Host Defense against Klebsiella pneumoniae Pneumonia Is Augmented by Lung-Derived Mesenchymal Stem Cells

This article is featured in Top Reads, p.999

Bacterial pneumonia remains a major global disease burden (1). Klebsiella pneumoniae, the Gram-negative bacterium, induces severe pneumonia associated with substantial parenchymal damage in the lungs. In recent years, the spread of carbapenem-resistant K. pneumoniae strains, which cause ≥50% mortality, has made treatment more difficult (2, 3). Moreover, K. pneumoniae is known to cause life-threatening lung infections in patients with diabetes and in alcohol consumers (4–6). Thus, the emergence of multidrug-resistant K. pneumoniae strains necessitates the investigation of alternative therapeutic options.

Mesenchymal stem cells (MSCs), adult multipotent stromal cells, have shown potential as therapeutic agents with significant clinical application toward a variety of diseases, including those affecting the lung. MSCs are a favored cell type in preclinical and clinical research due to their distribution in numerous organs, convenient isolation and propagation under in vitro conditions, extensive proliferative and differentiable capabilities, paracrine functions, and immunosuppressive activities (7–14).

Bone marrow–derived MSCs (BM-MSCs) have been widely studied because of their potential as a cell-based immunotherapy for a broad range of diseases, including autoimmune diabetes, severe graft-versus-host disease, myocardial infarction, Crohn’s disease, pulmonary hypertension, hepatic failure, acute renal failure, endotoxin-induced acute lung injury, sepsis, acute respiratory distress syndrome, bleomycin-induced lung injury and fibrosis, chronic obstructive pulmonary disease, allergic asthma, bronchiolitis obliterans, bronchopulmonary dysplasia, and bacterial pneumonia (9, 11–16). Recent studies using preclinical animal models have demonstrated that BM-MSCs have potent beneficial effects for lung injury caused by infection, which are attributed to a number of MSC properties, including their immunomodulation, secretion of various reparative growth factors and cytokines, and augmentation of host defense mechanisms (11, 17, 18). In contrast with the well-established use of BM-MSCs in cellular therapy, much less is known about the reparative and paracrine functions of lung MSCs (LMSCs). Moreover, although cells with MSC characteristics (19–27) have been isolated from the lungs of rodents, sheep, and humans, few studies have tested the therapeutic potential of these cells in animal models of lung diseases, including elastase-induced pulmonary emphysema (23, 24, 28), allergic asthma (29), and LPS-induced acute respiratory distress syndrome (30).

To our knowledge, the role of LMSCs in bacterial pneumonia has not been investigated. Therefore, we evaluated the therapeutic potential of purified stem cell Ag-1⁺ (Sca-1⁺) LMSCs in a mouse model of bacterial pneumonia using the Gram-negative bacterium, K. pneumoniae. In the current study, we cultured total lung cell digests from C57BL/6J mice for an extended time to generate plastic-adherent cells and then used Sca-1 as a phenotypic marker to purify a specific population of LMSCs from the lung cell culture outgrowth with the help of a fluorescence-activated cell sorter. Unlike Sca-1⁺ bronchoalveolar stem cells (31, 32), the LMSCs purified in this study could be expanded and cultured (without growth factors) in noncoated (not coated with Matrigel or collagen) plastic cell culture dishes under nondifferentiating conditions. These Sca-1⁺ LMSCs exhibit several characteristics typical of classical MSCs (8, 33) and differentiate into adipogenic and osteogenic lineages in vitro. These Sca-1⁺ LMSCs exhibit antibacterial properties and more potently suppress the growth of K. pneumoniae in vitro than do bone marrow–derived neutrophils (BMDNs). In addition, Sca-1⁺ LMSCs possess the ability to phagocytose and kill K. pneumoniae in vitro. Adoptive transfer of these cells into a mouse model of K. pneumoniae–induced pneumonia improved survival, minimized inflammation and lung injury, and augmented bacterial clearance in the lungs and in extrapulmonary organs.

The following Abs were used to characterize Sca1⁺ LMSCs, as well as for Western blot, immunofluorescence, and confocal microscopic studies. These Abs were obtained from Abcam (anti-GAPDH Ab, K. pneumoniae Ab, and goat anti-rabbit IgG H&L [Alexa Fluor 594] secondary Ab), Cytoskeleton, Inc., Denver, CO (Acti-stain 488 phalloidin), eBioscience (PE-conjugated donkey anti-mouse IgG, FITC-conjugated anti-mouse IgG, PE-conjugated anti-mouse CD103, anti-mouse CD105 PE, PE-conjugated anti-mouse CD123, PE-conjugated anti-mouse CD135, anti-mouse CD117 [c-kit] PE, FITC-conjugated anti-mouse CD54, and FITC-conjugated anti-mouse/rat ICOS), BD Pharmingen (FITC rat anti-mouse CD45, FITC rat anti-mouse CD34, PE rat anti-mouse CD73, FITC rat anti-mouse Ly 6-A/E, FITC rat anti-mouse CD31, PE rat anti-mouse I-A/I-E, FITC hamster anti-mouse CD11c, FITC rat anti-mouse CD44, FITC rat anti-CD11b [clone M1/70], and PE anti-mouse complement receptor of the Ig superfamily [CRIg; clone 17C9]), BioLegend (PE anti-mouse TLR4 and FITC anti-rat CD90/mouse CD90.1 [Thy 1.1]), Cell Signaling Technology (LC3 A/B rabbit Ab, Beclin-1 rabbit mAb, ATG7 rabbit mAb, HRP-conjugated anti-mouse IgG, and HRP-conjugated anti-rabbit IgG), Miltenyi Biotec (FcR blocking reagent), and Thermo Fisher Scientific (CD18 [LFA-1 β] mAb [M18/2], PE).

Pathogen-free C57BL/6J mice (8- to 12-wk-old male and female mice; purchased from The Jackson Laboratory, Bar Harbor, ME) were used in all experiments. C57BL/6J mice were fed mouse chow and water ad libitum and were kept under specific pathogen-free conditions with a 12-h dark and 12-h light cycle. Animal experiments were performed in accordance with National Institutes of Health guidelines for the use of animals with the approval of the Louisiana State University (LSU) Institutional Animal Care and Use Committee.

In brief, K. pneumoniae (strain 43816, serotype 2; American Type Culture Collection, Manassas, VA) was grown as previously described (34, 35). The bacterial suspension was then diluted to the desired concentration (1 × 10³ CFUs/50 µl of sterile PBS/mouse [for acute pneumonia study] or 5 × 10³ CFUs/75 µl of sterile PBS/mouse [for survival study]).

The Sca-1⁺ LMSCs were isolated from the lungs of 8-wk-old male C57BL/6J mice. Lung tissues were digested using the lung digestion methods described in our earlier publication (36). In brief, we perfused the left and right ventricles of the heart with 15 ml of Dulbecco’s PBS (DPBS; containing 20 IU/ml heparin sodium; Sigma-Aldrich). The trachea and upper airways were carefully removed, and five whole lungs were digested with 20 ml of dissociation buffer containing Collagenase Type II (Worthington Biochemical Corporation, Lakewood, NJ), Dispase II (Sigma-Aldrich), and DNase I (Sigma-Aldrich) for 45 min at 37°C in a CO₂ incubator. The lung suspension was diluted with an equal volume of complete α MEM (CMEM; containing 9% FBS, 9% donor equine serum, l-glutamine [2 mM], penicillin [100 IU/ml], streptomycin [100 μg/ml], and amphotericin B [0.25 μg/ml; Sigma-Aldrich]). Next, the lung digest was filtered using a 70-µm nylon cell strainer (Corning, NY), which was washed twice with DPBS and centrifuged at 1800 rpm for 10 min at 4°C. The supernatant was discarded. Erythrocytes were removed by lysing the cell pellet with RBC lysis buffer (eBioscience, San Diego, CA) for 5 min at room temperature. The lysis reaction was quenched by adding 25 ml of CMEM and centrifugation at 1800 rpm for 10 min at 4°C. Cells were resuspended in CMEM and cultured in 25-cm² plastic tissue culture flasks (Genesee Scientific Corporation, El Cajon, CA). The medium was removed and replaced with fresh CMEM every 2–3 d. After 7–10 d of culture (passage 1), the adherent cells were washed with DPBS and then trypsinized using 0.25% trypsin EDTA solution (Life Technologies, Carlsbad, CA) for 2–3 min at 37°C in a CO₂ incubator. The cells were washed with CMEM and centrifuged at 1800 rpm for 10 min at 4°C. Lung cells were then resuspended in CMEM and cultured for another 2 × 7–10 d (passages 2 and 3) in 182-cm² plastic tissue culture flask (Genesee Scientific). After passage 3, the adherent lung cells were trypsinized, washed with DPBS, and centrifuged. This process was repeated once more. The cell pellet was suspended in cell staining buffer (Sigma-Aldrich) and then incubated with FITC-conjugated Ly 6-A/E (Sca-1) anti-mouse Ab or FITC-conjugated IgG control Ab for 30 min on ice. The cells were washed with cell staining buffer, centrifuged, and resuspended in 2 ml of cell staining buffer. The Sca-1⁺ cells were sorted using a FACSAria II cell sorter (BD Biosciences, San Jose, CA). The flow-sorted Sca-1⁺ lung cells were washed with CMEM, centrifuged twice, resuspended in CMEM, and cultured in plastic tissue culture flasks for another three (passages 4–6) to six passages (passages 7–9) to generate more cells. At passage 6, the adherent flow-sorted Sca-1⁺ cells were washed with DPBS, trypsinized, and analyzed for differentiation capacity, as well as the expression of MSCs and other cell surface Ags. These well-characterized passage 6 Sca-1⁺ LMSCs were used for the in vitro and in vivo experimental analyses.

Flow cytometric analysis was conducted to analyze the phenotypes of sorted Sca-1⁺ lung cells (passage 6). Flow-sorted Sca-1⁺ lung cells (passage 6) were first blocked with mouse FcR blocking reagent (0.25 µg/ml) for 15 min on ice. Cells were then incubated with primary Abs (1–4 µg/ml for 30 min on ice) conjugated to FITC or PE. Mouse IgG PE and IgG-FITC Abs were used as isotype controls. Cells were analyzed using a FACSCalibur analyzer (BD Biosciences, San Jose, CA) with FlowJo software (FlowJo, Ashland, OR). Abs used and manufacturers are detailed in the Abs section. After characterization of these flow-sorted passage 6 Sca-1⁺ lung cells using differentiation and proliferation assays and phase-contrast and electron microscopies (see later), we cultured these Sca-1⁺ LMSCs for another 3 passages (passage 9). We analyzed the expression of various MSCs and other cell surface Ags on these passage 9 Sca-1⁺ LMSCs using flow cytometry. We have also characterized the morphology and ultrastructural features of passage 9 Sca-1⁺ LMSCs using phase-contrast microscopy and transmission electron microscopy (TEM), respectively.

The colony-forming capacity (i.e., CFUs) of purified Sca-1⁺ cells (passage 6) was determined using the protocol described by Peister et al. (37). In brief, Sca-1⁺ cells (in triplicates) were cultured at a density of 1000 cells in 10 ml of CMEM in a 100-mm petri dish for 21 d, and the medium was replaced every 3 d. On day 21, the plates were rinsed and stained with 3% crystal violet (Sigma-Aldrich) for 20 min.

The differentiation potential of purified Sca-1⁺ cells (passage 6) in response to adipogenic and osteogenic lineages was assessed using the method described by Peister et al. (37). In brief, the cells were cultured (50 cells/cm²) in 58-cm² dishes and incubated in CEM for 10 d. Osteogenic differentiation was induced by culturing the adherent cells in IMDM (containing FCS [10%], horse serum [10%], penicillin [100 IU/ml], streptomycin [100 μg/ml], l-glutamine [12 mM], β-glycerol phosphate [20 mM; Sigma-Aldrich], thyroxine [50 ng/ml; Sigma-Aldrich], dexamethasone [1 nM; Sigma-Aldrich], and ascorbate 2-phosphate [0.5 μM; Sigma-Aldrich]). The medium was changed every 3 d for 3 weeks. After 3 weeks, adherent cells were washed with DPBS, fixed with 10% formalin, and then stained with Alizarin Red S reagent (Sigma-Aldrich).

For adipogenic differentiation, cell cultures in 24-well plates were stimulated with a specific adipogenic-differentiation medium (IMDM containing FCS [10%], horse serum [10%], penicillin [100 IU/ml], streptomycin [100 μg/ml], l-glutamine [12 mM], insulin [5 μg/ml; Sigma-Aldrich], indomethacin [50 μM; Sigma-Aldrich], dexamethasone [1 × 10⁻⁶ M; Sigma-Aldrich], and 3-isobutyl-1-methylxanthine [0.5 μM; Sigma-Aldrich]). The medium was changed every 3 d for 3 weeks. After 21 d, the adherent cells were washed with DPBS, fixed with 10% formalin, and then stained with 0.5% Oil Red O (Sigma-Aldrich) reagent for 20 min at room temperature. Lipid droplets were then visualized using a light microscope.

To assess proliferation, we cultured Sca-1⁺ LMSCs (passage 6) at a density of 5 × 10³ cells or 20 × 10³ cells in 2 ml CMEM/well in 12-well plates (Genesee Scientific). Adherent cells from three random wells were trypsinized at the indicated time points, stained with trypan blue (MP Biomedicals, Santa Ana, CA), and counted using a hemocytometer. Growth curves were plotted using the mean values. Experiments were performed two times.

For light microscopy, we cultured the flow-sorted Sca-1⁺ LMSCs (passages 6 and 9) in CMEM (using 12-well plates) for 1, 2, and 5 d. Adherent cells were stained using 1% methylene blue (Electron Microscopy Sciences, Hatfield, PA) and examined under a light microscope.

For scanning electron microscopy studies, cells in CMEM (5 × 10⁴ cells in 0.5 ml of CMEM; for Sca-1⁺ LMSCs [passage 6]) or DMEM (containing 10% FBS, penicillin [100 IU/ml], and streptomycin [100 μg/ml] for 3T3 mouse embryonic fibroblasts [3T3 MEFs WT; CRL-2752; purchased from American Type Culture Collection, Manassas, VA]) were cultured on Thermanox coverslips in 24-well plates. After 5 d of culture, the coverslips containing the adherent cells were washed with sterile DPBS and fixed with electron microscope (EM) fixative for 2 h. Cells were then washed with distilled water, dehydrated through an ethanol series, and dried using a Critical Point Drier (Polaron E3000, USA). The dried samples were then coated with gold-palladium in EMS 550X Sputter Coater and examined using a JEOL JSM-6610LV scanning electron microscope (JEOL USA, Peabody, MA).

For TEM studies, Sca-1⁺ LMSCs (passages 6 and 9) and 3T3 MEFs were cultured for 5 d in 182-cm² flasks using CMEM, as described earlier for the scanning electron microscopy study. On day 5, adherent cells were trypsinized, washed several times with DPBS, and centrifuged (2000 rpm for 8 min at 4°C). Cell pellets were fixed in primary electron microscope fixative (pH 7.4; 1.6% paraformaldehyde, 2.5% glutaraldehyde, 0.03% CaCl₂ in 0.05 M cacodylate buffer) for 2 h, washed with 0.1 M cacodylate buffer supplemented with sucrose, and embedded in 3% agarose. Agarose blocks were cut into 1-mm³ cubes, washed in 0.1 M cacodylate buffer, and fixed in 1% osmium tetroxide (in 0.1 M cacodylate buffer) for 1 h. Samples were then washed in water and stained with 2% uranyl acetate (in 0.2 M sodium acetate buffer, pH 3.5) for 2 h, dehydrated in ascending ethanol concentrations (30–100%) and propylene oxide, and fixed in Epon-Araldite mixture. Blocks were then sectioned using Ultratome Leica EM UC7 (Leica Microsystems, Buffalo Groove, IL). Ultrathin sections (80 nm) were stained with lead citrate for 5 min and examined using a JEOL JEM-1400 TEM (JEOL USA, Peabody, MA). All reagents for scanning electron microscopy and TEM studies were purchased from Electron Microscopy Sciences (Hatfield, PA).

We generated macrophages by culturing bone marrow–derived (from 8- to 12-wk-old female C57BL/6J mice) cells with recombinant murine M-CSF (50 ng/ml in DMEM containing 10% FBS; PeproTech, Rocky Hill, NJ) for 8 d (34). On day 8, we trypsinized the adherent bone marrow–derived macrophages (BMDMs) and cultured them with Sca-1⁺ LMSCs (total of 0.5 million passage 6 Sca-1⁺ LMSCs/0.5 ml of DMEM containing 10% FBS in 24-well plates) at different ratios (Sca-1⁺ LMSCs to macrophages at 1:1 to 1:4 ratios) in the presence or absence of LPS (055:B5 [Sigma-Aldrich]; 500 ng/ml) for 24 h. The levels of TNF-α and IL-1β in the culture supernatants were measured using Ready-SET GO ELISA kits (eBioscience, San Diego, CA). To determine whether Sca-1⁺ LMSCs were able to regulate the production of proinflammatory cytokines by LPS-stimulated neutrophils in a cell-cell contact-dependent manner, we purified neutrophils from bone marrow (BMDNs) of female C57BL/6J mice using the EasySep mouse neutrophil enrichment kit (STEMCELL Technologies, Cambridge, MA). Then, we cultured the Sca-1⁺ LMSCs with BMDNs (total of 0.5 million cells/0.5 ml of DMEM containing 10% FBS in a 24-well plate) at different ratios (Sca-1⁺ LMSCs to BMDNs at 1:2 and 1:4 ratios) in the presence or absence of LPS (500 ng/ml) for 8 h. The level of TNF-α in the culture supernatants was measured by ELISA (eBioscience).

The antimicrobial activity of Sca-1⁺ LMSCs and BMDNs against K. pneumoniae was determined in vitro at different time points. First, we treated the Sca-1⁺ LMSCs (passage 6) or BMDNs (each 2.5 × 10⁵ cells/0.2 ml of DMEM supplemented with 10% FBS) with cytochalasin D (10 μg/ml; Sigma-Aldrich) for 20 min to block phagocytosis prior to infection. K. pneumoniae (in 50 μl sterile PBS) was added (1 or 10 multiplicities of infection [MOIs]) to the wells containing Sca-1⁺ LMSCs or BMDNs, and the cell culture plates were incubated at 37°C for various time periods (1–6 h). The ratios of bacterium to Sca-1⁺ LMSCs or BMDNs were 1:1 and 10:1. In some experiments, we added K. pneumoniae (10 MOI) to the wells containing LMSCs supernatant (0.1 ml supernatant and 0.1 ml DMEM supplemented with 10% FBS/well) alone or wells containing both LMSCs supernatant (0.1 ml/well) and BMDNs (2.5 × 10⁵ cells/0.1 ml of DMEM supplemented with 10% FBS). At designated time points, the plates were centrifuged at 600 × g for 4 min, and supernatants were collected, diluted in sterile PBS, plated onto MacConkey agar plates, and incubated overnight at 37°C.

To investigate whether infection with K. pneumoniae can induce the secretion of antimicrobial compounds, we cultured the Sca-1⁺ LMSCs (1 million cells/ml DMEM [containing 10% FBS]) with 20 MOIs of opsonized K. pneumoniae for 4 h. We centrifuged the culture supernatants at 2000 rpm for 8 min at 4°C and quantified the antimicrobial compounds cathelicidin-related antimicrobial peptide (CRAMP; MyBioSource.com, San Diego, CA), leukotriene B4 (LTB4; Cayman Chemical, Ann Arbor, MI), and lipocalin-2/NGAL (R&D Systems, Minneapolis, MN) in the culture supernatants using ELISA.

To investigate whether treatment with CRAMP can inhibit the extracellular growth of K. pneumoniae (38), we incubated the bacteria in tryptic soy agar broth (5 × 10³ CFUs/100 µl/well in 96-well plates) in the presence or absence of murine CRAMP (Chi Scientific, Maynard, MA) (20 µg/well) for 2 h at 37°C. Then, the bacterial cultures were serially diluted in sterile PBS and plated onto MacConkey agar plates.

The extracellular hydrogen peroxide (H₂O₂) was quantified using Fluoro H₂O₂ kit (Cell Technology, Hayward, CA). In brief, the passage 6 LMSCs (1 × 10⁵ cells/well in 100 µl of HBSS using a 96-well cell culture plate) were infected with 20 MOIs of opsonized K. pneumoniae for 30 min. Postinfection, 50 µl of reaction mixture was added to each well and incubated for 10 min at room temperature in the dark. H₂O₂ was used as the standard. The plates were read at excitation and emission wavelengths of 530 and 570 nm using a fluorescence plate reader.

Opsonophagocytic killing assay was conducted as described by Davis et al. (39) with modifications. In brief, K. pneumoniae was grown to midlog phase, PBS washed, and resuspended in +++ solution. The bacteria (200 × 10⁷ CFUs in 400 μl +++ solution) were preopsonized with 1600 μl of 3–4 wk rabbit complement (Pel-Freez, Rogers, AR) for 30 min at 37°C with constant shaking. The opsonized K. pneumoniae (1 × 10⁷ CFUs in 50 µl) were added to wells containing Sca-1⁺ LMSCs (passage 6) or BMDMs (each set of 0.5 × 10⁶ cells in 450 µl of DMEM containing 10% FBS/well in 12-well plates), and the plates were incubated for 1 h at 37°C in a CO₂ incubator. Cell culture plates were centrifuged at 600 × g for 4 min at 4°C, the medium was removed, and the extracellular bacteria were removed by washing the plates three times with sterile PBS, as well as by centrifuging the plates at 600 × g. Then the cells were cultured with 0.5 ml DMEM (containing 10% FBS and gentamicin sulfate [final concentration, 250 μg/ml]; Sigma-Aldrich) for 1 h. After 1 h, the medium was removed and the cells were cultured with fresh DMEM (0.5 ml/well) containing 10% FBS and gentamicin sulfate (final concentration, 125 μg/ml) for another 2 h (4-h time point) at 37°C in a CO₂ incubator. After each time point, the antibiotic was removed by serial (three times) washing with PBS. Adherent cells were lysed with 0.1% Triton X-100 (120 µl/well; Sigma-Aldrich) for 10 min, serially diluted in PBS, and plated onto MacConkey agar plates. The number of viable bacteria was determined after incubating the plates at 37°C for 16 h.

We also confirmed the uptake of live K. pneumoniae by Sca-1⁺ LMSCs using immunofluorescence microscopy and immunofluorescence confocal microscopy. We opsonized the live K. pneumoniae with rabbit complement by using the protocol described earlier. Then, we inoculated the Sca-1⁺ LMSCs (passage 6) with 20 MOIs of opsonized K. pneumoniae and cultured the infected cells in Millicell EZ slide (EMD Millipore Corporation, Burlington, MA) at 37°C in a CO₂ incubator for 3 h. Plates were centrifuged, and adherent cells were washed three times with PBS. The adherent cells were fixed with 2% paraformaldehyde overnight, washed thrice with PBS, permeabilized with 0.1% Triton X-100 for 10 min, and blocked with 1% BSA (in PBS) for 1 h. To localize K. pneumoniae in Sca-1⁺ LMSCs, we incubated cells with anti–K. pneumoniae Ab (2 µg/ml in PBS containing 0.1% BSA) overnight at 4°C. Cells were washed three times with PBS containing 0.02% Tween 20 and then incubated with preadsorbed goat anti-rabbit IgG H&L (Alexa Fluor 594) secondary Ab (2 µg/ml in PBS containing 0.1% BSA) for 1 h at room temperature. The cells were washed three times with PBS and stained with Acti-stain 488 phalloidin by using the protocol developed by the manufacturer (Cytoskeleton, Inc.). Cells were counterstained with DAPI (VECTASHIELD Mounting Medium with DAPI; Vector Laboratories, Burlingame, CA). Images were acquired using a Carl Zeiss immunofluorescence microscope (Carl Zeiss Microscopy, Thornwood, NY) and Leica TCS SP8 confocal microscope (Leica Microsystems, Buffalo Grove, IL) with a 488-nm argon laser and a 594-nm laser.

Heat inactivation of K. pneumoniae (strain 43816) was accomplished by incubating 3 × 10⁹ CFUs of bacteria in 3 ml of sterile PBS for 1 h at 70°C using a water bath (40). BMDMs and Sca-1⁺ LMSCs (1 million cells/0.5 ml of DMEM containing 10% FBS) were cultured (in a 12-well plate) separately either with medium or with 20 and 50 MOIs of heat-inactivated and FITC-labeled K. pneumoniae (HI Kp-FITC) for 1 h at 37°C in a CO₂ incubator. After 1 h, the plates were centrifuged and washed two times with PBS. The adherent cells were trypsinized, washed three times with PBS, and resuspended in PBS. One hundred microliters of filtered (using 0.22-µm filter) trypan blue was added to 1 ml of cell suspension before measurement using a flow cytometer. Sca-1⁺ LMSCs cultured for 1 h with DMEM (containing 10% FBS) and then quenched with trypan blue were used as negative control. The percent uptake of HI Kp-FITC by Sca-1⁺ LMSCs and BMDMs was analyzed using flow cytometry.

Sca-1⁺ LMSCs were labeled with isotype IgG control Abs (PE-conjugated anti-mouse IgG; FITC-conjugated anti-mouse IgG), complement receptor 3 (CR3) Abs (FITC rat anti-CD11b [clone M1/70)]; CD18 [LFA-1 β] mAb [M18/2], PE], or CRIg Ab (PE anti-mouse CRIg) for 30 min on ice. The cells were washed with PBS, fixed in 2% paraformaldehyde, and analyzed using a flow cytometer.

Total intracellular reactive oxygen species (ROS) were measured using Cell Meter Fluorimetric Intracellular Total ROS Activity assay kit *Deep Red Fluorescence* (AAT Bioquest, Sunnyvale, CA). In brief, the LMSCs were seeded in Costar black wall/clear-bottom 96-well plates at a density 5 × 10⁴ cells (in 100 µl DMEM containing 10% FBS). The cells were infected with 20 MOIs of opsonized K. pneumoniae for 15 and 30 min. At the end of each time point, ROS Brite 670 working solution (100 µl/well) was added to the wells and incubated for 30 min. An increase in fluorescence was measured at excitation and emission wavelengths of 650 and 675 nm using a fluorescence plate reader.

We cultured the cells (0.5 million cells/0.5 ml DMEM containing 10% FBS/well in 12-well plates) in the presence or absence of 5 mM 3-methyladenine (3-MA; Millipore Sigma, St. Louis, MO) for 1 h as reported previously (41). Then, the cells were infected with 20 MOIs of opsonized K. pneumoniae for 1 and 4 h. The plates were centrifuged and washed three times with sterile PBS. The adherent cells were lysed with 0.1% Triton X-100, serially diluted, and plated onto MacConkey agar plates. The levels of different autophagy-associated proteins in the 3-ME–treated LMSCs/3-ME–treated and K. pneumoniae–infected LMSCs (3 million cells/3 ml DMEM containing 10% FBS/well in six-well plates) were analyzed using Western blot. LMSCs cultured with medium alone were used as controls. For Western blot analysis, the adherent cells were lysed with Nonidet P-40 cell lysis buffer (Invitrogen, Frederick, MD) and centrifuged at 2000 rpm for 8 min at 4°C using an Eppendorf centrifuge. The protein concentration in the lysates was quantified using Pierce BCA Protein Assay kit (Thermo Scientific, Rockford, IL) to ensure that equal amounts of proteins were loaded onto 12% SDS-PAGE. Then the proteins were electrophoretically transferred onto a polyvinylidene difluoride membrane (Millipore), and membranes were incubated with appropriate primary Abs (LC3 A/B Ab, Beclin-1 rabbit mAb, ATG7 rabbit mAb, anti-GAPDH Ab) and then with appropriate secondary Abs (HRP conjugated to either anti-mouse IgG or anti-rabbit IgG) linked to HRP conjugates. Blots were developed using an ECL kit (Amersham Biosciences UK, Buckinghamshire, UK). For autophagy induction experiments, cells (0.5 million cells/well in 0.5 ml of 10% FBS) were cultured in the presence or absence of 80 µg/ml rapamycin (InvivoGen, San Diego, CA) for 2 h before infecting them with 20 MOIs of opsonized K. pneumoniae.

We infected female C57BL/6J mice with K. pneumoniae as previously described (34). At 4 h postinfection with K. pneumoniae, mice were anesthetized and given either passage 6 Sca-1⁺ LMSCs (750,000 cells in 50 μl sterile DPBS/mouse intratracheally [i.t.]) or 3T3 MEFs (750,000 cells in 50 μl sterile DPBS/mouse i.t. as a control for LMSCs) or sterile DPBS (50 μl/mouse i.t. as a control for bacteria). The mice were sacrificed at 48 h postinfection, and the samples were collected from each mouse for bronchoalveolar lavage fluid (BALF) and phenotyping and lung histopathology, as well as to enumerate the bacterial burden in different organs.

The trachea was cannulated using a 20-gauge shielded i.v. catheter (42). BALF was collected four times from each mouse by instilling 0.8 ml of sterile PBS (containing 0.2 U of heparin and protease inhibitor mixture [one tablet/10 ml]; Sigma-Aldrich). To quantify the total WBCs (TWBCs), we centrifuged 3 ml of BALF (from each mouse) at 1800 rpm for 8 min at 4°C. The cell pellet obtained from the BALF of each mouse was resuspended in 3 ml of PBS, and the TWBCs were enumerated using a hemocytometer. To analyze different inflammatory cell populations in the BALF, we centrifuged BALF onto Superfrost/plus microscopic slides (Fisher Scientific, Pittsburgh, PA) using a Wescor Cytopro Cytocentrifuge (Hyland Scientific, Stanwood, WA). The slides were stained using Quick-Dip kit (Mercedez Scientific, Lakewood Ranch, FL). Differential counts were performed on 300 cells, as described in standard cytological techniques (43).

Different cytokines, chemokines, and growth factors in the lung supernatants were measured using Milliplex MAP mouse cytokine/chemokine magnetic bead assay (EMD Millipore Corporation) and Milliplex MAP mouse angiogenesis and growth factor magnetic bead assay (EMD Millipore Corporation).

The lungs were inflated with 1% low-melting agarose (Sigma-Aldrich) and fixed in 4% phosphate-buffered formalin for 18–24 h. The whole lung was embedded in paraffin, and sections (5-μm thickness) were stained with H&E. The H&E-stained lung sections in each group were coded, and representative images (whole lung) were acquired with a scanner (NanoZoomer 2.0-HT slide scanner; Hamamatsu Photonics) using NDP scan v3.1.7 software (lens magnification, ×40). These lung sections were scored for the extent of inflammation/injury by a veterinary pathologist in a blinded fashion. We used the following scoring system (scores 0–3) to determine inflammation/injury in the lungs: 0, no inflammation in lung tissue; 1, <10% of the lung tissue was infiltrated by inflammatory cells; 2, ∼10–50% of the lung tissue was infiltrated by inflammatory cells; 3, >50% of the lung tissue was infiltrated by inflammatory cells.

We labeled the passage 6 Sca-1⁺ LMSCs with 10 µM CFSE (Thermo Fisher Scientific-Invitrogen, Waltham, MA) according to the manufacturer’s protocol. We administered either PBS or CFSE-labeled Sca-1⁺ LMSCs (750,000 cells in 50 μl sterile DPBS/mouse i.t) into the lungs of K. pneumoniae–infected (1 × 10³ CFUs/50 µl of sterile PBS/mouse i.t.) mice at 4 h postinfection. We sacrificed the mice 4 h (group I: Kp + PBS; Kp + CFSE-Sca-1⁺ LMSCs) and 48 h (group II: Kp + PBS; Kp + CFSE-Sca-1⁺ LMSCs) postinfection and inflated the lungs with 1% molten agarose. We fixed the lungs in 4% buffered formalin and embedded them in paraffin. Images (20 per lung section) of the unstained lung sections (5-µm thickness) were acquired using a Leica DM6 B (Allendale, NJ) fluorescence microscope (lens magnification, ×20). In each image, the number of CFSE-positive cells (green) was counted manually.

Mouse organs (lungs, spleen, and liver [left lower lobe]) were harvested and homogenized in 1 ml of sterile PBS using a Retsch Tissue Lyser (Vender Scientific, Newtown, PA) for 2 min. Aliquots of homogenates and BALF were serially diluted in sterile PBS and plated onto MacConkey agar plates. The viable counts of K. pneumoniae were enumerated after incubating the plates overnight at 37°C.

For the survival study, we infected female C57BL/6J mice with K. pneumoniae (5 × 10³ CFUs in 75 μl sterile DPBS/mouse by oropharyngeal [o.p.] route), and then these mice were treated with either passage 6 Sca-1⁺ LMSCs (1 × 10⁶ cells in 75 μl sterile DPBS/mouse o.p.) or 3T3 MEFs (1 × 10⁶ cells in 75 μl sterile DPBS/mouse o.p.), or with sterile DPBS (75 μl/mouse o.p.) at 4 and 24 h postinfection. Survival in each group was recorded every 12 h for up to 15 d. Survival experiments were performed twice.

Statistical analysis was performed using GraphPad Prism 8.3.0 (GraphPad Software, La Jolla, CA). Specific statistical testing is described in the accompanying figure legends. The error bar represents mean ± SEM. A p value < 0.05 was considered statistically significant. Significant differences are indicated by *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

We digested lung tissues using the protocol described in our previous publication (36) and sorted Sca-1⁺ cells using a fluorescence-activated cell sorter. The Sca-1⁺ cells constituted as much as 20.3% of the total adherent lung cell population (Supplemental Fig. 1C, right panel). Flow cytometric analysis revealed that the purified and expanded passage 6 Sca-1⁺ cells strongly express MSC/stromal cell Ags CD44, CD73, CD105, CD106, and CD123, although the expression of MSC/stromal cell Ag CD90.1 (Thy 1.1), as well as CD34 and TLR4 Ags, was weak (Supplemental Fig. 1D). In contrast, the purified Sca-1⁺ cells were negative for endothelial marker CD31 and hematopoietic cell markers CD11b and CD45 (Supplemental Fig. 1D). In addition, these cells were negative for MHC class II, CD11c, CD54, CD103, CD117 (c-kit), ICOS, and CD135 Ags (Supplemental Fig. 1D). Similar to passage 6 Sca-1⁺ LMSCs (Supplemental Fig. 1D), passage 9 Sca-1⁺ LMSCs expressed higher levels of MSC/stromal cell Ags CD105 and CD106, and low levels/weak expression of MSC/stromal cell Ags CD44, CD73, and CD123, as well as CD34 and TLR4 Ags, on their surface (Supplemental Fig. 2). The expression of MSC/stromal cell marker CD90.1 reached to control level in passage 9 Sca-1⁺ LMSCs (Supplemental Fig. 2). Passage 9 Sca-1⁺ LMSCs were also negative for MHC class II, CD31, CD45, CD54, CD11c, CD11b, CD103, CD117 (c-kit), ICOS, and CD135 Ags (Supplemental Fig. 2).

As indicated in Supplemental Fig. 3, the flow-sorted and expanded Sca-1⁺ lung cells (passage 6) formed colonies (Supplemental Fig. 3A) and differentiated into Alizarin Red S–positive osteocytes (Supplemental Fig. 3B, bottom left panel) or Oil Red O–positive adipocytes (Supplemental Fig. 3B, bottom right panel) when cultured in differentiation medium. Control cells grown in CMEM did not stain with either dye (Supplemental Fig. 3B, upper panels).

We cultured the flow-sorted Sca-1⁺ LMSCs (passage 6) in CMEM (for 1, 2, and 5 d) and stained the plastic adherent cells using methylene blue (Fig. 1A, bottom panels). Within 24 h of initial plating, the purified Sca-1⁺ LMSCs were able to attach to the plastic and appeared as spindle-shaped cells arranged in parallel or whorls (Fig. 1A, upper panels). Moreover, these cells exhibited high proliferative potential in culture (Fig. 1B). We further characterized (after 5 days of culture in CMEM) the morphology and ultrastructural features of Sca-1⁺ LMSCs and 3T3 MEFs (as control) using scanning electron microscopy and TEM. Sca-1⁺ LMSCs (passage 6) are, as shown by scanning electron microscopy, to be flattened cells with lamellipodia and many thin filopodia (Fig. 1Ci) similar in morphology to cultured 3T3 MEFs (Fig. 1Di). TEM images revealed the following ultrastructural features: cell bodies that are either rounded, oval, or polygonal, and nuclei that are centric or eccentric in both Sca-1⁺ LMSCs (Fig. 1C) and 3T3 MEFs (Fig. 1D). The nuclei are large, pale, mostly euchromatic, and occupy most of the cell body of Sca-1⁺ LMSCs (Fig. 1C) when compared with 3T3 MEFs (Fig. 1D). In both Sca-1⁺ LMSCs and 3T3 MEFs, (a) the nuclei contain clefts and invaginations of the nuclear envelope; (b) the cell membranes are irregular with lamellipodia and thin filopodia; (c) the cellular organelles are distributed throughout the cytoplasm; (d) the rough endoplasmic reticulum is prominent, containing moderately electron-dense material; and (e) mitochondria are numerous, appearing to be mostly elongated and occasionally organized in small groups of three to four elements. Another frequent finding was the presence of a large number of autophagosomes, endosomes, lysosomes, and vesicles in the cytoplasmic compartment of Sca-1⁺ LMSCs (Fig. 1C) compared with 3T3 MEFs (Fig. 1D). To investigate whether Sca-1⁺ LMSCs were able to maintain such morphological and ultrastructural features after subsequent passages, we cultured the passage 9 Sca-1⁺ LMSCs in CMEM for 5 d and characterized them using phase-contrast microscopy and TEM. Similar to passage 6 Sca-1⁺ LMSCs, the passage 9 Sca-1⁺ LMSCs are plastic-adherent and spindle-shaped cells (Supplemental Fig. 4A). TEM images (Supplemental Fig. 4B) show that passage 9 Sca-1⁺ LMSCs are ultrastructurally similar to passage 6 Sca-1⁺ LMSCs (Fig. 1C), except that (a) their nuclei are mostly heterochromatic in nature, and (b) they contain relatively fewer numbers of lysosomes in their cytoplasm.

Sca-1⁺ LMSCs exhibit an immunomodulatory effect on innate immune cells in response to LPS challenge in vitro. We cocultured the Sca-1⁺ LMSCs with BMDMs for 24 h or with BMDNs for 8 h in the presence or absence of LPS (500 ng/million cells/ml) at different ratios (1:1, 1:2, or 1:4 LMSCs/BMDMs or LMSCs/BMDNs) and quantified inflammatory cytokines in culture supernatants. Inflammatory cytokine production was attenuated in both BMDMs and BMDNs in the presence of LMSCs (Fig. 2A–C). Treatment with LMSCs remarkably inhibited LPS-induced secretion of TNF-α (Fig. 2A) and IL-1β (Fig. 2B) by BMDMs and TNF-α by BMDNs (Fig. 2C) at all ratios tested. Neither BMDMs nor BMDNs secreted TNF-α or IL-1β under unstimulated conditions (data not shown). Further, Sca-1⁺ LMSCs also did not secrete TNF-α and IL-1β under unstimulated conditions or after stimulation with LPS (data not shown).

Bacterial killing by neutrophils, macrophages, mast cells, and γ δ T cells plays a critical role in clearance of K. pneumonia (35, 44–46). First, we performed extracellular killing assays by infecting cytochalasin D–treated Sca-1⁺ LMSCs and BMDNs with K. pneumoniae (1 and 10 MOIs) for various time periods. Both Sca-1⁺ LMSCs and BMDNs showed similar rates of killing when inoculated with a low MOI (1 MOI) of K. pneumoniae at all time points examined (Fig. 3A). However, only Sca-1⁺ LMSC- and Sca-1⁺ LMSC-derived supernatants were efficient at killing K. pneumoniae at 1.5-h (Fig. 3B) and 3-h (Fig. 3C) time points when cultured with a higher MOI (10 MOI) of K. pneumoniae. Interestingly, the Sca-1⁺ LMSC-derived supernatants exhibited more potent killing of K. pneumoniae than Sca-1⁺ LMSCs at 3 h (Fig. 3C). Intriguingly, the combination of Sca-1⁺ LMSC-derived supernatant with BMDNs increased the extracellular killing effect of BMDNs at the 3-h time point (Fig. 3C), although BMDNs did not appear to have potent killing ability, and it is mostly due to LMSC-derived supernatant. At 6 h, both Sca-1⁺ LMSC- and Sca-1⁺ LMSC-derived supernatant, as well as BMDNs, failed to inhibit the growth of K. pneumoniae (Fig. 3D). To investigate whether increased bacterial killing by Sca-1⁺ LMSCs was associated with enhanced production of antimicrobial compounds, we cultured the Sca-1⁺ LMSCs with 20 MOIs of opsonized K. pneumoniae for 4 h and quantified antimicrobial compounds (CRAMP, LTB4, and lipocalin-2/NGAL) in the culture supernatants using ELISA. As shown in (Fig. 3, Sca-1⁺ LMSCs secrete CRAMP (Fig. 3E), LTB4 (Fig. 3F), and lipocalin-2/NGAL (Fig. 3G) under unstimulated conditions, and all of them increased at 4 h postinfection with K. pneumoniae. Importantly, treatment with murine CRAMP remarkably inhibited the extracellular growth of K. pneumoniae in vitro (Fig. 3H). In addition, infection with K. pneumoniae also induced the secretion of H₂O₂ by Sca-1⁺ LMSCs (Fig. 3I).

Next, we compared the phagocytic uptake and intracellular bacterial killing abilities of Sca-1⁺ LMSCs and BMDMs using an opsonophagocytic assay. We performed the intracellular killing assay using 20 MOIs of opsonized K. pneumoniae for 1 and 4 h and found that BMDMs were more efficient in the phagocytotic uptake of K. pneumoniae than Sca-1⁺ LMSCs at the 1-h time point (Fig. 4A). However, Sca-1⁺ LMSCs displayed more efficient intracellular killing of K. pneumoniae compared with that of BMDMs at 4 h postinfection (Fig. 4A). We also confirmed the presence of K. pneumoniae in Sca-1⁺ LMSCs using both confocal and fluorescence microscopies. Imaging using confocal (Fig. 4B, indicated by yellow arrow in right panel) and immunofluorescence microscopy (Fig. 4C, indicated by yellow arrows) showed the presence of this Gram-negative bacterium in Sca-1⁺ LMSCs infected with opsonized K. pneumoniae. These findings prompted us to further investigate the phagocytotic capacity of LMSCs by flow cytometry using HI Kp-FITC. Representative flow cytometric analyses for the percent uptake of HI Kp-FITC by Sca-1⁺ LMSCs and BMDMs are depicted in (Fig. 4D and 4E, respectively. Similar to the opsonophagocytosis assay using live K. pneumoniae (Fig. 4A), the percent of BMDMs involved in phagocytosis after 1 h of HI Kp-FITC incubation was significantly higher when compared with Sca-1⁺ LMSCs (Fig. 4F). Next, we evaluated the surface expression of CR3 (CD11b/CD18) and CRIg in Sca-1⁺ LMSCs, which are known to be involved in the phagocytosis of complement opsonized bacteria (47, 48). Flow cytometric analysis revealed the expression of CRIg in Sca-1⁺ LMSCs (Fig. 4G), whereas LMSCs were negative for CR3 (Fig. 4G). Quantification of intracellular ROS using Cell Meter Fluorimetric ROS Activity assay kit also shows increased production of intracellular ROS by LMSCs at 15 min postinfection with opsonized K. pneumoniae over that of control LMSCs (Fig. 4H).

To investigate whether autophagy is associated with intracellular survival of K. pneumoniae in Sca-1⁺ LMSCs, we first pretreated the Sca-1⁺ LMSCs with either 3-MA or rapamycin and then performed an opsonophagocytosis assay using 20 MOIs of opsonized K. pneumoniae. We found that pretreatment of Sca-1⁺ LMSCs with the autophagy inhibitor, 3-MA, remarkably suppressed the intracellular replication of K. pneumoniae at 1 h postinfection (Supplemental Fig. 5A). In contrast, pretreatment with the autophagy inducer, rapamycin, significantly enhanced the intracellular replication of K. pneumoniae during early time of infection (1 h) in Sca-1⁺ LMSCs (Supplemental Fig. 5B). However, in both groups at 4 h postinfection, Sca-1⁺ LMSCs were able to clear most of the phagocytosed K. pneumoniae (Supplemental Fig. 5A, 5B). No live intracellular K. pneumoniae was detected in Sca-1⁺ LMSCs pretreated with 3-MA (Supplemental Fig. 5A), and few live K. pneumoniae were observed in lysates from Sca-1⁺ LMSCs pretreated with rapamycin at 4 h postinfection (Supplemental Fig. 5B).

To determine whether autophagy inhibition results in decreased levels of key autophagy-associated proteins, we extracted proteins from control Sca-1⁺ LMSCs, K. pneumoniae–infected Sca-1⁺ LMSCs, and 3-MA–pretreated and K. pneumoniae–infected Sca-1⁺ LMSCs and carried out immunoblotting with Abs against Beclin-1, ATG7, and LC3. All three of these autophagy-associated proteins were abundantly expressed in Sca-1⁺ LMSCs at baseline (Supplemental Fig. 5C). Infection with K. pneumoniae further increased the levels of Beclin-1, ATG7, and LC3 at 1 h postinfection in Sca-1⁺ LMSCs (Supplemental Fig. 5C). However, pretreatment with 3-MA decreased the levels of these autophagy-associated proteins in Sca-1⁺ LMSCs at 1 h postinfection (Supplemental Fig. 5C). These results suggest activation of the autophagy pathway occurs after opsonophagocytic uptake of K. pneumoniae by Sca-1⁺ LMSCs.

Retention of CFSE-labeled Sca-1⁺ LMSCs in the lungs of K. pneumoniae–infected mice was evaluated using fluorescence microscopy. The numbers of CFSE-labeled Sca-1⁺ LMSCs were comparatively higher in the lung sections from the mice that were sacrificed at 4 h after LMSC administration (2.26 ± 0.26 cells/image) than those mice that were sacrificed at 44 h after LMSC transplantation (1.69 ± 0.23 cells/image) (Supplemental Fig. 6A, 6B). Moreover, Sca-1⁺ LMSCs formed contiguous clusters of four to seven cells (appearing as green in Supplemental Fig. 6A, 6B) interspersed throughout the lung parenchyma of the K. pneumoniae–infected mice.

To determine the effect of treatment with Sca-1⁺ LMSCs on K. pneumoniae–induced lung inflammation, we analyzed TWBCs, as well as different inflammatory cell populations, in BALF from mice 48 h after K. pneumoniae infection/Sca-1⁺ LMSC treatment. Treatment with Sca-1⁺ LMSCs markedly decreased inflammatory cell infiltration into the BALF of mice infected with K. pneumoniae (Fig. 5A). The TWBC number in BALF at 48 h after K. pneumoniae infection of mice treated with PBS was significantly higher than those treated with Sca-1⁺ LMSCs (Fig. 5A). In both K. pneumoniae–infected and K. pneumoniae–infected/Sca-1⁺ LMSC-treated groups, the vast majority of infiltrating cells in the BALF were neutrophils. Treatment with Sca-1⁺ LMSCs reduced the infiltrating neutrophils in the K. pneumoniae–infected group. In contrast, macrophages were the major inflammatory cell population in the BALF of the infection control (PBS) group (Fig. 5A). However, administration of 3T3 MEFs had no significant effect on TWBCs, as well as different inflammatory cell populations, in the BALF of K. pneumoniae–infected mice (Supplemental Fig. 7A). To determine the regulatory effect of Sca-1⁺ LMSCs on inflammatory cytokine and chemokine responses in the lung, we analyzed the protein levels of different cytokines, chemokines, and growth factors in lung supernatants. As expected, lung supernatants from the infection control (PBS) mice showed very low levels of inflammatory markers, whereas lung supernatants from those infected with K. pneumoniae showed higher levels of cytokines (IL-6, TNF-α, IL-1α, and IL-1β) and chemokines (MIP-1α, MIP-2, MCP-1, and KC) (Fig. 5B). Treatment with Sca-1⁺ LMSCs significantly reduced the levels of IL-6, TNF-α, IL-1α, IL-1β, MIP-1α, MIP-2, MCP-1, and KC in the lungs of the K. pneumoniae–infected group (Fig. 5B). However, treatment with Sca-1⁺ LMSCs did not alter the levels of LIX, GM-CSF, IFN-γ, IL-10, Angiopoietin-2, or SDF1 (Fig. 5B). Next, we measured the extent of inflammation in H&E-stained lung sections from K. pneumoniae–infected (Fig. 5C, upper left panel) and Sca1⁺ LMSC-treated (Fig. 5C, bottom left panel) mice. We found that the lung pathology scores were lower in the lung sections from K. pneumoniae–infected and Sca1⁺ LMSC-treated mice when compared with K. pneumoniae–infected mice that received PBS (Fig. 5D). H&E-stained lung sections from K. pneumoniae–infected and Sca1⁺ LMSC-treated mice showed attenuated inflammatory foci (Fig. 5C, bottom left panel) containing lower numbers of inflammatory cells (Fig. 5C, bottom magnified image), as well as granular microcolonies of bacteria (indicated by arrows in (Fig. 5C), when compared with the K. pneumoniae–infected mice that received PBS (Fig. 5C, upper panels).

To evaluate the effect of Sca-1⁺ LMSCs on bacterial burden in the lungs and in extrapulmonary organs, we determined the viable bacterial counts in the lungs, BALF, spleen, and liver. Interestingly, we found reduced bacterial burden in the lungs (Fig. 6A) and BALF (Fig. 6B) of Sca-1⁺ LMSC-treated mice when compared with K. pneumoniae–infected and PBS-treated mice. Consistent with the findings in the lungs, Sca-1⁺ LMSC-treated mice also showed reduced dissemination to the spleen (Fig. 6C) and liver (Fig. 6D). Similar to 3T3 mouse lung fibroblast-mediated effects observed during Escherichia coli–induced pneumonia (49), transplanted 3T3 MEFs did not significantly influence the bacterial burden in the lungs, BALF, and extrapulmonary organs of K. pneumoniae–infected mice (Supplemental Fig. 7B–E).

We next explored whether the enhanced bacterial clearance in LMSC-treated mice contributes to survival during bacterial pneumonia caused by K. pneumoniae. At an inoculum of 0.5 × 10⁴ CFUs, ∼50% of PBS-treated mice and 3T3 MEF-treated mice (as control) succumbed to infection at 96 h (Fig. 6E). Analysis by log-rank (Mantel-Cox) test showed enhanced survival in Sca-1⁺ LMSC-treated mice compared with PBS and 3T3 MEF-treated mice (Fig. 6E). Furthermore, the median survival for Sca-1⁺ LMSC-treated mice was 204 h (8.5 d) compared with 96 h (4 d) for PBS- and 3T3 MEF-treated mice (Fig. 6E).

Although the therapeutic potential of BM-MSCs is currently being examined in clinical trials for lung diseases (ClinicalTrials.gov), understanding tissue-resident LMSCs in lung diseases remains elusive. Although reports suggest that only a small subset of MSCs is responsible for their therapeutic properties, MSCs have conventionally been studied as a polyclonal product of cell culture. Interestingly, Corselli et al. (50) have suggested that enrichment or purification of well-characterized functional subsets could potentially enhance the efficiency of current clinical trials of MSCs. Therefore, in this study, we used Sca-1 as a stem cell marker to isolate a subset of MSCs from a mixed population of mouse lung cells (LMSCs). These purified Sca-1⁺ cells appear to be a homogeneous population of LMSCs that express mesenchymal Ags and differentiate into osteocytes and adipocytes when subjected to specific differentiation conditions. Moreover, these LMSCs are highly replicative in vitro. More importantly, unlike bronchoalveolar stem cells (31, 32), these Sca-1⁺ LMSCs can be cultured and expanded on noncoated plastic dishes. Purified Sca-1⁺ LMSCs show similar morphological features and, to a certain extent, surface marker profiles as MSCs isolated from bone marrow, spleen, brain, lungs, pancreas, kidneys, muscle, and thymus (8, 23).

There is limited knowledge concerning the morphology and ultrastructural features of LMSCs (26). The most striking features of Sca-1⁺ LMSCs are that they are rich in autophagosomes, endosomes, lysosomes, vesicular elements, and long mitochondria. A recent study (51) using TEM showed the presence of a large number of autophagosomes in the cytoplasmic compartment of undifferentiated human BM-MSCs. TEM analysis of Sca-1⁺ LMSCs revealed ultrastructural features common for all cells with high proteosynthetic levels. However, unlike Sca-1⁺ LMSCs (this study) and human BM-MSCs (51), the cytoplasm of adipose-derived MSCs (52) contains large numbers of small lipid droplets and electron-dense lamellar bodies.

MSCs also exhibit broad immunoregulatory properties and have the ability to influence both innate and adaptive immune responses. The mechanisms by which BM-MSCs and other tissue-derived MSCs exert their immunomodulatory effects are thought to be mediated primarily through the release of soluble factors and/or direct cell-cell contact (15, 53). Much of what is known about the immunoregulatory properties of MSCs has been learned through coculture of MSCs with innate and adaptive immune cells. Recent reports have shown that MSCs interact with the innate immune system resulting in both anti-inflammatory and proinflammatory effects (54). Further, MSCs secrete numerous soluble factors that exert immunosuppressive effects by modulating both innate (macrophages, monocytes, neutrophils, dendritic cells, mast cells, myeloid-derived suppressor cells, and NK cells) and adaptive (CD4⁺ T cells, FOXP3 T regulatory cells, CD8⁺ T cells, B cells, and Th17 cells) immune responses (55–57). However, when compared with MSCs from other tissue sources, very little is known about the immunomodulatory role of LMSCs (29, 30, 58–60). In this study, we evaluated the immunoregulatory properties of Sca-1⁺ LMSCs by coculturing them with BMDMs and BMDNs, two key innate immune cells that play an important role in host protection against bacterial infection. Our experimental results show that treatment with Sca-1⁺ LMSCs remarkably inhibited the LPS-induced secretion of TNF-α and IL-1β from BMDMs and TNF-α from BMDNs. Similarly, murine BM-MSCs, as well as BM-MSC–derived supernatants, inhibited the secretion of inflammatory cytokines TNF-α, IL-12p70, IL-6, and IFN-γ by LPS-stimulated peritoneal macrophages (61). Importantly, unlike BM-MSCs (62), Sca-1⁺ LMSCs did not secrete the inflammatory cytokine TNF-α, when left unstimulated or after stimulation with LPS. However, the mechanism(s) by which Sca-1⁺ LMSCs suppress inflammatory cytokine production by BMDMs and BMDNs remains to be elucidated.

Apart from immunomodulating properties, BM-MSCs are also endowed with the ability to directly affect the infectious agent. These antimicrobial effects are mediated through the secretion of antimicrobial compounds (LL-37, β defensin-2, lipocalin-2) (49, 63, 64) or through the intensification of phagocytosis of monocytes and macrophages (65). In contrast, nothing is known about the antimicrobial properties of LMSCs. Bacterial killing by neutrophils (35), macrophages (44), mast cells (45), and γ δ T cells (46) is a major contributor to the clearance of K. pneumoniae. Therefore, we compared the bacterial killing ability of Sca-1⁺ LMSCs with that of BMDNs and BMDMs in vitro. Our results show that both Sca-1⁺ LMSCs and BMDNs exhibit similar extracellular killing effects when cultured with low MOI of K. pneumoniae. Moreover, when cultured with a higher MOI of bacteria, only LMSCs and LMSC-derived supernatants were able to kill K. pneumoniae at 1.5- and 3-h time points. However, at a later time point (6 h), LMSCs, LMSC-derived supernatant, and BMDNs failed to kill K. pneumoniae. Sca-1⁺ LMSCs secrete large amounts of antimicrobial compounds, such as CRAMP, LTB4, and lipocalin-2, even under resting conditions, an effect that was increased after infection with K. pneumoniae. Intriguingly, treatment with CRAMP peptide inhibited the extracellular growth of K. pneumoniae in vitro. LMSCs also secrete increased amounts of H₂O₂ postinfection with K. pneumoniae, which might have partly contributed to the inhibition (66) of the extracellular growth of this Gram-negative bacterium in vitro.

Although multiple studies have shown the contribution of soluble factors secreted by MSCs to killing of bacterial pathogens, not much is known about the direct interaction and subsequent phagocytic uptake and intracellular killing of infectious agents by MSCs. Recently, two studies have demonstrated binding and phagocytic uptake of Mycobacterium tuberculosis by BM-MSCs (67, 68). Another study identified the direct interaction and internalization of both Gram-negative (E. coli) and Gram-positive (Streptococcus pyogenes and Staphylococcus aureus) bacteria by adipose-derived MSCs (69). In our study, we compared the phagocytotic uptake and intracellular killing abilities of Sca-1⁺ LMSCs and BMDMs using opsonized, live K. pneumoniae in vitro and found that both Sca-1⁺ LMSCs and BMDMs were efficient in the phagocytotic uptake of K. pneumoniae. Interestingly, Sca-1⁺ LMSCs were more effective at controlling the intracellular replication of K. pneumoniae than were BMDMs at 4 h postinfection.

Complement plays an important role in the opsonization of circulating pathogens. CR3, CR4, and CRIg are involved in the phagocytosis of opsonized pathogens (47, 48). The B7 family–related protein, CRIg, is selectively expressed on tissue-resident macrophages (liver Kupffer cells and peritoneal macrophages) and dendritic cells (70) and is a major receptor for phagocytosis of complement-opsonized bacteria in tissue-resident macrophages (48). Our data show increased expression of CRIg in Sca-1⁺ LMSCs, suggesting the possible involvement of CRIg in the phagocytotic uptake of opsonized K. pneumoniae by Sca-1⁺ LMSCs.

Both phagocytosis and autophagy are cellular catabolic pathways that use lysosomes to digest particles (71). Recent studies reported interactions between phagosomes formed by phagocytosis and autophagosomes formed by autophagy (71–73). The interactions between these two forms of cellular eating processes represent a new dimension in host defense against both extracellular and intracellular bacterial pathogens. Our study also shows that activation of the autophagic pathway in LMSCs after phagocytotic uptake of the opsonized K. pneumoniae and pretreatment with the autophagy inhibitor, 3-MA, results in an enhanced ability of cells to control K. pneumoniae intracellular replication 1 h postinfection, as evidenced by lower numbers of visible K. pneumoniae in LMSCs. Conversely, pretreatment with the autophagy inducer rapamycin impairs the ability of the cell to control K. pneumoniae replication early after (1 h) infection, as evidenced by greater numbers of viable K. pneumoniae in LMSCs. In addition, the autophagy-associated proteins Beclin-1, ATG7, and LC3 are highly expressed in LMSCs, and infection with K. pneumoniae further increases the levels of Beclin-1, ATG7, and LC3 at 1 h postinfection. Further, pretreatment with 3-MA decreases the levels of these autophagy-associated proteins in LMSCs at 1 h postinfection, suggesting that autophagy is activated during the phagocytotic uptake of K. pneumoniae by LMSCs. Most bacterial pathogens are rapidly degraded in the phagolysosome, and ROS are critical weapons in the phagocyte arsenal (74). Interestingly, phagocytotic uptake and intracellular killing of K. pneumoniae are clearly associated with increased production of intracellular ROS in LMSCs. Our data also show that LMSCs are more effective than BMDMs at clearing phagocytosed K. pneumoniae because almost all bacteria were gone by 4 h postinfection.

Next, we further confirmed the ability of Sca-1⁺ LMSCs to phagocytize live K. pneumoniae in vitro using immunofluorescence microscopy and confocal microscopy. In addition, flow cytometric analysis also revealed the phagocytic uptake of HI Kp-FITC by both Sca-1⁺ LMSCs and BMDMs. Thus, the phagocytic activity of Sca-1⁺ LMSCs and BMDMs is likely to contribute to bacterial clearance/elimination during acute K. pneumoniae infection. Interestingly, human BM-MSCs also exhibit marked autophagy, and small interfering RNA knockdown of autophagy inhibitor beclin-1 and inhibition of NO have been found to enhance the intracellular survival of M. tuberculosis (67).

Although the existence of region-specific stem or progenitor cells in the murine respiratory tract has been demonstrated (75–78), the identification and localization of LMSCs in different regions of the murine and human lungs has been challenging because of the lack of specific markers. Accordingly, the role of endogenous lung resident MSCs in normal tissue homeostasis and disease settings is still unclear. A recent study (79) demonstrated the presence of abundant numbers of LMSCs in the lung parenchymal tissues of humans. However, sensitization and challenges with chicken OVA (80) increased the numbers of LMSCs in the lungs of mice, suggesting that the allergens may act through lung-resident MSCs to promote processes that are involved in the pathogenesis of asthma. Few studies have quantified the efficiency of LMSCs transplantation. Badri et al. (81) demonstrated the in vivo retention and interaction of the PKH-26 or DsRed lentivirus-labeled human LMSCs with alveolar epithelial cells in murine lungs. These human LMSCs were found to persist in murine lungs for up to 6 months. In another elegant study, Jun et al. (58) have identified ATP-binding cassette superfamily G member 2 as a LMSC marker to localize and study the immunomodulatory and protective effects of these cells in vivo in both murine and human tissues. Bleomycin treatment induced the loss of these endogenous LMSCs and elicited inflammation, fibrosis, and pulmonary arterial hypertension. Rescuing the deficient endogenous LMSC population via i.v. administration of LMSCs attenuated lung injury in mice (58). Another study (82) using fluorescence in situ hybridization analysis revealed that the primary MSCs in human transplanted lungs are perivascularly located in CD90/CD105⁺ tissue-resident cells. Interestingly, all MSC samples isolated from biopsies taken as long as 16 years posttransplantation showed donor sex karyotype (82). This finding contributes further evidence that these transplanted LMSCs are long-lived and/or self-renewing populations of lung stroma. However, it is yet to be shown clinically that improved MSC persistence leads to enhanced efficacy of MSC-based therapies.

The observed therapeutic effect of BM-MSCs is mediated by the production of immunomodulatory factors during the initial days after BM-MSC transplantation (83). A number of growth factors secreted by BM-MSCs and other tissue-derived MSCs are critical for neoangiogenesis (84), and vascular endothelial growth factor has been shown to be a key mobilizer of cardiac stem cells in the infarcted heart (85). BM-MSC administration stimulates the proliferation of cardiac stem cells in the infarcted heart of pigs (86) and increased the proliferation potential of bronchoalveolar stem cells in a mouse model of bronchopulmonary dysplasia (87). However, no prior studies have closely examined the effects of LMSCs on lung-resident stem or progenitor cells in vivo. In our study, we observed the retention of increased numbers of CFSE-labeled Sca-1⁺ LMSCs in the lung parenchyma of K. pneumoniae–infected mice at 4 h after administration when compared with the lungs of K. pneumoniae–infected mice that were sacrificed at 44 h after LMSCs transplantation.

Despite that the role of BM-MSCs and other tissue-derived MSCs in inflammatory and infectious diseases is fairly well established (11, 12, 14), the therapeutic potential of lung-derived stem/MSCs in bacterial infection and endogenous immune responses is unclear. Nonetheless, most studies have used the Gram-negative bacterium E. coli to infect mice or rats, reporting reduced lung inflammation, pulmonary edema, alveolar epithelial permeability, and enhanced bacterial clearance after the i.t. or i.v. administration of MSCs (49, 65, 88, 89). E. coli infection in animals is usually associated with acute onset of severe lung inflammation and clearance of bacteria and related more to acute lung injury than pneumonia. However, K. pneumoniae infection in animals induces a gradually evolving inflammatory response with a steadily increasing bacterial load in the lungs and extrapulmonary organs (34, 90, 91), which largely resembles the clinical scenario of bacterial pneumonia. i.v. infusion has been used as a preferred route of cell delivery for a large number of preclinical studies and is the favored route of administration in human clinical trials as well (11, 53, 92, 93). A major pitfall of i.v. administered MSCs is their inability to migrate across pulmonary capillaries to the lung stroma (94, 95), and these i.v. injected MSCs have short half-lives (94). In our study, we administered a single dose of Sca-1⁺ LMSCs to mice by i.t. route for acute bacterial pneumoniae and two doses of Sca-1⁺ LMSCs o.p. for the survival studies.

Our study using purified Sca-1⁺ LMSCs showed multiple beneficial effects of these cells in pneumonia caused by the Gram-negative bacterium, K. pneumoniae. Treatment with Sca-1⁺ LMSCs reduced the levels of inflammatory cells, predominantly neutrophils and macrophages, in the BALF of K. pneumoniae–infected mice. Interestingly, this effect was clearly associated with reduced monocyte/macrophage/neutrophil-attracting chemokines, such as MIP-1α, MIP-2, KC, MCP-1, and the inflammatory cytokines TNF-α, IL-1β, IL-1α, and IL-6, in the lungs of K. pneumoniae–infected and LMSC-treated mice. Our results are similar to the results obtained using BM-MSCs, which attenuated the level of neutrophils in the BALF of mice infected with E. coli (65), K. pneumoniae (96), and Streptococcus pneumoniae (97). Similar to what was seen with BM-MSCs (96), treatment with LMSCs did not alter the level of the anti-inflammatory cytokine IL-10 in the BALF of K. pneumoniae–infected mice. Interestingly, in the in vivo models of bacterial pneumonia and acute lung injury/acute respiratory distress syndrome induced by live bacteria, BM-MSCs demonstrated the ability not only to attenuate inflammation but also to improve bacterial clearance (49, 65, 97). The study by Hackstein et al. (96) showed the protective effect of purified murine BM-MSCs in a mouse model of pneumonia caused by hypervirulent K. pneumoniae, which was associated with reduced pulmonary inflammation and improved survival in mice; however, treatment with BM-MSCs did not reduce bacterial growth in the lungs or extrapulmonary organs of the mice. In this study, a single dose of purified PαS MSCs was i.t. administered into K. pneumoniae–infected C57BL/6J mice, and the survival of the mice was monitored up to 5 d. Another recent study by Perlee et al. (91) showed profound immunomodulatory effects of i.v. administered human adipose-derived MSCs in the lungs, resulting in reduced lung inflammation and bacterial burden in the lungs and distal organs, but did not influence distal organ injury during pneumosepsis caused by K. pneumoniae. However, this study did not investigate whether i.v. administration of adipose-derived MSCs can protect the mice from death caused by K. pneumoniae–induced pneumosepsis in mice. In our study, administration of Sca-1⁺ LMSCs was found to not only attenuate pulmonary inflammation but also promote marked bacterial clearance in the lungs and in extrapulmonary organs, as well as rescue mice from death caused by K. pneumoniae. As anticipated, administration of 3T3 MEF neither decreased the levels of inflammatory cells in the BALF nor attenuated the bacterial growth in the lungs and extrapulmonary organs of K. pneumoniae–infected mice. In our study, we monitored the survival of the mice for up to 15 d. Similar to the protective effects obtained using BM-MSCs (96), i.t. delivery of Sca-1⁺ LMSCs rescued 85% of the mice from death caused by K. pneumoniae at 4 d postinfection, whereas 25% of the LMSC-treated mice survived at 15 d postinfection. These different levels of MSC-mediated protection observed during K. pneumoniae infection in our study and other studies (91, 96) are likely due to the source of MSCs used; dose, route, time, and the frequency of administration of MSCs, and the amount of the inoculum (CFUs) of K. pneumoniae used to infect the animals; as well as the days taken to monitor the survival of the mice. Future studies are warranted to investigate whether the immunomodulatory, antibacterial, and reparative/regenerative potential of Sca-1⁺ LMSCs can be enhanced by altering the dose, route, time, and frequency of delivery, as well as by preconditioning of the Sca-1⁺ LMSCs with bacterial LPS, inflammatory cytokines, or hypoxia (98). Also, additional studies are required to determine whether Sca-1⁺ LMSCs can induce host protection in the setting of influenza virus–associated bacterial pneumonia (99), as well as in pneumonia caused by the multidrug-resistant K. pneumoniae (100) and Gram-positive bacterial species, including S. pneumoniae.

We thank Marilyn Dietrich at LSU School of Veterinary Medicine (SVM) for helping with flow cytometry, and Yuliya Sokolova and Xiaochu Wu at LSU Shared Instrumentation Facility for assistance with light and electron microscopic studies. We thank Mariano Carossino at LSU School of Veterinary Medicine for helping with NanoZoomer-HT.

The authors have no financial conflicts of interest.