NEW YORK (GenomeWeb) – A team led by researchers at the University of Texas MD Anderson Cancer Center has identified a protein marker that in initial studies appears to detect pancreatic cancer with 100 percent accuracy.
The finding, detailed in a study published today in Nature, offers the potential for better early detection of the disease, which could significantly improve patient outcomes, Raghu Kalluri, an MD Anderson researcher and senior author on the paper, told GenomeWeb.
He noted that while additional studies remain to be done, the marker — the exosome associated protein glypican-1 (GPC1) — has thus far detected pancreatic cancer in a 250-patient cohort with specificity and sensitivity of 100 percent.
"I can't tell you if when we do 5,000 patients the analysis will remain the same, but 250 patients is already a large sample set for a cancer like pancreatic cancer," Kalluri said, adding that he and his colleagues were surprised at the marker's performance.
"It's very rare to see on such a large sample set this level of specificity and sensitivity," he said. "But data is data. We didn't throw out any samples; all the samples analyzed are represented. So there is no statistical bias or anything in the analysis."
The Nature researchers focused their biomarker search on exosomes, secreted, membrane-bound particles that are released by cells and contain molecular content like proteins and nucleic acid from their cell of origin.
(For an example of similar exosome-based biomarker work in preterm birth, see this week's GenomeWeb story on NX Prenatal.)
"It has been known for a long time that pretty much all cell types make exosomes, including cancer cells, and people have been reporting that cancer cells make exosomes at a higher number than normal cells, probably because of all the defects that cancer cells are accumulating," Kalluri said.
He noted that if it were then possible to specifically isolate exosomes from cancer cells, then the protein and nucleic acid information within these exosomes could provide information on a person's cancer status.
Using a mass spec-based proteomic analysis of exosomes from breast cancer cells (MDA-MB-231), fibroblasts (HDF and NIH/3T3), and normal cells (MCF10A and E10), the researchers looked for proteins specific to cancer exosomes that would let them isolate these vesicles. They identified 48 such proteins, ultimately honing in on GPC1, a cell surface proteoglycan that they verified via immunoblotting and IG-TEM was found only in cancer exosomes.
Kalluri and his colleagues began their analysis in breast cancer cells, but as they began screening GPC1 against other cell types they found that, in fact, pancreatic cancer cell lines generated exosomes with higher amounts of GPC1 than breast cancer cell lines.
As they demonstrated in the Nature study, breast cancer can be detected by measuring GPC1-positive exosomes, but not with the accuracy the marker demonstrated in pancreatic cancer.
"So we diverted out attention to look at pancreatic cancer," Kalluri said, noting that larger-scale studies in breast cancer that took into account different subtypes and stages of the disease might find particular forms of the cancer where GPC1 performs particularly well.
In pancreatic cancer, the researchers found that GPC1-positive exosomes could with 100 percent accuracy distinguish between patients with stage I-IV pancreatic cancer from healthy donors and patients with benign pancreatic diseases. They also found that GPC1-positive exosome levels correlated with metastatic disease burden, with levels higher in patients with distant metastatic disease than in patients with metastases restricted to the lymph nodes and patients with no metastases. They also found a decrease in patient GPC1 levels after surgical resection.
Additionally, using mouse models, they found GPC1 levels correlated with cancer growth and, in fact, enabled the detection of cancer before the appearance of MRI detectable masses.
The findings, Kalluri said, suggest that GPC1-positive exosomes could provide a significant improvement over existing pancreatic cancer biomarkers like the protein carbohydrate antigen 19-9 (CA19-9), which is less sensitive and also struggles to distinguish between cancer and benign pancreatic disease. Pancreatic cancer has poor outcomes in large part because the disease is often detected only in its late stages. A marker like GPC1 capable of accurate early detection could improve patient survival by allowing for surgery before the disease has spread, Kalluri said.
Beyond that, the GPC1-positive exosomes contain a wealth of other molecular information that could prove useful in treating the disease, he noted. For instance, in the Nature paper, the researchers found that they were able to match KRAS driver mutations present in these exosomes to KRAS mutations present in the patient tumor itself.
"This is not a biomarker that ends its utility once you detect [the cancer]," Kalluri said. "Once you detect [the GPC1-positive exosome], you can isolate it. And now the exosome has nucleic acids, proteins, microRNA, and all of those things that will help us learn what stage the cancer is and what mutations it has collected. So we think that the detection of cancer exosomes is a beginning of the value of the biomarker."
He said that he and his colleagues are planning a larger study for the marker and also hope that publication of their findings will inspire other groups to take up further validation efforts including studies comparing it head-to-head with CA19-9.
Kalluri predicted that clinicians would be able to begin using the marker in the relatively near future. "I don't think it will take years to get this into practice," he said, adding that he and his colleagues are with MD Anderson exploring options for commercial development of the marker, but they have no firms plans yet.