A total of 15 male Sprague-Dawley rats (age, 8–9 weeks; weight, 200–250 g) provided by Samtako Bio Korea (Osan, Korea) were used. All experimental procedures were approved by the Institutional Animal Care and Use Committee in Daejeon St. Mary's Hospital, Catholic University of Korea (Daejeon, Korea). All rats were housed under a 12 h light/dark cycle, a 50–60% humidity, and an ambient temperature of 22–24°C. In addition, rats received ad libitum access to food and water. All procedures were conducted by the same individual to minimize variation. In order to induce ALI, LPS extracted from Escherichia coli 055:B5 (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) diluted in saline was used (20 mg/kg). ALI rats were injected intraperitoneally with LPS (5 mg/kg). In the control group, sham intervention was performed using the same amount of saline. Human MSCs were provided by The Catholic Institute of Cell Therapy (Seoul, Korea). The cells were preserved with Dulbecco's modified Eagle's medium containing 1,000 mg/l glucose, sodium bicarbonate, and pyridoxine (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) in a humidified atmosphere of 5% CO² at 37°C. The cells at passages 3–4 were isolated for in vivo experiments. All 15 rats were assigned randomly to one out of the following three groups (n=5/group): Saline-treated controls, LPS-induced ALI with saline (ALI) and LPS-induced ALI with BM-MSC (LPS+BM-MSC). At 30 min following administration with LPS, BM-MSCs (2×10⁶; 100 µl) or saline (100 µl) were injected slowly into the tail vein over 20 min.

Rats were sacrificed at 6 h following administration of saline or BM-MSCs. Blood was harvested via cardiac puncture and plasma was centrifuged for 10 min at 3,000 × g at 37°C. Plasma samples were frozen at −70°C prior to analyze alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate, blood urea nitrogen (BUN), and creatinine (CREA) using an IDEXX VetTest^® Chemistry Analyzer (IDEXX Laboratories, Inc., Westbrook, ME, USA).

The trachea was incised and bronchoalveolar lavage (BAL) fluid was obtained from the right lung. Total cells were counted using the LUNA automated cell counter (Logos Biosystems, Annandale, VA, USA) following the manufacturer's instructions. An aliquot of 200 µl of the diluted 500 µl pellet was cytospinned at a speed of 180 × g at 4°C, transferred to a slide, and stained with Wright-Giemsa stain at 24°C for 6 min. The 100-cell differential count was performed for estimating the percentage of neutrophils under a light microscope (magnification, ×400; Olympus Corporation, Tokyo, Japan) in 4 ideal slide zones. Rat left upper lobe lung tissue was fixed with 10% formalin overnight at 24°C, embedded in paraffin, and stained with hematoxylin and eosin at 24°C for 1 min. Each lung section was assessed independently by two clinical pathologists using microscopy (magnification, ×100) to evaluate the severity of lung injury. The lung injury score (LIS) comprises four components (hemorrhage, alveolar capillary congestion, inflammatory cells infiltrating the interstitium or airspace, and the alveolar wall thickness), with each component scored on a 5 point scale (0 = minimal damage, 1 = mild damage, 2 = moderate damage, 3 = severe damage, 4 = maximal damage). The LSI is the sum of all four component scores (9). The left lower lobe was frozen at −70°C prior to analysis of miRNA expression.

Total RNA was isolated from the rat lung tissue using TRI Reagent (Molecular Research Center, Inc., Cincinnati, OH, USA) following the manufacturer's protocols. The total RNA pellet was dissolved in nuclease-free water and its quantity and yield was estimated using a 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA). Rat miRNA expression profiling was performed using miRCURY LNA miRNA PCR Assays (Exiqon; Qiagen GmbH, Hilden, Germany). The seventh generation array included with this assay contains ~3,100 capture probes, covering all human, mouse and rat miRNAs annotated in miRBase (www.mirbase.org), as well as all viral miRNAs related to these species. Processed microarray slides were scanned using a G2565CA microarray scanner system (Agilent Technologies, Inc.) and imported using Feature Extraction software ver. 10.7.3.1 (Agilent Technologies, Inc.). The fluorescence intensities of each slide were quantified according to the Exiqon protocol. The results of miRNA expression were calculated as the mean ± standard error of the mean. Target prediction for functional estimation of the differentially expressed miRNAs was conducted using miRanda (34.236.212.39/microrna/home.do) and Targetscan ver. 7.0 (http://www.targetscan.org). The target lists of dysregulated miRNAs were submitted separately to the functional annotation tool provided by the Database for Annotation, Visualization, and Integrated Discovery (DAVID, https://david.ncifcrf.gov), ver. 6.7. The predicted targets were annotated according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (19,20).

To visualize the predicted target genes associated with dysregulated miRNAs, the Network Analyzer plug-in (apps.cytoscape.org/apps/with_tag/networkanalysis) of Cytoscape 3.6 was used (21).

The Kruskal-Wallis test and a one-way analysis of variance for non-normally distributed data was used to test the median difference between each variable in the three groups. A post hoc Tamhane's T2 test was performed for pairwise comparison of subgroups. The box-and-whisker plot was used to present the data distribution in each figure. A line is drawn inside the box at the median and the box portion of the plot is defined by two lines at the 25th percentile and 75th percentile. The distance between the lower (25th percentile) and upper (75th percentile) lines of the box is defined as the inter-quartile range. The two whisker boundaries indicate the 10th (lower) and 90th (upper) percentiles. MedCalc Statistical Software Version 17.6 (MedCalc Software bvba, Ostend, Belgium) was used for statistical investigation. P<0.05 was considered to indicate a statistically significant difference.

The presence of moderate pulmonary injuries including hemorrhage, congestive alveolar capillaries, inflammatory cell infiltration, and alveolar wall thickening in the LPS group were revealed via histopathological examination, compared with the LPS+BM-MSC group. The LIS was used to estimate the influence of BM-MSCs on lung injury. Similar to the histopathological examination, LIS (median, 10) in LPS rats was significantly higher than in controls (P<0.05). LIS (6) in BM-MSC rats was significantly lower than in LPS rats (P<0.05). The total cell count and neutrophil percentage in BAL fluid were counted to evaluate the protective role of BM-MSCs in LPS-induced ALI. As a result, BM-MSCs markedly reduced the number of total cell count (control, 0.3; LPS, 2.4; LPS+BM-MSC, 1.3), and significantly reduced the neutrophil percentage (Control, 1; LPS, 91; LPS+BM-MSC, 60; P<0.05) in the BAL fluid compared with the LPS group (Fig. 1).

Organ damage was estimated by measuring serum biochemical indicators 6 h following administration with LPS. The levels of four analytes (excluding ALT) were significantly elevated by LPS (Fig. 2). The levels of liver enzymes ALT and AST released into circulation upon injury were lower in ALI treated with BM-MSCs than in ALI only (ALT, P=0.223; AST, P<0.05). In particular, AST was significantly decreased (control, 32; LPS, 300; LPS+BM-MSC, 70). The concentration of lactate, typically used as an indicator of tissue hypoperfusion, was lower (control, 1.3; LPS, 5.2; LPS+BM-MSC, 2.7) in ALI with BM-MSCs compared with ALI alone (P<0.05). In kidney injury, blood urea nitrogen (BUN; control, 15; LPS, 49; LPS+BM-MSC, 18) and creatinine (CREA; control, 0.2; LPS, 0.5; LPS+BM-MSC, 0.2) levels were also significantly lower in ALI with BM-MSCs, compared with ALI alone (BUN, P<0.05; CREA, P<0.05).

miRNA expression profiling was performed to identify the alteration in miRNAs in the lungs of rats with LPS-induced ALI. A total of 128/690 rat miRNAs were expressed differently between the ALI and control groups. They included 68 upregulated and 60 downregulated miRNAs, respectively. Furthermore, 15 miRNAs were significantly upregulated or downregulated (fold-change ≥2) in the ALI group, compared with the control group (P<0.05). Five of these miRNAs (miR-760-3p, miR-223-3p, miR-449c-3p, miR-503-3p and miR-142-3p) were upregulated and 10 (miR-100-5p, miR-199a-5p, miR-99a-5p, miR-199a-3p, miR-181a-5p, miR-497-5p, miR-191a-5p, miR-28-5p, miR-3065-5p and miR-196b-3p) were downregulated following LPS treatment (Table I).

Among 690 rat miRNAs, 95 were differentially expressed between ALI in the BM-MSCs group and the control group. They included 66 upregulated and 29 downregulated miRNAs. Furthermore, nine miRNAs were significantly upregulated or downregulated in the ALI group, compared with the control group (fold-change ≥2; P<0.05). Among the nine miRNAs, five (miR-1843-3p, miR-323-3p, miR-183-5p, miR-182 and miR-196b-3p) were upregulated and four (miR-547-3p, miR-301b-5p, miR-503-3p and miR-142-3p) were downregulated following treatment with BM-MSCs (Table II). To investigate the effects of BM-MSCs in ALI, the inversely expressed miRNAs in ALI with BM-MSCs compared with ALI were selected. Three miRNAs were inversely expressed in the two groups. The expression of two of these miRNAs (miR-503-3p and miR-142-3p) was increased in the LPS group, and decreased in the BM-MSC group. The miR-196b-3p was downregulated in the LPS group, but upregulated in the BM-MSC group.

Gene ontology and KEGG pathway annotation analyses via DAVID ver. 6.7 revealed annotated KEGG pathways for the altered expressed miRNAs (Tables III and IV). The miRNA pathways were associated with inflammation, the immune response and cellular apoptosis.

It was observed that miR-503-3p and miR-142-3p were associated with myeloid/lymphoid or mixed-lineage translocated to, 1; cyclin T2; and granzyme B gene is a serine protease with a notable role in the rapid induction of target cell apoptosis (22). It was also predicted that miR-503-3p and miR-196-3p were correlated with muscleblind-like protein 1 and DNA damage regulated autophagy modulator 1 (DRAM1) genes, whereas miR-142-3p and miR-196b-3p were associated with activator of heat shock protein ATPase 2, tyrosine-protein kinase ABL2 (ABL2), homeobox protein Nkx2-3 (Nkx2-3), Ras-responsive element-binding protein 1 (RREB1) and Musashi RNA binding protein 2.