Macrophages redirect phagocytosis by non-professional phagocytes and influence inflammation

To address potential cross-regulation among phagocytes, we initially tested a panel of eleven soluble mediators that have been linked to tissue repair, inflammation dampening, and tissue morphogenesis¹³, for their ability to modulate engulfment by LR73 fibroblasts, a non-professional phagocytic cell line. Of the eleven factors tested, only insulin-like growth factor 1 (IGF-1) significantly dampened apoptotic cell uptake, at concentrations reported in mouse and human serum (100–600 ng/mL)¹⁴ (Fig. 1a, 1b). Other factors such as EGF, FGF2, VEGF, PDGF-AA and PDGF-BB did not alter engulfment even over a range of concentrations (Extended Data Fig. 1a), although they all elicited early downstream signaling events (Extended Data Fig. 1b–e). The engulfment dampening effect of IGF-1 was also seen with airway epithelial cell lines BEAS-2B and 16HBE14o-, and the endothelial cell line SVEC-40 (Fig. 1c, Extended Data Fig. 1f, g).

The IGF-1 effect was not due to masking phosphatidylserine (PtdSer) on the apoptotic cells (Fig. 1d). LR73 cells express the IGF-1 receptor (IGF-1R) and IGF-1 treatment elicited phosphorylation of Akt, a signaling molecule downstream of IGF-1R (Fig. 1b). In blood, IGF-1 is bound to IGF binding proteins (IGFBPs) that stabilize and sequester IGF-1¹⁵. Addition of IGFBP3 with IGF-1 restored the phagocytic capability of LR73 cells (Fig. 1e), suggesting that binding of active IGF-1 to IGF-1R was necessary. A neutralizing antibody against the human IGF-1R reversed the engulfment-dampening effect of IGF-1 on BEAS-2B cells (Fig. 1f). Further, OSI-906 (a small molecule kinase inhibitor of IGF-1R currently tested in clinical trials¹⁶), and another inhibitor, NVP-AEW541, rescued the engulfment capacity of LR73 cells treated with IGF-1, with concomitant decrease in phosphorylation of IGF-1R and Erk (Fig. 1g and Extended Data Fig. 2a)¹⁷. IGF-II and insulin, which share structural similarity with IGF-1 but have lower affinities to IGF-1R¹⁸, could also reduce apoptotic cell uptake (Extended Data Fig. 4), albeit at higher concentrations. Thus, productive signaling through the IGF-1R is necessary for IGF-1 to dampen apoptotic cell uptake.

In contrast to the inhibitory effect on apoptotic cell uptake, IGF-1 enhanced the uptake of PtdSer containing liposomes of 150–200nm in size (Fig. 1h). This was not seen with EGF or VEGF. IGF-1 enhancement required IGF-1R signaling, as the inhibitor OSI-906 reduced the increased liposome uptake (Fig. 1h and 1i). Thus, IGF-1 can redirect phagocytosis by non-professional phagocytes, suppressing uptake of larger apoptotic cells while enhancing internalization of smaller particles.

The IGF-1 effect was reversible, as washing phagocytes pre-treated with IGF-1 resulted in near complete restoration of engulfment (Fig. 1j). The effect of IGF-1 on phagocytes was also rapid, as adding IGF-1 simultaneously with apoptotic cells inhibited engulfment similar to pre-treated cells (data not shown), suggesting interference at early step(s) during phagocytosis. Blocking some early signaling pathways downstream of IGF-1R, such as Akt, mTOR, Erk or PI3-kinase¹⁹, did not rescue of the IGF-1 effect (Extended Data Fig. 3a–d). Although IGF-1 can activate RhoA, which is known to dampen apoptotic cell engulfment²⁰, inhibiting RhoA-mediated signaling did not reverse IGF-1-mediated engulfment suppression (Extended Data Fig. 3e,f). Overexpressing activated Rac1 (which promotes apoptotic cell engulfment^21,22) bypassed attenuation of apoptotic cell uptake (Fig. 1k), suggesting that IGF-1R acts at or before a step regulated by active Rac1. Since apoptotic cell uptake requires both Rac1-dependent actin polymerization and de-polymerization²³, we explored actin regulation via IGF-1. Cytochalasin D (cytoD) that promotes actin depolymerization²⁴. Phagocytosis of apoptotic cells was potently inhibited by cytoD⁴ (data not shown). While cytoD did not affect the basal uptake of liposomes by LR73 cells, it blocked the IGF-1 induced increase in liposome uptake (Fig. 1l). Of note, cytoD treated LR73 cells appeared morphologically normal for the duration of the assay. Latrunculin A, which promotes actin depolymerization via a different mechanism²⁵, also reversed IGF-1 mediated enhancement of liposome uptake, without affecting the basal uptake of liposomes (Fig. 1m). Additionally, Arp2/3 complex can regulate phagocytosis through the formation of branched actin networks; however, CK-666, a small molecule inhibitor of Arp2/3²⁶, had minimal effect on the IGF-1 mediated increase of liposome uptake (Extended Data Fig. 3g). Collectively, IGF-1 mediated modulation of phagocytosis involves rapid and reversible modification of F-actin/G-actin dynamics, but likely not Arp2/3 mediated functions.

We next tested the effect of IGF-1 on professional phagocytes. Adding IGF-1 to the macrophage cell lines J774 or IC-21, or primary macrophages (bone marrow-derived or resident peritoneal macrophages), did not affect corpse uptake, even with supra-physiological concentrations of IGF-1, at every time point analyzed (Fig. 1n–p, and Extended Data Fig. 4a, d). Peritoneal macrophages exposed to IGF-1 also had comparable liposome uptake as control treated cells (Fig. 1q). Importantly, IGF-1R expression was confirmed in all macrophage cell types analyzed (Extended Data Fig. 4b) and IGF-1 initiated early signaling in macrophages, as assessed by IGF-1R and Akt phosphorylation (Fig. 1n, Extended Data Fig. 4a). These data suggest that IGF-1 mediated modulation of phagocytosis does not extend to macrophages.

High levels of serum IGF-1 is linked to IGF-1 production in the liver, but macrophages can also produce IGF-1 in response to the cytokine IL-4^27,28, raising the possibility that within a tissue, IGF-1 release from macrophages could regulate nearby non-professional phagocytes. We first confirmed IGF-1 secretion by peritoneal macrophages treated with IL-4 (Fig. 2a). Furthermore, resident peritoneal macrophages exposed to apoptotic Jurkat cells (but not live cells) produced IGF-1 (Fig. 2a). IGF-1 protein induction appears to be from newly transcribed Igf1 message (Extended Data Fig. 5a). Macrophage-produced IGF-1 could also suppress apoptotic cell uptake by LR73 cells, reversed by IGFBP3 addition (Fig. 2b).

To test the communication between phagocytes in vivo, we chose the lung, as alveolar macrophages and airway epithelial cells reside in close proximity, and both engulf apoptotic cells⁴. Alveolar macrophages, similar to the resident peritoneal macrophages, are derived from fetal monocytes²⁹, and are readily isolated via CD11c⁺ Siglec F⁺ expression, while the airway epithelial cells can be isolated and tracked via available genetic tools³⁰. IGF-1R expression was prominent in the airway epithelial cells (Extended Data Fig. 5b) but was also detectable at lower levels in alveolar macrophages. To test the effect of IGF-1 in vivo, we used mice where airway epithelial cells are specifically marked by YFP. When either apoptotic cells or liposomes were administered intranasally with IGF-1 (Fig. 2c), YFP⁺ airway epithelial cells had decreased apoptotic cell engulfment, but enhanced liposome uptake (Fig. 2d, e). Importantly, lung macrophages from the same IGF-1 treated mice were unaffected in their ability to engulf these targets (Fig. 2d, e). Thus, IGF-1 can differentially affect engulfment by professional and non-professional phagocytes closely residing in the same tissue.

We next tested IGF-1 release by alveolar macrophages in vivo. Intranasal administration of IL-4 or apoptotic Jurkat cells resulted in an increased level of IGF-1 in the bronchoalveolar lavage (BAL) fluid (Fig. 2f). When we tested IL-13 or IL-5 (cytokines linked to lung homeostasis and inflammation³¹), IL-13 (whose receptor shares a common subunit with the IL-4R³²), but not IL-5, induced IGF-1 production (Fig. 2f). Since the source of IGF-1 is difficult to distinguish in these experiments, we generated LysM-Cre/Igf1^fl/fl mice ^33,34, targeting IGF-1 deletion in the myeloid lineage. LysM-Cre/Igf1^fl/fl mice showed loss of Igf1 mRNA in alveolar macrophages (Extended Data Fig. 5c) and IGF-1 induction after IL-4, IL-13 or apoptotic cell treatment (Fig. 2H). This suggested the macrophage/myeloid population as the predominant source of inducible IGF-1 in the lung.

To investigate IGF-1/IGF-1R signaling in inflammation, we used a model of airway inflammation induced by the allergen house dust mite (HDM), where apoptotic cell recognition by airway epithelial cells influences inflammation^4,35. We administered HDM intranasally in the allergen sensitization and challenge stage, modeling the natural route of allergen encounter (Fig. 3a). When we first tested LysM-Cre/Igf1^fl/fl mice, these mice proved unsuitable, as we continued to detect variable levels of IGF-1 in the BAL fluid after HDM administration (likely due to leakage of serum IGF-1 during inflammation). Therefore, we targeted instead the IGF-1R on the epithelial cells. We crossed Igf1r^fl/fl mice³⁶ with CCSP-_rtTA/tetO-Cre transgenic mice^30,37, the latter driving Cre under the Club cell secretory protein (CCSP) promoter in the epithelial cells of the trachea, bronchi and bronchioles^30,37. This CCSP-_rtTA/tetO-Cre strain allows for inducible Cre expression through doxycycline administration via drinking water, thereby allowing normal development and gene deletion just prior to allergen exposure. We observed near complete loss of IGF-1R on epithelial cells of CCSP-Cre/Igf1r^fl/fl mice after doxycycline treatment (Fig. 3b). After sensitization and challenge with low-endotoxin HDM, the CCSP-Cre/Igf1r^fl/fl mice had greater airway inflammation based on several parameters. First, there was marked increase in eosinophils and CD4⁺ T-cells in the BAL fluid (Fig. 3c) and trending increase in inflammatory cells in lungs (Extended Data Fig. 6a). Second, lung-draining lymph nodes were larger with more CD4⁺ T-cells (Fig. 3d). Third, lung sections showed increased peribronchial and perivascular cellular infiltration (Fig. 3e, f) and greater mucus accumulation after HDM treatment (Fig. 3g, h). Fourth, HDM-treated CCSP-Cre/Igf1r^fl/fl mice displayed increased airway reactivity after methacholine challenge, a measure of the bronchial hyper-responsiveness^38,39 (Extended Data Fig. 6b). There were more apoptotic cells in lung sections (cleaved caspase 3 staining), likely due to greater inflammation (Extended Data Fig. 6c). These data suggested a requirement for IGF-1R in the airway epithelial cells in minimizing inflammation in this model.

These observations were initially surprising, as we were expecting the loss of IGF-1R on airway epithelial cells to improve apoptotic cell clearance and thereby attenuate, rather than worsen, inflammation. This prompted us to examine the temporal requirement of IGF-1/IGF-1R signaling in epithelial cells during allergen exposure. To distinguish the requirement for IGF-1/IGF-1R signaling at the sensitization versus challenge phase, we administered doxycycline at different times: deleting IGF-1R expression before allergen sensitization (Fig. 4a), or after the initial allergen sensitization but before the allergen challenge phase (Extended Data Fig. 7a). Mice with IGF-1R deletion prior to allergen sensitization showed a significant increase in inflammatory parameters (Fig. 4b), while deleting IGF-1R after the sensitization phase had minimal effect on disease severity (Extended Data Fig. 7b, c).

Airway epithelial cells encountering allergen can produce cytokines, such as TSLP, CSF-2/GM-CSF (affecting dendritic cell maturation), as well as IL-33 and IL-25 that drive type 2 innate lymphoid cells (ILC2s) to proliferate and produce IL-13⁴⁰. Bronchial epithelial cells from human asthmatics produce elevated levels of IL-6, IL-8, and CSF-2⁴¹, and TSLP released from airway epithelial cells (mice and human) can exacerbate airway inflammation⁴². When we assessed the cytokine profile in the BAL fluid of CCSP-Cre/Igf1r^fl/fl mice sensitized with HDM, levels of IL-6 and TSLP were greater (Fig. 4d, e). We did not find increased levels of IL-33, as seen with Rac1 deletion in airway epithelial cells⁴, but this may be due to differences in inflammatory pathways regulated by the loss of IGF-1R versus Rac1. These results suggest that IGF-1R in airway epithelial cells limits inflammatory cytokines during the initial antigen exposure/sensitization phase.

As for trigger(s) for IGF-1 release from the alveolar macrophages in the sensitization phase, while IL-4 is often considered a classic Th2 cytokine from the adaptive immune response^43,44, both IL-4^45–51 and IL-13⁴⁰ can be produced during the allergen sensitization phase by resident mast cells, basophils, and/or ILC2s. In fact, IL-4 was enhanced (along with IL-5 and eotaxin-1) in the BAL fluid of CCSP-Cre/Igf1r^fl/fl mice sensitized with HDM (Fig. 4c). Thus, IL-4 and/or apoptotic cells present during the initial allergen exposure could elicit IGF-1 production from alveolar macrophages.

Alveolar macrophages are reported to release microvesicles containing anti-inflammatory mediators during smoking-induced lung injury⁵². Since IGF-1 enhances liposome uptake by airway epithelial cells, we tested whether microvesicles from alveolar macrophages may impact airway epithelial cell response to HDM. Using differential centrifugation and filtration, we isolated and characterized microvesicles from alveolar macrophages (Fig. 4f). First, negative stain and cryo-electron microscopy revealed membrane bound spherical structures >100nm in diameter (Fig. 4g, h). Second, via ImagestreamX^™ analysis, particles from an alveolar macrophage cell line or primary mouse alveolar macrophages carry markers of lung macrophage origin, CD11c or SiglecF, with variable Annexin V staining (Fig. 4i). Third, tunable resistive pulse sensing analysis of microvesicles revealed a mean particle size of 357±148.5nm (Fig. 4j), within the reported 100–1000nm range for microvesicles⁵³. Further, IL-4 treatment of macrophages increased microvesicles secretion (Extended Data Fig. 8a). Also, uptake of these microvesicles by BEAS-2B airway epithelial cells was enhanced by IGF-1 (Fig. 4k). Importantly, adding alveolar macrophage derived microvesicles to HDM-treated BEAS-2B cells resulted in significantly lower TSLP, CSF2, IL6, and IL8 induction (Fig. 4l). To further explore this, we performed RNA-seq of BEAS-2B cells treated with HDM ± microvesicles. While HDM treatment increased several known genes associated with asthma in humans^54–57, such as FGF2, KLF4, Interferon-Induced Protein with Tetratricopeptide Repeats 2 (IFIT2), and Pentraxin 3 (PTX3), adding microvesicles from alveolar macrophages suppressed transcription of these genes in the epithelial cells (Fig. 4m, Extended Data Fig. 8b). Thus, microvesicles released from alveolar macrophages acting on airway epithelial cells exposed to allergen could dampen airway inflammation.

The data presented here provide several new insights on phagocytosis and tissue inflammation. First, it has long been known that professional and non-professional phagocytes reside in proximity and can engulf dying cells and debris, yet communication between these phagocytes was not understood. The data presented here identifies a rapid, transient, and reversible regulation, wherein soluble IGF-1 from macrophages influences the type of particle uptake by epithelial cells. This transient effect might allow the macrophages to temporarily ‘redirect’ the non-professional phagocytes toward other function(s). Such a possible hierarchy in cell clearance could provide temporal and spatial cross-communication within a given tissue¹⁰. Second, our data identifies a two-part regulation of epithelial cells by macrophages which impacts airway inflammation, i.e., the secretion of IGF-1 that redirects particle uptake, and the release of microvesicles that dampen inflammatory cytokine production by epithelial cells (Extended Data Fig. 9). Third, IGF-1 is a growth factor widely linked to growth, cellular proliferation, and aging^58–61. Global IGF-1 deletion in mice results in dwarfism and perinatal lethality, and IGF-1 mutations are linked to human diseases^34,62,63. This work identifies a previously unappreciated biological function for the IGF-1/IGF-1R axis as a modulator of airway hyper-responsiveness to allergens with potential therapeutic relevance.