GSE215376: Vascularization of pluripotent stem cell-derived islet organoids improves β-cell function

scRNA-seq Viewer
Info Studies were incorporated in a barcode, matrix, feature format. This format can be found on the 10x Genomics website for processing of single cell studies to obtain the gene expression matrices. For each study, the metadata incorporated in GEO were manually curated into profiles and the samples were separated based on applicable groups and conditions. The expression data from the cells for the samples within the same profile and condition were aggregated into an expression matrix with the cell barcodes having the sample name appended to it to ensure unique cell names.

Description (from GEO)

Submission Date: Oct 13, 2022

Summary: Blood vessels play a critical role in pancreatic islet health and function, yet current culture methods to generate islet organoids from human pluripotent stem cells (SC-islets) lack a vascular component. Here, we engineered 3D vascularized SC-islet organoids by assembling SC-islet cells, human primary endothelial cells (ECs) and fibroblasts both in a non-perfused model and a microfluidic device with perfused vessels. Vasculature improved stimulus-dependent Ca2+ influx into SC-β-cells; a hallmark of β-cell function that is blunted in non-vascularized SC-islets. We show that an islet-like basement membrane is formed by vasculature and contributes to the functional improvement of SC-β-cells. Furthermore, cell-cell communication networks based on scRNA-seq data predicted BMP2/4-BMPR2 signaling from ECs to SC-β-cells. Correspondingly, BMP4 augmented the SC-β-cell Ca2+ response and insulin secretion. The here-described vascularized SC-islet models will enable further studies of crosstalk between β-cells and ECs and serve as an in vivo-mimicking platform for disease modeling and therapeutic testing.

GEO Accession ID: GSE215376

PMID: No Pubmed ID

Select Condition to compute Visualization and Analysis
Info Preprocessing and downstream analysis were computed using the scanpy Python library and the steps of processing followed the Seurat vignette. Cells and genes with no expression or very low expression were removed from the dataset based on a predefined threshold. The data was then normalized across the expression within the cells and log normalized. The top 2000 highly variable genes were extracted to be used for downstream analysis. For each of these aggregated data matrices, the clusters were computed using the leiden algorithm. Scanpy was then used to compute the PCA, t-SNE, and UMAPs. The points in the plots are labelled by their corresponding cell type labels. The cell type labels were computed using the wilcoxon method as the differential gene expression method. The top 250 genes were then used for enrichment analysis against the CellMarker library in order to determine the most appropriate cell type label with the lowest p-value.
PCA, tSNE, and UMAP Visualizations Labeled by Cell Types:
Gene Expression Data Explorer
Enter gene symbol:

Select conditions below to toggle them from the plot:

Cell Types Cell Samples
Differential Gene Expression Analysis
Info Differential gene expression can be computed for a single cell type labeled group of cells vs the rest. These include wilcoxon, DESeq2, or characteristic direction.
Select Cluster for differential gene expression analysis:
Select differential expression analysis method: