CXCR4 Receptor Overexpression in Mesenchymal Stem Cells Facilitates Treatment of Acute Lung Injury in Rats^*

Reagents

The expression construct for CXCR4 was developed by cloning the rat CXCR4 coding sequence into the GFP lentiviral vector pCDH-CMV-MCS-EF1-copGFP (System Biosciences; Mountain View, CA) at XbaI and EcoRI restriction sites (Invitrogen). Constructs were isolated from bacteria with the plasmid small kit without endotoxin (Omega Bio-tek; Norcross, GA) and transfected with packaging plasmids pLP1, pLP2, pLP/VSV-G (ViraPower Lentiviral Expression Systems; Invitrogen) with Lipofectamine 2000 into 293T cells (a gift from Professor Yan Yaping; Tianjin Medical University) in DMEM with glucose (Invitrogen).

Rat MSCs were cultured in SD rat bone marrow MSC dedicated complete medium (Cyagen Biosciences; Guangzhou, China). In vitro migration assays were performed in 8-μm hanging Transwell chambers (Corning China; Shanghai, China) with SDF-1α (PeproTech; Rocky Hill, NJ). Hematoxylin-eosin staining dye (Nanjing Jiancheng Bioengineering Institute; Nanjing City, China) was used to stain cells.

The following antibodies were used for immunocytochemistry: CXCR4 rabbit anti-rat antibody, VCAM-1 (vascular cell adhesion molecule-1), and ICAM-1 (intercellular adhesionmolecule-1) rabbit anti-rat antibodies (Santa Cruz Biotechnology, Inc.; Dallas, TX); vWF and SP-C rabbit anti-rat antibodies (Beijing Biosynthesis Biotechnology Co.; Beijing, China); Ki67 rabbit anti-rat antibody (Abcam; Cambridge, MA); Cy3-labeled donkey anti-rabbit fluorescence secondary antibody (Jackson ImmunoResearch Laboratories, Inc.; West Grove, PA). DAPI (Sigma) was used for staining of nuclei. IL-6, VEGF, IL-10, and TNF-α enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems; Minneapolis, MN) were used for the detection of factors in cell supernatants or bronchoalveolar lavage (BAL) fluid.

Ethics Statement

All animal protocols were approved by the Animal Care Committee in Dalian Medical University (Dalian, China) and performed according to institutional guidelines.

Animals

Sprague-Dawley rats (age 4–6 weeks) were purchased from the experimental animal center of Dalian Medical University (SCXK (Liaoning) 2008-0002).

Primary Culture and Identification of MSCs

Male rats were anesthetized (10% urethane for 10 min), abdomens were disinfected with 75% alcohol, and long bones (femur and tibia) of the two hind limbs were prepared for the isolation of MSCs. Both ends of each long bone were cut off, and the marrow cavity was rinsed with low glucose DMEM repeatedly. The supernatant was centrifuged at 1200 rpm for 6 min, and the pelleted cells were collected as the rat MSCs. Cells were counted and plated (5 × l0⁵) in 25-cm² flasks and cultured at 37 °C, 5% CO₂ in SD rat bone marrow MSC dedicated complete medium containing 10% fetal bovine serum (FBS). The medium was changed at 24 and 48 h, 50 and 100%, respectively, and the non-adherent cells were discarded. The attached cells were labeled P0, and the medium was changed once every 3 days. When cells became 90% confluent, cultures were dissociated with trypsin (1–1.5 ml 0.25% trypsin) at 37 °C 2–3 min, and serum was used to inactivate the enzyme. Cells were expanded, and at the third passage (P3), cells were characterized for cell surface markers by flow cytometry with the following fluorescently labeled antibodies: CD29-PE, CD34-FITC, CD44-FITC, CD45-FITC, and CD90-FITC (11). Quantitative flow cytometric analysis was also performed to determine the percentages of P0 through P3 cells positive for CXCR4 and CXCR7, a second receptor for SDF-1α on MSCs. Cell surface adhesion molecules VCAM-1 and ICAM-1 (23, 24) were examined by immunocytochemistry.

Cloning of CXCR4 into Recombinant Lentiviral Vector

CXCR4 was amplified from rat mRNA by RT-PCR, gel-purified, and ligated with T4 DNA ligase into the pMD18-T vector. The ligation was transformed into competent Escherichia coli, and the transformation was plated onto an ampicillin plate. Colonies were picked 16 h later and inoculated into LB medium. Bacteria were grown in shaking culture at 37 °C for 18 h. Plasmid was extracted by the alkaline lysis method, digested with restriction endonucleases XbaI and EcoRI, and confirmed by PCR (pT-CXCR4). pT-CXCR4 was sequenced and ligated to the lentiviral vector pCDH-CMV-MCS-EF1-copGFP, which carries the green fluorescent protein (GFP) gene, to develop a construct co-expressing rat CXCR4 and GFP. pT-CXCR4 and the lentiviral vector were digested with XbaI and EcoRI, respectively. The targeting genes and the vector segment were recloned and ligated (T4 DNA ligase). Plasmid was isolated and digested to confirm that the CXCR4 gene fragment had inserted into the vector correctly. The final constructs were named L.v.-CXCR4 (co-expression CXCR4-GFP) and L.v.-GFP (expression GFP only) (25).

Lentiviral Infection of MSCs with Expression Constructs

L.v.-CXCR4-GFP and L.v.-GFP, along with packaging plasmids pLP1, pLP2, and pLP/VSV-G, were transfected into 293T packaging cells with Lipofectamine 2000 (Invitrogen). Green fluorescence was observed in both groups under a fluorescence microscope within 24 h. Supernatants containing lentivirus were collected from 293T cells at 36 and 72 h after transfection, centrifuged at 3000 rpm for 20 min to remove cellular debris, and filter-sterilized (0.22 μm). Supernatants were centrifuged at 5000 rpm for 1 h for a cut-off of 100 kDa to obtain concentrated viral titers. MSCs (P3) were infected with concentrated viral supernatant, and expression of CXCR4 was detected by immunocytochemistry 3 days after transfection.

In Vitro Migration Assay

Hanging transwell assay chambers (8 μm pore diameter) were set up in a 24-well plate. CXCR4- (5 day infection), GFP-, or Mock-infected MSCs were suspended in 100 μl (1 × 10⁴ cells) and placed in the upper chamber. Varying concentrations of SDF-1α (50, 100 ng/ml in 600 μl) were added to the lower chamber. Transwell assays were incubated at 37 °C in 5% CO₂ for 5 h, and the number of cells that crossed and adhered to the membrane was determined. The chamber was rinsed, and a cotton swab was used to remove cells, which did not migrate, from the plating surface of the membrane. Adherent cells were fixed with 4% paraformaldehyde for 15 min at room temperature, stained 10 min with 0.5% crystal violet, and counted. The number of transmembrane cells in 10 different fields under the microscope was averaged (26).

Growth Curves

CXCR4-, GFP (control MSCs)-, and Mock-infected MSCs were seeded on day 5 at 6 × 10³/well in a 96-well plate, and the medium was changed once every 2 days. Cells from 6 wells were dissociated, counted, and averaged each day over 9 days. For the cell growth curve, the average cell numbers (vertical axis) were plotted as a function of time in days (horizontal axis).

Differentiation of MSCs into Lung Tissue

Conditioned medium (CM) for differentiation experiments was prepared from lung tissue inflammation induced by LPS in cells in vitro. Male rats (120–150 g) were anesthetized (10% urethane), the chest was opened, and the animals were perfused. The lungs were removed, and the fresh, perfused tissue was ground and pressed through a sieve (70 μm) for dissociation into single cells. Cells harvested from lung tissue were expanded and seeded into 6-well plates at a density of 5 × l0⁴/ml in 2 ml per well of complete medium containing 10% FBS. After 24 h, fresh complete medium or complete medium plus LPS (10 μg/ml) for inducing inflammation was added to 6 wells each (27). The cell supernatant was collected 4 h later, centrifuged at 3000 rpm for 10 min to remove cellular debris, and stored at −80 °C for subsequent analysis. These supernatants were the control and LPS-induced lung injury CM.

For differentiation in CM, MSCs were seeded in 24-well plates at a density of 5×l0⁴/well, and 50% CM was added 24 h later. The CM was changed once every 2 days for 8 days. At the end of 8 days, immunocytochemistry was used to determine expression of SP-C and vWF, markers for alveolar epithelial and endothelial cells (respectively), in the differentiated MSCs (28).

In Vivo Model of ALI

Rats (200–250 g) were injected intraperitoneally with LPS (10 mg/kg) (29–31) and sacrificed 72 h later. Success of the model was based on the following parameters: volume of the BAL fluid, the levels of inflammatory cytokines in serum, wet/dry weight ratio of lung tissue, and histological examination. Animals were divided into four groups of 10 animals per group: control (saline), intravenous injection of 0.2 ml PBS; ALI/model (LPS-induced acute lung injury), LPS intraperitoneal injection (10 mg/kg), 0.2 ml PBS through tail vein after 1 h; GFP-MSCs, LPS intraperitoneal injection (10 mg/kg), GFP-MSCs (1 × 10⁶/rat) through tail vein after 1 h; CXCR4-MSCs, LPS intraperitoneal injection (10 mg/kg), CXCR4-MSCs (1 × 10⁶/rat) through tail vein after 1 h. An additional 10 rats were prepared for harvesting fluids and lung tissue from animals transplanted with GFP- and CXCR4-MSCs to detect the degree of homing of MSCs after 2 weeks.

Collection of BAL Fluid and Lung Tissue Samples

BAL fluid preparation was performed as previously described (27, 29, 30). Briefly, rats were fixed in a supine position and anesthetized with ether cotton in a centrifuge tube set in the mouth and nose. The sternum was cut to expose the chest, the trachea, bronchus, and lung were dissociated, and both lungs were ligatured. The left bronchus was intubated, and three cycles of perfusion extraction were performed with 1.5 ml of saline. Lavage was performed for 30 s, and 1.0–1.2 ml BAL fluid was collected and centrifuged at 2000 rpm for 10 min. Protease inhibitor was added, and the BAL samples were stored at −80 °C for subsequent analysis. Neutrophils per ml were counted in PBS. The trachea was cannulated, and BAL was performed 6 times with 1 ml of 0.1 mm EDTA, phosphate-buffered saline. After counting the cells in the BAL fluid, cells were cytospun onto glass slides and stained with Hemacolor (EMD Chemicals; Gibbstown, NJ) for differential cell counting.

The right upper lung was cut, and the wet/dry weight ratio was determined. The right lower lung was immersed into ice-cold 4% paraformaldehyde (24 h), dehydrated, and embedded in paraffin. Paraffin blocks were sectioned (5 μm), and sections were slightly dried at room temperature followed by 40 °C constant temperature. Right middle lobes were embedded in OCT after freezing at −80 °C for frozen sections.

Detecting Cytokines by ELISA

ELISA was used to detect cytokines TNF-α, IL-6, and cytokines IL-10 and IL-6 and VEGF in BAL and supernatants, respectively, according to the manufacturer's protocols (R&D Systems).

Lung Histopathology

Sections (4 μm) from paraffin-embedded biopsies were H&E-stained and histologically examined. For each group, 10 biopsies from 5 rats were selected, and lung tissue injury was assessed based on the following histological changes (12, 31): alveolar congestion, alveolar hemorrhage, inflammatory infiltration, alveolar wall thickening, and hyaline membrane formation. The degree of injury was scored according to the following scale: no injury (0), 1–25% injury (1), 25–50% injury (2), 50–75% injury (3), 50–75% injury (4).

Lung Wet/Dry Weight Ratio

The right upper lobe of the lung was removed, and the wet weight was recorded. The lungs were placed in an incubator at 55 °C for 24 h to obtain the dry weight, and the wet-to-dry ratio was calculated (32, 33).

Detection of MSCs Homing to Lung Tissue

Frozen sections (10 μm) were prepared from right middle lung specimens. Sections were fixed in 4% paraformaldehyde (30 min), permeabilized with 0.1% Triton X-100 (20 min), and blocked with 3% BSA (1 h at room temperature). Primary antibodies (rabbit anti-rat CXCR4 and Ki67, 1:100) were added and incubated overnight at 4 °C. For detection, Cy3-labeled secondary antibody (1:150) was added, and slides were incubated in the dark (1 h). Sections were subsequently stained with DAPI (5 min in the dark), mounted with anti-fluorescence quenching mounting medium, and viewed under fluorescence. Slides were rinsed three times with PBS after each incubation. Five rats per group were analyzed, 5 sections per rat were selected, and 10 fields were photographed in each section to determine the number of CXCR4-MSCs (CXCR4⁺/GFP⁺/DAPI⁺) homing to the lungs/site of injury (34, 35). Cells triple-labeled with Ki67⁺/GFP⁺/DAPI⁺ were identified as MSCs with proliferation potential. Cells were counted with Image J software (free download available at rsb.info.nih.gov).

Statistical Analysis

Statistical analysis was performed with SPSS version 16.0 (2007; Chicago, IL). Data are expressed as the mean ± S.D., and one-way analysis of variance with Bonferroni correction for multiple comparisons was used to analyze the differences between the groups.

Confirmation of Rat MSCs with Cell Surface Markers

The cells obtained from rat bone marrow were characterized phenotypically and molecularly at P3. At P3, cells were adherent in serum-containing medium and had the long, spindle-shaped morphology typical of an MSC (Fig. 1A). Flow cytometry was used to determine expression of cell surface markers that typically characterize MSCs. The majority of cells were positive for CD29, CD44, and CD90 and negative for CD45 and CD34, indicating that the isolated cells were MSCs and could be used in subsequent experiments (Fig. 1B). Immunocytochemistry was subsequently used to detect the expression of stromal cell markers VCAM-1 and ICAM-1. The MSCs were positive for both surface adhesion molecules; the staining was uniformly relatively weak for VCAM-1 but strong for ICAM-1 (Fig. 1C).

Expression of CXCR4 in MSCs

The expression of CXCR4 and CXCR7 was examined first in MSCs at P0 by flow cytometry. Results indicated that only a small proportion of MSCs expressed CXCR4 (11.2%) and CXCR7 (16.8%) and that the expression of both receptors was gradually diminished as cells were expanded in vitro (7.1% and 6.4% in P3 cells, respectively; Fig. 2A).

The MSCs at P3 were infected with the two different lentiviral constructs. The total number of cells infected was determined based on expression of the GFP gene, which was also contained in the constructs. Within 3 days after infection, GFP (green) expression was first observed in MSCs under fluorescence microscopy (Fig. 2B). The infection with either construct was extremely efficient as the proportion of GFP expressing MSCs was 93.7% and 92.8% for CXCR4-GFP and GFP, respectively (Fig. 2C). Immunocytochemistry and flow cytometric analysis was subsequently used to determine the expression of CXCR4 in each cell population. Although the chemokine receptor was expressed in the three cell populations, CXCR4 (red) was highly expressed only in CXCR4-MSCs (Fig. 2D). The percentage of CXCR4-positive cells was as high as 90.2% in CXCR4-MSCs in contrast to 7.6% in GFP-MSCs and 7.5% in mock-MSCs (Fig. 2E).

Overexpression of CXCR4 Enhances Migration of MSCs in Vitro

CXCR4 is a receptor for the chemokine SDF-1α, and the chemokine is chemotactic for cells expressing the receptor. In these experiments the hanging Transwell chamber was used to determine whether CXCR4 expression influenced the migration of the rat MSCs in response to the chemokine SDF-1α in vitro. Cell counts revealed that proportionally more of the CXCR4-MSCs than the GFP-MSCs migrated to SDF-1α (Fig. 3, A and B). The results indicated that CXCR4 expression significantly enhanced migration of MSCs in response to the chemokine SDF-1α.

CXCR4 Does Not Affect the Proliferative Capacity of MSCs

Proliferation is tightly linked to self-renewal capabilities of stem cells. To determine whether CXCR4 altered the proliferation of MSCs, growth curves were derived for CXCR4-, GFP-, and Mock-MSCs to detect any differences in the rates of proliferation in vitro. At 5 days after infection, cells were replated and counted every second day over 9 days. The curves displayed no significant differences in proliferation among three cell types (p > 0.05; Fig. 4A). These results indicated that gene transfer and/or expression of CXCR4 does not affect the proliferative capacity of MSCs.

CXCR4 Gene Transfer Does Not Affect the Differentiation Potential of MSCs

To test whether CXCR4 gene transfer influenced differentiation, MSCs were incubated with CM made from lung cells cultured in vitro and treated with LPS. MSCs were evaluated after 8 days for expression of marker proteins SP-C and vWF for alveolar epithelial and vascular endothelial cells, respectively. By immunofluorescence, the number of cells expressing each of the markers was determined for CXCR4-, GFP-, and Mock-MSCs (Fig. 4B). Quantitative analysis showed that all cell populations differentiated with equal capability into lung and endothelial cell types. These results indicated that CXCR4 gene transfer did not affect the ability of MSCs to differentiate (p > 0.05; Fig. 4C).

Overexpression of CXCR4 Promotes MSCs to Secrete Cytokines

An important property of MSCs in the process may be the cytokines secreted when attracted to the site of injury. To detect the cytokines secreted by MSCs after lentiviral infection, ELISA was used to detect the level of VEGF and IL-6 in the supernatants of CXCR4-, GFP-, and Mock-MSCs in vitro. MSCs with expression of CXCR4 secreted high levels of VEGF and IL-6 relative to GFP-MSCs (p < 0.01; Fig. 4, D and E). These results indicated that CXCR4 expressing MSCs might ultimately display enhanced paracrine effects at a site of injury.

Overexpression of CXCR4 Enhances the Inhibition of Lung Tissue Inflammation in ALI

To study the role of CXCR4-MSCs in the treatment of ALI in rats, CXCR4- or GFP-MSCs (1 × 10⁶/rat) were transplanted via tail vein 1 h after intraperitoneal injection of LPS (10 mg/kg) into rats. Lavage fluid from bronchial alveoli was removed 72 h later. As a measure of the inflammatory response to ALI among the different treatment groups, the levels of TNF-α, IL-6, IL-10, and the neutrophil count were determined.

The levels of each of the molecules and the neutrophil count followed the expected profiles for inflammation in the ALI model. TNF-α, IL-6, and the neutrophil count were significantly increased in the untreated ALI animals relative to controls, whereas IL-10, an anti-inflammatory molecule, remained unchanged (p < 0.01; p < 0.01; p < 0.05; Fig. 5).

The injection of MSCs, either GFP- or CXCR4-MSCs, altered the inflammation profile of induced ALI in the LPS-treated animals. In both GFP- and CXCR4-MSCs treated animals, the inflammatory molecule TNF-α was reduced relative to the levels in the untreated ALI animals (p < 0.05; p < 0.01; Fig. 5), but CXCR4-MSCs had a greater effect (p < 0.05; Fig. 5A). However, there was no statistically significant difference in IL-6 levels in treated relative to untreated ALI animals (Fig. 5B). In contrast, the anti-inflammatory molecule IL-10 increased significantly when GFP- and CXCR4-MSCs were injected compared with the untreated ALI animals. Again, the change with CXCR4-MSCs was more significant (p < 0.05; Fig. 5C). Finally, the neutrophil count was reduced in ALI animals treated with GFP- or CXCR4-MSCs. However, there was no significant difference between GFP- and CXCR4-MSCs, indicating that CXCR4 had no apparent advantage in suppressing the neutrophil count (Fig. 5D). The results indicated that MSCs in general influenced inflammation, but compared with GFP-MSCs, CXCR4-MSCs had a more significant effect on the expression of TNF-α and IL-10 in the target site of tissue damage.

Overexpression of CXCR4 Enhances MSC Inhibition of Lung Injury

At a molecular level, the MSCs appeared to suppress inflammation, but the final test was whether pathological differences were still evident. To assess at a pathological level events in treated ALI animals, lung tissue was histologically examined 72 h after LPS and cell injection and compared with tissue from untreated animals. Pathological lesions were scored on H&E staining. Protein liquid exudation mainly caused by pulmonary vascular hemorrhage was apparent as well as hyperemia and neutrophil infiltration of lung tissue in ALI animals. In addition, pathological changes such as alveolar wall thickening and fracture had occurred to some degree (Fig. 6A). The score of pulmonary tissue injury was significantly increased compared with the control animals (3.06 ± 0.25; p < 0.01; Fig. 6B).

Tissue specimens from GFP-MSC-treated animals exhibited reduced bleeding and protein liquid exudation relative to untreated ALI animals, but the pathological changes of alveolar wall thickening were still observed. Overall, pathological lesions were less severe when compared with ALI animals (p < 0.05). CXCR4-MSCs improved pulmonary vascular hemorrhage, pulmonary congestion, and tissue protein liquid exudation as well as the thickening of the alveolar wall in alveolar fracture. The pathological score was 2.20 ± 0.13 (versus untreated ALI animals; p < 0.01). CXCR4-MSCs appeared to suppress pathological damage more effectively than GFP-MSCs. The pathological differences between these two groups reached statistical significance (p < 0.01).

To investigate the underlying mechanisms of CXCR4-MSCs in the treatment of invasive injury by ALI in rats, the wet/dry weight ratio and total protein content in BAL were also examined. The wet/dry ratio increased significantly compared with the normal group 72 h after intraperitoneal injection of LPS (Fig. 6C; p < 0.01). When treatment with GFP- and CXCR4-MSCs by injection via tail vein was performed, the ratio decreased significantly compared with untreated ALI animals (p < 0.05, p < 0.01). Furthermore, the change with CXCR4-MSCs was significantly greater than with GFP-MSCs (5.13 ± 0.37 versus 6.54 ± 0.25; p < 0.05).

The results of the determination of the total protein content in BAL demonstrated that protein in BAL from untreated ALI animals was significantly higher than in controls (p < 0.01) and that GFP- and CXCR4-MSCs significantly reduced the concentration of protein in BAL (Fig. 6D; p < 0.05, p < 0.01). These results taken together demonstrated that treatment with CXCR4-MSCs was more effective in reducing the lung pathology score, the wet/dry ratio, and the total protein content in BAL.

Overexpression of CXCR4 Facilitates Homing of MSCs to Damaged Lung Tissue

One of the putative mechanisms for suppression of inflammation by the CXCR4-MSCs is that they simply have greater facility to migrate to the injured tissue. Tissues were, therefore, examined for differences in the number of cells that had reached the damaged lung tissue 2 weeks after transplantation of CXCR4- and GFP-MSCs by immunocytochemistry and quantitative analysis. GFP-positive cells were found in lung tissue in both cases, and the vast majority of transplanted GFP⁺ MSCs were Ki67-positive (Ki67⁺/GFP⁺/DAPI⁺, Fig. 7A), a marker for cell proliferation (36). This result indicated that the two cell types in vivo had a high proliferation potential (87.05 ± 9.39% and 85.89 ± 11.08% of CXCR4- and GFP-MSCs, respectively, Fig. 7C). High expression of CXCR4 was observed only in the GFP⁺ cells in lung tissue where CXCR4-MSCs had been transplanted (CXCR4⁺/GFP⁺/DAPI⁺, Fig. 7B). Moreover, the number of GFP⁺ cells in the CXCR4-MSCs group was significantly higher than in rats injected with GFP-MSCs (10.83 ± 1.85/mm² versus 3.07 ± 0.98 cells/mm², p < 0.01, Fig. 7D). These results indicated that high expression of CXCR4 facilitated the homing of MSCs to damaged lung tissue. The homing of MSCs in vivo may thus be an important mechanism influencing the efficacy of CXCR4-MSCs in this model of ALI.