All three of the independent derivations of HoxB8-conditional progenitor cell lines analyzed were derived using the same protocol. As described previously (Cohen et al., 2021), a lentivirus-based construct with a tamoxifen-inducible Hoxb8 oncogene (Salmanidis et al., 2013) (a gift from Paul Ekert, Murdoch Children’s Research Institute) with either a puromycin or hygromycin resistance cassette was used to stably express HoxB8 in bone marrow hematopoietic stem and progenitor cells (HSPCs) isolated from C57BL/6 J mice using a kit (Stemcell Technologies). HSPCs were cultured with recombinant murine stem cell factor (SCF) and interleukin-3 (IL-3) for 2 days, exactly as described (Wang et al., 2006), and then subjected to spin-infection (2000 RPM, 60 min, 30°C) with lentivirus in a 24-well plate. Cell culture media consisted of Opti-MEM containing GlutaMax (ThermoFisher), 30 μM beta-mercaptoethanol (Sigma-Aldrich), 10% fetal bovine serum (Gemini Bioproducts), penicillin/streptomycin (Gibco), and non-essential amino acids (Gibco). To establish conditionally-immortalized progenitor lines, cells were cultured in media supplemented with 50 ng/ml SCF (BioLegend), 100 nM Z-4-hydroxytamoxifen (Tocris), and either puromycin (Sigma-Aldrich) or hygromycin (Sigma-Aldrich) for up to 2 weeks. In some cases, we used conditioned media from CHO cells that secrete recombinant murine SCF (a gift from Patrice Dubreuil, Centre de Recherche en Cancérologie de Marseille).

To generate gene knockout cell lines, we used pLentiCRISPR v2 vector, a gift from Feng Zhang (Addgene plasmid #52961). The following 20-nucleotide single-guide RNA (sgRNA) targeting sequences were inserted into pLentiCRSIPR v2: CGG​AAG​CGA​GGT​GCA​GAC​CG (Itga4), AGT​GAC​ATA​GAG​AAT​CCC​AG (Itgb1), AAT​GTC​ATC​GCG​GCG​CTC​AC (Cxcr2); Tln1 and Itgb2 targeting have been previously described (Wilson et al., 2017). Lentivirus was produced using a HEK293T Lenti-X cell line (Takara Bio). HoxB8-conditional progenitors were transduced with lentivirus in the presence of 5 μg/ml polybrene (Sigma-Aldrich) in a 24-well plate subjected to centrifugation at 800 x g for 60 min.

These studies were approved by the Lifespan Animal Welfare Committee (Approval # 5017-19, Office of Laboratory Animal Welfare Assurance #A3922-01) and were conducted in accordance with Public Health Service guidelines for animal care and use. Experiments used 8-14-week old C57BL/6 or CD45.1 (B6. SJL-PtprcaPepcb/BoyJ) mice (The Jackson Laboratory). Water and standard chow were available ad libitum. For cell transplantation, CD45.1 mice were anesthetized with 3% isoflurane and received 1 × 10⁸ (unless otherwise indicated) HoxB8-conditional progenitors via retro-orbital i. v. injection.

Unless otherwise specified, the following antibodies were from BioLegend: anti-Ly6G (1A8), anti-CD45.1 (A20), anti-CD45.2 (104), anti-cKit (ACK2), anti-CXCR4 (FAB21651, R&D Systems), anti-CD11a (M17/4), anti-CD11b (M1/70), anti-CD18 (M18/2), anti-CD29 (HMB1-1), anti-CD49d (R1-2), anti-CD49e (5H10-27), anti-β7 integrin (FIB27), anti-CD44 (IM7), anti-L-selectin (MEL-14), anti-CD16/32 (93), anti-Ly6C (HK1.4), anti-CXCR2 (SA044G4), anti-CD101 (Moushi101, Invitrogen), anti-FcγRIV (9E9), anti-CD162 2PH1, BD Biosciences).

At the indicated time after transplantation of HoxB8-conditional progenitors, blood samples were collected by saphenous venipuncture in K3-EDTA tubes (Sarstedt). Blood samples were subjected to erythrocyte lysis with ammonium chloride (BioLegend). For harvesting bone marrow, mice were euthanized and then femurs and tibias were flushed with ice cold PBS containing 5 mM EDTA and 0.5% BSA (PEB buffer) through 70-μm nylon filters. Samples in PEB buffer were labeled with antibodies as indicated, washed extensively with PEB buffer, and then analyzed by flow cytometry using a MACSQuant Analyzer 10. Data analysis was performed using FlowJo software (BD Biosciences).

On day 7 post-transplant, an initial blood sample was collected from CD45.1 recipient mice. Mice then received intraperitoneal injection of 1 ml 4% thioglycollate broth (Sigma-Aldrich). Approximately 30 min prior to the end point, a “post-thio” blood sample was collected. At 6 h after injection of thioglycollate, mice were euthanized and the peritoneal lavage was collected by flushing the peritoneal cavity with 5 ml ice cold PEB buffer. Blood and peritoneal lavage samples were labeled with antibodies against CD45.1, CD45.2, Ly6G, CD18 and/or CXCR2 as indicated, and then analyzed by flow cytometry as described above.

Bone marrow was harvested as described above and split into samples of approximately 2.5 × 10⁵ cells each in Hanks’ balanced salt solution containing Ca²⁺/Mg²⁺ (HBSS) and 10% FBS. Some samples were pretreated with 10 μg/ml cytochalasin D prior to adding 10 μg/mL S. aureus bioparticles conjugated with a pHrodo-Green fluorescent probe (ThermoFisher Scientific). After allowing phagocytosis by incubating samples at 37°C for 60 min, cells were washed twice with ice cold PEB buffer. Samples were labeled with antibodies against CD45.2 and Ly6G before analysis by flow cytometry as described above.

A GFP-expressing strain of S. aureus, USA300-sGFP (a gift from Alexander Horswill, University of Colorado Denver), was grown overnight in tryptic soy broth, then diluted 1:10 in fresh broth and grown to log phase. After extensive washing in PBS, the colony forming unit (CFU) density of USA300-sGFP was quantified by measuring its optical density at 600 nm (OD600) on a spectrophotometer. Anti-coagulated blood samples from recipient mice were subjected to lysis of erythrocytes with ammonium chloride (BioLegend) and then labeled at room temperature with antibodies against Ly6G and CD45.2. After washing, samples were resuspended in HBSS containing 10% FBS. Vehicle control or USA300-sGFP (6 × 10⁵ CFU/sample) was added to samples and they were incubated at 37°C for 15 min to allow S. aureus uptake. To remove remaining extracellular S. aureus, samples were washed in the presence of 10 μg/mL lysostaphin, pelleted, and resuspended in HBSS containing 10% FBS. A portion of the sample was transferred to ice for analysis by flow cytometry as the initial 0 min time point. The remaining sample was incubated at 37°C, then transferred to ice for analysis by flow cytometry as described above.

Anti-coagulated blood samples from recipient mice were diluted 10-fold in HBSS containing Ca²⁺/Mg²⁺. Samples were then loaded for 15 min with 2.5 μg/ml dihydrorhodamine 123 (DHR123; Cayman Chemical), a probe that fluoresces upon exposure to reactive oxygen intermediates. After DHR123 loading, diphenyleneiodonium (DPI; Sigma Aldrich) was added to the appropriate samples at a concentration of 10 μM just prior to stimulation of samples with either heat-killed S. aureus (HKSA) or 100 ng/ml phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich). Samples were stimulated for either 15 min (PMA) or 30 min (HKSA). At the end of stimulation, samples were centrifuged to pellet cells, subjected to red blood cell lysis as described above, and then stained on ice with antibodies against CD45.2, CD45.1, and Ly6G. Samples were analyzed by flow cytometry as described above.

All statistical analyses were performed using Prism 8 (GraphPad). Data throughout are presented as the group mean and standard deviation. One-way or two-way analysis of variance (ANOVA) was used to compare the differences between group means and post-hoc analysis was performed using a Bonferroni multiple comparison test. A paired or unpaired (as appropriate) Student’s t-test was used for statistical comparisons limited to two groups.

Since the approach for conditionally immortalizing murine myeloid progenitors by exogenous expression of HoxB8 was initially described (Wang et al., 2006), others have shown that such progenitor cell lines can be transplanted into irradiated recipient mice and subsequently generate mature, terminally differentiated myeloid leukocytes in the in vivo environment (Gautam et al., 2013; Redecke et al., 2013; Orosz et al., 2021). Recently, we reported that a HoxB8-conditional progenitor cell line that we derived is able to produce substantial numbers of donor graft-derived neutrophils in mice that were not subjected to any conditioning of the hematopoietic niche (Cohen et al., 2021). To understand if the engraftment of HoxB8-conditional progenitors in naïve, unconditioned mice was a unique feature of this particular progenitor cell line, we compared its engraftment to two additional and distinct lines of HoxB8-conditional myeloid progenitors using congenic CD45.1 recipient mice. Each of these three cell lines tested were independently generated using the same protocol, by lentiviral delivery of a tamoxifen-inducible Hoxb8 transgene to hematopoietic stem and progenitor cells (HSPCs) isolated from the bone marrow of C57BL/6 J mice (CD45.2⁺). To determine whether three distinct HoxB8-conditional progenitor lines would home to and take up residence in the bone marrow, we transplanted them separately in CD45.1 mice that had not received any prior conditioning regimen. For two of the independently-established progenitor lines, line 2 (P2) and line 3 (P3), we observed very few donor-derived cells present in the recipient mouse bone marrow at 4 days after transplant (Figure 1A). This was in contrast to the robust engraftment of progenitor line 1 (P1), as measured by the fraction of CD45.2⁺ cells in bone marrow (among all CD45⁺ cells) that were derived from transplanted donor progenitors (Figure 1A). These results indicate that HoxB8-conditional progenitor line 1 has a unique capacity for hematopoietic-intact engraftment.

To further characterize the engraftment of HoxB8-conditional progenitor cell line 1, we used CD45.1 recipient mice and performed a kinetic analysis of donor-derived CD45.2⁺ cells in the bone marrow. In this experiment, CD45.1 received 1 × 10⁸ HoxB8-conditional progenitors on day 0. The fraction of all bone marrow CD45⁺ cells that are derived from transplanted donor cells peaks around day 7 after transplant and are nearly undetectable by day 14 (Figure 1B), following kinetics similar to those observed by others in irradiated recipient mice (Orosz et al., 2021). The absolute number of CD45.2⁺ donor-derived cells in the bone marrow follows a similar trajectory (Figure 1B), suggesting that a relatively small fraction of transplanted HoxB8-conditional progenitors initially engraft and that it is their proliferation within the hematopoietic niche that drives the substantial donor chimerism that develops over the first week following transplant. Furthermore, we observed a dose-dependence between the number of transplanted progenitors and the fraction of neutrophils in the bone marrow and peripheral blood that were donor-derived on day 7 after transplant (Figure 1C). At longer time points of 45 days and 6 months, we did not detect any donor-derived cells in the bone marrow and there was not a detectable long-term impact on the host stem/progenitor cell composition of the bone marrow (Supplementary Figure S1). These data suggest that once exogenous tamoxifen-driven HoxB8 expression is turned off, donor progenitors enter a terminal differentiation program and their existence is then solely determined by the mature neutrophil lifespan.

The derivation of HoxB8-conditional progenitors from bone marrow HSPCs is likely to produce a heterogeneous mix of conditionally-immortalized myeloid progenitor states. To further understand whether the disparate engraftment among the three HoxB8-immortalized cell lines was due to the presence of a subpopulation with an intrinsic ability to engraft in the bone marrow niche, we generated single-cell clones of progenitor line 1 and then analyzed the engraftment of four distinct clonal progenitor lines that were transplanted into naïve CD45.1 recipient mice. We analyzed peripheral blood longitudinally to assess the peak fraction of donor-derived neutrophils in the periphery among the different clonal lines tested, as there was some variability between progenitor clones in their kinetics of engraftment and differentiation. In addition, we utilized a transplantation dose of 3 × 10⁷ HoxB8-conditional progenitors in order to better resolve potential differences between progenitor line 1 and the clonal progenitor lines. Across four different single-cell clones, we observed a range in the peak fraction of peripheral blood neutrophils that were derived from transplanted donor progenitors (Figure 1D). Clones 6 and 25 exhibited a reduced peak frequency of donor-derived neutrophils in the periphery compared to the parental line 1, whereas the donor fraction for clones 2 and 7 was comparable to that of the parental progenitor line 1 (Figure 1D). These data suggest the presence of a subset of highly-engraftable progenitors within the parental HoxB8-conditional progenitor line 1. As these analyses involved serial sampling of the peripheral blood, we accrued a substantial number of data points to be able to assess whether neutrophils generated from transplanted donor progenitors displace those generated via endogenous host granulopoiesis. There was no detectable relationship between the host-derived neutrophil count in the blood and the fraction of all neutrophils that were derived from the donor (Supplementary Figure S1).

To gain insight into potential cell surface receptors that may regulate engraftment or define a subset of highly-engrafting neutrophil progenitors, we performed flow cytometry analyses of HoxB8-conditional progenitor lines and single-cell clones to evaluate expression of cKit; C-X-C chemokine receptor 4 (CXCR4); integrin subunits CD11a, CD11b, and CD29; CD44, L-selectin, CD16/32, CD101, Ly6C, and Ly6G. Most of these receptors and markers exhibited homogeneous expression levels within each line (Supplementary Figure S2) and many had similar expression across the different progenitor cell lines analyzed (Figure 2). All progenitor lines lacked expression of Ly6G and CD101 (Supplementary Figure S2). Notable differences between progenitor lines include higher cKit expression for the poorly engrafting line 2 and disparate expression levels of several receptors/markers (CD11a, CD29) on progenitor line 3 (Figure 2). There was a prominent Ly6C^high population within line 1, line 3, and all of the clonal progenitor lines derived from progenitor line 1 (Figure 2). Compared to line 1, line 2 had a reduced fraction of Ly6C^high progenitors while line 3 exhibited a near 100% fraction of the Ly6C^high population (Figure 2). However, the frequency of the Ly6C^high subpopulation exhibited no apparent correlation with the intrinsic engraftment capacity of the respective HoxB8-conditional progenitor cell lines. Together, these data suggest that conditional HoxB8 expression in myeloid progenitors may lead to the immortalization of several different types of progenitors.

Prior to taking up residence in the hematopoietic niche, transplanted HSPCs must first traffic from the circulation to the bone marrow. Broadly, HSCs (hematopoietic stem cells) and HPCs (hematopoietic progenitor cells) transplanted into conditioned recipients undergo engraftment that depends on the integrin VLA-4 (Papayannopoulou et al., 1995; Potocnik et al., 2000; Qian et al., 2007; Rettig et al., 2012). To determine whether VLA-4 plays a role in the engraftment of HoxB8-conditional progenitors, we used CRISPR/Cas9 to generate progenitor cell lines deficient in either Itga4 or Itgb1 that encode the α4 and β1 subunits of VLA-4, respectively (Supplementary Figure S3). Congenic CD45.1 mice received a 1:1 mix of wild-type and either Itga4 ^−/− or Itgb1 ^−/− HoxB8-conditional progenitors. At 5 days post-transplant, we measured the relative fraction of wild-type and knockout CD45.2⁺ donor-derived cells in the bone marrow (Figure 3A). Interestingly, in the respective recipient mice, the abundance of Itga4 ^−/− donor cells was similar to that of wild-type donor cells (Figure 3A). In contrast, Itgb1 ^−/− HoxB8-conditional progenitors exhibited a level of engraftment that was significantly less than that of wild-type (Figure 3A). These data indicate that VLA-4 is not essential for the engraftment of HoxB8-conditional progenitors, but that another β1 integrin subtype plays a role in this process. Flow cytometry analyses of the Itga4 ^−/− progenitor cell line indicated that CD29 (β1 integrin subunit) remains expressed at the cell surface (Supplementary Figure S3), indicating that other β1 integrins are expressed. Wild-type, but not Itgb1 ^−/− , progenitors expressed CD49e and CD49f that pair with CD29 to form α5β1 and α6β1 integrins, respectively (Supplementary Figure S3). Wild-type, but not Itga4 ^−/− , progenitors also expressed α4β7 integrins (Supplementary Figure S3).

Extravasation of some leukocyte subsets from blood vessels is dependent on β2 integrins and a recent study suggests that β2 integrins contribute to transplanted HSPC homing (Mei et al., 2020). In contrast to our observations of the reduced engraftment of β1 integrin-deficient HoxB8-conditional progenitors, Itgb2 ^−/− donor progenitors (Supplementary Figure S3) were present in the bone marrow at similar frequencies as wild-type when transplanted at a 1:1 ratio (Figure 3A). Talin binds to the cytosolic tail of all integrin β chains and is required for integrin activation (Kim et al., 2011). Similar to β1 integrin-deficient HoxB8-conditional progenitors, talin knockout (Tln1 ^−/−) progenitors exhibited significantly reduced engraftment compared to wild-type control (Figure 3A).

If transplanted donor HoxB8-conditional progenitors truly engraft and take up residence in the bone marrow niche, even if that residence is transient, one would expect to observe progenitor cell proliferation and progressive differentiation towards terminal leukocyte states. We have previously shown that cells in the bone marrow derived from transplanted HoxB8-conditional progenitors that were loaded with a cell proliferation dye exhibit a greater than 20-fold dilution of the dye within 3 days, indicating in vivo proliferation of donor progenitors prior to their differentiation (Cohen et al., 2021). Analyses of all donor-derived cells in the bone marrow at days 4 and 7 after transplant indicate that their composition shifts from being split between cKit⁺Ly6G^low progenitors and cKit⁻Ly6G^high neutrophils at day 4, to predominantly cKit⁻Ly6G^high neutrophils at day 7 (Figures 3B,C). Taken together, these data suggest that HoxB8-conditional progenitors home to the bone marrow and engraft in the hematopoietic niche, where they transiently proliferate and generate large numbers of mature neutrophils.

Several studies have demonstrated that HoxB8-conditional progenitors can produce mature neutrophils following their transplantation into irradiated mice (Gautam et al., 2013; Redecke et al., 2013; Orosz et al., 2021). To further probe the dynamics of neutrophils generated in vivo from transplanted HoxB8-conditional progenitors, we first performed experiments using a model of sterile peritonitis induced by intraperitoneal injection of thiogylcollate. First, CD45.1 mice received transplantation of wild-type HoxB8-conditional progenitors and given sufficient time to produce robust chimerism in the blood. In response to peritonitis, we observed, as indicated by the similar levels of donor chimerism in the blood (before and after thioglycollate injection) and in the inflamed peritoneal cavity (Figure 4A). These data indicate that donor-derived neutrophils are recruited to a site of inflammation to an equal extent as endogenous host neutrophils.

Neutrophils mature in the bone marrow and their subsequent mobilization to the peripheral circulation is regulated by C-X-C chemokine receptor 2 (CXCR2) (Eash et al., 2010). To further understand whether neutrophils derived in vivo from transplanted progenitors are mobilized to the periphery and recruited to a site of inflammation via mechanisms that have been described for primary neutrophils, we performed several types of analyses in mice that had received a mix of wild-type and gene knockout donor progenitors. In these studies, CD45.1 mice received both wild-type and Cxcr2 ^−/− HoxB8-conditional progenitors and we restricted our analyses to CD45.2⁺Ly6G^high donor neutrophils (Supplementary Figure S4), as nearly 100% of wild-type neutrophils (both host- and donor-derived) express CXCR2 (Supplementary Figure S5). At day 8 post-transplant, donor neutrophils of both genotypes were present in the bone marrow and peripheral blood of recipient mice. However, there was a significant reduction in the relative fraction of those donor neutrophils that were Cxcr2 ^−/− in the blood compared to the bone marrow (Figure 4B). These data indicate that homeostatic egress of mature donor-derived neutrophils to the periphery is regulated by CXCR2 in a similar manner to that of primary endogenous neutrophils (Eash et al., 2010).

To evaluate the role of CXCR2 in the mobilization of donor neutrophils in response to inflammatory stimuli, we analyzed the cellular composition of peripheral blood before and after mice received intraperitoneal thioglycollate to induce a sterile peritonitis. Similar to the analyses of homeostatic neutrophil egress to the periphery, the CD45.1 mice first received a transplant of a mix of CD45.2-expressing wild-type and Cxcr2 ^−/− HoxB8-conditional progenitors. The mix of donor progenitors for transplantation was composed of approximately one-third wild-type and two-thirds Cxcr2 ^−/− progenitors, as we wanted to be able to detect changes in donor neutrophil fraction with the expectation that CXCR2 is involved both in neutrophil mobilization and recruitment, as is known for primary murine neutrophils (Cacalano et al., 1994; Eash et al., 2010). At 90 min after inducing inflammation, there was a significant decrease in the fraction of donor-derived CXCR2-deficient neutrophils in the blood (Figure 4C), indicating a CXCR2-dependent mobilization of donor neutrophils from the bone marrow. There was a further reduction in the fraction of donor-derived Cxcr2 ^−/− neutrophils relative to wild-type in the peritoneal lavage, relative to the blood (Figure 4C). Taken together, these data confirm that both the stimulated mobilization and recruitment of neutrophils generated in vivo from transplanted HoxB8-conditional progenitors occurs via CXCR2-dependent mechanisms.

Endogenous host neutrophils and those derived in vivo from transplanted donor HoxB8-conditional progenitors express similar levels of adhesion and signaling receptors involved in regulating neutrophil recruitment, including PSGL-1, L-selectin, CD11a, CD11b, CXCR2, and FcγRIV (Supplementary Figure S5). These data are consistent with a recent study, with the exception that we observed equal expression of CD11b and FcγRIV on donor and host neutrophils (Supplementary Figure S5), whereas those receptors were elevated on neutrophils when produced in irradiated recipient mice (Orosz et al., 2021). Two of the major types of β2 integrins expressed by neutrophils, LFA-1 (αLβ2) and Mac-1 (αMβ2), play key roles in neutrophil recruitment by mediating adhesion to inflamed endothelium (Ley et al., 2007). To understand whether donor-derived neutrophils are recruited to sites of inflammation via adhesion mechanisms that have been well described for host-derived primary murine neutrophils, we again used a model of thioglycollate-induced peritonitis in mice that received a 1:1 mix of wild-type and β2 integrin-deficient (Itgb2 ^−/−) HoxB8-conditional progenitors. While there was a similar proportion of wild-type and Itgb2 ^−/− donor neutrophils in the blood both before and after inducing inflammation, we observed a significant decrease in the fraction of β2 integrin-deficient donor-derived neutrophils in the peritoneal lavage (Figure 4D). These data indicate that, like primary host neutrophils, neutrophils derived in vivo from transplanted HoxB8-conditional progenitors are recruited to the inflamed peritoneum via β2 integrin-dependent mechanisms.

Previous studies have characterized the effector functions of mature neutrophils derived in vitro and in vivo from HoxB8-conditional progenitors (McDonald et al., 2011; Weiss et al., 2018; Zehrer et al., 2018; Chu et al., 2019; Saul et al., 2019; Orosz et al., 2021). Some of these studies are limited to evaluating only the function of neutrophils derived from HoxB8-conditional progenitors in vitro or lack a controlled in vivo comparison with endogenous host primary neutrophils. Since the unique line of HoxB8-conditional progenitors that we describe in this study engrafts in unconditioned mice, we were therefore able to directly assay both host and donor neutrophil functions under internally controlled conditions. Using CD45.1 mice that received CD45.2⁺ donor progenitors, we assayed neutrophil effector functions using blood samples. First, we exposed blood samples to Staphylococcus aureus bioparticles conjugated to a pHrodo-Green dye that fluoresces upon internalization into the neutrophil phagolysosome, using a low particle:cell ratio so as to better resolve potential differences between host- and donor-derived neutrophils (Supplementary Figure S6). We observed that neutrophils derived in vivo from transplanted HoxB8-conditional progenitors had a similar capacity for S. aureus bioparticle phagocytosis that was inhibited by the actin-disrupting agent cytochalasin D (Figure 5A).

To determine whether neutrophils are able to kill S. aureus after uptake, we employed an assay based upon the intracellular killing event being shown to result in the bleaching of GFP expressed by a strain of S. aureus (Schwartz et al., 2009). Donor-derived blood neutrophils were able to eradicate live S. aureus to a similar extent as host-derived neutrophils (Figure 5B). Neutrophil intracellular killing of S. aureus occurs in a manner dependent on the NADPH oxidase (NOX2). Both host- and donor-derived neutrophils underwent NOX2-mediated respiratory burst in response to heat-killed S. aureus or PMA, and sensitive to the NOX2 inhibitor DPI (Figure 5C). Together, these data indicate that neutrophils generated from HoxB8-conditional progenitors following their transplantation into naïve mice have competent host defense effector functions.