Mice were housed in specific pathogen-free barrier facilities. Mice were used in accordance with protocols approved by the Institutional Animal Care and Use Committee at the Institute of Hematology, Chinese Academy of Medical Sciences. C57BL/6 mice, Lkb1 ^f/f mice, AMPKα1 ^f/f mice and Cd11c ^Cre mice were all purchased from Jackson Laboratories (Bar Harbor, ME, USA). The mice were backcrossed with C57BL/six mice for at least seven generations. Lkb1 ^f/f mice and AMPKα1 ^f/f mice were crossed with Cd11c ^Cre mice to generate Cd11c ^Cre Lkb1 ^f/f mice and Cd11c ^Cre AMPKα1 ^f/f mice, respectively. Mice were used at 6–8 weeks old, unless otherwise indicated.

For analysis of cell surface markers, single cell suspensions were prepared from BAL fluid, lung, medLN, kidney, liver and thymus. To obtain BAL fluid cells, mice were sacrificed and pinned upright to a dissection board. The tracheas were exposed, and a small opening was cut with surgical scissors. A 1 ml syringe with a needle was inserted into the tracheal opening, and the alveolar cavity was washed with 500 µl PBS at least six times. BAL fluid was obtained and transferred to a 10 ml tube. The lung, medLN, kidney, liver and thymus single-cell suspensions were prepared using the same procedure as follows. The target tissues were finely ground with a filter, and the red blood cells were lysed with 200 µl erythrocyte lysate (Solarbio) for 10 min at 4°C. Leukocytes were washed with PBS to obtain single cell suspensions.

Flow cytometry antibodies and their clone numbers are listed in Table S2 . Cell surface staining was performed at 4°C for 30–40 min in the dark. The cells were stimulated with phorbol myristate acetate (50 ng/ml) and ionomycin (500 ng/ml) for 4–5 h for intracellular cytokine staining. Intracellular staining with Ki67 was performed using Foxp3 staining kits (eBioscience). The antibodies we used were purchased from Biolegend, BD Bioscience, eBioscience, and Invitrogen. The indicated AM populations were sorted by FACSAria III (BD Biosciences) from BAL fluid or lung. The purity of sorted populations was >99%, unless otherwise indicated. Data were obtained on a FACSCanto II (BD Biosciences) and analyzed using FlowJo software (FlowJo LLC, Ashland, Oregon).

Recipient CD45.1⁺CD45.2⁺ mice (8–10 weeks old) were lethally irradiated twice with 400 cGy. Irradiated recipient mice were then intravenously injected with 1 × 10⁷ BM cells, which were from Lkb1 ^f/f (CD45.1⁺) mice and Cd11c ^Cre Lkb1 ^f/f (CD45.2⁺) mice at a ratio of 3:1. Chimeric mice were housed in sterile caging for 8 weeks to allow for reconstitution.

According to the instructions of the manufacturers, Annexin V and PI staining were performed with an apoptosis detection kit (Biolegend) to test the apoptosis of AMs from Lkb1 ^f/f mice and Cd11c ^Cre Lkb1 ^f/f mice by flow cytometry (Canto II, BD Biosciences).

Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice were intranasally challenged with 5 × 10⁴ colony-forming units (CFU) per animal of S. aureus in 50 μl PBS. The mouse body weight was monitored until the mouse died or up to 6 days. The secretion of inflammatory cytokines by T cells in the lung and medLN and the percentage of neutrophils and eosinophils in the lung and BAL fluid were analyzed by flow cytometry 3 days after challenge. BAL fluid was obtained for culture on day 3 after challenge to calculate CFU.

Encapsome and Clodrosome are the products of Encapsula NanoSciences (SKU# CLD-8901). The clodronate was encapsulated with liposomes to form a multilamellar liposome suspension, which is called Clodrosome. Encapsome is the control liposome of Clodrosome, in which clodronate was not added to the liposomes. The larger particles were removed from the liposomes, and all the preparation processes were performed under sterile conditions. When treated with Clodrosome, the phagocytic cells in animals recognize liposomes as invasive foreign particles and remove liposomes by phagocytosis. Then, the liposomes released the clodronate into the cytosol, leading to cell death. The recognition and phagocytosis mechanism of Encapsome was the same as that of Clodrosome. Since the Encapsome did not contain clodronate, the phagocytic cells were not killed.

Lkb1 ^f/f mice were intranasally challenged with 100 μl (500 μg) Encapsome or Clodrosome (5 mg/ml, from Encapsula NanoSciences, SKU# CLD-8901) per mouse once a day, one to two times a week.

Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice were intraperitoneally anesthetized with chloral hydrate (Solarbio) and then intranasally challenged with 25 μg HDM (Greer Laboratories Inc., Lenoir, NC, USA) per animal in 50 μl PBS on days 0, 2, 4, 7, 9, 11, 14, 15, and 16 (12). On day 17, the mice were sacrificed, and the secretion of inflammatory cytokines by T cells in the lung and medLNs and the percentage of neutrophils and eosinophils in the lung and BAL fluid were analyzed by flow cytometry.

CD45⁺ CD11c⁺ SiglecF⁺ AMs in the lungs of Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice were sorted for RNA sequencing. The purity of sorted populations was more than 99%. RNA was extracted using TRIzol reagent (Invitrogen) for RNA sequencing analysis on the BGISEQ500 platform (BGI-Shenzhen, China). DESeq2 (v1.4.5) was used for the differential expression analysis (18) with a Q value ≤0.05. RNA transcriptome sequencing datasets were analyzed and edited using GlueGO in Cytoscape software (19). We have submitted the RNA-Seq datasets to the Gene Expression Omnibus (GEO), and the accession number is GSE167349.

Samples were harvested from the lungs on day 3 after S. aureus infection and fixed with 10% formalin immediately. Samples were washed using 70% ethanol and embedded in paraffin. Then, they were cut into 6-μm thick slices and stained with hematoxylin and eosin. The presented data are from individual lung. All slices used for analysis were encoded and read blindly. Photomicrographs were taken at a 10 × and 40 × magnification.

Two investigators quantified the lung injury score blindly according to published criteria of American Thoracic Society Documents (20). Each sample was analyzed for at least five fields. Five independent parameters of lung injury scoring system were as follows: A. neutrophils in the alveolar space; B. neutrophils in the interstitial space; C. hyaline membranes; D. proteinaceous debris filling the airspaces; E. alveolar septal thickening. The calculation was: scores = [(20 × A) + (14 × B) + (7 × C) + (7 × D) + (2 × E)]/(number of fields × 100). The resulting score is a continuous value between zero and one (inclusive).

Data were analyzed using GraphPad Software (Prism 5.00, San Diego, CA, USA). The statistical significance of the difference between the two groups was calculated, and the P-value was determined. According to the number of comparison groups, Student’s t-test or two-way ANOVA was performed. P <0.05 was considered statistically significant. *P <0.05; **P <0.01; ***P <0.001.

CD11c is highly expressed on AMs, and studies have shown that Lysm ^Cre-mediated recombination results in inefficient gene deletion in AMs (1, 4, 21). Therefore, we used Cd11c ^Cre Lkb1 ^f/f mice, in which Lkb1 was deleted in CD11c⁺ cells, including DCs and macrophages, to explore the function of Lkb1 in AMs. Interestingly, we found that the numbers of AMs and CD103⁺ DCs were prominently decreased in bronchoalveolar lavage (BAL) fluid and/or lungs from Cd11c ^Cre Lkb1 ^f/f mice ( Figures 1A–D ). Conversely, lung interstitial macrophages (IMs), lung CD11b⁺ DCs and DCs in other tissues, including the kidney, liver and thymus, were unaffected ( Figures 1C, D and Figures S1A–C ). We then investigated whether cell-intrinsic or cell-extrinsic factors contributed to AM reduction using a mixed bone marrow chimera model in which Lkb1-deficient (CD45.2⁺) bone marrow (BM) cells and wild type (CD45.1⁺) BM cells were mixed at a ratio of 3:1 and then transferred into CD45.1⁺ CD45.2⁺ host mice. We observed significant impairment only in AM and CD103⁺ DC accumulation ( Figures 1E, F ), indicating that the homeostatic defects of AMs and CD103⁺ DCs were cell intrinsic.

The development of AMs, including the stepwise process from monocytes, prealveolar macrophages (pre-AMs) and mature AMs in neonatal mice, is essential for maintaining the AM population (1). To determine whether the development of AMs was retarded and participated in their absence, we analyzed the numbers of monocytes, pre-AMs and mature AMs in the lungs of Cd11c ^Cre Lkb1 ^f/f and Lkb1 ^f/f mice on DOB (day of birth) and PND7 (postnatal day 7). The mature AM population was comparable between Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice on DOB. Although the percentage and absolute number of mature AMs in Cd11c ^Cre Lkb1 ^f/f mice were significantly reduced, we observed that the percentages and absolute numbers of monocytes, fetal macrophages and pre-AMs were comparable in Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice at PND7 ( Figures 2A–D ). These results indicated that the reduction in the AM pool occurred postnatally but not prenatally. Since the AM population is autonomously maintained by proliferative self-renewal throughout life (1), we thus examined whether self-renewal was impaired and contributed to AM reduction in Cd11c ^Cre Lkb1 ^f/f mice. Results showed that Lkb1-deficient AMs exhibited increased apoptosis and reduced proliferation compared to wild type AMs ( Figures 2E, F ). These results suggest that impaired self-renewal rather than AM development contribute to AM reduction in Cd11c ^Cre Lkb1 ^f/f mice.

The AM pool is maintained at constant levels to ensure lung homeostasis, respiratory function and the pulmonary response of scavenging inhaled pathogens (22). Therefore, we detected the impact of decreased AM number on the pulmonary response using an S. aureus pneumonia model. Compared to Lkb1 ^f/f mice, Cd11c ^Cre Lkb1 ^f/f mice presented much more body weight loss and severe inflammation by pathologic analysis ( Figures 3A–C and Figure S2A ). Moreover, we evaluated the bacterial scavenging capacity of AMs by culturing BAL fluid from Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice challenged with S. aureus pneumonia. BAL fluid from Cd11c ^Cre Lkb1 ^f/f mice produced more colonies than that from Lkb1 ^f/f mice, indicating that Cd11c ^Cre Lkb1 ^f/f mice had defects in bacterial scavenging capacity ( Figure 3D ). In addition, we observed an increased percentage of T helper 17 (Th17) cells in the lung and mediastinal lymph node (medLN) in Cd11c ^Cre Lkb1 ^f/f mice ( Figures 3E, F ). There is evidence that IL-17 enhances neutrophil accumulation under inflammatory conditions (23, 24). Indeed, an increased frequency of neutrophils and a significant reduction in eosinophils were observed in the lungs and BAL in Cd11c ^Cre Lkb1 ^f/f mice ( Figures 3G, H ). These results indicate that Th17/neutrophilic inflammation was driven by Lkb1 ablation in AMs. We next determined whether loss of AMs also contributes to neutrophilic lung inflammation during S. aureus pneumonia. Lkb1 ^f/f mice intranasally administered Clodrosome but not control liposomes-Encapsome exhibited approximately 70% depletion of AMs with no change in DCs, including CD11b⁺ DCs or CD103⁺ DCs ( Figures S3A–D ). Consistent with the phenotype in Cd11c ^Cre Lkb1 ^f/f mice, Lkb1 ^f/f mice without AMs also exhibited more severe pathology, impaired bacteria-scavenging capacity and an imbalance between eosinophils and neutrophils in the lung during S. aureus pneumonia ( Figures 3B–D, G, H and Figure S2A ). Although the potential role of CD103⁺ DCs may not be completely excluded, AM deletion in Lkb1 ^f/f mice caused almost the same degree of changes in lung pathology and neutrophil and eosinophil variation as observed in Cd11c ^Cre Lkb1 ^f/f mice; thus, we concluded that AMs exerted a dominant role in S. aureus infection and might be involved in maintaining the balance between eosinophils and neutrophils in the lung during S. aureus pneumonia.

We also utilized an asthma model to investigate whether the lung inflammatory response was functionally affected in Cd11c ^Cre Lkb1 ^f/f mice. Lkb1 ^f/f and Cd11c ^Cre Lkb1 ^f/f mice were challenged with house dust mite (HDM) allergen intranasally. We observed a significant reduction in IL-4-producing T helper 2 (Th2) cells and an increased frequency of IL-17-producing Th17 cells in the lungs and medLNs of Cd11c ^Cre Lkb1 ^f/f mice ( Figures 4A, B ). The typical feature of Th2-polarized allergies is the robust recruitment of eosinophils, whereas IL-17 increases neutrophil recruitment (25). Remarkably, we observed that eosinophil accumulation was decreased, while the percentages of neutrophils were significantly increased in BAL fluid and lungs from Cd11c ^Cre Lkb1 ^f/f mice ( Figures 4C, D ). Moreover, Cd11c ^Cre Lkb1 ^f/f mice manifested more severe pathology in the lung than Lkb1 ^f/f mice ( Figures 4E, F ). These results indicate that the immunoprotective function is impaired in Cd11c ^Cre Lkb1 ^f/f mice during allergic inflammation. AMs were directly deleted by intranasal administration of Clodrosome in Lkb1 ^f/f mice to verify the crucial role of AMs in asthma. Although Lkb1 ^f/f mice with AM depletion developed more severe pathology and displayed defects in recruiting eosinophils in the lung ( Figures S4A–D ), no significant increase in neutrophils was observed in the lungs of Lkb1 ^f/f mice in the absence of AMs. These results suggest that AMs exert a dominant role in protecting the host from allergic inflammation and participate in regulating the balance of neutrophils and eosinophils during lung inflammation.

Adenosine monophosphate-activated protein kinase (AMPK) is an important downstream target of Lkb1 for regulating metabolism (14). To investigate whether AMPK was involved in AM reduction caused by Lkb1 deletion, we generated Cd11c ^Cre AMPKα1 ^f/f mice in which AMPK was specifically deleted in CD11c⁺ cells. However, we found that the numbers of AMs or DCs were not significantly different between AMPKα1 ^f/f and Cd11c ^Cre AMPKα1 ^f/f mice ( Figures S5A–D ). These results indicate that the function of Lkb1 in maintaining AM abundance is independent of AMPK.

To explore the potential molecular mechanisms of Lkb1 in regulating the homeostasis and functions of AMs, we further analyzed the transcriptome sequencing of AMs sorted from Cd11c ^Cre Lkb1 ^f/f and Lkb1 ^f/f mice. Remarkably, there were 508 transcripts with a ≥2- or ≤-2-fold change in Lkb1-deficient AMs. Expression of a wide variety of genes critically involved in immune function was increased in Lkb1-deficient AMs, including those encoding secreted cytokines and chemokines (Il1b, Il33, Cc19), their related receptors (Il1rl1, Il17re, C3ar1) associated with immune activation, factors in the acute inflammatory response (Ptgs2) (26), and molecular and chemokine receptors in neutrophil migration and chemotaxis (Itgb3, Itgam, Pecam1, Ccxr2, Ccr7) (27). The increased expression of inflammatory mediators in Lkb1-deficient AMs might contribute to the more severe pathology in Cd11c ^Cre Lkb1 ^f/f mice during S. aureus infection and allergic inflammation. Four genes were observed to have significantly reduced expression in Lkb1-deficient AMs, including Igf-1, Fbp1, Prkn and Ugtla2. Igf-1 enhances phagocytosis and bacterial killing in AMs (28, 29), which could explain the impaired bacterial scavenging capacity in AMs from Cd11c ^Cre Lkb1 ^f/f mice during S. aureus infection. The Ugtla2 gene regulates glycogen/glucose level and promotes the storage of glycogen (30), and the Fbp1 gene encodes fructose-1,6-bisphosphatase 1, which inhibits glycolysis and tumor growth (31). Pathogenic variants in Prkn led to mitochondrial autophagy (32). In addition, some genes critically participating in lipid metabolism exhibited increased expression, such as fatty acid synthesis (Dgat2) (33) and transport (Slc27a4) and lipid synthesis, storage and β-oxidation (Acsl3, Dgat2, Hilpda) (34). Inhba and Osm genes, which inhibit cell proliferation, and the G0s2 gene, whose function is promoting apoptosis, were increased in Lkb1-deficient AMs ( Figures 5A, B ) (35). Furthermore, GO (Gene Ontology) enrichment and pathway analysis revealed that positive regulation of the apoptotic signaling pathway was upregulated ( Figure 5C , Table S1 ), which could explain the higher apoptosis proportion in Lkb1-deficient AMs. In addition, some pathways related to immune responses and metabolism were also enriched in Lkb1-deficient AMs ( Figure 5C ). Immune response-related pathways, including positive regulation of the (acute) inflammatory response and the IL-17 signaling pathway, were upregulated in Lkb1-deficient AMs, confirming the crucial function of Lkb1 in the suppression of the inflammatory response. Metabolism-related pathways, including positive regulation of the lipid biosynthetic process pathway, long-chain fatty acid import into cells and lipid storage, were upregulated in Lkb1-deficient AMs. The pathway for regulating the ATP biosynthetic process was downregulated in Lkb1-deficient AMs, which could be the potential mechanism through AM reduction occurs. These results suggest that Lkb1 operates as a pivotal mediator of AM immune protective function and metabolic homeostasis in the lung.