SAA drives proinflammatory heterotypic macrophage differentiation in the lung via CSF-1R-dependent signaling

Reagents

Cell culture medium (RPMI 1640), FBS, trypsin EDTA, and gentamicin were purchased from Life Technologies (Carlsbad, CA, USA). Acradine Orange/Ethidium Bromide (Life Technologies), Hemacolor for microscopy (Merck Millipore, Billerica, MA, USA), TaqMan Fast Universal PCR Master Mix, TaqMan preoptimized target primers, high-capacity RNA-to-cDNA kit (Life Technologies), RNeasy kit (Qiagen, Hilden, Germany), and human recombinant SAA (rSAA; PeproTech, Rocky Hill, NJ, USA) were dissolved in sterile 0.1% BSA in dH₂O to a stock concentration of 1 mg/ml and stored at −80°C prior to use. WRW4 was sourced from Tocris Biosciences (Ellisville, MO, USA). Human GM-CSF and CSF-1 (PeproTech Rocky Hill, NJ, USA) were dissolved in 0.1% BSA in dH₂O at a stock concentration of 10 μg/ml. The Vybrant Phagocytosis Kit was purchased from Life Technologies. All antibodies were purchased from eBioscience (San Diego, CA, USA) unless otherwise stated. Anti-human antibodies; phycoerythrin (PE)-conjugated CD163 (clone eBioGH/61), and mouse IgG1_K isotype control were used to determine cell surface expression of receptors. Viability was determined by annexin V (BD Biosciences, Franklin Lakes, NJ, USA) or DAPI (Life Technologies). The PKH26PCL kit was purchased from Sigma-Aldrich (St. Louis, MO, USA). All other reagents were purchased from Sigma-Aldrich unless otherwise stated.

Differentiation of blood monocyte-derived macrophages (MDMs) and THP-1 cells

Healthy volunteers were recruited and all subjects gave written informed consent as approved by the Royal Melbourne Hospital Ethics committee. For experiments involving the comparison of healthy control subjects and patients with COPD, subject demographics are summarized in the Supplemental Table S1. Venous blood was collected in BD vacutainers containing sodium heparin (Becton Dickinson, Franklin Lakes, NJ, USA). Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll Paque Plus (GE Healthcare, Little Chalfont, UK, USA), washed, and resuspended in MAC buffer (2 mM EDTA, 0.5% BSA in PBS). Blood monocytes were purified from the PBMCs using the Monocyte Isolation kit (Miltenyi Biotec, Auburn, CA, USA) according to manufacturer's instructions.

Differentiation was promoted by culturing cells in complete medium (10% FBS in RPMI 1640 supplemented with 25 mM HEPES, 2 mM l-glutamine, 50 μM β-mercaptomethanol, 1.5 g/L sodium bicarbonate, 1 mM sodium pyruvate, and 5 μg/mL gentamicin) containing 10 ng/ml CSF-1 and 10 ng/ml GM-CSF in the absence or presence of rSAA at 0.1 μg/ml (8 nM) at 37°C in humidified atmosphere with 5% CO₂. Cells were maintained for 7 d with replacement of differentiation medium on d 3. On d 7, cells were harvested with trypsin-EDTA, and a viability cell count was performed. In a separate set of experiments, MDMs were also differentiated in the presence of the ALX/FPR2-selective antagonist WRW4. Cells were pretreated with WRW4 (5 μM) for 15 min prior to the addition of rSAA. On generation of MDMs, 3 × 10⁵ cells were retained for quantitative PCR (qPCR). In addition, differentiated MDMs were harvested and replated in 24-well format (2.5×10⁵ cells/well) and stimulated with LPS (1 μg/ml) for 2 h prior to analysis by qPCR. The human THP-1 cell line was also maintained in the same complete medium and differentiated under identical conditions in a subset of experiments.

RNA purification and qPCR

RNA was extracted using the RNeasy Plus Mini kit according to the manufacturer's instructions (Qiagen) and as described previously (21). RNA concentrations were quantified using NanoDrop 1000 (Thermo Fisher Scientific, Waltham, MA, USA). Total RNA was used as a template to generate cDNA with the High Capacity RNA-to-cDNA Kit, and qPCR was performed using predeveloped TaqMan primers and the ABI Prism 7900HT sequence detection system (Life Technologies). Threshold cycle values were normalized to 18S rRNA, and the ΔΔC_t relative expression method was used to generate fold increase above the control group as described previously (22). Each assay was performed in triplicates.

Phagocytosis and efferocytosis assays

Phagocytosis was assessed using fluorescein isothiocyanate (FITC)-labeled Escherichia coli (K12 strain) bioparticles using a modified protocol derived from the Vybrant Phagocytosis Assay Kit (Life Technologies). Briefly, 20 μl of reconstituted FITC-E. coli was incubated with 1 × 10⁵ MDMs in 100 μl complete medium for 45 min at 37°C. MDMs were retrieved by centrifugation (400 g, 5 min, 4°C), and nonphagocytized E. coli was quenched with trypan blue incubation for 1 min. Fluorescence quantitation of FITC was performed by flow cytometry using an LSR Fortessa cell analyzer (BD Biosciences), and data were collected for 5 × 10⁶ viable cell events. Macrophages were resolved from E. coli alone on the basis of forward-scatter size, and forward- and side-scatter gates were set to exclude aggregates. Propidium iodide (PI) staining was also used to exclude dead cells. Positive selection gates for phagocytosis were set as FITC fluorescence greater than that of single E. coli bioparticles alone, so that positive events represented phagocytosis of multiple E. coli (Supplemental Fig. S1).

For efferocytosis, cryopreserved blood lymphocytes isolated from the same donor used to generate MDMs were induced to undergo apoptosis using a UV transilluminator (100% UV for 15 min) followed by incubation at 37°C for 2 h. Apoptosis was confirmed using annexin V-FITC/PI staining. Apoptotic lymphocytes were labeled using the PKH26-PCL Red Fluorescent Cell Linker Kit (Sigma-Aldrich) as per the manufacturer's instructions. Next 4 × 10⁵ apoptotic lymphocytes were incubated with 1 × 10⁵ MDM for a final ratio of 4:1. Fluorescence quantitation of PKH26 was performed by flow cytometry using an LSR Fortessa cell analyzer (BD Biosciences), and data were collected for 5 × 10⁶ viable cell events. For quantitation of efferocytic macrophages, MDMs were resolved from lymphocytes alone on the basis of forward-scatter size, and PI staining was used to exclude dead cells and forward and side scatter gates were set to exclude aggregates. Positive selection gates for efferocytosis were set as PKH26 fluorescence equal to or greater than that of single lymphocytes alone so that positive events represented efferocytosis of apoptotic lymphocytes (Supplemental Fig. S1). Data were analyzed using FlowJo X software (TreeStar, Ashland, OR, USA).

In vivo SAA challenge of mice and administration of anti-CSF-1R

Specific pathogen-free (SPF) male BALB/c mice were obtained from the Animal Resource Centre (Perth, Australia) and maintained in SPF conditions at the University of Melbourne. All experimental protocols were approved and conducted in compliance with the National Health and Medical Research Council of Australia. Mice (6–8 wk old) were lightly anesthetized by inhalation of Penthrane vapor, and each treatment was administered into the lungs of mice using the intranasal method, as described previously, in a volume ≤ 50 μl (19). Recombinant SAA (2 μg) and/or 100 μg/mouse of rat anti-mouse IgG2a mAb (isotype control, in PBS) or anti-CSF-1R mAb (AFS298 in PBS) was administered intranasally, as described previously (23). Anti-CSF-1R was instilled into the lungs 2 h prior to rSAA treatment. At the specified times, brochoalveolar lavage (BAL) was performed via tracheotomy, and lung homogenates were obtained as described previously (17). To avoid nonspecific binding of Abs to FcRγ, FACS buffer containing anti-mouse CD16/32 mAb was added to all primary stains.

Surface marker analysis by flow cytometry

For CD163 surface expression on MDMs, cells were dislodged using accutase, washed twice in FACS buffer (2% FCS in PBS), and incubated with FACS buffer containing mAb or isotype control for 30 min at 4°C in the dark. After antibody incubation, cells were washed twice in FACS buffer and retrieved by centrifugation (400 g, 3 min, 4°C). Analysis was performed using the LSR Fortessa cell analyzer (BD Biosciences), and data were collected for up to 2 × 10⁵ viable cell events.

For in vivo staining of mice at the specified times, BAL was performed via tracheotomy, and lung homogenates were obtained as described previously (17). To avoid nonspecific binding of Abs to FcRγ, FACS buffer containing anti-mouse CD16/32 mAb (Mouse BD Fc Block; 2.4G2; BD Biosciences) was added to all primary stains. All antibodies were purchased from BD Biosciences unless otherwise stated. Allophycocyanin-conjugated CD11b and PE-Cy7-conjugated CD11c was purchased from eBioscience. A strict gating strategy was used to determine different macrophage cell populations as follows: single-cell gate (exclusion of aggregates; FSC-H vs. FSC-A), viability (PI exclusion), granularity/size cell gate (FSC-A vs. SSC-A), and specific surface marker gates. Macrophages and neutrophils were initially gated as single, live, large cells. Macrophages were deemed to be autofluoresent cells that were F/480⁺ and CD11c⁺. The level of CD11b expression was used to further distinguish macrophage subpopulations. Neutrophils were classified as F4/80⁻, CD11C⁻, and Ly6G^high. Cells were sorted on the Aria II Cell Sorter (BD Biosciences Australia). FlowJo software (TreeStar) was used to analyze data.

Statistical analysis

Statistical analyses were conducted using GraphPad 5.0 (GraphPad, San Diego, CA, USA), and normally distributed data were expressed as means ± sem. Significance between treatment groups was tested using an unpaired t test or 1-way or 2-way ANOVA, and values of P < 0.05 were considered statistically significant. For the phagocytosis and efferocytosis assay, a paired t test was used.

SAA promotes a mixed M1/M2 profile during monocyte to macrophage differentiation

Differentiation of blood monocytes in the presence of SAA (at 8 nM) resulted in a morphologically distinct macrophage population characterized by a heavily vacuolated cytoplasm when compared to MDMs differentiated in the absence of SAA (Fig. 1A–C). qPCR analysis revealed that the levels of IL-6 and IL-1β transcript were 6- and 4-fold higher, respectively, in the MDMs generated in the presence of SAA relative to control (Fig. 1D, E). Detailed kinetic analysis of THP-1 monocytes demonstrates a rapid induction of IL-6 and IL-1β within 6 h, peaking at 24 h and remaining elevated at d 7 (100- and 10-fold above vehicle-treated cells, respectively; Supplemental Fig. S2A, B). In addition, the M2 marker CD163 was significantly increased in MDMs by SAA (∼6-fold, Fig. 1F) with kinetic analysis revealing a gradual induction of CD163 over 6 d (Supplemental Fig. S2C). Concentration response curve analysis in THP-1 monocytes differentiated under identical conditions generated an EC₅₀ value of 8.1 nM for SAA-induced CD163 mRNA expression (Supplemental Fig. S2D). Increasing concentrations of SAA (8 and 80 nM) also significantly increased cell surface expression of CD163, as determined by flow cytometry (Supplemental Fig. S2E).

SAA also stimulated expression of IL-6 and CD163 in MDMs differentiated from COPD blood monocytes, which was a very similar response when compared to healthy controls (IL-6, 8.0±1.6 vs. 8.4±3.6, and CD163, 7.9±1.3 vs. 6.7±2.2 -fold increase, control vs. COPD-derived MDMs; Fig. 2A, B). In addition, IL-10 expression was increased by SAA in MDMs isolated from subjects with COPD (7.7±3.1-fold increase above vehicle-derived MDMs; Supplemental Fig. S2F). Since IL-10 is a key antiinflammatory gene that can promote expression of M2 markers, its association with CD163 and IL-6 was investigated. IL-10 and CD163 levels were positively associated (P<0.05, r=0.79; Fig. 2D), whereas no significant correlation was observed between IL-10 and IL-6 expression (P=0.23, r=0.54; Fig. 2C). Levels of systemic SAA in matching serum samples collected at time of monocyte isolation were not associated with the in vitro response to SAA, as assessed by directly comparing serum SAA levels with IL-6 (Fig. 2E) and CD163 (Fig. 2F) expression.

The functional characteristics of SAA-generated MDMs were assessed by measuring the phagocytic (uptake of E. coli) and efferocytic (uptake of apoptotic lymphocytes) capacity of the cells. SAA significantly increased the percentage of macrophages that phagocytosed FITC-labeled E. coli compared to vehicle control (SAA: 78.88±5.11% vs. veh: 45.12±5.68%; Fig. 3A) by an average of 2-fold (Fig. 3C). This indicates that SAA is a potent promoter of phagocytic function. When assessing efferocytosis (using matching MDMs and PKH26-labeled apoptotic lymphocytes) SAA-induced MDMs significantly increased (∼2.4-fold increase, P=0.006) their ability to engulf dying lymphocytes (SAA: 40.79±8.51% vs. veh: 18.51±4.78%; Fig. 3B, D).

Maximal activation of SAA-derived MDMs is dependent on ALX/FPR2, and stimulation with LPS demonstrates enhanced production of IL-6 and IL-1β monokines

Since SAA can potentially interact with multiple receptors, including ALX/FPR2, TLR2, and CD36, the peptide antagonist WRW4 was used to selectively target ALX/FPR2. As shown in Fig. 4A, B, SAA-induced expression of IL-6 and IL-1β was significantly reduced by 38 ± 11 and 37 ± 10%, respectively, in MDMs treated with WRW4. The effect of ALX/FPR2 antagonism on the phagocytic and efferocytic function of MDMs was also determined. In contrast to gene expression, no significant reduction was observed in phagocytosis (SAA: 83.4±2.9% vs. SAA-WRW4: 75.4±6.6%, P=0.31; Fig. 4C) or efferocytosis (SAA: 47.3±6.7% vs. SAA-WRW4: 40.3±11.3%, P=0.62; Fig. 4D). MDMs were also stimulated with LPS to further characterize their activation status. LPS markedly increased IL-6 expression by 223-fold in control MDMs, and this was further increased to 436-fold in MDMs generated with SAA, representing a 96% increase in IL-6 expression (Fig. 5A). IL-1β expression was also increased by 22-fold in control MDMs by LPS, and this was further enhanced to 31-fold in MDMs generated with SAA, representing a 41% increase in IL-1β expression (Fig. 5B).

In vivo appearance of an SAA-induced proinflammatory macrophage subpopulation that is dependent on CSF-1R signaling

Lung macrophage subpopulations were investigated in BALB/c mice in response to in vivo stimulation with SAA, which initiates neutrophilic inflammation via an ALX/FPR2 dependent mechanism (19). In control mice, the majority (>95%) of resident alveolar macrophages were CD11c^highCD11b^low (Fig. 6A), and, as previously shown were CD115⁺ and Ly6c⁻ (6). After challenge with SAA the number of resident CD11c^highCD11b^low cells did not significantly change in the lung tissue (Fig. 6C) or BAL compartment (Fig. 6D). In contrast, a CD11c^highCD11b^high population appeared following SAA challenge (Fig. 6B) and was significantly elevated in the lung tissue (Fig. 6E) and BAL (Fig. 6F) at the 24–72 h time points, disappearing by d 7 postchallenge. To investigate the consequences of chronic SAA exposure on macrophage lung subpopulations, BALB/c were challenged 1×/wk over 5 wk with either vehicle or SAA, with the final challenge given 7 d prior to analysis of the BAL compartment. A significant increase in BAL CD11c^highCD11b^high macrophages was observed in SAA-treated mice, where numbers increased 3.3-fold above vehicle-treated mice (Fig. 7A), demonstrating the persistence of the population due to chronic SAA exposure. A cell-sorting strategy was applied to isolate CD11c^highCD11b^low and CD11c^highCD11b^high macrophages from mice treated with vehicle or SAA in the chronic challenge model. On ex vivo stimulation with LPS, the resident CD11c^highCD11b^low population from vehicle- or SAA-treated lungs expressed similar levels of IL-6, IL-1β, and G-CSF (Fig. 7B–D). In contrast, CD11c^highCD11b^high macrophages isolated from SAA-challenged mice generated significantly higher levels of IL-6 (245- vs. 15-fold increase; Fig. 7B), IL-1β (22- vs. 10-fold increase; Fig. 7C) and G-CSF (27- vs. 9-fold increase; Fig. 7D) in comparison to CD11c^highCD11b^low control macrophages stimulated with LPS.

Signaling through the CSF-1R can contribute to the differentiation of myeloid progenitors into heterogeneous populations of macrophages. To determine the importance of CSF-1R receptor signaling on the appearance of CD11c^highCD11b^high cells, a neutralizing antibody to CSF-1R was administered intranasally 2 h prior to challenge with SAA. Blocking CSF-1R did not significantly alter the resident CD11c^highCD11b^low population (Fig. 8A); however, it markedly reduced the emergence of the SAA-induced CD11c^highCD11b^high population (71% reduction, SAA-treated isotype control vs. anti-CSF-1R; Fig. 8B) and BAL neutrophils (80% reduction, SAA-treated isotype control vs. anti-CSF-1R; Fig. 8C).