|Cancer||Cell death||Cell cycle||Cytoskeleton||Exo/endocytosis||Differentiation||Division||Organelles||Signalling||Stem cells||Trafficking|
Cell Biology International (2008) 32, 959965 (Printed in Great Britain)
Optimization of an effective directed differentiation medium for differentiating mouse bone marrow mesenchymal stem cells into hepatocytes in vitro
Xiao‑Lei Shia, Liang Maoa, Bi‑Yun Xub, Ting Xiea, Zhang‑Hua Zhua, Jun‑Hao Chenb, Lei Lib and Yi‑Tao Dinga*
aDepartment of Hepatobiliary Surgery, Affiliated Drum Tower Hospital, Medical College of Nanjing University, No. 321 Zhongshan Road, Nanjing 210008, PR China
bScientific Research Department, Affiliated Drum Tower Hospital, Medical College of Nanjing University, PR China
We have used a uniform design to explore the most effective directed differentiation medium (MEDDM) for differentiating mouse bone marrow mesenchymal stem cells (mMSCs) into hepatocytes. Our methods involved arranging eight differentiation medium groups following uniform design. Flow cytometry was used to evaluate the percentage of ALB+ and CK18+ cells in each group. Factors and their concentrations in the MEDDMs were then identified. The MEDDMs were evaluated by their ability to differentiate mMSCs into hepatocytes by RNA and protein expressions and synthesis functions. FGF at 35
Keywords: Bone marrow mesenchymal stem cells, Hepatocyte, Differentiation, Uniform design.
*Corresponding author. Tel.: +86 25 83304616x66866; fax: +86 25 83317016.
Stem cells are cells with clonogenic and self-renewing capabilities which can differentiate into multiple cell types (Weissman, 2000). One particular stem cell in bone marrow has been shown to differentiate into hepatocytes (Theise et al., 2000a,b; Alison et al., 2000; Lagasse et al., 2000; Jian et al., 2002). The differentiation process of stem cells occurs in a specific microenvironment. Although complex, the process is initiated by adhesion-related events through signaling in the extracellular matrix, which contains soluble ligands controlling cell growth, differentiation and morphogenesis (Hemler and Rutishauser, 2000; Birchmeier and Rosenthal, 2000). Several cytokines, including fibroblast growth factor-4 (FGF-4), oncostatin M (OSM), hepatocyte growth factor (HGF) and epithermal growth factor (EGF), have been identified as controlling hepatic differentiation and maturation (Wells and Melton, 1999; Taga and Kishimoto, 1997; Somerset et al., 1997; Kamiya et al., 2001). Previously, we have successfully obtained hepatocyte-like cells through directed differentiation of mouse bone marrow mononuclear cells by using FGF-4, OSM, HGF and EGF as differentiation factors (Shi et al., 2005). However, the cytokines that affect the process of stem cell differentiation into hepatocytes in vitro have not been identified. In this paper, we identified and optimized the composition and concentration of the most effective differentiation medium. Using this medium, we demonstrate the differentiation of mouse bone marrow mesenchymal stem cell into hepatocytes using mRNA analysis, protein expression and cell functionality examination.
2 Materials and methods
2.1 Isolation and characterization of mouse bone marrow mesenchymal stem cells (mMSCs)
Bone marrow cells were prepared as previously described (Levite et al., 1991). Briefly, 5–7
Bone marrow cells were inoculated at 5
Flow cytometry was used for analysis of cell surface molecules. Harvested P4 cells were washed with PBS containing 5% FBS. A total of 2
2.2 Selection of the most effective directed differentiation medium (MEDDM)
The uniform design method was adopted for the experiments, which involved four factors at four different levels. The four factors were hepatocyte growth factor (HGF, denoted by x
Medium containing different compositions of the four differentiation factors (FGF-4, OSM, HGF and EGF) was designed using the uniform design system U
The uniform design experiment medium composition
2.3 Flow cytometry analysis of ALB and CK18 positive cells
Cells at day 21 were detached from the flasks and stained for flow cytometry analysis. After immunofluorescent staining, cells were then fixed with 2% paraformaldehyde and left in solution until being analyzed with an FACSCalibur (Becton Dickinson). Primary antibodies included anti-mouse albumin (DakoCytomation, CA, USA) and anti-mouse CK18 (Pharmingen, CA, USA). Fluorescein (FITC) and phycoerythrin (PE)-conjugated antibodies (Sigma–Aldrich) were used as immunofluorescent secondary antibodies.
2.4 Directed differentiation of mMSCs with the MEDDM
The mMSCs were cultured in the MEDDM or the basic medium (as negative control) at 5
Using trizol reagent (Sigma–Aldrich), the total RNA was extracted from fresh mMSCs, cultured mMSCs and directed differentiated mMSCs in the MEDDM, respectively. C57BL/6 mouse hepatocytes were used for positive control. cDNA was synthesized and amplified using a one-step RT-PCR kit (Clontech Laboratories, CA, USA). For albumin (ALB), cytokeratin18 (CK18), alpha fetal protein (AFP) and transthyretin (TTR), the following reaction conditions were used: reverse transcription at 50
Primers used for amplification were as follows: AFP (184 Albumin (148 CK18 (127 TTR (105 β-Actin (587
All primers were synthesized by Shanghai Bioengineer Company, China. mRNA levels were normalized to β-actin. The amplified products were subjected to electrophoresis in 1% agarose gels and stained with ethidium bromide.
Regular culture mMSCs and differentiated mMSCs with the MEDDM were coated on glass slides and fixed with 4% paraformaldehyde for 10
Mouse hepatocytes were used as positive controls and negative controls were designed by using PBS instead of primary antibody. Slides were observed with a Zeiss Axiovert fluorescence microscope (Carl Zeiss, Jena, Germany).
2.7 Albumin ELISA
Cell culture medium at various time points was collected as samples for albumin ELISA following standard protocol. Mouse hepatocytes were used as positive controls. Briefly, the ELISA plate wells were coated with the mouse albumin antibody (1:100, Dako) and incubated for 60
2.8 Urea assay
Urea concentrations were measured with a colorimetric assay kit (Randox Life Sciences, Co. Antrim, UK). The mMSCs were plated at 5
3.1 Phenotypic analysis of mMSCs
The expression of different cell surface molecules, including CD29, CD44, CD45 and CD90, was determined by flow cytometry. For P4 cells, over 90% cells were CD45 negative, while CD29, CD44, CD90 were positive.
3.2 Optimization of the effective directed differentiation medium
Table 2 summarized the cell differentiation results of these medium as indicated by the percentage of cells expressing the liver-specific proteins, i.e., ALB (denoted by y
Flow cytometry analysis of cells cultured in different media (mean ± SD)
For designing the composition of the medium, the following second-order polynomial model equation was used for the selection of variables: (1)
The following equations were obtained from the studies: (2) (3)
To apply differentiation factors to the regression models (4)
According to the above statistical analysis, 35
3.3 FGF-4 and OSM are effective enough to differentiate mMSCs into hepatocytes 3.3.1 Gene expressions of liver specific markers To assess the directed differentiation of mMSCs into hepatic lineages, mRNA transcriptions of endodermal and liver specific genes, including ALB, CK18, TTR and AFP, were examined. Results suggested that all genes were transcribed in mMSCs cultured in the MEDDM, while they were non-detectable in either fresh mMSCs or regular culture mMSCs (Fig. 1).
Fig. 1 The mMSCs cultured in MEDDM expressed liver-specific genes on the mRNA level. (Lane 1: fresh mMSCs; lane 2: mMSCs cultured in MEDDM; lane 3: mMSCs cultured in regular medium; lane 4: mouse hepatocytes as positive control). RT-PCR was conducted as described in the Section
3.3.1 Gene expressions of liver specific markers
To assess the directed differentiation of mMSCs into hepatic lineages, mRNA transcriptions of endodermal and liver specific genes, including ALB, CK18, TTR and AFP, were examined. Results suggested that all genes were transcribed in mMSCs cultured in the MEDDM, while they were non-detectable in either fresh mMSCs or regular culture mMSCs (Fig. 1).
The mMSCs cultured in MEDDM expressed liver-specific genes on the mRNA level. (Lane 1: fresh mMSCs; lane 2: mMSCs cultured in MEDDM; lane 3: mMSCs cultured in regular medium; lane 4: mouse hepatocytes as positive control). RT-PCR was conducted as described in the Section
3.3.2 Protein expressions of liver specific markers
To further confirm the directed differentiation of mMSCs into hepatocytes, immunofluorescence was conducted to determine the expression of ALB and CK18 proteins. The results (Fig. 2) clearly demonstrate that ALB resides in the cytoplasm and cell membrane, while CK18 is scattered in the cytoplasm. ALB and CK18 expressions were also found to correlate with the degree of mMSCs differentiation. For example, on day 21, the ratio of ALB-positive cells reached 82.83
Immunofluorescent staining of liver marker proteins (ALB-FITC, green; CK18-TRITC, orange under the same filter) in mMSCs cultured in the MEDDM (A) or regular medium (B) after 21 days. (1000×).
3.3.3 Hepatocyte functional activity
Albumin secretion, one characteristic of hepatocytes, was measured at various time points throughout cell differentiation. The mMSCs cultured in regular medium did not secrete any albumin, while mMSCs cultured in MEDDM produced albumin in a time-dependent manner. At day 15, the amount of albumin secretion reached its peak (Fig. 3).
Albumin secretion of mMSCs was analyzed by albumin ELISA. Black circle symbols represented mMSCs cultured in the MEDDM. White circle symbols represented mMSCs cultured in regular medium, which was used as the negative control. Black triangle symbols represented primary mouse hepatocytes, which was used as the positive control.
To check hepatocyte functionality, urea production was examined. While mMSCs cultured in regular medium did not produce urea, cells cultured in the MEDDM produced urea after three days and the production of urea was found to be time-dependent (Fig. 4).
Urea production of mMSCs was analyzed by urea assay. Black circle symbols represented mMSCs cultured in MEDDM. White circle symbols represented mMSCs cultured in regular medium, which was used as the negative control. Black triangle symbols represented primary mouse hepatocytes, which was used as the positive control.
Liver development proceeds through multiple distinct steps with several growth factors and cytokines involved in each step in vivo. In mice, the initial event in liver development occurs at E9 and the foregut endoderm was directed to become the liver through interaction with the cardiogenic mesoderm (Zaret, 2000, 2001). During this process, growth factor FGF-4 is involved in endoderm specification and hepatic differentiation (Wells and Melton, 1999). Another factor, OSM, also plays an important role in hepatic maturation during the mid- to late-fetal stages (Taga and Kishimoto, 1997). In addition, HGF was reported to be an important component for liver development and differentiation, involved in the postnatal hepatic maturation as a paracrine factor (Somerset et al., 1997). Other cytokines such as EGF, insulin and TGF have also been reported to play crucial roles in liver development and regeneration (Kamiya et al., 2001).
Cell differentiation in vitro, however, may be different from the in vivo process. It is not known which cytokines affect the process of stem cells' differentiation into hepatocytes in vitro, though some researchers have suggested that the main players include HGF (Wang et al., 2004; Oh et al., 2000; Fiegel et al, 2003); HGF and EGF (Miyazaki et al., 2002); or HGF and FGF (Schwartz et al., 2002). Could all these factors – HGF, FGF-4, EGF and OSM (Wells and Melton, 1999; Taga and Kishimoto, 1997; Somerset et al., 1997; Kamiya et al., 2001; Zaret, 2000) – regulate the differentiation process in vitro? Do these factors interact with each other? This research addressed these questions and attempted to identify the most effective directed-differentiation medium through uniform design. Liver-specific mRNA, protein expression and cell functional assay were utilized for the validation of the MEDDM. Previous data from our lab clearly demonstrated hepatocyte-like cell generation when culturing mouse bone marrow mononuclear cells in a directed differentiation medium containing factors FGF-4, OSM, HGF and EGF (Shi et al., 2005). Based on previous results, new compositions of medium containing various amounts of HGF, FGF-4, EGF and OSM were set up for studies. Since ALB and CK18 are important markers of hepatocytes (Schwartz et al., 2002; Pan et al., 1998; Tateno and Yoshizato, 1996), they were chosen as two response variables for analyzing the relationship between the four factors (HGF, FGF-4, EGF and OSM) and variables.
Among the three common experimental designs (comprehensive, orthogonal and uniform), the uniform experimental design approach was selected. Though the comprehensive approach could be more accurate since all levels of all factors are chosen, the number of experiments could easily become unmanageable (Atkinson and Donev, 1992). For an experimental design of four factors and four levels, the number of experiments would be 256 every time if the comprehensive approach had been used. Orthogonal design utilizes a suit of specific orthogonal tables, and arrays the most representative experiments. Orthogonal design can reasonably decrease the number of experiments without compromising the information gained from the study. On the other hand, to avoid interaction effects (Goupy, 1993; Fang, 1980; Zeng, 1994), 66 experiments would have had to be run if the orthogonal design had been chosen. The uniform design of four factors and four levels, however, requires significantly less experiments. The key point for uniform design is to choose the most representative factors so as to obtain as much information as possible by a few experiments (Zeng, 1994; Fang et al., 2000). Uniform design approach was chosen because of the following advantages: (1) the number of experiments is less, thus it is economical; (2) it takes interaction effects into consideration; (3) regression models can be established to reveal the relationship between response elements and factors in the experiment domain; (4) experimental conditions can be optimized by identifying the optimal combination of factors at different levels; and (5) uniform design table is easily accessible for users on the web at http://www.math.hkbu.edu.hk/UniformDesign (Fang et al., 2000; Liang et al., 2001).
The U8 (44) system was chosen as the suitable uniform design of four factors and four levels. Accordingly, eight different mediums were set up to study the differentiation of mMSCs into hepatocytes. At the end of the differentiation period (day 21), flow cytometric analysis was conducted to evaluate the percentage of ALB and CK18 positive cells in each group. The stepwise regression analysis data suggested that the percentage of ALB-expressed cells correlated with OSM; the percentage of CK18-expressed cells correlated with both FGF-4 and OSM. These results are in accordance with the functions of different factors in liver development (Wells and Melton, 1999; Taga and Kishimoto, 1997; Zaret, 2000, 2001).
To explore the effect of the MEDDM, we identified the directed differentiated cells at gene level, protein level and cell's synthetical and metabolic function. We selected AFP, ALB, CK18 and TTR as markers of hepatocytes. AFP is a marker of endodermal differentiation as well as an early fetal hepatic marker. Expression of ALB starts in early fetal hepatocytes and reaches the maximal level in adult hepatocytes (Pan et al., 1998). CK18 is a cytoskeletal protein and expressed in mature hepatocytes. TTR represents endodermal differentiation and is expressed throughout liver maturation (Makover et al., 1989). We detected all these markers at the gene or protein level in the course of cell-directed differentiation. Cells cultured in the MEDDM also produced albumin, synthesized urea and glycogen in a time dependent manner. Although renal tubular epithelium cells and hepatocytes can synthesize urea, only hepatocytes can synthesize albumin (Dunn et al., 1991). Therefore, the combined data suggest that directed differentiated mMSCs through MEDDM have hepatocyte function. Furthermore, the differentiation efficiency of the new medium was much higher than previously reported (Shi et al., 2005). These data validated our experimental design approach and confirmed the effectiveness of the MEDDM containing FGF-4 and OSM in the differentiation of mMSCs into functional hepatocytes. Next, we will serve these functional hepatocytes as resources for bioartificial liver or hepatocyte transplantation.
The uniform experimental design approach was suitable for screening the most effective differentiation medium of mMSCs. The MEDDM containing 35
This work was supported by the National Natural Science Foundation of China (No.: 30371391) and the National Natural Science Foundation of Jiangsu Province, PR China (No.: BK2003008).
Atkinson AC, Donev, AN. Optimum experimental designs. 1992:
Atkinson AC, Donev, AN. . Genes that control the development of migrating muscle precursor cells 2000:12:6:725-30
Fang KT. The uniform design: application of number-theoretic methods in experimental design. Acta Math Appl Sin 1980:3:363-72
Fiegel HC, Lioznov, MV, Cortes-Dericks, L, Lange, C, Kluth, D, Fehse, B. Liver-specific gene expression in cultured human hematopoietic stem cells. Stem Cells 2003:21:98-104
Goupy JL. Methods for experimental design, principles and application for physicists and chemists. 1993:
Lagasse E, Connors, H, Al-Dhalimy, M, Reitsma, M, Dohse, M, Osborne, L. Purified hematopoietic stem cells can differentiate into hepatocytes in vivo. Nat Med 2000:6:11:1229-34
Jiang Y, Jahagirdar, BN, Reinhardt, RL, Schwartz, RE, Keene, CD, Ortiz-Gonzalez, XR. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature 2002:418:6893:41-9
Kamiya A, Kinoshita, T, Miyajima, A. Oncostatin M and hepatocyte growth factor induce hepatic maturation via distinct signaling pathways. FEBS Lett 2001:492:1–2:90-4
Levite M, Meshorer, A, Reisner, Y. A rapid method for obtaining murine bone marrow cells in high yield. Bone Marrow Transpl 1991:8:3:225-7
Makover A, Soprano, DR, Wyatt, ML, Goodman, DS. An in situ hybridization study of the localization of retinol-binding protein and transthyretin messenger RNAs during fetal development in the rat. Differentiation 1989:40:1:17-25
Miyazaki M, Akiyama, I, Sakaguchi, M, Nakashima, E, Okada, M, Kataoka, K. Improved conditions to induce hepatocytes from rat bone marrow cells in culture. Biochemical and Biophysical Research Communication 2002:298:24-30
Oh SH, Miyazaki, M, Kouchi, H, Inoue, Y, Sakaguchi, M, Tsuji, T. Hepatocyte growth factor induces differentiation of adult rat bone marrow cells into a hepatocyte lineage in vitro. Biochemical and Biophysical Research Communication 2000:279:500-4
Schwartz RE, Reyes, M, Koodie, L, Jiang, Y, Blackstad, M, Lund, T. Multipotent adult progenitor cells from bone marrow differentiate into functional hepatocyte-like cells. J Clin Invest 2002:109:1291-302
Shi XL, Qiu, YD, Wu, XY, Xie, T, Zhu, ZH, Chen, LL. In vitro differentiation of mouse bone marrow mononuclear cells into hepatocyte-like cells. Hepatology Research 2005:31:4:223-31
Tateno C, Yoshizato, K. Growth and differentiation in culture of clonogenic hepatocytes that express both phenotypes of hepatocytes and biliary epithelial cells. Am J Pathol 1996 Nov:149:5:1593-605
Theise ND, Badve, S, Saxena, R, Henegariu, O, Sell, S, Crawford, JM. Derivation of hepatocytes from bone marrow cells in mice after radiation-induced myeloablation. Hepatology 2000:31:1:235-40
Wang PP, Wang, JH, Yan, ZP, Hu, MY, Lau, GK, Fan, ST. Expression of hepatocyte-like phenotypes in bone marrow stromal cells after HGF induction. Biochemical and Biophysical Research Communication 2004:320:712-6
Received 25 October 2007/21 December 2007; accepted 2 April 2008doi:10.1016/j.cellbi.2008.04.013