[Show abstract][Hide abstract]ABSTRACT: Lectin-glycan-glycosyltransferase relationships and correlations of lectin array intensities with glycosyltransferase expression patterns. Lectins with FDR<0.05 are colored red. The glycosyltransferases in the expression signature are indicated by a circle in the column "Expression signature" in "Gene expression". The Pearson's correlation coefficients between the lectin signal intensities and the expression profiles of the corresponding glycosyltransferases are listed, together with the significance probabilities. The original lectin array data can be obtained by request to HT or JH.
[Show abstract][Hide abstract]ABSTRACT: Correlation coefficient matrix for all cells. Pearson’s correlation coefficients between 51 cells for the expression profiles of all genes were calculated. The abbreviations used are the same as those listed in Figure 1 and additional file 1.
[Show abstract][Hide abstract]ABSTRACT: List of 2,502 genes in the expression signature, together with the fold-changes in expression levels and FDR values. The fold-change values are listed for the minimum values among the four sets of comparisons between iPSCs and SCs (+, iPSCs>SCs; -, iPSCs<SCs), and the FDR values shown are the maximum values among these sets.
[Show abstract][Hide abstract]ABSTRACT: Locations of the glycosyltransferases detected in the present study in the pathways of “Glycan Biosynthesis and Metabolism”. The glycosyltransferases listed in Table 1 were allocated to the pathways in “1.7 Glycan Biosynthesis and Metabolism” of the KEGG GLYCAN program (http://www.genome.jp/kegg/pathway.html#glycan). The glycosyltransferases and epitopes related to differentiation are indicated by red-colored boxes and red lines, respectively, in each pathway (see the text for details).
[Show abstract][Hide abstract]ABSTRACT: Clustering for all cells by another method. Another clustering was performed by the WARD method, instead of the complete linkage method of Figure 1, with Euclidean distance, and was visualized using the Java TreeView 1.1.0 software. The gene expression values are displayed as normalized log ratios. The abbreviations used are the same as those listed in Figure 1 and additional file 1.
[Show abstract][Hide abstract]ABSTRACT: Cell lines and numbers of passages analyzed in the present study. The following abbreviations are used for the human somatic cell (SC) and induced pluripotent stem cell (hiPSC) sources: AM, amniotic membrane; PAE, placental artery endothelial; UtE, uterine endometrium; and MRC, MRC-5 cell line. The AM and MRC cell lines were named previously [22,23]. The number of passages for each cell line is indicated by the letter ‘p’ followed by an Arabic number.
[Show abstract][Hide abstract]ABSTRACT: Schematic representation of the procedure used to obtain the network signature. The procedure for obtaining the network signature from the expression signature is shown schematically. The detailed procedure is as follows: 1) We first prepare the information for the gene sets to which the transcriptional factors bind, as deduced from the ChIP-on-chip experiments ; 2) Next, we prepare the information for the gene sets that were classified using knowledge of biological functions ; 3) The large gene sets in step 1 are divided into smaller subsets, according to the classification scheme of the gene sets in step 2; 4) If at least one gene in the expression signature is included in each gene subset in step 3, then the subset is regarded as a reference network; 5) In each reference network, the enrichment probability of the genes in the expression signature is tested with a significance probability of 0.05. Thus, we narrow down the network signature from the reference networks, in terms of gene numbers; 6) The significant reference networks identified in step 5 are further tested by calculating the graph consistency probability, which assesses the consistency between the network structure and the expression data for the constituent genes . In this step, we further refine the network signature, in terms of both the network structure and the extent of gene expression; 7) Finally, we define the network signature, using the reference networks that passed the tests in steps 5 and 6.
[Show abstract][Hide abstract]ABSTRACT: Cross-validation of cell classification. The classification accuracy was evaluated by leave-one-out cross-validation (LOOCV) on the nearest-neighbor classifier, based on the Pearson's correlation distance.
[Show abstract][Hide abstract]ABSTRACT: Number matrix for common genes. The numbers of genes that were different between the iPSCs and SCs are listed on the diagonal of the matrix, and those that were shared between the four gene sets that showed expression differences between the iPSCs are listed above the diagonal. The abbreviations used are the same as those listed in Figure 1.
[Show abstract][Hide abstract]ABSTRACT: Generation of iPSCs from human PAE cells. (A) PAE cells from the arterial endothelium of a human placenta (a), and generation of hiPSCs through epigenetic reprogramming by retrovirus infection-mediated expression of OCT4, SOX2, KLF4, and c-MYC (b). (B) Expression patterns of the pluripotent cell markers, TRA-1-60, SSEA-4, NANOG, OCT3/4, and SOX2. The cell nuclei were stained with DAPI. (C) Hematoxylin-eosin staining of sections of teratomas generated by PAE-hiPSC implantation. The histological examination revealed that the tumors contain neural tissues (a: ectoderm), cartilage (b: mesoderm), and a gut-like epithelial tissue (c: endoderm).
[Show abstract][Hide abstract]ABSTRACT: Human iPS cells (hiPSCs) have attracted considerable attention for applications to drug screening and analyses of disease mechanisms, and even as next generation materials for regenerative medicine. Genetic reprogramming of human somatic cells to a pluripotent state was first achieved by the ectopic expression of four factors (Sox2, Oct4, Klf4 and c-Myc), using a retrovirus. Subsequently, this method was applied to various human cells, using different combinations of defined factors. However, the transcription factor-induced acquisition of replication competence and pluripotency raises the question as to how exogenous factors induce changes in the inner and outer cellular states.
We analyzed both the RNA profile, to reveal changes in gene expression, and the glycan profile, to identify changes in glycan structures, between 51 cell samples of four parental somatic cell (SC) lines from amniotic mesodermal, placental artery endothelial, and uterine endometrium sources, fetal lung fibroblast (MRC-5) cells, and nine hiPSC lines that were originally established. The analysis of this information by standard statistical techniques combined with a network approach, named network screening, detected significant expression differences between the iPSCs and the SCs. Subsequent network analysis of the gene expression and glycan signatures revealed that the glycan transfer network is associated with known epitopes for differentiation, e.g., the SSEA epitope family in the glycan biosynthesis pathway, based on the characteristic changes in the cellular surface states of the hiPSCs.
The present study is the first to reveal the relationships between gene expression patterns and cell surface changes in hiPSCs, and reinforces the importance of the cell surface to identify established iPSCs from SCs. In addition, given the variability of iPSCs, which is related to the characteristics of the parental SCs, a glycosyltransferase expression assay might be established to define hiPSCs more precisely and thus facilitate their standardization, which are important steps towards the eventual therapeutic applications of hiPSCs.
Full-text Article · Jun 2011 · BMC Systems Biology