Biotech Patents

After Mayo (2012), Myriad (2013), and Amgen (2023), § 101 patent eligibility in biotechnology is complex, but many biotech inventions remain patentable. Patent-eligible biotech inventions include: (1) cDNA and synthetic nucleic acid sequences — Myriad specifically preserved cDNA and non-naturally occurring sequences; (2) recombinant proteins with non-naturally occurring sequences or modifications; (3) engineered cell lines and genetically modified organisms (Diamond v. Chakrabarty, S.Ct. 1980 — 'anything under the sun made by man'); (4) CRISPR-Cas9 and other gene editing tools and methods as applied technology; (5) therapeutic compositions — drug formulations, vaccine compositions, gene therapy vectors; (6) specific antibodies characterized by CDR sequence, VH/VL sequence, or defined structural features; (7) assays and diagnostic kits as physical compositions; (8) methods that include a specific technical step transforming or analyzing a sample in a concretely described way, beyond the natural correlation itself; (9) bioinformatics algorithms applied to specific technical applications (not merely performing computation on biological data). Not patent-eligible: naturally occurring DNA sequences in isolation (Myriad); natural phenomena, laws of nature, and natural correlations as such (Mayo); naturally occurring proteins in isolation (§ 101 natural product); diagnostic methods claiming 'determine [biomarker] → correlate with disease → treat' without a specific, concrete inventive technical step beyond the natural correlation. The key distinction is whether the claim is directed to the natural phenomenon or to a specific technical application, machine, or composition that uses the natural phenomenon in a defined, non-obvious way.

Amgen Inc. v. Sanofi (598 U.S. 594, decided May 18, 2023) was a unanimous Supreme Court decision that invalidated Amgen's claims to a functional genus of antibodies targeting PCSK9 (a protein involved in LDL cholesterol regulation) under the enablement requirement of 35 U.S.C. § 112(a). Amgen's claims covered 'all antibodies that bind the PCSK9 protein at a specific epitope region and block PCSK9 from binding LDL receptors' — a functional definition covering potentially millions of different antibody structures. The specification disclosed 26 specific antibodies by structure and amino acid sequence. The Court held this was insufficient enablement because making and using the full scope of the claims — identifying and characterizing all antibodies in the claimed genus — required undue experimentation. The Court applied the Wands factors: quantity of experimentation required, nature of the invention's unpredictability, amount of direction provided, the existence of working examples, and representativeness of those examples to the full genus. Amgen's 26 examples were a tiny fraction of the millions of possible antibodies in the genus, and the Court rejected the argument that anyone skilled in making antibodies could find any member of the genus through routine screening. Practical impact: (1) functional genus antibody claims ('all antibodies that bind X and block Y') face severe enablement challenges without extensive structural characterization; (2) specifications must disclose a 'representative number' of structurally diverse examples covering the genus; (3) dependent claims with specific CDR sequences remain valid and valuable as fallback positions; (4) continuation strategies that add structural characterization as research progresses are more important than ever; (5) method of treatment claims using specifically described antibodies may provide protection without the genus enablement problem.

The written description requirement of 35 U.S.C. § 112(a) requires that the specification describe the claimed invention in sufficient detail that a person of ordinary skill in the art (POSITA) can reasonably conclude that the inventor had possession of the claimed invention at the time of filing. The Federal Circuit in Regents of the University of California v. Eli Lilly & Co. (119 F.3d 1559, Fed. Cir. 1997) established that for biological inventions, a claim to a genus of biological sequences or molecules requires adequate structural disclosure — the genus must be 'representatively' described by specific structural characteristics, not just by a functional goal. For DNA sequences: a claim to 'all cDNA sequences encoding insulin' requires structural characterization of the claimed sequences — it is not sufficient to describe a single rat insulin cDNA and then claim all mammalian insulin cDNAs, even if all mammals make insulin. For antibodies: the written description requirement (as applied alongside Amgen's enablement holding) requires structural disclosure of a representative number of species in the claimed genus. The USPTO's written description guidelines provide that the specification must 'draw a clear line around the claimed subject matter' — a functional definition without structural characterization does not draw that line. Practical requirements for biotech written description: (1) file a Sequence Listing with SEQ ID NOs for all claimed sequences; (2) characterize structural representatives across the genus (conserved motifs, variable region sequences, structural features); (3) include working examples demonstrating that diverse species within the genus can be made and used; (4) describe the structural boundaries of the claimed genus explicitly.

Yes — CRISPR-Cas9 and related gene editing inventions are patentable. The basic chemistry and biology of CRISPR-Cas9 was developed simultaneously by the Broad Institute (MIT/Harvard) and the University of California, Berkeley laboratory of Jennifer Doudna and Emmanuelle Charpentier. A landmark patent interference proceeding at the USPTO resolved in favor of the Broad Institute for CRISPR-Cas9 as applied in eukaryotic cells (Interference No. 106,048, final judgment affirmed by the Federal Circuit in 2022). UC Berkeley holds patents on the fundamental CRISPR-Cas9 system, and both institutions license to different entities in different fields. What can be patented around CRISPR: (1) specific guide RNA sequences (by SEQ ID NO) for defined genomic targets; (2) Cas9 variants and modifications with improved properties (high-fidelity Cas9, base editors, prime editors); (3) CRISPR delivery vehicles (viral vectors, lipid nanoparticles, RNP complexes); (4) therapeutic applications of CRISPR editing for specific diseases (ex vivo cell editing, in vivo gene correction); (5) CRISPR-based diagnostic platforms (e.g., SHERLOCK — Specific High Sensitivity Enzymatic Reporter UnLOCKing, developed by the Broad); (6) cell products produced by CRISPR editing. Important considerations: (a) the core CRISPR-Cas9 patent landscape is contested between Broad and UC Berkeley — any commercial use requires freedom-to-operate analysis and likely licensing from both portfolios; (b) CRISPR editing of human embryos and germline editing raise ethical and regulatory issues beyond patent law; (c) patent term — the earliest CRISPR patents filed in 2012–2013 will expire around 2032–2033, absent PTA or continuation extensions.

After Mayo Collaborative Services v. Prometheus Labs (566 U.S. 66, 2012), diagnostic method patents face serious § 101 eligibility challenges — but eligible diagnostic patents are still being granted. Mayo struck down claims that essentially claimed 'use the natural correlation between drug metabolite levels and therapeutic response to optimize dosing' — the Court held this was directed to the natural law itself, with no inventive concept beyond the law. To draft a surviving diagnostic claim after Mayo: (1) Anchor the claim to specific, concrete technical steps that go beyond mere observation of the natural correlation. The diagnostic method should specify the precise assay technology — for example, specific PCR primers (by SEQ ID NO), defined qPCR conditions, specific immunoassay antibodies and detection chemistry, or a defined next-generation sequencing panel. Generic steps like 'determine the level of biomarker X' are dangerous; concrete steps like 'performing targeted amplicon sequencing of genomic regions comprising variants [list] using primer pair [SEQ ID NOs] on a sample isolated from the patient' are stronger. (2) Include treatment steps that are integrated with the diagnostic. Claims that end with 'treat the patient' are more resistant than claims ending with 'correlate the level with disease risk.' A specific dosing algorithm or combination therapy regimen tied to the diagnostic finding goes beyond the natural correlation. (3) Claim the diagnostic kit as a composition — antibodies, probes, reagents in defined combinations are compositions of matter and are not subject to Mayo's natural law analysis. (4) Use the 2019 USPTO Revised Guidance Step 2A Prong 2 (practical application) to your advantage — claims that integrate the diagnostic into a larger technical process (an automated analysis system, a medical device that measures and displays a risk score, a specific software algorithm applied to genomic data) are stronger. (5) Pursue claims in multiple categories — diagnostic method (claim the assay process) + composition (claim the kit) + system (claim the automated platform) + treatment method (claim the treatment guided by the diagnostic). Even if one category faces § 101 challenges, others may survive.