NEW YORK (GenomeWeb News) – Researchers from Johns Hopkins Kimmel Cancer Center and elsewhere have demonstrated that they can directly detect cancer-related chromosomal alterations in patient blood samples by sequencing cell-free DNA without prior knowledge of alterations present in the actual tumor.
"Our new approach identifies both rearrangements and chromosomal arm alterations directly in patient plasma with no previous interrogation of tumor DNA," Rebecca Leary, a post-doctoral researcher at the Johns Hopkins Kimmel Cancer Center's Ludwig Center for Cancer Genetics and Therapeutics, told GenomeWeb Daily News.
In a proof-of-principle study, appearing online today in Science Translational Medicine, Leary and her colleagues described the sequencing and analytical approaches they used to assess blood samples from 10 individuals with colon or breast cancers and as many unaffected individuals.
The approach did not pick up any obvious chromosomal rearrangements or copy-number changes in samples from the healthy controls. In contrast, though, it did unearth chromosomal alterations in circulating cell-free DNA from all of the individuals with cancer. Those alterations included chromosome-level rearrangements, losses and gains of chromosome arms, and, in some cases, amplifications affecting suspected cancer drivers.
"The sensitivity and specificity of this approach are dependent on the amount of sequence data obtained," the study's authors wrote, "and are derived from the fact that most cancers harbor multiple chromosomal alterations, each of which is unlikely to be present in normal cells."
For the time being, Leary noted that the level of sequence coverage needed to see very early-stage cancers may preclude its use as an early diagnostic test, though that may change as sequencing costs continue to decline.
The group is now keen to try out the blood-based sequencing approach in more cancer patients to determine how well suited it is for finding new cancers, guiding cancer treatment, and/or tracking patient progression, relapse, or recurrence after treatment.
"We're very excited about the next step," Leary said. "We'll need to bring this into larger clinical trials to really figure out the best application."
In a study published in Science Translational Medicine in 2010, members of the same Johns Hopkins team showed that it was possible to detect chromosomal rearrangements in circulating tumor DNA from individuals with breast or colorectal cancers using a method dubbed personalized analysis of rearranged ends, or PARE.
For that work, though, researchers first needed access to samples and sequences from the tumor itself, making it better suited as a strategy for tracking tumor treatment or recurrence than for detecting and characterizing tumors and their potential vulnerabilities.
With their latest approach, the investigators have come up with ways to not only find chromosomal rearrangements in circulating tumor DNA, but also to uncover chromosome gains and losses — all in the absence of sequence data from the original tumor.
"Our previous work, including the PARE approach, was based on identifying alterations in tumor DNA and developing PCR assays to detect those alterations in patient plasma," Leary said. "These studies showed that we could detect patient-specific alterations in cell-free plasma DNA, and we wondered if we could find these alterations by directly sequencing patient plasma."
To explore that possibility in their new study, researchers started with blood samples from three individuals diagnosed with late-stage breast cancer, seven individuals with late-stage colorectal cancer, and 10 unaffected control individuals.
After isolating the total cell-free DNA from each sample, they then performed massively parallel paired-end sequencing on this DNA, generating one lane of Illumina HiSeq 2000 sequence data per sample. These reads were subsequently mapped to the human reference genome, revealing potential somatic rearrangements in the form of paired reads that failed to map appropriately to the reference.
By applying stringent filtering criteria to this data, the researchers moved from millions of oddly mapped reads down to just 14 chromosomal rearrangement candidates. All 14 of these apparent rearrangements appeared in patient blood samples, with rearrangement candidates turning up in plasma from nine of the 10 patients.
Sequencing-based analyses of matched tumor and normal samples verified that the chromosomal changes picked up by the blood-based method were also present in corresponding tumor samples, but not in patients' unaffected tissues.
Meanwhile, for their copy number analyses, the team profiled chromosome gains and losses using an approach reminiscent of the digital karyotyping that co-senior author Victor Velculescu, co-director of the Johns Hopkins cancer biology program, and his colleagues described nearly a decade ago in the Proceedings of the National Academy of Sciences.
"The sequencing libraries generate paired-end tags," Leary said, "but you can also look at single tags — so one tag from each fragment — and use those tags in a method similar to digital karyotyping to count the number of tags mapping to any region in the genome."
For copy number analyses in the current study, researchers focused on gains and losses affecting whole chromosome arms, using a plasma aneuploidy score developed with existing SNP chip data on three dozen colorectal cancers and 45 breast cancers.
In all 10 patients, sequence data generated from cell-free plasma DNA pointed to the presence of chromosome arms gains and losses. None of the healthy controls had copy-number profiles that met the team's threshold for aneuploidy.
Such rearrangements and copy-number shifts seem to be features shared by other cancer types, too, Leary said. In additional, unpublished work, she said, the team has been finding telltale signs of chromosomal instability in virtually every tumor type tested.
And though the cell-free plasma DNA sequencing approach described in the current study does not offer the sequence-level resolution required to see small mutations such as insertions and deletions, findings from the study indicate that the approach can pick up amplifications involving apparent cancer-driver genes. For instance, the team uncovered an amplification involving ERBB2 in blood from one of the colorectal cancer patients and a CDK6 amplification in samples from one of breast cancer patients.