HoxB8 neutrophils replicate Fcγ receptor and integrin‐induced neutrophil signaling and functions

Abstract Neutrophils are short‐lived, terminally differentiated leukocytes that form an essential part of host immunity and play a key role in acute and chronic inflammation. The analysis of these important cells is hindered by the fact that neutrophils are not amenable to culture, transfection, or transduction. Conditionally HoxB8‐immortalized mouse hematopoietic progenitors are suitable for in vitro differentiation of a range of myeloid cells, including neutrophils. Integrins and FcγRs are cell surface receptors, the ligation of which is required for a range of neutrophil functions that are important in health and disease. We show here that HoxB8 neutrophils express major neutrophil integrins and FcγRs. They respond to FcγR and integrin stimulation in a manner that is comparable with primary neutrophils, in terms of intracellular signaling. HoxB8 neutrophils also perform a range of FcγR/integrin‐dependent neutrophil functions, including, generation of reactive oxygen species, degranulation, and chemotaxis. Our findings suggest that HoxB8 neutrophils represent a faithful experimental model system for the analysis of Fc and integrin receptor‐dependent neutrophil functions.


INTRODUCTION
Neutrophils, the most abundant circulating leukocytes in man, are an essential component of the innate immune system. 1 Terminally differentiated neutrophils are constantly renewed in the bone marrow. Neutrophils are programmed to undergo constitutive apoptosis after only a short time in the circulation, with estimated lifespans between 1 and 5 d. 6,7 Following isolation, neutrophils rapidly undergo apoptosis, with significant cell death observed after only 6-12 h in culture. Although large numbers of primary neutrophils can easily be obtained, this places a distinct limit onto in vitro experimentation. In particular, neutrophils are not amenable to genetic modification. Genetically modified mice represent the perhaps most widely used model system for the analysis of neutrophil functions. In addition to permitting the in vitro characterization of freshly isolated primary neutrophils carrying any genetic modification of interest, mice also offer the advantage of functional in vivo analyses. The generation of new mouse strains is, however, time consuming and expensive.
Conditional immortalization of mouse bone marrow or fetal liverderived hematopoietic progenitor cells with a retrovirally expressed HoxB8-ER fusion construct has been described. 8 13,14 In the context of in vitro analyses performed by a number of different laboratories, HoxB8 neutrophils efficiently phagocytosed yeast and zymosan particles, 13 released cytokines 14,15 and generated extracellular traps. 16 The functionality of HoxB8 neutrophils in integrin and/or Fc R-mediated adhesion-dependent processes, however, remains to be explored.
We show here that HoxB8 neutrophils represent a useful model for the analysis of integrin and Fc R-stimulated neutrophil functions.
Hence, HoxB8 neutrophils exhibit similar Fc R and integrin receptor expression as primary neutrophils. HoxB8 neutrophils chemotax, and their Fc R and integrin receptor stimulation induces molecular signaling events as well as production of ROS and degranulation in a manner that resembles primary neutrophils.

MATERIALS AND METHODS
All reagents were of the highest available grade and with lowest possible endotoxin level. Cell culture reagents were from (Life Technologies, Paisley, UK) cultureware from (Corning, Amsterdam, The Netherlands) and other reagents from (Sigma, Gillingham, UK) unless indicated otherwise.

Cell culture
Cells were cultured at 37 • C with 5% CO 2 in a humidified incubator.
HoxB8 conditionally immortalized progenitors were generated from bone marrows of C57Bl/6 mice essentially as described 8 ; see supplemental information for a detailed description. HoxB8 progenitors were differentiated in the presence of GM-CSF, resulting in mixed neutrophils (∼60%) and monocytes/macrophages. Where indicated, HoxB8 neutrophils were passed over a discontinuous percoll gradient (25%, 42%, and 51% 17 ) for increased purity (>80%). Prior to experiments cells were washed into assay buffer.

Isolation of primary neutrophils from bone marrow
Mouse bone marrow-derived neutrophils (BMNs) were prepared from hind legs of C57Bl/6 mice by employing a discontinuous percoll gradient as previously described. 18

Preparation of ICs and of fibrinogen-coated plastic
Insoluble ICs (HSA and rabbit polyclonal IgG to HSA) were prepared as mentioned in Chu et al. 19 Immobilized ICs (BSA and rabbit polyclonal to BSA) and fibrinogen-coated tissue culture plastic for adhesion-dependent activation were prepared as mentioned in Gambardella et al. 18

Analysis of signaling events
Phosphorylation of Akt/PKB (Thr 308) and Erk (Thr 202, Tyr 204) was analyzed by Western blotting with phosphospecific antibodies (Cell Signaling Technology, Leiden, The Netherlands) as described. 19

Analysis of cell surface receptors
Surface integrins and Fc Rs were determined by flow cytometry with specific antibodies for mouse integrins or Fc Rs (see Table 1). Data were collected on a LSR Fortessa (BD Biosciences, Oxford, UK) and analyzed using FlowJo V10 software.

ROS production assays
Total ROS production was measured in real-time by chemoluminescence on a Cytation plate reader (BioTek Instruments, Swindon, UK) in 96-well polystyrene plates (Nunc, Paisley, UK) essentially as previously described, 18

Degranulation
Degranulation of neutrophils that had been plated into appropriately coated wells or that had been stimulated by N-formyl-methionylleucyl-phenylalanine (fMLF) in the presence of cytochalasin B were assayed as in 20 by in-gel zymography of neat supernatants and immunoassay of diluted supernatants. Due to a lack of mouse lactoferrin standards, lactoferrin release was expressed in arbitrary units.

Chemotaxis
Neutrophil chemotaxis in a 0-300 nM gradient of fMLF was analyzed using a Dunn chemotaxis chamber as mentioned in Vermeren et al. 20 Cells were time-lapse imaged using an inverted (Leica, Newcastle, UK) RMDIB microscope with temperature-controlled chamber.
Images were captured every 30 s using micromanager software and an Orca camera (Hamamatsu, Welwyn Garden City, UK). Paths of individual neutrophils were tracked using the "Manual Tracking" plug-in into ImageJ, for subsequent analysis of the tracks using the "Chemotaxis" plug-in (Ibidi). Tracks from time-lapse movies generated on three separate days were combined for analysis.

RESULTS AND DISCUSSION
We aimed to identify a model system that would be amenable to the analysis of molecular events in adhesion-dependent neutrophil signal-ing, and replicate neutrophil functions that are controlled by Fc Rs and/or integrins. We initially tested DMSO-induced neutrophil-like PLB-985 cells, but had to realize that they are only a poor model for signaling events induced downstream of integrin and/or Fc R ligation (Supplemental Fig. 1). This might be because PLB-985 cells lack expression of Fc RIII/CD16 21 a finding that was confirmed by ourselves (Supplemental Fig. 2). We therefore set out to test whether HoxB8 neutrophils might represent a suitable model for neutrophil signaling and functions that depend on integrins and/or Fc Rs. We generated HoxB8 neutrophil populations from conditionally immortalized bone marrow-derived progenitor cells. HoxB8 progenitor populations readily differentiated upon withdrawal of 4OHT, with neutrophils identified by their characteristic nuclear morphology in cytocentrifuge preparations (Fig. 1A).

HoxB8 neutrophils express neutrophil integrins and Fc Rs
Mouse neutrophils constitutively express two activating Fc Rs, Fc RIII, and Fc RIV, as well as the inhibitory Fc RIIB. We compared Fc R expression by BMNs and differentiated HoxB8 neutrophils neutrophils. Both expressed all of these integrins at comparable levels (Fig. 1C).

Stimulating HoxB8 neutrophil integrins and Fc Rs induces cellular signaling
Fc Rs and integrins utilize convergent signaling pathways in neutrophils. 3,22 We were keen to identify a cellular model that replicates signaling events that occur upon integrin and/or Fc R stimulation in primary neutrophils. As a control, we analyzed fMLFstimulated intracellular signaling events indirectly by Western blotting of cell lysates with phosphospecific antibodies specific for the acti-vated forms of PKB/Akt (a read-out for phosphoinositide 3-kinase) and Erk. fMLF stimulation resulted in rapid phosphorylation of Akt/PKB and Erk in BMNs and HoxB8 neutrophils ( Fig. 2A). To analyze integrin-dependent stimulation, we next plated the cells onto the synthetic integrin ligand polyArg-Gly-Asp (pRGD), which mediates robust neutrophil stimulation even in the absence of costimulation of another receptor. 23 This triggered phosphorylation of Akt/PKB and Erk in HoxB8 neutrophils to a comparable extent as was observed with freshly prepared BMNs (Fig. 2B). We also stimulated the cells by plating them onto immobilized ICs, again obtaining comparable fold activations of Akt/PKB and Erk with HoxB8 neutrophils and with BMNs (Fig. 2C). Finally, we also stimulated the cells with insoluble ICs. This again resulted in Akt/PKB phosphorylation and Erk phosphorylation with HoxB8 cells and with freshly prepared BMNs (Fig. 2D). These results, in combination with the poorer responses obtained with PLB-985 (Supplemental Fig. 1) led us to conclude that HoxB8 neutrophils do indeed represent a suitable model system for the analysis of Fc R/integrin-induced neutrophil signaling. We next tested whether HoxB8 cells are able to perform neutrophil functions that require integrins or that are induced downstream of integrin and/or Fc R ligation.

HoxB8 neutrophils generate ROS in response to integrin and/or Fc R stimulation
HoxB8 neutrophils have already been shown to produce ROS when stimulated with PMA 12 or with serum-opsonized zymosan. 13 We confirmed that PMA stimulation of HoxB8 neutrophils resulted in significant ROS production as determined in a luminol-amplified chemoluminescence assay, which has been shown to detect mostly the production of superoxide anions. 24 Using this sensitive functional assay, we were able to show that the same number of HoxB8 cells produced the same ROS in response to PMA stimulation (Fig. 3A), regardless of whether they had or had not been subjected to neutrophil enrichment by discontinuous percoll gradient. PMA-induced ROS production by HoxB8 neutrophils was abrogated by pretreatment of HoxB8 neutrophils with the NADPH oxidase inhibitor DPI (Fig. 3B, E). We also tested ROS production in response to integrin and/or Fc R ligation, and found HoxB8 neutrophils also produce significant, NADPH oxidase-dependent ROS upon being plated onto immobilized ICs or onto the synthetic integrin ligand pRGD (Fig. 3C, D, F, G).

HoxB8 neutrophils degranulate in response to integrin and/or Fc R stimulation
Degranulation, in particular of secondary granules, is poor in neutrophil-like cells (e.g. HL60 and PLB-985). 21,25 To our knowledge, degranulation had not yet been tested with HoxB8 neutrophils.
We therefore stimulated HoxB8 neutrophils by plating them onto pRGD or immobilized ICs to test whether this caused HoxB8 neutrophils to release gelatinase from tertiary granules. As a control, we also stimulated the cells with fMLF in the presence of cytochalasin B, which drives significant degranulation to the outside of the cell. 26 This treatment triggered strong gelatinase granule release in HoxB8 neutrophils, with weaker gelatinase release from HoxB8 neutrophils observed with the more physiologic stimuli (Fig. 4A). We also analyzed degranulation from secondary granules by determining lactoferrin release. Lactoferrin release was also induced by plating of the HoxB8 neutrophils onto pRGD or immobilized ICs, or, under the control conditions, fMLF in the presence of cytochalasin B (Fig. 4B). We note, that lactoferrin, but not gelatinase release was induced more potently by immobilized ICs than by pRGD. In summary, we have

HoxB8 neutrophil chemotaxis
Reaching inflammatory sites by chemotaxis, neutrophils are among the fastest moving single cells in the body. All forms of cell migration be able to chemotax in transwells, 13 which are thought to support integrin-independent chemotaxis. 27 We performed HoxB8 neutrophil chemotaxis in Dunn chambers, where cells migrate between sheets of glass in a gradient of chemoattractant in an integrin-dependent fashion. Our experiments showed directional migration of HoxB8 neutrophils toward fMLF (Fig. 5A, B). The total accumulated and Euclidian distances travelled by HoxB8 neutrophils in these experiments were in keeping with those of BMNs in Dunn chambers analyzed under the same conditions. 18 During migration, chemotaxing HoxB8 neutrophils displayed the typical polarized morphology that is characterized by a distinct leading edge and a trailing end (Fig. 5C). This suggests that HoxB8 cells represent a suitable model system for the analysis of neutrophil chemotaxis.
In summary, we have shown here that HoxB8 neutrophils, unlike neutrophil-like differentiated PLB-985 cells, were able not only to replicate Fc R and/or integrin-dependent signaling events, but also neutrophil functions, including degranulation from secondary and tertiary granules, triggered downstream of integrin and/or Fc R ligation.
HoxB8-immortalized hematopoietic progenitors are easy to generate from mouse bone marrow stem cells; conditionally immortalized progenitors can be kept in culture for months. HoxB8 progenitors can be generated from existing mouse mutants, including those that are embryonic lethal, with reports that fetal liver and ES cells can also be used to generate the progenitors, 28  All authors approved of the final version of the paper.