NEW YORK (GenomeWeb) – A new study suggests the accuracy of transcriptional profiles produced from single cells does not seem to suffer significantly when samples are frozen beforehand.
Researchers from Spain and Israel did RNA sequencing on nearly 700 individual cells nabbed from fresh samples and on more than 800 cells that had been frozen in the cryoprotectant dimethyl sulfoxide (DMSO) or liquid nitrogen for up to six months. Their results, appearing in Genome Biology yesterday, suggest freezing does damage a proportion of cells. But the transcriptional profiles discerned from the remaining cells were on par with those generated by single-cell RNA sequencing on fresh samples.
"We can now store patient material along the course of treatment without the need of immediate sample processing," senior author Holger Heyn, a single-cell genomics leader at Barcelona's Institute of Science and Technology's national center of genome analysis (CNAG-CRG), said in a statement.
Heyn noted that the team's hospital collaborators are now starting to archive samples by freezing in the presence of cryoprotectants, "providing us access to samples that were previously out of scope."
While most single-cell RNA-seq approaches involve cells from samples processed straightaway after they are obtained, Heyn and his co-authors explained, the ability to assess RNA transcripts in single cells at a different time or place than where the samples were originally connected "enables complex experimental designs and widens the scope of accessible specimens."
To look at the possibility of profiling transcripts in cells from previously frozen samples, the researchers profiled 1,486 individual cells from a wide range of human, mouse, and dog cell lines and primary tissue samples, relying on Illumina HiSeq 2000 or 2500 instruments in combination with a single-cell RNA-seq protocol called MARS-seq that focuses on 3'-end RNA library preparation.
The team used fluorescence-activated cell sorting to isolate individual cells from four human, mouse, or dog cell lines, peripheral blood samples, and a range of primary tumor samples.
Although cryopreservation was linked to damage affecting roughly one-fifth to more than 60 percent of sorted cells, the researchers reported that single-cell RNA-seq results for 670 individual cells from fresh samples and 816 viable single cells from previously frozen samples produced similar gene numbers, transcript numbers, and cell type-specific expression profiles.
They saw similar consistency between cells from fresh and previously frozen samples when they expanded the analysis to include hundreds more cells from cell lines or primary tissue samples assessed using MARS-seq or a full-length RNA sequencing protocol called Smart-seq2, though the K562 human cell line produced heterogeneous results across both conditions.
For his part, Heyn noted that cryopreservation might be a feasible way for dealing with such batched effects, since it allows for parallel analyses and comparisons of cells from different batches of a frozen sample.
The team confirmed the general consistency between single-cell RNA-seq in fresh and frozen samples by profiling individual cells from a single mouse colon sample that was halved and assessed after fresh preparation or after a week of freezing. It also applied both approaches to mouse and human tumor samples, supporting the notion that frozen samples can yield useful single-cell RNA-seq data.
"The method constitutes a straightforward and powerful tool to broaden the scope of single-cell genomics study designs," the authors concluded, noting that "cryopreservation can be readily implemented into standard single-cell genomics workflows, without modifications of established protocols."