By Deborah Borfitz and Stan Gloss
December 15, 2021 | A growing number of companies are endeavoring to replicate the success of gene sequencing juggernaut Illumina in the burgeoning field of proteomics. A top contender is Nautilus Biotechnology, which has spent the past few years drumming up excitement about the potential for comprehensively profiled proteins to make therapeutic and diagnostic development more effective and time efficient.
The goal of Nautilus Biotechnology is to “democratize access to the proteome” for all researchers through a platform, powered by artificial intelligence, which will “literally change biology,” according to CEO Sujal Patel, who co-founded Nautilus in 2016 with Stanford professor Parag Mallick. It allows scientists to look at multiple modifications on the same protein molecule, they explain, a notable advance given that the pattern of those modifications is believed to govern how and where a protein works—a level of specificity that is otherwise unavailable.
This capability helps differentiate diseased from healthy tissues, says Mallick, who serves as the company’s chief scientist. It is also useful in mechanism-of-action studies and as a companion diagnostic in prognostic studies on drug effectiveness.
Using the same method of repeatedly interrogating each molecule with a different class of affinity reagents, the platform can also be used to detect parts of many overlapping proteins that in combination are “shockingly specific,” Mallick says. “With the right machine learning framework layer on top, there is sufficient information in that repeated probing information to uniquely identify 95% of the proteome,” up from the current industry standard of 8%.
Nautilus announced earlier this year it had become a publicly traded company, a reflection of growing investor appetite for development-stage companies in the proteomics space. As of the end of the last quarter, Nautilus had $374 million in cash to build out the platform, reports Patel.
As also announced only last month, Nautilus entered into a strategic development and supply partnership with Abcam to build reagents to augment internal capabilities. Nautilus Biotechnology’s single-molecule protein analysis platform is sample-agnostic and will work with any biological reagent, says Mallick.
Abcam has an enormous catalogue of antibodies already being used by life science companies, including Genentech that last year began leveraging them in Nautilus’ single-molecule measurement tool, notes Patel. Similarly, Amgen and an investigator at MD Anderson Cancer Center more recently began working with Nautilus to investigate specific proteins and proteoforms of interest.
“Our primary mission is to build a platform that can be leveraged by biopharma and diagnostic companies, not jointly be in the business of making drugs,” Patel says. But those sorts of partnership opportunities will likely be created over the long term based on the recognized value of proteomic data throughout the drug discovery and development process, he adds.
Nautilus intends to commercialize easy-to-use, push-button instrumentation that scientists everywhere can use to access the proteome, Patel says. Indications are that the proteomics sector, which has been growing by about 12% annually over the past decade, has the potential to explode into a nearly $100 billion market in the coming decade, he adds.
In another three to five years, Patel predicts, “everyone will have to have one of these machines... to get full proteomic information and figure out how they are going to leverage machine learning and understand what's going on at the cellular level.”
Proteomics is “the most valuable source of biological insight, which we don't have access to today,” Patel continues. About 90% of marketed drugs and most molecular diagnostics target proteins, and all the body’s cellular machinery is defined by proteins. Their shape, and how they are modified, drive cellular functions.
In terms of its composition, the market is 50% proteomics organizations, much of it therapeutic development, and roughly 20% is academic and research organizations, says Patel. The other 30% includes agricultural applications such as crop protection from environmental stresses.
The Big Idea
As Patel recalls, he first met Mallick in 2004 when he was directing clinical proteomics at Cedars-Sinai Medical Center and a top customer of Isilon, the scale-out network-attached storage platform Patel had founded three years earlier and subsequently acquired by EMC. Mallick moved to Stanford about a decade ago to start a lab “sitting at the intersection of computing and biochemistry,” whose primary focus was precision and personalized medicine.
Mallick essentially pitched Patel the idea to start Nautilus, which he said would “revolutionize biomedicine,” Patel continues. It only took about a week to convince one another to go for it.
Mallick has worked in proteomics for decades, Patel notes, and is intimately familiar with many of the different applications and workflows—and why they will be transformative. Mass spectrometry is the current gold standard in proteomics, but a multitude of companies are working on alternative methods to better find and describe proteins with greater molecular precision, he says, including companies focused on the sequencing of peptides and targeted assay panels closer to clinical application.
Mass spectrometry has many “incredible use cases” in metabolomics, food safety, and metallurgical analysis, Patel notes. But it also has many significant limitations, including low sensitivity and a dynamic range that is “dramatically” limited, requiring proteins to be chopped into peptides. “You lose tons of information.”
There have been “very slow” incremental improvements to the technology, he adds, but mass spectrometry is so hard to do that extremely specialized expertise is required to use it effectively. The mission of Nautilus Biotechnology is to build an instrument so labs everywhere can more readily and holistically leverage proteomics in their research—be it to develop a more effective therapeutic, build a precision or companion diagnostic, or determine what therapy is best for a particular indication.
On the diagnostic front, the Nautilus proteomics platform looks to address a two-part problem, says Patel. The first is the “needle in a haystack” ordeal of finding rare biomarkers, and the other is the availability of an assay sensitive enough to detect those biomarkers in low concentrations.
The clinical market is a big but longer-term opportunity, Patel adds. “We have a highly sensitive platform that can look at 10 billion molecules of sample at once and, at single-molecule level, count up what protein biomarkers are there. And that has significant implications when you're trying to find something early or look at tissue leakage markers in blood serum or perhaps another biofluid.”
Unlike existing approaches, for example enzyme-linked immunosorbent assay, the Nautilus platform works by gluing down individual protein molecules from a sample and sequentially interrogating them using different affinity reagents. These are typically antibodies given ultra-bright labels to aid in the measurement of specific protein molecules and mutations, explains Mallick.
He likens the analysis method to a giant chess board to which molecules are affixed, with coordinates indicating their position on that board. “If I can identify each of the proteins at each coordinate, I get my quantification by just counting, and definitionally you can’t be more sensitive than a single molecule counter,” Mallick says.
The approach “fundamentally changes the nature of the problem,” he continues, referencing the move from a bulk to a single-molecule assay characterizing the heterogeneity of proteoforms as well as how proteins are being quantified. It has utility whether looking deeply at individual protein molecules or doing broad-scale analysis of intact proteins using multiple reagents specific to commonly occurring epitopes only three to four amino acids in length.
As used by Genentech, Amgen, and the MD Anderson investigator, the Nautilus platform is “analyzing the proteoform landscape of a particular protein target… at a level that can’t be done with mass spectrometry or with any other assay or solution out there in the marketplace,” says Patel. In the MD Anderson use case, the quantity and patterns of post-translational modifications (PTMs) on specific oncology protein targets of interest are being measured pre- and post-treatment.
PTMs are “decorations” put on proteins that change their structure and function—including how it moves and interacts with other proteins—and there are hundreds of different types, says Mallick, including methylation, phosphorylation, and glycosylation. “Those types of modifications are a huge part of how the protein does its job, and many proteins are modified at potentially several places.”
Conditions involving “ourself versus ourself” proteins include neurodegenerative, autoimmune, and heart diseases as well as cancer, he adds, are where specificity becomes particularly important. The therapeutic window is very narrow relative to diseases where bacteria or viruses need to be differentiated from human proteins.
“By being able to look at proteoforms, you are able to potentially find things that are super specific, serum troponin [protein often associated with heart attacks] being a great example,” says Mallick. The pattern of those modifications is also important in differentiating cardiac arrest from cardiac distress, which can also elevate troponin levels but may originate from exercise not disease.