoplasmic antibodies-induced respiratory burst in hu- 963, 2002 man neutrophils. J Am Soc Nephrol 13: 1740–1749, 2002 12. Goldschmeding R, van der Schoot CE, ten Bokkel Huinink 26. Pfister H, Ollert M, Froehlich LF, Quintanilla-Martinez L, D, Hack CE, van den Ende ME, Kallenberg CG, von dem Colby TV, Specks U, Jenne DE: Anti-neutrophil cytoplas- Borne AE: Wegener’s granulomatosis autoantibodies iden- mic autoantibodies (ANCA) against the murine homolog tify a novel diisopropylfluorophosphate-binding protein in of proteinase 3 (Wegener’s autoantigen) are pathogenic in the lysosomes of normal human neutrophils. J Clin Invest vivo. Blood 104: 1411–1418, 2004 84: 1577–1587, 1989 27. Kaplan-Pavlovcic S, Cerk K, Kveder R, Lindic J, Vizjak A: 13. Niles JL, McCluskey RT, Ahmad MF, Arnaout MA: Wege- Clinical prognostic factors of renal outcome in anti-neutro- ner’s granulomatosis autoantigen is a novel neutrophil phil cytoplasmic autoantibody (ANCA)-associated glo- serine proteinase. Blood 74: 1888–1893, 1989 merulonephritis in elderly patients. Nephrol Dial Transplant 14. Witko-Sarsat V, Cramer EM, Hieblot C, Guichard J, Nus- 18[Suppl 5]: v5–v7, 2003 baum P, Lopez S, Lesavre P, Halbwachs Mecarelli L: Pres- 28. Slot MC, Tervaert JW, Franssen CF, Stegeman CA: Renal ence of proteinase 3 in secretory vesicles: Evidence of a survival and prognostic factors in patients with PR3- novel, highly mobilizable intracellular pool distinct from ANCA associated vasculitis with renal involvement. Kid- azurophil granules. Blood 94: 2487–2496, 1999 ney Int 63: 670–677, 2003 15. Rarok AA, Stegeman CA, Limburg PC, Kallenberg CGM: 29. Weidner S, Geuss S, Hafezi-Rachti S, Wonka A, Rupprecht Neutrophil membrane expression of proteinase 3 (PR3) is HD: ANCA-associated vasculitis with renal involvement: J Am Soc Nephrol 16: ???–???, 2005 Membrane Proteinase 3 in Wegener’s Granulomatosis 9

An outcome analysis. Nephrol Dial Transplant 19: 1403– 33. Elghetany MT: Surface antigen changes during normal 1411, 2004 neutrophilic development: A critical review. Blood Cells 30. Abdel-Salam B, Iking-Konert C, Schneider M, Andrassy K, Mol Dis 28: 260–274, 2002 Hansch GM: Autoantibodies to neutrophil cytoplasmic an- 34. Bender JG, Unverzagt KL, Walker DE, Lee W, Van Epps tigens (ANCA) do not bind to polymorphonuclear neutro- DE, Smith DH, Stewart CC, To LB: Identification and com- phils in blood. Kidney Int 66: 1009–1017, 2004 parison of CD34-positive cells and their subpopulations 31. Heeringa P, Cohen Tervaert JW: Reply from the authors. from normal peripheral blood and bone marrow using Kidney Int 66: 2090–2091, 2004 multicolor flow cytometry. Blood 77: 2591–2596, 1991 32. Lieber JG, Webb S, Suratt BT, Young SK, Johnson GL, 35. Hao QL, Shah AJ, Thiemann FT, Smogorzewska EM, Keller GM, Worthen GS: The in vitro production and char- Crooks GM: A functional comparison of CD34 ϩ CD38- acterization of neutrophils from embryonic stem cells. cells in cord blood and bone marrow. Blood 86: 3745–3753, Blood 103: 852–859, 2004 1995