Mesenchymal stem cells (MSCs) ameliorate injury and accelerate repair in many organs, including the kidney, although the reparative mechanisms and interaction with macrophages have not been elucidated. This study investigated the reparative potential of human bone marrow-derived MSCs and traced their homing patterns following administration to mice with ischemia-reperfusion (IR) injury using whole body bioluminescence imaging. The effect of MSCs on macrophage phenotype following direct and indirect coculture was assessed using qPCR. Human cytokine production was measured using multiplex arrays. After IR, MSCs homed to injured kidneys where they afforded protection indicated by decreased proximal tubule kidney injury molecule-1 expression, blood urea nitrogen, and serum creatinine levels. SDS-PAGE and immunofluorescence labeling revealed MSCs reduced collagen α1(I) and IV by day 7 post-IR. Gelatin zymography confirmed that MSC treatment significantly increased matrix metalloproteinase-9 activity in IR kidneys, which contributed to a reduction in total collagen. Following direct and indirect coculture, macrophages expressed genes indicative of an anti-inflammatory “M2” phenotype. MSC-derived human GM-CSF, EGF, CXCL1, IL-6, IL-8, MCP-1, PDGF-AA, and CCL5 were identified in culture supernatants. In conclusion, MSCs home to injured kidneys and promote repair, which may be mediated by their ability to promote M2 macrophage polarization.
- ischemia-reperfusion injury
- mesenchymal stem cells
since the initial excitement surrounding the multilineage potential and self-renewal properties of mesenchymal stem (stromal) cells (MSCs), their therapeutic potential to elicit tissue regeneration has been explored experimentally and in a wide range of clinical applications (45). MSCs are capable of modulating inflammation through interacting with a variety of immune cells (53, 68). These immunomodulatory properties, in combination with their tissue-regenerative capabilities, have created great enthusiasm for these cells to be used as a treatment for a wide variety of pathological conditions ranging from autoimmune to chronic inflammatory diseases (for a review, see Refs. 45, 62, and 68). MSCs reside in most postnatal organs and tissues and can be isolated and expanded in culture (13). Unlike embryonic stem (ES) cells and induced pluripotent stem (iPS) cells, MSCs typically do not form tumors following transplantation in rodents and are free of the ethical limitations associated with ES cell research.
Human MSCs have been shown to ameliorate the symptoms of inflammatory diseases in rodent models (4, 9, 24, 27, 41, 70); however, the mechanisms responsible for their protective and regenerative effects are not completely understood. The interaction of MSCs with macrophages may play a vital role in their downstream anti-inflammatory and immunomodulatory effects, yet the specific cell cross talk MSCs have with infiltrating macrophages and damaged kidney cells, along with the cytokines that contribute to their unique immunomodulatory properties, remains poorly defined.
MSCs secrete a broad range of cytokines, including macrophage chemoattractants, as well as a variety of factors with renoprotective and reparative capabilities. These include anti-inflammatory, antiapoptotic, mitogenic, antifibrotic, and proangiogenic agents, which most likely govern repair via paracrine and endocrine pathways (5, 19, 21, 67). In a setting of acute kidney injury (AKI), transplanted MSCs localized within peritubular capillaries, adjacent to the renal tubules, and glomeruli (56). However, the survival of MSCs and timing of administration leading to the interplay between MSCs and macrophages, along with their ability to modify the tissue microenvironment in a setting where aberrant wound healing-induced collagen accumulation leads to fibrosis, have yet to be elucidated.
Macrophages comprise a heterogeneous population that is governed by the inflammatory cues in the surrounding microenvironment (54). Although initially recognized as contributing to the pathogenesis of kidney injury, macrophages may also play a vital role in the remodeling phase of kidney regeneration following acute damage (30, 61, 63). Subsequently, macrophages have been broadly classified into one of two opposing polarization states: classically activated “M1” and alternatively activated “M2” populations (38). M1 macrophages secrete numerous proinflammatory cytokines and are involved in pathogen clearance whereas M2 macrophages secrete anti-inflammatory cytokines that mediate wound healing and tissue remodeling (38).
This study investigated the therapeutic potential of human bone marrow (BM)-derived MSCs in conjunction with their homing patterns following intravenous (iv) administration to mice with ischemia-reperfusion (IR) injury using whole body bioluminescence imaging. In addition, the effect of MSCs on macrophage phenotype and the soluble factors produced following direct and indirect coculture experiments were assessed.
MATERIALS AND METHODS
Mesenchymal stem cell culture.
Human BM-derived MSCs purchased from the Tulane Center for Stem Cell Research and Regenerative Medicine (Tulane University, New Orleans, LA) and enhanced green fluorescent protein (eGFP) and firefly luciferase (fluc) eGFP+fluc+ MSCs were cultured as previously described (47). Karyotype analysis was performed on MSCs at passage 3 (Southern Cross Pathology, Clayton, Australia). The clonogenic potential of MSCs was tested using a colony-forming unit-fibroblast (CFU-F) assay, and colonies were stained with 3% (wt/vol) crystal violet (Sigma-Aldrich, St. Louis, MO).
To demonstrate multilineage differentiation potential, MSCs were differentiated toward osteogenic, adipogenic, and chondrogenic lineages using a human functional identification kit (R&D Systems, Minneapolis, MN). Following differentiation, osteocytes were stained with Alizarin red S (Sigma-Aldrich), adipocytes with fatty acid binding protein-4 (FABP-4; R&D Systems), and chondrocytes with aggrecan (R&D Systems).
Immunophenotypic analysis of MSCs by flow cytometry was performed using the following fluorochrome-conjugated anti-human antibodies: CD73-PE, CD90-PerCP-Cy5.5, CD105-Alexa Fluor 647, CD14-APC (eBioscience), CD19-FITC, CD34-APC, CD45-APC, and HLA-DR-FITC. All antibodies were purchased from BD Biosciences (San Jose, CA) unless otherwise indicated. Cell population data was acquired using a FACSCanto II flow cytometer (BD Biosciences) and analyzed using Flowlogic Software (Inivai Technologies, Mentone, Australia).
All animal studies were approved by the Monash University Animal Ethics Committee, which adheres to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. For IR injury, male 6- to 8-wk-old C57BL/6J mice (Monash Animal Services, Clayton, Australia) were anesthetized with 2.5% (vol/vol) inhaled isoflurane (Abbott Australasia Pty, Kurnell, Australia), and injury was induced by clamping the left renal pedicle for 40 min (unilateral) or both renal pedicles for 25 min (bilateral) with a microvascular clamp (0.4–0.1 mm; S&T Fine Science Tools, Foster City, CA) through a flank incision. Following reperfusion, mice were injected iv with 1 × 106 MSCs resuspended in 120 μl PBS or a vehicle control (120 μl PBS alone). A third group of mice served as a sham-operated control group, whereby the animals were anesthetized and a flank incision made without clamping the renal pedicle. Mice that received bilateral IR injury were placed in metabolic cages to obtain 24-h urine samples. Urinary kidney injury molecule (Kim)-1 was measured with a Kim-1 mouse ELISA (Abcam, Cambridge, UK). Concentrations of blood urea nitrogen (BUN) and serum creatinine were measured 3 days post-IR using the i-STAT CHEM8+ cartridges and the i-STAT system (n = 8; Abbott, Ontario, Canada).
Mice (n = 5) were anesthetized with 2.5% (vol/vol) isoflurane, injected intraperitoneally (ip) with 200 μl d-luciferin (15 mg/ml in PBS; VivoGlo Luciferin, Promega, San Luis Obispo, CA) and imaged 10 min after injection using the Xenogen IVIS 200 system (Xenogen, Alameda, CA) on days 0 (1-h post-MSC injection), 1, and 3 post-IR. Regions of interest (ROI) were drawn, and fluc luminescent signal intensities were analyzed using Living Image 3.2 software (Xenogen).
Histology and immunofluorescence labeling.
Histopathology was assessed on formalin-fixed, 4-μm-thick paraffin sections stained with hematoxylin and eosin (H&E). Semiquantification of histopathology was performed after taking five fields of view/kidney section within the corticomedullary region (n = 3; 3 sections/mouse; ×400). Proximal tubular damage and protein cast formation were assessed, and the percentage of kidney damage was graded on a scale of 0 to 4 (refer to Table 1).
To assess proliferation, kidney sections were stained with mouse anti-PCNA (DakoCytomation, Glostrup, Denmark) and rabbit anti-mouse Ki67 (Abcam) primary antibodies followed by Alexa Fluor 488 donkey anti-mouse (Molecular Probes, Eugene, OR) and Alexa Fluor 555 goat anti-rabbit (Molecular Probes) secondary antibodies. For proximal tubule Kim-1 expression, immunohistochemical staining was performed with rat anti-mouse Kim-1 (R&D Systems) using the avidin-biotin complex (ABC) method as described previously (43). The area of 3,3′-diaminobenzidine staining per unit area of tissue was measured using a custom macro from the image-analysis software ImageJ/FIJI, version 1.48d. Areas of positive staining were quantified in five nonoverlapping, randomly selected fields of view (n = 3, 3 sections/mouse; ×400 magnification).
For the visualization of type IV collagen, kidney sections were stained with a goat anti-human collagen type IV primary antibody (Southern Biotech, Birmingham, AL) followed by an Alexa Fluor 647 chicken anti-goat antibody (Molecular Probes) and for macrophage staining, a rat anti-mouse F4/80 antibody (AbD Serotec, Oxford, UK) followed by an Alexa Fluor 555 goat anti-rat antibody (Molecular Probes). Sections were counterstained with 4,6-diamidino-2-phenylindole (Molecular Probes) and viewed with a Provis AX70 fluorescence microscope (Olympus, Tokyo, Japan). Fluorescence images were captured with an F-view II digital camera (Soft Imaging System, Munster, Germany).
Hydroxyproline, SDS-PAGE, and zymographic analyses.
A kidney from each animal was divided into portions containing both cortex and medulla for use in each assay. The total collagen content (% collagen content/dry weight tissue) in the kidney (n = 3/group) was measured using a hydroxyproline assay as previously described (49). In brief, kidneys were lyophilized to measure dry weight, hydrolyzed in 6 M hydrochloric acid, and hydroxyproline levels were determined by measuring the absorbance of hydrolyzed samples at 558 nm, using a Digital Spectrophotometer (Varian, Palo Alto, CA) (50). Total collagen content was determined by multiplying the hydroxyproline measurements by a factor of 6.94.
SDS-PAGE analysis was used to detect changes in interstitial collagen subtypes within the kidney (50). The supernatants from pepsin-digested kidneys were analyzed on 5% (wt/vol) acrylamide gels with 3.5% (wt/vol) acrylamide stacking gels. The α1(III) chains were separated from the α1(I) collagen chains with interrupted electrophoresis with delayed reduction of type III collagen. The gels were stained with 0.1% Coomassie blue R-250 overnight at 4°C and then destained with 30% (vol/vol) methanol containing 7% (vol/vol) acetic acid. Densitometry was performed with a calibrated imaging densitometer (Gel Scan-710, Bio-Rad, Hercules, CA), and data were analyzed using Quantity-One software (Bio-Rad).
Matrix metalloproteinase (MMP)-2 and MMP-9 activity was assessed by gelatin zymography (65). Zymographs consisted of 7.5% (wt/vol) acrylamide gels containing 1 mg/ml gelatin. The gels were stained with 0.1% (wt/vol) Coomassie blue R-250 overnight at 37°C and then destained with 7% (vol/vol) acetic acid. Clear bands indicated gelatinolytic activity, where the enzymes had digested the substrate. Densitometry of these MMP bands was performed, and data were analyzed using Quantity-One software.
MSC and macrophage coculture.
BM was isolated from male 6- to 8-wk-old C57BL/6J mice and cultured in DMEM/F12 (Invitrogen, Camarillo, CA) supplemented with 10% FBS, 10 mM l-glutamine, 100 μg/ml penicillin/streptomycin, and 100 U/ml mouse recombinant colony-stimulating factor (CSF)-1 (Chiron) to generate macrophages. On day 7, the purity of the BM-derived macrophages was >95% when checked by flow cytometry.
For coculture experiments, macrophages were primed with 120 ng/ml of IFN-γ (R&D Systems) and 10 ng/ml LPS (Sigma-Aldrich) to induce an M1 phenotype or with 20 ng/ml IL-4 (Invitrogen) to induce an M2 phenotype. The macrophages were then washed with PBS before MSCs were plated indirectly, on a 0.4-μm pore size Transwell (Corning Life Sciences, Pittson, PA), or directly and cultured for 48 h. Following 24 h of coculture, 1 ml of the coculture supernatant was collected and screened for human MSC-derived cytokines, using a MILLIPLEXMAP Human Cytokine/Chemokine Panel (Millipore).
Real-time quantification PCR gene expression analysis.
Macrophages were sorted by fluorescence-activated cell sorting (FACS) from the cocultures using the conjugated anti-mouse antibodies CD45-FITC (BD Biosciences) and F4/80-APC (BD Biosciences). RNA was extracted using an RNeasy Micro Kit (Qiagen, Hilden, Germany) according to the manufacturer's guidelines. RNA samples were reverse transcribed using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA), and real-time quantitative PCR (qPCR) for each target gene was performed in duplicate on cDNA samples using TaqMan Universal PCR Master Mix (Applied Biosystems) and TaqMan Gene Expression Assays (Applied Biosystems; see Table 2). The threshold cycle (Ct) values were normalized against endogenous control β-actin to determine ΔCt.
Statistical analyses of the data were performed using GraphPad Prism software version 5.0 (GraphPad Software, San Diego, CA). An unpaired t-test was used to analyze data between two groups. Comparisons among three groups were performed by one-way ANOVA followed by Tukey's multiple comparison tests. All data were expressed as means ± SE. P < 0.05 was considered statistically significant.
Characterization of MSCs.
Human MSCs were initially characterized to confirm their cellular identity using the minimal criteria established by Dominici et al. (14). In vitro, cultured MSCs adhered to plastic, had a spindle-shaped morphology (Fig. 1A), displayed a normal karyotype (Fig. 1B), and formed CFU-F (Fig. 1C). Functionally, MSCs differentiated into osteocytes, adipocytes, and chondrocytes as evidenced by positive staining with Alizarin red (Fig. 1D), FABP-4 (Fig. 1E), and aggrecan (Fig. 1F), respectively. Finally, MSCs were uniformly positive for the canonical MSC markers CD73, CD90, and CD105 and lacked the expression of the hematopoietic markers CD14, CD19, CD34, CD45, and HLA-DR (Fig. 1G).
MSCs home to the injured kidney following unilateral and bilateral IR injury.
eGFP+fluc+ MSCs were FACS sorted to enrich for the number of eGFP+fluc+ MSCs (Fig. 2A), with the purity of the postsorted cells also determined by flow cytometry (Fig. 2A). eGFP expression was confirmed visually using fluorescence microscopy (Fig. 2B). Using a noninvasive bioluminescent imaging technique, eGFP+fluc+ MSCs were tracked in vivo following iv administration immediately following surgery in mice with unilateral or bilateral IR injury and in sham-operated control mice (see diagram in Fig. 2C).
Following sham surgery, MSCs accumulated only in the lungs, likely the result of being trapped in the pulmonary capillaries (Fig. 3A). Bioluminescence measurements in the sham-operated control mice decreased over the 3-day time course (2.038 × 107 photons·s−1·cm−2·sr−1 on day 0 to 3.362 × 106 photons·s−1·cm−2·sr−1 on day 1 as per mouse in Fig. 3A). No signal was detected at day 3. In contrast, following unilateral and bilateral IR injury, MSCs homed to the site of damage via two routes: directly to the kidney(s), as detected at the day 0 imaging time point (Fig. 3, B and D, respectively), or to the kidney(s) via the lungs (Fig. 3, C and E, respectively). The localization of the MSCs in the kidney was confirmed by imaging the lateral aspect of the mouse (images not shown) before the kidneys were excised and imaged ex vivo. Examples of each of the MSC homing patterns with detected fluc signals in sham and IR mice are shown in Fig. 3. The fluc signal following direct homing to the kidney with unilateral IR injury was marginally decreased from day 0 to day 1 (4.436 × 107 to 3.828 × 107 photons·s−1·cm−2·sr−1) and further by day 3 (1.953 × 107 photons·s−1·cm−2·sr−1; Fig. 3B). In contrast to the unilateral model, the fluc signal with bilateral IR injury gradually increased from 5.109 × 107 photons·s−1·cm−2·sr−1 on day 1 to 1.706 × 108 photons·s−1·cm−2·sr−1 on day 3. At 7 days post-IR, the fluc signal was no longer detected in either the unilateral or bilateral models. In the mice where MSCs were observed to accumulate in the lungs before migrating to the damaged kidney(s) following unilateral or bilateral IR injury (Fig. 3, C and E), the majority of injected cells had localized in the lungs at 1 h postadministration. However, the MSCs further migrated from the lungs to the injured kidney(s) (imaged on days 1 and 3), with the majority of cells being present in the kidney(s) at day 3. Again, at day 7, no cells were detected.
MSCs promote structural and functional regeneration.
Compared to sham-operated mice, at 7 days following IR injury there was widespread tubular epithelial cell damage within the kidney, evidenced by numerous protein casts, interstitial matrix expansion, and extracellular matrix deposition along with a marked infiltration of inflammatory cells (Fig. 4A). In contrast, the administration of MSCs to mice with IR injury promoted structural regeneration, including reduced inflammation and reestablishment of the tubular epithelium. Semiquantitative examination of kidney sections revealed a significant reduction in the number of protein casts (P < 0.001) and proximal tubule epithelial cell damage (P < 0.001; Fig. 4B) by 5 days following MSC injection. Structural regeneration of the MSC-treated kidneys was associated with a significant increase in tubular epithelial cell proliferation demonstrated at the day 3 time point, as assessed with Ki67 and PCNA immunostaining (Fig. 4, C and D). This MSC-mediated repair was further evidenced by functional recovery. BUN and serum creatinine concentrations were measured 3 days post-MSC administration (Fig. 5, A and B). At 3 days after bilateral IR surgery, BUN levels had increased over twofold compared with sham-operated controls (18.1 ± 1.9 vs. 8.4 ± 0.4 mmol/l; P < 0.001) and serum creatinine 1.5-fold higher than sham levels (34.7 ± 3.1 vs. 22.4 ± 2.1 μmol/l; P < 0.05). In MSC-treated mice, both the BUN and serum creatinine concentrations were comparable to baseline measurements and were significantly lower than the vehicle-treated controls (Fig. 5, A and B). In addition, immunohistochemical staining revealed increased expression of Kim-1, a marker of proximal tubular injury, on the apical membrane of proximal tubule cells 3 days after IR injury, compared with sham-operated kidneys (Fig. 5C), while Kim-1 expression was markedly reduced in MSC-treated mice. Notably, urinary Kim-1, assessed by an ELISA, was significantly increased at 7 days post-IR compared with sham-operated control mice (P < 0.001) but returned to baseline levels in MSC-treated mice (P < 0.01; Fig. 5D).
MSCs reduce collagen accumulation in the injured kidney.
MSC therapy following IR injury reduced interstitial collagen accumulation as assessed by hydroxyproline assay, SDS-PAGE, and type IV collagen immunofluorescence labeling. IR injury resulted in a gradual but significant increase in the total collagen concentration at 3 (P < 0.05), 5 (P < 0.001), and 7 (P < 0.001) days postinjury compared with sham-operated controls (Fig. 6A). At 5 days post-IR injury, MSC treatment significantly decreased the total renal collagen concentration (P < 0.05) compared with vehicle-treated mice. SDS-PAGE revealed the predominant interstitial collagen subtypes within the kidney were type 1 collagen [α1(I) and α2(I) monomers and dimers of two α1(I) chains (β11) or α1(I) and α2(I) monomers (β12)], and a small amount of type V collagen (Fig. 6B). Scanning densitometry further revealed a decrease in the accumulation of the collagen subtype α1(I) in MSC-treated kidneys compared with the vehicle-treated controls at 5 and 7 days postinjury (Fig. 6B), which reached significance (P < 0.05) at day 7. Immunofluorescence microscopy was utilized to visualize type IV collagen and macrophage (F4/80) localization within the kidney (Fig. 6D). At day 7, an accumulation of interstitial collagen was evident in vehicle-treated kidneys. In comparison, type IV collagen appeared as a delicate framework surrounding the glomeruli and reepithelialized tubules of MSC-treated kidneys, with a pattern of expression comparable to kidneys from sham-operated control mice.
Gelatin zymography revealed that IR injury resulted in a significant increase in latent and active MMP-2 levels compared with sham-operated control kidneys at both 5 (P < 0.001) and 7 (P < 0.001) days postinjury (Fig. 6C). In comparison, the latent and active forms of MMP-2 in the MSC-treated kidneys remained significantly lower than in the vehicle-treated kidneys at both days 5 and 7. Active MMP-9 was also significantly increased in vehicle-treated kidneys at 3 (P < 0.001), 5 (P < 0.01), and 7 (P < 0.05) days postinjury compared with the sham-operated kidneys (Fig. 6C). Notably, MSC treatment resulted in a significant increase in active MMP-9 at 3 days postinjury (P < 0.05) compared with its vehicle-treated counterpart.
MSCs alter macrophage phenotype following in vitro coculture.
Direct and indirect coculture of MSCs with macrophages resulted in an MSC-dependent polarization of macrophages toward an M2 phenotype. BM-derived murine macrophages that had been stimulated to display an M1 or M2 phenotype in vitro were cocultured with MSCs for 48 h either directly or indirectly using a Transwell coculture system (see diagram in Fig. 7A). qPCR analysis of macrophage gene expression showed that the direct coculture of M1 macrophages with MSCs caused an upregulation of the M2-associated gene, Arg1 (Fig. 7B). Another M2-associated gene, Ccl2, was also upregulated following the indirect coculture of M1 macrophages with MSCs (Fig. 7B). Furthermore, an enhanced expression of the M2-associated genes, Arg1, Chi3l3, Ccl2, and Fizz1 (also known as Retnla), was observed following both the direct and indirect coculture of M2 macrophages with MSCs (Fig. 7B).
The MSC-macrophage coculture medium was then screened using a panel of human cytokines and chemokines (Table 3). The human soluble factors EGF, granulocyte macrophage colony-stimulating factor (GM-CSF), CXCL1, IL-6, IL-8, monocyte chemotactic protein (MCP)-1, PDGF-AA, and CCL5 were detected in the coculture supernatants, suggesting these factors may play a role in the MSC-mediated shift in macrophage polarization.
The therapeutic efficacy of MSCs derived from various sources including BM (41), adipose (11), umbilical cord (7), embryos (69), and Wharton's jelly (15) to treat cisplatin- (5, 17, 41)-, glycerol (18, 42)-, unilateral ureteral obstruction (UUO) (2, 34, 35, 46)-, and IR (28, 51, 56, 57)-induced experimental models of AKI have been investigated (for a review, see Ref. 64). However, the mechanisms by which MSCs elicit repair remain largely unknown. Following injury, MSCs have the capacity to migrate along an inflammatory cytokine gradient, governed largely by chemokines and their receptors, to the site of damage (18, 23, 33, 58). The present study demonstrated that MSCs administered to sham-operated mice migrated directly to the lungs, where they remained and were cleared within 3 days. In comparison, MSCs administered to mice following IR had the potential to home directly to the injured kidney(s), where they remained for up to 3 days postadministration and exerted beneficial effects over the longer term. These findings are consistent with previously published work (59). While some MSCs still traveled to the lung, these cells retained the ability to migrate to the injured kidney(s) within the first 3 days following IR injury. The localization of the MSCs in the lungs of mice has been confirmed in previous studies in other experimental models (29, 44). Cell size is believed to contribute to the initial entrapment of the MSCs within the pulmonary capillaries, due to the small diameter of the vessels. In addition, adhesion molecules expressed by MSCs and the corresponding receptors expressed on the lung endothelia may also contribute to the MSC lung entrapment and dislodgment (44). Although tissue-specific homing has been demonstrated in a number of different conditions, long-term engraftment of the MSCs has rarely been shown. Consequently, several studies have investigated strategies aimed at enhancing the MSC migratory properties, survival, and consequently regenerative capacity through preconditioning with various growth factors such as IGF-1 (66), glial cell-derived neurotrophic factor (GDNF) (52), melatonin (40), exposure to hypoxia (20), or genetic modification (10, 12, 16, 69, 71).
The current study utilized the xenogeneic transplantation of MSCs into immunocompetent mice without the use of immunosuppressant agents. Although there are extensive data demonstrating the immunomodulatory properties of MSCs in vitro, it is unclear why these cells remain tolerated by the host's immune system following xenogeneic transplantation (3). In the current study, the possibility that the host's immune system cleared the transplanted MSCs by the day 7 time point cannot be discounted. Nevertheless, numerous studies have demonstrated extraordinary regenerative efficacy following successful transplantation of human MSCs into mice in several disease models (31). However, the type of MSC transplantation (allogeneic vs. autologous), tissue of origin (BM, adipose, umbilical cord), isolation method (enzymatic vs. mechanical), delivery route (systemic vs. local), dose, and timing of administration are also key factors that may influence the renoprotective effect of MSC therapy and need to be carefully considered before clinical application. For example, in an experimental model of glomerulonephritis, the administration of MSCs improved renal function but resulted in long-term maldifferentiation into glomerular adipocytes (26). These findings raise considerable concerns surrounding the safety of MSC-based therapies, and so it is imperative that studies looking into their long-term safety and unwanted differentiation are performed.
In this study, we demonstrate that following migration to the kidney in response to IR injury, MSCs promoted tubular epithelial cell proliferation, resulting in structural repair and tissue remodeling concurrent with a reduction in collagen. MMPs are enzymes that are involved in extracellular matrix remodeling via collagen degradation (8). The identification of an MSC-induced increase in MMP-9 at day 3 and decrease in MMP-2 at days 5 and 7 provides insight into the temporal pattern of MSC-mediated tissue remodeling. In addition to structural improvement, proximal tubular Kim-1 expression and urinary Kim-1 levels were assessed, as a direct measure of kidney injury. Both demonstrated significant improvements in the severity of injury at 3 and 7 days post-MSC treatment. Kim-1 is a sensitive AKI biomarker useful for detecting early disease onset and can provide useful insight into the state of injury before the production of classic indicators of nephrotoxicity, such as serum creatinine levels (22, 60).
MSCs have unique immunomodulatory properties and their trophic effects on T, B, natural killer, and dendritic cells have been thoroughly investigated (53, 68). However, the effect of MSCs on macrophage polarization and the consequences of this cell-cell interaction in altering the proinflammatory course of injury remains largely unknown. Our findings are consistent with other studies that have demonstrated the ability of MSCs to polarize macrophages toward an M2 phenotype in vitro (1, 25, 32, 36, 37). However, the influence MSCs have on the phenotypic and functional characteristics of macrophages is often variable. For example, MSCs have been shown to both upregulate and inhibit the expression of IL-6. Similarly, macrophage phagocytic activity has been both enhanced and suppressed by MSCs (1, 25, 32, 37).
Li et al. (32) demonstrated that MSC repair requires the infiltration of macrophages after the induction of IR injury. Given this important observation, we show herein that MSCs significantly enhanced the expression of M2-associated macrophage genes in both M1 and M2 macrophage subsets in vitro. Furthermore, MSC-induced M2 polarization was evident in both direct and indirect coculture systems, indicating that the alteration of macrophage phenotype was mediated through paracrine mechanisms. Screening of the coculture supernatants detected the presence of MSC-derived EGF, GM-CSF, CXCL1, IL-6, IL-8, MCP-1/CCL2, PDGF-AA, and RANTES/CCL5, all of which, except for EGF, GM-CSF, CXCL1, and PDGF-AA, have previously been shown to promote M2 polarization (1, 6, 39, 48, 55). Interestingly, CXCL1 was only detected in the direct coculture system, indicating that its production required direct cell-to-cell contact. Conversely, RANTES/CCL5 was only detected in the Transwell coculture system, signifying that the direct cell-to-cell contact inhibited the release of this chemokine. Although the enhancement of an M2 phenotype was facilitated through paracrine mechanisms, with the exception of EGF and RANTES/CCL5, direct coculture did result in increased levels of MSC-secreted soluble factors.
In summary, whole body bioluminescence imaging to trace MSCs delivered to mice with unilateral or bilateral IR injury demonstrated a unique pattern of infiltration where MSCs either homed directly to the injured kidney(s) or mobilized from the lungs to the injured kidney(s). MSC therapy was renoprotective and promoted kidney repair, as indicated by decreased proximal tubule Kim-1 expression and urinary Kim-1 levels. In addition, MSC therapy stimulated somatic tubular epithelial cell proliferation and significantly reduced aberrant collagen accumulation, resulting in improved kidney function. This highlights the therapeutic potential of MSCs in ameliorating the progression of kidney disease, of which established fibrosis is a common characteristic. MSCs are thought to elicit repair through paracrine and/or endocrine mechanisms that modulate the immune response, leading to tissue repair and cellular replacement. Our results provide important insights into the production of various cytokines, chemokines, and enzymes resulting from macrophage-MSC interactions and how these govern the inflammatory and remodeling phases of AKI. However, determining the optimal delivery methods for engraftment, testing long-term safety, and understanding their ability to modify the tissue microenvironment in a setting of progressive fibrosis require further consideration.
This study was funded by a Project Grant (1003806) from the National Health and Medical Research Council (NHMRC) of Australia. C. S. Samuel is supported by Monash University Mid-Career and NHMRC Senior Research Fellowships.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: A.F.W., C.S.S., and S.D.R. provided conception and design of research; A.F.W., T.M.W., M.B.G.K., N.L.P., and C.S. performed experiments; A.F.W., T.M.W., M.B.G.K., and C.S.S. analyzed data; A.F.W., T.M.W., M.B.G.K., and S.D.R. interpreted results of experiments; A.F.W. prepared figures; A.F.W. drafted manuscript; A.F.W., T.M.W., C.S.S., and S.D.R. edited and revised manuscript; A.F.W., T.M.W., M.B.G.K., N.L.P., C.S., C.S.S., and S.D.R. approved final version of manuscript.
We thank Professor Claude Bernard for intellectual input and Keith Schulze from Monash Micro Imaging, Monash University, for technical support.
- Copyright © 2014 the American Physiological Society