Myeloid Cell–Specific Deletion of AMPKα1 Worsens Ocular Bacterial Infection by Skewing Macrophage Phenotypes

Bacterial endophthalmitis is a serious eye infection that can lead to vision loss due to severe intraocular inflammation and tissue damage (1–4). Staphylococci, especially Staphylococcus aureus, are the primary culprits, and although they account for only 10–15% of cases, they often lead to poor outcomes (5). Typically, during ocular surgeries or trauma, S. aureus gains entry into the eye, releasing virulence factors such as toxins and cell wall components such as peptidoglycan and lipoteichoic acids (6). Host immune cells recognize these bacterial elements through pattern recognition receptors and trigger a complex inflammatory immune response (3, 7). Moreover, the blood–retinal barrier, which is crucial for ocular defense, becomes compromised, allowing a large influx of immune cells, predominantly neutrophils and monocytes/macrophages, into the vitreous chamber (8, 9). While this influx is necessary to combat bacterial growth, it also leads to excessive inflammation, which, if unchecked, damages intraocular tissues, especially the neurosensory retina, resulting in irreversible vision loss (2, 6, 10). Unfortunately, there is no standardized treatment to control intraocular inflammation, and visual outcomes often remain suboptimal even after aggressive antibiotic treatments and vitrectomy surgeries (11, 12). Moreover, the increase in antibiotic-resistant ocular pathogens poses a significant challenge in managing these infections (13, 14). Therefore, research aimed at understanding the mechanisms governing inflammation initiation and resolution holds promise for identifying potential therapeutic targets to prevent and treat bacterial endophthalmitis.

Given the intricate nature of host–pathogen interactions, we employed an integrated approach combining metabolomics and transcriptome analysis of S. aureus–infected retina tissues to identify novel immunomodulatory and antimicrobial targets (15–17). Our findings revealed significant impacts on various pathways during endophthalmitis, including inflammatory and immune responses, antimicrobial defense mechanisms against bacteria, endoplasmic stress, cell trafficking, and death, as well as energy metabolism, notably involving disruptions in NAD⁺ and lipid biosynthesis (1, 15, 16, 18). Within the realm of energy metabolism, AMP-activated protein kinase (AMPK) has emerged as a key player in maintaining host energy balance (19). AMPK not only governs cellular energy metabolism but also regulates immune responses by modulating immune signaling pathways (20–22). This cross-talk between AMPK and immune signaling pathways has crucial implications for reshaping the metabolic and functional profiles of immune cells (23–29).

In our previous work, we demonstrated that both genetic (using AMPKα1 global knockout [KO] mice) and pharmacological (with AICAR) activation of AMPK confer protection against endophthalmitis (1). However, a limitation of this study was the inability to discern the cell-specific contribution of AMPK to the pathogenesis of bacterial endophthalmitis. Given that the primary immune cells involved in endophthalmitis are of myeloid origin, we aimed to investigate the specific role of AMPKα1 in myeloid cells during endophthalmitis. In this study, employing myeloid-specific AMPKα1 KO mice (LysM^Cre;AMPK α1^fl) and bone marrow reconstitution chimera approaches, we found that the loss of AMPKα1 in myeloid cells worsens bacterial endophthalmitis, highlighting the importance of myeloid-specific AMPK in bacterial endophthalmitis.

B6.SJL-Ptprc^a Pepc^b/BoyJ (PepBoy) mice were obtained from The Jackson Laboratory (Bar Harbor, ME), while wild-type (WT) and AMPKα1 global KO mice (B6/129 background) were used as described (1). AMPKα1^fl mice were purchased from The Jackson Laboratory (no. 014141) and crossed with LysM^Cre mice (no. 004781) to generate a myeloid-specific loss of AMPKα1 referred to as LysM^Cre;AMPKα1^fl. Six- to 8-wk-old mice of both male and female sex were used. They were housed in a controlled-access Division of Laboratory Animal Resources facility at the Kresge Eye Institute, maintained under a 12-h light/12-h dark cycle at a temperature of 22°C, and provided ad libitum access to water and LabDiet rodent chow (PicoLab; LabDiet, St. Louis, MO). All procedures were conducted per the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee of Wayne State University (protocol no. IACUC-22-04-4557).

The S. aureus strain RN6390 was cultured in tryptic soy broth and on tryptic soy agar (TSA) plates (Sigma-Aldrich, St. Louis, MO). Bacterial endophthalmitis was induced in mice, as described previously (3, 30). Briefly, mice were anesthetized and intravitreally injected with S. aureus (500 CFU/eye) using a 34G needle under a dissecting microscope within a biosafety cabinet. Contralateral eyes received sterile PBS injections and served as controls. Disease progression was monitored through slit lamp examination and other modalities. Scotopic electroretinography (ERG) was used to determine retinal function using the Celeries ERG system (Diagnosis LLC, Lowell, MA), as per the manufacturer’s recommendations and standardized in our prior studies (16, 31).

Mouse bone marrow–derived macrophages (BMDMs) and neutrophils (BMDNs) were isolated according to previously established protocols (32). Briefly, mice were euthanized, and bone marrow was flushed from the tibias and femurs using RPMI 1640 media supplemented with 10% FBS and 0.2 mM EDTA. The cells were then centrifuged at 400 × g for 5 min at 4°C to pellet them. RBCs were lysed by adding 0.2% NaCl (hypotonic) solution for 20 s, followed by the addition of 1.6% NaCl (hypertonic) and subsequent centrifugation. The cell pellets were rinsed with RPMI 1640 media, resuspended, seeded in RPMI 1640 media supplemented with 10% FBS, 100 U/ml penicillin, 100 mg/ml streptomycin, and 10 ng/ml M-CSF for macrophage differentiation, and cultured at 37°C in 5% CO₂. At 6 d post differentiation, BMDMs were seeded at a concentration of 4 × 10⁶ cells/ml in 60-mm petri dishes for in vitro experiments.

For BMDN isolation, bone marrow cell pellets were resuspended in 1 ml of PBS and overlaid on a Histopaque gradient of 1119 and 1077 (Sigma-Aldrich, St. Louis, MO), followed by centrifugation at 600 × g at 25°C for 30 min without centrifuge break. The layer between Histopaques 1119 and 1077 was collected and washed twice with RPMI 1640 media by centrifugation at 400 × g for 7 min, and BMDNs were cultured as described for macrophages. For in vitro infections, cells were challenged with S. aureus RN6390 at a multiplicity of infection (MOI) of 10:1 for the indicated time point.

The eyes were enucleated following S. aureus infection at 24, 48, and 72 h postinfection. Whole-eye lysates were prepared in 250 μl of PBS by beating with stainless steel beads in a TissueLyser (Qiagen, Valencia, CA). A 50-μl aliquot of the tissue homogenate was serially diluted and plated on TSA to quantify the viable bacterial load. The colonies were counted the next day, and the results are reported as the mean number of CFU/eye.

For in vivo samples, eyes were enucleated, and lysates were prepared using a TissueLyser as described above. The tissue homogenates were then centrifuged at 10,000 × g for 20 min at 4°C, and the resulting supernatant was used for cytokine estimation. In the in vitro studies, conditioned media from both the S. aureus–infected and the vehicle control group, were collected and preserved for cytokine measurements. Mouse ELISA was conducted to quantify the inflammatory cytokines IL-1β, IL-6, TNF-α, and CXCL2 according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN).

The eyes were enucleated and preserved in modified Davidson’s fixative for subsequent histopathological analysis. The embedding, sectioning, and staining with H&E were carried out by Excalibur Pathology (Oklahoma City, OK).

Total RNA was extracted from both the neural retina and BMDMs using TRIzol according to the manufacturer’s guidelines (Thermo Fisher Scientific, Rockford, IL). Subsequently, the RNA was reverse transcribed using a Maxima first-strand cDNA synthesis kit following the manufacturer’s protocol (Thermo Fisher Scientific, Rockford, IL). Quantitative assessment of gene expression was conducted via SYBR Green-based quantitative PCR (qPCR), employing gene-specific primers, on the StepOnePlus real-time PCR system (Applied Biosystems, Foster City, CA). Data analysis was performed utilizing the 2^−ΔΔCt method, as previously described (33). The expression levels of genes in the test samples were normalized to endogenous β-actin controls.

BMDM cells (10⁶ cells/dish) were cultured in 60-mm petri dishes in RPMI 1640 media. Subsequently, the cells were exposed to S. aureus at an MOI of 10:1 for 2 h. After the incubation period, the cells were washed and treated with gentamicin (200 μg/ml) for 2 h to eliminate extracellular bacteria. The absence of extracellular bacteria was confirmed through dilution plating and enumeration of CFU. At 2 h after gentamicin treatment, the cells were washed with RPMI 1640 and incubated with fresh media containing gentamicin (200 μg/ml) for 1, 3, or 6 h. To quantify internalized bacteria, following the designated incubation periods, the cells were washed three times with PBS and lysed using 0.01% Triton X-100. The lysed cells were collected by scraping and centrifuged at 7500 × g for 5 min. The resulting cell pellets were washed twice with PBS through centrifugation at 7500 × g for 5 min each time. The pellets were then resuspended in 1 ml of sterile PBS, serially diluted, and plated onto TSA plates for enumeration of bacterial counts.

BMDMs were cultured in a four-well chamber slide (Thermo Fisher Scientific, Rochester, NY). Following treatment, the cells were infected with S. aureus at an MOI of 10:1 for 3 h. At the specified time points, the infected cells were fixed with 4% PFA in PBS at 4°C. After washing with PBS, the cells were permeabilized and blocked with 1% (w/v) BSA and 0.4% Triton X-100 in PBS for 1 h at room temperature. Subsequently, the cells were incubated overnight at 4°C with primary Abs against inducible NO synthase (iNOS) and arginase-1 (Arg1; 1:100 dilution, Cell Signaling Technology, Danvers, MA). Following the removal of the primary Abs, the cells were thoroughly washed with PBS and then incubated with anti-mouse/rabbit Alexa Fluor 485/594–conjugated secondary Abs (1:200 dilution) for 1 h at room temperature. Finally, the cells were washed four times (5 min each) with PBS, and the slides were mounted in Vectashield antifade mounting medium containing DAPI (Vector Laboratories, Burlingame, CA) before visualization using a Keyence microscope (Keyence, Itasca, IL).

Protein samples were obtained by directly lysing cells in radioimmunoprecipitation assay (RIPA) buffer containing protease and phosphatase inhibitor cocktails. The total protein concentration of the cell lysates was determined using a Micro BCA protein assay kit (Thermo Fisher Scientific, Rockford, IL). Subsequently, protein samples (30–40 µg) were separated via SDS-PAGE (8 or 12%) and transferred onto a nitrocellulose membrane (0.45 μm) using a wet transfer system. The membranes were then blocked in 5% (w/v) dry milk in TBST for 1 h at room temperature. Following blocking, the membranes were incubated with primary Abs against iNOS, Arg1, and heat shock protein (HSP)-90 (Cell Signaling Technology, Danvers, MA) according to the manufacturer’s instructions, followed by incubation with appropriate HRP-conjugated secondary Abs. After four TBST washes, the membranes were exposed to SuperSignal West Femto chemiluminescent substrate and imaged using an iBright fluorescence imager (Thermo Fisher Scientific, Rockford, IL). Densitometry analysis was carried out using ImageJ software (National Institutes of Health, Bethesda, MD).

Flow cytometry was conducted following previously established methods to assess the adoptive transfer and infiltration of neutrophils into the eyes (3). In brief, retinas from euthanized mice were isolated and digested with Accumax (Millipore) for 10 min at 37°C. A single-cell suspension was generated by gently triturating the cell mixture with a 22G needle and subsequently passing it through a 40-μm cell strainer (BD Falcon, San Jose, CA). The cells were then blocked using Fc Block (BD Biosciences) for 30 min, followed by three washes with PBS containing 1% BSA. Subsequently, the cells were stained by incubating them with CD45.1-allophycocyanin–, CD45.2-PE–, and Ly6G-FITC–conjugated Abs (BD Biosciences) on ice for 20 min in the dark. After incubation, the cells were washed and suspended in PBS supplemented with 1% BSA. Intracellular AMPK staining was performed in the macrophages and neutrophils using the polyclonal PE-conjugated p-AMPK (p-AMPKα1-Thr¹⁷²) (Bioss, Woburn, MA). Briefly, the cells were fixed with an eBioscience intracellular fixation and permeabilization buffer set (Thermo Fisher Scientific, Rockford, IL) and stained for 30–45 min at 4°C and then washed before acquisition. Flow cytometric analysis was performed using an Accuri C6 flow cytometer and accompanying software (BD Biosciences), and data analysis was conducted using FlowJo v8 software (BD Life Sciences).

The adoptive transfer of naive bone marrow cells into AMPKα1 KO mice (with the CD45.2 allele) and B6.SJL-Ptprc^a Pepc^b/BoyJ mice (PepBoy, used as WT mice with the CD45.1 allele) was carried out via the busulfan method (34). Initially, recipient mice received a single preconditioning dose of 20 mg/kg busulfan (Cayman Chemicals, Ann Arbor, MI) i.p. daily until a cumulative dose of 100 mg/kg was achieved. Subsequently, all mice were i.v. injected with 1 × 10⁷ donor bone marrow cells obtained from the recipient mouse via tail vein injection in 200 μl of PBS. Starting at 2 wk after transplantation, ∼20–30 µl of blood was collected from the lateral saphenous vein, and CD45.1 versus CD45.2 expression was assessed in peripheral blood using flow cytometry. Efficient engraftment of donor bone marrow (>80%) was observed at 6 wk posttransplantation. Following successful engraftment, the mice were subjected to endophthalmitis as described above.

For RNA sequencing (RNA-seq), total RNA was extracted from the BMDMs of the AMPKα1 KO and WT mice in biological triplicates using RNeasy mini kits (Qiagen, Hilden, Germany). The sequencing procedures were conducted by Novogene. RNA-seq libraries were constructed from high-quality RNA using NEBNext Ultra II RNA library prep kits (New England Biolabs, Ipswich, MA) following the manufacturer’s instructions. Sequencing was performed on the NovaSeq 6000 platform as paired-end 150-bp reads (Illumina), with a sequencing depth of ∼30 million reads per sample. Differential gene expression analyses were conducted using DESeq2, while Gene Ontology enrichment analysis was carried out using PANTHER and KEGG pathways analyses. The RNA-seq data are deposited in the National Institutes of Health Sequence Read Archive with the accession number PRJNA1143903 (https://dataview.ncbi.nlm.nih.gov/object/PRJNA1143903?reviewer=5315cto6crratth096nfc5d4eg).

Statistical analyses were conducted using Prism version 9.2 (GraphPad Software, San Diego, CA). Group comparisons were assessed using one-way ANOVA followed by Tukey multiple comparison post hoc tests, and data are presented as mean ± SD. A confidence interval of 95% was applied for all statistical tests. A p value <0.05 was considered statistically significant. Unless specified otherwise, all experiments were performed at least three times.

All relevant data are within the manuscript, figures, and supporting data files.

The hallmark of bacterial endophthalmitis is the infiltration of myeloid cells such as neutrophils and monocytes into the eye (6). To investigate the role of AMPK in these cells during ocular infections, we induced S. aureus endophthalmitis in both AMPKα1^fl (WT) and LysM^Cre;AMPKα1^fl mice using a low-dose infection model, in which 500 CFU of S. aureus were administered. Although this dosage resolved in WT mice, it led to severe endophthalmitis in LysM^Cre;AMPKα1^fl mice (3, 16). We monitored the progression of endophthalmitis through daily eye examinations, retinal function testing, histopathological analysis, and estimation of intraocular bacterial burden and inflammation. Compared to AMPKα1^fl mice, endophthalmitis worsened in LysM^Cre;AMPKα1^fl mice, indicated by a time-dependent increase in corneal haze, opacity, and hypopyon formation (Fig. 1A). Histopathological analysis revealed increased retinal folding, tissue disintegration, and significant cellular infiltrates in the vitreous chamber of LysM^Cre;AMPKα1^fl mouse eyes (Fig. 1B). ERG demonstrated a notable decrease in both a- and b-wave amplitudes in LysM^Cre;AMPKα1^fl eyes at various time points (Fig. 1C). Additionally, the bacterial burden was relatively greater in LysM^Cre;AMPKα1^fl mice (Fig. 1D). Overall, a greater percentage of eyes exhibited damage due to S. aureus endophthalmitis in LysM^Cre;AMPKα1^fl mice compared with AMPKα1^fl mice (Fig. 1E).

Furthermore, in addition to direct retinal damage caused by bacterial toxins, uncontrolled inflammation resulting from excessive infiltration and activation of innate immune cells such as neutrophils contributes to bystander tissue damage. To evaluate intraocular inflammation, we analyzed the expression of inflammatory cytokines at both the mRNA (Fig. 2A) and protein (Fig. 2B) levels at various time points. Notably, myeloid-specific deletion of AMPKα1 led to higher expression of inflammatory cytokines. For instance, the transcript levels of Il-6 and Cxcl2 were elevated at 72 h, while the protein levels increased at both 48 and 72 h. Conversely, AMPKα1-deficient macrophages exhibited consistently increased transcript levels of the inflammatory cytokines IL-1β and TNF-α at all time points, with protein levels peaking at 48 or 72 h. Additionally, we investigated the inflammatory response of BMDMs from both AMPKα1^fl and LysM^Cre;AMPKα1^fl mice, revealing increased levels of inflammatory mediators in AMPKα1 KO BMDMs, consistent with the in vivo data (Supplemental Fig. 1). Taken together, these findings suggest that AMPKα1 deficiency in myeloid cells enhances inflammation and exacerbates endophthalmitis.

To elucidate the mechanisms underlying the heightened inflammatory response of S. aureus–infected AMPKα1-deficient mice, we performed RNA-seq analysis in the BMDMs, which express relatively higher levels of AMPK (Supplemental Fig. 2). This analysis revealed ∼3516 differentially expressed genes (DEGs) in S. aureus–infected AMPKα1 KO BMDMs compared with WT, with 1665 genes showing upregulation and 1848 displaying downregulation (Fig. 3A). Gene set enrichment analysis indicated significant alterations in pathways associated with inflammation, metabolism, endoplasmic reticulum stress, and cholesterol biosynthesis, with notable downregulation of bacterial defense pathways in AMPKα1-deficient macrophages (Fig. 3B). Further analysis of DEGs based on the encoded proteins corroborated these findings, that is, a reduction in pathways related to inflammation resolution and responses to LPS and bacteria (Fig. 3C). A heat map was constructed to visualize the highly downregulated and upregulated genes involved in inflammatory and antimicrobial immune defense (Fig. 3D).

Given the increased bacterial burden observed in S. aureus–infected LysM^Cre;AMPKα1^fl mice, we hypothesized that this might stem from the downregulation of antimicrobial genes such as S100a7, S100a8, and S100a10 in the RNA-seq data. To verify this, we examined their expression via qPCR and found that their levels were significantly lower in AMPKα1 KO BMDMs (Supplemental Fig. 3A). Consistently, LysM^Cre;AMPKα1^fl BMDMs exhibited reduced intracellular phagocytic killing of S. aureus in the gentamicin protection assay (Supplemental Fig. 3B). Collectively, these findings underscore the role of AMPKα1 in regulating antibacterial activity by enhancing phagocytic killing and the induction of antimicrobial molecules during bacterial infection.

Our RNA-seq analysis revealed that a diminished response to bacterial infection corresponds with the exacerbated endophthalmitis observed in AMPKα1-deficient mice. Given the pivotal role of macrophages in inflammation resolution via the transition from a proinflammatory (M1) to an anti-inflammatory (M2) phenotype, we examined the expression of these markers in infected retinal tissue and cultured BMDMs. Our data demonstrated relatively higher transcript levels of the classical M1 markers Cox2, iNos, and Il23 in the retinas of LysM^Cre;AMPKα1^fl mice compared with WT mice (Fig. 4A). Conversely, the levels of M2 phenotype markers (Arg1, Fizz1, and Yam1/2) were downregulated in retinas of LysM^Cre;AMPKα1^fl mice. To corroborate these findings, we assessed the proinflammatory (M1) and anti-inflammatory (M2) signatures in BMDMs isolated from LysM^Cre;AMPKα1^fl and AMPKα1^fl mice, challenging them with S. aureus. Immunostaining of S. aureus–infected LysM^Cre;AMPKα1^fl macrophages revealed increased levels of iNOS (Fig. 4C) and decreased expression of Arg1 (Fig. 4D). Similarly, Western blot analysis also demonstrated increased protein levels of iNOS (Fig. 4E, 4F) and decreased Arg1 levels (Fig. 4G, 4H) in AMPKα1-deficient macrophages. These results further support the notion that AMPKα1 regulates the balance between the M1 and M2 phenotypes, whereby its deficiency promotes the inflammatory phenotype in the eye.

To illustrate the protective function of myeloid AMPKα1 during endophthalmitis, we used a bone marrow chimera approach, in which bone marrow from donor PepBoy mice (CD45.1 allele) were transferred to recipient AMPKα1-deficient mice (CD45.2 allele) (WT → KO) and vice versa (KO → WT) (Fig. 5A, schematic). This adoptive transfer yielded >80% efficiency of bone marrow transfer in the recipient mice. Once stable bone marrow chimeras were established, S. aureus endophthalmitis was induced, and a time-course study was conducted. Eye imaging revealed decreased opacity and corneal haze in the bone marrow of AMPKα1 KO mice with WT (PepBoy) bone marrow compared with those with AMPKα1 KO bone marrow in WT mice (Fig. 5B). Electroretinography analysis demonstrated better retention of a- and b-wave amplitudes in AMPKα1 KO mice with WT bone marrow (Fig. 5C). Furthermore, transfer of WT chimera in AMPKα1 KO mice mitigated disease severity, as evidenced by decreased corneal opacity (Fig. 5D) and reduced intraocular bacterial burden (Fig. 5E).

Subsequently, we assessed intraocular inflammation and found that the WT → KO chimera exhibited significantly lower levels of inflammatory markers such as IL-1β, IL-6, TNF-α, and CXCL2 (Fig. 5F). This reduction correlated with a decrease in neutrophil infiltration in AMPKα1 KO chimera (Fig. 5G) compared with PepBoy mice with AMPKα1-deficient bone marrow (Fig. 5H). Thus, myeloid-specific AMPKα1 contributes to inflammation resolution and mitigates bacterial endophthalmitis (Fig. 6).

Bacterial endophthalmitis remains the most common vision-threatening complication of eye surgeries and ocular trauma (4). Owing to the rapid progression and poor prognosis of endophthalmitis, early treatments with intravitreal injections of empirical antibiotics remain the main treatment option for bacterial endophthalmitis (35). Antibiotics, while destroying bacteria, may release lipopolysaccharides (36, 37), lipoteichoic acid, and peptidoglycan from bacterial cell walls, including those of S. aureus (37), and evoke an inflammatory response (2, 38). To suppress inflammation, intravitreal injections of corticosteroids are recommended, but the usage of steroids along with antibiotics for endophthalmitis has been marred with controversial outcomes in humans (39–43) and animal studies (44–48). Moreover, steroid-related side effects, including elevated intraocular pressure, cataract formation, and retinal detachment, further hamper their inclusion in the management of endophthalmitis (49). Thus, there is an urgent need to develop nonimmunosuppressive anti-inflammatory therapies. In this study, we demonstrated an anti-inflammatory and protective role of myeloid-specific AMPK in bacterial endophthalmitis.

AMPK is an evolutionarily conserved heterotrimeric kinase with α, β, and γ subunits that primarily regulate cellular energy homeostasis (23). The AMPK complex is activated either by increases in cellular AMP or ADP or by phosphorylation on Thr¹⁷² by upstream kinases, such as CaMKK-β, LKB1, or TAK1 (50–55). This enhanced activity, in turn, controls signaling pathways that regulate protein synthesis, cell division, and intracellular membrane trafficking. Recent studies on AMPK regulation indicate that it not only functions as an intracellular energy sensor but also functions as a stress sensor to maintain intracellular homeostasis during infectious, autoimmune diseases and skeleton muscle regeneration (56–58). Previously, using global AMPKα1 KO mice, we reported that loss of AMPKα1 increased the severity of bacterial endophthalmitis. However, the cell-specific role of AMPKα1 was not determined. In this study, our data showed that myeloid cell–specific AMPKα1 plays a protective role during endophthalmitis as evidenced by worse disease outcomes in LysM^Cre;AMPKα1^fl mice (Fig. 6). Moreover, the protective effects of AMPKα1 on myeloid cells are exhibited via several potential mechanisms.

In this study, we observed that the bacterial burden was greater in myeloid-specific AMPKα1-deficient mice. As we reported that activation of AMPKα and its downstream ACC in retinal Müller glia regulate autophagy-mediated clearance of S. aureus (17), we hypothesized that lack of AMPKα will impair pathogen clearance. Notably, phagocytosis of the pathogen activates the host autophagy initiation complex and the upstream regulatory components LKB1 and AMPKα1, which regulate autophagy induction through their kinase activities (59). Deletion of AMPKα1 in myeloid cells results in susceptibility to exacerbated bacterial infection and host tissue cell death. In support of these findings, our data showed that BMDMs from LysM^Cre;AMPKα1^fl mice had reduced phagocytic killing of S. aureus. Moreover, AMPKα1 deficiency led to a significant decrease in the bacterial infection-induced expression of antimicrobial peptides, which might also contribute to greater bacterial burden in the eyes. Our results corroborate previous studies reporting the increased phagocytic activity (Escherichia coli and Legionella pneumophila) of macrophages upon pharmacological activation of AMPK via AICAR or metformin treatment (60, 61). Moreover, the defect in phagocytosis following AMPKα1 ablation in macrophages is associated with their inability to undergo phenotypic transition, explaining the defect in bacterial clearance. A recent study has shown that restoring AMPK activity accelerates recovery from bacterial pneumonia (62). Similarly, we also reported that AMPK activation inhibits viral replication in retinal cells (63). These studies highlight the multifaceted roles of AMPK in viral and bacterial infections (64) through the regulation of cellular metabolism (65, 66).

Macrophages are also highly plastic innate immune cells, and environmental stimuli or bacterial infection can generate macrophages with a range of different phenotypes and functions (67). The classical activation of macrophages generates M1 macrophages with proinflammatory properties, whereas alternative activation generates M2 macrophages with anti-inflammatory properties (68). M1 macrophages exacerbate tissue injury whereas M2 macrophages promote the resolution of inflammation upon tissue injury (68, 69). A careful examination of LysM^Cre;AMPKα1^fl revealed impaired inflammation resolution, leading to a subtle but significant increase in the expression of M1 macrophage phenotype markers (iNos and Il23 expression) with the endophthalmitis progression. In contrast, the same mice showed a decrease in M2 macrophage phenotype markers (Arg1 and Fizz1), corresponding to an anti-inflammatory state, resulting in accelerated inflammation and endophthalmitis. Interestingly, a previous study showed that AMPKα1 mediated the regulation of macrophage skewing at the time of inflammation resolution during skeletal muscle regeneration (58, 70). Our data showing higher levels of inflammatory mediators in LysM^Cre;AMPKα1^fl mice eyes indicate the predominance of the M1 milieu, which results in inflammation-mediated retinal tissue damage.

To further confirm the role of AMPKα1 in ocular infection, we generated bone marrow chimeras and showed the ability of AMPKα1-competent bone marrow–derived cells to rescue AMPKα1 KO from severe endophthalmitis. Consequently, impaired inflammation resolution was observed in KO → WT chimera where WT macrophages were replaced with macrophages lacking AMPKα1 expression, whereas a benefit was observed in WT → KO chimera expressing AMPKα1 in macrophages. Moreover, the extent of endophthalmitis resolution in the AMPKα1 KO recipients was similar to those in WT mice. This indicated that AMPKα1 expressed in infiltrating/residential myeloid cells contributes to endophthalmitis resolution. Among myeloid cells, our data showed that macrophages had relatively higher levels of AMPK compared with neutrophils, suggesting their role in inflammation resolution during endophthalmitis. However, in KO → WT chimera there was an increased neutrophil infiltration, and our in vitro data showed that AMPK-deficient macrophages produce higher levels of chemokines such as CXCL2. Therefore, we postulate that higher levels of chemokines in KO → WT chimera led to increased neutrophil infiltration and more disease severity. Besides macrophages, the activation and production of chemokines by retinal resident immune cells such as microglia also promote neutrophil infiltration during endophthalmitis (71). Moreover, neutrophils are also reported to have N1 (inflammatory) and N2 (anti-inflammatory) phenotypes (72, 73), and AMPK deficiency could skew their phenotype toward N1, leading to increased disease severity, which needs further investigation. Despite the extensive use of LysM-Cre mice to study myeloid cells in various diseases (58, 74, 75), we acknowledge that it does not allow the distinction among myeloid-derived cell types, including monocytes/macrophages, polymorphonuclear neutrophils, and microglia (retinal-resident macrophages). Therefore, complementary approaches such as macrophage or neutrophil depletion in chimeric mice could decipher their specific roles in the pathogenesis of bacterial endophthalmitis.

In summary, our study demonstrated the protective role of myeloid cell–specific AMPKα1 during bacterial endophthalmitis. These effects are mediated by enhanced phagocytic clearance of bacteria and the promotion of inflammation resolution via inducing M2 macrophage phenotype. Therefore, AMPKα1 activation can be targeted therapeutically to treat bacterial endophthalmitis.

The authors have no financial conflicts of interest.

We are grateful to other laboratory members for helpful discussions and editing of the final manuscript.