Two pharmaceutical giants head to the Supreme Court in a dispute over a patent covering cholesterol-lowering antibodies on March 20. In Amgen Inc. v. Sanofi, the justices will consider whether Amgen’s patent is too broad to be valid under federal patent rules.  

Amgen has patents on monoclonal antibodies that lower LDL, or “bad,” cholesterol. The antibodies work by binding to a particular protein in the body (PCSK9), which then blocks that protein from destroying receptors that extract cholesterol from the bloodstream.

Amgen and other companies already have patents covering particular antibodies defined by their amino acid sequences. Those narrow patents are not in dispute here. Instead, this case involves a broader patent Amgen obtained, which covers any monoclonal antibody that binds to particular sweet spots on the PCSK9 protein and blocks that protein from binding to LDL receptors. In other words, the patent covers a genus defined by how the antibodies work. The patent gives details for 26 example antibodies and specifies how to run further processes to identify others from a pool of potentially thousands or even millions of candidates. However, the patent also goes far beyond those 26 examples, covering any antibody within the broad genus.

One of the rules for a valid patent is that it must enable others to make and use the invention. On Monday the justices will analyze whether Amgen’s broad patent satisfies that rule.

Background

Patent attorneys love to talk about the quid pro quo of patents. Under the bargain, the government grants an inventor a monopoly in the form of a patent. In exchange, the inventor must disclose the invention to the world in terms sufficient “to enable any person skilled in the art … to make and use” the invention. This law, 35 U.S.C. §112(a), is known as the “enablement” requirement, which forms the heart of this dispute. 

The enablement requirement allows the public to use the invention after the patent expires. It also ensures that the patent claims only what the inventor has actually invented and described to the public.

A century ago, the Supreme Court clarified that a patent fails the enablement test if a person skilled in the art would have to resort to “painstaking experimentation” or “elaborate experimentation” before being able to make or use the invention. The U.S. Court of Appeals for the Federal Circuit, which hears the nation’s appeals in patent cases, further clarified that the patent must teach how to make and use “the full scope of the claimed invention without ‘undue experimentation.’”

The dispute here

The question here is how to apply the enablement requirement when a patent claims a broad genus of things, this genus is defined by how the things work, and only a small number of the things are likely to work at all.

Sanofi claims that what Amgen did in this case was “claim a lot and disclose a little.” Amgen’s patent covers any monoclonal antibody that binds to particular spots on the PCSK9 protein and successfully blocks the protein. The patent covers many other possibilities beyond the 26 examples it describes (including Sanofi’s infringing antibody, which falls within the scope of Amgen’s patent but is not one of Amgen’s 26 examples). For those other possibilities, the patent explains a process in which a scientist creates antibodies and checks whether they work. Amgen describes this process as “screening”; Sanofi calls it “trial and error.”

Amgen contends that courts should not consider the “cumulative time and effort” that it may take someone to “reach the full scope” of the patent, meaning the effort required to make a complete set of antibodies covered by the patent. Yet Amgen agrees that patents “must reasonably enable the entire scope of the claim” and that the enablement must be “commensurate with the scope of the claims.” 

Sanofi, on the other hand, argues that the patent must enable someone “to make the entire invention claimed, not just a subset, without the need for any significant independent experimentation.” It asserts that a patent fails the enablement requirement if the public “cannot predictably produce specific undisclosed embodiments of the claimed invention” without significant trial and error.

In a “friend of the court” brief, the federal government largely sides with Sanofi. the government asserts that a patent fails the enablement requirement if it requires someone “to engage in the same trial-and-error process the inventor undertook to produce her innovation in the first place.”

Will “genus” patents survive? 

Although the case is nominally about the test applicable for enablement under 35 U.S.C. § 112, the real question is whether broad “genus” patents will remain viable. 

A typical patent may cover one or a small handful of specific devices or methods for doing something. By contrast, a genus patent covers a group of related things, often defined by how the things work. Genus patents are particularly common in the pharmaceutical, chemical, and biotech fields. 

In recent years, the Federal Circuit has invalidated almost every genus patent that it faced. A group of law professors published a recent article reporting that in the past 30 years, only a small minority of genus claims survived the Federal Circuit. 

Companies and organizations have lined up on both sides of the issue. Many friend-of-the-court briefs urge the court to rule for Amgen and hold that genus patents pass the enablement test. They argue, among other things, that it makes no sense for the law to require disclosure of things the inventor doesn’t know (for example, to require Amgen to disclose how to make antibodies that Amgen never actually made). Other briefs argue that broad genus patents go too far precisely because they cover things the inventor doesn’t know about. 

Is this really a dispute about the meaning of “undue experimentation”?

Several issues lurk in the background in this case. Much of the action seems to stem from the “undue experimentation” limitation. Although both sides fight over the text of 35 U.S.C. § 112, “undue experimentation” is not in the text of the statute at all. Yet both sides embrace that long-settled judicial gloss on the statute.

“Undue experimentation” captures part of the patent bargain. If the inventor made a broad invention and teaches the world how to make and use the broad invention, then the inventor should get a patent with a broad scope. If the inventor made only a narrow invention, and teaches the world about only a narrow set, then the inventor should get only a narrow patent. The patent scope should be commensurate with the invention. If the patent claims a broad invention but discloses only a little, the “undue experimentation” limitation helps to check the proper scope.

A century ago, the Supreme Court faced a similar issue in The Incandescent Lamp Patent. In that case Thomas Edison was accused of infringing a light-bulb patent that covered filaments made out of any “carbonized fibrous or textile material.” But the patent included only a couple of examples (such as carbonized paper). After months of trial and error, Edison found that bamboo also worked as a filament, but thousands of other materials did not work at all. The Supreme Court did not allow the patentee to have a monopoly over every possible fibrous material because it took significant experimentation to find any other materials within the genus that worked.

In some sense, the background rates of success may affect the “undue experimentation” analysis. In Incandescent Lamp, under 1% of the tested materials actually worked. Even with the patent in hand, it took months of trial and error to find another material that worked. This fact suggests that the inventor did not actually invent a new genus of filaments, but instead invented only the few specific types described in the patent. Consider instead if Edison’s research showed that 50% of tested materials worked. In that scenario, the patent may have survived because a scientist would not have to conduct many experiments before finding another working species within the genus.

Consider also industry norms. Some have criticized Amgen’s disclosure for requiring “trial and error” experimentation, contending that trial-and-error experimentation is “undue experimentation.” Others counter that in monoclonal antibody research, trial-and-error experimentation is customary, and therefore is not “undue experimentation.”

We will know more on Monday about how the justices line up on the question of how much experimentation is “undue” when it comes to patents like this one.