Skip to main content
Premium Trial:

Request an Annual Quote

MIT-Harvard Team Develops Synthetic Peptide-Based Approach to Biomarker Discovery


A team led by researchers at the Massachusetts Institute of Technology and Harvard Medical School has developed a biomarker discovery technique using mass-encoded synthetic peptides to track changes in protease activity linked to disease states.

The method, which they detailed in a paper published last month in Nature Biotechnology, offers a potential route to improve early detection and monitoring of various diseases and could prove useful across a wide range of settings including point-of-care and direct-to-consumer, Gabriel Kwong, a postdoctoral researcher at MIT and first author on the paper, told ProteoMonitor.

Protein biomarkers have drawn great attention for their theoretical potential for early detection and monitoring of diseases like cancer. In practice, however, it has proven difficult to find clinically useful markers – in significant part because of their low abundance.

To get around this problem, the MIT-Harvard team tracked not protein markers produced by the body but, rather, levels of synthetic peptides that they injected and targeted to organs and areas of interest. Once inside the body, these peptides were degraded by native proteases and then excreted in urine. By measuring levels of the degraded synthetic peptides via mass spectrometry, the researchers could observe differences in protease activity between healthy and diseased subjects and identify signatures of this activity indicative of disease.

Because they are using synthetic peptides as reporters, the researchers can introduce them at levels detectable via conventional peptide quantification methods. This is potentially useful not only for situations where existing early detection markers are in very low abundance, but also for diseases that aren't known to result in shedding of any protein markers at all, Kwong said.

"We're very interested in using this technology for early detection, but … in cancer, for example, there are a lot of tumors that don't secrete a protein biomarker, at least that we know of," he said. "So we think that [we can use] these synthetic markers to interrogate larger pools of [diseases]."

To target and then quantify the peptides and corresponding protease activity, the researchers used a combination of nanotechnology and isobaric tagging. To target the synthetic peptides to their desired location, they attached them to iron oxide nanoworm nanoparticles, which carried them to the organ of interest – in this case the liver.

To make for convenient measurement of the degraded peptides on the other end of the process, the researchers attached the synthetic peptides to isobarically tagged mass reporters, allowing them to be quantified via mass spec in much the same way as samples tagged with isobaric labels like Thermo Fischer Scientific's TMT or AB Sciex's iTRAQ reagents.

Using sets of ten different peptides, Kwong and his colleagues monitored the development of liver fibrosis in mice, identifying subsets of peptides that could track liver fibrosis and resolution with an area under the curve of .91.

They used the same set of peptides to detect colorectal cancer, comparing the effectiveness of their markers to carcinoembryonic antigen, a protein marker for colorectal cancer currently in clinical practice. In a study of 10 mice, they found the synthetic peptide system could detect early-stage colorectal cancer with an AUC of .89, compared to .61 for CEA, and that it could identify low CEA-secreting tumors at sizes nine-fold smaller than was possible using CEA.

Despite the small sample size, the two experiments offer an intriguing proof of concept of the researchers' approach. "We're very excited about the prospects of trying to translate this technology" to humans, Kwong said.

He noted that to do so, considerably more work will be required, including safety studies of the reagents, but, Kwong said, the researchers hope to begin human studies in the next two to three years.

Currently they are developing the platform for monitoring liver metastases, he said. "If a colorectal cancer patient comes in, you resect the colon, and then we use [the published method] for monitoring liver metastases over time. The same thing with pancreatic cancer patients. That's one specific direction we're going with the strategy."

Another angle Kwong and his colleagues are pursuing is adapting the method to an immunoassay readout by including in the molecules' reporter region agents to which antibodies can bind.

That, he noted, would allow for wider access to the technique, among, for instance, hospital labs without access to high-end mass spectrometry equipment.

It might also make it possible to move the method into the home testing market, packaging it in a format similar to home pregnancy tests, Kwong said.

"All those are is essentially printed antibodies on a piece of paper," he said. "So we're designing [printed antibodies] to interface with our synthetic biomarkers so that we can sample urine and be able to tell if a patient has cancer or another type of disease."

Kwong added that the inexpensiveness of such a format might also make it amenable to use in resource-constrained countries for purposes like infectious disease testing.

In the Nature Biotechnology paper, the researchers maxed out at ten synthetic peptides. They have since upped that number to 20, Kwong said, and could, in theory, generate larger libraries.

Larger numbers would likely be necessary for more nuanced analyses, he said. "If we want to look at, say, breast cancer versus prostate cancer versus something else closely related, I'm guessing we would need to start on the order of around 50 candidate peptides that we could then boil down to a 10-element signature that could be very specific."

The team is also gathering data on the kinds of protease activity linked to various disease states – key knowledge for determining what peptide constructs to use to detect a given disorder.

"We're working closely [with colleagues] at MIT who have developed a proteomic method… to collect human tumors and do mass spec analysis of the matrix molecules," Kwong said. "And they have been able to provide us with lists of proteases that are upregulated in [specific disease] samples."

Another key to the process is the continued development of nanoparticle-based targeting techniques. In the recent study, the researchers took advantage of the liver's natural porousness to target their peptides to the organ, but, Kwong said, work is underway on targeting methods for other organs.

"For instance," he said, "if you design nanoparticles that are positively charged, they tend to accumulate in the lungs."

Another possible method, Kwong noted, is targeting particles to organ-specific integrins – transmembrane receptors that help cells attach to surrounding tissue. "So there are a variety of different ways to target different organs," he said.

The iron oxide nanoworms used in the Nature Biotechnology study are straightforward to make and fairly well established from a regulatory perspective, Kwong said, with US Food and Drug Administration-approved formulations and GMP guidelines in place. "I don't think synthesizing them on a large scale for testing in humans is an issue," he said.

What could prove an issue, however, are the peptides themselves. As Kwong noted, the technique uses panels of different peptides, each of which could be required by FDA to undergo safety studies.

"So instead of doing one safety study on a 10-mer batch, we [might] have to do a safety study on each individual peptide we're using," he said.

The researchers have filed two patents on the method, one covering the use of nanomaterials and the sampling of peptide degradation products in urine and the other on the mass encoding used in their reporter tags, Kwong said, adding that they would likely file additional patents as their research into the method progresses.