Functional characteristics of circulating granulocytes in severe congenital neutropenia caused by ELANE mutations

Peripheral blood from nine SCN1 patients, one SCN4 patient (the SCN4 patient is further described in the results section) and healthy adult controls were collected at the Children’s hospital, Chongqing Medical University, China. Informed consent was approved by the patients’ parents and controls in accordance with the Declaration of Helsinki and the study was approved by the Medical Ethics Committee of Children’s Hospital of Chongqing Medical University. Diagnosis of all patients was made by paediatricians. Two independent haematologists examined bone marrow aspirations and blood smears. Blood samples were obtained from the SCN1 patients; P01-P07 and P09 in the absence of G-CSF treatment and from P05, P06 and P08 shortly after G-CSF treatment (Filgrastim, 5–25 μg/kg/day). For each experiment, at least one healthy adult control was included and run in parallel for comparison. The experiments with purified eosinophils from healthy adult blood donors was approved by the Regional Ethical Board of Gothenburg, Sweden after informed written consent.

Genomic DNA was extracted from peripheral blood using the QIAamp DNA Mini Kit (Qiagen, China). At least 2 μg DNA was used to construct a targeted exome library (MyGenostics, China). A final library size of 350–450 bp, including adapter sequences, was selected and 243 genes associated with primary immunodeficiency diseases and other immune-related diseases were selected by a gene capture strategy, using the GenCap custom enrichment kit (MyGenostics). All mutations identified by NextSeq 500 sequencing were confirmed by Sanger sequencing.

Dextran and Ficoll-Paque were from GE-Healthcare Bio-Science (Sweden). Horseradish peroxidase (HRP) and superoxide dismutase (SOD) were from Boehringer Mannheim (Germany). Isoluminol, luminol, May-Grünwald, Giemsa, Wright’s stain and phorbol 12-myristate 13-acetate (PMA) were from Sigma (USA). The human myeloperoxidase (MPO) ELISA kit was from Immunology Consultants Laboratory Inc. (USA) and ionomycin was from Calbiochem (Germany). All flow cytometry monoclonal antibodies (mAbs) were from BioLegend (USA).

Granulocytes and peripheral mononuclear cells (PBMCs) from peripheral blood (5 mL from patients and controls, 100 mL from healthy donors for eosinophil purification) were isolated using dextran sedimentation and Ficoll-Paque centrifugation as described [16, 17]. Purified eosinophils were obtained by subjecting the granulocyte population to negative selection using magnetic beads coated with anti-CD16 mAbs (MACS, Miltenyi Biotec Inc., USA), according to manufacturer’s instructions. After isolation, cells were kept on ice in Krebs-Ringer phosphate buffer (KRG, pH 7.3; 120 mM NaCl, 5 mM KCl, 1.7 mM KH₂PO₄, 8.3 mM NaH₂PO₄ and 10 mM glucose) supplemented with Ca²⁺ (1 mM) and Mg²⁺ (1.5 mM).

Isolated cells were analysed for MPO content using a commercial ELISA kit as previously described [18]. For analysis of granulocyte composition, cells were stained with mAbs against human CD16b and CD49d, or mAbs against human CD45, CD11b, CD16, CD15, CCR3, Siglec-8, and CD14 (30 min, 4 °C) washed, and examined on a FACSCanto II (BD Biosciences). The NADPH-oxidase activity was determined using chemiluminescence (CL) [19] and measured in a six-channel Biolumat LB 9505 (Berthold Co., Germany; 1 mL system containing 1 × 10⁵ cells/mL), or a microplate reader (Synergy H1 Multi-Mode Reader, BioTek, USA; 0.2 mL system containing 3 × 10⁵ cells/mL). Intracellular CL was recorded in the presence of luminol (2 × 10^− 5 M) and SOD (50 Units/mL), and extracellular CL was recorded in the presence of isoluminol (2 × 10^− 5 M) and HRP (4 Units/mL) as described [19]. The CL values are presented as Mega counts per minute (Mcpm) for measurements performed in the Biolumat or relative light units (RLU) for measurements performed in the microplate reader.

Data analysis was performed using GraphPad Prism 7.0a (Graphpad Software, USA), except flow cytometry data which was analysed by FlowJo 10.3 (TreeStar Inc., USA). Statistical tests used are described in the figure legends for each figure and statistical significance is indicated by *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Nine SCN1 patients (seven males and two females, age range: 1 year and 4 months – 5 years and 9 months) from eight different families (P05 and P06 are twin brothers) were enrolled in the study. A total of eight different sporadic mutations in the ELANE gene were identified (Table 1), with P01 and P09 harbouring novel mutations. P01 carried a deletion of CG but an insertion of A (c.669-670CG > A), resulting in a frameshift mutation (p. C223fs), and P09 harboured a G deletion (c.593del G), resulting in a premature stop codon. The ANC for all patients was equal to or below 0.5 × 10⁹/L at a minimum of at least three separate occasions during 3 months without regular cyclic fluctuations (Table 2). Immunodeficiency was apparent for all patients, as indicated by recurrent bacterial and fungal infections (Table 3). Most patients experienced ulcers free from pus at least once, further supporting the low ANC [14]. In addition, one male patient diagnosed with SCN4, age 4 years and 1 month, with compound heterozygous mutations in the G6PC3 gene due to a stop codon in exon 2 (c.295C > T, p.Q99X) from the paternal side and a deletion in exon 6 (c.766-768del, p.256-256del) from the maternal side, was included.

Bone marrow examination performed during SCN1 diagnosis revealed limited numbers of band cells (BC) and neutrophilic segmented cells (SC; Figs. 1a-b), suggesting an early-stage maturation arrest of the neutrophil lineage. Compared to reference values, the mean percentage of neutrophils precursors, promyelocytes (PM) and myelocytes (MC) were higher than the normal range (Fig. 1b). Neutrophils and eosinophils are derived from a common myelocytic-committed progenitor, the myeloblast. The presence of eosinophils in bone marrow aspirations was evident and the percentage of all eosinophil precursors were higher than the reference values (Figs. 1a-b). The number of BC and mature segmented eosinophilic SC was 5–10 times higher than the normal range, and significantly increased as compared to the BC and SC neutrophils, strongly signifying hypereosinophilia in the bone marrow of the SCN1 patients. Of note, the bone marrow examination of the SCN4 patient did not reveal any eosinophils (Figs. 5a-b).

Peripheral blood counts were determined using Sysmex 800 or 2100, and data are presented as mean values for an average of > 30 blood samples per SCN1 patient collected in periods without G-CSF treatment (Table 2). The mean neutrophil numbers were reduced dramatically in all SCN1 patients as compared to the reference value (1.5–8.5 × 10⁹ /L) [20]. Yet, the mean numbers of neutrophils in patient samples, as measured with automated blood analysers, were above 0.5 × 10⁹/L, suggesting a mild disease, which contrasted the clinical manifestations (Table 3). For example, and in agreement with a previous report [6], all patient samples had increased levels of the acute phase C-reactive protein (CRP), displayed increased mean thrombocyte counts (higher than the normal range), and also high monocyte and eosinophil counts (above or at the upper level of the normal range), indicating frequent infections/inflammation. Of note, the mean eosinophil counts did not exceed the mean neutrophil counts for any patient, except for P08 (Table 2). Also, automated blood analysis of the SCN4 patient revealed reduced neutrophil counts (mean 0.78 × 10⁹/L) but normal eosinophil counts (mean 0.06 × 10⁹/L; data not shown).

The mean ANC (> 0.5 × 10⁹/L) and eosinophils from the automated blood analysers (Table 2) did not correlate with the number of mature neutrophils and eosinophils in the bone marrow (Figs. 1a-b) and was in disagreement with the disease severity (Table 3) [15]. Hence, we next performed blood smears to manually determine the granulocyte composition from four SCN1 patients and in parallel, these samples were again analysed by automated blood analysers. A large discrepancy was noticed regarding neutrophil and eosinophil counts obtained by these two methods. The blood analysers revealed significantly higher neutrophil numbers than eosinophils, whereas the blood smear examination demonstrated the opposite, i.e., significantly higher eosinophil numbers than neutrophils (Fig. 1c). The eosinophil counts obtained by manual counting of blood smears were also higher than the reference values for eosinophils (1–5% [21]), indicating eosinophilia in the blood of SCN1 patients. In summary, the data revealed by blood smear examination but not automated blood analysers clearly showed increased eosinophils in peripheral blood of the SCN1 patients, and agreed not only with the bone marrow examination but also with the clinical manifestations of the patients.

To further characterize the SCN1 patients’ granulocytes, we utilized the Ficoll-Paque separation method [16, 17]. Using 5 mL of blood from the SCN1 patients, Ficoll-Paque separation recovered granulocytes in numbers ranging from 0.1 × 10⁶ cells (P02) up to 3.76 × 10⁶ cells (P05 and P06 after G-CSF treatment). All experiments were performed on granulocytes from patients without G-CSF treatment if not indicated. To elucidate the presence of neutrophils and eosinophils in these granulocyte samples, we first screened for neutrophils by measuring MPO in P04, P05 and P06. MPO is a peroxidase abundantly present in neutrophil azurophil (primary) granules and monocytes, but not in eosinophils. The MPO levels in the SCN1 granulocytes were significantly decreased as compared to granulocyte samples from healthy controls (Fig. 1d), indicating that the SCN1 samples contained low amounts of neutrophils. We next tested for the presence of eosinophils in the Ficoll-Paque enriched granulocyte samples, by taking advantage of the fact that eosinophils are highly auto fluorescent compared to neutrophils when excited by the argon 488-laser and measured in the Fluorescein (FITC)-channel by flow cytometry [22, 23]. Compared with control granulocytes, SCN1 granulocytes were noticeably more auto fluorescent (Fig. 1e), which together with the reduced MPO levels strongly suggested more eosinophils than neutrophils in the SCN1 patients’ Ficoll-Paque enriched granulocytes as compared to controls. Of note, no increased auto fluorescence was seen in granulocytes isolated from the SCN4 patient (Fig. 5c).

Hereafter, the granulocytes were stained with mAbs that specifically recognize neutrophils and eosinophils. We used CD16b (expressed on neutrophils and basophils but not eosinophils), and CD49d (expressed on eosinophils but not on mature neutrophils) to classify cells isolated from two SCN1 patients. In agreement with the above results, a large proportion (~ 90%) of the SCN1 granulocytes were eosinophils (CD16b⁻CD49d⁺; Fig. 2a). In addition, isolated granulocytes from one SCN1 patients were also stained with a series of mAbs; CD45, CD11b, and CD15 that recognize both cell types, but also neutrophil-specific CD16 and eosinophil-specific Siglec-8 and CCR3. Our data show that the majority of CD45⁺CD11b⁺ granulocytes from SCN1 patients were CD15⁺CD16⁻. Further analysis of this population revealed high expression of CCR3 and Siglec-8, strongly signifying an enrichment of eosinophils with almost no neutrophils in granulocyte samples from the SCN1 patients (Fig. 2b). We also collected PBMCs and noticed that only 1% of the cells in the PBMCs were granulocytes (CD45⁺CD11b⁺CD14⁻), and that the majority of these (79%) were eosinophils (CD45⁺CD11b⁺; CD14⁻CD15⁺; CCR3⁺CD16⁻) and not neutrophils (Fig. 2c). Meanwhile, the percentage of the monocytes (CD45⁺CD11b⁺CD14⁺) were higher in the PBMCs of patients compared with those from controls (Fig. 2c), corroborating the data in Table 2. Taken together, these data support the blood smear-based cell counting (Fig. 1c) showing that eosinophils are the dominating granulocytes in the blood of SCN1 patients.

We next examined the function of the Ficoll-Paque enriched SCN1 granulocytes by measuring their capacity to generate the NADPH-oxidase derived ROS [19, 24]. Activation of granulocytes is associated with an increase in cellular consumption of molecular oxygen to generate highly reactive ROS through the NADPH-oxidase. ROS are not only strong bactericidal substances, but may also cause tissue damage and act as signaling molecules [19]. The Ficoll-Paque enriched SCN1 granulocytes released significantly higher amounts of extracellular ROS upon PMA stimulation (Fig. 3a-b), whereas the levels of PMA-induced intracellular ROS were significantly lower in SCN1 granulocytes as compared to that of the control granulocytes (Fig. 3c-d). Intracellular ROS production was further examined using ionomycin, a calcium ionophore that predominantly induces a translocation of cytosolic components of the NADPH-oxidase to the neutrophil granule membrane resulting in intracellular ROS production [25]. Our data clearly show that SCN1 granulocytes did not respond to ionomycin, unlike the control granulocytes (Fig. 3e-f). To examine if this was due to the fact that control granulocytes are dominated by neutrophils, and SCN1 granulocytes by eosinophils, we tested purified eosinophils obtained from healthy blood donors. The purified eosinophils, characterized by May-Grünwald/Giemsa staining (Fig. 4a), and auto fluorescence (Fig. 4b), also showed increased PMA-induced extracellular ROS production (Figs. 4c-d). In addition, purified control eosinophils displayed decreased PMA-induced intracellular ROS (Fig. 4e-f) and diminished ionomycin-induced intracellular ROS-production (Fig. 4g-h). That is, the ROS profile from purified control eosinophils was strikingly similar to the SCN1 granulocytes. In addition, the fact that ionomycin provoked normal intracellular ROS production in granulocytes isolated from the SCN4 patient (Fig. 5d) that primarily contained neutrophils (Fig. 5c), further supports the suggested difference regarding NADPH-oxidase activity between neutrophils and eosinophils. Also, when stimulating the SCN4 patients’ granulocytes with PMA the intracellular ROS was somewhat higher (Fig. 5e) and extracellular ROS somewhat lower (Fig. 5f) as compared to controls, i.e., in opposite to both SCN1 granulocytes (Fig. 3a-d) and purified eosinophils (Fig. 4e-h).

Taken together, these data not only confirm a dominance of eosinophils instead of neutrophils in the granulocyte population of SCN1 patients, but also show a normal NADPH-oxidase activity of SCN1 eosinophils.