NEW YORK – Having raised more than $100 million in funding over the last year, life science start-up Seer is working towards a 2021 launch of its ProteoGraph proteomics platform.
The ProteoGraph system uses nanoparticle-based enrichment of proteins in samples like human plasma to enable unbiased proteomic discovery experiments. According to Omid Farokhzad, CEO and founder of the Redwood City, California-based firm, Seer's technology allows for deep proteomic profiling at a level of throughput not currently possible with traditional mass spectrometry-based approaches.
The company this week published the first study demonstrating the technology. In a paper in Nature Communications, Seer employees and affiliated researchers used the platform to analyze 141 plasma samples from non-small cell lung cancer patients and healthy controls, identifying more than 2,000 proteins and generating a classifier that distinguished between the two sets of subjects with an area under the curve of .91.
The ProteoGraph platform is based on the observation that when incubated in a biological sample, nanoparticles collect proteins, which form a "corona." Given this, nanoparticles can serve as an enrichment tool, allowing researchers to pull proteins out of a sample, which they can then identify and quantify using technologies like mass spec.
This phenomenon was first observed by pharma researchers using nanoparticles for drug delivery. (Formerly a professor at Harvard Medical School, Farokhzad was also director of the Center for Nanomedicine at Brigham and Women's Hospital.) Traditionally, researchers have focused on trying to reduce this protein accumulation as it was thought to interfere with the nanoparticle-drug complexes' therapeutic effects. However, in recent years, some began to explore the possibility of using it for proteome profiling purposes.
Notable from a proteomic profiling perspective, is the fact that nanoparticles appear to collect proteins from across a large portion of the dynamic range present in samples like plasma. This means they could allow researchers to analyze low abundance proteins without using time-consuming and expensive steps like protein depletion and fractionation that are required for mass spec-based experiments that aim to go deep into the proteome.
With the nanoparticle-based approach, "you essentially grab proteins across that entire dynamic range," Farokhzad said.
Because different nanoparticles will, based on their physical and chemical properties, attract different sets of proteins, researchers could in theory use multiple sets of such particles to interrogate large portions of a proteome throughout much of its dynamic range.
Additionally, Farokhzad noted, this sampling is done in a 96-well plate format, making it rapid and automatable. In the Nature Communications study, the researchers profiled the 141 samples in under three weeks. Current mass spec methods can match or exceed that level of throughput but not at the same depth of coverage.
Affinity agent-based platforms like SomaLogic's Somascan and Olink's proximity extension assay also offer both high throughput and deep proteome coverage, with Somascan currently measuring around 5,000 proteins and Olink planning to expand its assay from its current count of around 1,500 proteins to 4,500 proteins by 2022.
Omead Ostadahn, president and chief operating officer at Seer, which he joined last month after leaving his position as chief product and marketing officer at Illumina, said, however, that the ProteoGraph system had the advantage of providing a more unbiased look at the proteome than the SomaLogic or Olink approaches, both of which rely on predefined panels of assays.
"Our roadmap is entirely predicated on increasing the diversity of our nanoparticles and essentially the range of proteoforms to which they can bind," he said. "And that is where I would argue that in terms of proteoform reporting, our numbers are likely to be substantially higher than what is possible with any other technology."
Farokhzad said that another potential advantage of the system compared to affinity-based approaches is the richness of information it can provide on different protein variants or post-translationally modified proteins.
"Take the protein ApoE and say that because of genetic variations and post-translational modifications a bunch of changes occur to that protein," he said. "Assume that some of those changes occur in parts of the protein that are not the site of the interaction [with a nanoparticle]. So the particle picks up all the different proteoforms of that protein, and then when they go to a mass spec and have the different peptides, you will identify these different proteoforms."
This ability to retain information on proteoform diversity follows the shift within proteomics in recent years towards focusing more on understanding the importance of modified forms of proteins as opposed to simply looking at expressions of a gene product as a whole.
Another area where Seer's system overlaps with trends in proteomics is its ability to identify protein-protein interactions by analyzing patterns in which proteins are found together in the same nanoparticle corona.
"Because the nanoparticles offer essentially a scaffold for proteins to assemble with each other the way they normally would biologically, the platform allows for protein-protein interactions to be mapped," Farokhzad said.
In addition to the lung cancer profiling work, the Nature Communications study also looked at the linearity and reproducibility of the system's measurements. In a spike-in experiment looking at four nanoparticles and peptides from three proteins, angiogenin, c-reactive-protein, and calprotectin, they observed good linearity across two orders of magnitude. Looking at reproducibility, they found that a 10 nanoparticle panel measured 761 proteins groups with a coefficient of variation below 20 percent.
The researchers used 60-minute LC-MS/MS runs to analyze the contents of the protein coronas captured by the nanoparticles. They used a conventional data-dependent acquisition bottom-up mass spec workflow for the analysis, but the authors noted that other mass spec methods including various data-independent acquisition approaches could be used. John Blume, Seer's chief scientist, said that the company also has top-down proteomic methods on its "radar screen," as such approaches could make even fuller use of the proteoform information captured by the ProteoGraph system.
While mass spec is currently the most natural fit as a read-out for the platform, it should be compatible with other protein analysis technologies that emerge in the future, Farokhzad said.
Seer plans to sell the system as a package called the ProteoGraph Product Suite that will contain the nanoparticles and required reagents along with an automated liquid handling system for processing samples and analysis software. Users will provide their own instrument, typically a mass spec, for protein detection.