AMPK-driven Macrophage Responses Are Autophagy Dependent in Experimental Bronchopulmonary Dysplasia

Abstract

The pathogenesis of bronchopulmonary dysplasia (BPD) remains incompletely understood. Recent studies suggest insufficient AMP-activated protein kinase (AMPK) activation as a potential cause of impaired autophagy in rodent and nonhuman primate models of BPD. Impaired autophagy is associated with enhanced inflammatory signaling in alveolar macrophages (AMs) and increased severity of murine BPD induced by neonatal hyperoxia exposure. The goal of this study was to determine the role of autophagy and AMPK activation in macrophage responses in murine BPD. C57BL/6J mice were exposed to neonatal hyperoxia starting on postnatal day (P)1 and treated with the AMPK activator 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) between P3 and P6. Mice were euthanized on P7, and markers of AMPK activation and autophagy were assessed by immunoblotting. Alveolarization was assessed using radial alveolar counts, mean linear intercept measurements, and quantification of alveolar septal myofibroblasts. Relative mRNA expression of M1-like and M2-like genes was assessed in AMs isolated from BAL fluid from wild-type, LysMCre⁻-Becn1^fl/fl, and LysMCre⁺-Becn1^fl/fl mice after neonatal hyperoxia exposure. AICAR treatment resulted in AMPK activation and induction of autophagic activity in whole-lung and BAL cell lysates and attenuated hyperoxia-induced alveolar simplification in neonatal lungs. AICAR-treated control but not Beclin1-deficient AMs demonstrated significantly decreased expression of M1-like markers and significantly increased expression of M2-like markers. In conclusion, pharmacologic activation of AMPK by AICAR resulted in induction of autophagy and played a protective role, at least in part, through attenuation of proinflammatory signaling in AMs via autophagy-dependent mechanisms in a murine model of BPD.

Bronchopulmonary dysplasia (BPD) is a unique form of chronic lung disease that develops in preterm infants because of interrupted in utero lung development perpetuated by postnatal lung injury (1). Despite efforts to minimize the deleterious effects of mechanical ventilation and supplemental oxygen therapy during neonatal intensive care, ∼40% of infants with birth weights less than 1,000 g develop BPD (2). Infants with BPD are at increased risk for rehospitalizations and long-term adverse effects on their pulmonary health (3, 4). They are also at high risk for feeding difficulties, failure to thrive, and neurodevelopmental impairment (5, 6). Currently, there is a paucity of evidence-based treatments for BPD because of an incomplete understanding of the underlying molecular mechanisms (7, 8).

Autophagy is a highly conserved and regulated cellular process for lysosomal degradation and recycling of damaged and long-lived organelles and macromolecules (9, 10). On the basis of recent studies from our laboratory and others, autophagy has emerged as a key regulator of lung development and a potentially protective pathway in BPD (11–13). In newborn murine lungs, basal autophagy levels are low and peak during alveolarization at approximately postnatal day (P)7. Hyperoxia exposure starting on P1 initially induces autophagic activity at P3. However, ongoing hyperoxia exposure subsequently leads to significantly decreased autophagy by P7. Studies in the autophagy reporter GFP-LC3 transgenic mouse model have shown that autophagic activity is primarily detected in type 2 alveolar epithelial cells (AEC2) and macrophages in newborn lungs (13). In murine models of BPD, induction of autophagy improves alveolarization, whereas autophagy deficiency results in greater lung injury and impaired alveolarization (12, 13). Furthermore, we identified a novel link between autophagy and inflammatory responses of alveolar macrophages (AMs) in a murine model of BPD because autophagy-deficient Beclin1-heterozygote mice exhibited increased production of proinflammatory, “M1-like” mediators in AMs as well as increased lung concentrations of IL-1β in response to neonatal hyperoxia exposure (13).

At the post-transcriptional level, AMP-activated protein kinase (AMPK) is one of the key regulators of autophagy (14, 15). AMPK is activated by upstream kinases in response to increases in intracellular AMP-to-ATP ratio and Ca²⁺ concentrations, switching off ATP-consuming anabolic pathways, such as mTOR complex 1 (mTORC1), while switching on ATP-producing catabolic pathways, such as autophagy (16). In our previous studies, insufficient AMPK activation emerged as a potential key driver of impaired autophagic activity in BPD (13). In the neonatal hyperoxia-induced lung injury (nHILI) murine model of BPD (mBPD), an increase in phospho-AMPK (p-AMPK) concentrations in the lung was detected after hyperoxia exposure between P1 and P3, whereas with longer exposures between P1 and P5 and between P1 and P7, there were no differences in p-AMPK concentrations between control and mBPD groups. Autophagic activity at these time points demonstrated parallel alterations to p-AMPK concentrations, with an initial increase followed by a significant decrease at P7. Similarly, in nonhuman primate (NHP) lungs with BPD, p-AMPK concentrations demonstrated a positive correlation with autophagic activity. Furthermore, other groups demonstrated improved alveolar and vascular development in a rat model of BPD with metformin, an indirect activator of AMPK (17–19). However, the mechanisms underlying the beneficial effects of AMPK in BPD remain incompletely elucidated. Herein, we investigated the effect of 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR), a direct AMPK activator (20, 21), in a murine model of BPD induced by neonatal hyperoxia exposure with a focus on the role of autophagy in AMPK-induced macrophage responses.

Methods

Additional details on the methods used for making these measurements are provided in the data supplement.

Murine Model of nHILI

C57BL/6J (JAX stock no. 000664), Becn1^flox/flox (JAX stock no. 028794) (22), and LysM-cre (JAX stock no. 004781) mice (23) were obtained from The Jackson Laboratory. Becn1^flox/flox mice were crossed with LysM-cre mice to generate myeloid cell–specific autophagy-deficient Becn1-knockout mice (LysMCre⁺-Becn1^fl/fl). LysMCre⁻-Becn1^fl/fl mice served as littermate control animals. Neonatal mice were exposed to hyperoxia as previously described (13). Pups were euthanized at P7. All animal procedures performed were preapproved by the Animal Care and Use Committee at Brigham and Women’s Hospital.

AICAR Treatment

After hyperoxia (75% O₂) exposure from P1 to P7, the newborn mice received intraperitoneal injections of AICAR (PeproTech) between P3 and P6 at a dose of 500 mg/kg/day dissolved in sterile PBS. The vehicle control group received an equal volume of sterile PBS.

Harvesting and Processing of Murine Lung Tissues

After euthanasia, murine lungs were perfused with PBS through the right ventricle. Both lungs were removed, snap frozen in liquid nitrogen, and stored at −80°C. In other cohorts of mice, lungs were inflated with 10% neutral buffered formalin at a constant pressure of 25 cm H₂O and embedded in paraffin.

Morphometric Analysis

Paraffin-embedded lung sections were stained with hematoxylin and eosin for assessment of general architecture and measurements of radial alveolar counts (RACs) and mean linear intercepts (MLIs) as previously described (13, 24, 25). ACTA2-expressing myofibroblasts were quantified as a marker of secondary septation (26).

Immunofluorescence

Immunofluorescence staining was performed as previously described (13). The details of the antibodies used are provided in Tables E2 and E3 in the data supplement.

BAL, Isolation, and Adhesion Purification of AMs

BAL was performed on P7 pups for isolation and adhesion purification of AMs. For immunoblot studies, the lavages were combined as indicated, and the pellet was lysed using protein extraction buffer after centrifugation. For adhesion purification, the lavages from individual mice were centrifuged, and the pellet was resuspended in RPMI 1640 media. BAL cells were cultured on 24-well plates or chamber slides with RPMI 1640 media (Gibco, Thermo Fisher Scientific) for 1 hour. The attached BAL AMs were used for RNA isolation or immunofluorescence staining and analysis.

RNA Extraction, cDNA Synthesis, and Quantification of Gene Expression

Total RNA isolation from BAL AMs was performed using the RNeasy Mini Kit (Qiagen). Quantitative real-time PCR analysis was performed using the comparative cycle threshold method.

SDS-PAGE and Immunoblotting

Snap-frozen neonatal murine lungs were homogenized in radioimmunoprecipitation assay buffer (Boston Bio Products) supplemented with protease inhibitor cocktail (Roche) using a Dounce homogenizer. Protein at 30 μg/lane was separated by SDS-PAGE followed by immunoblotting as previously described (13). Sources and dilutions of antibodies used in immunoblotting analyses are listed in Table E2 and E3.

Statistical Analyses

Significant differences between experimental and control groups were assessed using unpaired two-tailed t tests or one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. All analyses were performed using GraphPad Prism 8.4.2 (GraphPad Software Inc.).

Results

AICAR Treatment Induces AMPK Activation and Autophagy in Neonatal Hyperoxia-exposed Lungs

In a recent study, we reported that hyperoxia exposure of neonatal murine lungs resulted in impaired autophagic activity in association with insufficient AMPK activation leading to impaired alveolarization and enhanced inflammatory signaling (13). Hence, in the present study, we first aimed to determine whether AMPK mimetic AICAR can lead to activation of AMPK in nHILI. AICAR is taken up by the adenosine transporters into the cells and phosphorylated by intracellular adenosine kinase to form an AMP analog, AICAR monophosphate, which binds to and activates AMPK (20, 21). Newborn mice were subjected to hyperoxia between P1 and P7 and received daily intraperitoneal AICAR injections between P3 and P6. After euthanasia on P7, whole-lung homogenates were prepared and used for immunoblotting and densitometric analysis. AICAR-treated mouse lungs demonstrated AMPK activation as determined by significantly increased p-AMPK to AMPK concentrations as compared with vehicle-treated lungs (Figures 1A and 1B). AMPK activation occurred in conjunction with significant increases in autophagy markers LC3-I, LC3-II, beclin1 and phospho-beclin1 and significantly reduced p62 (also known as SQSTM1) concentrations in AICAR-treated lungs (Figures 1A and 1B). In previous studies, we had found significantly increased proinflammatory gene expression in AMs isolated from BAL of autophagy-deficient Beclin1^+/− mice. To determine whether AICAR treatment induced AMPK activation and autophagy in BAL cells, we next analyzed p-AMPK and p62 concentrations and LC3-II/LC3-I ratio on pooled BAL cell lysates harvested from neonatal hyperoxia–exposed mice at P7. We detected significant increases in p-AMPK concentrations and LC3-II/LC3-I ratio in association with significantly decreased p62 concentrations in BAL cells harvested from AICAR-treated mice as compared with vehicle-treated littermates (Figures 1C and 1D). These findings indicated that AICAR treatment resulted in activation of AMPK and autophagy in neonatal hyperoxia–exposed lungs and airway immune cells.

Figure 1.

5-Aminoimidazole-4-carboxamide ribonucleotide (AICAR) treatment induced AMP-activated protein kinase (AMPK) activation and autophagy in neonatal hyperoxia–exposed lungs. (A) Representative immunoblot analysis for markers of AMPK activation and autophagy using whole-lung homogenates harvested at P7 from vehicle control (VC) and AICAR-treated hyperoxia-exposed C57BL/6J mice. (B) Relative protein concentrations normalized to β-actin (ACTB) by densitometry are indicated for all blots except for the AMPK blot, for which the phospho-AMPK (p-AMPK)/AMPK ratio is shown. (C) Representative immunoblot analysis for p-AMPK, p62, LC3-I, and LC3-II using pooled BAL cell lysates (n = 3/lane) harvested from VC and AICAR-treated hyperoxia-exposed mice at P7. (D) Relative protein concentrations normalized to ACTB are indicated for p-AMPK and p62 blots, and LC3-II/LC3-I ratio is indicated for the LC3 blot. Data are shown as mean ± SEM; n = 5 per group. *P < 0.05 and **P < 0.01. P = Postnatal Day.

AMPK Activation Attenuated Hyperoxia-induced Alveolar Simplification in Neonatal Mice

We recently reported that neonatal hyperoxia exposure in mice impairs autophagic activity in the lung, which then acts in synergy with hyperoxia to exacerbate alveolar simplification. Diminished autophagic activity was also evident in NHP lungs with BPD (13). Because insufficient AMPK activation appeared to be the main driver of impaired autophagic activity in both murine and NHP models of BPD, we next assessed whether AICAR treatment would attenuate alveolar simplification in nHILI. As expected, morphometric analysis of hematoxylin and eosin–stained lung sections demonstrated decreased alveolarization as indicated by significantly decreased RAC and significantly increased MLI in hyperoxia-exposed lungs as compared with normoxia controls (Figures 2A–2C). AICAR treatment did not lead to significant changes in RAC or MLI under normoxic conditions but resulted in significant increases in RAC (P < 0.01) and significant decreases in MLI (P < 0.05) in hyperoxia-exposed murine lungs as compared with vehicle treatment (Figures 2A–2C), indicative of improved alveolarization. To corroborate these results, we next assessed the density of ACTA2-positive myofibroblasts localized at septal tips as an additional surrogate marker of alveolarization (26) and found an increased number of ACTA2-positive myofibroblasts at the alveolar septal tips in AICAR-treated lungs as compared with vehicle-treated controls (Figures 2D and 2E) (P < 0.001). Because failure to expand the capillary network is a well-known feature of BPD (27, 28), we next quantified the expression of CD31, a pan-endothelial cell marker, in whole-lung homogenates by immunoblotting and found a significant increase in CD31 expression in AICAR-treated lungs as compared with controls (Figures 2F and 2G) (P < 0.01). Taken together, the results of these complementary analyses indicate that AICAR treatment attenuates nHILI by enhancing alveolar and vascular development.

Figure 2.

AICAR treatment attenuated alveolar simplification and enhanced vascular development in hyperoxia-exposed neonatal murine lungs. (A) C57BL/6J mice were exposed to normoxia (21% O₂) or hyperoxia (75% O₂) between P1 and P7 and treated with VC or AICAR from P3 to P6. After euthanasia at P7, lungs were inflated with 10% neutral buffered formalin and embedded in paraffin. Hematoxylin and eosin–stained sections are shown. Scale bars, 50 μm. (B) Radial alveolar counts were assessed as a measure of alveolarization. n = 6–8 per group. (C) Mean linear intercepts were measured as a surrogate for alveolar diameter. n = 5 per group. (D and E) ACTA2-positive myofibroblasts localized at septal tips (indicated by circles) as a surrogate marker of secondary septation were quantified using ImageJ and NIS-Elements Basic Research software. n = 5 per group. Scale bars, 50 μm. (F) Representative immunoblot analysis for the expression of CD31, a pan–endothelial cell marker, in whole-lung homogenates of VC or AICAR-treated C57BL/6J mice. (G) Densitometric analysis of relative protein concentration of CD31 normalized to ACTB is shown. n = 5 per group. Data were obtained from a minimum of three separate litters and are shown as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001. ns = not significant.

AMPK Activation by AICAR Ameliorates Inflammatory Responses in nHILI

We recently reported that autophagy-deficient Becn1^+/− mice exhibited significantly increased expression of proinflammatory genes in AMs upon exposure to neonatal hyperoxia (13). To determine the extent to which AMPK activation is involved as an upstream driver of inflammatory responses in nHILI, we first quantified total neutrophil and macrophage numbers in AICAR- versus vehicle-treated lungs. AICAR-treated lungs demonstrated significant decreases in both neutrophil and macrophage accumulation compared with vehicle-treated control lungs (Figures 3A–3D). Overall, ∼65% of immune cells in the airspaces were macrophages in both the vehicle- and AICAR-treated lungs. Next, we assessed mRNA expression levels of a panel of M1- and M2-like mediators in AMs by quantitative real-time PCR. Our results demonstrated significantly decreased expression levels of proinflammatory (M1-like) genes, TNF-α, IL-1β, and C-X-C motif ligand 1 (CXCL1), in association with significantly increased expression levels of antiinflammatory, proreparative (M2-like) genes, CD206 and Fizz1, in AMs isolated from AICAR-treated mice as compared with vehicle-treated control animals (Figure 3E). In accordance with previous reports, CD206 expression was detected ubiquitously in AMs isolated from BAL (Figure E1A) (29, 30). Mean fluorescence intensity measurements demonstrated increased CD206 expression in AICAR-treated AMs as compared with controls (Figures E1A and E1B). These findings were corroborated by detection of significantly higher CD206 protein concentrations in BAL cell lysates from AICAR-treated mice as compared with vehicle-treated control animals (Figure 3F). Furthermore, we detected increased concentrations of CD206 in whole-lung lysates from AICAR-treated mice (Figure 3G), suggesting that the effect of AICAR on CD206 expression was not unique to AMs but also extended to proreparative tissue-resident macrophages. Collectively, these results imply AICAR-induced AMPK activation and autophagy as a key driver of protective macrophage responses in nHILI.

Figure 3.

AICAR-induced AMPK activation ameliorates inflammatory responses in hyperoxia-exposed neonatal mice. Immunofluorescence staining was performed on paraffin-embedded sections at P7 from C57BL/6J neonatal murine lungs treated with VC or AICAR with primary antibodies against myeloperoxidase (MPO) and FABP4 to assess neutrophil (A) and macrophage (B) accumulation in lung tissues, respectively. The secondary antibody was Alexa Fluor 594 goat antirabbit IgG (red). White arrowheads indicate MPO-expressing neutrophils. Quantitative assessment of neutrophil (C) and macrophage (D) counts per HPF was performed using ImageJ and NIS-Elements Basic Research software. Representative images are shown. n = 3–5 per group. Scale bars, 50 μm. (E) BAL fluid was obtained from VC and AICAR-treated hyperoxia-exposed C57BL/6J neonatal mice at P7, and alveolar macrophages were isolated by adhesion purification. Total RNA was isolated, and quantitative real-time PCR was performed for a panel of M1-like (TNF-α, IL-1β, and CXCL1) and M2-like (CD206, Arg1, and Fizz1) genes as indicated on the figure. n = 5–8 per group. Representative immunoblot and densitometric analysis for CD206 protein expression in pooled BAL cell lysates (F) (n = 3/lane) and whole-lung homogenates (G) harvested from VC and AICAR-treated hyperoxia-exposed mice at P7. Relative protein concentrations were normalized to ACTB. n = 3–5 per group. Data were obtained from a minimum of three separate litters and are shown as mean ± SEM. *P < 0.05, ***P < 0.001, and ****P < 0.0001. HPF = high-power field. Arg1 = arginase1; Fizz1 = found in inflammatory zone 1.

Modulation of AM Gene Expression by AICAR Is Autophagy Dependent

To determine the role of autophagy in AMPK-driven protective responses in AMs in mBPD, we generated mice with myeloid cell–specific deficiency of Becn1 by crossing Becn1^F/F mice with LysM-cre transgenic mice (Becn1^F/F;LysM-cre). Immunoblotting of BAL cell lysates from adult Becn1^F/F;LysM-cre mice demonstrated a ∼90% decrease in Beclin1 protein concentrations compared with Becn1^F/F controls, thus indicating efficient deletion of Becn1 in AMs (Figures 4A and 4B). Given the significantly increased IL-1β concentrations found in Beclin1^+/− AMs, we examined the expression of the NLRP3 (NOD-, LRR- and pyrin domain-containing protein 3) inflammasome, a critical regulator of IL-1β production in macrophages, and found significantly increased NLRP3 concentrations in Becn1^F/F;LysM-cre mice compared with controls (Figures 4A and 4B). In line with our previous studies in Becn1^+/− mice, AMs isolated from neonatal hyperoxia–exposed Becn1^F/F;LysM-cre mice exhibited significantly increased mRNA expression levels of proinflammatory genes, TNF-α, IL-1β, and CXCL1. Furthermore, Becn1^F/F;LysM-cre AMs displayed significantly decreased concentrations of the M2-like markers, CD206, Arg1, and Fizz1, as compared with littermate controls (Figure 4C). Next, to determine the role of autophagy in the modulation of the AM phenotype in AICAR-treated neonatal hyperoxia–exposed mice, Becn1^F/F;LysM-cre and control Becn1^F/F mice were subjected to neonatal hyperoxia and treated with intraperitoneal AICAR injections as described in the Methods. AMs isolated from Becn1^F/F;LysM-cre at P7 demonstrated significantly higher mRNA expression levels of TNF-α, IL-1β, and CXCL1 in association with significantly decreased concentrations of Arg1 and Fizz1 as compared with control mice (Figure 4D). Furthermore, CD206 concentrations, which are significantly decreased by autophagy deficiency in AMs, were similar between the AICAR and control groups. These results indicate that the antiinflammatory effects of AMPK activation on AMs in neonatal hyperoxia–exposed mice are primarily mediated in an autophagy-dependent manner.

Figure 4.

Antiinflammatory effects of AICAR on alveolar macrophage responses in neonatal hyperoxia are autophagy dependent. (A) Representative immunoblot analysis for the expression of Beclin1 and NLRP3 in pooled BAL cell lysates harvested from 6-week-old Becn1^F/F;LysM-cre mice and Becn1^F/F littermates. n = 3/group. (B) Relative protein concentrations normalized to ACTB. (C) BAL fluid was obtained at P7 from Becn1^F/F;LysM-cre mice and Becn1^F/F littermates and (D) AICAR-treated Becn1^F/F;LysM-cre mice and Becn1^F/F littermates after neonatal hyperoxia exposure, and alveolar macrophages were isolated by adhesion purification. Total RNA was isolated, and quantitative real-time PCR was performed for a panel of M1-like (TNF-α, IL-1β, and CXCL1) and M2-like (CD206, Arg1, and Fizz1) genes as indicated on the figure. n = 7–17 per group. Data were obtained from a minimum of three separate litters and are shown as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Discussion

In this study, we found that administration of the AMPK mimetic AICAR to murine pups during neonatal hyperoxia exposure attenuates alveolar simplification and proinflammatory signaling in AMs. After dose optimization studies, AMPK activation induced by AICAR resulted in increased expression levels of autophagy markers in whole-lung and BAL cell lysates. Furthermore, AMs isolated from BAL from AICAR-treated mice demonstrated decreased mRNA expression of proinflammatory and increased mRNA expression of antiinflammatory genes. In an effort to better understand the role of autophagy in AICAR-driven macrophage responses, we generated mice with Beclin1-deficient AMs and found that AICAR treatment failed to reverse the proinflammatory signaling in these cells after exposure to neonatal hyperoxia. Collectively, these results demonstrate that AMPK activation plays a protective role in murine nHILI at least in part by modulation of AM responses via control of autophagic activity.

As a key cellular energy sensor, AMPK is activated in response to a variety of conditions that deplete cellular energy sources, including hyperoxia. In previous studies, we and others had found that AMPK activation was initially induced with short-term hyperoxia exposure up to 3–4 days but then declined after P7 in murine and rat lungs (13, 17). These observations have suggested that the promotion of AMPK activation could ameliorate the BPD phenotype that results from neonatal hyperoxia exposure. This hypothesis was initially explored by Chen and colleagues, who employed the antidiabetic agent metformin, an indirect activator of AMPK, in an experimental model of BPD induced by 100% O₂ exposure for 10 days in rat pups (18). This study demonstrated that metformin prolonged survival and reduced pulmonary inflammation, coagulation, and fibrosis but did not have an effect on alveolar development or BPD-associated pulmonary hypertension. In a more recent study, Yadav and colleagues employed metformin in a similar rat model of BPD and reported improved alveolar and vascular development in metformin-treated rat pups (17). In accordance with this study, we found that treatment with AICAR, a direct activator of AMPK, led to attenuation of alveolar simplification as demonstrated by three complementary methods. We also found evidence for increased capillary density as indicated by increased CD31 concentrations in AICAR-treated lungs. Given that AMPK subunits are ubiquitously expressed in pulmonary endothelial cells and endothelial AMPKα1 knockdown decreases angiogenesis both in vitro and in a murine BPD model (31), it is possible that AICAR could have a direct trophic effect on pulmonary microvasculature. However, because autophagic activity is primarily detected in AEC2s and AMs in newborn lungs, such an effect would be unlikely to be mediated by AICAR-induced autophagy in endothelial cells. Notably, although metformin is a widely used antidiabetic drug, AICAR has also been employed in several clinical trials (32), thus underscoring the translational potential of AMPK activation for BPD therapeutics.

The mechanism(s) by which AMPK activation improves alveolarization in nHILI is not fully elucidated. One potential mechanism for the beneficial effect of AMPK activation might be the preservation of mitochondrial function and cellular energy reserves as suggested by Yadav and colleagues, who reported increased concentrations of PGC-1α (peroxisome proliferator-activated receptor-γ coactivator-1α), a key transcriptional cofactor involved in mitochondrial biogenesis, in rat pups treated with metformin during hyperoxia exposure (17). However, we did not detect any significant changes in PGC-1α concentrations in whole-lung lysates between control and AICAR-treated pups in our model (Figure E2). Instead, our studies suggest activation of autophagy and downregulation of inflammatory signaling as potential mechanisms by which AMPK activation ameliorates nHILI. AMPK activation promotes autophagy by directly activating Ulk1, which initiates the forming of the autophagosome, and suppressing mTORC1 activity through phosphorylation (16). Previous studies using autophagy-deficient mice and mice treated with an inhibitor of RAPTOR, a component of mTORC1, suggested that autophagy plays an important role in the regulation of alveolarization in mBPD (11). Several studies have suggested a causal association between enhanced inflammatory signaling originating from macrophages and alveolar simplification in BPD (8, 33–35). AMPK activation has been reported to exert antiinflammatory effects in several acute as well as chronic lung disease models (36–38). The antiinflammatory effects of AMPK have been shown to be mediated through downregulation of NF-κB and PI3K/AKT (39). We now show that AICAR-mediated AMPK activation decreases the expression of M1-like proinflammatory genes, including IL-1β, in AMs exposed to neonatal hyperoxia. Our finding of increased NLRP3 expression in autophagy-deficient BAL cells reveals another potential mechanism by which the AMPK–autophagy axis can regulate IL-1β signaling in BPD. We also detected increased expression of CD206, a marker of M2-like, antiinflammatory macrophages, at both the mRNA and protein levels. These results were obtained in mice generated by Cre-lox recombination using the LysM promoter, which was believed to be myeloid cell specific (23, 40). However, after the generation of the Becn1^F/F;LysM-cre mouse line, we became aware of low-level expression of LysM in AEC2s (41, 42), which precluded the use of this novel mouse model to further investigate the role of autophagy and inflammatory responses of AMs in nHILI.

In summary, in the murine model of nHILI, AICAR treatment resulted in AMPK activation in association with increased autophagic activity. AICAR treatment significantly improved alveolarization and vascular density while decreasing the expression of proinflammatory (M1-like) genes and enhancing the expression of antiinflammatory (M2-like) genes in AMs. Beneficial modulation of the AM phenotype by AICAR was autophagy dependent. Further investigations are warranted to fully explore the therapeutic potential of the AMPK/autophagy pathway in BPD.