OSU researchers recently published data that is expected to yield a new test based on blood-borne miRNAs for lung cancer progression and prognosis.
In a study published earlier this spring in The Proceedings of the National Academy of Sciences, Ohio State University's Carlo Croce and colleagues investigated miRNA expression profiles in tissue and plasma samples from a completed spiral-computed tomography screening trial, identifying a set of signatures in blood that could predict cancer much earlier than CT alone and distinguish more aggressive forms of the disease.
In the study, the team aimed to determine whether a difference could be seen between a patient with cancer and a patient without cancer, looking not at the tumor tissue but looking at the plasma," Croce told Gene Silencing News this week. "What we found in the paper is that we can see differences in miRNA expression in plasma in patients with cancer versus patients without cancer."
Using the markers, "we could detect a malignant lung cancer before it was detectable by any other means, which is very important, because if you can detect cancer by non-invasive methods before the cancer is seen, you have tremendous opportunity to cure it," he said.
MicroRNAs have been widely studied as diagnostic biomarkers for lung cancer and other pulmonary diseases. And while currently commercialized miRNA tests use tissue samples, the relative ease and safety of blood testing has spurred companies such as Rosetta Genomics and Asuragen to develop blood-based miRNA diagnostics.
"Cancer tissue you can retrieve... only by biopsy or by cutting it off, so it's very invasive," Croce said. "With plasma, you have to just draw blood."
In the study, Croce and his colleagues examined tissue and serum samples from two large randomized spiral-CT trials, the first a training set used to identify the miRNA expression profiles and the second an independent validation set from another randomized spiral-CT trial. The patients included 1,035 individuals aged 50 or older who were pack-a-day smokers for at least 20 years and underwent CT scanning once a year for five years. They had also provided blood samples throughout the screening process, which, Croce said, permitted the examination of their miRNA profiles long before CT screening found their cancer.
The researchers first analyzed 28 tumor tissue samples and 24 samples of normal lung tissue to identify the miRNA profiles that were predictive of cancer presence and prognosis. Then they performed high-throughput miRNA expression profiling of blood samples collected in the trial using TaqMan microfluidic cards from Applied Biosystems. In blood serum, they found a series of miRNA signatures that were either predictive of cancer or associated with prognosis.
The researchers found 16 ratios among 15 miRNAs that could predict cancer, identifying 18 of 20 individuals who were later diagnosed with cancer after spiral-CT with 90 percent sensitivity. Verified in the validation cohort, the same signature correctly identified 12 of 15 samples from patients with 80 percent sensitivity.
For prognosis, they found a signature that identified five out of five patients with poor prognosis in the training study and four of five in the validation group with 100 percent and 80 percent specificity, respectively, as well as a signature that identified five of 15 patients with good prognosis in the first set and five of 11 in the second.
Croce said there are still many questions about what the miRNAs found in blood represent — whether they have their own function, or are just degradation products of apoptosis. "There are a lot of investigations [into whether they are functional or not], but at the present time we don't know much.”
The study's identification of blood-based miRNA signatures revealed some miRNAs that had been previously known to be dysregulated in tumor tissue, like mir-21 and mir-221, Croce said. But "in the plasma, we had no clue which miRNA would be found to be different."
Interestingly, miRNAs deregulated in tissue specimens were "rarely detected in plasma samples," the authors reported in the study, "suggesting a predictive role of plasma miRNAs independent of tissue specimens.
"The levels of those few microRNAs can be used now to predict the occurrence of cancer and to predict whether the cancer is really bad or is really aggressive," said Croce, who added that he expects the signatures found to lead to an easy and accurate test, once the researchers have tied down the best technology and the most relevant miRNAs.
"We will continue to investigate more, to validate the signature, trying to find the easiest way to look at those changes," he said. "And then hopefully a very simple test will be developed, and that will be quite important."
Have topics you'd like to see covered in Gene Silencing News? Contact the editor at mashford [at] genomeweb [.] com.