Transcriptomic heterogeneity of cultured ADSCs corresponds to embolic risk in the host

Abstract

Stem cell therapy emerges as an effective approach for treating various currently untreatable diseases. However, fatal and unknown risks caused by their systemic use remain to be a major obstacle to clinical application. We developed a functional single-cell RNA sequencing (scRNA-seq) procedure and identified that transcriptomic heterogeneity of adipose-derived stromal cells (ADSCs) in cultures is responsible for a fatal embolic risk of these cells in the host. The pro-embolic subpopulation of ADSCs in cultures was sorted by gene set enrichment analysis (GSEA) and verified by a supervised machine learning analysis. A mathematical model was developed and validated for the prediction of embolic risk of cultured ADSCs in animal models and further confirmed by its application to public data. Importantly, modification of culture conditions prevented the embolic risk. This novel procedure can be applied to other aspects of risk assessment and would help further the development of stem cell clinical applications.

Keywords: Stem cells research; computational bioinformatics; transcriptomics.

Graphical abstract

Figure 1

Analysis of embolism after infusion of stem cells cultured in MF-media or in IL-media (A) Human ADSCs obtained from the same donor by the same procedure, but cultured in MF-media or IL-media before infusion into mice. (B) Numerical difference in mice showing typical symptoms of pulmonary embolism induced by hADSCs obtained from MF-media or IL-media. (C) Computed tomography of mouse lungs infused with hADSCs in comparison to saline control. (D) Histopathological examination of embolism formation in the lungs postinfusion of hADSCs cultured in either MF or IL medium. Black arrows indicate phlebothrombosis; Bar = 200 μm. (E) The density of emboli found in each 10× visual field (1 × 10⁶cells per group); ∗p < 0.05. (F) PKH26 + hADSCs and lectin + vessels in the lungs of mice infused with saline or hADSCs cultured in MF-media or in IL-media. (G) Number of PKH26 + hADSCs found in each 10× visual field (1×10⁶ cells per group). (H) Detected volume of PKH26 + hADSCs in each 1.0 mm of lung tissue obtained from mice after infusion of hADSCs cultured in MF-media or in IL-media. (I–L) Examination of four blood coagulation indexes (I) APTT, Activated partial thromboplastin time, (J) PT, Prothrombin time, (K) FIB-C, Fibrinogen-C, (L) TT, Thrombin time; ∗p < 0.05. Each data point represents a biological replicate; data are mean ± s.e.m. Images are representative of six biological replicates. Data were analyzed using Student’s unpaired two-tailed t-test when comparing two conditions and ANOVA when comparing multiple conditions.

Figure 2

Detection of heterogeneity of hADSCs by a new functional clustering procedure (A) Schematic diagram of the development of the functional clustering procedure. (B and E) Functional clustering using the gene set enrichment analysis algorithm from ssGSEA of hADSCs obtained from MF-media (B) shows 6 clusters, and from IL-media (E) showing 5 clusters. (C and F) Functional clustering using the area under curve algorithm from AUCell of the hADSCs above shows 6 clusters for the cells from the MF-media (C) and 5 clusters for the cells from the IL-media (F). (D and G) Functional clustering using the Seurat module score algorithm (Seurat) of the hADSCs above showing 6 clusters for the cells from the MF-media (D) and 5 clusters for the cells from the IL-media (G). There were 28,732 hADSCs analyzed in total.

Figure 3

Detection of pro-embolic hADSC subpopulation by a new functional clustering procedure and a machine learning algorithm (A) Heatmap of top differentially expressed genes in functional clusters of hADSCs cultured in the MF-media. (B) Heatmap of top differentially expressed genes in functional clusters of hADSCs cultured in the IL-media. (C) Supervised clustering of a subset of pro-embolic hADSCs cultured in the MF-media based on known genes involved in the embolism-related pathways. (D) Supervised clustering of a subset of pro-embolic hADSCs cultured in the IL-media based on known genes involved in the embolism-related pathways. (E) Schematic diagram of the machine learning algorithm training and testing to predict potential pro-embolic cells. (F) Supervised clustering of hADSCs cultured in the MF-media based on 13 key genes identified by the machine learning algorithm. (G) Supervised clustering of hADSCs cultured in the IL-media based on 13 key genes identified by the machine learning algorithm.