Background
Schering Corporation held foundational patents on recombinant DNA methods for producing human leukocyte interferon-alpha subtypes — naturally occurring antiviral proteins used as therapeutics. These patents covered processes for using recombinant DNA to produce specific natural interferon-alpha subtypes (such as IFN-α2a and IFN-α2b), which Schering marketed as Intron A. The patents claimed the recombinant production process for interferons having amino acid sequences identical or substantially identical to specific natural subtypes.
Amgen developed INFERGEN (interferon alfacon-1) — a synthetic “consensus” interferon. Consensus interferon is not found in nature; it was engineered by identifying the most common amino acid at each position across all known human interferon-alpha subtypes and combining those positions into a single optimized synthetic protein. The result was a protein more potent than any natural subtype but structurally distinct from all of them. Schering sued Amgen, arguing INFERGEN’s recombinant production infringed its patents. The district court found no infringement. Schering appealed.
The Court’s Holding
The Federal Circuit affirmed. Writing for the court, Judge Gajarsa held that Schering’s recombinant DNA patents did not cover Amgen’s production of consensus interferon because the claims were limited to producing proteins with sequences that duplicate natural interferon subtypes. Consensus interferon is a synthetic construct that duplicates no natural subtype; it was designed by a computational averaging process to be an optimal synthesis across subtypes, not a recombinant copy of any specific natural protein.
The court rejected Schering’s argument that “substantially identical” in its claims should be read broadly enough to encompass consensus interferon. Under proper claim construction, “substantially identical” referred to sequence identity with a natural subtype — not to functional similarity or overall potency. Consensus interferon was structurally and sequentially distinct from any natural subtype, even if it was broadly related to the interferon family. The court also addressed and rejected Schering’s doctrine of equivalents arguments, finding that the function-way-result test was not satisfied because the way consensus interferon was made (by computational optimization of a synthetic sequence) was fundamentally different from the patented method of recombinantly producing a naturally occurring subtype.
Key Takeaways
- Recombinant DNA patents covering the production of specific natural protein subtypes do not automatically extend to cover synthetic “consensus” or engineered proteins designed to outperform any natural variant.
- The term “substantially identical” in biological patents is construed based on the patent’s definition and context — typically requiring close sequence identity to the named natural counterpart, not merely functional equivalence.
- Engineered proteins optimized beyond any natural template can avoid infringement of patents directed to naturally occurring sequences even when operating in the same biological family.
- The case illustrates how structural claim drafting in biologic patents — focused on specific natural sequences — can leave gaps for engineered or synthetic alternatives.
- Doctrine of equivalents does not rescue a patent where the accused product uses a meaningfully different structural and manufacturing approach, even if the end result is similar therapeutic activity.
Why It Matters
Schering v. Amgen is an important decision for the pharmaceutical and biotechnology industry, illustrating the limits of foundational recombinant DNA patents when applied to later-generation engineered proteins. The interferons were among the first class of recombinant biologics approved for human therapy, and their patents represented some of the most commercially significant IP of the 1980s and 1990s. As protein engineering advanced, companies like Amgen began designing synthetic proteins that were superior to natural sequences but structurally distinct.
The Federal Circuit’s ruling confirmed that patent protection for natural proteins does not automatically follow engineered analogs, even when those analogs are designed with full knowledge of the natural proteins and serve the same therapeutic purpose. For biotech companies, the decision highlighted both an opportunity — engineering around natural-sequence patents by designing optimized synthetic variants — and a drafting lesson: foundational protein patents should include broader claims encompassing engineered variants if the patent holder wants protection against rational protein design workarounds.