As proteomics researchers’ biomarker discovery efforts begin to yield fruit, companies and individuals trying to collect IP on their markers are likely to face higher standards — and more competition — than ever before, legal experts and researchers told ProteoMonitor. In addition, the multiplexed, multi-instrument, and often multipurpose nature of mass spec-based biomarkers could cloud the IP surrounding disease indications.
The US patent office has recently upped the ante for patentability with a new set of standards demanding that scientists demonstrate a “specific, substantial, credible use” for disease markers, according to Laurie Axford, a partner specializing in biotech IP at Burns, Doane, Swecker, and Mathis. “The biggest obstacle for people seeking patent protection for [biomarkers] is that the standards have gotten higher,” Axford said: There is now a greater burden on inventors to validate and show a direct role for a biomarker in a disease, and to demonstrate that the marker has clear utility.
While making claims on a biomarker by association with other known markers or by just showing its presence in a few disease samples was once enough to claim patentability, Axford said, “now the patent office is saying it’s not enough — you have to prove it has the same function.”
For mass spec-based pattern markers, this means more pressure to actually identify the proteins behind the peaks, said Lisa Treannie, a partner in the IP group at Ropes and Gray. If in a patent application “you submit a read-out from a mass spec and you say, ‘a method of diagnosing whatever disorder by assessing a protein that has this functional characteristic, a peak of this level at this position,’ I think those are unlikely to be viewed favorably in the patent office.” She said the patent office would be more likely to demand an identity for the protein “because you can have a number of proteins that will give you similar mass spec results, and unless you know what it is that you’re actually looking for, how do you validate that what you have is the same as what someone else has?”
The bar was not always so high: CA125 — the old ovarian cancer marker that those promoting pattern markers are trying to replace — was granted a patent with the defining characteristic of a molecular weight obtained from a gel band, noted Emanuel Petricoin and Lance Liotta of the NCI-FDA Clinical Proteomics Program. “People filed patents on CA125 without having an amino acid sequence on it — they were able to objectify it by characteristics that were patentable,” Petricoin said. “So you can imagine that especially with ultra-high-resolution mass spectroscopy that an investigator could say, ‘here’s a protein ion fragment at 6,025.8793,’ and that becomes an accurate mass tag.”
But Treannie said that simply providing a molecular weight on its own is now not likely to be enough. “The patent office has raised the bar a little bit by saying, ‘unless you can tell us what the amino acid sequence of your protein is, or give us a commonly known name so that someone would know what the sequence is, we’re going to hold you to a very high standard of providing other identifying characteristics,’” she said.
A Shrinking Pool of Gold
The best-case scenario for a company seeking maximum patent protection is to patent a previously unknown protein, as this would essentially grant the company claims to any and all future uses of that protein. But such composition patents are not likely to be the dominant form of patent protection for biomarkers, according to Treannie. “Most of the patent protection that’s going to come out of this area is going to be methods,” she said. “Because a lot of proteins and genes are known, it just gets harder and harder as time goes by to actually find a novel gene or protein.”
But Petricoin and Liotta said they are staying ahead of this curve: They have filed patents on behalf of the government for “more than a couple thousand proteins, fragments, and peptides that never before have been known to exist in the blood,” Liotta said. He said these patent filings — some of which have been granted and some of which are still pending — are advertised in the federal registry and that non-exclusive license rights to the markers are being offered for a small fee to any company that has the infrastructure to develop them into a diagnostic or other clinical use. The pair has also filed broad patents for all of the methods involved in finding, harvesting, and collecting the markers. The goal, they said, was to get everything into the public domain as quickly as possible, “so that someone could come in and get a nonexclusive, cheap license that wouldn’t be a barrier, and anyone could do it,” Petricoin said. “The whole idea of government patents is not to make money, but to get it transferred out and to the public benefit.”
But companies obtain patents to make money — which means that other parties might not be as quick to get biomarkers out into the public domain. “Human nature being what it is, I think it’s likely that certainly in the beginning, people will want to stake out their territory,” Treannie said. “There’s just so much potential money at stake.” But she expressed hope that companies may well eventually choose a nonexclusive licensing path, rather than “make the entire biotech industry come after them and try to invalidate their patent” — a risk they would run if a major enabling technology or marker were tied up in an exclusive licensing deal.
Nevertheless, the avalanche of protein biomarker patents is going to come. “I think there’s every bit as much of a rush with proteins as there are with genes,” Axford said.
You Infringe Mine, I Infringe Yours
Given the multiple methods and multiple markers involved in a pattern biomarker, getting as broad and detailed a patent as possible will be key, Petricoin and Liotta said. “We write our claims up listing, say, 70 ways to measure the proteins, because we realize there are probably new technologies that haven’t been thought of yet,” Liotta said. “So we file the exact mass tag, the sequence, the identity, all of the above measured by ELISA, measured by mass spec, measured by capture array, measured by protein arrays.” The purpose is to cover as many technologies and instruments as possible under the claim, so that someone else cannot file a competing patent, he said. Still, there may be a limit to the breadth of the claims allowed, according to Treannie. “I don’t think the patent office is going to give broad method claims — I think they’re going to say a pattern generated by a particular method,” such as mass spec.
No matter how comprehensive a patent claim is, however, there are still several potential infringement scenarios that could come up. In one scenario, one company could file for a patent that covers essentially the same panel of biomarkers as another company’s claim, but claims a different method of finding the panel. In this case, Treannie said, the patent office would probably not have the search capabilities to determine that the panel was really the same, and so would not have reason to reject the claim. “I think what’s going to end up happening is, you’re going to get your claims, I’m going to get my claims … and that’s going to wind up being fodder for litigation down the road,” she said. Depending on the money involved, some sort of cross-licensing agreement or collaboration would probably result.
Complex infringement issues could also arise from the multiplexed nature of biomarker panels. If one company patented a collection of markers that overlapped in part with another company’s different panel of markers, then both companies could be granted patents, but at the same time each would be infringing on the other’s patent, according to Treannie. “There wouldn’t really be a basis for the patent office to prevent the second patent from granting, because … the combination of markers is both new and unobvious,” she said. The result would be “competing patents where neither can practice the other party’s invention.” But that doesn’t mean the patents would be useless: Each company’s patent would give it leverage for developing cross-licensing agreements, Treannie said.
Similar issues could arise from one company trying to market a certain set of markers for use in diagnosing a different disease state than the disease state that the company that originally patented for the panel indicated. That is, if one company held a composition patent for a marker that could be used as a diagnostic for breast cancer, another company could get a method patent using that marker to diagnose a different disease — but the first company would still own the rights to the marker itself, according to Axford. In this case, “your patent gives you the right to exclude others from practicing your invention, but it doesn’t exclude others from patenting different aspects of your invention,” she said. The upshot would be the same for the company patenting the second application: cross-license the marker, or face possible litigation.
The Means Versus the Ends
When it comes to marketing an actual diagnostic, the biomarkers themselves and their predictive value for disease are not the only elements of the diagnostic that carry IP. Every tube, mass spec, and computer algorithm is patented as well, and any company marketing a diagnostic that uses these elements would have to pay royalties for each instrument used. But if an instrument or algorithm is used in the biomarker discovery process and not in the marketed diagnostic, these royalties could be difficult to collect.
In the case of pattern recognition software, for example, “an investigator could use that method, discover the ions, and measure them by different methods and not be interfering with the underlying pattern recognition methods” when it came to the diagnostic, according to Petricoin. Although there might be a one-time patent infringement of the software in the discovery process, this infringement would be difficult to prove, he said.
Treannie said additionally that even if the company holding the patent could prove infringement, the damages would likely be minor.
But while Petricoin felt that this might render the pattern recognition patent “not very useful,” Treannie said these patents still held leveraging value. “A lot of companies would rather take a reasonable license to have a known risk, rather than take the risk of patent litigation,” she said.