Neutrophil plasticity is increasingly recognised, and granulocyte subsets such as the low-density neutrophils (LDNs) reportedly participate in self-resolving inflammation and cancer progression.1, 2 Sickle cell anaemia (SCA) is an inherited haemolytic disorder associated with chronic inflammationS1 and extensive neutrophil functional heterogeneity, whereby augmented neutrophil recruitment to the vascular wall can play a triggering role in vaso-occlusive mechanisms.3-5 We aimed to determine whether normal-density neutrophil (NDN) to LDN transition is modulated in SCA, and to characterise this leucocyte subset further, with a view to determining a potential role in the disease’s pathophysiology.

Adult patients with SCA (HbSS) (Total N = 36; Table S1) participated in the study. Healthy volunteers (N = 58) were age- and gender-matched when possible. Additionally, chimeric SCA mice were generated by the transplantation of bone marrow (BM) cells isolated from Berkeley transgenic mice (that express the mutated human HbS) into sublethally irradiated 8-week-old C57BL/6 mice.5 After analysis of BM reconstitution efficiency, only chimeric SCA mice with >97% HbS were used for experiments, aged 5–6 months.5,S2 For further information about subjects and human and animal ethics, please see Data S1.

The LDN were obtained from the peripheral blood (PB) of human subjects and from murine BM exudates and PB by density gradient centrifugation, as previously described.6 Flow cytometry was used to identify human CD66b+ or murine Ly6Ghigh neutrophils in the granulocyte (NDN) and the mononuclear (LDN) density layers (10 000 cells). Percentage of LDN was calculated from the total neutrophil pool. The percentage of circulating LDN was significantly higher in the PB of patients with SCA, compared to controls (11.7% vs. 2.0%, p < 0.0001; Figure 1A), demonstrating substantial expansion of this neutrophil subset in SCA. Augmentation of the LDN pool has also been reported in other inflammatory conditions, such as in human immunodeficiency virus (HIV) and severe coronavirus disease 2019 (COVID-19) infections, cancer, and sepsis.S3−5

Circulating and BM neutrophil subsets in human and murine SCA. (A) Percentage of LDN among total circulating human neutrophils (N = 23 CON; N = 18 SCA) determined by flow cytometry. (B) TGF-β1 levels in the plasma of human subjects (N = 21 CON; N = 16 SCA) determined by high-sensitivity enzyme-linked immunosorbent assay. (C) Flow cytometry gating strategy for determining the expression of CD206 on human NDN and LDN (D) Percentage of CD206+ cells and (E) expression of CD206 in human NDN and LDN subsets (N = 6 CON; N = 8 SCA). (F) Percentage of LDN among total circulating and BM neutrophils from chimeric SCA mice (N = 7 mice per group). (G) Percentage of CD206+ cells and (H) density expression of CD206 on murine NDN and LDN (N = 7 mice per group). PB blood was collected from human subjects in EDTA for plasma separation (B) and in heparin for neutrophil isolation (A, C–E). Neutrophils were obtained through centrifugation over a Ficoll gradient (Histopaque®; 1.077 and 1.119 g/dl). The granulocytic and mononuclear fractions were separated for the identification of NDN and LDN, respectively. Neutrophil samples were purified by positive selection with anti-CD66b microbeads (Miltenyi Biotec) after red blood cell lysis (155 mM NH4Cl, 10 mM KHCO3). EDTA PB samples were obtained from mice by cardiac puncture, and BM cells were flushed from femurs and tibias and subjected to the same gradient centrifugation. For flow cytometry (A, C–H; BD FACSCalibur), human CD66b+ or murine Ly6Ghigh neutrophils were identified in 10 000 cells acquired from the granulocyte (NDN) and mononuclear (LDN) density layers, and the LDN percentage was calculated as a proportion of all neutrophils in both layers. Eosinophils were excluded from analyses based on size/granularity. Conjugated antibodies: anti-human CD66b (Perc PC-Cy5.5), anti-human CD206 (APC), anti-mouse Ly6G (APC), anti-mouse CD206 (FITC). APC, allophycocyanin; BM, bone marrow; CON, healthy controls; EDTA, ethylenediamine tetra-acetic acid; FITC, fluorescein isothiocyanate; FSC-H, forward scatter height; LDN, low-density neutrophils; MFI, mean fluorescence intensity; NDN, normal-density neutrophils; PB, peripheral blood; Perc PC-Cy5.5, peridinin chlorophyll protein-cyanine 5.5; SCA, sickle cell anaemia; SSC-H, side scatter height; TGF-β1, transforming growth factor β1. The Mann–Whitney test was used to compare non-parametric data (A, B), and Student’s t-test compared parametric data (F). One-way ANOVA followed by Bonferroni or Kruskal–Wallis followed by Dunn were used in D, E and G, H. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001

The LDN subset is thought to comprise both mature and immature neutrophils. Transitory LDN increases in the circulation have been reported in acute inflammatory conditions,2, 7 returning to normal when inflammation is resolved, suggesting that these cells are released from the BM as part of an inflammation resolution process.8 Furthermore, levels of transforming growth factor β1 (TGF-β1) cytokine, proposed to mediate HDN–LDN transition,2, 9 were elevated in the plasma of a subgroup of our SCA population (Figure 1B), consistent with previous observations.S7

Having established significantly enhanced NDN to LDN switching in SCA, we investigated some of the properties of human LDN, enriched through anti-CD66b magnetic bead selection (Miltenyi Biotec, Bergisch Gladbach), initially determining the cell surface expression of CD206 by flow cytometry. The CD206 mannose receptor is a pattern recognition receptor and a predictor of phagocytic function on anti-inflammatory M2-type macrophages.10 CD206 has recently been identified on the surface of engulfed neutrophils in the brains of mice after stroke induction, purportedly conferring a proposed N2 phenotype.11 Although not exclusive to LDN, CD206 surface expression was significantly higher on LDN than on NDN, in both healthy and SCA individuals (Figure 1C–E), and even higher on SCA LDN, compared to control LDN (Figure 1D).

In murine SCA, a condition associated with accelerated granulopoiesis,S8 the proportion of LDN in the BM neutrophil population was considerably greater than in the PB neutrophil population (50.1% vs. 8.2%, p < 0.0001, Figure 1F), supporting the hypothesis that LDN may originate from the BM. Interestingly, in chimeric SCA mice, CD206 expression was considerably upregulated on circulating LDN compared to NDN, but not on BM LDN (Figure 1G,H). To our knowledge, this is the first association of CD206 receptor expression with the LDN population; upregulation of CD206 on neutrophils appears to occur on the LDN when in the circulation, rather than in the BM compartment, further supporting a role for the circulating environment in neutrophil plasticity and indicating a resemblance of these cells to the N2 tissue-resident neutrophil phenotype.7

Previous reports have suggested that the mature subset of LDN display a higher expression of the Mac-1 integrin subunit, CD11b,7, 12 which plays a role in phagocytosis and migration.1 However, in the context of SCA, further expression and activation of Mac-1 enables neutrophils to interact with endothelial ligands and to form heterocellular aggregates, promoting vaso-occlusive processes in the microcirculation.3, 4 We found that circulating human LDN from both control and SCA individuals display a significantly higher expression density of the CD11b subunit, compared to NDN (Figure 2A). Furthermore, CD11b activation (measured using an activation epitope-specific monoclonal antibody, CRBM1/5), was considerably higher on LDN of both control and SCA populations (Figure 2B). To confirm the pro-adhesive capacity of LDN, we compared the adhesions of human neutrophil subtypes to fibronectin, an extracellular matrix ligand, using adhesion assays.S9 LDN from both control and SCA subjects were more adhesive than respective NDN (Figure 2C).

Mac-1 integrin subunit (CD11b) and adhesive properties of NDN and LDN from patients with SCA. (A) Expression of CD11b (Mac-1 integrin subunit) and (B) activated CD11b epitope on human NDN and LDN (N = 6 CON; N = 8 SCA). (C) Percentage of human neutrophils adhered to immobilised FN ligand (N = 9 CON; N = 8 SCA). Human NDN and LDN were isolated and purified as described in Figure 1. Flow cytometry (A, B) was performed according to the method specified in Figure 1C. For adhesion assays (C), cells were allowed to adhere to a FN (20 µg/ml; 30 min; 37 °C) pre-coated plate; after plate washing, the percentage of adhered cells was calculated by measuring myeloperoxidase content. Conjugated antibodies: anti-human CD66b (Perc PC-Cy5.5), anti-human CD11b (PE), anti-activated human CD11b epitope (PE). CON, healthy controls; FN, fibronectin; LDN, low-density neutrophils; MFI, mean fluorescence intensity; NDN, normal-density neutrophils; PE, phycoerythrin; Perc PC-Cy5.5, peridinin chlorophyll protein-cyanine 5.5; SCA, sickle cell anaemia. One-way ANOVA followed by Bonferroni or Kruskal–Wallis followed by Dunn. *p p p 

While the LDN subset has a proposed immunosuppressive, and therefore tumour permissive, role in cancer and metastasis,13, 14 under certain circumstances, mature LDN demonstrate increased neutrophil extracellular trap (NET) release, suggesting a potential pro-inflammatory function.15 Neutrophils display notorious functional heterogeneity and, in SCA, microbiota-driven ageing of neutrophils shifts them to a pro-inflammatory C-X-C motif chemokine receptor 4 (CXCR4)hiCD62Llo phenotype that leads to NET formation and programmes their BM clearance.4 Our findings highlight some novel features of the recently identified LDN subset in non-tumour settings. Circulating human LDN present increased adhesive properties (probably mediated by Mac-1 activation) and upregulation of CD206, which may facilitate their eventual clearance by phagocytosis.

While the LDN subset does not appear to display significantly altered properties in patients with SCA compared to control individuals, LDN transition is greatly accelerated in SCA and patients demonstrate a significantly expanded pool of LDN, whose properties may make an important contribution to disease pathophysiology. Importantly, while some features of these cells may facilitate the resolution of inflammatory processes, in the context of SCA, their increased integrin-binding adhesive activity could contribute to vaso-occlusion initiation. Thus, understanding the plasticity and defining approaches to modulate the phenotype of neutrophils may hold potential for managing the clinical manifestations of SCA. Furthermore, future studies focussing on neutrophil function in SCA should be careful to consider both the NDN and LDN cellular fractions when examining the PB of patients.

ACKNOWLEDGMENTS

The authors would like to thank Elizabeth S.R. Somessari from the Nuclear and Energy Research Institute, National Nuclear Energy Commission, IPEN-CNEN, São Paulo, Brazil, for assistance with the irradiation of the animals used in this study.

CONFLICT OF INTEREST

No relevant conflicts of interest to disclose.

AUTHOR CONTRIBUTIONS

Lidiane S. Torres and Nicola Conran designed the research study. Merav E. Shaul and Zvi G. Fridlender contributed to experimental design. Lidiane S. Torres, Lediana I. M. Teles, Irene Santos and Flávia C. Leonardo carried out the laboratory work. Paula de Melo Campos, Bruno D. Benites, Sara T. Olalla Saad, Fernando F. Costa provided clinical expertise, patient care and data interpretation. Lidiane S. Torres and Nicola Conran analysed the data and wrote the paper. All authors have reviewed and approved the manuscript.

Funding information

This study was support by funds from São Paulo Research Foundation (FAPESP), Brazil (grants # 2017/14594-0 and 2014/00984-3).