Adipose tissue-derived stromal cells (ADSC) potentially differentiate into various cell types similar to bone marrow-derived mesenchymal stromal cells (BMSC). Unlike BMSC, ADSC can be harvested easily and repeatedly. However, the advantages of ADSC for cell transplantation in liver disease remain unclear. To investigate this, we developed a novel culture system for ADSC, as well as effective methods for transplantation of ADSC into mice liver. ADSC were isolated from subcutaneous adipose tissues of male C57BL6/J mice and cultured on plastic dishes with or without basic fibroblast growth factor (bFGF). In the in vivo study, ADSC isolated from green fluorescent protein-transgenic mice were transplanted into carbon tetrachloride-injured C57BL6/J mice liver. bFGF-treated ADSC expressed several liver-specific marker genes and demonstrated liver-related functions such as albumin secretion, glycogen synthesis, urea production, and low-density lipoprotein uptake. Importantly, pretreatment of ADSC with bFGF for 1 wk enhanced the repopulation rate of ADSC in mice liver, attenuated liver fibrosis, and restored normal serum alanine aminotransferase and albumin levels. The results indicate that basic FGF facilitates transdifferentiation of ADSC into hepatic lineage cells in vitro and that transplantation of bFGF-pretreated ADSC reduced hepatic fibrosis in mice. ADSC are a potentially valuable source of cells for transplantation therapy.
- hepatic lineage cells
- basic FGF
- cell transplantation
- α-smooth muscle actin
liver transplantation is one of the most effective treatments for end-stage liver disease. However, a shortage of suitable donor organs and the requirement for immunosuppression restrict its application. Effective therapies to replace liver transplantation are clearly needed.
Recent studies indicated that bone marrow-derived mesenchymal stromal cells (BMSC) can transdifferentiate into adipogenic, osteogenic, chondrogenic (24), neurogenic (35), myogenic (10), and hepatogenic (18, 23, 26, 28) cells under prescribed conditions. Mesenchymal stromal cells can be isolated from several organs including fetal tissue (6), umbilical cord blood (5), and adipose tissue (37). Adipose-derived stromal cells (ADSC) are similar to BMSC (8, 34), in that both have limited self-renewal ability and can be induced to various mesenchymal tissues and cells including those of hepatic lineage (4, 13, 37). Furthermore, unlike BMSC, ADSC can be repeatedly harvested by a simple and minimally invasive method and can be easily cultured (29). These characteristics are clear advantages of ADSC, making them potentially superior to BMSC as a cell transplantation source.
Basic fibroblast growth factor (bFGF) is essential for initiating liver development (15). We demonstrated previously that bFGF promotes the transdifferentiation of bone marrow cells into hepatic lineage cells in vitro (26). Ishikawa et al. (14) also reported that bFGF facilitates the differentiation of bone marrow cells into hepatic lineage cells in vivo. Thus we reasoned that ADSC could be differentiated into hepatic lineage cells in the presence of bFGF both in vitro and in vivo.
This study developed a primary culture system for mouse ADSC, seeking to differentiate ADSC into hepatic lineage cells in vitro using bFGF. We also investigated the hepatogenic transdifferentiation ability of ADSC in vivo using the carbon tetrachloride (CCl4)-induced liver-injury mouse model.
MATERIALS AND METHODS
Human hepatocyte growth factor (HGF) was provided by Sumitomo Pharmaceuticals (Osaka, Japan). Human bFGF was purchased from Invitrogen (Carlsbad, CA). Human oncostatin M (OSM) was purchased from Genzyme/Techne (Minneapolis, MN). Dimethyl sulfoxide (DMSO) was purchased from Wako Pure Medical (Tokyo, Japan). Knockout serum replacement was purchased from GIBCO-BRL (Grand Island, NY). Primary antibodies were obtained as follows: anti-mouse albumin (goat polyclonal) from Bethyl (Montgomery, TX), anti-human cytokeratin 18 (CK18) (mouse monoclonal) from Sigma (St. Louis, MO), anti-green fluorescent protein (GFP) (rabbit polyclonal) from Medical & Biological Laboratories (Tokyo), anti-alpha smooth muscle actin (α-SMA) (mouse monoclonal) from Dako (Kyoto, Japan), and anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (rabbit polyclonal) from Trevigen (Gaithersburg, MD). Secondary antibodies (Alexa 488-conjugated donkey anti-goat IgG, Alexa 594-conjugated goat anti-rabbit IgG, and Alexa 594-conjugated goat anti-mouse IgG) were purchased from Molecular Probes (Eugene, OR).
Isolation and culturing of ADSC.
C57BL/6J mice were purchased from Clea Japan (Tokyo), and the C57BL/6-Tg(CAG-EGFP)C15-001-FJ001Osb mice were kindly provided by Dr. Masaru Okabe (Genome Information Research Center, Osaka University, Osaka, Japan). ADSC were collected from the subcutaneous adipose tissue of 12-wk-old male C57BL/6J mice or C57BL/6-Tg(CAG-EGFP)C15-001-FJ001Osb mice as described previously (21). Briefly, subcutaneous adipose tissues were isolated from mice, minced into fine pieces in phosphate-buffered saline (PBS) containing antibiotic-antimycotic solution (Sigma), and incubated in Dulbecco's modified Eagle's medium (DMEM) containing 1 mg/ml collagenase type II and antibiotic-antimycotic solution at 37°C for 30 min. The tissues digests were filtered thorough sterile 250-μm nylon mesh, centrifuged at 600 g for 5 min, and resuspended; this process was repeated twice. ADSC were seeded onto culture dishes with DMEM/10% fetal calf serum (FCS) containing antibiotic-antimycotic solution. In this study, C57BL/6J mice ADSC were used for in vitro studies, and GFP-positive ADSC were used for the in vivo experiments.
ADSC second passages were plated onto plastic dishes in medium (DMEM supplement with 10% Knockout serum replacement, l-glutamine, MEM nonessential amino acids solution, antibiotic-antimycotic solution, and with or without growth factors; 20 ng/ml). The growth factors used were bFGF or the combination of HGF, OSM, and 0.1% DMSO added to the media on the 10th day after differentiation. Culture media were replaced twice a week.
RNA extraction and RT-PCR analysis.
Total RNA was extracted from cultured ADSC by using Sepasol-RNA I (nacalai tesque, Osaka, Japan) according to the instructions provided by the manufacturer. RT-PCR was performed using a Gene Amp RNA PCR kit (Applied Biosystems, Branchburg, NJ). PCR conditions were as follows: hot start for 5 min at 95°C, 40 cycles at 95°C for 15 s, 60°C for 15 s, and 72°C for 1 min. The sequences of the primers used in this study are available on request. Real-time PCR was performed on a LightCycler as described previously (16). Primers used in real-time PCR were obtained from Qiagen (Hilden, Germany).
ADSC were fixed with 4% paraformaldehyde at room temperature for 10 min, followed by incubation with blocking solution comprising PBS and 5% host serum (goat and donkey) for 30 min. The cells were allowed to react with primary antibodies [albumin (1:500), CK18 (1:500)] at room temperature for 1 h and then with secondary antibodies and DAPI (1:5,000) at room temperature for 30 min. Cells were washed with PBS between each step.
Immunohistochemistry and picrosirius red staining.
After deparaffinization, the sections were allowed to react with primary antibodies [GFP (1:500), albumin (1:500), α-SMA (1:50)] at room temperature for 1 h and then with secondary antibodies and DAPI at room temperature for 30 min. Specimens were washed with PBS between each step. Picrosirius red (Sigma) staining was used to detect collagen fibrils. The area of fibrosis stained by picrosirius red was quantified by using Adobe Photoshop (Adobe Systems, San Jose, CA).
Immunoblotting was performed as described previously (19), using antibodies against α-SMA (1:500 dilution) and GAPDH (1:2,000).
Evaluation of serum ALT and albumin levels.
Mice serum alanine aminotransferase (ALT) levels were measured by using a Wako Transaminase CII-Test kit. Serum albumin levels were measured by using an Albumin kit (Wako Pure Medical, Tokyo, Japan), following the protocol supplied by the manufacturer.
Albumin secretion and urea production by ADSC.
ADSC were cultured in dishes as described above, in the absence and presence of bFGF. The culture medium was replaced twice a week. Samples of culture supernatants were collected at 2 and 4 wk to measure albumin concentrations by ELISA (Albuwell M, ExoCell, Philadelphia, PA) and urea concentrations by using a Urea assay kit (DIUR-500) according to the instructions supplied by the manufacturer (BioAssay Systems, Hayward, CA). Bovine albumin present in the Knockout serum replacement did not cross-react with the anti-mouse albumin antibody used in the ELISA.
Uptake of LDL.
Low-density lipoprotein (LDL) uptake was assessed by incubating cells for 4 h at 37°C with 10 μg/ml Dil-Ac-LDL (Biomedical Technologies, Stoughton, MA). The assay was performed using the method provided by the manufacturer.
Periodic acid-Schiff staining.
Cells were fixed with 4% paraformaldehyde for 10 min, and incubated with or without 0.1% α-amylase for 1 h. Then, cells were oxidized in 0.5% periodic acid for 5 min and rinsed twice with water. Cells were then treated with Schiff's reagent for 15 min, rinsed with water, and the nuclei stained with hematoxylin.
Transplantation of GFP-positive ADSC into mice.
In this study, 500 μl/kg body wt of CCl4 was injected intraperitoneally into 8-wk-old recipient male C57BL/6J mice twice a week for 4 wk to induce permanent liver damage. Two days after completion of the CCl4 treatment, 1 × 105 GFP-positive ADSC second passages pretreated with bFGF (bFGF+ group; n = 10) or without bFGF (bFGF− group; n = 10) in DMEM/10% FCS containing antibiotic-antimycotic solution for 1 wk were diluted in 100 μl of PBS and then transplanted slowly into mice spleens using a 26-gauge needle. After ADSC injection, the same dose of CCl4 was continuously injected twice a week to maintain permanent liver damage for 4 more wk. Mice were euthanized 2 days after the final CCl4 injection, and blood and liver samples were obtained. The liver was either fixed with 4% buffered-paraformaldehyde for histological examination or immediately frozen in liquid nitrogen for RNA extraction. Mice treated with CCl4 for 8 wk but without transplanted ADSC were used as the CCl4 control group (n = 10). Mice treated with CCl4 for 4 wk but without ADSC transplantation were used as the 4W group (n = 4).
The Ethics Review Committee for Animal Experimentation of Osaka University Graduate School of Medicine approved the study protocol.
The results are presented as means ± SE. Differences between groups were examined for statistical significance using analysis of variance with Fisher's paired least significant difference test. Statistical significance was defined as P < 0.05.
Expression of growth factor receptor genes in ADSC.
RNA samples were extracted from ADSC cultured in DMEM with 10% FCS. The ADSC showed mRNA expressions of the following: FGF receptor-1 (FGFR1), an FGF-receptor subunit; OSMR, an OSM-receptor subunit; c-Met, an HGF-receptor subunit; and gp130, an OSM-receptor subunit. These results indicated that bFGF, OSM, and HGF can function through these receptors. Mouse liver RNA was used as a positive control (Fig. 1A).
ADSC morphology during transdifferentiation into hepatic lineage cells.
We next analyzed morphological changes in the ADSC during the differentiation protocol. Undifferentiated ADSC exhibited a fibroblast-like morphology (Fig. 1Ba). Culture with 20 ng/ml bFGF produced a gradual change in ADSC morphology from fibroblast-like to hepatocyte-like polygonal cells (Fig. 1B, b–d).
Liver-specific gene expressions in ADSC cultured with various growth factors.
We next assessed the mRNA expression of hepatic-specific markers in ADSC treated with or without growth factors to monitor the hepatic transdifferentiation progress. ADSC cultured with bFGF for 2 wk expressed α-fetoprotein (AFP), albumin, CK18, CK19, and phosphoenolpyruvate carboxykinase (PEPCK) (Fig. 1C). In addition, culture in bFGF for 4 wk induced expression of the transthyretin gene, a late-phase hepatic differentiation marker (Fig. 1D). ADSC cultured without bFGF (control) expressed CK19 gene, a marker of biliary epithelial cells. However, CK19 gene expression was lower in ADSC treated with bFGF for 2 wk and was further reduced in ADSC treated with bFGF for 4 wk (Fig. 1, C and D). Recently, Seo et al. (31) demonstrated transdifferentiation into hepatogenic cells of human ADSC cultured with the combination of HGF, OSM, and DMSO (31). In our study, mouse ADSC cultured with the same combination also showed hepatocyte-marker gene expression after 4-wk culture, but not after 2-wk culture (Fig. 1, C and D).
Quantitative albumin gene expression and albumin synthesis by cultured ADSC.
Real-time PCR revealed a gradual increase in albumin gene expression in ADSC cultured with bFGF (Fig. 2A). Furthermore, ELISA experiments showed that ADSC cultured in the presence of bFGF secreted significantly higher levels of albumin than cells cultured without bFGF (control) (Fig. 2B).
Immunocytochemistry of albumin and CK18 in ADSC treated with bFGF.
Analysis of protein expression by immunocytochemistry revealed the expression of albumin and CK18 proteins in ADSC cultured with or without bFGF for 4 wk (Fig. 3). Approximately 25% of the ADSC cultured with bFGF for 4 wk showed positive staining for both albumin and CK18 (Fig. 3B), whereas only a few cells were stained for albumin or CK18 in the absence of bFGF (Fig. 3A).
Hepatocytic function of transdifferentiated ADSC in vitro.
To assess whether hepatic lineage cells derived from ADSC are functionally competent, we analyzed ADSC cultured with or without 20 ng/ml of bFGF for 4 wk. Periodic acid-Schiff staining to reveal glycogen-storage ability showed positive staining in ∼5% of ADSC cultured with bFGF, whereas no staining was observed in cells without bFGF (Fig. 4A). This positive staining was diminished by amylase pretreatment. These experiments indicated that ADSC cultured with bFGF for 4 wk could store glycogen.
We next examined whether ADSC-derived hepatic lineage cells can uptake LDL by incubating differentiated ADSC with Dil-Ac-LDL. Approximately half the ADSC took up LDL, whereas undifferentiated ADSC did not (Fig. 4B). Finally, ADSC-derived hepatic lineage cells were tested for urea production. Undifferentiated ADSC showed no urea production, whereas cells cultured in the presence of bFGF secreted higher amounts of urea in the culture media in a time-dependent manner (Fig. 4C).
bFGF facilitated differentiation of transplanted ADSC into hepatic lineage cells.
GFP-positive cells were not found in spleens of all mice despite being injected with ADSC (data not shown). A few GFP-positive cells (0.3% of whole cells) were observed in the portal area of the bFGF− group mice livers, whereas none were observed in the CCl4 group mice livers. However, these GFP-positive cells in ADSC-transplanted mice livers showed only weak albumin protein expression (7.7% of GFP-positive cells) (Fig. 5, A and B). In contrast, many GFP-positive cells (2.3% of total cells) were observed in the portal area of the bFGF+ group mice livers. Moreover, these GFP-positive cells strongly expressed albumin protein (54.8% of GFP-positive cells) (Fig. 5C).
Liver function improves after ADSC transplantation.
We compared serum ALT and albumin levels across the three experimental groups. Serum ALT levels were significantly lower in bFGF− mice than in CCl4-induced mice and tended to be even lower in bFGF+ mice (Fig. 6A). Serum albumin levels were significantly elevated in bFGF− mice compared with CCl4 mice, and interestingly, were further elevated in bFGF+ mice compared with both the CCl4 and bFGF− groups (Fig. 6B).
Transplanted bFGF-pretreated ADSC attenuate liver fibrosis.
Liver fibrosis was induced after 8 wk of CCl4 treatment in mice of the CCl4 group. ADSC transplantation significantly enhanced this fibrosis in livers of bFGF− mice compared with the CCl4 group, but liver fibrosis was attenuated in bFGF+ mice livers compared with the other two groups (Fig. 7, A and B). The severity of liver fibrosis in the bFGF+ group was almost similar to that of the 4W group. This indicated that transplantation of bFGF-pretreated ADSC prevented further fibrosis in mice liver. Next, we performed GFP-α-SMA double immunostaining and identified transplanted ADSC-derived myofibroblasts in bFGF− mice livers (Fig. 7C). Immunoblotting revealed enhanced α-SMA protein level in bFGF− mice liver (Fig. 7D). Moreover, the livers of bFGF+ mice showed low expression levels of fibrogenic markers, transforming growth factor (TGF)-β1, collagen Iα1, and tissue inhibitor of matrix metalloproteinase 1 (TIMP1), but high expression level of the fibrolytic gene marker matrix metalloproteinase 13 (MMP13) (Fig. 7, E–H). The expression levels of collagen Iα1 and TIMP1 were significantly elevated in the liver of bFGF− mice compared with those of CCl4 mice.
bFGF pretreatment reduces fibrogenic gene expression and increases fibrolytic gene expression in ADSC.
Next, we analyzed the expression levels of fibrogenic and fibrolytic genes in ADSC pretreated with or without bFGF. ADSC pretreated with bFGF for 1 wk showed significantly low collagen Iα1 and TIMP1 gene expression levels and increased MMP13 gene expression level compared with untreated ADSC. In addition, a significantly low α-SMA gene expression level was noted in bFGF-pretreated ADSC. However, pretreatment of ADSC with bFGF enhanced TGF-β1 expression (Fig. 8).
The present study demonstrated that bFGF facilitates the transdifferentiation of ADSC into hepatic lineage cells in vitro and in vivo. Unlike BMSC, ADSC can be repeatedly harvested by a simple and minimally invasive method and can be easily cultured (29). These characteristics would make repeated ADSC transplantation easier than BMSC transplantation. Therefore, ADSC represent an equal, if not superior, potential source of undifferentiated cells for liver cell transplantation.
Liver development is a multistep process, mediated by various growth factors and cytokines. At the initial stage of liver development, the foregut endoderm commits to becoming future liver via interactions with the cardiogenic mesoderm. It is at this step that bFGF produced by mesodermal cells participates in hepatic differentiation (15), being required to induce the hepatic fate in the foregut endoderm (30). Previously, we demonstrated that bFGF promotes the transdifferentiation of bone marrow cells into hepatic lineage cells with the induction of transcription factors including hepatocyte nuclear factors and the GATA family of proteins (26). Hepatic expression of FGF genes is upregulated in the CCl4-induced liver injury mice early after bone marrow transplantation (22). Moreover, pretreatment of bone marrow cells with FGFs, especially bFGF, significantly accelerates the repopulation of bone marrow cells in mice liver (14). The present study demonstrated that bFGF promotes the transdifferentiation of ADSC into hepatic lineage cells both in vitro and in vivo.
Human ADSC were recently reported to transdifferentiate into hepatogenic cells under specific conditions (31). In support of that work, the present study also found that culture of ADSC in media containing HGF and OSM, with 0.1% DMSO added after induction of differentiation, could induce hepatocyte marker genes after 4 wk in culture. However, our cells showed such induction with bFGF at an earlier stage compared with the findings of the other group.
The volume of human bone marrow harvested under local anesthesia is generally limited to 40 ml. This volume contains on average 2.4 × 104 BMSC (3). By contrast, a typical harvest of adipose tissue under local anesthesia can easily exceed 200 ml and contains 1 × 106 ADSC (2). Thus lipoaspirate will yield ∼40-fold more stem cells than the equivalent possible harvest of bone marrow (33). Liposuction is one of the most popular cosmetic surgical procedures conducted worldwide (36), raising the possibility of simple and repeatable access to subcutaneous adipose tissue. In addition, ADSC are technically simple to isolate compared with BMSC and exhibit superior features in culture, such as colony frequency and proliferative ability (19).
Unlike BMSC, ADSC express CD34, one of the most well-established stem cell markers (36). After plating, CD34 expression decreased gradually in ADSC, and the expression of mesenchymal stem cell-associated marker, CD105, dramatically increased. These findings confirmed the stem cell capability of ADSC, further enhancing their stature as a cell-source alternative to BMSC.
In our study, transplanted ADSC were successfully engrafted into mice liver and did not disappear after additional 4-wk CCl4 treatment. GFP-positive transplanted ADSC were still alive and functional 4 wk after final CCl4 treatment, and the numbers of GFP-positive cells remained almost similar to those at the end of CCl4 treatment (data not shown). CCl4 is metabolized into toxic metabolic intermediates through hepatic cytochrome P-450 (20). In our in vitro study, ADSC still expressed albumin and CK19 after 4-wk bFGF treatment, indicating that ADSC still possessed the features of both the hepatocyte and biliary epithelial cell after 4-wk bFGF treatment. Although we did not assess the properties of engrafted ADSC in mice liver, it seemed that the transplanted ADSC were still immature and did not express sufficient cytochrome P-450 activity at the time of transplantation. In this regard, multipotent stromal cells synthesize a wide variety of growth factors and cytokines, exerting a paracrine effect on local cellular dynamics (7, 11). These bioactive substances might operate in our model, promoting cell survival, and hence improve liver function.
Recent reports claimed that BMSC transplantation attenuates liver fibrosis (12, 27). On the other hand, Russo et al. (25) reported that bone marrow contributed functionally and significantly to liver fibrosis, and they identified bone marrow-derived activated hepatic stellate cells. These results seem contradictory. In the present study, we examined whether ADSC transplantation had a fibrolytic effect on liver fibrosis. ADSC transplantation into our mice significantly enhanced liver fibrosis in bFGF− mice livers compared with the CCl4 group mice. In contrast, the transplantation of ADSC pretreated with bFGF attenuated liver fibrosis. In addition, bFGF− mice livers contained ADSC-derived α-SMA-positive cells and high hepatic protein level of α-SMA. bFGF stimulates MMP13 at the transcriptional level in human chondrocytes (13), suppresses the promoter activity of collagen I gene in osteoblast-like cells (9), and suppresses TIMP1 gene expression in human periodontal ligament cells (32). In this study, bFGF increased MMP13 expression and decreased the expressions of collagen I and TIMP1 in ADSC. Moreover, we found low protein levels of α-SMA, one of the markers of activated hepatic stellate cells, in bFGF+ mice liver compared with bFGF− mice and found that bFGF decreased α-SMA gene expression in cultured ADSC. bFGF treatment downregulates TGF-β1-induced α-SMA protein expression in fibroblasts (1, 17). It is possible that this inhibitory effect of bFGF on TGF-β1 suppresses the activation of hepatic stellate cells. These antifibrotic effects of bFGF on ADSC might therefore reduce the problem of liver fibrosis in cell transplantations.
In conclusion, bFGF induced transdifferentiation of ADSC into hepatic lineage cells both in vitro and in vivo. Moreover, the transplantation of bFGF-pretreated ADSC contributed to the regression of liver function and fibrosis in CCl4-induced mice liver fibrosis. ADSC can be easily and reproducibly harvested from patients by a simple and minimally invasive method. ADSC-based transplantation for patients with end-stage liver disease could therefore form an important part of therapeutic strategies in this clinical arena.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society