Potential conflict of interest: Nothing to report.
Oval cells and small hepatocytes (SHs) are known to be hepatic stem and progenitor cells. Although oval cells are believed to differentiate into mature hepatocytes (MHs) through SHs, the details of their differentiation process are not well understood. Furthermore, it is not certain whether the induced cells possess fully mature functions as MHs. In the present experiment, we used Thy1 and CD44 to isolate oval and progenitor cells, respectively, from D-galactosamine-treated rat livers. Epidermal growth factor, basic fibroblast growth factor, or hepatocyte growth factor could trigger the hepatocytic differentiation of sorted Thy1+ cells to form epithelial cell colonies, and the combination of the factors stimulated the emergence and expansion of the colonies. Cells in the Thy1+-derived colonies grew more slowly than those in the CD44+-derived ones in vitro and in vivo and the degree of their hepatocytic differentiation increased with CD44 expression. Although the induced hepatocytes derived from Thy1+ and CD44+ cells showed similar morphology to MHs and formed organoids from the colonies similar to those from SHs, many hepatic differentiated functions of the induced hepatocytes were less well performed than those of mature SHs derived from the healthy liver. The gene expression of cytochrome P450 1A2, tryptophan 2,3-dioxygenase, and carbamoylphosphate synthetase I was lower in the induced hepatocytes than in mature SHs. In addition, the protein expression of CCAAT/enhancer-binding protein alpha and bile canalicular formation could not reach the levels of production of mature SHs. Conclusion: The results suggest that, although Thy1+ and CD44+ cells are able to differentiate into hepatocytes, the degree of maturation of the induced hepatocytes may not be equal to that of healthy resident hepatocytes. (HEPATOLOGY 2013)
The liver normally exhibits a very low level of cell turnover, but when loss of mature hepatocytes (MHs) occurs, a rapid regenerative response is elicited from all cell types in the liver to restore the organ to its initial state. The loss may occur as a result of toxic injury, viral infection, trauma, or surgical resection. Because hepatocytes are the major functional cells of the liver, large-scale hepatocytic loss becomes a trigger for regeneration, and replication of existing hepatocytes is generally the quickest, most efficient way to compensate for the lost functions. However, when the replication of hepatocytes is delayed or entirely inhibited, hepatic stem/progenitor cells (HPCs) are activated.1-3 As HPCs, oval cells and small hepatocytes (SHs) are well known. Oval cells were first reported to be cells that possessed an ovoid nucleus and scant cytoplasm.4 The appearance of oval cells has been reported in rat livers treated with hepatotoxins, such as 2-acetylaminofluorene (2-AAF), combined with partial hepatectomy (PH) and D-galactosamine (GalN).1,5–7 In GalN-induced rat liver injury, it has been shown that oval cells appear in the periportal area and differentiate into MHs through basophilic small-sized ones.8,9 Oval cells show a wide range of phenotypic heterogeneity, and cytokeratins (CKs) 7 and 19, alpha-fetoprotein (AFP), CD34, c-kit, and Thy1 have been reported as markers for them.1,2,5–7
On the other hand, SHs are a subpopulation of hepatocytes, and cells isolated from healthy adult rats10,11 and human livers12 can clonally proliferate to form colonies and differentiate into MHs in vitro.11,13 Recently, we identified CD44 as a specific marker of SHs.14 In GalN-treated rat livers, CD44+ cells appear near the periportal area between Thy1+ oval cells and resident hepatocytes soon after the emergence of Thy1+ oval cells.15 In addition, we previously showed that Thy1+ oval cells differentiate into hepatocytes through CD44+ cells.15,16 Our data suggested that cells sequentially converted from Thy1+CD44− to Thy1+CD44+ and then to Thy1−CD44+ cells during the process of hepatocytic differentiation of oval cells.15,16 Furthermore, sorted Thy1+ and CD44+ cells could repopulate host livers when they were transplanted into rat livers treated with retrorsine (RET) and two-thirds PH.
Although most oval cells are thought to differentiate into MHs, the details of their differentiation process, such as factors for hepatic commitment, characteristics of intermediate cells, and their fates are not well understood. In addition, it has not been elucidated whether the induced hepatocytes differentiate to possess the same capabilities as MHs. In the present experiment, we aimed to clarify which factors might induce hepatocytic differentiation of Thy1+ cells and to examine how Thy1+ cells could differentiate into hepatocytes through CD44+ cells. In addition, we examined whether the Thy1+ and CD44+ cells could differentiate into fully MHs, as with those in the healthy adult liver.
Male F344 rats (dipeptidylpeptidase IV [DPPIV]+ strain; Sankyo Lab Service Corporation, Inc., Tokyo, Japan), weighing 150-200 g, were used. All animals received humane care, and the experimental protocol was approved by the committee on laboratory animals according to Sapporo Medical University guidelines. For GalN-injured livers, GalN (75 mg/100 g body weight [BW] dissolved in phosphate-buffered saline [PBS]; Acros, Geel, Belgium) was intraperitoneally (IP) administered.14 For the transplantation experiment, female F344 rats (DPPIV− strain; Charles River Laboratories, Wilmington, MA) were (IP) given two injections of RET (30 mg/kg BW; Sigma-Aldrich Chemical Co., St. Louis, MO), 2 weeks apart,17 and 4 weeks after the second injection, two-thirds PH was performed (RET/PH liver). Sorted DPPIV+ cells (5 × 105 cells/0.5 mL) were transplanted into RET/PH livers (DPPIV−) through the spleen (at least 3 rats per group).
Isolation and Culture of Cells.
Rats were used to isolate hepatic cells by the collagenase perfusion method, as previously described.18 After perfusion, the cell suspension was centrifuged at 50 × g for 1 minute. The supernatant and the precipitate were used for sorting Thy1+ and CD44+ cells and preparing MHs, respectively. The procedure used for cell sorting was as previously described,15 with some modifications. Antibodies (Abs) used for cell sorting are listed in Supporting Table 1. Thy1+CD44+ cells were sorted from CD44+ cell and Thy1+ cell fractions by using anti-Thy1 or CD44 Abs, respectively, and both were pooled. Furthermore, Thy1+ and CD44+ cells were also separated from CD44− and Thy1− cell fractions, respectively. After the number of viable cells was counted, 1 × 105 viable cells were plated in 12-well plates (Corning Inc., Corning, NY) and cultured in the medium listed in Supporting Table 2. The medium was replaced with fresh medium thrice-weekly.
To examine whether cells in the colonies could fully differentiate into MHs and form functional bile canaliculi (BCs), Thy1+CD44− and Thy1−CD44+ cells sorted from GalN-D3 and SHs derived from a healthy liver were cultured for 10 days. Thereafter, some dishes were treated with Matrigel (BD Biosciences, San Diego, CA) for 10 days. To enhance the organoid formation of the colonies, as previously reported,19 colonies were separated from dishes by using Cell Dissociation Solution (Sigma-Aldrich), and colonies (2 × 103) were replated on collagen-coated dishes. Cells were cultured in the induction medium (Supporting Table 2) for 14 days. Cloning rings were used to isolate total RNA of each colony. At least two separate experiments were performed, and more than five colonies were investigated.
GeneChip Analysis, RNA Isolation, and Real-Time Polymerase Chain Reaction.
Details are shown in the Supplementary Methods.
For detecting CD44+ colonies, cells were fixed with cold absolute ethanol at 10 days after plating, and immunocytochemistry (ICC) for CD44 was carried out. Details of staining were previously reported.15 The numbers of CD44+ colonies at days 5 and 10 were counted, and positivity was calculated. Three separate experiments were performed. To measure the labeling index (LI), 40 µM of 5-bromo-2′-deoxyuridine (BrdU) were added to the medium 24 hours before fixation. In double ICC for CD44 and BrdU, a combination of the avidin-biotin peroxidase complex method (Vectastain ABC Elite Kit; Vector Laboratories Inc., Burlingame, CA) and the alkaline phosphatase method was used. For fluorescent immunohistochemistry, sliced liver samples were frozen using isopentane/liquid nitrogen, and materials were kept at −80°C until use. All Abs used for immunostaining are listed in Supporting Table 1. Sections were embedded with 90% glycerol including 0.01% p-phenylenediamine and 4,6-diamidino-2-phenylindole. A confocal laser microscope (Olympus, Tokyo, Japan) was used for observation, and findings were analyzed using DP Manager (Olympus).
Treatment With Fluorescein Diacetate.
As previously reported,20 fluorescein diacetate (FD; Sigma-Aldrich) was dissolved in dimethyl sulfoxide, and the solution was diluted with the culture medium. Then, 0.25% FD was added to the medium, and the dish was rinsed three times with warm PBS. Fluorescent images were immediately photographed using a phase-contrast microscope equipped with a fluorescence device (Olympus).
Enzyme Histochemisty for DPPIV.
To identify donor cells, enzyme histochemistry for DPPIV was carried out. DPPIV enzyme activity was detected as previously described.15 DPPIV+ foci in livers were photographed using a microscope equipped with a CCD camera, and the area of each focus was measured using ImageJ software (http://rsb.info.nih.gov/ij/index.html).
All data were analyzed using Turkey-Kramer's multiple comparison test. Level of statistical significance was P < 0.05. Experimental results are expressed as the geometric mean ± standard deviation.
Characterization of Isolated Cells From Livers Treated by GalN.
As previously reported,15 Thy1+ cells differentiated into hepatocytes through a CD44+ intermediate state, as shown with clonally cultured Thy1+ cells and cell transplantation. This transition likely happened in the GalN-treated rat liver as well. GeneChip data (Affymetrix, Inc., Santa Clara, CA) indicated that the immature hepatocyte markers, Dlk21 and AFP were up-regulated in Thy1+CD44+ and Thy1−CD44+ cells, whereas markers related to hepatic differentiation were gradually up-regulated during the transition from Thy1-D2 to CD44-D4 cells (Fig. 1). The results also suggested that most D2-Thy1+ cells were not committed to the hepatic lineage. This is consistent with our previous finding that Thy1+ cells isolated from GalN-D2 could form a few epithelial cell colonies in the standard medium for SH induction, whereas those from GalN-D3 certainly formed colonies consisting of CD44+ cells. Therefore, we considered the possibility that Thy1+ cells became the hepatocyte lineage between D2 and D3. To specify the factors that trigger hepatic commitment, we compared expression patterns of genes related to receptors of growth factors and cytokines and selected 12 candidates (Table 1).
Table 1. Effects of Growth Factors and Cytokines on the Formation of Epithelial Cell Colonies
Induction of Epithelial Cell Colonies by Growth Factors.
D2-Thy1+ cells were cultured in the medium supplemented with each factor. To elucidate the formation of epithelial cell colonies, ICC for CD44 was conducted 10 days after plating. Of the 12 candidates, only epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), and hepatocyte growth factor (HGF) could induce colonies (Fig. 2A; Table 1). CD44 expression of cells varied among the colonies, and some colonies consisted of cells with low expression of CD44 (CD44− cells). Next, we examined whether bFGF and/or HGF could enhance the formation and expansion of colonies in the culture with EGF (Fig. 2B). Compared to EGF only (control), the addition of bFGF or HGF did not enhance the frequency of colony formation. In the combination of EGF and bFGF or HGF, the number of cells per colony increased to twice as many as in the control (Fig. 2C). In addition, the combination of the three factors also dose dependently increased the number of cells per colony. These results suggested that a certain number of Thy1+ cells possessed the ability to differentiate into hepatic cells, and that the induction was initiated by EGF, bFGF, and/or HGF.
Induction of Epithelial Cell Colonies by Extracellular Matrix.
Because CD44 is one of the receptors of hyaluronic acid (HA)22,23 and because SHs can selectively proliferate on HA,18 we investigated whether extracellular matrix (ECM) affected the frequency of emergence and phenotype of colonies derived from D2-Thy1+ cells. Sorted D2-Thy1+ cells were cultured on dishes coated with type I collagen, fibronectin, laminin, and HA, and ICC for CD44 was performed 10 days after plating. When cells were cultured in the medium supplemented with EGF, frequency of colony formation was significantly higher for cells on HA-coated dishes than for the control (Fig. 3A), but no difference was observed in the number of cells per colony among the dishes with each ECM (Fig. 3B).
Growth Ability and CD44 Expression of Cells Sorted From GalN-D3.
Thy1+CD44− (Thy1), Thy1+CD44+, and Thy1−CD44+ (CD44) cells sorted from a GalN-D3 liver were cultured in the medium with EGF for 10 days. Double ICC for CD44 and BrdU was carried out (Fig. 4A–C). The frequency of colony formation was more than four times higher for CD44 cells than for both Thy1 and Thy1+CD44+ cells (Fig. 4D), and the average number of cells per colony was significantly larger for CD44 cells than for Thy1 and Thy1+CD44+ cells (Fig. 4E). The percentages of BrdU+ cells were approximately 70% and 80% in colonies derived from Thy1 and CD44 cells, respectively (Fig. 4A–C, F). Growth ability of Thy1+CD44+ cells was also intermediate between those of Thy1 and CD44 cells.
Intensity and the localization of CD44 varied among cells forming colonies. In spite of the origin of sorted cells, CD44 protein was usually expressed in cell membranes between cells (Fig. 4C). Some colonies consisted of cells with CD44 protein localized in both the cell membrane and cytoplasm (Fig. 4A, B). The latter type of colony was often observed in the culture of Thy1 cells. CD44 positivity of Thy1+ cells in a colony was approximately 65% at day 5 and increased to approximately 80% at day 10 (Fig. 4G).
Next, to examine whether acquisition of CD44 expression in Thy1+ cells was also correlated to growth ability of cells in vivo, cell transplantation was carried out. D2-Thy1+, D3-Thy1+CD44−, D3-Thy1+CD44+, D3-Thy1−CD44+, and D4-CD44+ cells (5 × 105 cells/rat) isolated from GalN-treated livers were intrasplenically transplanted into RET/PH-treated rats. One month after transplantation, the number of cells in foci derived from D4-CD44 was much larger than in those from Thy1-expressing cells (Fig. 4H, I). The growth rate of engrafted cells increased in correlation with the expression of CD44 and time after GalN treatment. In addition, no types of donor-derived (Y chromosome+) cells, other than hepatocytes, could be found in recipient livers (Supporting Fig. 1).
Gene Expression of Hepatic Markers of Epithelial Cell Colonies Derived From D3-Thy1 Cells.
To elucidate the characteristics of cells in colonies derived from D3-Thy1 cells, quantitative polymerase chain reaction (qPCR) of cells was performed for each colony, which was separated from the culture dish using a cloning ring. Phase-contrast photos were taken of every colony, and qPCR was performed. Gene-expression patterns of the colonies were roughly divided into two groups by the level of CD44 expression. Results for a representative colony in each group are shown in Fig. 5 (results for other colonies are shown in Supporting Fig. 2). Although intensity of gene expression varied among colonies, all colonies expressed CD44 messenger RNA (mRNA). Compared to cells in CD44low colonies, those in CD44high colonies showed not only relatively high expression of albumin and hepatocyte nuclear factor (HNF)-4α, but also suppression of CK-19 expression. Although coexpression of CD44 and Thy1 was observed in many Thy1-derived colonies, some large cells (arrowheads) exhibited no expression of either protein (Fig. 5E, a–d). There were colonies consisting of a mixture of albumin (Alb)+ and CK-19+ cells (Fig. 5E, e–h), though most cells expressed Alb and only a few cells exhibited CK-19 (Fig. 5E, i–l). To induce maturation of cells in Thy1-derived colonies, cells were treated with Matrigel. The treatment dramatically decreased Thy1 expression and increased levels of both HNF-4α and Alb (Supporting Fig. 3A). In addition, a marked increase of Alb secretion was also observed in cells with Matrigel (Supporting Fig. 3B). However, neither CD44 nor CK-19 expression was changed by the treatment.
Induction of Maturation in Cultured Cells.
Next, we examined whether the newly generated hepatocytes could reconstruct hepatic organoids with highly differentiated functions. To enhance the organoid formation of colonies, colonies derived from D3-Thy1, CD44, and SHs from a healthy liver were replated on collagen-coated dishes to increase their density. In contrast to the Matrigel treatment, this procedure resulted in natural organoid formation, which consisted of piled-up cells with BCs. The expression of CCAAT/enhancer-binding protein (C/EBP)-α was ICC examined, and expression of genes related to hepatic differentiated functions was investigated by qPCR. In addition, to certify the function of the newly formed BCs, FD was added to the culture medium and the ability to secrete fluorescence into BCs was examined.
In spite of their origins, some cells in colonies became large and piled up with time after replating. Morphologically, BCs and cyst-like structures were observed in colonies, similar to those in colonies formed by SHs derived from healthy rat liver.13,19,20 However, ICC for C/EBP-α revealed that the numbers of positive nuclei in Thy1- and CD44-derived colonies were smaller than in SH-derived colonies (Figs. 6A, B). Results of qPCR for each colony derived from Thy1, CD44, and SHs revealed that gene expression of cytochrome P450 1A2 (CYP1A2), tryptophan 2,3-dioxygenase (TDO), and carbamoylphosphate synthetase I (CPS-I), which are regarded as indicators for differentiated hepatic functions, was significantly higher in SH than CD44 or Thy1 (Fig. 6C). However, expression of tyrosine aminotransferase (TAT) was not different among cells.
In a SH-derived colony, fluorescence was secreted into BCs and accumulated in cysts (Fig. 7, e,f). The networks of BCs were well developed, corresponding to the regions of piled-up cells. On the other hand, in the colonies derived from both Thy1 (Fig. 7A, a,b) and CD44 cells (Figs. 7A, c,d), part of the region consisting of piled-up cells had a green, patch-like appearance. This phenomenon indicated the retention of fluorescence in the cytoplasm. In addition, to quantitatively compare structural differentiation among cells, the total length of BCs was measured in each colony. Total length of BCs was significantly larger in the SH-derived colony than in the Thy1- and CD44-derived colonies (Fig. 7B). These results suggested that newly generated hepatocytes derived from Thy1 and CD44 were not as mature as those from SHs.
Hepatocytic Differentiation of Thy1-Positive Cells.
Thy1 was first identified as a marker of oval cells by Petersen et al.24 and then widely used in experiments with HPCs. Recently, a question was raised about the validity of Thy1 as a marker for oval cells.25,26 It was reported that Thy1 was not a marker of oval cells, but of hepatic myofibroblasts and/or stellate cells. Although the issue regarding whether Thy1 is a marker for hepatic stem/progenitors is open to debate, we recently found that some Thy1+ cells isolated from GalN-treated livers differentiated into CD44+ hepatocytes through Thy1+CD44+ cells.15 These results suggested that the population of Thy1+ cells was heterogeneous, and that it contained putative hepatic stem cells possessing the ability to differentiate into hepatocytes. Interestingly, colony formation was clearly observed in the culture of cells from GalN-D3, whereas it was rarely observed in Thy1+ cells isolated from GalN-D2, suggesting that Thy1+ cells became the hepatic lineage between D2 and D3. In the present experiment, we demonstrated that EGF, fibroblast growth factor (FGF), and HGF might trigger the commitment to the hepatic lineage of Thy1+ cells, some of which possess capability as putative stem cells. It has been reported that transforming growth factor (TGF)-α, HGF, and FGF play important roles in stem/progenitor cell-mediated liver regeneration. Indeed, the growth factors are transcriptionally up-regulated during the period of active proliferation and differentiation of progenitor cells in rat liver27–29 and appear to drive the early proliferation of the progenitor cell compartment.27,30 Interestingly, it has been reported that stellate cells, which proliferate concomitantly and in close contact with progenitor cells,31 appear to be the main source of TGF-α, HGF, and acidic FGF, whereas the corresponding cognate receptors are strongly expressed in progenitor cells, suggesting that the regulation of progenitor cell proliferation and differentiation by growth factors occurs primarily in a paracrine manner.2,5 On the other hand, it is well known that priming factors are necessary for the emergence and proliferation of HPCs.5,7 A correlation between the severity of liver disease and the magnitude of the response of hepatic progenitor cells has been reported and inflammatory cytokines, such as tumor necrosis factor alpha (TNF-α), interleukin (IL)-6, and interferon-gamma (INF-γ), were suggested to play central roles as priming factors in rodents.32–36 However, in the present experiment, TNF-α, IL-6, and INF-γ could not induce the epithelial differentiation of Thy1+ cells or enhance their expansion. This might be because D2-Thy1+ cells have already been primed by the inflammation induced by GalN.
Expression of CD44 in Thy1-Derived Cells.
In this experiment, we demonstrated that the growth and degree of hepatocytic differentiation of Thy1+ cells were correlated with the expression of CD44. The results of GeneChip analysis demonstrated that the expression of genes related to hepatocytic differentiation, which were absent in GalN-D2-Thy1 cells, progressively increased in the order D3-Thy1, D3-Thy1+CD44+, and D4-CD44 cells. Although results of qPCR showed that the degree of expression of the genes of interest was different among epithelial colonies derived from D3-Thy1 cells, the cells in the CD44high colonies showed high expression of Alb and HNF-4α, compared to cells in CD44low colonies. In fact, CD44 expression in Thy1+ cells increased with time in culture.
Acquisition of CD44 expression in Thy1+ cells was also correlated with growth ability of cells. Growth of cells in CD44+ cell-derived colonies was clearly faster than that of cells in colonies derived from Thy1-expressing cells. High growth activity of CD44+ cells was also shown in the cell transplantation experiment. One month after transplantation, the number of cells in the foci derived from D4-CD44 was much larger than in those from Thy1-expressing cells. In general, the growth speed of cells shows an inverse correlation with degree of cell differentiation, and less-differentiated cells can proliferate much faster than differentiated cells. However, in the present experiment, although CD44+ cells were more differentiated than Thy1+ cells, growth speed of CD44+ cells was higher than that of CD44− cells (data not shown). At present, we cannot explain these findings, and further experiments will be required to clarify the regulatory mechanism of CD44 expression.
Restricted Maturation of HPCs.
We previously reported that SHs derived from the healthy liver could spontaneously differentiate into hepatocytes that showed typical features of MHs and reconstructed three-dimensional (3D) structures by interacting with hepatic nonparenchymal cells.11 In 3D structures, a complicated network of BCs is formed, and, when FD is added, fluorescence is secreted to BCs and expands all over the colony.11,20 In the present experiment, despite their origins, cells could become large and pile up to form 3D structures that were morphologically similar to the hepatic organoids previously reported.11,13 However, fluorescence was mostly retained in the cytoplasm of cells derived from both Thy1+ and CD44+ cells. Cytoplasmic retention of fluorescence indicates that cellular polarity is not well established, so that BCs cannot be well reconstructed. The short length of BCs also showed the incomplete maturation of both Thy1- and CD44-derived cells. Furthermore, compared to the organoids derived from SHs, CD44 expression remained and the ratio of C/EBP-α+ cells was lower in organoids derived from both Thy1+ and CD44+ cells. The lower expression of CYP1A2, TDO, and CPS-I genes in organoids from Thy1 and CD44 than in those from SHs also indicated the immaturity of stem/progenitor cell-derived hepatocytes. These results demonstrated that the newly generated hepatocytes derived from Thy1+ and CD44+ cells might not have acquired the same highly differentiated functions as MHs. On the other hand, we previously reported that, compared to MHs, the repopulation efficiency of the liver by transplanted Thy1+ cells was very low, and that most Thy1-derived foci disappeared within 2 months after transplantation.16 Similar results were shown for the transplantation of CD44+ cells. Those results indicate that HPCs may not be able to survive for a long time. Detailed histological analysis at 2 weeks after transplantation revealed that the percentage of C/EBP-α+ cells in the early CD44-derived foci was lower than that in MH-derived foci. In addition, the size and shape of the cells and the distribution of the DPPIV+ membrane were more irregular in Thy1- and CD44-derived foci than in MH-derived foci, which meant that BCs were not connected among cells, and that sinusoids were indistinct.16 These cell-transplantation results may reflect the in vitro data shown in the present experiments. Shafritz et al.6 summarized previous transplantation experiments using HPCs, and found that cells showed very low efficiency of engraftment and repopulation, regardless of the condition of the recipient liver. Thus, hepatocytes induced from embryonic stem cells and other stem/progenitor cells have not yet matured to the stage at which they can efficiently repopulate the liver of an adult. In other words, to use HPCs in regenerative medicine, the state of differentiation of the cells used for cell transplantation may be very important, and a procedure for assessment of the maturation should be immediately developed.
The authors thank Ms. Minako Kuwano and Ms. Yumiko Tsukamoto for their technical assistance. The authors also thank Mr. Kim Barrymore for his help with the manuscript for this article.
This work was supported by the Ministry of Education, Culture, Sports, Science, and Technology, Japan, Grant-in-Aid for Scientific Research (C) (19566021; to N.I.), Grants-in-Aid for Young Scientists (B) (22790385, to N.I.; and 19790294, to J.K.), a grant from the Yuasa Memorial Foundation (to T.M.), and Grants-in-Aid for Scientific Research (B) (22390259, to K.H.; and 21390365, to T.M.), a program for developing the supporting system for upgrading the education and research (to T.M.).