A team of researchers at Oslo University Hospital has used arrays and other technologies to identify several gene signatures and biomarkers that could be used to diagnose and treat a number of cancers.
Led by Ragnhild Lothe, head of the molecular genetics group in the department of cancer prevention, the team has identified a dozen DNA methylation markers related to colorectal cancer that were recently licensed to Oxford Gene Technology (BAN 2/28/2012).
Another potential test that has emerged from Lothe's lab is called ColoGuideEx. Described in a recent paper in the journal Gut, the group used Affymetrix exon-level microarrays to obtain gene expression profiles for more than 300 samples from patients with different stages of colorectal cancer. The team identified a 13-gene expression classifier predictive of relapse among patients with stage II colorectal cancer.
The lead author on the study was Trude Ågesen, a PhD student in Lothe's lab.
Lothe told BioArray News this week that Inven2, the hospital's technology transfer office, is in discussions with "a few companies that have a potential interest in commercializing" the signature.
According to Lothe, the team applied a statistical model that identified a small number of genes that do not have a high level of covariance, so they each contribute to the prognostic information. The signature has also been validated in a second cohort of samples in Norway and in an external dataset from the US and Australia, Lothe said.
Affymetrix exon-level arrays replaced another array platform in Lothe's lab, Applied Biosystems' 1700 Chemiluminescent Expression Analysis System. ABI, now a Life Technologies division, discontinued the system in 2007 and stopped supporting it in 2009 (BAN 7/15/2008).
According to Guro Lind, head of the epigenetics group within Lothe's lab and a coauthor on the ColoGuideEx paper, the lab performed most of its initial array work on the 1700, including the experiments that led to the 12 markers that OGT licensed last month. When ABI discontinued the platform, Lothe's group acquired the remainder of ABI's array inventory, not only to complete the colorectal cancer-related projects, but two others, one focused on bladder cancer and another where the team profiled different tissues from cancers from 17 parts of the body.
"We just wanted to have enough arrays to finish up the projects because moving to another platform with another project would have been very challenging," Lind told BioArray News this week. "All of these [ABI] arrays were [run] several years ago," she added.
Bladder Cancer Panel
As with its work in colorectal cancer, the group's work in bladder cancer has also yielded some promising biomarkers. Working with researchers at the Portuguese Oncology Institute in Porto, Portugal, the team used gene-expression arrays to identify DNA methylation markers for bladder cancer.
Ultimately, a methylation panel consisting of three genes — GDF15, TMEFF2, and VIM — correctly identified bladder cancer tissues with 100 percent sensitivity and specificity. In urine samples, the panel achieved a sensitivity of 94 percent and specificity of 100 percent. The gene panel could also discriminate bladder cancer patients from both healthy individuals and renal or prostate cancer patients. The work was featured in a paper in Clinical Cancer Research last year.
Lind said that the researchers are now working on collecting more samples to validate the group's original findings.
Lothe said that the bladder cancer project is one of the "most promising projects" in her lab and said that a "well-known company" in the US has expressed interest in licensing the gene panel. She did not name the firm, but said that there is a market need for such a test.
"The importance of this is that a bladder cancer patient is one of the most expensive patients of all because they have to come for clinical checkups between 15 and 20 times during their disease because they have a high frequency of recurrence," said Lothe. "If you are able to have a precise, noninvasive test based on a patient's urine sample, that would be helpful," she said.
Lind said that such a test would most likely be performed using real-time PCR. She explained that smaller panels are favored, given the limited amount of DNA available from noninvasive sample materials.
"In a noninvasive test, we will have to use DNA from fecal or blood samples," said Lind. "That type of DNA is scarce in the samples, so I think you can compromise the results of the test by spreading out your template on too many assays, instead of picking the ones that you know will give robust results and limiting it down to the most promising ones," she said.
Even with regards to the dozen markers OGT licensed, Lothe said that an optimal early-stage test should consist of fewer markers.
"If you have too many, you will not have enough template in your noninvasive sample to actually be able to detect it," Lothe said. "So if you have at least five markers, you are both biologically and technically robust, which the single marker tests are not," she said. "The intention is to make this panel suitable for screening studies on a noninvasive level instead of colonoscopy."
Lind said that the lab is currently working with OGT to generate an early-stage colorectal test that can be used on noninvasive material.
In addition to Affy exon-level arrays and real-time assays, another technology Lothe's group has at its disposal is a fusion gene array platform developed in the lab of Rolf Skotheim, who heads the genome biology group in the department of cancer prevention.
According to a paper published last year, the content was culled from published fusion genes, including those reported only in individual studies and samples, and fusion genes resulting from deep sequencing of cancer genomes and transcriptomes.
From the total set of 548 fusion genes, the group designed an array of 599,839 oligonucleotides, targeting both chimeric transcript junctions as well as sequences internal to each of the fusion gene partners. According to the paper, Roche NimbleGen manufactured the chip.
Using the array, the researchers looked for the presence of fusion genes in a series of 67 cell lines representing 15 different cancer types. Data from 10 leukemia cell lines with known fusion gene status were used to develop an automated scoring algorithm, and in five cell lines the correct fusion gene was the top scoring hit, and one came second.
Lothe said that Inven2 has sought to license the fusion gene array to commercial partners but has not yet found success. "My impression is that the companies are thinking more about next-generation sequencing," she said, adding that she thinks "there is a gap and that such an array for fusion breakpoints for leukemia, for instance, would be of diagnostic interest."
Lothe's lab is also thinking about next-generation sequencing as well as newer arrays for studying DNA methylation, such as Illumina's Infinium HumanMethylation27 BeadChip.
"I think [array] technology has improved and a lot of great data has been generated using the Illumina arrays, but we haven't started using them yet," said Lind. "The question now is whether to go for that or whether to go for next-generation sequencing," she said. "The costs are almost the same now, at the same time the level of the data you get is much better with sequencing compared with arrays, where you just get the ratios."
Lothe said that her lab has acquired an Illumina sequencer and is performing RNA-seq and exome sequencing on it, but has not yet published the data.
Have topics you'd like to see covered in BioArray News? Contact the editor at jpetrone [at] genomeweb [.]com