A complete list of the reagents, cells, and animals and their commercial sources is provided ( Supplemental Table 1 ). Briefly, a clinical isolate of nontypeable Haemophilus influenzae (NTHi) was provided by Dr. Hong Wei Chu (National Jewish Health). Moraxella catarrhalis (M. catarrhalis), Streptococcus pneumoniae (S. pneumoniae), and Pseudomonas aeruginosa (P. aeruginosa) strain PAO1 were obtained from ATCC. LPS from P. aeruginosa was obtained from Sigma-Aldrich. Chocolate agar and sheep blood agar plates, brain heart infusion broth, and Todd-Hewitt broth (THB) were obtained from BD Biosciences. Tryptic soy broth (TSB), tryptic soy agar, and hemin were obtained from MP Biomedicals. β-Nicotinamide adenine dinucleotide hydrate, gentamicin, and DMEM for the biofilm assay were obtained from Sigma-Aldrich. Phloretin was obtained from Cayman Chemicals. Crystal violet stain was obtained from Fisher Scientific. RAW 264.7 and NCI-H292 cells were obtained from ATCC. Normal and COPD human bronchial epithelial (HBE) cells were obtained from MatTek.

NTHi and M. catarrhalis bacterial strains were grown overnight on chocolate agar plates and S. pneumoniae was grown on sheep blood agar plates (37°C, 15h, 5% CO₂). PAO1 was grown in TSB (37°C, 15h, with shaking). A portion was subcultured in fresh TSB (37°C, 2h, with shaking), centrifuged (4000 RCF, 5 min), and washed in PBS. Bacterial concentrations were determined by optical density (OD₆₀₀) and were confirmed by serial dilutions on the appropriate agar plates. A detailed table of bacterial growth conditions for each assay is provided ( Supplemental Table 2 ).

The antibacterial activity of phloretin was examined in a cell-free system by incubating multiple doses of phloretin with bacteria in 96-well plates. Vehicle, phloretin (0.1-1 mM, diluted in PBS), or gentamicin (100 µg/mL) was incubated with control (appropriate bacterial diluent), NTHi (10⁶ colony forming units (CFU), in brain heart infusion broth supplemented with hemin and β-nicotinamide adenine dinucleotide hydrate (sBHI)), M. catarrhalis (10⁵ CFU, in THB), S. pneumoniae (5x10⁵ CFU, in THB supplemented with yeast extract), or PAO1 (10⁵ CFU, in 20% TSB) at a 1:1 v/v. Microtitre plate lids were coated with 0.5% Triton X-100 in ethanol to prevent condensation. Plates were incubated (37°C, 18 h) and the OD₅₇₀ was determined hourly using a plate reader (BioTek). Prior to a reading, the plate was shaken for 10 min by the plate reader. Following the assays, the bacteria mixtures were serially diluted and plated on their corresponding agar plates to manually count the CFU. Two to three independent trials were conducted with each condition performed in triplicate for each assay (n=6-9 total wells per condition).

Quantification of crystal violet stained biofilm was determined using a microtiter plate assay, as modified by O’Toole (O’Toole, 2011). Vehicle (0.1% DMSO), 100 µM phloretin, or 100 µg/mL gentamicin in PBS was added to control (PBS), NTHi (10⁶ CFU in sBHI), M. catarrhalis (10⁶ CFU in THB), S. pneumoniae (10⁵ CFU in THB), or PAO1 (10⁶ CFU in DMEM) at a 1:1 v/v in 96-well plates. Plates were incubated under the same conditions (37°C, 16 h, 5% CO₂) except for PAO1, which was incubated without CO₂. Cells were stained with 0.5% crystal violet (22°C, 20 min with shaking). Plates were dried (22°C, 16 h) and bound dye was released from the cells with methanol and quantified by measuring the OD₅₇₀. Two independent trials were conducted with each condition performed in triplicate for each assay (n=6 total per condition).

NCI-H292 cells (a human pulmonary epithelial cell line) or RAW 264.7 cells (a mouse macrophage cell line) were grown in 75 cm² tissue culture flasks in RPMI-1640 (ATCC) or DMEM (Thermo Fisher Scientific), respectively, supplemented with 10% FBS and penicillin (100 U/mL) and streptomycin (100 µg/mL). HBE cells derived from normal and COPD donors were seeded onto collagen-coated T75 flasks and grown in bronchial epithelial growth medium (Lifeline Cell Technology). For experiments, cells were seeded into 12-well plates (NCI-H292 cells: 5,000 cells/cm²; HBE cells: 5,000 cells/cm²), or 96-well plates (RAW 264.7 cells: 312,500/cm²). Upon confluency, medium was replaced with supplement-free medium prior to experiments (18h). All cultures were maintained at 37°C in 5% CO₂/95% air.

The antibacterial activity of phloretin against NTHi was measured in the presence of NCI-H292 cells using an assay adapted from previous studies (Morey et al., 2011; Euba et al., 2017). Cells were pretreated for 1 h with 0.1 mM phloretin and exposed to NTHi (3x10⁴ CFU/mL, 500 µL) diluted in 20% sBHI in RPMI. After 4 h, the cells were washed twice with PBS and lysed with 0.1% w/v saponin (Alfa Aesar); this was considered the adherent and intracellular collection. Another collection of treated cells were washed twice with PBS and treated with 200 µg/mL gentamicin (1 h) prior to lysing; this was considered the intracellular collection. Two independent trials were conducted with biological replicates performed in triplicate for each assay (n=6 total). Each replicate was serially diluted and plated on chocolate agar and incubated overnight (37°C, 15 h, 5% CO₂). The CFU was enumerated to compare the number of bacteria in each collection.

The anti-inflammatory activity of phloretin against the COPD pathogens in RAW 264.7 cells was measured. Cells were pretreated 1 h with vehicle (0.1% DMSO) or 0.1 mM phloretin and exposed to NTHi (2x10⁵ CFU/mL, 50 µL), M. catarrhalis (2x10⁵ CFU/mL, 50 µL), S. pneumoniae (2x10⁵ CFU/mL, 50 µL), or PAO1 (2x10⁴ CFU/mL, 50 µL) for 2h. Tumor necrosis factor (TNF) was measured in the supernatant of cells by ELISA (BioLegend). Two independent trials were conducted with biological replicates performed in triplicate (n=6 total) and 2 technical replicates for the ELISA.

Normal and COPD HBE cells were pretreated with 0.1 mM phloretin (1 h) and exposed to 500 µL 1.5x10⁷ CFU/mL of NTHi. RNA was extracted from HBE cells after 2 h with TRI Reagent (Sigma-Aldrich) and reverse transcribed into first-strand cDNA (Bio-Rad). RT-qPCR was performed using TaqMan gene expression master mix (Life Technologies) and primers with a thermal cycler (Applied Biosystems) under the following conditions: 50°C (2 min), 95°C (10 min), and 40 cycles of 95°C (15 sec) and 60°C (1 min). Interleukin (IL) 6, IL8, C-C motif chemokine ligand 20 (CCL20), and ribosomal protein (RPL32) transcripts were analyzed. Relative gene expression was calculated using the 2^-ΔΔCT method normalized to RPL32. Two independent trials were conducted with biological replicates performed in triplicate (n=6 total), each with 3 technical replicates for the RT-qPCR assays.

RAW 264.7 cells and HBE cells were pretreated with vehicle (0.1% DMSO) or 0.1 mM phloretin (1 h) and exposed to medium control or bacteria. Lactate dehydrogenase activity was measured using CytoTox-ONE reagent (Promega). Two independent trials were conducted with biological replicates performed in triplicate (n=6 total), with no technical replicates.

Animal studies were conducted with approval of the Institutional Animal Use and Care Committee of the University of Pittsburgh. FVB/NJ mice (Jackson Laboratories, female and male, 8 wk) were housed under pathogen-free conditions and maintained on an AIN-93G diet (Envigo) for 3 days prior to the start of the experiment. Mice were then fed an AIN-93G standard diet without or with the addition of 0.157% phloretin ad libitum for 1 week. Diet was administered within 2 weeks upon receipt from the manufacturer to maintain freshness and efficacy. Single colonies of NTHi were incubated in sBHI (37°C, 1.5 h, with shaking). The culture was centrifuged (4000 RCF), washed twice in PBS, and adjusted to an OD₆₀₀ of 2. Mice were inoculated intratracheally with PBS or 10⁵ CFU of NTHi. The dose was confirmed by plating serial dilutions on chocolate agar plates. After 24 h, mice were anesthetized with sodium pentobarbital. The dorsal aorta was severed and the diaphragm punctured. Bronchoalveolar lavage (BAL) fluid was collected by cannulating the trachea and washing the lungs with 1 mL of PBS, reinserted twice.

To analyze CFU content, lungs were harvested and homogenized in 500 µL PBS. Serial dilutions of the homogenate and BAL fluid were plated on chocolate agar and CFU was counted. Results presented are total CFU in the lung homogenate. BAL was also analyzed for chemokine (C-X-C motif) ligand 1 (CXCL1 aka GRO-1) protein by a Luminex-based assay (Research And Diagnostic Systems).

All cell-free and in vitro data is expressed as mean ± SEM of 2-3 independent trials, with every condition performed in triplicate within each trial. In vivo data is expressed as mean ± SEM of 2 independent trials, with 5-6 mice tested within each trial. Dot plots represent results from the combined trials, with each biological replicate presented as a single dot. Data was analyzed using GraphPad software and significant difference (P<0.05) between groups was assessed by t-test or analysis of variance (ANOVA) with multiple comparisons test (Dunnett’s, Tukey’s).

A microtiter plate-based antibacterial assay was used to examine the antibacterial capability of phloretin against the COPD pathogens. Bacteria were incubated with vehicle, 0.1-1 mM phloretin, or 100 µg/mL gentamicin (antibiotic control). Phloretin inhibited NTHi, M. catarrhalis, S. pneumoniae, and PAO1 growth ( Figure 1A ). Inhibition was significant at ≥ 0.1 mM for NTHi, M. catarrhalis, and S. pneumoniae, and ≥ 1 mM for PAO1. Microtitre results were confirmed by manual CFU counts of the wells following the completion of the assays (18 h) and were comparable to the growth curves, where phloretin had antibacterial activity against all of the bacterial strains in a dose-dependent manner, except for S. pneumoniae ( Supplemental Table 3 ). This may be due to a loss of viability of the bacteria. When the CFU was manually enumerated at the beginning of the stationary phase (11 h), phloretin (0.3-1 mM) displayed antibacterial activity against S. pneumoniae ( Figure 1A ). In another abiotic assay, biofilm biomass was measured by crystal violet staining (OD₅₇₀). Phloretin (0.1 mM) inhibited biofilm formation of NTHi, M. catarrhalis, and S. pneumoniae but not PAO1 ( Figure 1B ).

The antibacterial effect of phloretin was examined in co-culture with NCI-H292 cells, a human pulmonary epithelial cell line. Over time, phloretin inhibited NTHi adhesion to the cells over 8 h. After 4 h exposure, phloretin reduced adherent bacteria but did not have an impact on bacterial invasion at this time ( Figure 1C ). Phloretin was previously found to non-cytotoxic to NCI-H292 cells at this dose, as measured by lactate dehydrogenase release and ATP levels (Birru et al., 2020).

LPS is a component of the cell wall of gram-negative bacteria that is recognized and bound by Toll-like receptors (TLRs) on cell surfaces, initiating an innate immune response (Lee and Kim, 2007; Knobloch et al., 2011). NTHi contains short-chain LPS, or lipooligosaccharide (LOS), with structural variant oligosaccharide components attached to its core region (Choi et al., 2014) and can bind to TLR4 in lung epithelial cells (Birru et al., 2020) and macrophages (Lorenz et al., 2005). S. pneumoniae is a gram-positive bacterium that does not express LPS. However, TLR2 and TLR4 recognize lipoteichoic acid, present on its cell wall, and pneumolysin, which it secretes (Sánchez-Tarjuelo et al., 2020). Upon TLR activation, macrophages secrete cytokines to stimulate activation and migration of neutrophils and other immune cells to the site of injury (Wu et al., 2009). To evaluate the anti-inflammatory effects of phloretin, we exposed RAW 264.7 macrophages to the COPD associated pathogens and measured TNF secretion as a marker of inflammation. TNF protein released from the cells into the supernatant was measured at 2 h, prior to the phloretin-induced inhibition of bacterial growth ( Figure 1 ). Each pathogen increased TNF secretion relative to vehicle control ( Figure 2A ). Phloretin pretreatment inhibited pathogen-induced TNF secretion in RAW 264.7 cells, indicating its anti-inflammatory role. In pathogen exposed cells at 24 h, vehicle treated cells had a significant increase in TNF secretion while levels in phloretin treated cells remained at baseline and were comparable to the vehicle plus control treatment group (data not shown). This may be a result of both the anti-inflammatory and antimicrobial activity of phloretin. At this concentration, phloretin did not induce cytotoxicity in RAW 264.7 cells ( Supplemental Figure 2A ). Furthermore, the bacteria did not impact cell viability ( Supplemental Figure 2A ).

To confirm these results in a human model of infection, we used HBE cells derived from normal and COPD subjects. Airway epithelial cells release cytokines in response to pathogens to induce chemotaxis of inflammatory mediators to the site of injury. In HBE cells, we measured IL6, an inflammatory mediator, IL8, a neutrophil chemoattractant, and CCL20, a lymphocyte chemoattractant, which have all been found to be induced by respiratory epithelial cells following NTHi exposure (Clemans et al., 2000; Amatngalim et al., 2017). NTHi increased IL6, IL8, and CCL20 transcript levels in HBE cells. IL6 and CCL20 transcript levels were significantly higher in COPD HBE cells compared to NHBE cells. Phloretin suppressed these biomarkers in both NHBE and COPD HBE cells ( Figure 2B ). At this concentration, phloretin did not induce cytotoxicity in HBE cells ( Supplemental Figure 2B ). NTHi had a small but significant effect on cell viability in NHBE cells but not COPD HBE cells ( Supplemental Figure 2B ).

To examine its antibacterial effects in vivo, mice were supplied a diet supplemented with phloretin (0.157%, 1 wk) and exposed intratracheally to NTHi (24 h). Serial dilutions of the bronchoalveolar lavage fluid and lung tissue homogenate were plated to quantify the bacterial load by enumerating the CFU. NTHi CFU was reduced in the BAL fluid and lung tissue of mice exposed to phloretin ( Figure 3A ). Previously, we showed that NTHi exposure induced neutrophil infiltration into the lungs, which was inhibited by phloretin treatment (Birru et al., 2020). To strengthen these results, we measured CXCL1, a neutrophil chemoattractant, in BAL fluid from NTHi-treated mice. CXCL1 increased after NTHi exposure and was reduced with a phloretin diet ( Figure 3B ). The reduction in neutrophils and CXCL1 in mouse lung was in agreement with the inhibition of IL8 in HBE cells. Together these findings indicate that phloretin displays both bacteriostatic and anti-inflammatory activity against COPD associated pathogens in vitro and in vivo.