Characterization and reduction of non-endocrine cells accompanying islet-like endocrine cells differentiated from human iPSC

Hideyuki Hiyoshi, Kensuke Sakuma, Noriko Tsubooka-Yamazoe, Shinya Asano, Taisuke Mochida, Junji Yamaura, Shuhei Konagaya, Ryo Fujii, Hirokazu Matsumoto, Ryo Ito, Tarō ToyodaScientific 2022 · DOI

TL;DR

This study tackles the challenge of non-endocrine cell contamination in human induced pluripotent stem cell (iPSC)-derived pancreatic islet cells (iPIC), which are crucial for diabetes cell therapy. By using single-cell RNA sequencing, researchers identified specific markers and pathways in these contaminating cells, leading to the development of novel strategies to selectively reduce their numbers, thereby improving the purity and potential efficacy of iPIC for therapeutic use.

What Problem Does This Paper Address?

The ultimate goal of treating Type 1 diabetes is to restore insulin production. One promising avenue is differentiating human pluripotent stem cells (like iPSCs) into pancreatic islet-like endocrine cells (iPIC). While this approach offers an unlimited cell source, a significant hurdle is the inevitable presence of non-endocrine cells as byproducts of the differentiation process. These contaminating cells can proliferate, leading to enlarged grafts and potentially unwanted outcomes in vivo. Previous research has lacked a comprehensive understanding of the identity and characteristics of these non-endocrine cells, making it difficult to develop effective purification strategies. This paper aims to fill that gap by characterizing these cells at a single-cell level and using that information to devise methods for their reduction.

Methods

The core of this study relies on single-cell RNA sequencing (scRNA-seq) to analyze the transcriptomes of individual cells within iPIC populations. This allowed for the identification of distinct cell clusters and their gene expression profiles. To further refine the analysis and identify specific non-endocrine populations, the researchers employed:

  • Flow Cytometry: Used for initial characterization and protein expression analysis of iPIC.
  • Reference Component Analysis (RCA): A technique to assess the similarity of cell clusters to known tissue types or cell lines, helping to infer the developmental origins of non-endocrine cells.
  • Trajectory Analysis (Monocle 2): Used to infer developmental pathways and relationships between different cell populations.
  • Quantitative PCR (qPCR): Employed to validate the expression of specific genes identified through scRNA-seq in bulk cell populations.
  • In vitro drug treatments: Various compounds targeting specific pathways (FGFR inhibition, PLK inhibition, glycolysis inhibition) were used to assess their efficacy in selectively reducing non-endocrine cells.
  • In vivo implantation studies: iPICs were implanted into diabetic mice to assess their long-term efficacy, morphology, and the impact of any remaining non-endocrine cells.

The novelty lies in the comprehensive single-cell characterization of these contaminating cells and the subsequent development of targeted reduction strategies based on these findings.

Datasets

The study utilized the following datasets:

  • Human iPSC-derived pancreatic islet cells (iPIC): Generated in-house using a multi-step differentiation protocol.
  • Modified iPSC lines (s6-iPIC-A and s6-iPIC-B): These were generated with variations in the differentiation protocol to enrich for non-endocrine cells, aiding in their identification.
  • Reference human islet samples: Used as a benchmark for comparison with iPSC-derived cells.

The scRNA-seq data was generated from a total of 19,969 cells across these samples, with 17,100 cells used after quality control. This data was deposited in the Gene Expression Omnibus (GEO).

Key Insights and Ideas

  • Non-endocrine cells are not a single entity: The study revealed that non-endocrine cells within iPIC are heterogeneous, comprising at least two distinct populations:
  • Proliferating cells: These cells are actively dividing, contributing to potential graft overgrowth.
  • Cells with mixed lineage traits: These cells exhibit pancreatic markers alongside markers associated with liver (e.g., FGB) or intestinal (e.g., AGR2) lineages, suggesting a partial retention of developmental plasticity.
  • Specific molecular targets identified: Researchers pinpointed specific genes and pathways that are differentially expressed in non-endocrine cells compared to endocrine cells. Notably, FGFR2, PLK1, and LDHB were found to be significantly enriched in non-endocrine populations.
  • Targeted inhibition reduces non-endocrine cells: The identification of these specific targets enabled the development of novel therapeutic strategies. Inhibiting FGFR2 (using PD-166866), PLK1/4 (using GSK 461364, CFI-400945), or glycolysis (using 2-DG) selectively reduced the number of non-endocrine cells without significantly harming the desired endocrine cells.
  • FGFR2 inhibition is a key mechanism: PD-166866, a pan-FGFR inhibitor, was found to primarily target FGFR2 in non-endocrine cells, suggesting this as a crucial pathway for their survival or proliferation.
  • Glycolysis as a metabolic vulnerability: The higher expression of LDHB in non-endocrine cells points to a greater reliance on glycolysis for energy. Inhibiting glycolysis with 2-DG effectively reduced non-endocrine cells, though it also impacted endocrine cells to a lesser extent.

Key Results

  • In vivo efficacy of iPIC: Implanted iPIC grafts demonstrated long-term functionality, normalizing blood glucose levels in diabetic mice for up to 36 weeks post-implantation.
  • Non-endocrine cell characterization: scRNA-seq identified non-endocrine cells as a heterogeneous population, including proliferating cells and those with mixed liver/intestinal traits.
  • Identification of key markers: FGFR2, PLK1, and LDHB were identified as specific markers for non-endocrine cells.
  • Selective reduction of non-endocrine cells:
    • PD-166866 (FGFR inhibitor) reduced non-endocrine cells by approximately 50-70% in vitro.
  • PLK inhibitors (GSK 461364, CFI-400945) also effectively reduced non-endocrine cells.
  • 2-DG (glycolysis inhibitor) reduced non-endocrine cells by over 80% at 10 mM, while reducing endocrine cells by about 37%.
  • Validation of reduction strategies: Compounds used in the iPIC differentiation protocol (PD-166866 and TR06141363) were shown to have already reduced non-endocrine cell contamination.

Limitations and Future Work

  • In vivo sample size for safety assessment: The authors note that the in vivo safety assessment of iPIC was inconclusive due to an insufficient sample size, suggesting further studies are needed.
  • 2-DG limitations: While effective, 2-DG treatment also reduced endocrine cell numbers, albeit to a lesser extent. Future work could focus on developing LDHB subunit-specific inhibitors to further refine this metabolic targeting approach.
  • Clinical translation of reduction strategies: While PD-166866 and TR06141363 show promise for clinical application, further optimization and validation are required for large-scale cell manufacturing.
  • Understanding the "mixed lineage" cells: The precise developmental origin and fate of the non-endocrine cells exhibiting liver or intestinal traits warrant further investigation.
  • Combining reduction strategies: The authors suggest that combining previously reported methods with their newly identified strategies could further improve cell quality for large-scale production.

Summary updated

Abstract

The differentiation of pancreatic endocrine cells from human pluripotent stem cells has been thoroughly investigated for their application in cell therapy against diabetes. Although non-endocrine cells are inevitable contaminating by-products of the differentiation process, a comprehensive profile of such cells is lacking. Therefore, we characterized non-endocrine cells in iPSC-derived pancreatic islet cells (iPIC) using single-cell transcriptomic analysis. We found that non-endocrine cells consist of (1) heterogeneous proliferating cells, and (2) cells with not only pancreatic traits but also liver or intestinal traits marked by FGB or AGR2. Non-endocrine cells specifically expressed FGFR2, PLK1, and LDHB. We demonstrated that inhibition of pathways involving these genes selectively reduced the number of non-endocrine cells in the differentiation process. These findings provide useful insights into cell purification approaches and contribute to the improvement of the mass production of endocrine cells for stem cell-derived cell therapy for diabetes.
BibTeX
@inproceedings{hiyoshi2022characterization,
  author = {Hideyuki Hiyoshi and Kensuke Sakuma and Noriko Tsubooka-Yamazoe and Shinya Asano and Taisuke Mochida and Junji Yamaura and Shuhei Konagaya and Ryo Fujii and Hirokazu Matsumoto and Ryo Ito and Tarō Toyoda},
  title = {{Characterization and reduction of non-endocrine cells accompanying islet-like endocrine cells differentiated from human iPSC}},
  booktitle = {Scientific},
  year = {2022},
  doi = {10.1038/s41598-022-08753-5},
}
Figure