NEW YORK (GenomeWeb) – Understanding the function of rare pancreatic cells, beyond just the cells that produce insulin, may be key to understanding the pathophysiology for type 2 diabetes and lead to new treatment and management strategies, according to researchers from the Jackson Laboratory for Genomic Medicine.
In a study published this month in Genome Research, the Jackson Lab team used single-cell RNA sequencing to study the various cell types that constitute pancreatic islets — the clusters of hormone-producing cells that regulate blood glucose.
Although the majority of cells in islets are glucagon-producing alpha and insulin-producing beta cells, a less common cell type may play a key role in the manifestation of diabetes, the researchers found. These rare cell types, in particular delta cells, may behave more like a "commander or general, while the alpha and beta cells are more the foot soldiers," Michael Stitzel, senior author of the study, told GenomeWeb.
In standard RNA-seq experiments of pancreatic islets, gene expression is dominated by the alpha and beta cells, since they are the most common, Stitzel said. But, with growing evidence that the other cell types such as delta, gamma, and epsilon play important roles in diabetes, Stitzel's team wanted to look at them individually.
On average, around 54 percent of cells in islets are beta, 35 percent are alpha, 11 percent are delta, a few percent are gamma, and less than 1 percent are epsilon.
In the study, the researchers wanted to look at gene expression in the different types of cells in the islets and also look at how it differed between individuals with type 2 diabetes and those unaffected by diabetes.
The team obtained pancreatic islets from eight organ donors — three with diabetes and five without — and used Fluidigm's C1 system to capture 1,050 single cells. Next they constructed libraries using Clontech's SMARTer chemistry and sequenced transcriptomes on Illumina's NextSeq 500.
An initial analysis found that 72 cells had low levels of gene expression and other qualities indicative of cell death, so were removed from further analyses.
To identify specific cell types and verify that they were analyzing expression from single cell, the researchers first looked for unique marker genes specific to each cell and eliminated cells expressing more than one marker gene, since those were likely to be duplicates.
A total of 617 single cells expressed one marker gene. Next, the researchers examined how the cells clustered based on 1,824 highly expressed genes. They found that the cells tended to cluster by type, with unique transcripts for specific cell types even though the cells were from different individuals.
In addition, the researchers found gene expression signatures specific to each cell type. They identified previously reported genes specific to beta cells and alpha cells. However, they also found genes specific to delta cells, including a transcription factor that may govern delta cell identity and the function of which has previously been linked to type 2 diabetes from a genome-wide association study. That transcription factor, HHEX, was expressed exclusively by the delta cells. In addition, the team found that delta cells expressed a gene BCHE, which encodes for an enzyme that breaks down acetylocholine and ghrelin, "providing a mechanism for delta cells to exert local inhibition of islet-influencing endocrine signals," the authors wrote.
The researchers identified a long intergenic non-coding RNA of unknown function that was specific to gamma cells, as well as a gene related to cell adhesion and another gene that encodes an inhibitor of DNA-binding protein. In addition, they identified 30 signature genes related to cell surface proteins.
Stitzel said that some of the delta and gamma cell gene expression findings were surprising because some of the genes they identified were previously thought to be expressed solely in beta cells. "It implicates these other cell type in the genetics of diabetes," he said. For instance, of 1,683 previously reported beta-specific genes, only 115 displayed enriched expression compared to other cell types. For example, the gene HADH is typically associated with beta cells. When that gene is mutated it leads to hypersecretion of insulin. However, in this study, the researchers found that it is also expressed in delta cells.
The group also found interesting findings related to gamma cell expression. Previous studies have found that gamma cells regulate energy homeostasis. In the Genome Research study, the Jackson Lab team found "interesting parallels" in expression between gamma cells and serotonergic neurons, which influence anxiety, mood, sleep, and satiety. They found that gamma cells express FEV/PET1, a serotonergic transcription factor.
Finally, the researchers compared the single-cell transcriptomes from diabetic and non-diabetic individuals, identifying cell type-specific differences. For instance, Stitzel said that expression of insulin was previously thought to be significantly lower in beta cells from individuals with diabetes compared to individuals without. In this study, however, the researchers found that while that was still generally true, the difference in insulin expression was not as large as expected. Rather, he said, it appeared that individuals with diabetes had fewer overall beta cells. This was a finding, he said, that could only be discovered via single-cell sequencing, since bulk sequencing would just pinpoint the amount of expression, as opposed to the number of cells. In addition, he added that this was a very preliminary finding that would need to be validated in additional studies involving more individuals and more cells.
Going forward, Stitzel said that his group planned to do single-cell sequencing in greater numbers of individuals, as well as bulk sequencing of the specific cell types. Stitzel said that separating the cells by type and then sequencing would be a valid approach to look for cell type-specific expression differences, but the researchers decided to do single-cell sequencing in this study to see if they could find evidence of subpopulations within a cell type. He said his team would pursue both approaches in the future.
In addition, he said the team is also continuing to evaluate other single-cell sequencing technology. In this study they used the Fluidigm C1 system, but Stitzel said he is also interested in testing 10X Genomics' platform as well as the Drop-Seq technology developed by Steven McCarroll's group at Harvard Medical School.