Identification and removal of unexpected proliferative off-target cells emerging after iPSC-derived pancreatic islet cell implantation

Abstract

Differentiation of pancreatic endocrine cells from human pluripotent stem cells (PSCs) has been thoroughly investigated for application in cell therapy against diabetes. In the context of induced pancreatic endocrine cell implantation, previous studies have reported graft enlargement resulting from off-target pancreatic lineage cells. However, there is currently no documented evidence of proliferative off-target cells beyond the pancreatic lineage in existing studies. Here, we show that the implantation of seven-stage induced PSC-derived pancreatic islet cells (s7-iPICs) leads to the emergence of unexpected off-target cells with proliferative capacity via in vivo maturation. These cells display characteristics of both mesenchymal stem cells (MSCs) and smooth muscle cells (SMCs), termed proliferative MSC- and SMC-like cells (PMSCs). The frequency of PMSC emergence was found to be high when 10⁸ s7-iPICs were used. Given that clinical applications involve the use of a greater number of induced cells than 10⁸, it is challenging to ensure the safety of clinical applications unless PMSCs are adequately addressed. Accordingly, we developed a detection system and removal methods for PMSCs. To detect PMSCs without implantation, we implemented a 4-wk-extended culture system and demonstrated that putative PMSCs could be reduced by compound treatment, particularly with the taxane docetaxel. When docetaxel-treated s7-iPICs were implanted, the PMSCs were no longer observed. This study provides useful insights into the identification and resolution of safety issues, which are particularly important in the field of cell-based medicine using PSCs.

Keywords: cell-based therapy; induced pluripotent stem cell–derived pancreatic islet cells; long-term safety; type 1 diabetes mellitus.

Fig. 1.

Unexpected abnormal outgrowth in s7-iPIC grafts is outside the pancreatic lineage and continues to proliferate after implantation. (A) Schematic representation of s6-iPIC and s7-iPIC differentiation and subcutaneous implantation using fibrin gel. DE, definitive endoderm; PGT, primitive gut tube; PF, posterior foregut; PP, pancreatic progenitors; EP, endocrine progenitors; Act A, activin A; CHIR, CHIR99021; K-Cyc, KAAD-cyclopamine; Nic, nicotinamide; ROCKi, ROCK inhibitor; RA, retinoic acid; ALK5i, ALK5 inhibitor II; LDN, LDN-193189; GSI, γ-secretase inhibitor; N-cys, N-acetyl cysteine. (B–G) Cell implantation experiments in streptozotocin-injected and normal NOD-scid mice. Mice were implanted with s6-iPICs or s7-iPICs (2.0 to 4.0 × 10⁶ cells/mouse) embedded in a fibrin gel into the subcutaneous space. (B) Hematoxylin and eosin (HE)-stained sections 24 wk after implantation. The Left image shows a normal s7-iPIC graft, and the two Right images show a s7-iPIC graft with abnormal outgrowth. The high-magnification image is an enlarged image of the area enclosed by the dotted line in the low-magnification image. Black scale bars indicate 500 μm at low magnification and 200 μm at high magnification. The images are representative of dozens of samples showing similar results. (C) Details of abnormal outgrowth appearance frequency after s6-iPIC and s7-iPIC implantation. (D–F) Immunohistochemical images of an s7-iPIC graft with abnormal outgrowth at 24 wk postimplantation. White scale bars indicate 500 μm at low magnification and 100 μm at high magnification. HuN; human nucleus. Images were taken from serial sections of the same sample and are representative of dozens of samples showing similar results. (G) Graft weight of samples without abnormal outgrowth (n = 82, combined total number of s6-iPIC and s7-iPIC grafts); and samples with abnormal outgrowth (n = 47, combined total number of s6-iPIC and s7-iPIC grafts). See SI Appendix, Fig. S1A for graft weight distribution according to the type and number of implanted cells. Data are shown as the mean ± SD. ***P < 0.001, Aspin–Welch test.

Fig. 2.

The unknown cell population that cannot be classified as known cell types in single-cell analysis could be responsible for abnormal outgrowth. (A) Schematic representation of vitro and vivo s6-iPIC samples subjected to scRNA-seq. (B) Macroscopic photographs of vivo s6-iPIC samples. From Left to Right, alginate gel–embedded s6-iPICs before implantation, the graft was subcutaneously engrafted or retrieved 2 mo after implantation, and the graft was retrieved 6 mo after implantation. The area surrounded by a dotted line indicates the area where implanted cells have engrafted subcutaneously. The white bar indicates 5 mm. (C) Cell distribution on t-SNE projections for four combined samples and each sample. The four samples were broken down as follows: one sample of vitro s6-iPICs, two samples of vivo s6-iPICs (2 and 6 mo after implantation), and one sample of reference human islets. (D) Shared nearest neighbor clustering identified five clusters specific to Vitro s6-iPICs (1, 2, 3, 10, and 11), six clusters specific to vivo s6-iPICs (0, 4, 7, 13, 15, and 17), and seven clusters specific to human islets (5, 8, 9, 12, 14, 19, and 20). Clusters 6 and 16 were shared by vivo s6-iPICs and human islets. Cluster 18, a population of proliferative cells, was shared by all the samples. See SI Appendix, Fig. S2C for further details. Each cluster was assigned to a known cell type using the characteristic gene expression shown in SI Appendix, Fig. S2D. β, β-fate cells or actual β-cells; α, α-fate cells or actual α-cells; EC, enterochromaffin-fate cells; ε, ε-fate cells; NEP, nonendocrine progenitor cells; Acinar, acinar-like cells or actual acinar cells; Duct, duct-like cells or actual duct cells; Unknown, unknown cells that could not be assigned to known cells; Stromal, stromal cells; VE, vascular endothelial cells; Neural stem, neural stem cells; Mitotic, Mitotic cells.

Fig. 3.

Abnormal outgrowth consists of proliferative MSC- and SMC-like cells (PMSCs). (A) Reclassified cell populations identified by RCA on the t-SNE projection. (B) Heatmap of tissues and cell lines with similarity to cells annotated as green (clusters 15 and 9 in Fig. 2D) in the RCA. See SI Appendix, Fig. S3A for a full RCA heatmap. (C) Bubble plot of the top 20 differentially expressed genes in cluster 15 (P < 0.05, fold-change > 1.2, and pct.2 < 0.05). Color intensity indicates average relative expression levels. The bubble size indicates the percentage of expressing cells. (D) Single-cell gene expression of MSC- and cluster 15–specific SMC-related markers on the t-SNE projection. (E and F) Immunohistochemical images of an s7-iPIC graft with abnormal outgrowth at 24 wk postimplantation. White scale bars indicate 500 μm at low magnification and 100 μm at high magnification. Images were taken from serial sections of the same sample as in Fig. 1 D–F and are representative of dozens of samples showing similar results. See SI Appendix, Fig. S6A for immunohistochemical data on other PMSC-related markers. (G) Hierarchical clustering analysis. Cluster 15 was most closely related to the nonendocrine progenitor cells in vitro s6-iPICs (cluster 11) and most distantly related to the β-cell population (clusters 8, 7, and 16).

Fig. 4.

An EGF-supplemented extended culture system exposes a cell population (putative PMSCs) that closely resembles in vivo PMSCs. (A) Schematic representation of extended culture, subsequent flow cytometry analysis, and scRNA-seq using s6-iPICs. (B and C) Representative flow cytometry plots illustrating the protein expression of s6-iPICs before and after extended culture. The numbers in each plot diagram show the percentage of each population. The reproducibility was confirmed in three independent experiments. (D) Cell distribution on uniform manifold approximation and projections (UMAPs) for seven combined samples and for each sample. We reanalyzed the four samples shown in Fig. 2C with the addition of the following three samples: one sample of vitro s6-iPICs cultured without PD-166866 and two samples (technical duplicates) of s6-iPICs after 4 wk of extended culture including EGF treatment. Cells in the samples after extended culture were classified into three types based on gene expression intensity: PDX1⁻/CHGA⁻, PDX1⁺/CHGA⁻, and CHGA⁺ populations. See SI Appendix, Fig. S9 A–C for further details. (E) Newly classified cell populations identified by RCA on the UMAP. The PDX1⁻/CHGA⁻ population in the extended culture samples, PMSCs in Vivo s6-iPICs, and stromal cells in human islets were classified as “pink” and “brown”. SI Appendix, Fig. S9 A and B shows the position of PMSCs in vivo s6-iPICs and stromal cells in human islets on UMAPs. (F) Heatmap of tissues and cell lines similar to cells annotated as pink and brown in the RCA. See SI Appendix, Fig. S9D for a full RCA heatmap. (G) Single-cell gene expression of PMSC markers on the UMAP.

Fig. 5.

Cisplatin and docetaxel effectively remove putative PMSCs through mechanisms other than kinase inhibition. (A) Schematic representation of s6-iPIC derivatives induced with or without additional compound treatment and subsequent extended culture and flow cytometry analysis. (B–E) Representative flow cytometry plots illustrating protein expression before and after extended culture of s6-iPIC derivatives (B and D). Live cell counts and relative cell numbers of PDX1⁻/CHGA⁻, PDX1⁺/CHGA⁻, and CHGA⁺ populations in each s6-iPIC derivative postextended culture. The number of cells in control s6-iPICs was set to “1” (C and E). The number of cells in each population was calculated from flow cytometry results (percentage of each population) and live cell counts. Data are shown as the mean ± SD (n = 3, technical replicates). The reproducibility of the effect of docetaxel was confirmed in three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. s6-iPICs, Dunnett’s test. ^$P < 0.05, ^$$P < 0.01, Aspin–Welch test.

Fig. 6.

Docetaxel treatment of s7-iPICs abrogates the appearance of off-target cells while showing therapeutic efficacy. (A) Schematic representation of docetaxel-treated s7-iPIC differentiation and subcutaneous implantation using fibrin gel. (B–J) Cell implantation experiments in streptozotocin-injected NOD-scid mice. Mice were implanted with docetaxel-treated s7-iPICs (3.0 to 4.5 × 10⁶ cells/mouse) embedded in a fibrin gel into the subcutaneous space. (B and C) Blood glucose and plasma human C-peptide levels after docetaxel-treated s7-iPIC (4.0 × 10⁶ cells/mouse) implantation. Data are shown as mean ± SD (sham: n = 5→1, docetaxel-treated s7-iPICs: n = 18→13). The decrease in the n number was caused by unexpected death. See SI Appendix, Fig. S12 A–D for efficacy data of the other iPICs. (D and E) Plasma glucose, and human C-peptide levels during the oral glucose tolerance test 23 wk after docetaxel-treated s7-iPIC (4.0 × 10⁶ cells/mouse) implantation. Data are shown as mean ± SD (sham: n = 1, docetaxel-treated s7-iPICs: n = 13). (F) HE-stained sections 24 wk after implantation. Black scale bars indicate 500 μm at low magnification and 200 μm at high magnification. The images are representative of dozens of samples showing similar results. (G–J) Immunohistochemical images of a docetaxel-treated s7-iPIC graft at 24 wk postimplantation. White scale bars indicate 500 μm at low magnification and 100 μm at high magnification. HuN; human nucleus. Images were taken from serial sections of the same sample and are representative of dozens of samples showing similar results. See SI Appendix, Fig. S12 E–H for HE and immunohistochemical images on other iPICs. (K) Details of PMSC, large cyst (>500 μm in diameter), and small cyst (<500 μm in diameter) appearance frequency after fibrin gel–embedded subcutaneous implantation of s6-iPIC, s7-iPIC, docetaxel-treated s6-iPIC, and docetaxel-treated s7-iPIC. The parts of s6-iPICs and s7-iPIC data are previously shown in Fig. 1C and are repeated here for easy comparison between all groups. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. s7-iPICs, Fisher’s exact test.