Research ArticleMACROPHAGES

Identification of a nerve-associated, lung-resident interstitial macrophage subset with distinct localization and immunoregulatory properties

See allHide authors and affiliations

Science Immunology  27 Mar 2020:
Vol. 5, Issue 45, eaax8756
DOI: 10.1126/sciimmunol.aax8756

Airway-hugging macrophages

Effective immune defense in the lungs relies on myeloid cells that phagocytose, process, and present foreign substances that enter the airways including pathogens. Ural et al. studied a subset of pulmonary interstitial macrophages expressing CD169 that are predominantly found within the bronchovascular tree. These nerve- and airway-associated macrophages (NAMs) are self-renewing, yolk sac–derived cells requiring colony-stimulating factor 1 (CSF1) for their development. NAMs serve an immunoregulatory role during responses to lung inflammation. These findings provide a deeper insight into the specialized myeloid subsets that contribute to maintaining pulmonary homeostasis.


Tissue-resident macrophages are a diverse population of cells that perform specialized functions including sustaining tissue homeostasis and tissue surveillance. Here, we report an interstitial subset of CD169+ lung-resident macrophages that are transcriptionally and developmentally distinct from alveolar macrophages (AMs). They are primarily localized around the airways and are found in close proximity to the sympathetic nerves in the bronchovascular bundle. These nerve- and airway-associated macrophages (NAMs) are tissue resident, yolk sac derived, self-renewing, and do not require CCR2+ monocytes for development or maintenance. Unlike AMs, the development of NAMs requires CSF1 but not GM-CSF. Bulk population and single-cell transcriptome analysis indicated that NAMs are distinct from other lung-resident macrophage subsets and highly express immunoregulatory genes under steady-state and inflammatory conditions. NAMs proliferated robustly after influenza infection and activation with the TLR3 ligand poly(I:C), and in their absence, the inflammatory response was augmented, resulting in excessive production of inflammatory cytokines and innate immune cell infiltration. Overall, our study provides insights into a distinct subset of airway-associated pulmonary macrophages that function to maintain immune and tissue homeostasis.


Tissue-resident macrophages are myeloid cells that are distributed in virtually all organs in mice and humans. Recently, resident macrophages in mucosal tissues have received more attention, and emerging evidence suggests that these cells play critical roles in maintaining immune and tissue homeostasis under steady-state or inflammatory conditions (14). In addition, studies have identified the importance of tissue-specific microenvironment in regulating the gene expression profile of resident macrophage populations (58). The local proliferation and maintenance of macrophages depend on environmental signals such as colony-stimulating factor 1 (CSF1), granulocyte-macrophage CSF (GM-CSF), or transforming growth factor–β (TGF-β) (911).

CD169+ tissue-resident macrophages are located in secondary lymphoid tissues and nonlymphoid tissues such as the intestine (1216). In lymphoid tissues, these macrophages are positioned strategically to serve as gatekeepers for phagocytosis and clearance of invading pathogens from the blood (17, 18) and lymph (19, 20) and to promote erythropoiesis in the bone marrow (BM) (13). In the intestine, these macrophages promote the recruitment of inflammatory monocytes after dextran sodium sulfate–induced colitis (12).

The lung is a complex organ with specialized structures to allow for adequate gas exchange. Currently, it is well established that the lung harbors two distinct populations of macrophages, alveolar macrophages (AMs) and interstitial macrophages (IMs), with the IMs composed of several subsets (10, 21). Pulmonary macrophages have been implicated in maintaining lung homeostasis by immune surveillance and clearance of dead cells, debris, and invading pathogens. AMs, the most studied lung-resident macrophage population, arise from the yolk sac but, over time, are replaced by fetal liver monocytes (2224). Development and maturation of AMs are highly dependent on GM-CSF (23, 25, 26). Typically, IMs are located in the interstitium, along with dendritic cells and lymphocytes. Recently, these IMs have been classified into two distinct populations (21) or three subpopulations (27) mainly based on their phenotype and gene signature. However, our understanding of the anatomical/spatial positioning, ontology, maintenance, and function of IM subsets during homeostasis and infection is still limited.

We focused our study on a population of CD169+ lung-resident macrophages that are primarily localized around the large bronchiolar airways and adjacent to airway-associated nerves. This unique nerve- and airway-associated macrophage (NAM) population is morphologically and transcriptionally distinct compared with AMs. NAMs are self-renewing, yolk sac derived, and require the CSF1-CSF1R (colony stimulating factor 1 receptor) signaling axis for development and survival. Our understanding of the different roles of lung macrophages is largely based on studies that have used global macrophage-depleting techniques such as clodronate liposomes or use of CD11c-DTR or CD11b-DTR mice that can deplete macrophages, dendritic cells, and/or other myeloid cell subsets (2831). These studies have largely attributed the outcomes to the role of AMs; however, the contribution of NAMs in these outcomes has not been considered. Our results suggest that NAMs substantially contribute to regulating inflammation after its induction in the lung, whereas AMs primarily play a critical role in viral clearance. Thus, our study provides new insights into the specialized functional roles of individual subsets of lung-resident macrophages.


Characterization of a CD169+ IM population in lung tissue

We began our study by visualizing the distribution of tissue-resident macrophages within lung tissue. Wild-type (WT) and CD169-tdTomato frozen lung sections were immunostained and imaged by confocal microscopy. AMs expressed CD169 and CD11c and were localized inside the alveolar spaces throughout the lung parenchyma (Fig. 1A and fig.S1A, purple arrowheads). However, we also observed a population of IMs that expressed CD169 but not CD11c (Fig. 1A and fig. S1A, white arrowheads). This IM population was primarily localized around the airways in close apposition to the airway epithelial cells (Fig. 1A and fig. S1A, white arrows). The characteristic pattern of localization of these IMs is illustrated in Fig. 1B, in which a whole lung lobe from CD169-tdTomato mice was cleared and imaged using a Zeiss Light Sheet microscope for CD169-tdTomato–expressing cells (Fig. 1B). Note that because of their close association with airway epithelial cells throughout the lung, the bronchial tree was readily illuminated as though the epithelial cells themselves were specifically stained. In addition to their distinct geographical localization, these macrophages also exhibited a distinct morphology compared with AMs. Unlike AMs that are large and rounded, NAMs appeared elongated and had dendritic processes (Fig. 1A, purple and white arrowheads). To determine whether this population of IMs is conserved between species, we also analyzed human lung tissue, which exhibited a similar population of CD169+CD11c IMs with similar morphology and localization near the airways (Fig. 1C and fig. S1C). Because CX3CR1 is a chemokine receptor that is expressed by several tissue-resident macrophage populations such as arterial macrophages and microglia (3234), we immunostained CX3CR1-eGFP (enhanced green fluorescent protein) mice lung sections for CD169 and CD11c. We found that CX3CR1 was expressed on CD169+ airway-associated IMs (white arrowheads) and CD169 IMs (green arrows) but not AMs (purple arrowheads) (Fig. 1D and fig. S1D). Because gut-resident macrophage populations in muscularis are associated closely with enteric nerves (35, 36), we determined whether these subsets of pulmonary macrophages interacted with pulmonary nerves. Consistent with other nerve-associated resident macrophages, the airway-associated macrophages were observed closely interacting with β-tubulin and tyrosine hydroxylase (TH)–expressing nerves (Fig. 2 and fig. S2), which heavily innervated the tissue areas around the large airways. Therefore, we have termed this IM population NAMs.

Fig. 1 Airway-associated IMs display distinct morphology and localization.

(A) Stitched (left) and single (right) confocal images of C57BL/6 naïve lungs immunostained for EpCAM (blue), CD11c (green), and CD169 (red). AMs (CD11c+CD169+) (purple arrows) and NAMs (CD169+CD11c) (white arrows) depicting two distinct populations of CD169+ macrophages. Scale bars, 400 μm (stitched images) and 80 μm (single image). (B) Zeiss Light Sheet Z.1 image of clarified CD169-tdTomato lung. Scale bar, 300 μm. (C) Healthy human lungs immunostained with CD169 (red) and CD11c (green). AMs (CD11c+CD169+) (purple arrows) and NAMs (CD169+CD11c) (white arrows). (D) CX3CR1-eGFP lungs immunostained with CD11c (green) and CD169 (red). NAMs (CD169+CD11cCX3CR1+) (white arrows), and AMs (CD169+CD11c+CX3CR1) (purple arrows). AW, airways. Scale bar, 80 μm. Data are representative of two to three independent experiments (n = 3 to 5).

Fig. 2 Airway-associated IMs interact closely with nerves.

Confocal images of C57BL/6 naïve lungs immunostained for EpCAM (purple), βIII-tubulin (green), CD11c (cyan), and CD169 (red). Magnified views of the boxed areas in subsequent isosurfaced images (left to right). Scale bars, 100, 50, 20, and 7 μm (left to right). Data are representative of three to four independent experiments (n = 3).

Phenotypic and genomic profiling of pulmonary macrophages

We next examined the phenotypic properties of NAMs using imaging and flow cytometry to compare these cells with AMs and other IMs (CD11c+CD169; henceforth referred to as CD169 IMs). Further phenotypic analysis revealed that NAMs expressed F4/80, CD169, MerTK, and CD64 (fig. S3A and table S1). Moreover, we observed that CD169 IMs (CD11c+CD169) and NAMs (CD169+CD11c) appeared to be two distinct subsets of IMs (fig. S3A). To understand molecular differences between the two major populations of CD169+ macrophages, we conducted bulk RNA sequencing (RNA-seq) to analyze the transcriptome of sorted NAM and AM subsets isolated from lungs of the same mice (fig. S3B, gating strategy). Reproducibility between biological replicates was verified through Pearson’s correlation coefficient analysis (fig. S3C). Principal components analysis (PCA) revealed that NAMs have a distinct gene expression profile compared with AMs and CD169+ marginal zone splenic-resident macrophages [CD169+ marginal zone macrophage (MZM)] (fig. S3D; splenic macrophage transcriptomics data were previously published and used here for comparison) (18). Differentially expressed gene (DEG) analysis [false discovery rate (FDR) ≤ 0.05 and fold change (FC) ≥ 2] revealed 1690 DEGs between AMs and NAMs (fig. S3E). Moreover, we observed that NAMs were heavily skewed toward a regulatory genetic profile compared with AMs, characterized by high expression of Retnla (resistin-like molecule alpha/FIZZ1), C1q, Il10, and several other M2 macrophage–related genes (fig. S3F) as well such as Bmp2, which is up-regulated in certain populations of macrophages that interact with local nerves (35, 36). Gene ontology (GO) enrichment analysis (37) of up-regulated genes in NAMs compared with AMs revealed the terms immune processes and immune response related among the top five most significant biological processes (ranked according to P value). The genes related to immune processes and responses accounted for more than 30% of all up-regulated genes (table S2). In addition, genes related to macrophage activation, macrophage chemotaxis, and macrophage differentiation were significantly overrepresented in the up-regulated genes (table S3). GO enrichment analysis of down-regulated genes among NAMs versus AMs showed no enrichment of immune-related processes among the top 40 most significant biological processes (ranked according to P value). Among the down-regulated genes, the most significantly overrepresented ones were related to metabolism such as regulation of lipid, fatty acid, and reactive oxygen species metabolism (table S4). In summary, the GO analyses for both, up- and down-regulated genes, indicated NAMs to be significantly enriched in immunoregulatory genes.

To ensure that NAMs were a homogeneous IM subpopulation that was truly distinct from other pulmonary macrophage subsets, we validated our bulk RNA-seq results by performing single-cell RNA-seq (scRNA-seq) (Fig. 3). Lung-resident macrophages (MerTK+CD64+) (38) were sorted, and scRNA-seq was performed using the 10x Genomics platform. Using dimensionality reduction through t-distributed stochastic neighbor embedding (t-SNE) analysis, we identified five potential clusters within the lung tissue–resident macrophage populations. Although t-SNE analysis indicated that clusters 1 and 2 were separate populations, a careful gene expression analysis indicates that these two populations are AMs with minor transcriptome heterogeneity (Fig. 3B). The NAMs were identified as cluster 3 by the high expression of genes such as C1qa, C1qb, C1qc, major histocompatibility complex class II (MHCII) genes (H2-Eb1, H2-Ab1, and H2-Aa), Mgl2, Cd83, Apoe, Pf4, and Tmem176a (Fig. 3, B to E) from the population sequencing data. Cluster 4 contained cells that highly expressed cell cycle genes and stem cell–related genes, which is suggestive of a proliferative progenitor population (Fig. 3, A and F). Gene expression profile of cluster 5 likely represents a separate population of IMs distinct from the NAMs. This cluster was enriched for cells that expressed genes such as Icam2, Ly6a, and Lyve1 but failed to express high levels of MHCII genes. Together, these data clearly indicate that NAMs exhibit a distinctive gene signature, morphology, and geographical localization that set them apart from the AMs and other macrophage subsets in the lung.

Fig. 3 Seurat analysis of unbiased scRNA-seq of pulmonary macrophages.

(A) t-SNE plot identifying five clusters of MerTK+CD64+ pulmonary macrophage subsets with distinct gene expression profiles. (B and C) Heat maps of hierarchical cluster analysis of clusters 1 through 5. (B) Top 42 DEGs between clusters 1, 2, and 3. (C) Top 78 DEGs between clusters 3 and 5. (D) t-SNE plots of specific genes highly enriched in cluster 3 (NAMs). (E) Violin plots showing expression pattern of additional genes that were enriched in cluster 3 (NAMs) compared with all other clusters. (F) Violin plots showing expression pattern of genes that were enriched in cluster 4 compared with all other clusters.

NAMs originate from yolk sac and depend on CSF1-CSF1R signaling for development

It has been previously demonstrated that macrophage origins and developmental pathways are linked to their specific functions in their tissues of residence; thus, we next determined the ontogeny and development of NAMs (9, 3841). To evaluate the origin of NAMs, we crossed Cx3cr1CreER with a tdTomato stop-floxed reporter strain (42). Treatment of these mice with tamoxifen (TAM) at embryonic day 8.5 (E8.5) induced the irreversible expression of the tdTomato reporter in yolk sac–derived CX3CR1+ cells and their progeny, which allowed us to determine whether the origin of NAMs was embryonic and/or postnatal (34). Through confocal microscopy and flow cytometry analysis, we found that most of the NAMs from Cx3cr1-CreEr-Rosa26-tdTomato mice were positive for tdTomato expression (Fig. 4A and fig. S4A), suggesting that NAMs are embryonically derived and more specifically through the yolk sac. NAMs, AMs, and IMs were all negative for the tdTomato expression in the absence of TAM administration (fig. S4B).

Fig. 4 NAMs are yolk sac derived and require CSF1-CSF1R axis for development.

(A) Immunofluorescence of E19.5 (left) and 6-week-old (right) E8.5 TAM-treated Cx3cr1-CreEr-Rosa26-tdTomato lungs labeled by tdTomato+ (blue), CD169 (red), and CD11c (green). (B to E) Immunostaining of CD169 (red) and CD11c (green) of lung (B and D) and flow cytometry analysis of NAMs, CD169IMs, and AMs (C and E) of Csf2−/− (left) and Csf1op/op (right) mice. Scale bars, 80 μm. Data are representative of two independent experiments. Data shown as means ± SEM (n = 4 to 5, **P < 0.01 and ****P < 0.0001, Student’s t test). KO, knockout.

Recent studies have shown that tissue-resident macrophages originate from yolk sac or fetal liver progenitors (22, 43, 44). During embryonic development, mice lacking the master regulator PU.1 (45) or CSF1R (25) fail to develop tissue-resident macrophages, such as microglia (4548) and Langerhans cells (45, 47, 49). In addition to its importance in development, CSF1 is crucial for the local proliferation and repopulation of macrophages (25, 44, 50). Because PU.1 was required for the development of yolk sac macrophages (45), we confirmed the expression of PU.1 in NAMs (fig. S4C), thus further confirming NAMs as a yolk sac–derived macrophage. Next, we sought to determine which growth factor is important for the development of NAMs. Although GM-CSF (encoded by Csf2 gene) is known to be a crucial growth factor for the development and maintenance of AMs (23), we did not observe any defect in the number or localization of NAMs in the lungs of Csf2−/− mice (Fig. 4, B and C). We also did not observe any reduction in the number of CD169 IMs in the absence of GM-CSF. However, NAMs expressed high levels of CSF1R (fig. S4D), and their numbers were markedly reduced in Csf1r−/− (fig. S4, E and F) (51) mice. To further evaluate the importance of CSF1R for NAM maintenance, we treated mice with anti-CSF1R blocking antibody and analyzed cell numbers by flow cytometry. Our results indicated that CSF1R was crucial for the maintenance of NAMs but not for CD169 IMs or AMs (fig. S4G). Moreover, NAM numbers were also substantially reduced in mice that lacked the ligand for CSF1R (Csf1op/op mice) (Fig. 4, D and E) (52). This suggests that NAMs are dependent on the CSF1-CSF1R axis rather than the GM-CSF–GM-CSFR axis for development and/or maintenance. The growth factor(s) critical for CD169 IMs remain to be identified.

NAMs are self-renewing, lung-resident macrophages

Tissue-resident macrophages have been shown to repopulate from either a self-renewing progenitor population or through circulating BM monocytes (41, 53). Our results thus far have demonstrated that NAMs are yolk sac derived and require the CSF1-CSF1R signaling axis for development and/or survival. We next determined whether NAMs were truly a tissue-resident cell population. To this end, we parabiosed CD45.1 and CD45.2 C57BL/6 congenic mice and analyzed the lung and spleen at 5 or 12 weeks after parabiosis. We found that, consistent with their circulatory characteristic, B and T cells equilibrated to a 1:1 ratio (50%) in lymphoid organs. However, as previously shown for AMs (23, 44), NAMs also failed to receive input from circulating precursors, and most of the cells in the lungs were of recipient origin indicating their tissue residency (Fig. 5A). It is noteworthy that compared with NAMs, a small percentage of the CD169 IM subset also failed to circulate and exhibited tissue residency but still had the potential to be derived from a circulatory pool (Fig. 5A).

Fig. 5 NAMs are self-maintaining resident macrophages.

(A) Chimerism of lung AMs (F4/80+, CD169+, and CD11c+), NAMs (F4/80+, CD169+, CD11c, and Ly6C), and CD169 IMs (F4/80+, CD169, and CD11c+) (left) and splenic T and B cells (right) from CD45.1-CD45.2 parabionts at 5 (top) and 12 (bottom) weeks after chimerism. (B) Immunostaining of CD169 (red), Ki-67 (green), and CD11c (cyan) in naïve C57BL/6 lungs reveals active proliferation (yellow arrows). Scale bars, 80 μm (left) and 40 μm (right). Data are representative of two experiments. Data shown as means ± SEM (n = 3 to 6, **P < 0.01 and ****P < 0.0001, paired Student’s t test).

Given that NAMs are tissue-resident macrophages, we next aimed to understand the mechanisms that regulate NAM maintenance postnatally and in adulthood during homeostatic conditions. Initially, we investigated whether NAMs undergo homeostatic proliferation by staining naïve lung tissue with Ki-67. As shown in Fig. 5B, several NAMs (orange arrows) stained positive for Ki-67 at any given time, suggesting maintenance of the self-renewal capabilities during adulthood. Unlike macrophages in the intestines (53) that are replenished over time by CCR2+ BM monocytes, WT and CCR2−/− mice exhibited similar cell numbers and localization of NAMs (Fig. 6A). Although these data indicated that CCR2+ BM monocyte precursors were not required for the maintenance of NAMs, they did not eliminate the potential for these cells to develop from or be maintained by the progeny of BM hematopoietic stem cells (HSCs). To address this, we used an inducible HSC-specific genetic labeling system (54, 55). Transgenic Pdzk1ip1-CreER animals expressing a stop-floxed tdTomato reporter were treated with a single dose of TAM, followed by lung confocal imaging and flow cytometry (Fig. 6B and fig. S4H) to determine whether any NAMs were tdTomato+. Although many other immune cells in the lungs were labeled, very few (<10%) of the NAMs expressed tdTomato, indicating that HSCs did not contribute to the maintenance of NAMs during adulthood (Fig. 6B).

Fig. 6 NAMs are maintained independent of CCR2+ BM monocytes.

(A) CD169 (red) and CD11c (green) immunostaining (left) and flow cytometric analysis (right) depicting similar localization and frequency of NAMs (F4/80+, CD169+, CX3CR1+, CD11c, and Ly6C) between WT and CCR2−/− mice. (B) CD169 (red) and CD11c (green) immunostaining (left) and flow cytometric analysis (right) of Pdzk1ip1CreERT2-tdTtomato lungs. AMs (CD169+CD11c+) (yellow arrows), NAMs (CD169+CD11c), and HSCs originating cells (tdTomato+). (C) Confocal image (left) and flow cytometric analysis (right) of WT and CCR2−/− BM cells engrafted to DT-treated and DT-irradiated CD169-DTR mice at 3 days after irradiation. Control (left), CCR2−/− BM (center), and WT BM (right). Scale bars, 80 μm. All data are representative of two to three independent experiments. Data shown as means ± SEM (n = 4 to 5, *P < 0.05, and ***P < 0.001, Student’s t test). i.p., intraperitoneal.

Next, we determined whether the NAMs can be repopulated and the mechanisms controlling the repopulation of NAMs following depletion in adult mice. To this end, we treated CD169-DTR transgenic mice (56) with a single dose of diphtheria toxin (DT) followed by irradiation and transferred WT or CCR2−/− BM cells. We allowed hematopoietic reconstitution for 8 weeks before analyzing the lungs of the BM chimeras. The number and localization of NAMs were again similar between WT and/or CCR2−/− BM chimeras (Fig. 6C), suggesting that even after depletion, CCR2+ BM precursors were not required for the repopulation of NAMs in adult animals. Nevertheless, other small numbers of BM-derived monocyte precursor cells could repopulate an empty niche created by NAM depletion. BM chimera experiments indicated that depleted NAMs could be repopulated by BM-derived progenitor cells (fig. S4I). Together, these results suggest that NAMs are tissue-resident cells that develop from the yolk sac and are maintained into adulthood by homeostatic proliferation and/or replenishment from a local tissue progenitor population.

NAMs respond robustly to inflammatory stimuli

The localization of NAMs around airways suggested that these macrophages serve as a “gatekeeper” function within the complex structure of the lung and predicts that the NAMs will be one of the first immune cell types to interact with substances and infectious agents entering the lungs by inhalation (Fig. 1 and fig. S1). Thus, we next attempted to understand the physiological role of CD169-expressing lung-resident macrophages (AMs and NAMs) during influenza infection with multiple transgenic mouse models to deplete CD169+ macrophage subsets in the lung tissue. As indicated previously, we used the CD169-DTR mice to deplete NAMs and AMs and crossed the CD169-cre mice to B6N.129P2-Cx3cr1tm3 (DTR) Litt/J mice (referred to as NAM-DTR) to specifically deplete NAMs but not AMs. In addition, we crossed the CD169-cre mice to CD11c stop-floxed DTR mice to deplete AMs but not NAMs (referred to as AM-DTR). Using these transgenic mouse models, we can specifically dissect the role of each CD169+ macrophage subset in the lung tissue. CD169-DTR, NAM-DTR, or AM-DTR mice were treated with DT intratracheally to locally deplete specific macrophage subsets in the lung. We confirmed that NAMs were exclusively depleted in NAM-DTR mice after DT administration, whereas the AM numbers and other innate cell types were not affected (fig. S5, A and B, top). Moreover, NAMs and AMs were both depleted in the CD169-DTR mice (fig. S5B, bottom). To investigate the role of CD169+ macrophages in the lung tissue, we infected WT, CD169-DTR, AM-DTR, and NAM-DTR mice 2 days after DT treatment with a sublethal dose of PR8 influenza virus (a murine-adapted influenza A strain) and measured the viral titers 3 days postinfection (dpi). Our results indicated that the absence of AMs results in higher viral titers (Fig. 7A). However, NAMs were not vital in reducing the viral load due to influenza infection. To further validate these observations, we infected WT mice with a PR8 influenza virus encoding a GFP-tagged NS1 (NS1-GFP) (57) to allow visualization of the cells that can engulf influenza virus. Early after infection, confocal analysis revealed that AMs but not NAMs could engulf the influenza virus (Fig. 7B, white arrows). As we investigated the weight loss and survival for CD169-DTR and NAM-DTR mice, we observed that CD169-DTR mice (58) lost significantly more weight in the first few days of infection and died between days 7 and 10 after infection. NAM-DTR mice lost significantly more weight compared with WT mice (Fig. 7C), and some showed signs of becoming moribund. However, most mice did not succumb to the infection as in the mice depleted of both AMs and NAMs. Immunostaining of 4 dpi WT lung tissue with Ki-67 showed that NAMs were actively proliferating (fig. S6A) in response to the ongoing active infection. Interleukin-10 (IL-10) is an important regulatory cytokine that is essential for preventing inflammatory damage in mucosal tissues (59). Therefore, we infected IL-10–GFP mice with PR8 and analyzed the lung tissue by confocal microscopy. Through using the colocalization and isosurfacing functions of the Imaris software, we were able to analyze IL-10–producing cells. As shown in Fig. 7D, unlike AMs, NAMs were the primary producer of IL-10, suggesting that this is one mechanism by which NAMs regulate lung inflammation. These results indicated that, in the absence of AMs, mice were more susceptible to influenza infection. Overall, these results suggest that AMs are functioning to help clear the infection, whereas NAMs are important in regulating the infection-induced inflammatory process in vivo, which confirmed our genomic data and the ability of NAMs to secrete immunosuppressive factors such as IL-10 (Fig. 7D).

Fig. 7 NAMs regulate the inflammatory response and react robustly to influenza infection.

(A) FFU of C57BL/6, CD169-DTR, AM-DTR, and NAM-DTR lungs 3 dpi with 100 EID50 of PR8. (B) Confocal images of C57BL/6 lungs 24 and 48 hours after infection with NS1-GFP–expressing PR8 and counterstained with CD11c (blue) and CD169 (red). Scale bars, 50 and 20 μm. (C) Weight loss (left) and survival (center) curve of C57BL/6 and CD169-DTR mice and weight loss (right) of C57BL/6 and NAM-DTR mice after PR8 intranasal infection. (D) Confocal imaging of EpCAM (purple), CD169 (red), and CD11c (blue) in IL10-GFP mice after PR8 infection. Colocalization of CD169 and IL-10 pseudo-colored in white (left) and zoomed (right). Scale bars, 20 and 80 μm. Data are representative of two independent experiments. Data shown as means ± SEM (n = 3 to 5, **P < 0.01, ***P < 0.001, and ****P < 0.0001, two-way ANOVA with Bonferroni’s posttest and Student’s t test).

Human and mouse studies have demonstrated that lethality associated with influenza infection is due to immunopathology caused by elevated infiltration of innate immune cells and excessive cytokine production (60). Because little is known about the physiological significance of NAMs, we next determined how these cells would respond to inflammatory cues without the complication of a replicating pathogen such as influenza virus. To this end, we treated mice with the TLR3 ligand polyinosinic:polycytidylic acid [poly(I:C)] for two consecutive days and analyzed the lung tissue. NAMs responded robustly to the poly(I:C) treatment as evidenced by the remarkable expansion of NAMs early after treatment (fig. S6B). This increase in numbers could be a result of local intrinsic proliferation of NAMs or infiltration and differentiation of a monocyte population or both. The fact that even during the inflammation-induced proliferation, NAMs remained primarily localized around the airways suggest the possibility of local proliferation. It was confirmed by Ki-67 staining after poly(I:C) treatment, revealing actively proliferating NAMs (fig. S6B) similar to those observed after influenza infection. It was clear that NAMs responded robustly to inflammatory stimuli as judged by their remarkable expansion, which was confirmed by flow cytometric analysis (fig. S6C). To understand whether CCR2+ monocytes were responsible for the increase in the number of NAMs, we treated WT and CCR2−/− mice with poly(I:C) and analyzed the number of NAMs by flow cytometry. After poly(I:C) treatment, the number of NAMs was similar between the WT and CCR2−/− mice (fig. S6D), which further indicated that most of the expansion of NAMs was likely the result of local active proliferation.

Absence of NAMs leads to elevated inflammation

Given the elevated expression of regulatory genes by NAMs, we next investigated whether the inflammatory response was altered after selective depletion of NAMs in the lung. We treated WT and NAM-DTR mice with poly(I:C) after DT administration and analyzed the cytokine and chemokine production in lung tissue and bronchoalveolar lavage (BAL) fluid 8 hours after treatment. In the absence of NAMs, we observed a significant increase in proinflammatory cytokines and chemokines such as IL-6, CCL2, CCL3, and CCL5 (fig. S6, E and G). In accordance with the increased production of CCL2 and CCL3, we also observed an increased recruitment of monocytes in both lung and BAL (figs. S6, F and H). Because we observed that NAMs preferentially produced IL-10 after PR8 infection, we treated IL-10–GFP mice for two consecutive days with poly(I:C) and analyzed the lung tissue by confocal microscopy. As shown in figs. S6 (I and J), NAMs were the primary producers of IL-10, which further confirmed IL-10 as a potential mechanism by which NAMs regulate inflammation in the lungs.

Overall, our study highlights the complexity that underlies the biology of the several different subsets of tissue-resident macrophages in the lung. Furthermore, our results provide ample evidence for the need to reevaluate the precise roles of each subset of tissue-resident macrophages in the lungs under steady-state and inflammatory conditions in vivo.


Tissue-resident macrophages play critical roles in maintaining immune and tissue homeostasis in mucosal organs (14). Moreover, the tissue-specific microenvironment can regulate the genetic landscape of macrophage populations (58). Although studies primarily using BM-derived macrophages have been critical in increasing our understanding of macrophages in response to specific stimuli in vitro, these roles remain to be elucidated among tissue-resident macrophage subsets in local tissue environment in vivo. In the current study, we used multiple technical approaches including imaging, flow cytometry, and transcriptomics to investigate tissue-resident macrophages in the lung. Initially, our imaging studies revealed the existence of a unique subset of IMs that are distinct from AMs in many different respects. NAMs exhibit a distinctive morphology characterized by elongated cell body and protruding dendritic processes and are primarily localized around airways, whereas the AMs are distributed throughout the lung alveolar spaces. The NAMs associated closely with nerves present in the bronchovascular tree. These nerves stained positive for TH, suggesting that sympathetic fibers are present in these pulmonary nerve bundles. The reasons for this association could be many, and future studies will likely unravel the physiological significance of this close neuroimmune interaction in the lungs. It is possible that the NAM production of bone morphogenic protein 2 (Bmp2), which is highly expressed in NAMs but not by AMs, activates the neurons to provide the growth factor CSF1 needed for the survival of NAMs (36). This mechanism may be important for the autonomic contractions of the airway smooth muscle required for proper respiration, similar to the peristaltic movement of the intestinal tissue (35, 36). We identify NAMs in both murine and human lungs. In the human lung, NAMs expressed the M2 macrophage–associated protein CD206 and localized to airways, indicating that these cells may serve a similar function in mice and humans. In addition to the distinct localization and morphology, NAMs express unique surface proteins (i.e., CX3CR1) and a distinct gene signature when compared with AMs. The considerable differences in the gene expression profile of NAMs and AMs at the single-cell level and even at the population level were notable and strongly suggested that these two cell types are distinct populations. Moreover, scRNA-seq analysis validated our bulk population RNA-seq data and clearly indicated that NAMs represent a homogeneous subpopulation of a subset of IMs that are distinct from other pulmonary macrophage subsets. Our fate-mapping studies clearly showed that NAMs originate in the yolk sac and require CSF1 and not GM-CSF for development, which indicated that NAMs develop, and are maintained independently of, AMs. Furthermore, antibody-blocking studies suggested that NAMs were the only subset of lung macrophages that require CSF1R for maintenance in adulthood. The NAMs appeared to be self-maintained, and their numbers were not reduced in CCR2-deficient mice, which was similar to a previous study (44) where it was demonstrated that pulmonary tissue-resident macrophages (especially the AMs) failed to circulate and were maintained independently of CCR2-expressing monocyte precursors. Furthermore, as shown before for AMs (44), once depleted, the NAMs are repopulated by BM-derived precursors, albeit independent of CCR2+ monocytes. Although our results with CCR2-deficient mice suggest that NAMs are not replaced by monocytes, it is still possible that CCR2-independent monocyte populations may replace NAMs over time.

Although our current study primarily focused on a particular lung IM subset, NAMs, note that our knowledge of the different types of pulmonary IMs remains incomplete. Only recently, studies have begun focusing on identifying different subsets of IMs (21, 27). These studies, as well as ours, have yielded different results with respect to phenotype, development, and maintenance of IM subsets. In particular, a recent study (21) used bulk and scRNA-seq to identify two different subsets of IMs: a Lyve1loMHCIIhi population that interacted with nerve bundles and a Lyve1hiMHCIIlo subset that associated with blood vessels. We observed two IM populations that were CX3CR1+CD169hiCD11c (NAMs) that associated with nerves and CX3CR1+CD169CD11c+ (CD169 IMs). Upon assessment of protein surface expression, we found that the NAMs expressed high levels of MHCII and low levels of Lyve1 (table.S1), whereas the CD169 IMs express lower levels of MHCII and higher levels of Lyve1. We also observed that the NAMs (Lyve1loMHCIIhi) were the subset of IMs that preferentially interacted with pulmonary sympathetic nerves. We also compared our scRNA-seq data with a recently published study by Cohen et al. (61) where they performed an unbiased extensive scRNA-seq of the entire lung. NAMs expressed considerably higher levels of Cx3cr1, C1q genes, C3ar1, and F13a1 in comparison with other macrophage populations. Whereas genes such as Itgax (CD11c) or Ear1 are not expressed in NAMs but were highly expressed in AMs. Thus, NAMs were very similar to the population of macrophages they identified as macrophage I. In agreement with our results, Cohen et al. (61) postulated that the gene signature of the macrophage I subset of macrophages suggested an embryonically derived population that emerges early in pregnancy and not by fetal liver or other monocyte precursors.

Although our results were similar with respect to the localization, gene signature, and ontogeny of IM subsets in the lung, Chakarov et al. (21) observed that their IM populations were replaced quickly over time in adulthood by circulating monocytes. In our study, timed pregnancy experiments revealed that most of the NAMs were yolk sac derived and were not replaced by circulating monocytes at least within a few months of birth. In accordance with this finding, we failed to notice any reduction in NAM frequency or numbers in CCR2 knockout mice by imaging or flow cytometry studies. In addition, our studies with the Pdzk1ip1-CreER mice also show that only a small fraction of NAMs are replaced by BM HSC precursors after TAM treatment. In contrast, a much larger percentage of AMs were replaced by HSC precursors. Furthermore, several recent studies in other tissues including intestines and skin have shown that macrophage subsets associated with nerves also exhibit self-renewing capacity (62, 63).

These differences between the recent studies are likely due to the complex nature of the lung tissue. Accordingly, further studies will be required to begin to parse out how the different subsets of macrophages develop and function in the lung. In addition, we know little about how the external and internal stimuli including microbiome, genetics, inflammatory environment, and xenobiotics affect the ontogeny, lineage commitment, and maintenance of macrophage subsets in vivo.

Besides eradicating pathogens (18), macrophages also play critical roles in regulating inflammation in mucosal organs (28, 64). Serious complications due to inflammatory disorders or respiratory infections result from severe lung inflammation and tissue damage caused by excessive innate and adaptive immune responses (6470). Thus, there is a critical need to understand the cellular and molecular mechanisms that regulate inflammation and pathology in vivo. Such knowledge will be important for developing new therapeutic strategies to control and prevent deleterious immunopathology in the lung. Results from our study suggest that NAMs that express a strong regulatory gene profile play an important role in regulating inflammation in the lungs in vivo and may serve as a target for therapy for inflammatory diseases. NAMs exhibited a robust proliferative response to poly(I:C) treatment and remained localized to the large airways. Furthermore, selective depletion of NAMs during poly(I:C) treatment resulted in increased inflammatory cytokine production and infiltration of innate immune cells, demonstrating that these cells are critical for regulating many aspects of the immune response and tissue homeostasis following inflammatory stimuli in the lung.

To investigate further the importance of AMs and NAMs during respiratory infections, we used three different mouse models in which we depleted AMs and NAMs (CD169-DTR), AMs alone (AM-DTR), or NAMs alone (NAM-DTR). Influenza virus titers were significantly higher in CD169-DTR and AM-DTR mice when compared with both WT and NAM-DTR mice. Our imaging studies showed that AMs harbored GFP-expressing virus unlike NAMs. Thus, these studies demonstrated that AMs are vital for viral clearance in vivo, and NAMs are important for regulating the virus-induced inflammation. In addition, this explains why CD169-DTR mice failed to control the influenza infection and died compared with NAM-DTR mice that survived longer.

Our results provide essential insights into the ontology, development, maintenance, and molecular functions of a previously undescribed subset of resident macrophages in the lung. NAMs regulate inflammation in the lung both during steady state and after induction of inflammation. Considering the putative importance of these macrophages and given that we know little about these cells, our study will open new avenues of research and therapy with respect to respiratory infections and other inflammatory lung diseases. Furthermore, our results will prompt a reevaluation of our understanding of the role various macrophage subsets play in the lung during inflammatory conditions.


Study design

The goal of this study was to provide insights into a distinct subset of nerve- and airway-associated pulmonary macrophages that function to maintain immune and tissue homeostasis. We used a combination of genomic, in situ microscopy, flow cytometric, and in vivo fate- mapping and functional assays to study these macrophages in mice and in human lung tissue samples. The number of replicates for each experiment is indicated in the figure legends.


All the mouse strains used were on a C57BL/6 background. A full list is provided in table S5. To generate CD169-tdTomato mice, CD169-cre mice were crossed to Rosa26-tdTomato mice. To generate the NAM-DTR mice, CD169-cre mice were crossed to CX3CR1-flox–DTR. To generate the AM-DTR mice, CD169-cre mice were crossed to CD11c-flox–DTR. To generate CX3CR1CRE/ERT2/tdTomato embryonic fate-mapping mice, CX3CR1CRE/ERT2 were crossed to Rosa26-tdTomato. In all experiments, adult female mice were used with some exceptions (e.g., in Fig. 4A, mice from both genders were used for embryogenesis experiments). All mice were housed with food and water ad libitum under a 12-hour dark/light cycle in a specific pathogen–free facility at UConn Health and New York University Langone Health. All animal experimental procedures were performed in accordance with protocol reviewed and approved by the Institutional Animal Care and Use Committee at UConn Health and New York University and in accordance with guidelines from the National Institutes of Health, the Animal Welfare Act, and the U.S. Federal Law.

Lung tissues

R. Tigne (Duke University) provided deidentified human lung tissue samples. E. Richard Stanley (Albert Einstein School of Medicine) provided CSF1R−/− mouse lung tissues. B. Reizis provided TAM induced Pdzk1ip1-tdTomato lung tissues (54).

Animal procedures

HSC studies

Transgenic Pdzk1ip1-CreERT (Map17-CreERT) mice (54) were bred to Rosa26 stop-floxed tdTomato mice (the Jackson Laboratory). Pdzk1ip1-CreER-tdTomato mice at 8 to 10 weeks of age (for confocal imaging) and 3 to 5 months of age (for flow cytometry) were treated once with 1 mg of TAM in sunflower oil by oral gavage. The expression of tdTomato fluorescence was measured by BM aspirate 3 days after TAM treatment to confirm recombination of the Rosa26 locus in HSCs. HSC contribution to the production of mature hematopoietic lineages was assessed monthly by flow cytometry analysis of the peripheral blood. Animals were euthanized 6 months after TAM treatment, and tissues were harvested for confocal microscopy and flow cytometry analysis.

CSF1R blocking studies

To block CSF1R, 2 mg of anti-CSF1R (clone AFS98, BioXCell) or immunoglobulin G2a control was administered intraperitoneally to C57BL/6 WT mice on days −2 and −1 and harvested on day 0.

Depletion studies

DT (Sigma-Aldrich) was prepared in distilled water and administered intraperitoneally (40 ng/g body weight) or intratracheally (50 ng per mice for CD169-DTR mice and 100 ng per mice for NAM-DTR mice) for the depletion studies. Mice were then rested for 48 hours before additional analyses described below were performed.

Timed pregnancy in utero labeling

Cx3cr1-CreER mice that express cre recombinase estrogen receptor fusion protein under the control of CX3CR1 promoter were crossed with tdTomato stop-floxed reporter strain. After TAM treatment (42), nuclear translocation occurs to allow tdTomato fluorescent protein expression in the cells that express the cre recombinase [Cx3cr1CreER mice (B6.129P2 (Cg)-Cx3cr1tm2.1 (cre/ERT2) Litt/WganJ ) crossed to tdTomato stop-floxed mice (B6; 129S6-Gt(ROSA) 26Sor tm9 (CAG-tdTomato) Hze/J); referred to as Cx3cr1CRE/ERT2/Rosa26-tdTomato]. To label the yolk sac–derived macrophages, (Z)-4-hydroxytamoxifen (75 μg/g; OH-TAM, Sigma-Aldrich) was administered intraperitoneally to CX3CR1CRE/ERT2/Rosa26-tdTomato pregnant females at E8.5. Pups were removed at E19.5 via C-section and fostered.

Poly(I:C) stimulation

Poly(I:C) (Invivogen) was prepared in sterile endotoxin-free physiological water, and 50 μg was administered intranasally for one to three consecutive days as indicated in the figures.

Influenza infection

L. Cauley provided H1N1 (PR8) virus. H1N1 (PR8) virus stock was also purchased from Charles River Laboratories. Mice were infected with 150 to 750 EID50 (egg infectivity dose 50%) of PR8 virus through an intranasal route and analyzed in various time points as indicated in the main text. The NS1-GFP virus was provided by A. Garcia-Sastre. Mice were infected with 103 plaque-forming units and analyzed in various time points as indicated in the main text.


Sex- and age-matched congenic CD45.1 and CD45.2 C57BL/6 J mice were parabiosed as described in (71). Mice were rested for 5 or 12 weeks before analysis.

BM transplantation

DT (40 ng/g body weight) was administered once for cell depletion to CD169-DTR mice. Two days later, mice were lethally irradiated with 1000 gray. Two hours after irradiation, animals were reconstituted with 2.5 × 106 WT or CCR2−/− BM cells through intravenous injection. Mice were rested for 8 to 10 weeks before analysis.

Focus formation unit assay

Tissue homogenization

Lungs were harvested at designated time points and washed in 1× phosphate-buffered saline (PBS) before placement in 2-ml tubes containing 500 μl of sterile 1× PBS and one 5-mm stainless steel bead (QIAGEN). Lung tissue was homogenized with the Retsch MM400 mixer at 30 Hz for 2 min. After homogenization, lung debris was centrifuged at 15,000 rpm for 10 min. Supernatants were transferred to new tubes.

Focus formation unit assay

Virus titers were determined on MDCK (Madin-Darby canine kidney) cells using a focus formation (FOCI) assay. MDCK cells (6 × 10 4 cells/cm2) were seeded in 12-well culture plates and incubated at 37°C in 5% CO2 for 24 hours. Cells were washed twice with 1× Dulbecco's PBS (DPBS) before adding 200 μl of a serial 10-fold dilution of mouse lung extracts [dilution prepared in 1× DPBS and 0.3% bovine serum albumin (BSA)]. After 1 hour of incubation at 37°C, the inoculates were removed from the cells by washing twice with 1× DPBS. Overlay medium [1 ml per well; Dulbecco’s modified Eagle’s medium; 0.7% Oxoid purified agar, 0.1% DEAE–dextran hydrochloride, 0.03% BSA, 0.01% fetal bovine serum (FBS), 1% penicillin/streptomycin, and TPCK-trypsin (1 μg/ml)] was added to the cells, followed by incubation at 37°C in 5% CO2 for 72 hours.

Subsequently, the cells were fixed with 4% formalin in 1× PBS (500 μl per well) for 30 min at room temperature. After washing twice with 1× PBS, the cells were permeabilized using 1× PBS containing 0.1% Triton X-100 for 5 min at room temperature. The cells were washed twice with 1× PBS; blocked with 1× PBS, 0.2% fish gelatin, and 0.2% skim milk powder; and washed again twice with 1× PBS. The primary antibody (anti-influenza A nucleocapsid) and the secondary antibody (goat anti-mouse horseradish peroxidase) were diluted 1:1000 in 1× PBS with 0.2% fish gelatin. Primary antibody (200 μl per well) was added and incubated with rocking overnight at 4°C. The cells were washed twice with 1× PBS and incubated with 200 μl of secondary antibody for 1 hour with rocking at room temperature. After washing again twice with 1× PBS, cells were stained by incubating with 200 μl of True Blue peroxidase substrate and incubated until blue spots from infected cells appeared (about 15 min). Last, cells were washed twice with ddH2O and air-dried. Foci were counted, and viral titers were calculated as focus formation units (FFU) per milliliter.

Preparation of cell suspensions, flow cytometry, and cell sorting

Single-cell suspensions of spleen and lung tissues were prepared for flow cytometry. Lung tissues were injected and incubated with RPMI 1640 (Lonza) media containing collagenase IV (1.8 mg/ml; Gibco) or Liberase TM (2 mg/ml; Sigma-Aldrich), 10% FBS (Gibco), 0.2% CaCl2, 0.2% MgCl2, deoxyribonuclease I (Roche), and 1% HGPG [1 mM Hepes, 5 mM l-glutamine, penicillin/streptomycin (10,000 U/ml), and gentamicin (5 μg/ml) (pH 7.5)] for 30 min at 37°C. Collagenase digestion was inactivated by addition of RPMI 1640 media containing 1 mM EDTA and 10% FBS. Lung tissue was dissociated into single-cell suspensions, and tris-buffered ammonium chloride [Trizma HCl (20.6 g/liter; Sigma-Aldrich) and NH4Cl (8.3 g/liter; Thermo Fisher Scientific)] was used to lyse red blood cells. Cell suspensions were enriched for CD3 and B220 cells by using an EasySep mouse biotin positive selection kit protocol (STEMCELL Technologies). Splenocytes were prepared similarly without the enrichment protocol. Spleen and lung cells were resuspended in fluorescence-activated cell sorting (FACS) buffer [1× PBS (Gibco), 5% FBS, and 0.5% sodium azide]. Fc receptors were blocked with anti-CD16/32 Fc block antibody (BioLegend) and stained with the indicated antibodies in the above table for 20 min in 4°C. Cells were fixed with 2% paraformaldehyde (PFA) for 20 min in 4°C and resuspended in FACS buffer. Cell suspension was processed on Becton Dickinson LSRII or Bio-Rad ZE5 Yeti instrument, and data were analyzed using FlowJo software. For cell sorting, lung tissue was treated in the same way without fixation. After staining, cell pellet was directly resuspended in 1× PBS with 1 mM EDTA and 2% FBS. Cells were sorted on Becton Dickinson FACSAria II instrument.

Tissue preparation for immunofluorescence and confocal imaging

In the embryogenesis studies, pregnant mice were euthanized by cervical dislocation, and embryos were removed. All the other mice were euthanized by CO2 exposure. Lung tissues were dissected from the embryos and adult mice and fixed in PFA, lysine, and periodate buffer [0.05 M phosphate buffer, 0.1 M l-lysine (pH 7.4), NaIO4 (2 mg/ml), and PFA] overnight at 4°C. The next day, tissues were dehydrated in 30% sucrose overnight at 4°C and subsequently embedded in optimal cutting temperature compound (OCT) media. Frozen tissue sections were sectioned using Leica CM1850/Leica 3050S at a thickness of 20 μm. Fc receptors were blocked with anti-CD16/32 Fc block antibody (BioLegend) diluted in PBS containing 2% serum and 2% FBS for 1 hour at room temperature. Sections were stained with the indicated antibodies listed in table S5 that were diluted in PBS containing 2% goat serum and 2% FBS for 1 hour at room temperature. For the intracellular staining, all the antibodies including the Fc block were diluted in PBS containing 2% serum, 2% FCS, 0.05% Tween-20, and 0.3% Triton X-100. Sections were washed with the same buffer. Images were acquired using a Zeiss LSM 780 FCS/NLO and Zeiss LSM 880 confocal microscope (Carl Zeiss). Image analysis was conducted on acquired images as follows: two images per section, two sections per slide, and two slides per animal of n = 3 animals. The imaging data were processed and analyzed using Imaris software version 8.3.1 (Bitplane, Oxford Instruments). Where noted, colocalization, isosurface, and spot detection functions were used in the Imaris software.


Tissue clearing by the CLARITY method followed standard protocols (72, 73) Briefly, for passive CLARITY, mice were transcardially perfused with 20 ml of ice-cold 1× PBS followed by 20 ml of ice-cold 4% PFA. Perfused lungs were postfixed in 4% PFA overnight in 4°C, followed by overnight incubation in acrylamide-thermal initiator cocktail at 4°C. Next, acrylamide-thermal initiator cocktail was degassed with nitrogen gas and incubated at 37°C for hydrogel polymerization. Upon hydrogel formation, tissues were washed three times in 1× PBS, followed by incubation in 8% SDS solution at 37°C for 24 hours and 4% SDS solution on the subsequent days until tissue became transparent. Subsequently, tissues were washed three times in 1× PBS and incubated in refractive index matching solution (RI = 1.46) overnight before imaging. Passive CLARITY processed whole lung lobes were imaged using the Zeiss Light Sheet Z.1 microscope, and the image was acquired using the Zen for Light Sheet Z.1 software from Carl Zeiss.

RNA extraction and RNA-seq analysis

Lung cell suspensions from 20 naïve C57BL/6 mice were pooled for the RNA-seq analysis. Cells were obtained and sorted as indicated above. The RNAeasy Plus Micro Kit (QIAGEN, Valencia, CA) was used to extract total RNA from the sorted cells. RNA-seq libraries were prepared using the Illumina TruSeq Stranded mRNA Library Kit, and the individual libraries were pooled together. The samples were run over four lanes on the Illumina HiSeq2500 with 101 paired-end reads. Trimmomatic v0.32 was used to remove adapter sequences and bases with base quality lower than three from the ends. Also, using sliding window methods, bases of reads that did not qualify for a window size of 4 and a mean quality of 15 were trimmed, and reads with length shorter than 36 base pairs were dropped. The quality of the data was assessed before and after trimming using FastQC v0.10.0. Trimmed reads were mapped to reference genome (gencode.mm10) with TopHat v.2.0.13 and bowtie2 v.2.2.3, and –G option on Cufflinks v.2.2.1 was used for transcript assembly and fragments per kilobase of transcript per million (FPKM) quantification. Data were normalized using log2 transformation with 1 added to the raw signal (FPKM + 1) followed by quantile normalization using the preprocessCore package in R. The libraries were prepared and pooled at the Center for Genome Innovation at UConn Storrs, CT. The samples were run on the Illumina HiSeq2500 at Macrogen. Preliminary data analysis and quality checks were performed by Macrogen.

RNA-seq quality control

A quality control check was performed for each gene: For the six samples, if more than one FPKM value of a given gene was 0, the gene was not included in the analysis. Therefore, from the total of 47,443 genes, 31,726 were excluded and only 15,717 genes were used for statistical analysis. To ensure reproducibility between biological replicates, Pearson correlation coefficients were calculated on the log2(FPKM + 1) values where coefficient values for replicates ranged from 0.97 to 0.99. Using each sample’s log2(FPKM + 1) value, two-dimensional PCA and hierarchical cluster analysis on the 15,717 expressed genes showed clear separation between naïve CD169+ MZMs, NAMs, and AMs under steady-state conditions.

Analysis of DEGs

For differential expression analysis between NAMs and AMs at steady state, statistical analysis was performed using FC and independent t test per comparison pair. The up-regulated genes were filtered using FC ≥ 2 and FDR < 0.05. These DEGs are graphically depicted in a heat map using z-scores. Hierarchical clustering distances were calculated using the Euclidean method, and clustering was performed using the Ward.D2 method. The top 50 up-regulated genes that had an FC ≥ 70 are graphically depicted in heat maps using z-scores. For the identified up- and down-regulated DEGs in each comparison group, we performed a GO enrichment analysis using geneXplain software platform (, v. 2017.1) and its Proteome database. For all the analyses, we only considered GO terms with P ≤ 10−4 to be significant.

Single-cell RNA-seq

Single-cell suspensions from naïve lung tissues were processed with Liberase TM as indicated previously. For single-cell sorting, after surface staining, the cell pellet was directly resuspended in 1× PBS with 1 mM EDTA and 2% FBS. Cells were sorted on Sony biotechnology SY3200 sorter instrument. Cells were gated for single, live CD45+ cells. Then, MerTK+CD64+ cells were sorted for the scRNA-seq analysis. The sorted cellular suspensions were loaded on a 10x Genomics Chromium instrument to generate single-cell gel beads in emulsion. Cells (10,500) were loaded per channel with an expected yield of 6000 cells. scRNA-seq libraries were prepared using the Single Cell 3’ Reagent Kits v2 (Chromium Single Cell 3’ Library and Gel Bead Kit v2, PN-120237), Single Cell 3’ Chip Kit v2 PN-120236 and i7 Multiplex Kit PN-120262″ (10x Genomics) (74), and the Single Cell 3’ Reagent Kits v2 user guide (manual part no. CG00052 Rev A). Libraries were run on a NovaSeq 6000.

Alignment, barcode assignment, and UMI counting

The Cell Ranger Single Cell Software Suite (version 1.3) was used to perform sample demultiplexing, barcode ad unique molecular identifier (UMI) processing, and single-cell 3′ gene counting.

scRNA-seq analysis

The single-cell count matrix was exported from 10x Genomics cell ranger version 3.0.0 output and subsequently analyzed in R version 3.4.1 Single Candle. Clustering and differential expression analysis were performed with the standard functions from Seurat version 2.3.4. Expression markers for each cluster were considered significant with an FDR of 5% and log2 FC as specified in each figure.

Cytokine and chemokine analysis

BAL was obtained by flushing the airway with 4 ml of saline. Cells and the supernatant were separated by centrifugation. The supernatant was collected in centrifugal filter tubes (Millipore) and concentrated by additional centrifugation steps. Lung lysate and BAL supernatant were measured through 23-plex array analysis (Bio-Rad) for cytokine and chemokine protein levels.

Statistical analysis

GraphPad PRISM 5-8 was used for the statistical analyzes. The analyses were conducted using paired or unpaired two-tailed Student’s t test (*P < 0.05, **P < 0.01, ***P < 0.001, and N.S. for not significant). Where noted, one-way and two-way analysis of variance (ANOVA) with Bonferroni’s post hoc test was used to analyze more than two groups.


Fig. S1. CD169+ macrophages reside in mouse and human lung tissue.

Fig. S2. NAMs are associated with TH-expressing neurons.

Fig. S3. Gating and sorting strategy for NAMs and AMs.

Fig. S4. NAMs originate from the yolk sac require the CSF1-CSF1R axis, and CCR2+ monocytes are not required for the replenishment of NAMs.

Fig. S5. Efficiency of cell deletion in NAM-DTR and CD169-DTR mice.

Fig. S6. NAMs proliferate following poly(I:C) treatment and influenza infection.

Table S1. Major markers and expression profile for AMS, NAMs, and CD169 IMs.

Table S2. Top 5 GO-enriched biological processes terms on up-regulated genes in NAM versus AM.

Table S3. Highly enriched macrophage activation, macrophage chemotaxis, and macrophage differentiation GO terms in up-regulated genes in NAM versus AM.

Table S4. Highly enriched GO terms in down-regulated genes in NAM versus AM.

Table S5. List of mouse lines, reagents, and software used.

Table S6. Raw data file (in Excel spreadsheet).


Acknowledgments: We would like to thank The NYU Langone Genome Technology center and the NYU Langone Microscopy Laboratory, which is partially supported by the Cancer Center Support Grant P30CA016087 at the Laura and Isaac Perlmutter Cancer Center for scRNA-seq help, R. Tigne (Duke University) for providing human lung samples and E. Richard Stanley (Albert Einstein School of Medicine) for providing tissues from CSF1R−/− mice. We would also like to thank X. Pham (UConn Health), J. Rutkowski (NYU Langone Health), and E. Jellison (UConn Health Flow Cytometry Facility) for assistance in performing experiments; the Center for Cell Analysis and Modeling at UCONN Health and the Microscopy Laboratory at NYU Langone Health for help with confocal imaging; and L. Cauley at UConn Health for providing influenza virus. Funding: This work was supported by NIH grants AI143861 and AI097375 to K.M.K. and AG049074 to B.R. Author contributions: K.M.K. directed the study. B.B.U., S.T.Y., and K.M.K. designed the study. B.B.U., S.T.Y., P.D.-Y., M.d.V, O.A.P., Q.P., and L.M. performed experiments. C.M.S., G.J., B.R., P.L., and M.D. provided reagents and analytical tools. J.C.D., P.V.-L., and T.S. provided help with bioinformatics. B.B.U., S.T.Y., and K.M.K. analyzed the data and wrote the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The RNA-seq data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus database under accession number GSE146683. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Immunology

Navigate This Article