Amgen Inc. v. Sanofi

598 U.S. 594

Opinion

PRELIMINARY PRINT

Volume 598 U. S. Part 2 Pages 594–616

OFFICIAL REPORTS OF

THE SUPREME COURT May 18, 2023

REBECCA A. WOMELDORF reporter of decisions

NOTICE: This preliminary print is subject to formal revision before the bound volume is published. Users are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D.C. 20543, pio@supremecourt.gov, of any typographical or other formal errors. 594 OCTOBER TERM, 2022

Syllabus

AMGEN INC. et al. v. SANOFI et al.

certiorari to the united states court of appeals for the federal circuit No. 21–757. Argued March 27, 2023—Decided May 18, 2023 This case concerns patents covering antibodies engineered by scientists that help reduce levels of low-density lipoprotein (LDL) cholesterol, sometimes called bad cholesterol because it can lead to cardiovascular disease, heart attacks, and strokes. To treat patients with high LDL cholesterol, scientists explored how antibodies might be used to inhibit PCSK9—a naturally occurring protein that binds to and degrades LDL receptors responsible for extracting LDL cholesterol from the blood- stream. Two pharmaceutical companies—Amgen and Sanof—each de- veloped a PCSK9-inhibiting drug. In 2011, Amgen obtained a patent for the antibody employed in its drug, and Sanof received one covering the antibody used in its drug. Each patent describes the relevant anti- body by its unique amino acid sequence. The dispute in this case con- cerns two additional patents Amgen obtained in 2014 that relate back to the company's 2011 patent. These later-issued patents purport to claim for Amgen “the entire genus” of antibodies that (1) “bind to spe- cifc amino acid residues on PCSK9,” and (2) “block PCSK9 from binding to [LDL receptors].” 872 F. 3d 1367, 1372. As part of its submission to the patent offce, Amgen identifed the amino acid sequences of 26 antibodies that perform these two functions. Amgen then described two methods—one Amgen called “the roadmap” and a second it called “conservative substitution”—that scientists could use to make other an- tibodies that perform the binding-and-blocking functions described in the claims. After Amgen obtained the 2014 patents, it sued Sanof for infringe- ment. Sanof replied that it was not liable to Amgen for infringement because Amgen's relevant claims were invalid under the Patent Act's “enablement” requirement. That provision requires a patent applicant to describe the invention “in such full, clear, concise, and exact terms as to enable any person skilled in the art . . . to make and use the [inven- tion].” 35 U. S. C. § 112(a). Sanof characterized the methods Amgen outlined for generating additional antibodies as amounting to little more than a trial-and-error process of discovery, and thus contended that Am- gen's patents failed to meet the enablement requirement because they sought to claim for Amgen's exclusive use potentially millions more anti- bodies than the company had taught persons skilled in the art to make. Both the district court and the Federal Circuit sided with Sanof. Cite as: 598 U. S. 594 (2023) 595

Held: The courts below correctly concluded that Amgen failed “to enable any person skilled in the art . . . to make and use the [invention]” as defned by the relevant claims. Pp. 604–616. (a) The patent “bargain” describes the exchange that takes place when an inventor receives a limited term of “protection from competi- tive exploitation” in exchange for bringing “new designs and technolo- gies into the public domain through disclosure” for the beneft of all. Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U. S. 141, 150–151. From the Patent Act's beginnings, Congress has sought to ensure the beneft of this bargain for the public by requiring the patent applicant to deposit a “specifcation . . . so particular . . . as not only to distinguish the invention or discovery from other things before known and used, but also to enable a workman or other person skilled in the art or manu- facture . . . to make, construct, or use the same.” 1 Stat. 110. Over time, Congress has left this “enablement” obligation largely intact. This Court has addressed the enablement requirement many times, and its decisions in O'Reilly v. Morse, 15 How. 62, The Incandescent Lamp Patent, 159 U. S. 465, and Holland Furniture Co. v. Perkins Glue Co., 277 U. S. 245, reinforce the simple statutory command: If a patent claims an entire class of processes, machines, manufactures, or composi- tions of matter, the patent's specifcation must enable a person skilled in the art to make and use the entire class. In Morse, for example, the Court held that one of the claims in Morse's patent for a telegraphic system was “too broad, and not warranted by law.” 15 How., at 113. The problem was that the claim covered all means of achieving tele- graphic communication, yet Morse's specifcation did not describe how to make or use them all. See id., at 113–117. In Incandescent Lamp, inventors of an “electric lamp” with an “incandescing conductor” made of “carbonized paper” claimed that a lamp created by Thomas Edison infringed their patent because it used bamboo as a conductor. The Court sided with Edison because the rival inventors, rather than confn- ing their claim to carbonized paper, “made a broad claim for every f- brous and textile material.” 159 U. S., at 472. That broad claim “might” have been permissible, the Court allowed, if the inventors had disclosed “a quality common” to fbrous and textile substances that made them “peculiarly” adapted to incandescent lighting, but they did not. Ibid. Finally, in Holland Furniture, a company that had developed a starch glue that was similar enough to animal glue to be used for wood veneering included a claim in its patent covering all “starch glue which, [when] combined with about three parts or less . . . of water, will have substantially the same properties as animal glue.” 277 U. S., at 251. The specifcation described the key input—the “starch ingredient”—in terms of its “use or function” rather than its “physical characteristics or chemical properties.” Id., at 256. The problem, as the Court put it, 596 AMGEN INC. v. SANOFI

was that “[o]ne attempting to use or avoid the use of [the] discovery as so claimed and described functionally could do so only after elaborate experimentation” with different starches. Id., at 257. All this is not to say a specifcation always must describe with particu- larity how to make and use every single embodiment within a claimed class. It may suffce to give an example if the specifcation also dis- closes “some general quality . . . running through” the class that gives it “a peculiar ftness for the particular purpose.” Incandescent Lamp, 159 U. S., at 475. Nor is a specifcation necessarily inadequate just be- cause it leaves the skilled artist to engage in some measure of adapta- tion or testing. See, e. g., Wood v. Underhill, 5 How. 1, 4–5. A specif- cation may call for a reasonable amount of experimentation to make and use a claimed invention, and reasonableness in any case will depend on the nature of the invention and the underlying art. See Minerals Separation, Ltd. v. Hyde, 242 U. S. 261, 270–271. Pp. 604–612. (b) Turning to the patent claims at issue in this case, Amgen's claims sweep much broader than the 26 exemplary antibodies it identifes by their amino acid sequences. Amgen has failed to enable all that it has claimed, even allowing for a reasonable degree of experimentation. Amgen's claims bear more than a passing resemblance to the broadest claims in Morse, Incandescent Lamp, and Holland Furniture. While Amgen seeks to monopolize an entire class of things defned by their function—every antibody that both binds to particular areas of the sweet spot of PCSK9 and blocks PCSK9 from binding to LDL recep- tors—the record refects that this class of antibodies does not include just the 26 that Amgen has described by their amino acid sequences, but a vast number of additional antibodies that it has not. Amgen insists that its claims are nevertheless enabled because scien- tists can make and use every functional antibody if they simply follow the “roadmap” or “conservative substitution.” These two approaches, however, amount to little more than two research assignments. The “roadmap” merely describes step-by-step Amgen's own trial-and-error method for fnding functional antibodies. Not much different, “conserv- ative substitution” requires scientists to make substitutions to the amino acid sequences of antibodies known to work and then test the resulting antibodies to see if they do too. Amgen's alternative arguments lack merit. Amgen first suggests that the Federal Circuit erred by confating the question whether an invention is enabled with the question how long may it take a person skilled in the art to make every embodiment within a broad claim. But the Federal Circuit made clear that it was not treating as dispositive the cumulative time and effort required to make the entire class of anti- bodies. Amgen next argues that the Patent Act supplies a single, uni- Cite as: 598 U. S. 594 (2023) 597

versal enablement standard, while the Federal Circuit applied a higher standard to Amgen's claims that encompass an entire genus of embodi- ments defned by their function. The Court agrees in principle that there is one statutory enablement standard, but the Federal Circuit's treatment in this case is entirely consistent with Congress's directive and this Court's precedents. Finally, while Amgen warns that a ruling against it risks destroying the incentives that lead to breakthrough in- ventions, since 1790 Congress has included an enablement mandate as one feature among many designed to achieve the balance it wishes to strike between incentivizing inventors and ensuring the public receives the full beneft of their innovations. In this case, the Court's duty is to enforce the statutory enablement requirement according to its terms. Pp. 612–616. 987 F. 3d 1080, affrmed.

Gorsuch, J., delivered the opinion for a unanimous Court.

Jeffrey A. Lamken argued the cause for petitioners. With him on the briefs were Michael G. Pattillo, Jr., Sarah J. Newman, Jonathan P. Graham, Stuart L. Watt, Wendy A. Whiteford, Emily C. Johnson, Keith R. Hummel, Andrei Harasymiak, and Sarah C. Columbia. Paul D. Clement argued the cause for respondents. With him on the brief were C. Harker Rhodes IV, David K. Barr, George W. Hicks, Jr., Nathan S. Mammen, Matthew M. Wolf, Dani el L. Reisner, Deborah Fishman, and Victor ia L. Reines. Colleen E. Roh Sinzdak argued the cause for the United States as amicus curiae urging affrmance. With her on the brief were Solicitor General Prelogar, Principal Deputy Assistant Attorney General Boynton, Deputy Solicitor Gen- eral Stewart, Daniel Tenny, Brian J. Springer, Thomas W. Krause, Farheena Y. Rasheed, Amy J. Nelson, and Monica B. Lateef.*

*Briefs of amici curiae urging reversal were fled for AbbVie Inc. by Jonathan S. Massey; for the Alliance of U. S. Startups and Inventors for Jobs et al. by Robert P. Taylor; for the Chemistry and Law Division of the American Chemical Society by James C. Carver, Justin J. Hasford, Matthew J. Hlinka, and Kendrew H. Colton; for Diversifed Researchers 598 AMGEN INC. v. SANOFI

Opinion of the Court

Justice Gorsuch delivered the opinion of the Court. The development of antibody drugs has yielded life- changing therapies. Individuals across the world now rely on antibody drugs to treat conditions ranging from Crohn's disease to cancer. This case concerns patents covering anti- bodies that help reduce levels of low-density lipoprotein cho- lesterol, sometimes called LDL cholesterol (for the obvious reason) or bad cholesterol (because it can lead to cardiovas- cular disease, heart attacks, and strokes).

and Innovators by Jeffrey P. Kushan, Joshua J. Fougere, Matthew Maho- ney, and Steven J. Horowitz; for the Fynder Group, Inc., by Matthew J. Dowd; for GSK plc by John M. Desmarais and Todd L. Krause; for Instil Bio, Inc., by John R. Labbe; for the Intellectual Property Law Association of Chicago by Charles W. Shifey, Robert H. Resis, John M. Augustyn, and Tiffany Gehrke; for Intellectual Property Professors by Mark A. Lemley, pro se; for the National Association of Patent Practitioners, Inc., by Xhavin Sinha and Louis J. Hoffman; for the New York Intellectual Prop- erty Law Association by Melvin C. Garner, Robert M. Isackson, Robert J. Rando, and Charles E. Miller; and for REGENXBIO Inc. et al. by William G. Jenks. Briefs of amici curiae urging affrmance were fled for the American Intellectual Property Law Association by James D. Crowne; for Arnold Ventures et al. by Lauren A. Degnan; for the Association for Accessible Medicines by Matthew S. Hellman and Adam G. Unikowsky; for Eli Lilly and Co., et al. by Katherine A. Helm, Michael H. McGinley, and Duane C. Marks; for Fresenius Kabi USA, LLC, by Neil Lloyd; for Genentech, Inc., et al. by Melanie L. Bostwick and Karen Boyd; for Intellectual Prop- erty Law Professors et al. by Edgar H. Haug; for Law Professors by Jeremy P. Oczek and Joshua D. Sarnoff, pro se; for Pfzer Inc. by Dimi- trios T. Drivas, John P. Scheibeler, and Amit H. Thakore; for the Public Interest Patent Law Institute by Nina Srejovic and Alex Moss; for Small and Medium Biotechnology Companies by Mark A. Perry, Zachary D. Tripp, and Robert B. Niles-Weed; for Unifed Patents, LLC, by Lisa M. Ferri, Richard J. McCormick, Andrew J. Pincus, and Jonathan Stroud; for Viatris Inc. by Steffen N. Johnson and Richard Torczon; for Robin Feldman by Gideon A. Schor and Ms. Feldman, pro se; and for Gregory Paul Winter et al. by Sophie F. Wang and Eric J. Marandett. Briefs of amici curiae were fled for the High Tech Inventors Alliance et al. by John D. Vandenberg and Sarah E. Jelsema; and for the Intellec- tual Property Owners Association by Vanessa Yen and Evan Diamond. Cite as: 598 U. S. 594 (2023) 599

The case comes to us this way. Several years ago, peti- tioners (Amgen) obtained two patents. Together, these pat- ents claim a monopoly over all antibodies that (1) bind to specifc amino acids on a naturally occurring protein known as PCSK9, and (2) block PCSK9 from impairing the body's mechanism for removing LDL cholesterol from the blood- stream. Soon after receiving these patents, Amgen sued re- spondents (Sanof) for infringement. In response, Sanof ar- gued that the patents were invalid under § 112 of the Patent Act. That provision requires a patent applicant to describe its invention “in such full, clear, concise, and exact terms as to enable any person skilled in the art . . . to make and use the [invention].” 35 U. S. C. § 112(a). Sanof contended that Amgen's patents failed to meet this standard because they sought to claim for Amgen's exclusive use potentially mil- lions more antibodies than the company had taught scientists to make. In the end, both the district court and Federal Circuit sided with Sanof. The question we face is whether to disturb their judgment. I A The immune system produces antibodies as a defense to foreign agents called antigens. When a particular antigen— a virus, for example—enters the body, the immune system generates antibodies to attack it. In a successful attack, the antibodies target and bind to the antigen, stopping it from causing harm to the body. See Brief for Sir Gregory Paul Winter et al. as Amici Curiae 8 (Winter Brief); M. Lemley & J. Sherkow, The Antibody Patent Paradox, 132 Yale L. J. 994, 1001–1002 (2023). Antibodies are incredibly diverse. Some scientists esti- mate that there may be as many unique antibodies as there are stars in the galaxy. See id., at 1003; see also B. Briney, A. Inderbitzin, C. Joyce, & D. Burton, Commonality Despite Exceptional Diversity in the Baseline Human Antibody Rep- ertoire, 566 Nature 393, 397 (No. 7744, Feb. 2019) (estimating 600 AMGEN INC. v. SANOFI

that the immune system could potentially generate up to a quintillion unique antibodies). This diversity shows up in both structure and function. Start with structure. “When scientists refer to an anti- body's `structure,' ” they may have in mind “several related concepts,” each of which describes “what an antibody is.” Winter Brief 10. Antibodies are made up of amino acids, and scientists commonly identify a particular antibody ac- cording to its specifc sequence of amino acids—what they call an antibody's “ `primary structure.' ” Id., at 9–10. But antibodies are not just linear chains of amino acids. As the atoms of the amino acids interact with each other, they cre- ate folds that result in complex three-dimensional shapes. Ibid. Scientists refer to an antibody's intricate topography as its “tertiary structure.” Id., at 10. An antibody's structure does much to dictate its function— its ability to bind to an antigen and, in some instances, to block other molecules in the body from doing the same. “For an antibody to bind to an antigen, the two surfaces have to ft together and contact each other at multiple points.” Id., at 11. But just because an antibody can bind to an anti- gen does not mean that it can also block. To bind and block, the antibody must establish a suffciently broad, strong, and stable bond to the antigen. See ibid. Different antibodies have different binding and blocking capacities based on the amino acids that compose them and their three-dimensional shapes. See id., at 11–12. Despite recent advances, aspects of antibody science re- main unpredictable. For example, scientists understand that changing even one amino acid in the sequence can alter an antibody's structure and function. See id., at 14. But scientists cannot always accurately predict exactly how trad- ing one amino acid for another will affect an antibody's struc- ture and function. Ibid. As Amgen's expert testifed at trial: “ `[T]he way in which you get from sequence to that three-dimensional structure isn't fully understood to- Cite as: 598 U. S. 594 (2023) 601

day. It's going to get a Nobel Prize for somebody at some point, but translating that sequence into a known three- dimensional structure is still not possible.' ” Id., at 14–15.

B While the immune system naturally produces an army of antibodies to protect us from various harms, scientists are now able to engineer antibodies to assist in treating diseases. Some of these lab-made antibodies target not foreign agents but the body's own proteins, receptors, and ligands. “While naturally occurring in our bodies, these [proteins, receptors, and ligands] can also be involved in infammatory disorders, uncontrolled cell growth, or other biological pathways that may be associated with disease.” Id., at 8. One part of this effort has focused on the creation of anti- bodies to treat patients with high LDL cholesterol. A silent killer, LDL cholesterol can contribute to the formation of plaque in the arteries that may lead to cardiovascular dis- ease, heart attacks, and strokes. For many people with high LDL cholesterol, drugs called statins offer an effective treat- ment. For others, statins do not work well or come with unwelcome side effects. In those cases, a relatively new antibody-based treatment known as a PCSK9 inhibitor may be appropriate. See Amgen Inc. v. Sanofi, 872 F. 3d 1367, 1371 (CA Fed. 2017). PCSK9 is a naturally occurring protein that binds to and degrades LDL receptors. That can pose a problem because the body produces LDL receptors to perform the benef- cial function of extracting LDL cholesterol from the blood- stream. See ibid. Scientists have understood this much for some time. But it wasn't until fairly recently that they began exploring how antibodies might be used to inhibit PCSK9 from binding to and degrading LDL receptors as a way to treat patients with high LDL cholesterol. In the mid-2000s, a number of pharmaceutical companies began looking into the possibility of making antibodies to 602 AMGEN INC. v. SANOFI

target PCSK9. See Brief for Respondents 7; Brief for Ar- nold Ventures et al. as Amici Curiae 17–20. More precisely, they sought to create antibodies that could bind to a particu- lar region of PCSK9 called the “sweet spot.” See Brief for Petitioners 10–11. The sweet spot is a sequence of 15 amino acids out of PCSK9's 692 total amino acids. Id., at 11. By binding to the sweet spot, scientists found, an antibody could prevent PCSK9 from binding to and degrading LDL recep- tors. See id., at 10–11; Amgen, 872 F. 3d, at 1371. Eventually, Amgen developed a PCSK9-inhibiting drug that it marketed under the name Repatha, and Sanofi produced one it labeled Praluent. Each drug employs a dis- tinct antibody with its own unique amino acid sequence. See id., at 1371–1372; Brief for Respondents 8–10. In 2011, Amgen obtained a patent for the antibody employed in Repa- tha, and Sanof received one covering the antibody used in Praluent. See id., at 8, 9. Each patent describes the rele- vant antibody by its amino acid sequence. See ibid. Nei- ther of these patents is at issue in this case. Instead, our dispute focuses on two additional patents Amgen obtained in 2014 that relate back to the company's 2011 patent. See U. S. Patent No. 8,829,165 (Sept. 9, 2014); U. S. Patent No. 8,859,741 (Oct. 14, 2014). We refer to them as the '165 and '741 patents. In particular, this case re- volves around claims 19 and 29 of the '165 patent and claim 7 of the '741 patent. See 987 F. 3d 1080, 1082 (CA Fed. 2021). In these claims, Amgen did not seek protection for any particular antibody described by amino acid sequence. Instead, Amgen purported to claim for itself “the entire genus” of antibodies that (1) “bind to specifc amino acid resi- dues on PCSK9,” and (2) “block PCSK9 from binding to [LDL receptors].” Amgen, 872 F. 3d, at 1372. As part of its submission to the patent offce, Amgen iden- tifed the amino acid sequences of 26 antibodies that perform these two functions, and it depicted the three-dimensional Cite as: 598 U. S. 594 (2023) 603

structures of two of these 26 antibodies. 987 F. 3d, at 1083. But beyond that, Amgen only offered scientists two methods to make other antibodies that perform the binding and block- ing functions it described. The frst method is what Amgen calls the “roadmap.” Brief for Petitioners 13. At a high level, the roadmap directs scientists to: (1) generate a range of antibodies in the lab; (2) test those antibodies to determine whether any bind to PCSK9; (3) test those antibodies that bind to PCSK9 to determine whether any bind to the sweet spot as described in the claims; and (4) test those antibodies that bind to the sweet spot as described in the claims to determine whether any block PCSK9 from binding to LDL receptors. See id., at 13–14. The second method is what Amgen calls “conservative substitution.” Id., at 14, 17. This technique requires scientists to: (1) start with an anti- body known to perform the described functions; (2) replace select amino acids in the antibody with other amino acids known to have similar properties; and (3) test the resulting antibody to see if it also performs the described functions. See id., at 14–15. C Soon after receiving the '165 and '741 patents, Amgen sued Sanof for infringing them. Sanof replied that it was not liable to Amgen because the relevant claims were invalid as a matter of law. Invalid, Sanof said, because Amgen had not enabled a person skilled in the art to make and use all of the antibodies that perform the two functions Amgen de- scribed in its claims. See 987 F. 3d, at 1083–1085. While Amgen had identifed the amino acid sequences of 26 antibod- ies that bind to PCSK9 and block it from binding to LDL receptors, Sanof observed that Amgen's claims cover po- tentially millions more undisclosed antibodies that perform these same functions. And, Sanof argued, neither of the two methods Amgen had outlined for generating additional antibodies with the same functions enable a person skilled in 604 AMGEN INC. v. SANOFI

the art to do so reliably. Instead, Sanof submitted, those methods require scientists to engage in little more than a trial-and-error process of discovery. See id., at 1085. After lengthy proceedings, the district court granted Sa- nof judgment as a matter of law, concluding that the claims at issue “are not enabled.” 2019 WL 4058927, *13 (Del., Aug. 28, 2019). The Federal Circuit affrmed. 987 F. 3d, at 1088. It determined that “no reasonable factfnder could conclude” that Amgen had provided “adequate guidance” to make and use the claimed antibodies “beyond the narrow scope of the [26] working examples” it had identifed by their amino acid sequences. Ibid. In response to Amgen's peti- tion for certiorari, we agreed to take up the case. 598 U. S. ––– (2022). II The Constitution vests Congress with the power to “pro- mote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” Art. I, § 8, cl. 8. Right there in the text, one fnds the outline of what this Court has called the patent “bargain.” Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U. S. 141, 150 (1989). In exchange for bringing “new designs and technol- ogies into the public domain through disclosure,” so they may beneft all, an inventor receives a limited term of “pro- tection from competitive exploitation.” Id., at 151; see also The Federalist No. 43, p. 272 (C. Rossiter ed. 1961) (J. Madi- son) (explaining that in such cases “[t]he public good fully coincides . . . with the claims of individuals”). Congress has exercised this authority from the start. The Patent Act of 1790 promised up to a 14-year monopoly to any applicant who “invented or discovered any useful art, manufacture, . . . or device, or any improvement therein not before known or used.” Act of Apr. 10, 1790, § 1, 1 Stat. 110. Refecting the quid-pro-quo premise of patent law, the statute required the applicant to deposit with the Secretary Cite as: 598 U. S. 594 (2023) 605

of State a “specifcation . . . so particular . . . as not only to distinguish the invention or discovery from other things be- fore known and used, but also to enable a workman or other person skilled in the art or manufacture . . . to make, con- struct, or use the same.” § 2, ibid. The statute made clear that this disclosure would ensure “the public may have the full beneft [of the invention or discovery], after the expira- tion of the patent term.” Ibid. Even as Congress has revised the patent laws over time, it has left this “enablement” obligation largely intact. See 35 U. S. C. §§ 111, 112. Section 111 of the current Patent Act provides that a patent application “shall include . . . a specifcation as prescribed by section 112.” § 111(a)(2)(A). Section 112, in turn, requires a specifcation to include “a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art . . . to make and use the same.” § 112(a). So today, just as in 1790, the law secures for the public its beneft of the patent bargain by ensuring that, “upon the expiration of [the pat- ent], the knowledge of the invention [i]nures to the people, who are thus enabled without restriction to practice it.” United States v. Dubilier Condenser Corp., 289 U. S. 178, 187 (1933); see also Grant v. Raymond, 6 Pet. 218, 247 (1832) (Marshall, C. J.) (“This is necessary in order to give the pub- lic, after the privilege shall expire, the advantage for which the privilege is allowed, and is the foundation of the power to issue a patent”); Whittemore v. Cutter, 29 F. Cas. 1120, 1122 (No. 17,600) (CC Mass. 1813) (Story, J.) (“If therefore [the disclosure] be so obscure, loose, and imperfect, that this cannot be done, it is defrauding the public of all the consider- ation, upon which the monopoly is granted”). This Court has addressed the enablement requirement on many prior occasions. See, e. g., Wood v. Underhill, 5 How. 1 (1846); O'Reilly v. Morse, 15 How. 62 (1854); The Incandes- cent Lamp Patent, 159 U. S. 465 (1895); Minerals Separa- 606 AMGEN INC. v. SANOFI

tion, Ltd. v. Hyde, 242 U. S. 261 (1916); Holland Furniture Co. v. Perkins Glue Co., 277 U. S. 245 (1928). While the technologies in these older cases may seem a world away from the antibody treatments of today, the decisions are no less instructive for it. Begin with Morse. While crossing the Atlantic Ocean in 1832 aboard a ship named Sully, Samuel Morse found himself in conversation with other passengers about “experiments and discoveries” around electromagnetism. 15 How., at 68. “In the course of this discussion, it occurred to [Morse] that, by means of electricity, signs representing fgures, letters, or words, might be legibly written down at any distance.” Id., at 69. So clear was the idea in Morse's mind that, “[b]efore he landed in the United States, he had . . . drawn out in his sketch book . . . the form of an instrument for an electro- magnetic telegraph.” Ibid. Immediately upon his arrival in New York, Morse showed his brothers his sketches. See id., at 69–70. He spent the next few years refning his invention. See id., at 70–76. The “great diffculty” he faced was that “the galvanic cur- rent, however strong in the beginning, became gradually weaker as it advanced on the wire[,] and was not strong enough to produce a mechanical effect, after a certain dis- tance.” Id., at 107. By 1837, Morse had a solution: “com- bining two or more electric or galvanic circuits, with in- dependent batteries for the purpose of overcoming the diminished force of electro-magnetism in long circuits.” Id., at 109. Morse demonstrated his telegraph the following year at the Franklin Institute in Philadelphia, and he dis- played it soon after in Congress. See id., at 76. He re- ceived a patent in 1840, which reissued in 1848. See id., at 81–83. The litigation that brought Morse before this Court con- cerned a telegraphic system that Henry O'Reilly had in- stalled between Louisville and Nashville. See id., at 65. Morse sued O'Reilly for infringement, alleging that O'Reil- Cite as: 598 U. S. 594 (2023) 607

ly's system was “identical with” Morse's own. Id., at 66. O'Reilly mounted a number of defenses, including that Mor- se's patent was void because it lacked an adequate specifca- tion. See id., at 99–101, 112. Morse's patent included eight claims, and this Court had no trouble upholding seven of them—those limited to the telegraphic structures and systems he had designed. See id., at 85–86, 112, 117. But the Court paused on the eighth. That claim covered “the essence” of the invention, which Morse described as “the use of the motive power of the elec- tric or galvanic current . . . however developed for marking or printing intelligible characters, signs, or letters, at any distances.” Id., at 112 (internal quotation marks omitted). Leaving no doubt about this claim's scope, Morse stated plainly: “ `I do not propose to limit myself to the specifc ma- chinery or parts of machinery described in the foregoing specifcation and claims.' ” Ibid. The Court held the eighth claim “too broad, and not war- ranted by law.” Id., at 113. The problem was that it cov- ered all means of achieving telegraphic communication, yet Morse had not described how to make and use them all. See id., at 113–117; see also 3 Chisum on Patents § 7.03[1], pp. 7– 18 to 7–19 (2021). “[I]f the eighth claim . . . can be main- tained,” the Court concluded, “there was no necessity for any specifcation, further than to say that he had discovered that, by using the motive power of electro-magnetism, he could print intelligible characters at any distance.” 15 How., at 119. “[I]t will be admitted on all hands, that no patent could have issued on such a specifcation.” Ibid. Consider, too, Incandescent Lamp. For much of the 19th century, gas lamps helped illuminate streets and supple- mented candles inside homes, factories, offces, and theaters. But gas lighting had drawbacks. It took effort to ignite lamps each night and extinguish them each morning. Then there were the problems of soot and fumes. See R. Stross, The Wizard of Menlo Park 84–85 (2007) (Stross). By the 608 AMGEN INC. v. SANOFI

1870s, many had experimented with other forms of lighting, including incandescence and the arc light. 159 U. S., at 470. But these alternatives burned unreliably or with unbearable brightness. See id., at 470–471. The latter problem in par- ticular led one observer to lament this “new sort of urban star,” which shines “horrible, unearthly, obnoxious” light. R. L. Stevenson, A Plea for Gas Lamps, in Virginibus Pueris- que and Other Papers 295 (1881). Enter Thomas Edison. From his laboratory in Menlo Park, Edison and a team toiled to improve upon the prevail- ing method of incandescent lighting, which tended to employ carbon flaments. 159 U. S., at 471–473. The problem with carbon flaments was that they disintegrated rapidly. In a sense, “carbon contained in itself the elements of its own de- struction.” Id., at 471. Seeking an alternative, Edison tin- kered for a time with platinum, but it was expensive and diffcult to bring to the point of incandescence without melt- ing. Stross 78, 82. Eventually, Edison dispatched men across the globe to collect specimens of bamboo. Id., at 109– 110. One sample from Japan worked brilliantly because “[its] fbres [ran] more nearly parallel than in other species of wood.” 159 U. S., at 473. Satisfed, Edison arranged to have a Japanese farmer supply all of the bamboo he would ever need. Stross 110. But there was a catch. William Sawyer and Albon Man had obtained a patent for an “ `electric lamp' ” with an “ `in- candescing conductor' ” made of “ `carbonized fbrous or tex- tile material,' ” which they claimed was an improvement over conductors made of “ `mineral or gas carbon.' ” 159 U. S., at 466, 468. Sawyer and Man's patent had not won them commercial success. They had designed a lamp with a con- ductor made of carbonized paper, but the lamp proved defec- tive and quickly fell out of use. See id., at 471–472. Still, their failure did not stop them from seeking to share in some of Edison's success. Sawyer and Man alleged that Edison's lamp infringed their patent because it “made use of a fbrous Cite as: 598 U. S. 594 (2023) 609

or textile material, covered by the patent.” Id., at 471. What was that offending material? Bamboo. This Court sided with Edison. It held that Sawyer and Man's patent claimed much but enabled little. “Sawyer and Man supposed they had discovered in carbonized paper the best material for an incandescent conductor.” Id., at 472. But “[i]nstead of confning themselves to carbonized paper, as they might properly have done, and in fact did in their third claim, they made a broad claim for every fbrous or textile material.” Ibid. Even that broad claim “might” have been permissible, the Court allowed, if Sawyer and Man had disclosed “a quality common” to fbrous and textile substances that made them “peculiarly” adapted to incandes- cent lighting. Ibid. Had they done so, others would have known how to select among such materials to make an opera- ble lamp. But the record showed that most fbrous and tex- tile materials failed to work. Only through “painstaking ex- perimentation” did Edison discover that bamboo “answered the required purpose.” Id., at 475–476. The Court sum- med up things this way: “[T]he fact that paper happens to belong to the fbrous kingdom did not invest [Sawyer and Man] with sovereignty over this entire kingdom.” Id., at 476. The Court returned to these principles in Holland Furni- ture. There, the evidence indicated that animal glue has properties that have long made it excellent for wood veneer- ing. See 277 U. S., at 247. Seeking a substitute, Perkins Glue Company had developed and patented a starch glue similar enough to animal glue that craftsmen could also use it for wood veneering. See ibid. Yet Perkins's patent in- cluded a claim that went beyond the specifc starch glue it manufactured. See id., at 250–251. This claim covered all “starch glue which, [when] combined with about three parts or less by weight of water, will have substantially the same properties as animal glue.” Id., at 251. Perkins's specif- cation instructed gluemakers to choose a “starch ingredient” 610 AMGEN INC. v. SANOFI

with “such qualities” that it would yield a product “ `as good as animal glue' ” for wood veneering “when combined with three parts of water and with alkali.” Id., at 256. The Court held this broad claim invalid for lack of en- ablement. Id., at 258. The specifcation described the key input—the “starch ingredient”—in terms of its “use or func- tion” rather than its “physical characteristics or chemical properties.” Id., 256. And that left gluemakers in a bind. As the Court put it: “One attempting to use or avoid the use of Perkins' discovery as so claimed and described function- ally could do so only after elaborate experimentation” with different starches. Id., at 257. To be sure, the Court held, Perkins was entitled to its patent on the specifc starch glue it had invented. See id., at 255. The specifcation de- scribed that glue's “characteristic ingredient” with “particu- larity.” Ibid. But just as Morse could not claim all means of telegraphic communication, and Sawyer and Man could not claim all fbrous and textile materials for incandescence, Per- kins could not claim all starch glues made from whatever starch happened to perform as well as animal glue. To hold otherwise, the Court said, “would extend the monopoly be- yond the invention.” Id., at 258. Our decisions in Morse, Incandescent Lamp, and Holland Furniture reinforce the simple statutory command. If a patent claims an entire class of processes, machines, manu- factures, or compositions of matter, the patent's specifcation must enable a person skilled in the art to make and use the entire class. In other words, the specifcation must enable the full scope of the invention as defned by its claims. The more one claims, the more one must enable. See § 112(a); see also Continental Paper Bag Co. v. Eastern Paper Bag Co., 210 U. S. 405, 419 (1908) (“[T]he claims measure the invention”). That is not to say a specifcation always must describe with particularity how to make and use every single embodi- Cite as: 598 U. S. 594 (2023) 611

ment within a claimed class. For instance, it may suffce to give an example (or a few examples) if the specifcation also discloses “some general quality . . . running through” the class that gives it “a peculiar ftness for the particular pur- pose.” Incandescent Lamp, 159 U. S., at 475. In some cases, disclosing that general quality may reliably enable a person skilled in the art to make and use all of what is claimed, not merely a subset. See id., at 475–476.1 Nor is a specifcation necessarily inadequate just because it leaves the skilled artist to engage in some measure of ad- aptation or testing. In Wood, a patent claimed a process for making bricks by mixing coal dust into clay. 5 How., at 4. The patent included “a general rule” about the proportion of dust and clay to use and offered two alternative proportions “where the clay has some peculiarity.” Id., at 5. The Court upheld the claim, recognizing that “some small difference in the proportions must occasionally be required” given the va- rieties of clay. Ibid. Similarly, in Minerals Separation, the Court dismissed a challenge to a claimed process for sep- arating metal from mineral ores. 242 U. S., at 270. The record showed that “preliminary tests” were required to adapt the process to any particular ore. Ibid. Once more, the Court explained that “the certainty which the law re- quires in patents is not greater than is reasonable.” Ibid. And because the “composition of ores varies infnitely,” it

1 See also Béné v. Jeantet, 129 U. S. 683, 684–686 (1889) (rejecting claim to method of shrinking coarse hair because the specifcation failed to give “one skilled in chemistry such an idea of the particular kinds and character of the chemicals, or combination of chemicals, with the relative proportions of each, as would enable him to use the invention without having to resort to experiments of his own to discover those ingredients”); Corona Cord Tire Co. v. Dovan Chemical Corp., 276 U. S. 358, 385 (1928) (rejecting claims to process of treating rubber with “ `a disubstituted guanidine' ” because “between ffty and one hundred substances” ft that description and the specifcation did not disclose “any general quality common to di- substituted guanidines which makes them all effective”). 612 AMGEN INC. v. SANOFI

was “impossible to specify in a patent the precise treatment which would be most successful and economical in each case.” Id., at 271.2 Decisions such as Wood and Minerals Separation establish that a specifcation may call for a reasonable amount of ex- perimentation to make and use a patented invention. What is reasonable in any case will depend on the nature of the invention and the underlying art. See Minerals Separa- tion, 242 U. S., at 270–271; see also Mowry v. Whitney, 14 Wall. 620, 644 (1872) (“[T]he defniteness of a specifcation must vary with the nature of its subject. Addressed as it is to those skilled in the art, it may leave something to their skill in applying the invention”). But in allowing that much tolerance, courts cannot detract from the basic statutory re- quirement that a patent's specifcation describe the invention “in such full, clear, concise, and exact terms as to enable any person skilled in the art” to “make and use” the invention. § 112(a). Judges may no more subtract from the require- ments for obtaining a patent that Congress has prescribed than they may add to them. See Bilski v. Kappos, 561 U. S. 593, 602–603, 612 (2010). III With these principles in mind, we return to claims 19 and 29 of the '165 patent and claim 7 of the '741 patent. In doing so, we do not doubt that Amgen's specifcation enables the 26 exemplary antibodies it identifes by their amino acid se-

2 See also, e. g., Ives v. Hamilton, 92 U. S. 426, 429, 432 (1876) (upholding claim “for an improvement in sawmills” based on “curved guides at the upper end of the saw,” even though the specifcation did not “stat[e] the nature of the curve,” because a “good mechanic acquainted with the con- struction of sawmills, and having the patent and diagram before him, would have no diffculty in adopting the improvement, and making suitable curves”); Tilghman v. Proctor, 102 U. S. 707, 732–733 (1881) (upholding claim for process of separating fats and oils even though the specifcation “suggests a trial . . . with different degrees of heat so as to ascertain that which is best for each particular kind of fat”). Cite as: 598 U. S. 594 (2023) 613

quences. Even Sanof concedes that description is enough to allow a person skilled in the art to make and use those embodiments. See Tr. of Oral Arg. 68. But the claims be- fore us sweep much broader than those 26 antibodies. And we agree with the lower courts that Amgen has failed to enable all that it has claimed, even allowing for a reasonable degree of experimentation. While the technology at the heart of this case is thor- oughly modern, from the law's perspective Amgen's claims bear more than a passing resemblance to those this Court faced long ago in Morse, Incandescent Lamp, and Holland Furniture. Amgen seeks to monopolize an entire class of things defned by their function—every antibody that both binds to particular areas of the sweet spot of PCSK9 and blocks PCSK9 from binding to LDL receptors. The record refects that this class of antibodies does not include just the 26 that Amgen has described by their amino acid sequences, but a “vast” number of additional antibodies that it has not. 987 F. 3d, at 1085, 1088; see 2019 WL 4058927, *8 (“at least millions of candidates”); see also Tr. of Oral Arg. 52–53. Much as Morse sought to claim all telegraphic forms of com- munication, Sawyer and Man sought to claim all fbrous and textile materials for incandescence, and Perkins sought to claim all starch glues that work as well as animal glue for wood veneering, Amgen seeks to claim “sovereignty over [an] entire kingdom” of antibodies. Incandescent Lamp, 159 U. S., at 476. That poses Amgen with a challenge. For if our cases teach anything, it is that the more a party claims, the broader the monopoly it demands, the more it must enable. That holds true whether the case involves telegraphs devised in the 19th century, glues invented in the 20th, or antibody treatments developed in the 21st. To be fair, Amgen does not dispute this much. It freely admits that it seeks to claim for itself an entire universe of antibodies. Still, it says, its broad claims are enabled because scientists can make and 614 AMGEN INC. v. SANOFI

use every undisclosed but functional antibody if they simply follow the company's “roadmap” or its proposal for “conserv- ative substitution.” We cannot agree. These two approaches amount to little more than two research assignments. The frst merely de- scribes step-by-step Amgen's own trial-and-error method for fnding functional antibodies—calling on scientists to create a wide range of candidate antibodies and then screen each to see which happen to bind to PCSK9 in the right place and block it from binding to LDL receptors. See Part I–B, supra; 987 F. 3d, at 1088; 2019 WL 4058927, *10–*13. The second isn't much different. It requires scientists to make substitutions to the amino acid sequences of antibodies known to work and then test the resulting antibodies to see if they do too—an uncertain prospect given the state of the art. See Parts I–A, I–B, supra; 987 F. 3d, at 1088; 2019 WL 4058927, *10–*13. Whether methods like a “roadmap” or “conservative substitution” might suffce to enable other claims in other patents—perhaps because, as this Court suggested in Incandescent Lamp, the inventor identifes a quality common to every functional embodiment, supra, at 609—they do not here. They leave a scientist about where Sawyer and Man left Edison: forced to engage in “painstak- ing experimentation” to see what works. 159 U. S., at 475. That is not enablement. More nearly, it is “a hunting li- cense.” Brenner v. Manson, 383 U. S. 519, 536 (1966). Think about it this way. “Imagine a combination lock with 100 tumblers, each of which can be set to 20 different positions.” Brief for Intellectual Property Law Professors and Scholars as Amici Curiae 20. “Through trial and error, imagine that an inventor fnds and discloses 26 different suc- cessful lock combinations.” Ibid. But imagine, too, “that the inventor tries to claim much more, namely all successful combinations,” while instructing others “to randomly try a large set of combinations and then record the successful ones.” Id., at 20–21. Sure enough, that kind of “roadmap” Cite as: 598 U. S. 594 (2023) 615

would produce functional combinations. Id., at 21. But it would not enable others to make and use functional combina- tions; it would instead leave them to “random trial-and-error discovery.” Ibid. Like many analogies, this one may over- simplify a bit, but it captures the gist of the problem. Failing in its primary argument that it has enabled all of the antibodies it claims, Amgen tries a few alternative lines of attack. First, it suggests that the Federal Circuit erred by applying an enablement test unmoored from the statutory text. As Amgen sees it, that court confated the question whether an invention is enabled with the question how long may it take a person skilled in the art to make every embodi- ment within a broad claim. See Brief for Petitioners 24–29; see also id., at 2, 19–20, 30–36. We do not see it that way. While we agree with Amgen that enablement is not meas- ured against the cumulative time and effort it takes to make every embodiment within a claim, we are not so sure the Federal Circuit thought otherwise. That court went out of its way to say that it “do[es] not hold that the effort required to exhaust a genus is dispositive.” 987 F. 3d, at 1088 (em- phasis deleted). Instead, the court stressed, the problem it saw is the same problem we see: Amgen offers persons skilled in the art little more than advice to engage in “trial and error.” Ibid. (internal quotation marks omitted). In any event, we review judgments of the lower courts, not statements in their opinions. See Black v. Cutter Labora- tories, 351 U. S. 292, 297 (1956). Taking a similar tack, Amgen next argues that the Federal Circuit erroneously “raise[d] the bar” for enablement of claims that, like Amgen's, encompass an entire “genus” of embodiments defned by their function. Brief for Petition- ers 25 (internal quotation marks omitted). This is imper- missible, Amgen argues, because the Patent Act “provides a single, universal enablement standard for all invention[s].” Ibid. (internal quotation marks omitted). Here, too, we agree with Amgen in principle: There is one statutory en- 616 AMGEN INC. v. SANOFI

ablement standard. But, once more, we do not understand the Federal Circuit to have thought differently. Instead, we understand that court to have recognized only that the more a party claims for itself the more it must enable. As we have seen, that much is entirely consistent with Congress's directive and this Court's precedents. Finally, Amgen warns that an affrmance risks “destroy- [ing] incentives for breakthrough inventions.” Id., at 38. But striking the proper balance between incentivizing inven- tors and ensuring the public receives the full beneft of their innovations is a policy judgment that belongs to Congress. Since 1790, Congress has included an enablement mandate as one feature among many designed to achieve the balance it wishes. Our only duty in this case lies in applying that mandate faithfully. * Section 112 of the Patent Act refects Congress's judgment that if an inventor claims a lot, but enables only a little, the public does not receive its beneft of the bargain. For more than 150 years, this Court has enforced the statutory enable- ment requirement according to its terms. If the Court had not done so in Incandescent Lamp, it might have been writ- ing decisions like Holland Furniture in the dark. Today's case may involve a new technology, but the legal principle is the same. The judgment is Affrmed. Reporter’s Note

The attached opinion has been revised to refect the usual publication and citation style of the United States Reports. The revised pagination makes available the offcial United States Reports citation in advance of publication. The syllabus has been prepared by the Reporter of Decisions for the convenience of the reader and constitutes no part of the opinion of the Court. A list of counsel who argued or fled briefs in this case, and who were members of the bar of this Court at the time this case was argued, has been inserted following the syllabus. Other revisions may include adjustments to formatting, captions, citation form, and any errant punctuation. The following additional edits were made:

p. 597, line 14, “10 F. 4th 1016” is replaced with “987 F. 3d 1080” p. 609, line 9, “and” is replaced with “or”