NEW YORK (GenomeWeb News) – Researchers from the US and Canada reported online in Nature yesterday that they have successfully used single-cell sequencing to discern copy number patterns in individual cells from primary and metastatic breast tumors.
"Our study demonstrates that we can obtain robust high-resolution copy number profiles by sequencing a single cell and that by examining multiple cells from the same cancer we can make inferences about the evolution and spread of cancer," lead author Nicholas Navin and co-authors wrote.
Navin, who performed the research as a post-doctoral researcher in senior author Michael Wigler's Cold Spring Harbor lab, is currently a genetics researcher at the MD Anderson Cancer Center in Houston. He presented preliminary findings from the study at the Biology of Genomes meeting last spring.
The team used flow sorting combined with whole-genome amplification and massively parallel sequencing to sparsely sequence the genomes of 100 individual cells from a genetically heterogeneous breast cancer sample. Using the same approach, they also sequenced cells from another, more genetically homogeneous, breast tumor and its liver metastasis.
In the process, researchers found copy number patterns pointing to mutational spurts — rather than a gradual snowballing of mutations — in the primary tumors. Patterns in the metastatic sample, meanwhile, indicate that cancers cropping up at secondary sites in the body carry many of the same copy number patterns in the original tumor.
"When we looked at the metastasis, we saw a copy number profile that was highly, highly similar to the primary tumor," Navin told GenomeWeb Daily News. "What that tells us is that the metastatic tumor cells had actually diverged very late in the evolution of the primary tumor."
To find such copy number patterns, the researchers first flow sorted individual nuclei from breast cancer cells. They then did whole-genome amplification of single genomes, followed by massively parallel sequencing with the Illumina GAII to generate random reads covering each genome to about six times, on average.
From there, they got copy number data in the single cells using a variable length binning strategy based on simulations showing the expected number of reads across the genome, Navin explained.
The first primary breast tumor, which was known to be genetically heterogeneous, was diced into a dozen sections based on its anatomical features before the researchers used their single cell sequencing approach to assess 100 cells from six of the sections.
Copy number patterns in this polygenomic tumor revealed three main sub-populations of tumor cells, along with a diploid cell population, Navin noted. But rather than finding copy number patterns consistent with a gradual accumulation of mutations, the team's data pointed to a punctuated model of tumor evolution, he explained, in which sub-populations of tumor cells diverged and expanded without any obvious intermediate cell stages.
"We didn't see the gradual, intermediate cells," he said. "It's not that they're not there, but we think that they're rare or unstable, so you don't see them in the advance tumor."
Using a similar strategy, researchers also evaluated 52 single cells from a second primary tumor that was thought to be fairly genetically homogenous based on array comparative genomic hybridization and other tests, along with 48 single cells from that tumor's liver metastasis.
In contrast to the punctuated evolution patterns detected in both primary tumors, the researchers found a more linear model for the development of the metastasis, with the metastatic tumor carrying most copy number changes found in the corresponding primary tumor.
In addition, the team reported, both primary tumors tested contained so-called "pseudodiploid cells" harboring a host of genetically diverse rearrangements. These rearrangements didn't seem to overlap from one pseudodiploid cell to the next, Navin noted, and did not show up in aneuploid cell populations in the tumor samples.
Although more research is needed to understand these pseudodiploid cells, the researchers speculated that they might represent a population of cancer precursor cells that spur some of the genomic diversity that eventually leads to cancer development.
Those involved in the current study are now tweaking their single cell sequencing method using a barcoding strategy that allows for simultaneously sequencing of multiple cells on a single sequencing lane, Navin noted.
They plan to continue assessing cells from primary and metastatic breast tumors as well as single cells from other cancer types, including colon and prostate cancers. And down the road, Navin said, researchers hope to do complete, high coverage whole-genome sequencing of individual tumor cells — an approach that would complement the existing copy number analyses and provide a more refined view of mutations across the genomes of these cells.