Macrophages redirect phagocytosis by non-professional phagocytes and influence inflammation

Abstract

Professional phagocytes (such as macrophages) and non-professional phagocytes (such as epithelial cells) clear billions of apoptotic cells and particles on a daily basis. Although professional and non-professional macrophages reside in proximity in most tissues, whether they communicate with each other during cell clearance, and how this might affect inflammation, is not known. Here we show that macrophages, through the release of a soluble growth factor and microvesicles, alter the type of particles engulfed by non-professional phagocytes and influence their inflammatory response. During phagocytosis of apoptotic cells or in response to inflammation-associated cytokines, macrophages released insulin-like growth factor 1 (IGF-1). The binding of IGF-1 to its receptor on non-professional phagocytes redirected their phagocytosis, such that uptake of larger apoptotic cells was reduced whereas engulfment of microvesicles was increased. IGF-1 did not alter engulfment by macrophages. Macrophages also released microvesicles, whose uptake by epithelial cells was enhanced by IGF-1 and led to decreased inflammatory responses by epithelial cells. Consistent with these observations, deletion of IGF-1 receptor in airway epithelial cells led to exacerbated lung inflammation after allergen exposure. These genetic and functional studies reveal that IGF-1- and microvesicle-dependent communication between macrophages and epithelial cells can critically influence the magnitude of tissue inflammation in vivo.

Extended Data Figure 1. Erk1/2 is phosphorylated in LR73 cells stimulated with EGF, VEGF, PDGF AA/BB

(A) Representative engulfment assay in which LR73 cells were treated with indicated growth factors at increasing concentrations and assessed for engulfment of apoptotic thymocytes (n=3). (B–D) Serum-starved LR73 cells were stimulated with 100ng/mL of indicated growth factors for 10 minutes and the phosphorylation of Erk1/2 was determined by immunoblotting (n=2).

Extended Data Figure 2. 16HBE14o- cells and SVEC-40 cells engulf less apoptotic cells when exposed to IGF-1

(A) Representative engulfment assay in which the uptake of apoptotic thymocytes by 16HBE14o- human airway epithelial cells is dampened by IGF-1 treatment (n=3). (B) Engulfment of SVEC-40 endothelial cells treated with IGF-1 with apoptotic thymocytes as targets (n=3). Data represented as mean ± s.d. A p-value of <0.05 (indicated by one asterisk), <0.01 (indicated by two asterisks), or <0.001 (indicated by three asterisks).

Extended Data Figure 3. A second IGF-1r inhibitor, NVP-AEW541, also reverses inhibition of apoptotic cell engulfment due to IGF-1

(A) Engulfment of apoptotic thymocytes by LR73 cells treated with various doses of NVP-AEW541, a small molecule inhibitor of IGF-1r (n=3). (B) Representative Western blot of LR73 cells stimulated with IGF-1 and treated with increasing doses of NVP-AEW541 (n=2). Data represented as mean ± s.d.

Extended Data Figure 4. LR73 cells treated with recombinant insulin or IGF-II engulf less apoptotic cells

Engulfment of apoptotic thymocytes by LR73 cells that were treated with the indicated concentrations of human insulin (A) and human IGF-II (B) (n=2). Data represented as mean ± s.d.

Extended Data Figure 5. Blocking canonical signaling intermediates downstream of IGF-1 receptor signaling does not reverse the IGF-1 mediated engulfment suppression

(A–D) Engulfment of apoptotic thymocytes by LR73 cells treated with U2016 (Erk1/2 inhibitor) (A), MK-2206 (Akt1/2/3 inhibitor) (B), Rapamycin (mTOR inhibitor) (C), or Wortmannin (PI 3-Kinase inhibitor) (D) in the presence or absence of IGF-1 (n=2–3). Wortmannin has been previously demonstrated to inhibit phagocytosis of apoptotic cells and adding IGF-1 does not alter this inhibition. Data represented as mean ± s.d.

Extended Data Figure 6. Inhibition of Rho-kinase (ROCK) does not appear to rescue IGF-1 induced engulfment suppression in LR73 cells

(A–B) (Left panel) Engulfment assay of LR73 cells treated with various doses of either Rho kinase inhibitors Y27632 or GSK269962 in the presence or absence of IGF-1 (n=2–3). Initially, it appeared that inhibition of ROCK was able to partially rescue IGF-1 induced engulfment suppression. However, as ROCK inhibition basally increases phagocytosis of apoptotic cells (consistent with what has been previously reported), we normalized the change in phagocytosis for each inhibitor concentration to the appropriate control (right panel). After normalizing, we observed that ROCK inhibition did not increase corpse uptake in LR73 cells in the presence of IGF-1 more than the increase observed basally due to Rho kinase inhibition. Thus, inhibition of ROCK does not appear to rescue IGF-1 induced engulfment suppression. Data represented as mean ± s.d.

Extended Data Figure 7. Inhibition of Arp2/3 mediated functions does not appear to reverse IGF-1 mediated enhancement of liposome uptake

LR73 cells were treated with CK-666 at the concentrations indicated and then assessed for uptake of liposomes in the presence of IGF-1 (n=3). Data are represented as mean ± s.d.

Extended Data Figure 8. J774 and LR73 cells phosphorylate IGF-1r upon IGF-1 stimulation similarly

J774 cells (A) or LR73 cells (B) treated with 100ng/mL mouse IGF-1 treated were assess for their ability to engulf apoptotic thymocytes (upper) or serum-starved for 6 hours and stimulated with 100ng/mL mouse IGF-1 and assessed for phosphorylation of IGF-1R by Western blot (bottom). (C) Flow cytometry histograms of IGF-1r expression on J774 cells (left), BMDM (middle), and peritoneal macrophages (right) (n=3–4). Data represented as mean ± s.d.

Extended Data Figure 9. IGF-1 and insulin do not modulate apoptotic cell uptake in the IC-21 macrophage cell line

IC-21 cells treated with indicated concentrations of mouse IGF-1 (A) and human insulin (B) were assessed for their ability to engulf apoptotic thymocytes (n=2–3). Data represented as mean ± s.d.

Extended Data Figure 10. Production of IGF-1 by peritoneal macrophages after apoptotic cell or IL-4 stimulation is mostly likely new transcription

Peritoneal macrophages were either untreated, stimulated with rIL-4 or apoptotic Jurkat cells and Igf1 mRNA (top panels) and IGF-1 protein in the supernatant (bottom panels) were assessed in a time course (n.d. refers to not detected) (n=3). Data represented as mean ± s.d.

Extended Data Figure 11. IGF-1R expression in alveolar macrophages and airway epithelial cells

Lung sections from wild type mice were stained with antibodies against alveolar macrophages (Mac-2), airway epithelial cells (CC-10), and IGF-1R. (C) is another field showing colocalization of IGF-1R and Mac-2 staining.

Extended Data Figure 12. Alveolar macrophages from LysM-Cre/Igf1^fl/fl mice have no detectable Igf-1 transcript

Alveolar macrophages isolated from LysM-Cre/Igf1^fl/fl and littermate controls were assessed for Igf-1 mRNA expression (n=2 per group). Data represented as mean ± s.d.

Extended Data Figure 13. CCSP-Cre/Igf1r^fl/fl mice exposed to HDM show a trend toward greater immune cell infiltration in the lungs and greater apoptotic cells

(A) Total cell counts of lung CD3+ CD4+ T cells (left), CD3+ CD4+ CD44+ T cells (middle), and CD3+ CD4+ CD69+ T-cells (right panel) in the lungs of CCSP-Cre/Igf1r^wt/wt and Igf1r^fl/fl mice given the full HDM course. CCSP-Cre/Igf1r^fl/fl mice exposed to HDM clearly show a trend in increased T-cells in the lung compared to Igf1r^wt/wt mice, but due to the spread in the data among the many mice analyzed, the data did not achieve statistical significance. (B) (left) Representative histology images of cleaved caspase (CC3) staining in lung sections of mice given the full HDM course. Average CC3-positive cells per mouse are quantified on the right (n=3 per group). Black arrowheads indicate positive staining. Data represented as mean ± s.e.m.

Extended Data Figure 14. Schematic detailing various HDM administration timelines tested

Schematic describing the different time courses for Igf1r deletion from Club cells (induced via administration of doxycycline) and for the allergen HDM exposure.

Extended Data Figure 15. CCSP-Cre/Igf1r^wt/wt and Igf1r^fl/fl mice exposed to HDM for regimen #2 (the challenge phase) have no significant differences in airway inflammation

(A) Total cell counts of various populations in the BAL fluid of CCSP-Cre/Igf1r^wt/wt and CCSP-Cre/Igf1r^fl/fl mice given HDM for regimen #2 (the “challenge phase” time course). (B) Total cell counts of CD3+ CD4+ T cells of draining lymph nodes of CCSP-Cre/Igf1r^wt/wt and CCSP-Cre/Igf1r^fl/fl mice given HDM for regimen #2 (the “challenge phase” time course). Data represented as mean ± s.e.m.

Extended Data Figure 16. Alveolar macrophages exposed to IL-4 produce more microvesicles

(A) Microvesicles were harvested from either control or IL-4 treated MH-S alveolar macrophages and then counted using qNano (n=3) (B) Supernatants from (A) were assessed for IGF-1 production by macrophages. Data represented as mean ± s.e.m.

Extended Data Figure 17. Macrophage-derived microvesicles suppress gene expression in lung epithelial cells exposed to house dust mite

BEAS-2B cells were treated with HDM either in the presence or absence of microvesicles isolated from mouse alveolar macrophages for 3 hours and then assessed for expression of FGF2, KLF4, IFIT2, and PTX3 (n=6). Data represented as mean ± s.e.m.

Extended Data Figure 18. Potential model of alveolar macrophage and airway epithelial crosstalk

Exposure of airways to allergens, such as HDM, can cause apoptotic cell death as well as IL-4 and IL-13 production, from mast cells and type 2 innate lymphoid cells (ILC2). These cytokines, along with apoptotic cells, trigger alveolar macrophages to produce IGF-1. The released IGF-1 (A) then acts on the airway epithelium to elicit two actions: first, to decrease the uptake of apoptotic cells and second to enhance the uptake of macrophage-derived microvesicles. These microvesicles (B) dampen inflammatory cytokine production by the airway epithelial cells.

Figure 1. IGF-1 dampens apoptotic cell engulfment and enhances liposome uptake by non-professional phagocytes

a, Effect of 11 growth factors on apoptotic cell engulfment by LR73 cells (each normalized to vehicle control). b, (Left) Engulfment by LR73 cells treated with increasing concentrations of IGF-1. (Right) Immunoblotting for IGF-1R and Akt phosphorylation. c, Apoptotic cell uptake by BEAS-2B cells treated with human IGF-1. d, IGF-1 does not affect Annexin V binding to apoptotic thymocytes (n=6). e, f, Reversal of IGF-1 mediated engulfment inhibition of LR73 cells by IGBFP3 (e), or BEAS-2B epithelial cells by IGF-1R neutralizing antibody (f). g, IGF-1-mediated reduction in engulfment is reversed by the IGF-1R inhibitor OSI-906 (left), and immunoblotting for IGF-1R and Erk1/2 phosphorylation (right). h, Liposome uptake by LR73 cells treated with either IGF-1 (50ng/mL), EGF or VEGF (100ng/mL each). i, Enhanced liposome uptake by IGF-1 is reversed by OSI-906. j, LR73 cells pretreated with IGF-1 were washed, and incubated with apoptotic thymocytes in the presence or absence of IGF-1. k, Engulfment by LR73 cells transfected with RacG12V treated and with mIGF-1. l, m, Engulfment by LR73 cells treated with Cytochalasin D (l) or Latrunculin A (m) and incubated with liposomes with or without IGF-1. n, (Left) Phagocytosis of apoptotic thymocytes by J774 macrophage cells treated with IGF-1. (Right) Immunoblotting for IGF-1R expression and Akt phosphorylation. o, p, Apoptotic cell engulfment by bone marrow derived macrophages or resident peritoneal macrophages treated with IGF-1. q, Liposome uptake by resident peritoneal macrophages treated with IGF-1. For a–q, except d, n=3; representative experiment is shown and data are mean ± s.d. n.s.- not significant. p-value of <0.05 (*), <0.01 (**), or <0.001 (***). See Supplementary Figure 1 for uncropped immunoblots.

Figure 2. Macrophages produce IGF-1 during apoptotic cell clearance

a, (Left) IGF-1 secretion by peritoneal macrophages stimulated with IL-4, apoptotic Jurkat cells or live Jurkat cells. (Right) Apoptotic or live Jurkat cells do not produce IGF-1 (representative of n=3), mean ± s.d. b, J774 cells were treated with rIL-4. After 24hrs, half of the supernatant was assessed for IGF-1 secretion by ELISA, (right panel) and the other half was incubated with PBS or IGBP3 and then tested in the phagocytosis using LR73 cells (right) (representative of n=3), mean ± s.d. Note: addition of IGFBP3 increases the basal phagocytosis due to low basal levels of IGF-1 found in the supernatant of unstimulated J774 cells. c, d, e, CCSP-Cre mice with YFP⁺ Club cells were pre-treated with 1μg IGF-1 intranasally for 1hr and then administered targets with or without IGF-1. Uptake of apoptotic thymocytes (d, n=6 per group) or liposomes (e, n=3 per group) by alveolar macrophages and lung epithelial cells were then assessed. f, Wild-type mice were given PBS, IL-4, apoptotic cells, IL-5, or IL-13 intranasally for 2 consecutive days, and the BAL fluid assessed for IGF-1 (n=2–3 mice per group for cytokines, n=5, 8 for apoptotic cell instillation). g, LysM-Cre/Igf1^wt/wt or Igf1^fl/fl mice were given PBS, IL-4 or IL-13, or apoptotic cells intranasally, and BAL fluid assessed for IGF-1 (n=6, 6, 4 mice per group for rIL-4; n=6, 4, 4 for rIL-13; n=6, 9, 9 mice per group for apoptotic cell instillation). Data are mean ± s.e.m unless otherwise indicated.

Figure 3. Mice lacking IGF-1R in airway epithelial cells have exacerbated airway inflammation

a, Schematic of HDM induced allergic airway inflammation. b, Representative images showing IGF-1R expression in bronchial epithelial cells, and its loss in CCSP-_rtTA/tetO-Cre/Igf1r^fl/f mice treated with doxycycline. c, Numbers of eosinophils, alveolar macrophages, and CD4⁺ T cells in the BAL fluid of CCSP-Cre/Igf1r^wt/wt and Igf1r^fl/fl mice administered PBS or HDM (each dot represents a mouse). d, (Left) Representative lung draining lymph nodes from CCSP-Cre/Igf1r^wt/wt and CCSP-Cre/Igf1r^fl/fl mice that were given PBS or HDM. (Right) Total CD4⁺ Tcell counts from lymph nodes. e, f, g, h, Representative hematoxylin and eosin (H&E) images (e) or PAS staining (g) of lung sections from CCSP-Cre/Igf1r^wt/wt and CCSP-Cre/Igf1r^fl/fl mice given PBS or HDM (n=3–4 mice per condition). Representative histological scoring of inflammation (f) and PAS staining (h) (n=6–10 sections and 3 mice per condition). All data are presented as mean ± s.e.m.

Figure 4. IGF-1R expression in airway epithelial cells is required in the sensitization phase of airway inflammation

a, Schematic of IGF-1R deletion prior to the sensitization phase to assess its effect on early stages of inflammation. b, Numbers of eosinophils, alveolar macrophages, and CD4⁺ T cells in the BAL of CCSP-Cre/Igf1r^wt/wt and CCSP-Cre/Igf1r^fl/fl mice primed with PBS or HDM. c, d, e, Analysis of IL-4, IL-5, eotaxin-1, and IL-6 (via Luminex c, e, n=3 mice per group) and TSLP (by ELISA, d, n=2, 7, 9 mice per group) in the BAL fluid from representative CCSP-Cre/Igf1r^wt/wt and CCSP-Cre/Igf1r^fl/fl mice primed with PBS or HDM. f, Schematic of generation and isolation of alveolar macrophage derived microvesicles. g, h, Representative negative-stain EM (g) or cryo-EM (h) images of microvesicles isolated from mouse alveolar macrophages. Images show spherical membrane-bound structures of a range of sizes (yellow arrows). i, ImageStreamX^™ analysis of microvesicles isolated from mouse alveolar macrophage cell line and primary mouse alveolar macrophages and stained for representative alveolar macrophage markers. j, Tunable resistive pulse sensing analysis of microvesicles from alveolar macrophages using qNano, pore size 400nm, to determine frequency and sizing of microvesicles (representative of n=3). k, BEAS-2B cells treated with IGF-1 (100ng/mL) were assessed for uptake of alveolar macrophage derived microvesicles (n=4). l, BEAS-2B cells were treated with HDM either in the presence or absence of alveolar macrophage derived microvesicles for 3 hours and assessed for expression of TSLP, CSF2, IL6, IL8 (n=4). m, Heatmap of top 10 differentially expressed genes from RNA-seq analysis of BEAS-2B cells exposed to HDM with or without alveolar macrophage-derived microvesicles. Data presented as mean ± s.e.m. n.d. is not detected