Skip to main content
Premium Trial:

Request an Annual Quote

Inability to Widely Reproduce Gene Fusion Finding Highlights Validation Need

NEW YORK (GenomeWeb) – Weill Cornell Medical College researchers attempting to build upon the discovery by another group of a novel fusion gene in some triple-negative breast cancer patients found instead that that prior work's findings were limited to one patient.

A 2012 Nature paper from researchers at the Broad Institute reported uncovering a novel rearrangement in a portion of triple-negative breast cancer samples that appeared to present a possible therapeutic target. But Cornell researchers, led by Sandra Shin, have reported in Nature this week that they were unable to find this MAGI3-AKT3 fusion in their set of 236 triple-negative breast cancer samples, underscoring the need for such findings to be validated.

"The finding of MAGI3­-AKT3 fusion by respected researchers and published in a top journal offered great hope and a sense of excitement about developing a new therapeutic target for triple negative breast cancer," Shin, an associate professor of pathology and laboratory medicine at Weill Cornell, said in a statement.

Shin and her colleagues developed a tissue-based assay test to search for this fusion gene in their own collection of triple-negative breast cancer samples. Triple-negative breast cancer, they noted, is a highly aggressive disease with few available treatment options and poor survival outcomes

In the 2012 study, researchers led by the Broad's Matthew Meyerson sequenced the whole exomes of 103 human breast cancer samples from patients in Mexico and Vietnam that they compared to matched normal DNA as well as the whole genomes of 22 breast cancer-normal pairs.

As they reported in Nature at the time, they confirmed recurrent somatic mutations, and also homed in on a MAGI3­AKT3 fusion found in a triple-negative breast tumor. They confirmed this fusion gene using PCR followed by sequencing of the index case. Using RT-PCR followed by Sanger sequencing, they reported observing the fusion event in eight of an additional 235 breast cancer samples, including five out of 72 triple-negative samples.

Additionally, they noted that this fusion event appeared to lead to the constitutive activation of AKT kinase, which they said likely could be targeted by AKT inhibitors.

However, the Cornell team realized that the original set of researchers didn't confirm their findings at the genomic level for the additional cases they uncovered.

Shin and her colleagues searched for the MAGI3­-AKT3 rearrangement in 236 samples of triple-negative breast cancer from Cornell using fluorescent in situ hybridization and RT-PCR and in 84 cases from The Cancer Genome Atlas using FusionSeq. For a subset of 28 tumors, they tested multiple regions of the tumors to account for tumor heterogeneity.

They reported no evidence of the fusion in any of the tumors they studied.

Their sample size, the researchers noted, had sufficient power to detect even low, 3 percent frequency, gene rearrangements. They then concluded that the MAGI3­-AKT3 rearrangement is neither recurrent nor sub-clonal in triple-negative breast cancer.

"We clarified that the initially reported mutation was not recurrent and, with the exception of one case, was an artifact," said first author Juan-Miguel Mosquera, an assistant professor of pathology and laboratory medicine at Weill Cornell.

Shin and her colleagues alerted the Broad researchers to their findings two years ago, and also in Nature this week, Meyerson and his colleagues revisited their findings.

After re-analyzing their samples using the hybrid capture array ExomePlus followed by sequencing, Meyerson and his colleagues confirmed the MAGI3-­AKT3 fusion in frozen tissue from the index case, but failed to find it in any of the other 71 original triple-negative breast cancer samples or in an additional 370 tumor samples. They also failed to find the fusion gene in FFPE tissue from the index case.

Based on this, they reported that the fusion event in that index case seems to represent a sub-clonal population, one whose overall prevalence is much lower than they first estimated.

Meyerson and his colleagues said that their initial positive findings are likely due to intra-tumor heterogeneity as well as to contamination with fusion cDNA. They further noted that they originally found the fusion in 4 percent of cases and in no controls, though they only analyzed 12 negative controls.

"In retrospect, any such study, even by a straightforward method such as PCR, would be better powered by using a number of controls equal to the number of experimental samples," the Broad team wrote.

In a statement, the Cornell team said that theses findings underscore the need for validation.

"It is gratifying to know that our results have saved many investigators time, effort, and resources toward a goal that was ultimately not achievable," Shin added. "Just as importantly, patients have been spared from participating in a clinical trial that could not have led to a positive result."