By Monica Heger
Sequencing individual cells from a breast cancer tumor could reveal important insights into the development of cancer, reported Nicholas Navin at last week's Biology of Genomes meeting at Cold Spring Harbor Laboratory.
Navin, a graduate student in Michael Wigler's cancer genomics laboratory at CSHL, presented results from a study that sequenced 100 cells from a basal-like breast cancer tumor to reveal five major subpopulations of cells, three of which were cancerous. The findings support a clonal theory of development in cancer tumors.
"We found multiple clonal subpopulations rather than a series of gradual intermediates," said Navin. "It supports a clonal evolution model for tumor progression."
Navin said sequencing single cells of a tumor could enable the researchers to identify which mutations occur first in cancer development as well as which mutations spur the development of new subpopulations. He said the next steps are to sample tumors from other types of breast cancer as well as tumors from prostate cancer and chronic lymphocytic leukemia. Navin also said that improvements to the method may reveal more subtypes of cancer cells.
In order to isolate DNA from a single cell, the researchers first dissected the tumor into 12 pieces, removing sectors from different geographical regions of the tumor. They then isolated the nuclei, using flow-sorting machines to segregate the nuclei into 96-well plates. Then, using a single-cell whole-genome amplification kit developed by Sigma-Genosys, they amplified the DNA, and sequenced it on Illumina's Genome Analyzer.
They obtained 76-base pair reads using a single-end sequencing strategy and obtained around 12 million reads per cell, about 60 percent of which were mappable, although Navin said that improved throughput is now giving them closer to 20 million reads per cell. He said the reads that could not be mapped were likely adapter sequences from the initial amplification. The experiment cost around $100,000, about $1,000 per sequencing lane, said Navin, not including any labor costs.
The researchers were only able to cover between 5 and 10 percent of the genome, which Navin said was due to the initial amplification step. "You get about a one-million fold expansion, but the problem is the amount of DNA that is initially amplified does not cover the entire genome," Navin said. He said this could be due to chromatin structure, and that the histones may be blocking access to the DNA.
Nevertheless, Navin said the method still gave them interesting results. They found five total subpopulations of cells, three of which were cancerous. Also, each subpopulation was clustered within a specific region of the tumor. One cancer subpopulation was found only in the upper region of the tumor, while the two other cancer subpopulations were found in the lower part of the tumor, including a subpopulation of cells that had highly mutated KRAS genes, known to be involved in cancer. Additionally, the researchers detected mutations in the cancer genes EFNA5 and COL4A5.
Each subpopulation of cells was very clonal, which Navin said suggests that a tumor cell will acquire mutations until it reaches a "perfect state," where it is relatively stable. Then it will multiply, forming clones. Eventually, one of those cells will begin to acquire more mutations, until it reaches a stable state, and then it, too, will form clones. He added that the absence of any intermediate subpopulations supported that theory.
Aside from the three cancerous subpopulations, Navin said the CSHL researchers also found a population of unstable diploid cells that had random amplifications and deletions that were not shared with any of the other subpopulations. "We think they were an early form of precursor cells that didn't go on to form tumor cells," he said.
The next steps, said Navin, are to evaluate other types of primary tumors as well as metastatic tumors. Comparing primary tumors to metastatic tumors may offer additional insight into the progression of cancer, Navin said. He added that the group plans to study tumors from additional breast cancer subtypes, as well as prostate cancer and chronic lymphocytic leukemia.
In addition, Navin said that single-cell sequencing could have applications in other types of diseases such as autism, neurological diseases, and autoimmune disease.