It’s not often that Nature publishes on a new way of detecting cancer. Recently the journal reported that bits of cancerous cells – tiny blebs containing protein, RNA and DNA – can be measured in blood samples from patients with pancreatic cancer. These particles, exosomes, might serve as tumor indicators, or biomarkers.
The investigators purified and marked high levels of exosomes from blood of pancreatic cancer patients. Amazingly, hundreds of billions of these microscopic, virus-sized sacs were obtained from less than a teaspoon of blood, using standard laboratory methods. These exosomes could be checked for cancer-related proteins and genetic abnormalities. Such a biomarker would be very useful in managing and possibly detecting pancreatic cancer.
The bulk of the studies were accomplished at MD Anderson Cancer Center in Houston, led by Raghu Kalluri MD, PhD, a Professor and Chair of the Department of Cancer Biology, along with collaborators in Spain and Portugal. The newly-published findings have not been confirmed by others; the results need be reproduced by other research groups.
“I think it’s a really exciting development,” said Dr. David Linehan, a professor and surgical oncologist who heads the Department of Surgery at the University of Rochester Medical Center. He specializes in gastrointestinal tumors, including pancreatic cancer and was not involved with the Nature report.
“As a surgeon, my frustration is that only about 15% of pancreatic cancer patients can have surgery with curative intent. Typically by the time it’s diagnosed, it’s already spread,” he said. “Making more patients candidates for curative surgery would be a home run.”
This year, approximately 49,000 will receive a diagnosis of pancreatic cancer, and over 40,000 will die from it. It’s generally a tumor that affects older adults; it affects men slightly more than women; half of patients are under age 71 years. As things stand, only 7.2% of patients with pancreatic cancer are alive 5 years after diagnosis, according to the NCI.
Although pancreatic cancer is not common, listed by the NIH as the 12th most common malignancy and accounting for only 3% of tumors, it’s the fourth most common causes of cancer deaths in the United States. Unlike some cancer forms, like breast, colon, prostate, stomach cancer, most kinds of leukemia and lymphoma, and even lung cancer, for which U.S. death rates have been declining – and progress has been evident – death rates from pancreatic cancer have been going up. This comparative lack of progress is due to a persistent absence of effective treatments beyond surgery, which is rarely curative, combined with an increase in cases.
“Inability to screen has limited progress in this tumor type,” Linehan considered. “For colon cancer, we can find it by colonoscopy, and for breast cancer, with mammography,” he said. “Because pancreatic cancer is found late in most cases, it usually causes death.”
The key finding – why this paper is important – is what the researchers found in blood samples from humans with and without cancer: high levels of the marker (circulating exosomes with GPC1) in all serum checked from 190 people with pancreatic cancer (190 of 190 samples), compared with healthy donor samples (none, 0 of 100 samples, met the threshold). These differences were highly significant from a statistical perspective and clearly distinguishable, i.e. separable with a wide gap between negative and positive results, by the appearance and protein-containing patterns of the microscopic particles, which were generally between 100 and 110 nanometers wide. In other words, the differences weren’t subtle.
The main problem with the study is that the biomarker also “lit up” in blood from patients with breast cancer, in 24 of 32 serum samples tested. But even if the method stands as reported, if confirmed in other labs, it would facilitate molecular studies of pancreatic tumors from blood as has never been done previously. It could be a game-changer (or better put, a life-changer), enabling doctors to easily detect genetic abnormalities in pancreatic cancer that could be treated with new drugs, and which could be monitored without invasive procedures or repeat biopsies.
For pancreatic cancer patients after surgery, the experimental marker predicted prognosis: how long they lived before recurrence of malignancy, and survival. If reproduced, and if the exosome analysis were adapted to distinguish pancreatic cancer markers from those of other cancers, it could enable screening and improve cure rates for a notoriously lethal form of malignancy.
“What we are proposing is that this is a reliable biomarker for pancreatic cancer,” Kalluri said in a phone interview. He, the senior author on the Nature study, stated he has no relevant conflicts of interest relevant to this work at this time, he said. “We hope others will reproduce these findings.” There is no specific plan or company involved in scaling up or commercializing this work, Kalluri confirmed by email. “MD Anderson Cancer Center owns the intellectual property surrounding this discovery and is currently looking for business partners and participating in discussions for commercialization,” he added.
“It’s not that difficult at all to isolate these particles from patients’ blood,” Kalluri told me. “As technologies go, this is a fairly simple,” he emphasized. “You can use an ultracentrifuge or purchase a kit to purify exosomes from blood.” The samples were analyzed with standard equipment, including an electron microscope and flow cytometry machines, and evaluated using commercially-available antibodies. The specimens don’t have to be fresh, he indicated. Some of the serum samples they’ve tested for exosomes have been frozen for as long as 30 years.
“At the moment, we think its value is in monitoring,” he said. “We can follow exosomes in patients after treatment, looking at relapse after surgery, and drug resistance,” he said. “Depending how much someone drops the exosome level after surgery, it could predict survival.” The published findings suggest this biomarker is a more accurate predictor than CA19-9, a test used in current clinical practice. “The other aspect of the study with exosomes is that you can look at some of the mutations,” he said. “The mutations, like in RAS, can provide a signature for certain cancers.”
“For screening, that will take some work,” Kalluri said. “We don’t know if the breast cancer samples were positive because of the subtype of cancer. We need to analyze more samples. We may need to modify our system,” he added. “We don’t have anything at the moment to make it pancreatic specific.” The group did not look at other markers on the cancer exosome surface, he said. “We need to do further studies.”
“The idea of isolating exosomes from cancer cells has been around for a while,” Kalluri said. His lab has been working on this project for about four years. In previous studies, his group has demonstrated that nucleic acids in the malignant cell bits, the exosomes from cancer cells, can transfer genetic information to nonmalignant, recipient cells and change them, causing them to proliferate or bind malignant cells, among other phenomena. This kind of exosome movement involving RNA delivery may play a role in how cancer spreads, and was reviewed recently in the NEJM.
“There’s tremendous interest in exosomes,” Kalluri said. “It’s very hard to get a sense of what internal organ cells are doing through a blood evaluation,” he said. “Exosomes shed by internal tumors give us an idea of genetic and other changes the cells are undergoing. That is the potential advantage.”
The exosome studies represent one of several approaches to liquid biopsy, a burgeoning area of pathology being tried for many solid tumor types. For internal cancers like that of the pancreas, which are physically hard to sample, liquid biopsies can be especially helpful, because they provide information about the malignant cells – how they may be changing, such as acquiring new mutations or becoming resistant to a drug – without subjecting the patient to the risks and costs of an invasive procedure.
Last month, Exact Sciences announced a plan in partnership with the Mayo Clinic, discussed here, to analyze DNA obtained from “pancreatic juice,” to develop a screening tool for pancreatic cancer. But to get at those pancreatic secretions, the “juice,” doctors would have to perform endoscopy; they’d have to collect the sample by a tube inserted down the esophagus, through the stomach and into the near part of the small intestine, the duodenum. While that may be a fairly routine procedure for gastroenterologists, it’s not without risk or cost, and it’s a lot harder than just taking a blood sample. Which is why the exosome method may be so valuable.
Other methods of liquid biopsy, for the most part still investigational, include measuring circulating tumor cells (CTCs, about which I wrote in 2009, in assaying lung cancer) and studies of circulating, cell-free tumor DNA (cfDNA). “There is a possibility that cfDNA could be of great value,” Kalluri indicated. But it also comes from normal cells. “After we isolate cancer exosomes, the genetic material is more representative of the cancer-associated changes,” he said.
“For circulating tumor cells, you need fresh blood,” Kalluri said. “They’re difficult to isolate, you get small numbers, whereas in one milliliter of blood there are 100 to 500 billion exosomes,” he added.
“I think the most important thing that would come of this work is a way that we could diagnose pancreatic cancers when they’re operable,” Linehan said. “If a blood test were validated and became something routine that could be checked at an annual checkup, then we could pick up cases early and do surgery for cure in 50% or 75% of cases,” he said. “That would change everything.”
“Most people who get pancreatic cancer have no known risk factor,” Linehan said. So what’s needed is a very specific test. Otherwise if you screen most people for pancreatic cancer, the test would generate too many false positives, he considered.
“Obviously, prevention would be better. But early detection of pancreatic cancer is absolutely a critical, unmet need,” Linehan said. “The challenge is developing a screening test that’s reproducible, inexpensive and accurate,” he emphasized. “If this test is truly what it’s said to be, that would be very helpful to many patients.”
If the new test only proves useful in patients who have known pancreatic cancer after treatment, that would be less valuable, Linehan said. “If that’s all it can do, then it’s less exciting. The blockbuster we need is an early detection test,” he said. “After surgery, we already have another test, CA19-9, which is good for follow-up.”
“Most pancreatic tumors have RAS mutations,” Linehan said. “But you can find those in non-malignant cells, too.” If researchers could match genetic changes found in the exosomes to patients’ tumors, that would be helpful. “We aren’t to the point in pancreatic cancer that we’re treating patients for actionable mutations.” But it will happen, he said. “Hopefully and ultimately, if you could sequence individual tumors and monitor using blood, that would be great.”
There’s a lot of science in this report. But as liquid biopsies enter clinical practice, this is the kind of thing a patient in the not-so-far-off future might want to know, and ask her doctor about, I think it worth reviewing:
The investigators began by comparing the tiny sacs, exosomes, extruded by cancerous and non-cancerous cell types they could grow in the lab. After identifying 48 proteins particular to the cancer exosomes, they focused on one protein called glypican-1 (GPC1). This molecule was found at high levels in exosomes in fluid exposed to cancerous cells (from pancreatic cancer and also breast cancer types laboratory cells), and was uniformly low in exosomes in fluid exposed to cells that don’t cause cancer. The researchers also did experiments in mice. (The gist of those results is that mice implanted with no or smaller vs. larger tumors had proportionate levels of the blood biomarker, exosomes with GPC1).
To confirm their initial findings, the investigators examined additional blood samples to see if they could prospectively distinguish those from people with and without pancreatic cancer. As reported, all 20 healthy donors had low levels of the GPC1-containing exosomes, as did people with non-malignant, inflammatory conditions affecting the pancreas (18 with pancreatitis) and other abnormalities (8 with cystic adenomas). Five samples were taken from people with premalignant pancreatic cancer precursor lesions, and those 5 yielded intermediate levels of the biomarker. The researchers tested another 56 samples from patients with the most common form of pancreatic cancer, adenocarcinoma, and all 56 had high biomarker levels.
While it needs be confirmed, this was neither a trivial nor small study: the investigators tested blood samples from 250 people, humans with pancreatic cancer.
Finally, my critique: The paper’s abstract refers to finding particles in serum of pancreatic cancer patients “with absolute specificity and sensitivity, distinguishing healthy subjects and patients with a benign pancreatic disease from patients with early- and late-stage pancreatic cancer.” This sounds almost too good to be true. And while this it technically correct as stated, it’s only true in the limited context of evaluating samples from patients with known pancreatic disease and healthy controls. The summary skips mentioning that the biomarker assay was positive, also, in samples from breast cancer patients. A casual reader might think the test is absolutely specific for pancreatic cancer, which it would not be in the context of screening.
The investigators didn’t test exosomes from blood of patients with other malignancies. Based on what’s reported and known about the GPC1+ exosomes, it could be that they’d be positive in blood from patients with other kinds of cancer. Nonetheless, if you could simply do a few additional stains on the exosomes – using standard lab methods, with a panel of antibodies – and add genetic studies to boot, it seems like you could develop a pancreatic cancer-specific assay. This is not science fiction. (Far from it.)
And maybe there’s too much jargon here. But that’s part of the point, that this is (or should be) routine stuff in 2015; scientists have been using these kinds of methods, equipment and research tools for decades. You might wonder, as I did, why didn’t we study cancer exosomes 20 years ago?
Also on Forbes:
To receive my articles, follow me here at Forbes. I’m on Twitter @ElaineSchattner. For my public updates on Facebook hit the “follow” button.