Vista, Calif.-based glycobiology firm Sialix is developing diagnostics for breast, colon, and prostate cancers using auto-antibodies to the non-human sialic acid Neu5Gc as protein biomarkers.
The company, which has licensed intellectual property related to use of the auto-antibodies from the University of California, San Diego, is in the second year of a two-year, $800,000 Phase II Small Business Innovation Research contract supporting the development of the tests.
It is currently performing a 400-patient trial using samples provided by the National Cancer Institute's Early Detection Research Network.
According to Mark Hattendorf, Sialix's chief financial officer and chief operating officer, the company hopes to finish the trial by the end of the year and move toward developing the diagnostics as CLIA-based tests.
The ultimate goal, he told ProteoMonitor, is to offer the diagnostics "on a microarray reader than would be available to physicians for testing purposes."
An altered form of the sialic acid Neu5Ac, Neu5Gc has been linked to numerous types of cancers. Research by UCSD professor of medicine Ajit Varki, one of Sialix's founders and the chair of its scientific advisory board, has demonstrated that humans are unable to synthesize Neu5Gc.
However, consumption of Neu5Gc-rich foods – particularly red meat – leads to the incorporation of the glycan into human cell surfaces, and especially into carcinomas. This elevated level of Neu5Gc in carcinomas suggests that anti-Neu5Gc auto-antibodies could be useful cancer biomarkers.
To test this theory, Varki and UCSD assistant clinical professor of medicine Richard Schwab collaborated with Sialix scientists to screen sera from 386 cancer cases and controls on a sialoglycan microarray presenting 20 Neu5Gc-glycans and 20 paired control Neu5Ac-glycans to determine if antibodies for the Gc forms were more prominent in patients with carcinomas.
The results, which were detailed in a study published online this month in Cancer Research, suggest that antibodies against one of the 20 Gc glycans, Neu5Gc-sialyl-Tn, could be useful as protein biomarkers for cancer.
The finding "is actually what we would have predicted because sialyl-Tn is known somewhat as a cancer-specific glycan," Schwab told ProteoMonitor.
Since the study described in the Cancer Research paper, the researchers have built a larger array containing roughly 30 to 35 Gc and Ac glycan pairs, including a number of additional known cancer-specific glycans. That new array, Schwab said, "pretty much covers everything that's seen in humans" in terms of sialoglycans.
"We're only building up to di- and tri-saccharides, so the chemical space is actually not that large," he said, adding that, while in theory the screen could be expanded beyond tri-saccharides to larger glycans, "synthesizing these complex carbohydrates is actually incredibly difficult. A big limitation [in glycobiology] has always been just these technical limitations."
Construction of glycans for the array was done by UC-Davis researcher Xi Chen. Printing of the arrays was also a challenge, Sialix senior research scientist Nancy Hurtado-Ziola told ProteoMonitor, noting that while most array firms have experience with DNA and proteins, "almost no one has experience with glycan chains." Currently, the research is done with microarrays printed by Gaithersburg, Md.-based KamTek.
In the Cancer Research study, the anti-Neu5Gc-sialyl-Tn IgG antibody was able to detect carcinomas with an estimated AUC of 0.6. Setting the sensitivity at 20 percent gave a specificity of around 90 percent, Schwab said, noting that early detection screening tests with such low sensitivity can still be useful clinically, provided the specificity is high enough.
"The real issue we face in screening tests for cancer is specificity," he said "If we only detect 10 percent of cancer with a test, … if we were 100 percent specific that test would actually be OK. We can work on developing other tests that find other cancers, but the real bugaboo of the screening test is being highly specific."
Given that Neu5Gc in humans comes from diet, and epidemiological studies have shown that the effect of diet on cancer is around 15 percent, "the best case scenario is that this type of test might be useful for [detecting] maybe 15 percent of cancers," Schwab said. "And the number would probably be a little bit smaller than that."
In the initial study, the Neu5Gc-sialyl-Tn autoantibody alone was a better biomarker than other combinations of antibodies, but early data from the current 400-subject trial using the expanded microarray suggests a multi-marker approach could prove most effective, said Hurtado-Ziola.
"We have to finish this entire dataset before we can really speak to that, but as of right now it looks like there may be [autoantibody] combinations associated with different kinds of cancers," she said, suggesting these combinations may involve Neu5Gc-sialyl-Tn plus additional markers depending on the cancer.
While Sialix is investigating NeuGc autoantibodies for the detection of cancer, Schwab noted that it's possible they could also be used as a measure of cancer risk.
"One possibility is that these anti-Gc antibodies in people develop early in life" and their presence leads to chronic inflammation and cytokine release that actually stimulates the development of cancer, he said. In that case, anti-Gc antibody levels could be not just a biomarker for the presence of cancer, but for the future development of the disease.
"That would obviously be the most exciting thing because it would be something that we could intervene on," Schwab said. "If people eat a lot of Gc in their diet, there are ways we could impact that."
Testing this theory will require samples from "screened healthy populations followed for decades" to see who did and did not develop cancer," he said. "And there are really only a few groups that have [such samples], so gaining access to the samples is really at the top of our list."
"We've shown that we can discriminate between cases and controls," he said. "Next would be to prove in a retrospective study how well we can discriminate in a healthy population. That would help us determine how much this is working for early detection versus for cancer risk, and then based on those results we could design a prospective study where we would actually intervene."
"That would be the gold standard," he added. "If we can show that testing for our biomarker and then taking action based on that data can decrease mortality — that's the 30-year plan."
In the near term, Sialix plans to focus on development of cancer diagnostics, said Hurtado-Ziola, though the company would potentially be interested in licensing any further intellectual property generated by Schwab and Varki's work on the anti-Neu5Gc auto-antibodies.
Currently, Sialix generates revenue primarily through selling glycobiology research kits and services. In October 2007 it won a $150,000 phase I SBIR contract for preliminary work on assays to test for the auto-antibodies and last year had its first financing round, Hattendorf said. He declined to specify how much the round raised, but said it was under $1 million.
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.