"These findings are very exciting," said Jain. "The biomarker score provides predictive and prognostic information separate from and independent of clinical and pathologic tumor characteristics that oncologists have available today and which often provide only limited clinical value."

Hunting for new biomarkers


The study authors focused on genes regulating the function of centromeres and kinetochores - the essential sites on chromosomes that spindle fibers attach to during cell division - based upon results from earlier research by the Karpen group and other labs in the field. In normal cell division, microtubule spindles latch on to the kinetochores, pulling the chromosome's two chromatids apart.

What the Karpen team previously found in fruit flies is that the overexpression of a specific centromere protein resulted in extra spindle attachment sites on the chromosomes.

"This essentially makes new centromeres functional at more than one place on the chromosome, and this is a huge problem because the spindle tries to connect to all the sites," said Karpen. "If you have two or more of these sites on the chromosome, the spindles are pulling in too many directions, and you end up breaking the chromosome during cell division. So overexpression of these genes may be a major contributing factor to chromosomal instability, which is a hallmark of all cancers."

This chromosomal instability has long been recognized as a characteristic of cancer, but its cause has remained unclear.

To determine if centromeres play a role in chromosome instability in human cancers, the researchers analyzed many public datasets from the National Center for Biotechnology Information, the Broad Institute and other organizations that together contained thousands of human clinical tumor samples from at least a dozen types of cancers. The researchers screened 31 genes involved in regulating centromere and kinetochore function to find the 14 that were consistently overexpressed in cancer tissue.

The extensive records included information on DNA mutations and chromosome rearrangements, the presence and levels of specific proteins, the stage of tumor growth at the time the patient was diagnosed, treatments given, and patient status in the years following diagnosis and treatment. This allowed the researchers to correlate the centromere and kinetochore gene expression score (CES) with patient outcomes either with or without treatments.

Medical Research Homepage