NEW YORK (GenomeWeb News) – In a paper appearing online today in Science, researchers from Harvard Medical School and Helicos BioSciences have shown that certain satellite repeats are highly expressed in cancer but not normal cells — findings that they say may have applications for cancer detection and diagnostics.
The team used the Helicos sequencing platform to perform digital gene expression analyses of several pancreatic, colon, lung, and other cancer samples. In every tumor type tested, the researchers found examples of elevated satellite sequence expression from parts of the genome that are usually transcriptionally silent.
"I think the obvious implication is sort of the diagnostic potential of using satellites as a marker for cancer amongst normal tissue," co-lead author David Ting, a researcher affiliated with Harvard Medical School and the Massachusetts General Hospital, told GenomeWeb Daily News.
The researchers first identified these highly expressed satellite transcripts while working on strategies for sequencing circulating tumor cells, he explained. In the process, they found a discrepancy in the number of RNA transcripts aligning to the known mouse transcriptome when they compared mouse primary tumors with normal mouse tissues.
When the researchers looked more closely at where these transcripts were coming from, Ting said, they found that they corresponded to major satellite repeats in centromeric and telomeric regions of the mouse genome.
And though these major satellite sequences comprise just three percent or so of the mouse genome, Ting noted, the researchers found that about half of all RNA transcript reads in the first mouse pancreatic ductal adenocarcinoma tumor tested corresponded to these sequences.
To explore this in more detail, Ting and his co-workers performed digital gene expression analyses with the Helicos platform for nine more mouse pancreatic ductal adenocarcinoma samples, as well as mouse colon and lung cancer samples.
They also tested samples representing several types of human cancer, including pancreatic, prostate, lung, ovarian, and kidney cancers.
Overall, the researchers found that eight of the nine mouse pancreatic tumors and all 15 human pancreatic tumors tested showed dramatically enhanced satellite transcript expression.
In the mouse pancreatic tumors, for example, transcripts from satellite sequences represented between one percent and 50 percent of all of the transcripts detected in the samples, the researchers reported — a dramatic increase compared to typical mouse tissues.
Although the mouse and human satellite sequences have no sequence homology with one another, Ting explained, the team detected a spike in satellite sequence transcripts in human tumors as well. In particular, he explained, human tumors showed elevated expression of two types of satellite sequences: alpha satellites and HSATII satellites.
"We don't have the numbers to really, truly determine prevalence or sensitivity or specificity," Ting said, "but the data is compelling that this is going to be a pretty common feature in many human tumors."
More work is also needed to understand the underlying biology behind this satellite expression, Ting explained, and to determine whether it contributes to — or is simply a marker of — cancer. But clues so far suggest that satellite expression in cancer cells may be linked to stem cell-like traits, since the satellite sequence expression in mice and humans is typically limited to embryonic tissues and stem cells.
Moreover, Ting said, the team found around 300 so-called satellite-correlated genes that show elevated expression in humans and mice when satellite expression goes up. Of these, around half of the genes identified are known to be expressed in neural tissues, while another 25 percent or so are expressed in stem cells.
Whatever the mechanism, researchers are keen to begin applying their findings in a diagnostic setting, particularly because their experiments suggest pre-cancerous lesions, as well as primary and metastatic tumors, all show enhanced satellite sequence expression. That, in turn, suggests that gauging the expression of these transcripts may help in early detection of cancers.
Consequently, the researchers are now doing follow-up work looking at the potential of using satellite expression as a diagnostic tool — including studies aimed at assessing the potential of detecting such sequences in circulating tumor cells.
"Our hope is that this abnormality will serve as an important biomarker in cancer diagnosis and that it will also shed light on common mechanisms by which cancer develops," senior author Daniel Haber, director of the Massachusetts General Hospital Cancer Center, said in a statement.