By Monica Heger
Researchers at Life Technologies and the Gurdon Institute of Cancer and Developmental Biology at the University of Cambridge have developed a single-cell transcriptome sequencing technique to study gene expression changes as cells develop into embryonic stem cells from inner cell mass.
The technique, which Life Tech is currently developing as a kit, could be used to study the changes that occur at a single-cell level not only in normal development, but also in disease states like cancer or neurodegenerative diseases, the authors reported in a recent study in Cell Stem Cell.
The researchers built on a previous study reported last year in Nature Methods that used the single-cell RNA-seq protocol on embryos. In that study, the cells were about 50 times larger than stem cells, and in the current study, the authors validated the method on cells that contained only around 100 femtograms of mRNA, Kaiqin Lao, principal scientist in molecular biology at Life Technologies and a senior author on the paper told In Sequence.
Lao and his team sequenced the transcriptomes of 33 single cells in different developmental states on Life Technologies' SOLiD machine in order to study gene expression changes as cells from the inner cell mass of mouse blastocysts developed to form pluripotent embryonic stem cells.
The researchers first identified only cells from the inner cell mass that contained the genes necessary to develop into embryonic stem cells. They then analyzed single cells at five different stages during the cells' development to become embryonic stem cells. The goal was to study how gene expression changes as cells develop, and also to examine gene expression differences between cells at the same developmental stage.
The Cambridge researchers used primers developed by Life Technologies to capture the mRNA, convert it to cDNA, and then amplify it. Then researchers at Life Technologies performed the sequencing.
The Life Tech team obtained around 30 million to 40 million reads per cell, which Lao said was sufficient for the analyses in the study. "If you want to look at splicing variants, 20 million reads is too low. Forty million is what you aim for."
In the 12 cells they sequenced at the embryonic stem cell stage, they detected about 65 percent of the known transcriptome, which Lao said is the most likely the extent of gene expression in that early stage of development.
Lao said he is now working on improving the method, and the company will develop a kit to be used for "low-input" transcriptome sequencing, but does not yet have a release date.
First, the Life Tech team will have to optimize the protocol so that it is able to capture the entire length of every transcript. Currently, the primers used to isolate the mRNA are biased toward the 3' end of the transcript, so in the longer transcripts the 5' ends are missing. In this particular study that was not a problem, but for researchers who want to study transcription start sites, the method would have to be improved.
Lao said they were able to capture transcripts up to about 3 kilobases, or around 60 percent of the transcripts. He is now working on bumping the length up to 10 kilobases, by using different enzymes and reagents. He said that would be the biggest technical hurdle to overcome with the technique. "Pushing to 10 kilobases requires that everything go right," he said.
He said that he'd also like to make the method strand-specific and demonstrate it using a 96-well barcoded plate, so RNA from 96 different cells can be pooled and sequenced at once. Those two improvements should not be too difficult, Lao said, noting that there is already a protocol to make the method strand specific, but it just wasn't used in this study.
Using the technique, the researchers were able to show how gene expression changes as cells develop. They also identified two classes of microRNAs that seem to regulate pluripotency, as well as alternative splicing events that appear to play a role in regulating development.
The authors compared the transcriptomes of embryonic stem cells with cells from the inner cell mass and epiblast cells. They found clear differences in the molecular signatures of the cells at the different stages. In particular, they found 2,475 genes with over a 4-fold change in expression between the inner cell mass stage and the embryonic stem cell stage, and also around 2,362 genes with under a .25-fold change. In addition, they found 2,110 genes with a greater than 4-fold change between the epiblast and stem cell stage, and 1,170 genes with under a .25-fold change.
Comparing the expression levels between cells at the same stage, they found that the genes with mid-level expression had higher variability between cells than genes with higher levels of expression.
This finding demonstrates the power of single-cell transcriptome sequencing, said Nick Navin, a postdoctoral researcher at Cold Spring Harbor Laboratory who has used single-cell whole-genome sequencing to study tumor heterogeneity (IS 5/18/2010). "They show clearly that there are a lot of genes whose expression tends to vary within a specific stage from cell to cell," he said. "If they had been looking at populations of cells, any rare cells with different expression levels would not have been seen."
The researchers also found that at the embryonic stem cell stage, genes associated with cell fate were already becoming enriched. "It's showing that there's already some differentiation," said Navin. "Some cells are already starting to commit to their lineages."
He added that one especially interesting finding was that the authors found a large change in the expressed transcript isoforms as the ICM cells developed, including 128 transcript variants that were lost in the embryonic stem cells, and 169 that were not expressed in ICM, but were expressed in ESC.
Azim Surani, professor of physiology and reproduction at the Gurdon Institute and a senior author of the paper, told In Sequence via e-mail that the findings suggest that "the regulation of alternative splicing may be crucial for establishing and regulating transcriptional networks within individual cells during embryonic development."
Finally, the authors identified two classes of miRNAs, one that promotes differentiation and another that represses it. Surani said that the two sets of miRNAs may be working together to maintain the pluripotent state. "These two sets of miRNAs may allow ES cells to respond appropriately to different signaling molecules and culture conditions," he said.
He said his group is now using single-cell RNA-seq to study gene expression dynamics during early germ cell development to help understand epigenetic reprogramming events. He said the technique could also be useful to analyze cell heterogeneity within any cell population, including in the central neural system, hematopoietic and other stem cells, and tumor cells. It could be particularly useful for identifying the existence of cancer stem cells.
Navin also agreed that the technique would be particularly useful for studying cancer, due to its heterogeneity. "Combining this with whole-genome structural changes could be a very powerful method for understanding tumor subpopulations," he said.