NEW YORK (GenomeWeb) – Tumor genomes are not static and clonal mutations expand and retract in response to therapy — a process that can be monitored via a liquid biopsy, according to researchers from the University of Torino in Italy.
Reporting their results this week in Nature Medicine, the group used next-generation sequencing and droplet digital PCR blood-based assays to demonstrate that when patients acquire resistance mutations to a drug, those mutations may not remain permanently. Several weeks after stopping a targeted EGFR inhibitor, RAS mutations that metastatic colorectal cancer patients had acquired could no longer be seen in the blood and the drug once again became effective.
"When the clinician stops therapy because the patient is resistant, [mutant] KRAS disappears relatively fast from the blood," Alberto Bardelli, senior author of the paper and associate professor of molecular oncology, told GenomeWeb. "You can re-challenge the patient with another round of [EGFR inhibitor] therapy and the patient re-responds."
Bardelli stressed that so far the theory has only been tested on a small number of patients, and thus would have to be confirmed in much larger cohorts.
Going forward, he said the goal is to use the ctDNA NGS assay as a discovery tool to identify resistance mechanisms and then to use a droplet digital PCR assay to monitor those mutations within the context of a clinical trial.
In the Nature Medicine study, the researchers examined colorectal cancers. Genotyping is routinely done on patients with colorectal cancer because there are several targeted therapies available. However, tissue can often be difficult to obtain and does not allow for the monitoring of disease, so the researchers sought to develop methods of analyzing blood.
Bardelli said that the first step was to confirm that mutations in the blood correlated with mutations in the tumor. He said that because the goal of the study was to look at drug resistance, they took a targeted approach, focusing on the RAS family of genes. Mutations in RAS family genes confer resistance to EGFR inhibitors, which are approved to treat metastatic colorectal cancers in RAS wild-type patients.
They confirmed that mutations found in the blood correlated with mutations found in the tissue and also identified mutations in the blood of eight cases that were not seen in the tissue, which "suggests tumor heterogeneity," Bardelli said. "Looking in the blood we see more complexity."
Next, they wanted to see whether they could identify the primary mechanism of resistance in the blood of metastatic colorectal patients that had been given an EGFR targeted therapy but had not achieved clinical benefit. Even when EGFR inhibitors are given only to patients with wildtype RAS genes, the drugs are still only effective in 20 percent of patients.
To look for other resistance mechanisms, the researchers designed a 226-gene NGS panel. They analyzed ctDNA from eight patients that had developed resistance to EGFR inhibitors but were confirmed to have wildtype RAS genes.
As expected, the ctDNA assay did not find mutations to the RAS genes.
Four cases had an ERBB2 amplification, which has previously been associated with EGFR inhibitor resistance; one case had a FLT3 amplification; and one case had a MAP2K1 mutation that encoded the MEK1 protein, which has not yet been associated with resistance to EGFR inhibitors.
Finally, the group tested to see if the ctDNA assay could identify acquired resistance mechanisms in 16 patients that initially responded to the EGFR inhibitors cetuximab or panitumumab, but eventually developed resistance and stopped treatment.
In 11 of the patients they were able to detect KRAS mutations in blood samples that had been obtained at disease progression. In two cases they also found EGFR mutations, and in four cases they found KRAS or MET amplifications — all of which have been linked to resistance to EGFR inhibitors in colorectal cancer.
They continued to monitor the KRAS mutations and MET amplification using the ddPCR assay in six patients after stopping the EGFR inhibitor and switching to different treatment regimens.
Somewhat surprisingly, the researchers found that in all six patients, the percentage of mutated KRAS alleles that had emerged after being treated with EGFR inhibitors declined after the treatment was stopped. "KRAS disappears relatively fast," Bardelli said, in about three to four weeks. "That's kind of exciting in a way because you can start to ask whether the patient may then be prone to respond again to the therapy."
The researchers confirmed in cell lines that indeed the percentage of KRAS mutated alleles dropped in the absence of the drug.
The team next searched their clinical database to see whether any patient had undergone two rounds of cetuximab. They found two colorectal cancer patients that had initially been placed on the drug and responded, but stopped treatment after developing resistance and received additional lines of therapy, but then were placed back on cetuximab and again responded.
The researchers also identified three patients who had received multiple rounds of either cetuximab or panitumumab and for whom longitudinal tissue and blood samples were available. For all three patients, the proportion of KRAS-mutated alleles increased and decreased in the presence and absence of the anti EGFR drug.
The study provides "molecular evidence that re-challenge with EGFR blockade after a withdrawal period is effective, as it exerts genomic selection in metastatic [colorectal cancers]," the authors wrote.
Bardelli said that the results would have to be verified in larger numbers of patients, but are nonetheless promising. Patients with metastatic colorectal cancer that develop resistance to EGFR inhibitors often have few good drug options available, Bardelli said, and revert back to standard chemotherapy. This study suggests that resistance mutations are dynamic, so re-challenging patients with a previously effective drug after a "holiday" may provide clinical benefit and help slow disease progression.
However, Bardelli said it is unclear how durable the response would be to drugs that are recycled in this way. "Based on clinical analysis, we'd anticipate that the second response would be shorter," but that still needs to be verified in larger studies, he said.
Aside from colorectal cancers, Bardelli said that such a model of rising and falling resistance mutations may be relevant for other cancer types.
Going forward, he said that his group is using the NGS ctDNA assay to identify additional mechanisms of resistance and in clinical trials based on patients' molecular profile, as well as ddPCR to then monitor those patients once resistance mutations are identified.