Parsing the Pancreas – University of Copenhagen

DanStem > News and Events > Parsing the Pancreas

14 March 2017

Parsing the Pancreas


In this commentary, published in New England Journal of Medicine, DanStem professors Anne Grapin-Botton and Palle Serup cover recent results obtained with new single-cell RNA-Seq measurements of mouse and human pancreas cells.

The most widely used single-cell transcriptomics assay, adapted from an assay that analyzes cell populations, is RNA-seq (RNA sequencing); RNA-seq measures global gene expression by reverse transcribing RNA into complementary DNA and then sequencing it. This strategy has been used to compare gene expression between samples, for example between a tissue from a diseased patient or a healthy individual. However, if different cells have different gene expression changes, this is not seen with such a strategy.In the single-cell RNA-seq protocol, individual cells are typically captured in microfluidic devices or by means of flow cytometry. This strategy uncovers changes between different cells from an organ and can compare them individually to the cell profiles of a diseased organ. 

Tracking Down the Transcriptome of Pancreatic Cells.

Islet preparations from donor pancreata generally contain substantial numbers of non-islet cells that allow for a sampling of all pancreatic cell types after their capture by flow-activated cell sorting or in microfluidic devices. After complementary DNA synthesis, amplification, and next-generation sequencing, the transcriptomics data are analyzed by customized computer algorithms to gauge similarities among cells and cluster them into cell types, as well as to identify potential cell subgroups and transition states. t-Distributed stochastic neighbor embedding (tSNE) is one such algorithm, by which high-dimensional objects are modeled as two-dimensional or three-dimensional points, such that similar objects are represented by nearby points and dissimilar objects by distant points. These methods have allowed the identification of new markers of rare pancreatic cell types and subpopulations of ductal, acinar, and beta cells, as well as the identification of differential expression of many genes between beta cells from healthy persons and those with type 2 diabetes.

Analyses of single-cell RNA-seq data are substantially more challenging than analyses of data from similar experiments on cell populations, which have been facilitated by the development of analysis and quality-control pipelines that make use of customized computer algorithms. Notably, as described in four recent papers, this technique has been used to analyze large numbers of single cells from human islet preparations, which by their very nature contain lots of contaminating non-islet cells, allowing a sampling of all pancreatic cell types. These analyses identified all the known cell types, including rare cells such as the ghrelin-producing epsilon-cell. In addition, new subgroups among beta cells and duct cells were identified. “Importantly, the profile of some cells was found to change in type 2 diabetes patients and these could prove important for future disease studies”, says Palle Serup.

Grapin-Botton, Anne & Palle Serup (2017). Parsing the Pancreas. The New England Journal of Medicine, 376(9), 886-888, doi:10.1056/NEJMcibr1616217