NEW YORK (GenomeWeb News) – Second-generation sequencing techniques can help researchers organize cancer cells into phylogenetic groups that provide insights into the progression of cancer, according to new research.
In a paper appearing online this week in the Proceedings of the National Academy of Sciences, researchers from the UK used the Roche/454 FLX sequencer to look at the evolution of B-cell chronic lymphocytic leukemia (CLL) cells. In the process, the team came up with a new algorithm to weed out real mutations from sequencing errors. Their results suggest that it’s possible to combine sequencing with principles from evolutionary biology to understand the transition from early mutations to dominant tumor cells.
Because abnormal masses of tissues that develop during cancer sometimes contain cells with more than one cellular genotype, the researchers reasoned that these cells may display some traits and relationships normally associated with evolutionary biology. As these competing genotypes develop and evolve, they noted, some subclones will be more evolutionarily fit than others and will eventually take over the population. Understanding this evolution, in turn, could help expose the mutations driving cancer.
“Cancer encapsulates many of the tenets underpinning evolutionary biology,” the authors argued. “[D]irect demonstration of multiple genetically related subclones within a tumor and their phylogenetic relationships has been hampered by the lack of tools for the detection of rare genetic variants.”
To address this, the team used Roche/454 sequencing to identify genetic variants — including rare variants — present in 24 human B-cell CLL tumors from 22 different patients. Using nested PCR, they amplified the Ig heavy chain or IGH locus of tumor cells, a region of the genome that’s often dramatically rearranged during B-cell CLL.
In order to discern real variants from sequencing artifacts, the researchers developed a bioinformatic algorithm to weed out common sequencing errors. Once these artifactual insertions, deletions, and substitutions were accounted for, the researchers were able to identify genuine somatic mutations that were present in as few as one in every 5,000 cells.
Indeed, the samples tested contained anywhere from one dominant mutation to up to 18 different clonal populations with distinct somatic mutations. A quarter of the samples tested contained a dominant clone as well as one or more of these subclones. The sequences of these subclones differed from the dominant clone by as many as 18 bases.
Based on their subsequent experiments and analyses, the team speculated that these subclones “represent genuine stages in the clonal evolution of CLL in these patients.” For instance, they noted, the dominant clone often contains mutated bases that are absent in other subclones, suggesting some subclones may represent steps on an evolutionary path toward the dominant mutation.
Next, they looked at the evolutionary relationships between dominant clones and subclones, generating phylogenetic trees to compare two samples that contained many subclones. Intriguingly, they found three kinds of subclones in these samples. One group of subclones seemed to be intermediates en route to the dominant clones. Another appeared to have diverged on the road to becoming the dominant clone and a third group contained cells that were apparently evolving from dominant clones.
That indicates that some of the earliest cellular mutations contain driver mutations that push cells toward the dominant clonal form, the authors hypothesized. Once a dominant clone — with a selective advantage — establishes itself in the tumor, meanwhile, subsequent mutations are under negative selection pressure and may be less common.
“[F]or the dominant clone to become numerically predominant, it must have a selective advantage over its direct predecessors, suggesting the existence of at least one or more driver mutation in that clone, whether it be in the IGH locus or elsewhere,” the authors wrote. “[T]his finding would be consistent with the hypothesis that somatic hypermutation of the antigen receptor plays an early role in leukemia development.”
Although they noted that the results need to be verified in larger studies, the researchers expressed enthusiasm about the possibility of applying sequencing technology and phylogenetic trees to answer questions about the cellular evolution of cells in other types of cancer as well.