Sensing of metabolic signals via GPR183 promotes occupation of lung macrophage niches by monocytes

The lung is continuously exposed to the outside environment and therefore a common target of infection and inflammation. Macrophages are the most abundant immune cell type in the lungs and are strategically positioned in the interstitium and alveoli, where the gas exchange takes place (Aegerter et al., 2022; Bain and MacDonald, 2022; Evren et al., 2020; Kulikauskaite and Wack, 2020). Mouse studies have shown that tissue macrophages either derive from embryonic progenitors or from hematopoietic stem cell–derived monocytes. In steady state, tissue-resident mouse alveolar macrophages develop from fetal monocytes, acquiring their typical identity under the influence of the instructive cytokines GM-CSF (Gschwend et al., 2021; Guilliams et al., 2013; Shibata et al., 2001) and TGFβ (Yu et al., 2017) and transcription factors, such as PPARγ (Schneider et al., 2014), BHLHE40/BHLHE41 (Rauschmeier et al., 2019), and EGR2 (McCowan et al., 2021). In homeostatic conditions, tissue-resident macrophages are maintained in the alveolar niche mostly by local self-renewal independent of input from bone marrow monocytes (Guilliams et al., 2013; Hashimoto et al., 2013; Soucie et al., 2016; Subramanian et al., 2022; Yona et al., 2013). Tracking macrophage origin in humans undergoing lung transplantation suggests that human alveolar macrophages are replenished to a greater degree by blood monocytes than mouse alveolar macrophages (Byrne et al., 2020). Macrophages also reside in the interstitial tissue space of the lung, with distinct subsets of interstitial macrophages differing in anatomical location, lifespan, replenishment, and function (Chakarov et al., 2019; Dick et al., 2022; Gibbings et al., 2017; Li et al., 2024; Schyns et al., 2019; Ural et al., 2020). Interstitial lung macrophages require the cytokines macrophage-CSF (M-CSF) and TGFβ as well as MAF transcription factors (Peng et al., 2025; Vanneste et al., 2023). Compared with their counterparts in the alveoli, interstitial lung macrophages are more dependent on bone marrow–derived monocytes in the steady state (Chakarov et al., 2019; Dick et al., 2022).

Spontaneous replenishment of tissue-resident macrophages as well as perturbed organ homeostasis triggers the recruitment of monocytes from the circulation into tissues. In mice and humans, CCR2-dependent Ly6C^hi and CD14⁺ classical monocytes, respectively, can differentiate into monocyte-derived macrophages in the lung (Aegerter et al., 2022; Evren et al., 2021; Neehus et al., 2024; T’Jonck and Bain, 2023; Trzebanski et al., 2024) and other organs (Guilliams et al., 2018). An important driver of this differentiation is the state of the macrophage niche, made up of stromal cells like fibroblasts, endothelial cells, and epithelial cells, that together provide the signals that determine monocyte fate (Bonnardel et al., 2019; Guilliams et al., 2020). In steady-state conditions, and with minimal derangement of the niche, monocytes replacing tissue-resident macrophages as part of a normal replenishment process acquire a phenotype and function that is hardly distinguishable from the original tissue-resident macrophages that derive from embryonic progenitors (Gibbings et al., 2015; van de Laar et al., 2016). However, when organ homeostasis is perturbed, a monocyte differentiating in an inflammatory niche might become a pro-fibrotic or pro-inflammatory macrophage with functions that are distinct from steady-state tissue-resident macrophages in that niche (Iliakis et al., 2025; T’Jonck and Bain, 2023). In support of this scenario, tissue-resident lung macrophages are replaced by infiltrating blood monocytes in severe respiratory infections or in chronic inflammatory lung diseases like asthma and chronic obstructive pulmonary disease, where they can determine the outcome of lung injury on a spectrum from complete resolution after injury, over altered immune responsiveness to debilitating remodeling and fibrosis (Aegerter et al., 2020; Li et al., 2022a; Machiels et al., 2017; Mirchandani et al., 2022; Misharin et al., 2017; Ruscitti et al., 2024; Szabo et al., 2021; Theobald et al., 2024; Wendisch et al., 2021).

The molecular signals controlling the development and function of monocyte-derived macrophages in the lung niche in homeostasis and disease remain poorly understood. Furthermore, the local cues that guide monocytes to specific macrophage niches in the lung are largely undefined. G protein–coupled receptors (GPCRs) are an important receptor family that allows cells to respond to diverse signals from their surrounding tissue microenvironment. One such GPCR expressed in immune cells is GPR183 (also known as EBI2), a cell surface receptor for hydroxylated cholesterol metabolites, called oxysterols (Cyster et al., 2014; Dang and Reboldi, 2024). The GPR183 ligand 7α,25-dihydroxycholesterol is generated from cholesterol by the enzymes cholesterol 25-hydroxylase (CH25H) and 7α-hydroxylase (CYP7B1). The enzyme hydroxy-δ-5-steroid dehydrogenase, 3 β- and steroid δ-isomerase 7 (HSD3B7) degrades 7α,25-dihydroxycholesterol into bile acid precursors. Oxysterols recognized by GPR183 control the migration and positioning of immune cells within lymphoid (Baptista et al., 2019; Gatto et al., 2009, 2013; Li et al., 2016; Pereira et al., 2009; Yi and Cyster, 2013) and nonlymphoid organs (Bohrer et al., 2022; Emgård et al., 2018) as well as immune responses.

Here, using reporter and gene-deficient mice, we uncover GPR183 as a receptor for metabolic signals that determine monocyte localization within the lung and monocyte differentiation into macrophages after experimental niche depletion and lung injury. Macrophage development from monocytes required a non-hematopoietic source of GPR183 ligand with fibroblasts as the main producers of 7α,25-dihydroxycholesterol. GPR183 promoted the interaction of recruited monocytes with lung fibroblasts and their differentiation into alveolar and interstitial macrophages. Monocytes accumulating in the chronically inflamed airways of humans with bronchiectasis, a neglected but important lung disease, also expressed the oxysterol receptor GPR183. Our results provide a mechanism of how the sensing of cholesterol metabolites after macrophage depletion links niche signals to macrophage differentiation from monocytes.

We hypothesized that monocytes use GPR183 to sense metabolic signals in the lung. To explore this hypothesis, we first investigated the expression of GPR183 in lung monocytes and macrophages. Analysis of Gpr183^GFP/+ reporter mice demonstrated that Ly6C^hi and Ly6C^lo lung monocytes expressed Gpr183 mRNA (Fig. 1 A and Fig. S1 A). Lung dendritic cells and eosinophils also expressed Gpr183-GFP, whereas neutrophils were largely GFP^- (Fig. 1 A). B lymphocytes highly expressed Gpr183 mRNA, whereas CD4 and CD8 T cells showed lower expression (Fig. 1 B). Gpr183 was required for monocyte migration toward 7α,25-dihydroxycholesterol in vitro (Fig. 1 C), consistent with functional GPR183 expression on the cell surface. These results identify the GPR183 ligand 7α,25-dihdroxycholesterol as a chemoattractant for monocytes. Like monocytes, both CD206⁺ and CD206^- interstitial lung macrophages transcribed Gpr183 mRNA (Fig. 2, A and B; and Fig. S1 A). In contrast, Gpr183 mRNA was not expressed by resident alveolar macrophages (Fig. 2 A and Fig. S1 A) that are derived from fetal monocytes (Guilliams et al., 2013; Schneider et al., 2014). The number of Ly6C^hi lung monocytes, interstitial lung macrophages, and alveolar macrophages was not different in Gpr183^+/+ and Gpr183^−/− mice (Fig. 2 C). Therefore, GPR183 was not required for the homeostasis of resident macrophages in the steady-state lung.

We next asked whether alveolar macrophages derived from bone marrow monocytes expressed Gpr183. For this purpose, we generated bone marrow chimeras, where irradiated B6 recipient mice (CD45.1⁺) were transplanted with bone marrow from Gpr183^GFP reporter mice (CD45.2⁺) (Fig. 2 D). Whole-body irradiation leads to the replacement of resident lung macrophages (CD45.1⁺) by macrophages derived from donor monocytes (CD45.2⁺). Flow cytometry analysis revealed that CD45.2⁺ monocyte–derived lung macrophage populations, including transitional CD11c⁺SiglecF^mid alveolar macrophages, expressed Gpr183 at 3 wk after bone marrow transfer (Fig. 2 E and Fig. S1 B). In contrast, monocyte-derived alveolar macrophages with a mature CD11c⁺SiglecF^hi surface phenotype expressed lower amounts of Gpr183 mRNA (Fig. 2 E). Low Gpr183 mRNA expression was maintained by CD45.2⁺ monocyte–derived alveolar macrophages at 5 wk after bone marrow reconstitution and eventually downregulated at 12 wk (Fig. 2 F). Gpr183 downregulation was also observed in vitro when bone marrow cells were differentiated into macrophages (Fig. 2 G). Like their counterparts in the lung, Ly6C^hi bone marrow monocytes expressed Gpr183 mRNA, and Gpr183 expression was maintained in M-CSF–induced “generic” macrophages (Fig. 2 H). In contrast, alveolar macrophage–like cells generated in the presence of GM-CSF and TGFβ (Luo et al., 2021) had lower Gpr183 expression, especially the CD11b^lo subset (Fig. 2 H) that resembles resident alveolar macrophages in vivo. Taken together, these results show that lung macrophages in the interstitial niche express Gpr183, whereas monocyte-derived macrophages downregulate Gpr183 expression when becoming resident in the alveolar niche.

We next examined GPR183 expression during repopulation of the empty alveolar niche. For this purpose, we used a new mouse model, called “SPAM deleter mice” (Fig. 3 A), that allows specific depletion of alveolar macrophages in an inducible manner (Verwaerde et al., 2025). Accordingly, intratracheal diphtheria toxin (DT) administration induced the rapid depletion of resident CD11c⁺SiglecF⁺ alveolar macrophages that were replaced within 2 wk (Fig. 3, B and C). Alveolar macrophage depletion increased the cytokines GM-CSF and M-CSF and the chemokine CCL2 in the bronchoalveolar lavage (BAL) fluid of SPAM deleter mice within 1 day (Fig. S2, A and B). Furthermore, DT-induced alveolar macrophage depletion caused neutrophil influx into the alveolar space (Fig. 3 C) and an increase in TNFα protein, whereas amounts of IL-1α and IL-β protein were unchanged in the BAL fluid (Fig. S2, A and B).

To track the trajectory of lung monocytes during macrophage differentiation, we performed single-cell RNA sequencing of lung cells from SPAM deleter mice before and 0.5, 1, 2, 5, 8, and 14 days after DT-mediated alveolar macrophage depletion (Fig. 3 D). Uniform Manifold Approximation and Projection (UMAP) showed several clusters of non-hematopoietic cells as well as hematopoietic cells that were protected from intravascular anti-CD45 antibody labeling and therefore resident in the lung (Fig. 3 E and Fig. S2 C). Zooming in on monocyte and macrophage clusters revealed the rapid recruitment of monocytes 12 h after alveolar macrophage depletion and their timed progression toward becoming recruited alveolar macrophages (Fig. 3 F and Fig. S2, D and E). Consistent with our results in Gpr183^GFP/+ reporter mice (Fig. 1 A and Fig. 2, A, E, and F), Gpr183 mRNA expression was highest in lung monocytes, interstitial lung macrophages, and differentiating monocyte-derived macrophages in SPAM deleter mice (Fig. 3 G). In contrast, resident alveolar macrophages and mature alveolar macrophages derived from recruited monocytes (on day 14 after depletion) did not express Gpr183 (Fig. 3 G). We conclude that developing alveolar macrophages derived from monocytes express Gpr183 mRNA during irradiation-induced lung injury and after experimental depletion of alveolar macrophages. Once the lung has returned to homeostasis and monocyte-derived alveolar macrophages have become resident, Gpr183 expression is downregulated.

Gpr183 expression in Ly6C^hi monocytes, known precursors of tissue-resident macrophages, led us to ask whether GPR183 was required for the differentiation of injury-induced lung macrophages. Lung injury was induced by whole-body irradiation, which causes replacement of host hematopoiesis by donor bone marrow and depletion of resident lung macrophages. We employed competitive bone marrow chimeras to study lung macrophages derived from Gpr183-deficient and Gpr183-sufficient monocytes within the same microenvironment in vivo. For this purpose, irradiated B6 mice (CD45.1⁺) were reconstituted with a 1:1 mixture of Gpr183^+/+ (CD45.1.2⁺) and Gpr183^−/− (CD45.2⁺) or Gpr183^+/+ (CD45.2⁺) and Gpr183^−/− (CD45.1.2⁺) bone marrow cells. Gpr183^+/+ and Gpr183^−/− Ly6C^hi monocytes were present at similar frequencies in blood and lung 9–13 wk after bone marrow transfer, consistent with the bone marrow input ratio (Fig. 4, A and B; and Fig. S1 C). In contrast, Gpr183^+/+ monocyte-derived alveolar macrophages outcompeted their Gpr183^−/− counterparts at a ratio of ∼2:1 (Fig. 4, A and B; and Fig. S1 C). The skewed Gpr183^+/+ to Gpr183^−/− ratio showed that Gpr183-deficient interstitial lung macrophages also had a defect in repopulating the injured lung (Fig. 4, A and B; and Fig. S1 C). Over time, Gpr183-deficient alveolar macrophages were progressively outcompeted (∼9:1) by their WT counterparts, as shown by flow cytometry 19–23 wk after bone marrow transfer (Fig. 4 C). We also generated single-transfer Gpr183^+/+ (CD45.2⁺) → B6 (CD45.1⁺) and Gpr183^−/− (CD45.2⁺) → B6 (CD45.1) bone marrow chimeras to assess the role of GPR183 in a noncompetitive context. These experiments showed that Gpr183-deficient monocytes were able to replenish an empty lung macrophage niche in the absence of cell–cell competition (Fig. 4 D). Overall, these results demonstrate that the replacement of resident lung macrophages by monocyte-derived macrophages after irradiation-induced lung injury is supported by GPR183 in a cell-intrinsic manner.

We next asked whether oxysterols sensed by the surface receptor GPR183 regulate monocyte-to-macrophage differentiation in organs other than the lung. Analysis of competitive Gpr183^+/+/Gpr183^−/− chimeras showed that macrophage populations in the spleen, peritoneal cavity, and brain originated equally from Gpr183^+/+ and Gpr183^−/− bone marrow monocytes (Fig. 4 E and Fig. S3, A–D). The lower Gpr183^+/+ to Gpr183^−/− ratios indicated that GPR183 may inhibit rather than promote the development of monocyte-derived macrophages in the liver (Fig. 4 E and Fig. S3 E). We conclude that GPR183 predominantly promotes monocyte-to-macrophage differentiation in the lung.

To confirm these findings in models without genotoxic lung injury, we performed mixed bone marrow chimeras in SPAM deleter mice and in IM–DT receptor (DTR) (Tmem119^CreCx3cr1^LSL-DTR) mice that allow the inducible depletion of interstitial lung macrophages by DT (Vanneste et al., 2023). First, SPAM deleter mice (CD45.2⁺) were reconstituted with a 1:1 mixture of Gpr183^+/+ or Gpr183^−/− (CD45.1.2⁺) and B6 (CD45.1⁺) bone marrow cells. SPAM deleter recipients were treated with busulfan to replace host monocytes while preserving host alveolar macrophages (Misharin et al., 2017) before depleting them with DT 7 wk after bone marrow transfer (Fig. 5 A). Analysis at 4 wk after DT administration revealed that macrophages derived from Gpr183^+/+ bone marrow were more efficient than their Gpr183^−/− counterparts at repopulating the alveolar niche (Fig. 5, A and B). In contrast, Gpr183^+/+ and Gpr183^−/− bone marrow cells contributed equally to Ly6C^hi monocytes in the blood and lung (Fig. 5 B). Second, we employed a similar approach to determine whether Gpr183 was required for repopulating the interstitial lung macrophage niche. In thorax-shielded IM-DTR bone marrow chimeras (Fig. 5 C), the ratio of Gpr183^−/− to B6 interstitial lung macrophages was significantly lower than the ratio of Gpr183^+/+ to B6 interstitial lung macrophages 2 wk after DT administration (Fig. 5, C and D). Therefore, Gpr183-deficient monocytes had a reduced ability to fill the vacant interstitial macrophage niche. Ly6C^hi blood monocytes were not impacted by the absence of Gpr183, whereas Gpr183^+/+ bone marrow cells were more efficient at reconstituting Ly6C^hi lung monocytes in IM-DTR mice (Fig. 5 D). We conclude that GPR183 promotes the replenishment of alveolar and interstitial macrophage niches by bone marrow–derived monocytes.

Energy metabolism controls macrophage homeostasis and alveolar macrophages are dependent on oxidative phosphorylation (Wculek et al., 2023). However, we found that GPR183 was dispensable for homeostatic alveolar macrophage metabolism because, like their Gpr183^+/+ counterparts, Gpr183-deficient alveolar macrophages derived from monocytes in competitive Gpr183^+/+/Gpr183^−/− bone marrow chimeras were mainly fueled by oxidative phosphorylation (Fig. S4 A). Finally, we examined the impact of GPR183 on monocyte-to-macrophage differentiation in vitro. Congenically marked Gpr183^+/+ and Gpr183^−/− bone marrow cells were mixed 1:1 and cultured with either M-CSF or GM-CSF and TGFβ₁ in the absence or presence of the GPR183 ligand 7α,25-dihydroxycholesterol (Fig. S4 B). An equal Gpr183^+/+ to Gpr183^−/− ratio of macrophages was maintained in all culture conditions (Fig. S4, B and C). These data demonstrate that genetic deletion of Gpr183 does not cause a general defect in macrophage differentiation per se.

Since B and T cells also expressed Gpr183 (Fig. 1 B), we asked whether lymphocyte-expressed GPR183 contributes to promoting the development of monocyte-derived alveolar macrophages. To address this possibility, irradiated B6 (CD45.1⁺) recipient mice were transplanted with Rag1^−/− (CD45.2⁺) bone marrow mixed with either Gpr183^+/+ or Gpr183^−/− (CD45.1.2⁺) bone marrow (Fig. 6 A). In mixed Gpr183^−/−/Rag1^−/− chimeras, Gpr183 is expressed only in innate immune cells, whereas T and B lymphocytes lack Gpr183. As in competitive Gpr183^+/+/Gpr183^−/− bone marrow chimeras (Fig. 4 B), Gpr183-deficient alveolar macrophages were outcompeted in mixed Gpr183/Rag1^−/− chimeras by their Gpr183-sufficient counterparts (Fig. 6 B). This result indicated that Gpr183 expression in innate immune cells supports monocyte-to-alveolar macrophage differentiation. To address whether GPR183 expression by monocytes promotes their differentiation into alveolar macrophages, we created mixed Gpr183^+/+ (CD45.1.2⁺)/Ccr2^RFP/RFP(CD45.2⁺) and Gpr183^−/− (CD45.1.2⁺)/Ccr2^RFP/RFP(CD45.2⁺) bone marrow chimeras in B6 (CD45.1⁺) recipients (Fig. 6 C). CCR2 is required for monocyte egress from bone marrow into the blood (Serbina and Pamer, 2006). Therefore, lung monocytes in mixed Gpr183^−/−/Ccr2^RFP/RFP chimeras do not express Gpr183, whereas Gpr183 expression in other cells is preserved. The similar ratio of Gpr183^+/+/Ccr2^RFP/RFP and Gpr183^−/−/Ccr2^RFP/RFP alveolar macrophages (Fig. 6 D) suggested that Gpr183 expression in CCR2⁺ monocytes is required to confer a competitive advantage for monocyte-to-alveolar macrophage differentiation.

To further examine the role of GPR183 in lung macrophage differentiation, we analyzed competitive Gpr183^+/+/Gpr183^−/− chimeras at 3 wk after bone marrow reconstitution (Fig. 7 A). The normal Gpr183^+/+ to Gpr183^−/− ratios of Ly6C^hi monocytes in bone marrow and blood demonstrated that GPR183 was dispensable for monopoiesis and monocyte egress into the circulation (Fig. 7 B). During lung macrophage differentiation, monocytes first upregulate CD64 before downregulating Ly6C, which is then followed by the upregulation of CD11c and finally SiglecF surface expression (Li et al., 2022a; Ruscitti et al., 2024; Sabatel et al., 2017). Consistent with this notion, lung tissue of the Gpr183^+/+/Gpr183^−/− chimeras at 3 wk contained CD64⁺CD11b⁺Ly6C⁺SiglecF⁻ and CD64⁺CD11b⁺Ly6C⁻SiglecF⁻ monocyte-macrophages, as well as CD64⁺CD11c⁺SiglecF^lo-mid transitional alveolar macrophages (Fig. 7, A and B; and Fig. S1 B). We found that bone marrow–derived Gpr183^+/+ cells started to outcompete their Gpr183^−/− counterparts from the CD64⁺CD11b⁺Ly6C⁺SiglecF⁻ monocyte-macrophage stage, whereas Ly6C^hiCD64⁻ lung monocytes originated equally from Gpr183^+/+ and Gpr183^−/− bone marrow (Fig. 7, A and B). These data show that lack of GPR183 leads to a defect in early monocyte-to-lung macrophage differentiation.

Next, we determined the cellular source of the GPR183 ligand 7α,25-dihydroxycholesterol in the injured lung. First, we asked whether Ch25h-dependent generation of 7α,25-dihydroxycholesterol promotes reconstitution of the lung macrophage compartment by monocytes. To address this question, irradiated Ch25h^+/+ or Ch25h^−/− (CD45.2⁺) mice were reconstituted with an equal mixture of Gpr183^+/+ or Gpr183^−/− (CD45.1.2⁺) and B6 (CD45.1⁺) bone marrow (Fig. 7 C). In Ch25h-sufficient hosts, alveolar and interstitial lung macrophages derived from Gpr183^+/+ bone marrow outcompeted their Gpr183^−/−-derived counterparts ∼twofold after bone marrow transfer (Fig. 7 C), consistent with our results above (Fig. 4, A and B). In contrast, the competitive disadvantage of Gpr183^−/− lung macrophages was not observed in a Ch25h-deficient microenvironment (Fig. 7 C). These findings demonstrate that a radioresistant source of Ch25h is required for the GPR183-dependent repopulation of macrophage niches after lung injury. To confirm that Ch25h was predominantly expressed by radioresistant cells in the injured lung, we used bone marrow chimeras with either Ch25h^+/+ or Ch25h^−/− bone marrow donors and recipients (Fig. 7 D). We found that Ch25h mRNA expression in the lung was similar in irradiated Ch25h^+/+ hosts injected with either Ch25h^+/+ or Ch25h^−/− bone marrow (Fig. 7 D). In contrast, Ch25h^−/− hosts failed to support normal Ch25h mRNA expression in the lung 3–4 wk after bone marrow injection. These results showed that Ch25h in the injured lung is mainly derived from non-hematopoietic (radioresistant) cells. Taken together, these data show that a non-hematopoietic source of the GPR183 ligand 7α,25-dihydroxycholesterol supports the development of monocyte-derived lung macrophages.

To further define the cellular source of 7α,25-dihydroxycholesterol, we leveraged our single-cell RNA-sequencing data from SPAM deleter mice (Fig. 3, D and E). Given the radioresistant source of 7α,25-dihydroxycholesterol (Fig. 7, C and D), we focused on expression of enzymes regulating 7α,25-dihydroxycholesterol production in non-hematopoietic cell clusters. We reasoned that cells co-expressing Ch25h and Cyp7b1 are the main producers of 7α,25-dihydroxycholesterol. We found that lung fibroblasts expressed both Ch25h and Cyp7b1 with alveolar fibroblasts having the highest expression (Fig. 8 A and Fig. S2 C). Time course analysis showed that expression of Ch25h and especially Cyp7b1 in the lung of SPAM deleter mice was rapidly upregulated after alveolar macrophage depletion (Fig. 8 B). In contrast, the 7α,25-dihydroxycholesterol-degrading enzyme Hsd3b7 was downregulated (Fig. 8 B). These results are consistent with increased production of the GPR183 ligand 7α,25-dihydroxycholesterol after alveolar macrophage depletion. The peak of Cyp7b1 mRNA expression at 48 h after alveolar macrophage depletion (Fig. 8 B) coincided with the time point when recruited monocytes began to differentiate into alveolar macrophages (Fig. 3 B, C, and F; and Fig. S2, D and E). Further analysis demonstrated that fibroblasts upregulated both Ch25h and Cyp7b1 expression within 12 h after alveolar macrophage depletion (Fig. 8, C and D).

We confirmed these findings in the context of monocyte-to-interstitial lung macrophage differentiation using the IM-DTR model (Fig. 8, E and F). Single-cell RNA sequencing of niche cells showed that lung fibroblasts were the main non-hematopoietic cell type co-expressing Ch25h and Cyp7b1 (Fig. 8 G and Fig. S5 A). In contrast, mRNA expression of the GPR183 ligand-degrading enzyme Hsd3b7 was very low in fibroblasts, endothelial cells, and epithelial cells (Fig. 8 G). DT-induced depletion of interstitial lung macrophages triggered a rapid increase in Ch25h and Cyp7b1 expression by fibroblasts with a peak at 12–24 h (Fig. 8 H). These results further support the concept that the GPR183 ligand 7α,25-dihydroxycholesterol is produced by lung fibroblasts in response to resident macrophage depletion.

We next asked whether Ch25h⁺Cyp7b1⁺ fibroblasts also produce known factors that induce monocyte-to-macrophage differentiation. Analysis of the SPAM deleter single-cell RNA-sequencing data showed that Ccl2 mRNA expression increased rapidly after alveolar macrophage depletion (Fig. S5 B). CCR2, the receptor for CCL2, is required for the generation of monocyte-derived alveolar macrophages (Sabatel et al., 2017). Together with interstitial lung macrophages and monocytes, fibroblasts contributed to the peak of Ccl2 expression at 12 h after alveolar macrophage depletion (Fig. S5 B). Expression of Csf1, encoding M-CSF, the critical factor for monocyte-to-lung macrophage differentiation (Aegerter et al., 2022), showed a similar upregulation after alveolar macrophage depletion (Fig. S5 C). Csf1 mRNA expression was highest in neutrophils, but fibroblasts also expressed it (Fig. S5 C). Depletion of interstitial lung macrophages in the IM-DTR model also triggered Csf1 upregulation in fibroblasts (Fig. 8 H). Taken together, our results support the notion that lung fibroblasts provide the GPR183 ligand 7α,25-dihydroxycholesterol as well as chemokines and cytokines to induce the differentiation of GPR183⁺ monocytes into lung macrophages.

The preferential expression of GPR183 ligand-synthesizing enzymes by lung fibroblasts predicted that Ch25h⁺Cyp7b1⁺ fibroblasts attract monocytes via GPR183 to a niche in the lung where monocyte-to-macrophage differentiation occurs. To test this prediction, we took advantage of Gpr183^GFP/+Ch25h^tdTom double reporter mice to visualize Gpr183⁺ monocytes and Ch25h⁺ cells in the steady-state lung and on day 3 after LPS-induced acute lung injury (Fig. 9 A). Immunofluorescence microscopy showed that platelet-derived growth factor receptor α (PDGFRα)⁺ fibroblasts expressed Ch25h-Tomato (Fig. 9, B and C) and were closer to CD68⁺MHCII^- monocytes and CD68⁺MHCII⁺ macrophages that expressed Gpr183-GFP (Fig. 9, B–D), consistent with a preferential interaction between receptor-expressing and ligand-producing cells. The proximity between Gpr183⁺ monocytes or macrophages and Ch25h⁺ fibroblasts further increased after LPS-induced acute lung injury (Fig. 9 D). Overall, these results identify a niche in the lung where GPR183-expressing monocytes and macrophages interact with Ch25h⁺ fibroblasts.

To validate our findings in the human context, we investigated monocyte-derived macrophages in bronchiectasis, a structural airway disease characterized by irreversible dilatation of the bronchi and recurrent lung inflammation (Chalmers et al., 2018). For this purpose, we performed single-cell RNA sequencing of CD45⁺CD66⁻ BAL cells (excluding neutrophils) from six bronchiectasis patients (Table S1) and integrated our single-cell RNA-sequencing data with a dataset from BAL fluid of 10 healthy donors (Mould et al., 2021). UMAP analysis showed 11 transcriptionally distinct clusters, composed of myeloid cells as well as T and B lymphocytes (Fig. 10 A). Cells in cluster 0, 2, 4, 6, and 11 (Fig. 10 A) were identified as alveolar macrophages based on expression of the master transcription factor PPARG (Fig. 10 B) and their high gene signature similarity score (Fig. 10 C). Cluster 1 and 10 showed greater similarity to CD14⁺ lung monocytes (Fig. 10 C) and expressed the monocyte-related transcription factor KLF4 (Fig. 10 B). Cluster 10 was defined as GPR183-monocytes based on differential expression of GPR183 (Fig. 10 B).

To further track macrophage origin in bronchiectasis, we performed gene module score analysis, taking advantage of our previous work that established the gene signatures of human lung macrophages of fetal versus monocytic origin in humanized MISTRG mice (Evren et al., 2022). The gene signatures of macrophages in cluster 0, 2, 4, 6, and 11 were most similar to that of human lung macrophages derived from CD116⁺ fetal precursors (Fig. 10 D). In contrast, cluster 1 and 10 (GPR183-monocytes) had lower similarity to fetal-derived lung macrophages and a higher gene module score for hematopoietic stem and progenitor cell (HSPC)-derived lung macrophages (Fig. 10 D), consistent with a monocytic origin. GPR183-monocytes (cluster 10) and monocytic VCAN-macrophages (cluster 1) as well as T lymphocytes (clusters 3 and 5) were expanded in bronchiectasis compared with healthy airways (Fig. 10 E). Finally, we used the Slingshot algorithm (Street et al., 2018) to further explore developmental relationships between human monocytes and airway macrophages in bronchiectasis. The Slingshot analysis predicted a trajectory from GPR183-monocytes toward the macrophage clusters (Fig. 10 F), consistent with potential differentiation of GPR183-monocytes into human lung macrophages. Overall, these results demonstrate that human monocytes recruited into the airways express GPR183, supporting the notion that GPR183 plays a conserved role in human monocyte-to-lung macrophage development.

Macrophages reside in tissue-specific niches and cues from the local microenvironment shape their identity and function (Amit et al., 2016; Bain and MacDonald, 2022; Bleriot et al., 2020; Guilliams and Svedberg, 2021; Guilliams et al., 2020; Jenkins and Allen, 2021), yet little is known about the signals and receptors that govern occupation of macrophage niches in the lung. Here, we discovered GPR183 as a receptor for metabolic signals in the lung that promote the development of monocyte-derived macrophages. We propose the concept that increased niche “availability” due to resident macrophage depletion triggers GPR183 ligand production and monocyte recruitment as a physiological response to replenish lung macrophages. Accordingly, lung fibroblasts responded to local macrophage depletion by producing GPR183 ligand-synthesizing enzymes as well as other factors, such as CCL2 and M-CSF. Our results support the notion that oxysterol sensing via GPR183 guides monocytes to a niche that favors the interaction of monocytes and developing macrophages with fibroblasts. We postulate that cross talk with fibroblasts within the niche results in monocyte exposure to M-CSF, thereby stimulating the generation of transitional macrophages that differentiate into alveolar and interstitial lung macrophages.

Cell differentiation and function within tissues relies on cell localization in anatomical niches. Chemotactic signals and their receptors guide cells to specific tissue sites to support cell–cell interaction and communication. Previous work demonstrated that GPR183 positions immune cells within lymphoid structures to support adaptive immune responses and lymphoid tissue formation (Baptista et al., 2019; Ceglia et al., 2023; Emgård et al., 2018; Frascoli et al., 2023; Gatto et al., 2009; Li et al., 2016; Pereira et al., 2009). In addition, GPR183 supports immune cell homeostasis by providing access to cytokines (Gatto et al., 2013; Yi and Cyster, 2013). Furthermore, GPR183 promotes inflammatory responses in the intestine (Emgård et al., 2018) and the lung (Jia et al., 2018). GPR183-mediated cell recruitment also contributes to host defense against bacterial infection in the lung (Bartlett et al., 2020; Bohrer et al., 2022; Ngo et al., 2022). Finally, GPR183 has been reported to regulate macrophage infiltration and pro-inflammatory cytokine production during SARS-CoV-2 infection in mice (Foo et al., 2023). Here, we demonstrate that oxysterol sensing via GPR183 allows monocytes to detect tissue-derived signals and occupy macrophage niches in the lung. Our findings, therefore, provide a mechanism of how monocytes sense niche availability in the lung and how they differentiate into distinct types of lung macrophages.

We discovered that GPR183 was differentially expressed in macrophages according to their location within the lung: Interstitial macrophages expressed GPR183, whereas GPR183 was absent in alveolar macrophages in the homeostatic lung. In response to macrophage depletion, alveolar macrophages derived from bone marrow monocytes transiently expressed GPR183 while establishing residence in the alveolar niche. Bone marrow monocyte–derived alveolar macrophages downregulated GPR183 expression once the lung returned to homeostasis. This raises the possibility that factors present in the homeostatic alveolar environment cause GPR183 downregulation in macrophages, whereas the interstitial niche supports GPR183 expression. We speculate that the residence time of macrophages in the alveolar niche and the status of the alveolar niche (homeostatic versus inflamed, full versus empty) may control GPR183 expression. Finally, GPR183 downregulation may be a mechanism to ensure lung homeostasis by preventing resident alveolar macrophages from responding to oxysterols.

The requirement of GPR183 for the generation of monocyte-derived lung macrophages was revealed in a competitive setting. The role of receptors, such as GPR183, in regulating cell localization in vivo is best studied in a competitive context because otherwise compensatory mechanisms may mask the phenotype. In the absence of cell–cell competition, cells lacking a specific chemotactic receptor can eventually access specific sites within tissues through random migration independently of chemotactic gradients. Furthermore, even if only a few receptor-deficient cells reach a specific location, these cells can replenish the niche through increased proliferation as has been shown for Kupffer cell repopulation by monocytes (Bonnardel et al., 2019). Finally, redundant chemoattractant signals and their receptors can disguise the phenotype in cells lacking a single chemotactic receptor. This phenomenon has been shown for GPR183 and CCR6 in the context of lymphoid tissue-inducer cell positioning in the intestine (Howley et al., 2023, Preprint). It is therefore possible that GPR183 operates in synergy with CCR2 to direct monocytes to macrophage niches in the lung.

Lack of GPR183 caused a defect in both alveolar and interstitial lung macrophages derived from bone marrow monocytes. These data indicate that GPR183 acts before monocytes make a fate decision, i.e., before they become anatomically distinct lineages of lung macrophages. Consistent with this notion, the generation of transitional monocyte-derived macrophages was dependent on GPR183. Specifically, GPR183 was required from the stage when monocytes upregulate CD64 and downregulate Ly6C surface expression. These observations suggest that GPR183 regulates the development of CD64⁺Ly6C⁺ monocyte-macrophage intermediates that are dependent on CCR2 and M-CSF, but independent of GM-CSF (Ruscitti et al., 2024; Vanneste et al., 2023). Overall, our findings are consistent with the notion that oxysterol sensing via GPR183 controls the fate of recruited monocytes in the lung.

It is poorly understood how monocytes occupy macrophage niches once they have entered an organ, such as the lung. The lung has a unique anatomy with a narrow interstitial space between the alveoli and the blood vessels to facilitate the gas exchange. Furthermore, the alveoli represent a special macrophage niche as only ∼30% of the alveoli are occupied by macrophages (Neupane et al., 2020). In steady state, lung monocytes reside at the interface of blood vessels and alveoli, either in the perivascular region or within capillaries in the alveolar interstitium (Rodero et al., 2015). In the current study, we found that monocytes migrated toward the GPR183 ligand 7α,25-dihydroxycholesterol in vitro and that recruited GPR183⁺ monocytes in vivo were positioned in the proximity of Ch25h⁺ lung fibroblasts that likely provide 7α,25-dihydroxycholesterol.

Cholesterol is abundant in the lung, and we demonstrate that the oxysterol 7α,25-dihdroxycholesterol functions as a migratory cue for monocytes. We found that a non-hematopoietic source of the GPR183 ligand 7α,25-dihydroxycholesterol supports monocyte-to-lung macrophage differentiation. Furthermore, we show that lung fibroblasts produce the 7α,25-dihydroxycholesterol–synthesizing enzymes CH25H and CYP7B1 in response to the loss of resident lung macrophages. Therefore, fibroblasts are likely the main cellular source of 7α,25-dihydroxycholesterol in the lung, similar to what has been reported in lymphoid organs (Yi et al., 2012) and the intestine (Emgård et al., 2018). In support of this idea, we found that Gpr183⁺ monocytes/macrophages and Ch25h⁺ fibroblasts co-localized in the lung. Lung endothelial cells express CH25H (Madenspacher et al., 2023) but lack CYP7B1. It is therefore possible that CH25H⁺CYP7B1⁺ fibroblasts and CH25H⁺CYP7B1⁻ endothelial cells cooperate to produce 25-hydroxycholesterol that is then further metabolized into the GPR183 ligand 7α,25-dihydroxycholesterol by fibroblast-derived CYP7B1.

Depletion of resident lung macrophages in the SPAM deleter and IM-DTR models was sufficient to increase the GPR183 ligand 7α,25-dihydroxycholesterol as well as M-CSF and CCL2, two factors required for the development of interstitial and alveolar macrophages from monocytes (Joshi et al., 2020; Sabatel et al., 2017; Schlitzer et al., 2013; Vanneste et al., 2023). This raises the possibility that the loss of resident macrophages is sensed by lung fibroblasts, which stimulates fibroblasts to produce factors that replenish the depleted niche. Consistent with this notion, recent work demonstrated that sensing of physical space and cell density by fibroblasts controls macrophage number within the niche through M-CSF production (Zhou et al., 2022). Coordinated upregulation of Ch25h, Cyp7b1, Ccl2, and Csf1 after depletion of resident macrophages supports the concept that fibroblasts act as a hub that not only provides growth factors, such as M-CSF, according to the two-cell circuit proposed by Medzhitov (Zhou et al., 2018), but also a cholesterol metabolite to position monocytes and to support macrophage differentiation.

In conclusion, our study provides mechanistic insights into how metabolic signals recognized by GPR183 direct the migration of monocytes and instruct their differentiation into distinct types of lung macrophages. This knowledge is relevant to the aim of reprogramming macrophages to improve the outcome of lung disease in humans. Targeting GPR183 and cholesterol metabolism could be a new way to limit lung pathology, especially because GPCRs are excellent drug targets.

GPR183 was dispensable for the development of resident macrophages in the steady-state lung. However, it is possible that GPR183 supports the development of embryonic-derived lung macrophages in a competitive setting. In the current study, we used genetic and genotoxic models of macrophage depletion to investigate the role of GPR183 in monocyte-derived lung macrophages. It will be worthwhile extending our findings in future studies using more physiologically relevant contexts, such as after lung infection. In addition, a cell type–specific genetic loss of oxysterol enzymes would be required to directly demonstrate that GPR183 ligand produced by fibroblasts promotes the development of monocyte-derived lung macrophages.

Gpr183^−/− and Gpr183^GFP/+ mice were previously described (Emgård et al., 2018; Pereira et al., 2009). Ch25h^−/− mice were obtained from The Jackson Laboratory (strain #016263). Ch25h^tdTom mice (Frascoli et al., 2023) were crossed with Gpr183^GFP/+ mice to generate Gpr183^GFP/+Ch25h^tdTom double reporter mice. IM-DTR (Tmem119^CreCx3cr1^LSL-DTR) mice to deplete interstitial lung macrophages were described by Vanneste et al. (2023). SPAM deleter (Epx^CreSiglecf^Lox-DTR-Lox) mice to deplete alveolar macrophages were described by Verwaerde et al. (2025). C57BL/6 (B6) CD45.1⁺ (B6.SJL-Ptprc^aPepc^b/BoyJ) mice were obtained from in-house breeding. Both male and female mice at least 6 wk old were used for experiments with similar results. Co-housed Gpr183^+/+ littermates were used as controls for experiments with Gpr183^−/− mice. All mouse strains were housed in individually ventilated cages in specific pathogen-free conditions. Animal experiments were approved by the Linköping Ethics Committee (#20547-2021), the Animal Care and Use Committee of the University of Liège, Liège, Belgium (#2642), the Animal Ethics Committee of Ghent University (#LA1400091/LA2400526), as well as the Animal Care and Use Committee of UMass Chan Medical School (#202000151, 201900263) and performed in accordance with institutional guidelines.

BAL fluid was collected from patients with bronchiectasis at Karolinska University Hospital Huddinge, Huddinge, Sweden, by bronchoscopy. See Table S1 for information on bronchiectasis patients. The Swedish Ethical Review Authority approved the collection of human BAL fluid (#2019-00618, 2020-07094). Informed consent was obtained from all donors after giving verbal and written information.

In general, to generate bone marrow chimeras, recipient mice underwent whole-body irradiation with 2 × 5 Gy before receiving 2 × 10⁶ bone marrow cells by tail vein injection. Irradiated mice were kept on antibiotic prophylaxis (Sulfatrim) in drinking water for 3 wk. To determine Gpr183 mRNA expression in monocyte-derived alveolar macrophages, irradiated B6 mice (CD45.1⁺) were injected with bone marrow from either Gpr183^GFP/+ or Gpr183^GFP/GFP reporter mice (CD45.2⁺) and analyzed 3, 5, and 12 wk after bone marrow injection (Fig. 2, E and F). Irradiated B6 mice (CD45.1⁺) injected with bone marrow from either Gpr183^+/+ or Gpr183^+/− mice (CD45.2⁺) were used as controls. For competitive bone marrow chimeras, irradiated B6 CD45.1⁺ recipient mice were injected with a 1:1 mixture of Gpr183^+/+ (CD45.1.2⁺) and Gpr183^−/− (CD45.2⁺) or Gpr183^+/+ (CD45.2⁺) and Gpr183^−/− (CD45.1.2⁺) bone marrow cells. Competitive bone marrow chimeras were analyzed at 3 wk (Fig. 7, A and B), 9–13 wk (Fig. 4, A and B), 19–23 wk (Fig. 4 C), or 9–29 wk (Fig. 4 E) after bone marrow transfer. Single-transfer bone marrow chimeras (Fig. 4 D) were generated by injecting irradiated B6 mice (CD45.1⁺) with either Gpr183^+/+ or Gpr183^−/− bone marrow cells (CD45.2⁺) and analyzed 7–8 wk after bone marrow transfer. Mixed Gpr183^+/+/Rag1^−/− and Gpr183^−/−/Rag1^−/− chimeras (Fig. 6, A and B) were generated by injecting irradiated B6 mice (CD45.1⁺) with a 1:1 mixture of either Gpr183^+/+ (CD45.1.2⁺) and Rag1^−/− (CD45.2⁺) bone marrow or Gpr183^+/+ (CD45.1.2⁺) and Rag1^−/− (CD45.2⁺) bone marrow. Analysis of chimeras was performed at 10 wk after bone marrow transfer. Mixed Gpr183^+/+/Ccr2^RFP/RFP and Gpr183^−/−/Ccr2^RFP/RFP chimeras were generated in an analogous manner and analyzed 9–10 wk after bone marrow injection (Fig. 6, C and D). To determine the role of 7α,25-dihydroxycholesterol in monocyte-to-lung macrophage differentiation, irradiated Ch25h^−/− mice or Ch25h^+/+ controls (CD45.2⁺) were reconstituted with an equal mixture of Gpr183^+/+ (CD45.1.2⁺) and B6 (CD45.1⁺) or Gpr183^−/− (CD45.1.2⁺) and B6 (CD45.1⁺) bone marrow cells (Fig. 7 C). Chimeras were analyzed 7–15 wk after bone marrow transfer. To distinguish a hematopoietic versus a non-hematopoietic source of Ch25h, irradiated Ch25h^+/+ or Ch25h^−/− mice were reconstituted with either Ch25h^+/+ or Ch25h^−/− bone marrow cells as indicated (Fig. 7 D). Lungs were harvested 3–4 wk after bone marrow transfer for quantitative RT-PCR to determine Ch25h mRNA expression.

To deplete alveolar macrophages, Epx^CreSiglecf^Lox-DTR-Lox mice, named “SPAM deleter” mice (Verwaerde et al., 2025), were used (Fig. 3 A). In Siglecf^Lox-DTR-Lox mice, DTR is expressed in alveolar macrophages and eosinophils. To confer alveolar macrophage–specific DTR expression, Siglecf^Lox-DTR-Lox mice were crossed with Epx^Cre mice, which results in excision of DTR in eosinophils as they express eosinophil peroxidase (Epx). In Epx^CreSiglecf^Lox-DTR-Lox mice, DTR is only expressed in alveolar macrophages, thereby allowing their specific depletion after DT administration (Fig. 3, B and C). For depletion of alveolar macrophages, SPAM deleter mice received the following concentrations of intratracheal DT: 100 pg for flow cytometry (Fig. 3, B and C) and cytokines/chemokines in BAL fluid (Fig. S2, A and B), 1 ng for bone marrow chimeras (Fig. 5, A and B), and 40 ng for the single-cell RNA-sequencing experiment (Fig. 3, E–G; Fig. 8, A–D; Fig. S2, C–E; and Fig. S5, B and C). To generate SPAM deleter bone marrow chimeras, SPAM deleter mice (CD45.2⁺) were preconditioned by busulfan administration (30 mg/kg body weight by intraperitoneal injection) to preserve resident alveolar macrophages. Preconditioned SPAM deleter mice (CD45.2⁺) were then injected intravenously with a 1:1 mixture of 3 × 10⁶Gpr183^+/+ or Gpr183^−/− (CD45.1.2⁺) and B6 (CD45.1⁺) bone marrow cells (Fig. 5 A). 7 wk after bone marrow transfer, resident alveolar macrophages (CD45.2⁺) were depleted by intratracheal administration of 1 ng DT. 4 wk after DT administration, reconstitution of the alveolar macrophage compartment was assessed by flow cytometry. For depletion of interstitial lung macrophages, we used IM-DTR (Tmem119^CreCx3cr1^LSL-DTR) mice (Vanneste et al., 2023). To generate IM-DTR bone marrow chimeras, IM-DTR mice (CD45.1.2⁺) were preconditioned by thorax-shielded irradiation with two doses of 6 Gy using a 0.6-cm-thick lead cover to preserve resident interstitial lung macrophages as described (Vanneste et al., 2023). After irradiation, mice received 0.05 mg/ml of enrofloxacin (Baytril, Bayer) in drinking water for 4 wk. After preconditioning, IM-DTR mice (CD45.1.2⁺) were injected intravenously with a 1:1 mixture of 10 × 10⁶Gpr183^+/+ or Gpr183^−/− (CD45.2⁺) and B6 (CD45.1⁺) bone marrow cells (Fig. 5 C). 4 wk after bone marrow transfer, resident interstitial lung macrophages (CD45.2⁺) were depleted by a single intraperitoneal injection of 50 ng DT (List Biological Labs) as described (Vanneste et al., 2023). 2 wk after DT injection, repopulation of the interstitial lung macrophage compartment was examined by flow cytometry.

BAL fluid was obtained by inserting a catheter into the trachea and flushing with 5 × 1 ml warm buffer (0.5% FCS/2 mM EDTA/PBS). After collecting BAL fluid, lungs were harvested. For flow cytometry, harvested non-perfused lungs were digested in RPMI 1640/5% FCS with 0.2 mg/ml of collagenase IV (Sigma-Aldrich) and 0.02 mg/ml of DNase I (Sigma-Aldrich) for 60 min at 37°C, and immune cells were purified as previously described (Evren et al., 2021). Blood obtained by cardiac puncture was diluted in 200 U/ml heparin (Sigma-Aldrich), and red blood cell lysis was performed before staining cells for flow cytometry. Bone marrow cells were isolated from femur and tibia by flushing with RPMI 1640/10% FCS/1% L-glutamine/penicillin/streptomycin/55 mM 2-mercaptoethanol before red blood cell lysis. Spleen cells were obtained by cutting the spleen into small pieces, followed by digestion for 30 min as described above for lungs (but without density gradient centrifugation). Cells from the peritoneal cavity were obtained by lavage as described (Ray and Dittel, 2010). Liver cells were isolated using protocols by either Kong et al. (2024) or Daemen et al. (2021). Brains were dissected, cut into small pieces with a scalpel and enzymatically digested using Neural Tissue Dissociation Kit T (Miltenyi Biotec) for 30 min at 37°C. Myelin was removed using 38% Percoll (Sigma-Aldrich), and red blood cells were lysed using ACK lysis buffer (Thermo Fisher Scientific).

Single-cell suspensions from various organs were first stained in FACS buffer (PBS/2% FCS) for 30 min on ice with fluorochrome- or biotin-labeled antibodies (Table S2). Then, secondary staining with streptavidin–Brilliant Violet 711 or streptavidin–BUV395 (BD Biosciences) was performed for 30 min on ice. All antibodies were pre-titrated to minimize nonspecific staining. After surface staining, cells were stained with fixable viability dye eFluor 506 (Thermo Fisher Scientific) and then fixed in PBS/2% paraformaldehyde before acquisition on a LSR II Fortessa or Symphony A5 flow cytometer (BD Biosciences). Data were analyzed with FlowJo version 10 software (https://www.flowjo.com/solutions/flowjo/downloads).

Cytokines and chemokines in the BAL fluid from SPAM deleter mice were quantified by ELISA using kits from Thermo Fisher Scientific and R&D Systems following the manufacturer’s instructions.

Ly6C^hi monocytes were enriched from bone marrow by immunomagnetic selection with the Monocyte Isolation kit (Miltenyi Biotec) according to the manufacturer’s instructions. After isolation, monocytes were resensitized in migration medium (RPMI 1640/0.5% fatty acid–free BSA/10 mM HEPES) for 30 min at 37°C/5% CO₂. Then, monocyte migration toward the indicated concentrations of 7α,25-dihydroxycholesterol (Merck) was determined using 5 μM Transwells (Corning). Monocytes (0.5–1 × 10⁵ cells per well) were allowed to migrate for 3 h. Cells in the bottom chamber were harvested and mixed with CountBright Absolute Counting Beads (Thermo Fisher Scientific) for quantification of migrated cells by flow cytometry. The frequency of cells compared with input was used to define the chemotactic response to 7α,25-dihydroxycholesterol.

Bone marrow cells were isolated from the indicated mouse strains. Cells (5–10 × 10⁵ per well) were then cultured in 12-well plates containing medium (DMEM/10% FCS/1% L-glutamine/penicillin/streptomycin) supplemented with either 20 ng/ml M-CSF (Thermo Fisher Scientific) to generate generic macrophages or 20 ng/ml GM-CSF and 2 ng/ml TGFβ₁ (Thermo Fisher Scientific) to generate alveolar macrophage–like cells (Luo et al., 2021). Medium was refreshed on day 7. On day 9, cells were harvested using 0.05% trypsin-EDTA and a cell scraper and stained for flow cytometry as described above.

To determine the single-cell metabolism of alveolar macrophages, the SCENITH (Arguello et al., 2020) method was used. Briefly, BAL fluid was harvested from competitive Gpr183^+/+/Gpr183^−/− bone marrow chimeras in homeostatic conditions, i.e., 9–11 wk after bone marrow transfer. BAL cells from two to three mice were pooled and then seeded in 96-well V-bottom plates in 100 μl medium (RPMI 1640/10% FCS/1% pyruvate/GlutaMAX/penicillin/streptomycin) at 5 × 10⁴ cells per well. In the first out of two independent experiments, a sufficient number of cells allowed for technical duplicates. Cells were then incubated for 30 min at 37°C/5% CO₂ with various metabolic inhibitors as described (Arguello et al., 2020). Next, cells were incubated with 10 μM puromycin (Sigma-Aldrich) for another 30 min at 37°C/5% CO₂ and harvested. After surface staining, cells were fixed and permeabilized with the FoxP3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific) before intracellular staining with a 1:400 dilution of anti-puromycin-PE antibody (clone 2A4; BioLegend) at 4°C overnight. After flow cytometry, metabolic values were calculated as described (Arguello et al., 2020).

Accessory lobes of the lungs were homogenized in TRIzol reagent (Invitrogen) using a tissue homogenizer (PT 2500 E; Polytron) and total RNA extracted according to the manufacturer’s instructions. cDNA was synthesized from 1.6 to 2.0 μg of total RNA with the SuperScript First-Strand Synthesis System (Invitrogen). cDNA and primer-probe sets from either Applied Biosystems (Ch25h) or Sigma-Aldrich (Hprt) were used for quantitative RT-PCR on a QuantStudio 5 Real-Time PCR system (Applied Biosystems).

For single-cell RNA sequencing, frozen BAL samples from six patients with bronchiectasis were used (Table S1). As samples were collected sequentially from different patients, they were frozen to be able to process all samples together for library preparation (see below). BAL fluid from bronchiectasis patients was centrifuged and the pellet resuspended in RPMI 1640/5% FCS before freezing. Frozen BAL samples were thawed, treated with 50 U/ml Benzonase Endonuclease (Merck) for 20 min at 37°C, and then live CD45⁺CD66⁻ cells were purified by cell sorting. Neutrophils (CD66⁺) were excluded as they are prone to cell death, especially after thawing. Single-cell libraries from purified cells were prepared with the 10x Genomics Single Cell NEXT GEM 3′ Library version 3 kit according to the manufacturer’s instructions. Libraries were then sequenced on a NextSeq 2000 (Illumina) and mapped to the human genome (GRCh37) with the Cell Ranger 3.0.2 pipeline from 10x Genomics. The R software package Seurat 4.1.1 was used for analysis. To compare the distribution of airway macrophages and monocytes in health and disease, we integrated our bronchiectasis dataset with the healthy BAL dataset by Mould et al. (2021). For this purpose, the 10 healthy control samples by Mould et al. (2021) were merged into one Seurat object and the Seurat object downsampled to 10,000 total cells. Then, the downsampled object was merged with our six bronchiectasis samples. Next, low-quality cells and doublets were removed based on the following criteria: (1) <200 and >6,000 unique genes expressed in at least three cells, (2) >20% mitochondrial content, and (3) < 5% ribosomal content. The FindVariableFeatures function was applied to select 2,000 highly variable genes, followed by ScaleData to produce a principal component analysis and neighborhood graph with 50 principal components. Then, all cells were integrated using the FindIntegrationAnchors function in Seurat, and dimensionality reduction was applied to the integrated space. Louvain clustering was performed with the FindClusters function set to 0.8, resulting in 11 clusters with 14,437 cells (Fig. 10 A). Wilcoxon rank-sum test with logfc. threshold = 0.2 as well as min.pct = 0.1 and min.diff.pct = 0.2 was employed to identify cluster-specific differentially expressed genes. Clusters were assigned cell type identities based on the expression of known cell type–specific marker genes. The lymphoid clusters were then removed by creating a new Seurat object of only myeloid clusters (clusters 0, 1, 2, 4, 6, 7, 10, and 11) with 10,978 cells. The expression of selected genes was visualized in the cell clusters with the ViolinPlot function in Seurat. The GeneModuleScore function in Seurat was used to assess the similarity of the bronchiectasis cell clusters to human monocyte and macrophage populations in terms of their gene signatures. The following published single-cell RNA-sequencing datasets were used for the gene signature score analysis: Alveolar macrophages and CD14⁺CD16⁻ monocytes from human lung (Li et al., 2022b), as well as human lung macrophages either derived from CD116⁺ fetal precursors or from CD34⁺ HSPCs in MISTRG mice (Evren et al., 2022). To infer lineage trajectories between the monocyte and macrophage clusters in bronchiectasis, the R package Slingshot 2.4.0 (https://www.bioconductor.org/packages/release/bioc/html/slingshot.html) was used. The analysis was done in an unsupervised fashion, without defining any cluster as starting point or endpoint.

Lungs were harvested from SPAM deleter mice 0, 12, 24, and 48 h and on day 5, 8, and 14 after alveolar macrophage depletion with 40 ng DT intratracheally (Fig. 3 D). Four mice (two males and two females) were used per time point. CD45⁻ non-hematopoietic cells (50% of lung cells) as well as resident CD45⁺CD64⁺ (25% of lung cells) and CD45⁺CD64⁻ hematopoietic cells (25% of lung cells) were isolated by cell sorting. Lung-resident cells were identified as cells that were protected from intravascular labeling with 1 μg intravenously injected anti-mouse CD45 antibody (clone 30-F11). After cell isolation, up to 2 × 10⁶ cells per mouse were centrifuged (400 g, 5 min, 4°C) and then resuspended in 50 μl of staining mix in PBS containing 0.04% PBS, TruStain FcX Block (BioLegend), fluorescent surface antibodies, and mouse CITE-Seq antibody panel, a unique mouse TotalSeq-C cell hashing antibody (BioLegend) diluted 1:1,000. The CITE-Seq panel (Table S3) contains 192 unique oligo-conjugated antibodies and isotype controls (BioLegend, TotalSeq-C). Replicates were tagged and pooled using TotalSeq-C hashtag antibodies specific against mouse CD45 and MHC class I (BioLegend; TotalSeq-C0301, TotalSeq-C0302, TotalSeq-C0303, and TotalSeq-C0304). TotalSeq-C hashtag oligonucleotides (HTOs) were used to distinguish the origin of cell from different mice. After 30-min incubation on ice, cells were washed (400 g, 5 min, 4°C) and resuspended in PBS/0.04% BSA. Cells were sorted from single-cell suspensions using a FACS Aria II/III (BD Biosciences). Sorted single-cell suspensions were resuspended at an estimated final concentration of 1,200–2,000 cells/μl and loaded on a Chromium GemCode Single Cell Instrument (10x Genomics) to generate single-cell Gel beads-in-EMulsion. The DNA libraries were prepared using the GemCode Single Cell 5′ Gel Bead and Library kit, version Next GEM version 2 according to the manufacturer’s instructions (10x Genomics, User Guide CG000330). Size selection with SPRIselect Reagent Kit (Beckman Coulter) was used to separately amplify cDNA molecules for 5′ gene expression and cell surface protein construction. The cDNA content of pre-fragmentation and postsample index PCR samples was analyzed using the 2100 Bioanalyzer (Agilent). Sequencing libraries were loaded on an Illumina NovaSeq flow cell at VIB Nucleomics Core on the Illumina NovaSeq 6000 platform with a read configuration of 28 cycles for read 1, 10 cycles for i7 index, 10 cycles for i5 index, and 90 cycles for read 2, incorporating 1% PhiX as a sequencing control, pooled in an 90:10 ratio for the combined 5′ gene expression and cell surface protein samples, respectively. The Cell Ranger pipeline (10x Genomics, version 6.1.1) was used to perform sample demultiplexing and to generate FASTQ files for read 1, read 2, and the i5, i7 sample index for the gene expression and cell surface protein libraries. Read 2 of the gene expression libraries was mapped to the reference mouse genome GRCm 38.99. Subsequent barcode processing, unique molecular identifiers, filtering, and gene counting were performed using the Cell Ranger suite (10x Genomics). Single-cell RNA sequencing analysis was conducted in R, beginning with the complete, unfiltered count matrix. The analysis followed the workflow described by the Marioni and Theis laboratories (Luecken and Theis, 2019). Cells expressing fewer than 200 genes and genes detected in fewer than three cells were excluded from further analysis. HTO data were incorporated using the CreateAssayObject function from the Seurat package (version 3.1.5). HTO counts were normalized using the centered log-ratio method. Demultiplexing was performed using both HTODemux and MultiSeqDemux functions from Seurat. Quality control filtering was applied to remove outliers based on three metrics: number of expressed genes, library size, and mitochondrial gene proportion. Cells exceeding five median absolute deviations from the median for the number of expressed genes or total counts were excluded. Additionally, cells with a mitochondrial gene proportion exceeding five median absolute deviations above the median were removed. This filtering strategy was applied to be lenient. Finally, for each of the samples, the Seurat (version 4.0.2) workflow was used to normalize, cluster, and visualize the gene expression data. Clustering and visualization were performed with FindNeighbors(), FindClusters(), and RunUMAP(). To demultiplex the HTO-labeled data, we used MultiSeqDemux() from Seurat with autoThresh = TRUE. Annotation was performed using differentially expressed gene sets obtained with the FindAllMarkers function. Doublets clusters were identified as concomitant expression patterns for different cell types and excluded from the analysis. This was further validated using the DoubletFinder (version 2.0). Standard visualization was done using DimPlot, FeaturePlot, and DotPlot functions from Seurat. Gene expression over time for indicated genes was smoothed using Loess regression.

Lungs were collected from IM-DTR mice at 0, 12, 24, and 48 h following interstitial macrophage depletion induced by intraperitoneal injection of 50 ng DT, with three to four mice analyzed per time point (Fig. 8 E). Single-cell suspensions were obtained after enzymatic digestion and non-hematopoietic CD45⁻ niche cells were isolated by MACS using CD45 MicroBeads (Miltenyi Biotec). Lung endothelial cells, epithelial cells, and fibroblasts were then FACS sorted separately as CD45⁻CD31⁺EpCAM⁻, CD45⁻CD31⁻EpCAM⁺ and CD45⁻CD31⁻EpCAM⁻ cells, respectively. Sorted lung endothelial cells, epithelial cells, and fibroblasts were pooled at an equicellular ratio of 1:1:1. Cells were then barcoded per time point post-DT injection with TotalSeq-A anti-mouse Hashtag antibodies (BioLegend). Hashtag barcoded cells were washed, pooled, spun down, and resuspended in PBS with 0.04% UltraPure BSA (Thermo Fisher Scientific) at a final concentration of 500 cells/μl. Cell suspensions were loaded into the Chromium Controller (10x Genomics) at a target recovery of 3 × 10³ cells per time point post-DT injection. Cells were encapsulated and partitioned, and their polyA RNAs were captured and barcoded using Chromium Single Cell 3′ GEM, Library & Gel Bead Kit version 3 (10x Genomics). The cDNAs were amplified and libraries compatible with Illumina sequencers were generated using Chromium Single Cell 3′ GEM, Library & Gel Bead Kit version 3 (10x Genomics). For HTO library, 1 μl HTO additive primer v2 (0.2 μM stock) were added to the mix at the cDNA amplification step. The libraries were sequenced on an Illumina NovaSeq sequencer on an SP100 cell flow (read 1, 28 cycles; read 2, 76 cycles; index i7, 10 cycles; index i5, 10 cycles) with a sequencing depth of 20,000 reads per cell. Single-cell RNA-sequencing data were processed using Cell Ranger (version 3.0.2) for demultiplexing, alignment to the mouse reference genome (GRCm38/mm10), filtering, unique molecular identifier counting, and generation of gene-barcode matrices. Downstream analysis was carried out in R (version 4.0.3) using the Seurat package (version 4.0.0). For each sample, filtered matrices containing cell barcodes and gene features were used to construct Seurat objects. Quality control included removal of cells with fewer than 200 detected genes, genes expressed in fewer than three cells, and cells with over 20% mitochondrial gene content. Gene expression was normalized using the default LogNormalize method (scale factor: 10,000), followed by log transformation. The top 2,000 highly variable genes were identified using the vst method. Clustering was performed using the FindClusters function, and cell types were annotated with the SingleR package. Contaminating immune cells were excluded from the final dataset.

To induce local airway inflammation and monocyte recruitment, Gpr183^GFP/+Ch25h^tdTom double reporter mice received 50 μg LPS from Escherichia coli O55:B5 (Sigma-Aldrich) via the intranasal route (Fig. 9 A). Control mice received the same volume of PBS. On day 3 after intranasal LPS or PBS treatment, lungs were perfused with 20 ml cold PBS and slowly inflated with 1–1.5 ml 50% Optimal Cutting Temperature compound in PBS. Harvested lung were fixed in 1.3% PFA for 6 h at 4°C. After washing three times with PBS, fixed lungs were dehydrated by incubation in a successive sucrose gradient (10% → 20% → 30% sucrose in PBS). Finally, lungs were frozen in a slurry of ethanol on dry ice, embedded in OCT, and stored at −80°C before preparing 7-μm sections. Lung sections were first stained for 2 h at room temperature in blocking buffer (0.3% Triton X-100 [Merck] and 2% donkey serum in PBS) with rat anti-mouse CD68 (1:50) and rabbit anti-mouse PDGFRα (1:50) primary antibodies. After washing sections with PBS three times, anti-rat AF647 (1:500) and anti-rabbit AF750 (1:250) secondary antibodies were added in blocking buffer and incubated for 2 h in the dark at room temperature, then washed with PBS three times. Sections were further stained with anti-GFP AF488 (1:200) and anti-tdTom DyLight 550 (1:100) for 2 h in the dark at room temperature. After washing, samples were incubated with MHC-II AF700 antibody (1:100) for 2 h in the dark at room temperature. Finally, sections were washed three times with PBS and were mounted with 10 μl Fluoromount-G reagent (Invitrogen) containing 0.3% SYTOX blue nucleic acid stain (Invitrogen) and stored at 4°C in the dark until imaging. Confocal images were acquired using a Leica Stellaris 8 Inverted Confocal Microscope equipped with a White Light Laser, using a 40×/1.30 NA oil-immersion objective. Image resolution was set at 512 × 512 pixels with unidirectional scanning at 400 Hz. Cell types within lung sections (Fig. 9, B and C) were defined as monocytes (CD68⁺MHCII⁻), macrophages (CD68⁺MHCII⁺), and fibroblasts (PDGFRα⁺). Antibodies against GFP and tdTomato were used to visualize Gpr183⁺ and Ch25h⁺ cells, respectively. Microscopy images were analyzed using QuPath (version 0.5.1) (Bankhead et al., 2017). Out of focus regions and obvious artefacts were manually excluded prior to quantification. Nuclei were identified using the Sytox counterstain and segmented with the StarDist plugin (Schmidt et al., 2018, Preprint), applying the pretrained model dsb2018_heavy_augment.pb. QuPath measurement tables were exported and processed in Python, including nuclear centroid coordinates (X/Y, µm) and per-cell morphology and fluorescence-intensity features. Marker positivity was determined by supervised classification on a per-image basis to account for staining variability. For each marker channel (CD68, GFP, PDGFRα, MHCII, and tdTomato), a set of cells was manually annotated as positive or negative based on visual inspection, and a random forest classifier was trained using QuPath-derived features and then applied to all detected cells in the same image. Classification was performed independently for each marker, and cell populations were defined by combinatorial marker expression (monocytes: CD68⁺MHC-II⁻; macrophages: CD68⁺MHC-II⁺; PDGFRα⁺tdTomato⁺CD68⁻ fibroblasts; GFP⁺ versus GFP^- subsets). The abundance of each population was quantified per image. For the double-reporter experiments (Fig. 9), three biological replicates (mice) per condition were analyzed (LPS and control), with four images in total quantified for the LPS group and three images in total for the control group. Spatial proximity between populations was quantified using nearest neighbor distances computed from nuclear centroid coordinates with scipy.spatial.distance (SciPy version 1.15.2). For a given source population, Euclidean distances to all cells in a target population were computed, and the minimum distance was recorded for each source cell; distances <1 μm were excluded to minimize artefacts due to overlapping nuclei or segmentation uncertainty. For the analysis shown in Fig. 9 D, nearest neighbor distances were computed from each PDGFRα⁺tdTomato⁺CD68⁻ fibroblast to the closest CD68⁺MHC-II⁻ monocyte or CD68⁺MHC-II⁺ macrophage, such that each value contributing to the violin plots corresponds to one fibroblast. Smaller nearest neighbor distances indicate fibroblasts with a monocyte/macrophage in close proximity, whereas larger nearest neighbor distances reflect fibroblasts that are comparatively isolated from monocytes/macrophages in that field. Nearest neighbor distances were pooled across images within each experimental group and summarized as violin plots. Nearest neighbor analysis was selected because it provides an interpretable per-cell measure of local proximity that is well suited to centroid-based segmentation, avoids selecting an arbitrary neighborhood radius, and is robust to heterogeneous lung tissue geometry (airspaces and excluded regions).

Data are presented as mean and SEM in the figures. The number of biological replicates (n) and repeat experiments are stated in the figure legends. Statistical tests used were Student’s t, Welch’s t, or Mann–Whitney test (for comparison between two groups) and one-way ANOVA with post hoc testing using Tukey’s multiple comparison test (for multigroup comparisons). To determine statistical significance, two-sided testing was applied with α = 0.05. P values for statistical significance are shown in the figures. Statistical analysis was performed with GraphPad Prism 9 (https://www.graphpad.com/scientific-software/prism/).

Fig. S1 shows the flow cytometry gating strategy to identify immune cells in the mouse lung and is relevant for Figs. 1, 2, 4, 6, and 7. Fig. S2 (relevant for Fig. 3; Fig. 8, A–D; and Fig. S5, B and C) shows concentrations of cytokines and chemokines in the BAL fluid of SPAM deleter mice as well as marker genes used to delineate single-cell RNA-sequencing cell clusters in the lung of SPAM deleter mice. Fig. S3 shows the flow cytometry gating strategy to identify monocytes and macrophages in tissues other than the lung in Fig. 4 E. Fig. S4 (relevant for Fig. 4) shows the energy metabolism of alveolar macrophages isolated from competitive Gpr183^+/+/Gpr183^−/− bone marrow chimeras and the differentiation of Gpr183^+/+ and Gpr183^−/− bone marrow cells into macrophages in vitro. Fig. S5 shows marker genes to define non-hematopoietic lung cells in IM-DTR mice by single-cell RNA sequencing (relevant to Fig. 8, F–H) as well as expression of Ccl2 and Csf1 in SPAM deleter mice after alveolar macrophage depletion (relevant to Fig. 8, A–D). Table S1 (related to Fig. 10) contains information on the human BAL donors that were used for single-cell RNA sequencing. Table S2 (related to all figures) provides information on reagents, assays, biological samples, mouse strains, deposited data, and software used in the study. Table S3 (related to Fig. 3, D–G; Fig. 8 A–D; Fig. S2, C–E; and Fig. S5, B and C) presents a list of antibodies used for CITE-seq analysis of SPAM deleter mice.

Single-cell RNA-sequencing data (read counts) from bronchiectasis patients are available at GEO (accession number GSE217516). The raw sequencing data cannot be shared publicly to protect the identity of the human tissue donors as required by Swedish legislation and the European General Data Protection Regulation. The single-cell RNA-sequencing data from healthy BAL donors by Mould et al. (2021) that were used for data integration with the bronchiectasis dataset are available at GEO (accession number GSE151928). Single-cell RNA-sequencing data from lung cells of SPAM deleter mice and lung niche cells from IM-DTR mice are available at ArrayExpress (accession number E-MTAB-16590) and GEO (accession number GSE303395), respectively. All other data are available in the article itself and its supplementary materials and are also available upon reasonable request from the corresponding authors.

We would like to thank the Bioinformatics and Expression Analysis core facility, which is supported by the board of research at the Karolinska Institutet. We acknowledge Jason Cyster (University of California San Francisco, San Francisco, CA, USA) and João Pereira (Yale University, New Haven, CT, USA) for providing Gpr183^GFP/+ reporter mice. We thank Klas Blomgren, Ahmed Osman, and Alejandro Lastra Romero (Karolinska Institutet, Solna, Sweden) as well as Kees-Jan Pronk and Isabel Hidalgo (Lund University, Lund, Sweden) for providing bones from Ccr2^RFP/RFP reporter mice and Rag1^−/− mice, respectively.

This work was supported by grants from the Swedish Research Council (2019-01099 and 2022-01310), the Swedish Heart-Lung Foundation (20190198, 20220361, and 20240579), the Region Stockholm/Center for Innovative Medicine (20190152), and Karolinska Institutet (2-1060/2018, 2020-01438, 2021-00327, 2024-02407, and 2-977/2025) to T. Willinger In addition, the work was supported by a grant from Fonds Wetenschappelijk Onderzoek (G063123N) to Bart N. Lambrecht and by grants from the Fund for Scientific Research-FNRS (PDR-INTER T.0200.25), the FRFS-WELBIO (WELBIO-CR-2022A-10), the European Research Council (ERC CoG 2023 MoMacTrajectALI: 101124390), the FEDER, and the Baillet Latour Fund to T. Marichal, as well as by grants from the Swedish Heart-Lung Foundation (2020619, 20210434, and 20220478) to A. Bossios and the European Research Council (948692) to N.K. Björkström A. Reboldi is a Burroughs-Wellcome Investigator in the Pathogenesis of Infectious Disease and is supported by National Institutes of Health grant AI155727. Open Access funding provided by Karolinska Institutet.

Author contributions: Laura Bub: formal analysis, investigation, validation, visualization, and writing—original draft. Elza Evren: formal analysis, investigation, visualization, and writing—review and editing. Stijn Verwaerde: data curation, formal analysis, investigation, methodology, visualization, and writing—review and editing. Cecilia Ruscitti: investigation, visualization, and writing—review and editing. Domien Vanneste: formal analysis, investigation, methodology, software, visualization, and writing—review and editing. Pia Ghosh: resources. Yu Gao: investigation. Natalie Sleiers: investigation. Rong Deng: investigation. Marta López Montes: investigation and writing—review and editing. Kelsey Howley: investigation. Raphaël La Rocca: formal analysis and visualization. Annika Niehrs: investigation and writing—review and editing. Vassilis Glaros: investigation. Maria Reina-Campos: formal analysis and investigation. Barbro Dahlen: funding acquisition, resources, and supervision. Anna Smed-Sörensen: resources and writing—review and editing. Harald Lund: formal analysis and investigation. Taras Kreslavsky: supervision and writing—review and editing. Niklas K. Björkström: writing—review and editing. Andrea Reboldi: investigation, methodology, and resources. Apostolos Bossios: funding acquisition, project administration, resources, supervision, and writing—review and editing. Thomas Marichal: conceptualization, formal analysis, funding acquisition, methodology, project administration, resources, supervision, validation, visualization, and writing—review and editing. Bart N. Lambrecht: conceptualization, formal analysis, funding acquisition, methodology, project administration, resources, supervision, and writing—review and editing. Tim Willinger: conceptualization, formal analysis, funding acquisition, methodology, project administration, resources, supervision, visualization, and writing—original draft, review, and editing.