Screening with Proteomics

It's hard to swallow the fact that more than 35 years after US President Nixon declared the War on Cancer, there is only a 3 percent survival rate if you're diagnosed with pancreatic cancer. Scientists like Kim Kelly know that finding biomarkers capable of detecting this cancer at its earliest stages will save lives.

While a technique called phage display has been used for biomarker discovery, research published in PLoS Medicine and led by Kelly, an assistant professor at Harvard and Massachusetts General Hospital, has pushed it into the realm of proteomics. In her study, she used a phage display screen to select for peptides that bind preferentially to cell surface proteins on mouse pancreatic cancer cells. Using this screen, she and her team created a "multimodal" nanoparticle with the potential to find very early stage pancreatic cancer cells.

To start, they created a library of 1 billion bacteriophage clones, each of which was engineered to express a random seven-amino acid peptide sequence. "We used the virus, the bacteriophage, as a scaffold to display the random peptides," Kelly says. By incorporating random genetic material into the phage genome to display this protein onto one of the phage's coat proteins, "we've tricked it into being a scaffold, or making billions of different peptides for us."

Kelly then performed a screen to see which of these phages would bind most strongly to cancer cells. She mixed them with both mouse cancer cells and normal cells, isolated those that bound to the cancer cells, and washed away the rest. "Instead of screening all of them in different wells and using up a lot of tissue or material, we can put them all into one well of cells, or tissue, let them compete against each other, and wash away everything that doesn't stick," Kelly says. "Then we can figure out which peptide is doing that specific phenotype." After sequencing the best picks, followed up with several rounds of validation, they characterized one particular peptide that bound most strongly.

Knowing that the peptide was binding to something on the cancer cell's surface, they followed up with a pull-down assay that used the peptide as an affinity ligand. They came up with a 500-kDA protein that later mass spec analysis told them was plectin-1. While plectin-1 is normally found inside cells, it's found both on the inside and on the outer cell membrane of pancreatic cancer cells.

Hoping to move this work into the clinical realm, Kelly's group attached the plectin-1-targeted peptide to a magnetofluorescent nanoparticle to act as an in vivo probe. After injecting this particle into mice with early stage pancreatic cancer, they detected fluorescence in the abdominal region of the mice, suggesting uptake. Nanoparticles in control mice, however, didn't illuminate any regions of the pancreas.

It could be a very useful probe applied clinically since it's so specific, but Kelly thinks that might be a little bit down the road. Kelly estimates that it might be ready in the next year and a half, but "it's still a long way off from being in clinical trials," she says.

But the technique itself has broad potential right now. "Phage display has been used for finding markers, but this is a little bit different in that we're starting to query the proteome on cancers. So using this technique, it's not just limited to finding something that recognizes one cell type or the other," she says. "It can also tell us something about the status of the receptors and potentially signal transduction pathways that are aberrantly regulated on the cancer cell surface."