All reagents were prepared by using pure compounds in a laminar flow hood and with disposable plastic ware. HEPES ([4-(2-hydroxyethylpiperazine-1-ethanesulonic acid], sodium-heparine 170 U/mg and salts were purchased from Sigma-Aldrich, GmbH, Germany. Goat anti-human monoclonal IgE was from Caltag, USA; N-formyl-L-methionyl-L-leucyl-L-phenylalanine (fMLP), mouse monoclonal anti-human-IgE (clone GE), 4-brome-calcium ionophore A23187 and dimethylsulfoxide (DMSO) were purchased from Sigma-Aldrich GmbH, Germany. Anti-CD203c-PE (isotype IgG₁, clone 97A6) was purchased from Beckman Coulter Immunotech, USA, anti-CD69-APC (isotype IgG₁, clone FN50), anti-CD123-PECy5 (isotype IgG₁, clone 9F5), anti-CD13-APC (isotype IgG₁, clone WM15), anti-HLA-DR-PECy7 (isotype IgG2a, clone L243), anti-CD45-APCCy7 (isotype IgG₁, clone 2D1) and anti-CD63-FITC (isotype IgG₁, clone H5C6) were purchased from Becton Dickinson Pharmingen USA. Vials and test tubes were purchased from (BD Falcon, NJ, USA).

Basophils were collected by pooling the leukocyte buffy coats drawn from venous K₂-ethylen-diamino-tetra-acetic acid (EDTA) anticoagulated peripheral blood of at least four healthy, non allergic, screened subjects (blood donors) for each experiment 20. Blood samples were worked out within 2–3 hours from venous withdrawal using a differential centrifugation step procedure. Blood specimens were diluted 1:4 into an ice-kept or refrigerated (+2°C/+8°C) HEPES modified buffer ([4-(2-hydroxyethylpiperazine-1-ethanesulonic acid] 20 mmol/L; NaCl 127 mmol/L; KCl 5 mmol/L; sodium-heparin 5 UI/ml, pH 7.4) using sterile 14-ml polypropylene cap-equipped round bottom tubes and centrifuged at 700 g for 15 minutes at +4°C. Leukocyte buffy-coat layers, contaminated by erythrocytes, were individually collected, suspended in the HEPES-heparin (HBE) buffered solution and centrifuged at 400 g for 10 minutes at +4°C. Pelletted buffy-coats were washed out from surnatants, collected in a single polypropylene tube, suspended in the ice-cold HBE medium and centrifuged at 400 g for further 10 minutes. Finally pooled leukocyte buffy-coats were suspended to 1:4 v/v compared to whole blood starting volume with the refrigerated HBE buffer. In order to maintain basophils in a resting state during the whole preparation procedure, cells were treated with apyrogenic solutions and sterile disposable plastic ware and kept into ice until use in order to prevent any spontaneous activation 17. An aliquot of about 1 ml of the above HBE-suspended cell culture was transferred to a Bayer ADVIA 2120 automated hematocytometer for basophil counting and yield 28.

fMLP was dissolved in dimethylsulfoxide (DMSO) as a 2 × 10^-2 M stock solution, stored at -20°C and thawed before use. Each solution was freshly prepared by diluting fMLP in HBE supplemented with 5 mM CaCl₂ and with 2 mM MgCl₂ (HBC buffer) to 2× the final concentration, in order to make the indicated working solutions. Anti-human monoclonal IgE, purchased as mouse ascitic fluid with a concentration of 7.3 mg/ml (4.87 × 10^-5 moles/L), or as purified goat-anti human (0.5 mg/ml) in buffered saline with NaN₃, was diluted in HBC buffer two times the final molarity, in order to make the indicated working concentrations. Calcium ionophore A23187 was dissolved in DMSO at the stock concentration of 10^-3 mol/L and then dissolved in HBC buffer to 2× the final concentration, in order to make the indicated working solutions.

Ten minutes before incubation, the cell suspension was diluted 1:1 v/v with HBE and brought at 37°C; dilution was carried out to reduce homotypic aggregation following cell activation 29. Stimulation with agonists was performed at 37°C for the indicated time inside round bottom, cap-equipped polystyrene 5-ml (12 × 75 mm) plastic test tubes. 50 μl of agonist or of control HBC for resting samples were distributed in the test tubes and brought to 37°C. Then 50 μl of the cellular buffered suspension were added and incubation carried on for the indicated time. Cell culture homogeneity was maintained by gentle mixing test tubes every ten minutes. The incubation was then stopped by adding 100 μl of ice-cold HBE buffer supplemented with 2.8 mmol/L Na₃-EDTA and samples were put on ice until staining with monoclonal antibodies.

In order to choose the most suitable fluorochromes for antibodies, we followed previously reported settings 30. The protocol included tandem-dyes for cell phenotyping and small organic molecules or proteins (FITC, PE and APC) for the activation markers and used a 488 nm-633 nm two lasers equipped air flow cytometer (Becton Dickinson FACScanto). For tandem-dyes stability we followed manufacturer's instructions. Staining protocols were performed at 4°C for 20 minutes following manufacturer's instructions and according to this pattern: 10 μl/200 μl cell suspension for all the antibodies except for the following markers CD45-APCCy7 and HLA-DR-PECy7 (5 μl/200 μl cell suspension).

Soon after staining, samples underwent erythrocyte lysis according Tsang's protocol 31: lysis was performed with 3 ml of a +4°C refrigerated ammonium-chloride solution (NH₄Cl 155 mmol/L; Na₂HCO₃ 10 mmol/L, Na₃EDTA 0,10 mmol/L, pH = 7,2) for 2 minutes on ice, then cells were pelleted at 500 g for 5 minutes in a refrigerated centrifuge. Surnatants were removed and pellets gently resuspended in 0.5 ml of a BD-Isoflow phoshpate saline (PBS) balanced buffer, just ready for flow cytometry assay.

Basophils were gated as low side-scattered cells (SSC) in the CD45^dim lymphocyte area: the use of CD45 contributed to discriminate basophil area from other leukocytes better than FSC/SSC light-scattering, excluding cellular debris and to account for a further selective marker to phenotype basophilic cells 23. This region was investigated for HLADR and CD123 expression. Cellular events with a HLADR^{non expressing}/CD123^bright phenotype were then identified as basophils 2432. For each sample about 50,000 events were acquired in which, by applying this gating protocol, approximately 1500–2500 basophils in the gate were counted.

To evaluate intracellular CD63, basophils were stained as the above described methods excluding CD63-FITC. Then the cells underwent a lytic treatment, were fixed with 4% paraformaldehyde and permeabilized with 0.1% saponin and 0.09% sodium azide phosphate balanced modified buffer (BD CelLyse buffer) and finally stained with CD63-FITC as above. Samples were reconstituted in BD-Isoflow PBS balanced buffer and read.

Flow analysis was performed using a two laser BD FACScanto flow cytometer: this instrument has a 10.000 events/sec capability, six-color detection and 0,1% of sample carryover. Analyses were performed with a mean flow rate of 300–500 events/sec, setting an excess limit of 50,000 events to record in the basophil gate in order to evaluate the whole buffered suspension volume and having a proper estimation of cell recovery and reproducibility. Compensation followed cytometer manufacturer's instruction according an off-line procedure by applying automated electronics algorithms and preset templates, by using bi-parametric logarithmic dot plots, gate-specific tubes and single-tube data analysis, and optimizing FSC threshold and fluorochrome voltage as set up parameters.

To evaluate fluorochrome unspecific staining, isotype controls for anti-IgG₁ and anti-IgG_2α were introduced in the preliminary procedure to set up photomultiplier and instrument technical parameters; this control used a staining procedure carried out without introducing in the assay system the fluorochrome of interest was also performed.

Mean of fluorescence intensity (MFI) for each fluorochrome-labeled monoclonal antibody was calculated automatically with the cytometer software by averaging the total fluorescence of the marker in the basophil gate. As well percentage of activated cells was calculated by the software considering the CD63-FITC bright cells counted to the right of a threshold that was established including the main peak of fluorescence of a sample of resting cells. In order to reduce standard deviation due to brightly fluorescent cells respect to dimly ones, a logarithmic scale and the coefficient of variation to measure variability dispersion were used. When necessary, Bigos' formula to normalize brightness over background was applied 22.

The samples from a total of 82 healthy subjects (38 males, 44 females, mean age 46.34 yrs ± 5.67 SD, range 26 to 65 yrs) were used, in a total of 21 experiments. Starting from K₂-EDTA anticoagulated peripheral blood with a WBC ×10³/μl mean count of 6.41 ± 0.90 SD, corresponding to a basophil concentration mean of 37.87 ± 10.70 SD cells/μl, pooled buffy coats, having a mean count of 8.47 ± 2.09 SD WBCx10³/μl and an estimated basophil concentration of 92.25 ± 18.43 SD cells/μl, were obtained. This yield corresponds to an enrichment that is approximately 2.4 times in respect to starting whole blood.

Figure 1 shows a typical dot plot of basophil immunologic phenotyping and electronic capture. HLADR-PECy7/CD123-PECy5 plot, built using the CD45-dim cells in the lymphocyte area (see methods), allowed the definition of a well-isolated area of CD123^bright cells, not expressing HLA-DR (basophils) (Figure 1A) 32. This electronic cloud did not change its position upon activation with any agonist (here the result with anti-IgE is shown, Figure 1B, see also Table 1 [see Additional file 1]). This last point is particularly important in studies designed to investigate the dynamics of activation markers compared to a basal, resting state.

Table 1 [see Additional file 1] summarizes MFI results concerning a series of experiments performed in our laboratory. The data show the changes induced by the two agonists employed on the different markers. Gating markers (CD123 and HLADR) did not significantly change their membrane expression under activation, while CD45 was up-regulated by 2–3 times. Such an MFI increment extent was also observed with CD203c triggered by fMLP or by anti-IgE and with CD13 triggered with fMLP. CD63-FITC MFI, very low in resting cells (mean 619.76), increased by 24 times following fMLP and by 4.8 times following anti-human IgE stimulation. The mean percentage of CD63-FITC^bright cells increased from 3.12% to 33.44% following fMLP activation and from 3.12% to about 13.67% following anti-IgE stimulation. The N-aminopeptidase CD13 and the c-type lectin CD69 did not show, however, a reproducible activation pattern in our assay condition, which led us to focus on CD63 (lysosome-associated membrane protein-3 or LAMP-3) and CD203c (ecto-nucleotide-pyrophoshatase phosphodiesterase-3 or ENPP-3) markers hereafter.

Basophil response towards the different agonists was examined following changes in the mean fluorescence intensity (MFI) associated to specific membrane marker fluorochrome and by evaluating dot plots of the acquired events.

Figures 2 and 3 show a typical experiment of cell activation. Non activated basal (resting) basophils behaved as typical non expressing cells for CD63 but expressing CD203c, although at a low level (~3000 fluorescence units) (Figure 2A and 2B): fluorescence histograms (Figure 2E and 2G) exhibited a normally distributed population of both markers. Following 30 minutes of incubation with 100 nM fMLP, a significant fraction of cells (41.4%) showed a CD63 bright phenotype (Figure 2C and 2F), while almost all basophils up-regulated CD203c membrane expression (Figure 2D and 2H). The difference between CD63 and CD203c expression following cell activation was even more striking using anti-human IgE as stimulatory agent in a simultaneous assay of the same cell preparation (Figure 3). Basal (resting) level of CD63 (Figure 3A) and of CD203c (Figure 3B) and their respective fluorescence distribution (Figure 3E and 3G) were comparable to those of the resting cells previously described. Following incubation with 10 μg/ml anti-IgE, CD63 membrane expression occurred with a lesser extent than with fMLP triggering and at only 22.4% of the cells (Figure 3C and 3F), while the CD203c expression increased in all basophils and at the same extent as with fMLP triggering (Figure 3D and 3H).

These two major markers were examined in the same biparametric dot plot (Figure 4). Resting (CD63 negative) basophils showed a significant CD203c-fluorescence on their membrane (Figure 4A), as expected. Following fMLP (Figure 4B) and anti-human IgE (Figure 4C) stimulation, basophils evidenced complex activation patterns, in which a population of basophils expressing a CD63^dim/CD203c^bright phenotype is clearly evident. This might indicate: a) the presence of a CD203c^bright basophil sub-population lacking CD63 tetraspanin or b) that the CD63 dimly cells have this intracellular-associated protein 3 but are unable to up-regulate it on the membrane. To examine these issues, we investigated intracellular storage of CD63. Figure 5 shows that all basophilic cells, having a CD63^{non expressing} phenotype on their membrane (Figure 5A and 5B), have a CD63^expressing phenotype with intracellular staining (Figure 5C and 5D). This suggests that CD63 is present in all the cells but is upregulated (expressed on the cell membrane) only in a few of them.

To evaluate the responsiveness of basophilic cells in order to assess their normal function in our testing condition, dose response and time course were performed. In fMLP dose-response curves, CD203c appears the earlier marker to be activated by formylated peptides: at a dose of 3 × 10^-9 M, CD203c MFI increased from 873 to 2765 (growing up to three times) while CD63 reached the maximum of activation at 3 × 10^-8 M (changing MFI from 179 to 4020). Moreover, while CD63 up-regulation seems to go forward up to 10^-6 M fMLP, CD203c reached a maximum plateau at the dose in which CD63 began to increase (Figure 6A). On the other hand, with the agonist anti-IgE the two curves were parallel and there was no dissociation between the two markers (Figure 6B).

Time-related response patterns were investigated using an optimal stimulatory dose of formylated peptides (Figure 7). CD63 up-regulation occurred within the first minute of activation while CD203c membrane reached the plateau after about three minutes: the expression of both markers was slightly down-regulated after 60 minutes following agonist treatment. A 15 to 30 minutes time appears as optimal for the full expression of cell activation using those markers.

In order to better describe the dissociation of two markers observed with fMLP as agonist, the response of CD203c in CD63^{non expressing} basophils was evaluated (Figure 8). CD203c dose response in the basophilic cell population which did not express CD63 basophils was comparable to that evaluated by taking into account all the basophils (see also Figure 6A), indicating that sensitivity to fMLP did not depend on the differential CD63 expressing subpopulation.

To investigate a possible dissociation of two activation markers at the level of the intracellular signaling, we examined CD63 and CD203c dose response to the calcium ionophore A23187. CD203c response to A23187 was similar to the response exhibited by CD63, both markers reaching a peak close to 3 × 10^-6 M A23187 (Figure 9).