A comprehensive evaluation of murine and human ex vivo cultured alveolar macrophages

Abstract

Background: Alveolar macrophages (AMs) serve as the frontline defence in the lungs; however, comprehensive cell culture models for their study remain limited. Freshly isolated AMs are hindered by restricted quantities (mice) or challenges in accessing donors (humans). Recently, murine fetal liver-derived macrophages cultured with granulocyte-macrophages colony-stimulating factor were proposed as AMs-like macrophage (called MPI). Furthermore, recent technical progress improved the culture and expansion of primary murine AMs.

Methods: We examined three distinct in vitro models of alveolar macrophages: MPI, long-term culture-expanded AMs from mice (mAM) and cultured from human (hAM), aiming to compare and elucidate their respective advantages.

Results: We observed that: 1) isolated AMs from mice and humans can be cultured for several days (human) or months (mice) with minor loss of AM-specific surface markers expression over time; 2) MPI is a self-replicative macrophage model that is easy to culture but lacks the typical AM surface expression marker (i.e. SiglecF) and presents a constitutive pro-inflammatory phenotype; 3) responses to Toll-like receptor (TLR) agonists are consistent between MPI, mAM and hAM models but differences in magnitude should be considered; 4) phagocytic activity of MPI, mAM and hAM were similar; 5) major differences were observed between murine and human AMs regarding programmed cell death, especially for MPI; and 6) ecological and ethical consequences of these AM models are different and sometimes opposed and should be carefully assessed.

Conclusion: Our study provides a comprehensive comparison of ex vivo murine and human AM models and gives insight into the translational value of these different AM models.

FIGURE 1

Phenotype evolution of murine and human alveolar macrophage models in culture. a) Representative pictures of MPI (“Max Planck Institute cells”), cells in fresh murine bronchoalveolar lavage fluid (mBALF) from a naive mouse and murine alveolar macrophages (mAM) cultured for 4 months stained with May–Grünwald Giemsa. Scale bars: 25 µm. b) Expression of surface markers across single/live/CD45⁺ MPI, gated AMs in a fresh mBALF from a naive mouse and mAM cultured for 6 months depicted as representative pink histogram. Grey histogram represents matched fluorescence −1 negative control. c) Representative pictures of cells collected in a fresh hBALF from human 3–4 h after the flexible bronchoscopy procedure (upper picture) and cultured human AM (hAM) (Day 10, lower picture). May–Grünwald Giemsa staining. Scale bars: 25 µm. d) Representative flow cytometry dot plots showing CD206, HLA-DR, CD11b, CD169 and CD15 expression in single/live/CD45⁺ gated AMs in fresh hBALF and its isolated hAM after 10 days of culture. e) Expression of surface markers across gated AMs in a fresh hBALF and its isolated hAM after 10 days of culture depicted as a blue histogram. Grey histogram represents matched fluorescence −1 negative control. In parts a and b pictures are representative of two experiments (MPI and mAM) and three mBALF. In part c pictures represent one donor for fresh hBALF and cultured hAM, two technical replicates. In part d data are representative of two human donors for fresh hBALF and cultured hAM.

FIGURE 2

Immune responses of murine and human macrophage models to pro-inflammatory agonists. a) Fresh and cultured human alveolar macrophages (hAM) from three donors (patients #1–3) were collected by bronchoalveolar lavage (BAL), isolated and cultured ex vivo. Quantification by ELISA of tumour necrosis factor-α (TNF-α) and interleukin (IL)-6 in the supernatants of fresh hAM (freshly isolated from human bronchoalveolar lavage (hBALF), blue line) compared to cultured hAM (10 days of culture, red line) stimulated for 6 h or not with 20 µg·mL⁻¹ of Poly (I: C), 0.1 µg·mL⁻¹ of lipopolysaccharide (LPS) from E. coli O55:B5, 0.4 µg·mL⁻¹ of a recombinant flagellin, 1 µg·mL⁻¹ of Pam3CSK4 or 20 µg·mL⁻¹ curdlan. b–d) Quantification by ELISA of TNF-α, IL-6 and IL-1β in the supernatants of MPI (“Max Planck Institute cells”), murine AM (mAM) and hAM stimulated for 6 h or not with the same Toll-like receptors (TLRs) in a. e) Quantification by ELISA of IL-6 in the supernatants of MPI, mAM and hAM stimulated for 0, 1, 2, 4 and 6 h with 0.1 µg·mL⁻¹ of LPS from E. coli O55:B5. a) Data are represented for three donors (referred to as patient #), one to two technical replicates by stimulation condition. Results are expressed in fold changes compared to baseline (in log scale). b) Experiments were replicated as follows: MPI (n=3, three technical replicates), mAM (n=4, three technical replicates, except for IL-6 (n=3, three technical replicates)), hAM (n=3, one to two technical replicates). c) Experiments were replicated as follow: MPI (n=2, two technical replicates, except for at 0 and 6 h (n=3, three technical replicates)), mAM (n=2, three biological replicates, except for at 0 and 6 h (n=4, three technical replicates)), hAM (n=3, two technical replicates). Data are mean and individual values. Statistical analysis was performed using Kruskal–Wallis with Dunn's post-test. ns: non-significance. *p<0.05.

FIGURE 3

MPI (“Max Planck Institute cells”), murine alveolar macrophages (mAM) and human AM (hAM) are capable of efficiently phagocytosing bioparticles. a) Measurement of the phagocytic cells (positive for zymosan and measured using a two-colour IncuCyte SX5 cell-live analysis system of zymosan intake) over time (total duration of 2 h with assessment every 30 min). Each line represents the counts of positive cells, indicated as red cells in b. Cytochalasin D was used as a negative control and to subtract the count of background noise for each point. b) Micrographs showing cells containing red zymosan staining 2 h after exposure. Scale bars: 25 μm. c) Representative plots of positive cells for zymosan (phycoerythrin fluorescence), analysed by flow cytometry 6 h after exposure. Cytochalasin D was used as a negative control to fix the positive labeling threshold. d, e) Percentage of positive cells for zymosan (phycoerythrin fluorescence). Experiments were replicated in parts a and b as follows: MPI (n=2, duplicate), mAM (n=3, triplicate) and hAM (n=3, one to two technical replicates). Experiments were replicated in parts c and d as follows: MPI (n=4, two to three technical replicates), mAM (n=5, two to three technical replicates) and hAM (n=4, one to two technical replicates). e) Graph represents one donor for fresh hAM and cultured hAM (two technical replicates). a) Data are mean±sem. Statistical analysis was performed using a two-way RM ANOVA with Tukey's multiple comparisons test. d) Data are mean and individual values. Statistical analysis was performed using Kruskal–Wallis with Dunn's multiple comparisons post-test. ns: non-significant.

FIGURE 4

Murine alveolar macrophage (mAM) models showing differences in programmed cell deaths compared to human AM (hAM). a) Dead cell counts over time after BV6 (20 µm) or lipopolysaccharide (LPS; 100 ng·mL⁻¹)+ATP (5 mM) treatments. PI: propidium iodide. Quantification by ELISA of b) interleukin (IL)-1β and c) tumour necrosis factor-α (TNF-α) levels in the supernatants of MPI (“Max Planck Institute cells”), mAM and hAM stimulated or not for 6 h and 24 h. Experiments were replicated as follows: MPI (n=2–4, two to three technical replicates), mAM (n=3–7, two to three technical replicates), hAM (n=3, two technical replicates). a) Data are mean±sem. b and c) Data are mean and individual values. Statistical analysis was performed using Kruskal–Wallis with Dunn's post-test. ns: non-significant. *p<0.05, **p<0.005, ***p<0.0005, ****p<0.0001.