Recombinant SLPI was produced as previously described (14). Recombinant human eotaxin-1 (CCL11) was obtained from PeproTech (Rocky Hill, NJ, USA).

For SLPI staining, three monoclonal Abs were used: biotin-mouse-anti-human SLPI (clone: 31, Abcam); mouse-anti-human SLPI (clone 20409, Creative Diagnostics); mouse-anti-human SLPI (clone 20409, R&D); isotype control antibody included; biotin mouse IgG1 к (clone: MOPC-31C, BD Pharmingen, BD Biosciences); and mouse IgG1, κappa (clone B11/6) Abcam.

Human blood samples were collected from fully informed and consenting individuals (written consent was obtained from patients), and human studies were approved by the Jagiellonian University Institutional Bioethics Committee (#87/B/2014; 1072.6120.7.2018; 1072.6120.30.2020; 1072.6120.355.2020). In total, 5 EGPA patients (age 53.4 ± 16.2; M:F 3:2), 12 AD patients (age 37.8 ± 13.1; M:F 6:6), and 88 healthy individuals (age 34.9 ± 9.1; M: F, 82:6) were enrolled in the study. Patients with EGPA fulfilled the 1990 American College of Rheumatology (ACR) Classification Criteria for EGPA (15, 16). Enrollment occurred at the time of diagnosis or any time during the course of their disease (flare or remission). Disease relapse was based on physician-expert clinical judgment and defined as an increase in the activity of the disease. EGPA activity was attested by an increase in the Birmingham Vasculitis Activity Score (BVAS), requiring a dose increase, imitation, or reinstitution of glucocorticoids (GCs) and/or any immunosuppressive drug (17). All but one patient at the time of enrolment was characterized by absolute eosinophil counts (AEC) >1,500/μl. None of the patients at the time of enrollment was treated with strong immunosuppressive drugs (i.e., rituximab, cyclophosphamide, mycophenolate mofetil). The daily dose of GCs was ≤8 mg methylprednisolone. AD patients were diagnosed according to the criteria defined by Hanifin and Rajka (18). Disease severity in AD was measured according to the Severity Scoring of Atopic Dermatitis (SCORAD) and ranged from 26.5 to 83 (50.2 ± 15.2 the mean ± SD). Healthy control subjects had no clinical signs of dermatologic or allergic diseases.

Blood was collected into sodium citrate collection tubes and was subjected to isolation of granulocyte and PBMC fractions within 1 h of draw. PBMCs and granulocytes were isolated using a density gradient 1.077 g/ml centrifugation, with the aid of Pancoll (PAN Biotech). Granulocytes were recovered from the corresponding erythrocyte fraction of the Pancoll gradient as previously described (19). Eosinophils were further purified via negative magnetic sorting using the Human Eosinophil Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s recommendations. The purity of the isolated cells was examined by flow cytometry based on FSC and SSC parameters and low CD16 immunoreactivity. The eosinophil preparations were routinely ≥94% pure.

Granulocytes and PBMCs were stained with the following directly conjugated monoclonal mouse anti-human antibodies: anti-CD16 (APC-Cy7, clone: 3G8, BioLegend), anti-CD15 (BV421, clone: W63D, BioLegend), anti-CD14 (PerCP-Cy5.5, clone: MΦP9, BD Pharmingen, BD Biosciences), anti-CD3 (APC, clone: UCHT1, eBioscience), anti-CD19 (BV421, clone: HIB19, BD Horizon, BD Biosciences), anti-CD56 (BV421, clone: NCAM16.2, BD Horizon, BD Biosciences), and anti-CD45 (BV510, clone: HI30, BioLegend). For SLPI staining, granulocytes and PBMCs were first fixed with 3.7% formaldehyde and permeabilized with 0.1% saponin in PBS. Then, unless it is stated otherwise, cells were incubated with primary antibodies—monoclonal-biotin-mouse-anti-human SLPI (clone: 31) or biotin mouse IgG1 к isotype control (clone: MOPC-31C) followed by PE−streptavidin (BD Pharmingen, BD Biosciences). Where indicated, cells were incubated with monoclonal mouse-anti-human SLPI (Creative Diagnostics or R&D clone 20409), or isotype control (mouse IgG1, κappa, clone B11/6), followed by APC-conjugated polyclonal goat anti-mouse IgG1 (Abcam). Flow cytometry was performed on LSRII (BD Biosciences), and the data were analyzed using the software DIVA and FCS Express.

To assess cell viability, granulocytes from healthy donors (100 μl) were preincubated with 20 μg/ml of SLPI for 15 min, in media supplemented with 10% FCS, and then incubated for 2 h at 37°C (0.5 ml; 0.5 × 10⁶ cells/ml). Viability was determined using Annexin V FITC (BD Pharmingen) staining, followed by staining with CD16 mAbs and propidium iodide (PI, BD Pharmingen). Live cells were calculated as Annexin V^- and PI^-.

Purified granulocytes or eosinophils were seeded on Permanox plastic slides using the Nunc Lab-Tek Chamber Slide System (Thermo Fisher Scientific) or on poly-L-lysine-coated glass slides (0.8–2 × 10⁵ cells per coverslip) and incubated at 37^◦C, 5% CO₂, for 20 min in serum-free RPMI (RPMI 1640, Biowest). Cells were seeded on poly-L-lysine slides and firmly attached to the glass by a gentle centrifugal force (4.5g, 2 min) using a cytospin centrifuge (Cytospin 3, Shandon). Alternatively, immediately after isolation cells were deposited on microscope slides (18g, 10 min) using a cytospin cytocentrifuge.

Cells were then fixed with 3.7% formaldehyde, blocked overnight with 5% FBS, 1% BSA, 0.05% Tween 20, 2 mM EDTA in PBS at 4°C, treated with 0.1% saponin in PBS for 30 min at room temperature, and stained for SLPI using monoclonal-biotin-mouse-anti-human SLPI (clone: 31, Abcam) or biotin mouse IgG1 к isotype control (clone: MOPC-31C, BD Pharmingen, BD Biosciences), followed by PE−streptavidin (BD Pharmingen, BD Biosciences), and counterstained with Hoechst 33258 (Invitrogen) to visualize DNA (14).

Frozen 6–10-μm sections were prepared from skin biopsies of the AD patients, fixed in acetone, and stained with the anti-SLPI Abs or isotype control as above, and with FITC-labeled anti-CCR3 Abs (clone: 5E8, BioLegend) or FITC mouse IgG2b к isotype control (clone: MCP-11, BioLegend). The sections were counterstained with Hoechst 33258. Slides were mounted using Fluorescence Mounting Medium (Dako). Images were captured with a fully motorized fluorescence microscope (Nikon, Eclipse) and analyzed using the software NIS-Elements software (Nikon).

Granulocytes and/or sorted eosinophils (1.5 × 10⁶) were fixed overnight in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer at 4°C. Cells were then postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 1 h at room temperature and were then stained en bloc with 2% uranyl acetate aqueous solution for 1 h at room temperature. Samples were then embedded in epoxy resin (Poly/Bed 812; Polysciences, Inc., Warrington, PA, USA) after dehydration in graded ethanol series (50%–100%) and in propylene oxide. Ultrathin sections (65 nm) were cut using ultramicrotome (Leica EM UC7) and post-stained with uranyl acetate and lead citrate. The specimens were observed using a transmission electron microscope (JEOL JEM-2100) operating at an accelerating voltage of 80 kV.

Ultrathin sections of samples placed on nickel grids were incubated with 4% sodium metaperiodate for 10 min and washed in distilled water, followed by 1% aqueous periodic acid for 10 min and blocked in 1% fish skin gelatin (FSG) in PBS for 1.5 h. Sections were incubated with primary biotinylated mouse anti-human SLPI mAbs [IgG1, clone 31 (Abcam)] or biotin mouse IgG1 к isotype control (clone: MOPC-31C, BD Pharmingen, BD Biosciences), in 1% FSG/PBS overnight at 4°C, followed by a secondary antibody (monoclonal mouse anti-biotin IgG1 к, clone 3D6.6, Jackson ImmunoResearch) overnight at 4°C and a tertiary antibody (12 nm Colloidal Gold AffiniPure Donkey Anti-Mouse IgG, Jackson ImmunoResearch) for 2 h at room temperature. Sections were then fixed in 1% glutaraldehyde in PBS and examined in TEM.

Levels of SLPI were quantified in lysates of granulocytes, eosinophils, and plasma using a sandwich ELISA. 1 × 10⁶ cells were lysed in 150 μl RIPA lysis buffer (Sigma) with protease inhibitor cocktail (Roche). Lysates were centrifuged at 16,000 × g for 20 min to remove cells and cellular debris. ELISA stripes (MaxiSorp loose Nunc-Immuno Module, Thermo Scientific) were coated with mouse 5 μg/ml anti-human SLPI mAbs (clone: 20409, Creative Diagnostics) in PBS, overnight at 4°C. For cell lysates, stripes were also coated using monoclonal mouse IgG1 (ICIGG1) isotype control (B11/6, Abcam). The stripes were then washed with PBS with 0.05% Tween 20 and blocked with 3% biotin-free BSA (Roth) for 2 h at room temperature. Samples diluted in blocking buffer were added and incubated at room temperature for 2 h. Recombinant human SLPI was used as a standard. Bound SLPI was detected using biotinylated monoclonal mouse anti-human SLPI antibody (clone: 31, Abcam) followed by incubation with HRP-conjugated streptavidin (BD Pharmingen, BD Biosciences). The absorbance was measured at 450 and 630 nm (as a reference wavelength) using a Tecan Infinite M200 Plate Reader. The absorbance values for isotype controls were subtracted from the total values in each lysate sample and normalized on the basis of total protein levels determined using BCA assay (Sigma-Aldrich).

Granulocytes from healthy donors were preincubated with 20 μg/ml of SLPI for 15 min, in media supplemented with 10% FCS, followed by in vitro Transwell chemotaxis assay. 100 μl of cells (3 × 10⁵ cells/well) in RPMI with 10% FCS (Gibco) was added to the top well of 3-μm-pore transwell inserts (Costar or Thermo Fisher), and media or media plus chemoattractant 50 ng/ml CCL11 was added to the bottom well at a volume of 600 μl volume. Migration was assayed for 1.5 h at 37°C. The inserts were then removed, and cells that had migrated through the filter to the lower chamber were collected, stained with anti-CD45, anti-CD16, and anti-CD15 mAbs, and counted using flow cytometry.

Magnetically sorted eosinophils from healthy donors were seeded on microscopy plates (black uncoated μ-Plate 96-well, ibidi, Munich, Germany) and incubated for 15 min in RPMI supplemented with 10% FCS, or RPMI + 10% FCS and 20 μg/ml SLPI, followed by gentle spinning at 50g for 3 min. The cells were then imaged in Nikon Eclipse with a motorized stage, at ×10 magnification and at a rate of one frame/16 s for 30 min at room temperature. The files generated were analyzed using NIS-Elements software (Nikon). The migrating cells were recognized as those with an elongated shape that adhered to the surface and which displayed any movement in the above time frame. The cells were counted manually and calculated as a percentage of all cells within the fields of view (FOV) in a blinded fashion by two independent investigators.

Significance was evaluated using the software OriginLab (OriginLab Corporation, MA, USA), and statistical differences were determined using the ANOVA and Mann–Whitney or t-test. A p value at *p < 0.05, **p < 0.01, or ***p < 0.001 was considered statistically significant.

To determine the cellular sources of SLPI in human blood, we analyzed the SLPI levels in granulocytes and peripheral blood mononuclear cells (PBMCs) isolated from healthy donors using flow cytometry and anti-SLPI or isotype control Abs. The detection of eosinophils in the granulocyte fraction was based on high granularity revealed by the high side scatter parameter (SSC), and low/negative CD16 surface expression, whereas neutrophils were identified on the basis of lower SSC and CD16-positive staining Figure 1A ). This strategy discriminates effectively between eosinophils (96.9% of CCR3+ cells) and neutrophils (99.9% of CCR3- cells) ( Supplementary Figure 1 ). Both neutrophils and eosinophils were found to be strongly positive for SLPI ( Figures 1A, C ). Among PBMCs, we found only low-density granulocytes (LDGs) identified as CD15+, CD14- cells (20), but not monocytes (CD14+, CD3-), T cells (CD14-, CD3+), NK cells (CD14-, CD56+ CD3-, or CD56+ CD16+) or B cells (CD19+, CD3-) to be immunoreactive for SLPI ( Figures 1B, C ). Moreover, significant immunostaining for SLPI in both eosinophils and neutrophils was observed using three different anti-SLPI mAbs ( Figures 1A, C, D ).

An increased number of circulating eosinophils are associated with EGPA and AD. As anticipated, EGPA and AD patients presented with a significantly higher and a tendency to a higher percentage of eosinophils among circulating granulocytes, respectively ( Figure 2A ). These conditions were also found to manifest in significantly elevated SLPI levels in plasma ( Figure 2B ). Flow cytometry analysis showed that in common with healthy individuals, eosinophils and neutrophils from EGPA and AD patients express high levels of SLPI ( Figure 2C ). Moreover, eosinophils that infiltrate the affected skin of patients with AD were positive for SLPI ( Figure 2D ). Together, these data further point to a possible association between eosinophils and SLPI in these disorders.

Flow cytometry analysis of SLPI levels in granulocytes suggested that eosinophils contain less SLPI compared to neutrophils ( Figures 1A, C ). However, ELISA revealed that the lysates of purified eosinophils derived from healthy individuals contain different, but overall significantly higher, levels of SLPI protein compared to neutrophils ( Figure 3A ). Likewise, immuno-fluorescence microscopic examination of granulocytes from healthy donors showed overall stronger staining of eosinophils for SLPI compared to neutrophils from the same donors ( Figure 3B ). Moreover, fluorescence microscopy of the sorted eosinophils from the same donors demonstrated different levels of immunostaining for SLPI in cells that were immediately attached to microscope slides using a cytocentrifuge or were first cultured on poly-L-lysine for 20 min followed by a short cytospin or were cultured on a Permanox plastic surface in chamber slides for 20 min ( Figure 3B ). Together, these data suggested that eosinophils express more SLPI compared to neutrophils but demonstrate assay-dependent immunostaining for SLPI. The different immunoreactivity could result from condition-dependent subcellular localization and/or interaction with other molecules. Ultrastructural immunogold analysis using TEM localized SLPI mainly to the crystalline core compartment of the eosinophil granules ( Figures 4A–E, H ). Neutrophils, devoid of crystalline core (21), stored SLPI in granules, but in what appeared to be a less-compacted manner ( Figures 4F, G ). In addition, immuno-electron microscopy of eosinophils showed that in some granules SLPI is not restricted to the electron-dense crystalline core but is also detected in what appears to be an electron-lucent matrix ( Figures 4D, E, H ). We conclude that SLPI is present in partly different subcellular localizations in eosinophils and neutrophils.

High levels of SLPI in eosinophils suggest that this protein may be involved in the regulation of eosinophil functions, such as viability or migration. However, preincubation of eosinophils with 20 μg/ml of SLPI followed by 2-h incubation did not significantly affect the lifespan of eosinophils. 92.8 ± 3.6% and 91.2 ± 5.8% of eosinophils remained viable when cultured in the absence or presence of SLPI, respectively (n = 5 donors, viability at t₀ = 96.2 ± 1.9%).

Eosinophils are recruited into various peripheral locations under homeostatic, infectious, and inflammatory conditions. To determine whether SLPI regulates eosinophil migration, we incubated eosinophils from the peripheral blood of healthy donors with recombinant human SLPI in media supplemented with 10% FCS, or media with 10% FCS as a control, followed by two migration assays: time-lapse video microscopy and Transwell chemotaxis assay. Since media supplemented with 10% FCS contained approximately 4 mg/ml of the total protein by BCA assay, the addition of 20 μg/ml SLPI did not significantly change the total protein levels (protein levels were on average 4 ± 0.7 vs. 3.9 ± 0.5, mean ± SD, for samples w/o and with SLPI, respectively, n = 3). Using time-lapse video microscopy, we first analyzed the ability of SLPI to regulate the migration of purified, magnetically sorted eosinophils, by manually scoring the number of migrating cells. Video microscopy performed for 30 min demonstrated that a significantly higher number of eosinophils exhibited migratory behavior when they were pretreated with SLPI for 15 min prior to seeding in a tissue culture dish and video recording ( Figure 5A and Supplementary Video 1 ). Likewise, when the functional migratory responses of SLPI-treated or untreated granulocytes were analyzed using Transwell chemotaxis assay, we observed a similar tendency for a higher number of migrating eosinophils in the presence of SLPI at the base line ( Figure 5B ). In addition, eosinophils in granulocyte fractions incubated with SLPI were more efficient in their chemotactic response to eosinophil-specific CCL11 ( Figure 5C ). Together, these data suggest that SLPI regulates the migratory potential of eosinophils.