Ficoll-Paque (1.077 density), RPMI 1640, Hanks' balanced salt solution (HBSS), and fetal bovine serum (FBS) were purchased from WISENT Inc. (St-Bruno, QC, Canada). Terminal deoxynucleotidyl transferase was purchased from Amersham Biosciences Inc. (now part of GE Healthcare, Little Chalfont, Buckinghamshire, UK). Trizol reagents and the Superscript™ II Reverse Transcriptase (RT) kit were obtained from Invitrogen Corporation (Carlsbad, CA, USA). Oligo-dT primers and Taq DNA polymerase were purchased from PerkinElmer Life and Analytical Sciences (Woodbridge, ON, Canada). The goat polyclonal anti-human RANK-L immunoglobulin (Ig) G antibody (Ab) (sc-7627) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). The mouse monoclonal anti-human RANK Ab was obtained from Alexis Biochemicals (part of Axxora Life Sciences, Inc., San Diego, CA, USA). The rabbit polyclonal anti-human inhibitor of kappaB-alpha Ab (no. 9242) was purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). Human recombinant RANK-L was obtained from PeproTech (Rocky Hill, NJ, USA). The human RANK cDNA was a kind gift from Dr. Naoki Sakurai (Discovery Research Laboratory, Tanabe Seiyaku Co., Ltd., Yodogawa-ku, Osaka, Japan).

The institutional review board of the Université Laval (Québec, QC, Canada) approved the present study, and volunteers signed a consent form. Samples were collected in anticoagulant solution, and cells were isolated under sterile conditions. Cells were obtained from the human venous blood of healthy donors and from the SF of seven patients with RA (according to the revised criteria of the American College of Rheumatology). Characteristics of patients with RA were as follows: six women/one man, age at onset of symptoms 52.1 ± 16.2 years (mean ± standard deviation [SD]), time between onset of symptoms and the present study 4.6 ± 4.0 years (mean ± SD), clinical parameters at the present examination: erythrocyte sedimentation rate (ESR) 27.6 ± 15.6 mm (mean ± SD), and C-reactive protein 36.2 ± 31.7 g/l (mean ± SD). Four patients were positive for IgM-rheumatoid factor (RF), and three were negative for IgM-RF. Four patients had radiographic erosions, two had local osteoporosis of inflammatory joints, and one showed no radiographic symptoms. Three patients had undergone no treatment, one was taking non-steroidal anti-inflammatory drugs, one was taking 5 mg/day prednisone, and two were taking disease-modifying anti-rheumatic drugs. Due to limited quantities of SF and neutrophils from patients with RA and due to the requirement of large numbers of cells depending on the experiments performed (described below), it was not possible to systematically include the cells of the seven patients in all the experiments reported.

Blood was centrifuged (250 g, 15 minutes) and the platelet-rich plasma was removed. The peripheral blood polymorpho (neutrophils) and mononuclear leukocyte (PBML) fractions were obtained by centrifugation over Ficoll-Paque after dextran sedimentation 25. Remaining erythrocytes were eliminated by hypotonic lysis. SF neutrophils were directly obtained by centrifugation over Ficoll-Paque. After two washes, cells were counted and resuspended in culture medium. Differential cell counts of leukocytes were performed by cytofluorometry (EPICS-XL; Beckman Coulter, Fullerton, CA, USA) and Wright's and non-specific esterase stains. Neutrophil suspensions were more than 98% pure with no CD3-positive cells, and non-specific esterase-positive cells represented less than 0.2% of the cell population.

Cells were incubated in 12-well plates (2 ml/well) at 37°C and 5% CO₂ for up to 4 days. Two culture media were studied. The control medium (CM) was RPMI 1640 and 10% FBS, and the survival medium (SM) consisted of CM supplemented with 500 pM granulocyte-macrophage colony-stimulating factor (GM-CSF), 10 ng/ml interleukin (IL)-4, and 10 ng/ml TNF-α. The cytokines present in SM were chosen for their anti-apoptotic effects on neutrophils 102627. Cells and supernatants were collected from days 1 to 4. After centrifugation (5,000 g, 2 minutes), cell pellets were resuspended in 1 ml of Trizol for RNA isolation or sonicated in 0.5 ml of HBSS (no. 211–512) for enzyme immunometric assay (EIA) analysis of cell-associated materials. These samples were frozen at -20°C until assayed. When required, samples of neutrophil supernatants were concentrated by centrifugation over Amicon Ultra 10000 MW CO (Millipore Corporation, Billerica, MA, USA) at 5,000 g for 1 hour at 4°C. Normal peripheral blood neutrophils (10⁷/ml) were also incubated at 37°C and 5% CO₂ for 3 days in the presence of acellular SF from four of the seven RA patients described above and in the presence of acellular SF from two patients with OA. The incubation media consisted of 80% SF and 20% CM.

Total RNA was isolated from cells by means of the Trizol reagent, and RT reaction was performed with Superscript™ II RT according to the manufacturer's instructions. The cDNAs were amplified by polymerase chain reaction (PCR) using gene-specific primer pairs designed with Primer 3 software (Whitehead Institute for Biomedical Research, Cambridge, MA, USA) (Table 1). Each PCR was performed with one tenth of the volume of cDNA from the RT reaction, 10 μM forward and reverse primers, 200 μM dNTPs, 2.5 μl 10× PCR buffer (200 mM Tris-HCl pH 8.4, 500 mM KCl), 1 to 1.5 mM MgCl₂, 0.5 U Taq DNA polymerase, and autoclaved, distilled water to obtain a final volume of 25 μl. The number of cycles corresponding to the linear phase of amplification and the annealing temperature were optimized for each primer set (Table 1). The human β-actin transcript was used to standardize between PCRs. The PCR products were separated on a 1% agarose gel by electrophoresis in Tris acetic acid EDTA (ethylenediaminetetraacetic acid) buffer and visualized using ethidium bromide. The sequence of the amplified gene fragments was determined by direct sequencing.

The EIAs used were at two sites with horseradish peroxidase (HRP) as a tracer. Ninety-six-well plates were coated with either the human OPG/Fc Chimera (805-OS; R&D Systems, Inc., Minneapolis, MN, USA) or a monoclonal anti-human OPG Ab (MAB8051; R&D Systems, Inc.) in phosphate-buffered solution (pH 7.4). A biotinylated secondary goat anti-human RANK-L Ab (BAF626; R&D Systems, Inc.) or a compatible biotinylated secondary goat anti-human OPG Ab (BAF805; R&D Systems, Inc.) in phosphate-buffered solution (pH 7.4) containing bovine serum albumin (BSA) was used. Antigen-Ab complexes were detected by the addition of a streptavidin-HRP conjugate and tetramethylbenzidine as a substrate for HRP. Concentrations of RANK-L and OPG were obtained from a standard curve generated by known concentrations of human RANK-L and OPG. The detection limits were 15 and 7.5 pg/ml for RANK-L and OPG, respectively.

SF neutrophils (3 × 10⁶) from four patients with RA were solubilized in SDS sample buffer. The positive control was COS-7 cells transiently transfected (Fugene 6 transfection reagent; Roche Diagnostics, Indianapolis, IN, USA) with human RANK cDNA. Transfected cells were lysed in 1.5% Triton X-100 at 4°C for 5 minutes, and samples were analyzed on a 7% SDS-polyacrylamide gel. The proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore Corporation) at 4°C overnight. The membrane was blocked with 2% pig gelatin in tris-buffered saline (TBS) with 6% Tween 20 for 60 minutes, incubated with 0.2 μg/ml of a goat anti-human RANK IgG Ab (no. AF683; R&D Systems, Inc.), washed three times in TBS-Tween, and incubated with 0.04 μg/ml of a rabbit HRP-conjugated anti-goat IgG Ab (The Jackson Laboratory, Bar Harbor, ME, USA). After incubation in SF from patients with RA for 3 days (see above), healthy blood neutrophils were centrifuged at 600 g for 30 minutes on a percoll gradient to remove debris and dead cells 28, washed, resuspended in HBSS (15 × 10⁶ cells per milliliter), and stimulated at 37°C by 50 ng/ml TNF-α for 10 minutes or by 100 ng/ml RANK-L for 10 and 20 minutes. They were then transferred to 2× boiling Laemmli's sample buffer (1×: 62.5 mM Tris/HCl [pH6.8], 4% [wt/vol] SDS, 5% [vol/vol] 2-mercaptoethanol, 8.5% [vol/vol] glycerol, 2.5 mM orthovanadate, 10 mM para-nitrophenylphosphate, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 0.025% bromophenol blue). Proteins were separated on a 12% SDS-PAGE gel and transferred on a PVDF membrane. Immunoblotting was performed using 5% Blotto as a blocking agent. The primary Ab directed against I-κB-α was diluted 1:1,000 in TBS-Tween 5% BSA and incubated with the membrane for 1 hour. The goat HRP-conjugated anti-rabbit IgG Ab (The Jackson Laboratory) was diluted 1:20,000 and incubated with the membrane. The labeled Abs were detected by the ECL (enhanced chemiluminescence) detection system (GE Healthcare) and visualized on Kodak Biomax MR film (Eastman Kodak, Rochester, NY, USA).

The expression of RANK-L and RANK at the membrane was evaluated by cytofluorometry. Freshly separated healthy human blood or SF neutrophils from patients with RA and healthy neutrophils incubated in SFs were incubated with a goat anti-human RANK-L Ab (Santa Cruz Biotechnology, Inc.) followed by a fluorescein isothiocyanate (FITC)-conjugated anti-goat F(ab')₂ Ab. A normal goat IgG was used as control. To evaluate cell surface expression of RANK, healthy human blood neutrophils incubated in SFs were fixed and permeabilized using the Fixation/Permeabilization Solution kit (no. 554723) from BD Biosciences Pharmingen (San Diego, CA, USA). Briefly, non-specific staining of Fc receptors was blocked by 10% human decomplemented serum before cells were resuspended in 250 μl of Fixation/Permeabilization Solution (BD Cytofix/Cytoperm no. 554714) for 45 minutes at 4°C. Cells were then washed and permeabilized with BD Perm/Wash buffer in the presence of 10% mouse non-specific immune serum. Fixed and permeabilized neutrophils were then stained with a mouse monoclonal anti-human RANK Ab (ALX-804-212-C100) followed by a FITC-conjugated anti-mouse F(ab')₂ Ab. Corresponding controls with non-specific Abs were also performed.

Neutrophil viability was evaluated by the lactate dehydrogenase (LDH) release assay. Neutrophil suspensions after incubation were centrifuged (5,000 g, 1 minute). Supernatants and pelleted neutrophils were collected separately, and cells were lysed in 1% Triton X-100 buffer. Prior to colorimetric analysis at 340 nm, 1.25 ml of substrate (0.14 mg/ml NADH in 0.1 M sodium phosphate buffer, pH 7.35) and 50 μl of pyruvate solution were added to 50 μl of cell lysate or supernatant. Results were expressed as percentages of the ratio between the optical density values measured in supernatants and the total optical density value measured in cells plus supernatants. Viable neutrophils that did not release LDH at day 3 represented 32% and 39% in CM and SM, respectively.

Values were expressed as means ± standard error of the mean (SEM) of n experiments performed with cells from different donors. Statistical analyses were performed using GraphPad Instat 3.0 (GraphPad Software, Inc., San Diego, CA, USA). Non-parametric analysis with the Mann-Whitney test was used to compare the means of two groups. Paired groups were analyzed using the paired t test. Significance was set at a two-tailed p value of less than 0.05.

Freshly isolated neutrophils from SF of patients with RA expressed RANKL, OPG, TRAF6, and RANK as determined by semi-quantitative RT-PCR (Figure 1a). In contrast, freshly isolated peripheral blood neutrophils from healthy subjects expressed RANK-L and TRAF6, but not OPG and RANK (Figure 1a). PBMLs from healthy subjects expressed the four genes tested, and platelets expressed RANK-L, TRAF6, OPG, and not RANK (Figure 1b). The absence of OPG or RANK expression in healthy neutrophils was confirmed by qualitative PCR using an increasing number of cycles as indicated in Table 1 (data not shown). The fact that PBMLs expressed OPG and RANK mRNA enabled us to confirm that the neutrophil and platelet preparations were not contaminated by these cells.

Freshly isolated neutrophils from SF of patients with RA expressed not only the mRNA of the four genes studied (Figure 1a) but also the corresponding proteins RANK-L, OPG, and RANK (Figure 2). Cell-associated materials of SF neutrophils from patients with RA contained detectable amounts of RANK-L and OPG as measured by EIAs (Figure 2a,c). In contrast, cell-associated materials of healthy blood neutrophils contained 68 ± 13 pg/ml RANK-L and no OPG (n = 13). SF neutrophils obtained from patients with RA and incubated for up to 4 days in CM (as described in Materials and methods and Figure 2a) or in SM (data not shown) did not release RANK-L. However, freshly isolated SF neutrophils of patients with RA, as well as healthy blood neutrophils, expressed RANK-L protein on their plasma membrane, as evaluated by cytofluorometry (Figure 2b). Percentages of RANK-L-positive neutrophils freshly isolated from SF of patients with RA (n = 5) and from normal blood (n = 13) were 10.9% ± 4.3% and 2.9% ± 0.7%, respectively (p = 0.09). Moreover, the same SF neutrophils obtained from patients with RA and incubated for up to 4 days in CM showed a time-dependent release and accumulation of OPG in supernatants (Figure 2c). Finally, the expression of RANK protein was also determined by Western blot analysis in freshly isolated SF neutrophils of four patients with RA. The SF neutrophils of two of the four patients studied expressed a detectable band of an apparent molecular weight of 110 kDa. This band is specific for the polyclonal anti-human RANK Ab as shown in COS cells transiently transfected with a human RANK cDNA (Figure 2d). The amounts of OPG and RANK-L in SFs of the same patients with RA were also quantitated by EIA. SFs of patients with RA contained 6,990 ± 1,912 pg/ml OPG and 21 ± 12 pg/ml RANK-L (mean ± SEM, n = 7) with a RANK-L/OPG ratio of 0.003.

We next investigated whether in vitro conditions could mimic our in vivo observations (Figures 1 and 2). Neutrophils were incubated with cytokines that decrease neutrophil apoptosis and that are found in SFs from patients with RA 29. Healthy blood neutrophils were shown to express RANK-L mRNA under CM and SM conditions without any significant changes from day 1 to day 3 (Figure 3). The expression of OPG mRNA by neutrophils incubated in CM was not yet detectable on day 2 and appeared only after 3 days (Figure 3). The incubation of neutrophils in SM, however, strikingly upregulated the expression of this gene. The expression of OPG mRNA, which was absent at day 0 (Figure 1a), significantly increased at days 1, 2, and 3 (Figure 3). Control studies using different combinations of the cytokines were also conducted. Neutrophils incubated for 3 days in medium containing TNF-α alone expressed RANK-L but not OPG. When IL-4 was present, alone or in combination with GM-CSF or TNF-α, neutrophils expressed OPG with no change of RANK-L. GM-CSF alone had no effect on the expression of the genes tested (data not shown). Healthy blood neutrophils that do not express RANK at day 0 (Figure 1a) have the capacity to express RANK mRNA in vitro when incubated in SM (Figure 3). The expression of RANK by neutrophils was detectable from day 2 to day 3 (results observed in two donors out of nine healthy subjects studied). The expression of TRAF6 by normal blood neutrophils, on the other hand, was detected in all the conditions tested but decreased significantly at day 3 in the presence of SM (Figure 3). In contrast, PBMLs expressed OPG and RANK from day 0 (Figure 1b) to day 3 (data not shown). In these cells, a decrease in the expression of OPG and RANK was observed when incubated in CM and an increase was observed when incubated in SM. TRAF6 expression by PBMLs was similar in CM and in SM with no significant changes from days 1 to 3 (data not shown).

These findings were then confirmed at the protein level. Incubation of healthy blood neutrophils in CM or SM conditions for up to 3 days did not modify the membrane expression of RANK-L as evaluated by cytofluorometry and did not stimulate the release of detectable amounts of OPG in neutrophil supernatants as measured by EIA (data not shown). Prolongation of the incubation time up to 5 days in SM, however, led to an accumulation of OPG in the supernatants of these neutrophils (2.0 ± 0.4 pg/ml, n = 7). Moreover, the same neutrophil supernatants contained no RANK-L as evaluated by EIA (data not shown). In contrast, PBMLs cultured under similar conditions secreted RANK-L and OPG in the supernatants (data not shown), confirming that neutrophils and PBMLs expressed RANK-L and OPG differently.

Having established that inflammatory cytokines can stimulate healthy blood neutrophils to express OPG and RANK (Figure 3) and that SF neutrophils from patients with RA spontaneously expressed RANK-L, OPG, and RANK proteins (Figure 2), we next investigated the effect of SF on the expression of these genes by incubating healthy human blood neutrophils in the presence of SF from patients with RA (Figure 4). After 2 days of incubation of healthy blood neutrophils in medium containing 80% SF from patients with RA and 20% CM, membrane RANK-L significantly increased and was detected on 13.4% ± 4.7% of cells (n = 5) (versus 2.5% ± 0.8% neutrophils in CM alone, a percentage similar to that of freshly isolated neutrophils). Similar experiments with incubation medium containing SF from patients with OA revealed that 3.9% ± 0.5% of neutrophils expressed membrane RANK-L (n = 14) (Figure 4a). The difference in the expression of membrane RANK-L in healthy blood neutrophils incubated in SF from patients with RA versus patients with OA was significant (p = 0.019). Moreover, SF from patients with RA, but not patients with OA, activated healthy blood neutrophils to express OPG and RANK mRNAs as evaluated by RT-PCR (Figure 4b). Finally, SFs from patients with RA, but not from patients with OA, strongly activated healthy blood neutrophils to express RANK at the cell surface. Membrane RANK, which is not expressed by freshly isolated human blood neutrophils (data not shown), was detected on 15.3% ± 5.6% of cells after 3 days of incubation in the presence of SF of patients with RA (Figure 4c).

The release of active nuclear factor-kappa-B (NF-κB) secondary to the stimulation of RANK by RANK-L is associated with the phosphorylation of the inhibitory I-κB-α protein. Subsequently, I-κB-α decreased through its conjugation with ubiquitin and its degradation by proteasome. To determine whether RANK expressed on the surface of neutrophils was functional, healthy blood neutrophils preincubated 3 days in SF from patients with RA were stimulated by RANK-L (or TNF-α as a positive control) and total amounts of I-κB-α protein were evaluated by Western blotting. A time-dependent decrease of I-κB-α protein was demonstrated in the presence of RANK-L or TNF-α (Figure 5), confirming that the stimulation of cell surface RANK in neutrophils pretreated with SF from patients with RA was followed by intracellular signaling through, at least in part, the NF-κB pathway.