NEW YORK (GenomeWeb) – Researchers from the Massachusetts Institute of Technology have developed a high-throughput, yet portable single-cell RNA-sequencing approach that they said could be extended to include other types of analyses.
The method, dubbed Seq-Well, builds on droplet-based approaches like Drop-Seq and inDrop, but instead relies on nanowells to hold single cells, as the researchers reported today in Nature Methods. By using wells and other features, Seq-Well aims to circumvent some of the drawbacks of the droplet approaches such as their resistance to scaling, loading inefficiencies, and differences in lysis timing.
"We wanted to create something that was incredibly easy, simple, efficient, and portable," co-senior investigator Alex Shalek, an assistant professor at MIT, said in an interview.
Baylor College of Medicine's Chenghang Zong, who was not involved in the project, said he liked the approach. Zong has developed single-cell RNA-sequencing protocols, including the multiple annealing and dC-tailing-based quantitative single-cell RNA-seq approach.
"It should be very straightforward and efficient," he said of Seq-Well, adding that it is an alternative to Drop-Seq or inDrop, but with the advantage of high efficiency.
In Drop-Seq, single cells and beads with barcoded oligonucleotides are combined and encapsulated in oil. Within the aqueous bubble, the cell is lysed and the bead captures the freed RNA. Once they are stuck together, the emulsion is broken and the samples are pooled for reverse transcription, PCR amplification, and sequencing library creation.
While Seq-Well follows the same general schematic, it takes a parallel approach, Shalek said, and uses an array of nanowells. Because Drop-Seq and inDrop were more serial in nature, he noted, it could be as long as an hour from when the first cell's contents were lysed to when the last one's contents were, creating a delay.
The nanowells, which Shalek likened to a big ice cube tray, harbor the same barcoded beads as the ones used in Drop-Seq. But because the beads are about the same size as the well, only one bead fits in each well. This boosted their efficiency as each cell that was then loaded had a chance to be captured — in Drop-Seq, a bead or a cell would sometimes wind up alone in a droplet. He estimated that they could capture about 80 percent of cells from low-input samples with hundreds of cells.
After cells and beads are loaded into the wells, the researchers use a semipermeable membrane to cover the wells. Shalek described the membrane as a screen door: it keeps the beads and RNA trapped in the wells, but allows buffers to flow in and out.
After the cell are lysed and the beads have captured the released RNA, the membrane is peeled off, the beads are spun out, and the samples go through RT-PCR, library preparation, and sequencing, just as in Drop-Seq and similar approaches.
But Shalek said that this approach is much more portable and can be used for a greater range of samples, including precious ones.
In their paper, Shalek and his colleagues sought to demonstrate the portability of Seq-Well by bringing their platform into a biosafety level three lab where they profiled gene expression in primary human macrophages exposed to Mycobacterium tuberculosis and in unexposed macrophages. Out of the 40,000 macrophages they loaded, they recovered 14,218 with more than a thousand mapped transcripts. Clustering analysis of nearly 5,000 cells with more than 5,000 transcripts each divvied them into five groups, each of which exhibited shifts in gene expression, especially in gene sets linked to infection.
Because of the approach's higher efficiency, Shalek said, they could now work with samples that people could not have worked with before, like cerebrospinal fluid.
Zong said the use of a semipermeable membrane to cover the wells makes the device easy to adapt for other chemistries.
Still, Seq-Well has some drawbacks, he noted. Just like in Drop-Seq, the samples are pooled for sequencing. This means that if researchers find an interesting cluster of cells, they can't go back to them because they've been mixed together. "I think that's a general challenge to all the methods using barcode-based pooling prior to amplification," he said.
Similarly, the field is limited by sequencing depth, he added, especially as lowly expressed genes are typically the ones researchers are most interested in. "Those lowly expressed genes we sometimes still need to sequence at a reasonable depth," Zong said. "I think that's the question facing Drop-Seq, inDrop, as well as this."
Shalek noted that Seq-Well also allows for imaging, which could enable additional analyses. In one of their experiments, they loaded human peripheral blood mononuclear cells into the wells before the beads for seven-color imaging using a panel of fluorescent antibodies. Clustering of their subsequent sequencing results uncovered subpopulations that corresponded to the major human peripheral blood mononuclear cell types. The proportion of each subpopulation, as gauged by sequencing, also matched the on-array immunophenotyping results.
He added that he and his colleagues are now pursuing different strategies for mapping specific wells on the array to the sequencing results. "I think what you really need is ways of barcoding different positions," Shalek said, so a cell in position A1 in Grid 1, for example, contains a specific barcode.
At the same time, Shalek noted, coauthor Chris Love has developed a number of assays that could be combined with Seq-Well, including, for instance, a glass slide with antibodies to place over the wells rather than the membrane in order to capture what the cell is secreting.
"There are a lot of nice things that you can do if you have the microscope set up, but it's also simple to operate and you don't need to do anything crazy if all you get out at the end of the day is a bunch of single-cell profiles from your samples," Shalek added.