Development of genetic quality tests for good manufacturing practice-compliant induced pluripotent stem cells and their derivatives

Abstract

Although human induced pluripotent stem cell (hiPSC) lines are karyotypically normal, they retain the potential for mutation in the genome. Accordingly, intensive and relevant quality controls for clinical-grade hiPSCs remain imperative. As a conceptual approach, we performed RNA-seq-based broad-range genetic quality tests on GMP-compliant human leucocyte antigen (HLA)-homozygous hiPSCs and their derivatives under postdistribution conditions to investigate whether sequencing data could provide a basis for future quality control. We found differences in the degree of single-nucleotide polymorphism (SNP) occurring in cells cultured at three collaborating institutes. However, the cells cultured at each centre showed similar trends, in which more SNPs occurred in late-passage hiPSCs than in early-passage hiPSCs after differentiation. In eSNP karyotyping analysis, none of the predicted copy number variations (CNVs) were identified, which confirmed the results of SNP chip-based CNV analysis. HLA genotyping analysis revealed that each cell line was homozygous for HLA-A, HLA-B, and DRB1 and heterozygous for HLA-DPB type. Gene expression profiling showed a similar differentiation ability of early- and late-passage hiPSCs into cardiomyocyte-like, hepatic-like, and neuronal cell types. However, time-course analysis identified five clusters showing different patterns of gene expression, which were mainly related to the immune response. In conclusion, RNA-seq analysis appears to offer an informative genetic quality testing approach for such cell types and allows the early screening of candidate hiPSC seed stocks for clinical use by facilitating safety and potential risk evaluation.

Figure 1

Schematic depiction of the protocol used to study the differentiation potential of good manufacturing practice-compliant human induced pluripotent stem cells (hiPSCs). (a) Three homozygous human leucocyte antigen (HLA)-type hiPSC lines were differentiated into three germ layers (i.e., neuronal cells, cardiomyocytes, and hepatocyte-like cells) during early and late passages at three independent collaborating institutions. The differentiation of the neuronal cells, cardiomyocytes, and hepatocyte-like cells is illustrated in orange, green, and blue boxes, respectively. The differentiation protocols are briefly presented at the bottom of each box. (b) Analysis workflow for the RNA-seq data of the hiPSCs and their derivatives. Five different analyses were performed. For the detection of chromosomal aberrations using eSNP karyotyping, we employed dedicated tools in the eSNP karyotyping package for alignment and variant calling.

Figure 2

Detection of chromosomal aberrations in human induced pluripotent stem cells (hiPSCs) and their derivatives using eSNP karyotyping. Moving median plots of 151 single-nucleotide polymorphisms (SNPs) of expressed genes from the RNA-seq data of all 22 cell lines used in this study are shown. For neuronal cells, hiPSCs and their neuronal derivatives are shown in a single plot at early (a) and late (b) passages, respectively. Black, red, and blue in the plot indicate hiPSCs, rosette progenitors, and neuronal cells, respectively. For cardiomyocytes, hiPSCs and their cardiomyocyte derivatives are shown in a single plot at early (c) and late (d) passages, respectively. Black, blue, green, and red in the plot indicate hiPSCs, stage 1 differentiation, stage 2 differentiation, and stage 3 differentiation, respectively. For hepatocytes, hiPSCs and their hepatocyte derivatives are also shown in a single plot at early (e) and late (f) passages, respectively. Black, blue, green and red in the plot indicate hiPSCs, definitive endoderm (DE) cells, hepatic endoderm (HE) cells, and hepatocyte-like cells (HLC), respectively. Coloured bars represent FDR-corrected p-values. Positions with p-values < 0.01 are indicated with a black line.

Figure 3

Characterization of differentiated cells of three lineages and the original human induced pluripotent stem cells (hiPSCs) at the transcriptome level. Global transcriptome analysis of hiPSC lines and differentiated cells. (a) Relationship of transcriptome profiles among the cells. Sample-to-sample distance matrix with hierarchical clustering. (b) Principal component analysis (PCA) of all lines. Neuronal cells (purple circles), cardiomyocytes (green circles), and hepatocyte-like cells (orange circles) were differentiated from hiPSCs. hiPSCs and iPS-like cells that divided with hiPSCs are highlighted with red circles. (c–e) Scatter plot of log2-normalized read counts for differentiated lines, neuronal cells, cardiomyocytes, and hepatocytes in the final stages compared with the original hiPSC lines. Lineage-specific markers for each lineage (filled triangles) and hiPSC-specific markers (inverted filled triangles) are indicated. Up- and downregulated genes are also denoted by red and blue circles, respectively. PAX3, paired box 3; PAX6, paired box 6; KLF, Krüppel-like factor 4; GATA binding protein 4, GATA4; AFP, alpha foetoprotein; TBX1, T-box 1. The top five gene ontology (GO) categories for the common differentially expressed genes (DEGs) identified during the differentiation process for each lineage (i.e., neuronal cells (f), cardiomyocytes (g), and hepatocyte-like cells (h)) are shown. The heatmap for GO analysis indicates the expression of each gene at each stage during differentiation.

Figure 4

Time-course transcriptome analysis of differentiation during early and late passages. Comparison of expression profiles between early (red bars) and late (blue bars) passages at three time points [human induced pluripotent stem cells (hiPSCs), the middle stage of differentiation, and the final stage of differentiation]. Forty similar expression patterns (SEPs) and five differential expression profiles (DEPs) are shown in the top and bottom panels, respectively. The normalized expression values were applied to calculate relative expression. The number at the top of each graph indicates the number of clusters. ODEP: obscure DEP.

Figure 5

Gene ontology enrichment analysis of the genes of differential expression profiles (DEPs). Pie charts show the top-ranked GO terms of biological processes for the genes in Cluster 38 (left) and cluster 72 (right). The more significant the GO term, the larger the portion of the pie chart. Only GO terms with a p-value < 0.05 were considered.