NEW YORK (GenomeWeb) – Scientists at Cellular Research have developed a method for sequencing-based digital gene expression profiling in thousands of single cells without the need for robotics or microfluidic chips.
The approach combines barcoding of individual mRNA molecules from single cells with next-generation sequencing. In an article published online in Science today, company researchers demonstrated the ability of the method – called CytoSeq in the paper and Resolve on the company's website – to detect low-abundance transcripts and rare cells from samples of thousands of individual cells.
Cellular Research plans to launch an instrument, called the Resolve system, in 2016 to perform the sample prep portion of the method, following an early-access program this year.
The instrument, which will be priced competitively to existing technology, will perform the entire sample prep workflow, serving as a front end to a next-gen sequencer. The company currently works mainly with Illumina sequencers but plans to support other NGS platforms, too. "Our goal is to make this common enough and affordable enough so it can simply go in front of every experiment," Stephen Fodor, Cellular Research's co-founder and CEO, told GenomeWeb.
To perform the assay, users will add a cell suspension to single-use cartridges that are preloaded with consumables. These will be processed by the instrument to yield ready-to-go sequencing libraries.
Fodor said the instrument will initially be able to process on the order of 5,000 to 10,000 single cells in parallel, but the technology will scale over time to handle hundreds of thousands of cells simultaneously.
The company claims that its current sample prep consumables costs – less than $1 per cell – are two to three orders of magnitude lower than for commercially available microfluidics approaches. Company representatives declined to comment on details of this calculation.
The Resolve system will likely compete with Fluidigm's C1 Single-Cell Auto Prep system, which uses microfluidic chips to process single cells for genome and transcriptome analysis. At the moment, Fluidigm offers chips, called integrated fluidic circuits or IFCs, that can prepare up to 96 single cells at a time. Other competing methods use standard micotiter plates, Fodor said.
The Resolve system expands the capability of Cellular Research's existing products, which all rely on molecular indexing. A year ago, the company released the Pixel system, which can analyze the expression of a handful of genes in just a few single cells.
Later in 2014, the company launched an early-access program for its Precise assays, which can analyze gene expression of 96 or 384 samples, using standard microwell plates.
The Resolve system, now, will allow researchers "to truly look at large biological diversity," Fodor said.
"One of the main outcomes of this technology may be that it may democratize access to high-throughput single-cell expression profiling, so that it will no longer be a reserve of specialist centers, but instead become a much more widely distributed technology, possibly penetrating into most biomedical research institutes," said Bertie Göttgens, a professor of molecular hematology at Cambridge University. His lab has studied gene expression in large numbers of single cells using Fluidigm's technology, but was not involved in Cellular Research's work. "At this stage it is difficult to predict what exact technology will ultimately dominate the market," he said, "and as with all these techniques, labs who want to take advantage of this will need to have substantial bioinformatics resources to support the data analysis."
The CytoSeq or Resolve method, which the company also refers to as "gene expression cytometry," relies on the basic concepts of molecular indexing and stochastic labeling, first published by Fodor's team in 2011. Unique barcodes attach to individual DNA or RNA molecules to make them distinguishable, enabling researchers to trace back amplified molecules to the original molecule, which eliminates amplification bias.
Cellular Research announced last September that it received a patent related to the technology, US Patent No. 8,835,358, that covers "Digital counting of individual molecules by stochastic attachment of diverse labels."
The company has been working on CytoSeq, which takes the approach to a massively parallel format for single cells, for the last two years or so, Fodor said.
To perform CytoSeq, a suspension of cells is first loaded onto a custom array of microwells, each with a volume of about 20 picoliters. For their Science study, the researchers used arrays with up to 100,000 microwells, but larger arrays could be fabricated. The concentration of the cell suspension is adjusted so that about every tenth well receives a single cell.
Next, a library of 20-micrometer beads, each functionalized with up to hundreds of millions of oligonucleotides, is added so that most wells receive a single bead. Each oligo carries a universal PCR priming site, a molecular index that is the same for all oligos on a bead and identifies the cell, a molecular index that differs for each oligo on a bead and identifies the RNA molecule, and an oligo-dT tail to capture the mRNA.
The bead library is generated using a combinatorial split-pool synthesis method and has almost a million different cell labels and approximately 100,000 different molecular labels.
After the beads are added, a lysis buffer breaks open the cells and releases the mRNAs, which bind to the oligos on the adjacent bead. Following the mRNA capture step, the beads are magnetically removed from the microwells and pooled into a single tube. A reverse transcription reaction converts the mRNA into molecularly barcoded cDNA, which can be amplified and sequenced, either in bulk or after selecting specific target genes.
According to Fodor, the technology will have applications in developmental biology and disease research. "There are a lot of scientific questions to be asked about the role of cellular diversity and individual cells," he said. In terms of disease, the method could be used to study the heterogeneity of tumors or the role of cellular subtypes in immunological diseases, for example.
In their Science paper, the scientists tested CytoSeq's ability to tease out subtypes of cells from increasingly complex mixtures.
They first demonstrated that CytoSeq can sort individual cells from a mixture of two cell populations by sequencing cDNAs from a panel of 12 genes in about 800 cells.
Sequencing cDNAs from a panel of 111 genes in about 1,200 single cells, B cells from a healthy individual with a few lymphoma cells mixed in, they were also able to identify the lymphoma cells.
Increasing the complexity of the mix, they analyzed cDNAs from 98 genes in more than 600 human peripheral blood mononuclear cells, which contained monocytes, natural killer cells, and different subsets of T and B cells, and were able to identify the different cell types based on their gene expression profiles.
In addition, the researchers looked at the response of individual human T cells to an in vitro stimulus by sequencing cDNAs from a panel of 93 genes in about 3,500 stimulated and 1,500 unstimulated cells. They were able to identify two main groups of cytotoxic T cells and T helper cells in the stimulated cells, with additional smaller subsets.
Finally, to identify rare antigen-specific T cells, they analyzed more than 6,000 T cells from two donors that had or had not been exposed to cytomegalovirus. Again, they were able to identify several clusters of T cells, including five cells in one donor, and two cells in the other, that were likely CMV-specific.
Various aspects of CytoSeq can be further improved, according to Christina Fan, a senior scientist at Cellular Research. The company aims to increase the number of cells analyzed per experiment, she said, both by making larger microwell arrays and by increasing the complexity of the combinatorial bead library, and to make the technique more user friendly.
One of CytoSeq's limitations is that it requires cells in suspension, though Fan said methods exist for dissociating cells from tissues, such as tumors, which are already being used for flow cytometry.
Longer term, Cellular Research plans to develop CytoSeq for applications other than mRNA analysis, for example to study other types of RNA, DNA, or RNA and DNA together, Fodor said.