Tumorigenicity-associated characteristics of human iPS cell lines

Abstract

Human induced pluripotent stem cells (hiPSCs) represent promising raw materials of human cell-based therapeutic products (hCTPs). As undifferentiated hiPSCs exhibit intrinsic tumorigenicity properties that enable them to form teratomas, hCTPs containing residual undifferentiated hiPSCs may cause tumor formation following transplantation. We first established quantitative and sensitive tumorigenicity testing of hiPSCs dissociated into single cells using NOD/Shi-scid IL2Rγnull (NOG) mice by inhibiting apoptosis of hiPSCs with a Rho kinase inhibitor. To examine different features in tumorigenicity of various hiPSCs, 10 commonly available hiPSC lines were subjected to in vivo tumorigenicity testing. Transplanted hiPSC lines showed remarkable variation in tumor incidence, formation latency, and volumes. Most of the tumors formed were classified as immature teratomas. However, no signs of malignancies, such as carcinoma and sarcoma, were recognized in the tumors. Characteristics associated tumorigenicity of hiPSCs were investigated with microarray analysis, karyotype analysis, and whole exome sequencing. Gene expression profiling and pathway analysis supported different features of hiPSC lines in tumorigenicity. hiPSC lines showed chromosomal abnormalities in some lines and 61-77 variants of cancer-related genes carrying effective nonsynonymous mutations, which were confirmed in the COSMIC databases. In this study, the chromosomal abnormalities and cancer-related gene mutations observed in hiPSC lines did not lead to the malignancy of tumors derived from hiPSCs. Our results suggest that the potential tumorigenicity risk of hCTPs containing residual undifferentiated hiPSCs is dependent on not only amounts of undifferentiated hiPSCs but also features of the cell lines used as raw materials, a finding that should be considered from the perspective of quality of hCTPs used.

Fig 1. Tumor incidence of the 201B7 hiPSC line.

201B7 cells were subcutaneously injected into NOG mice at the indicated doses (0, 1 × 10², 1 × 10³, or 1 × 10⁴ cells) with Matrigel. The 201B7 cells were transplanted as clumps (A), single cells with 1 × 10⁶ mitomycin C-treated normal human neonatal dermal fibroblasts (NHDF) (B), or single cells with 1 × 10⁶ mitomycin C-treated NHDF and 10 μM Y-27632 (C). Tumor formation was examined for 16 weeks. Six (A, B) or ten (C) mice were used in each group.

Fig 2. Tumor incidence of mice subcutaneously injected with 10 hiPSC lines.

Dissociated single cells of hiPSC lines (201B7, 253G1, 409B2, 454E2, HiPS-RIKEN-1A, HiPS-RIKEN-2A, HiPS-RIKEN-12A, DYR0100, HYR0103, and mc-iPS) were subcutaneously transplanted into NOG mice at 3 × 10⁴ cells with Matrigel and 1 × 10⁶ mitomycin C-treated NHDF in the presence of 10 μM Y-27632. Tumor formation was examined for 16 weeks. Six mice were used in each group.

Fig 3. Time course of tumor volume in mice subcutaneously injected with 10 hiPSC lines.

Dissociated single cells of hiPSC lines (201B7, 253G1, 409B2, 454E2, HiPS-RIKEN-1A, HiPS-RIKEN-2A, HiPS-RIKEN-12A, DYR0100, HYR0103, and mc-iPS) were subcutaneously transplanted into NOG mice at 3 × 10⁴ cells with Matrigel and 1 × 10⁶ mitomycin C-treated NHDF in the presence of 10 μM Y-27632. Tumor size was observed for 16 weeks. Six mice were used in each group.

Fig 4. Representative images of teratomas subcutaneously formed in NOG mice injected with hiPSCs in a mixture of Matrigel, mitomycin C-treated NHDF, and 10 μM Y-27632.

Excised tumor samples were stained with hematoxylin and eosin. Low power view of teratoma (A), ectodermal neuroepithelia (B), melanocytes (C), mesodermal cartilage (D), and endodermal intestinal tract-like ducts (E).

Fig 5. Karyotype analysis of the 10 hiPSC lines.

In vitro cultured hiPSC lines (201B7, 253G1, 409B2, 454E2, HiPS-RIKEN-1A, HiPS-RIKEN-2A, HiPS-RIKEN-12A, DYR0100, HYR0103, and mc-iPS) were subjected to karyotype analysis. 454E2 and mc-iPS lines showed trisomy 12. Chromosomal aberrations are encircled.