The Supreme Court held that Regeneron could not patent a whole range of genetically engineered mice when its patents taught a skilled person how to make only some of them. The mice were intended as platforms for developing antibodies for human treatment. For a product claim, the disclosure must enable substantially all the products within the claim; a groundbreaking idea that would benefit products which cannot yet be made is not enough.
The patents and the claimed mice
Regeneron owned two European patents (UK), the 287 Patent and its divisional, the 163 Patent, both with a priority date of 16 February 2001. It alleged that Kymab infringed claim 1 of the 163 Patent and claims 5 and 6 of the 287 Patent by offering its Kymouse to the pharmaceutical industry. Henry Carr J found infringement proved, but held all three claims invalid for insufficiency. It was common ground that their validity turned on claim 1 of the 163 Patent.
Law Teacher writing support
Writing a law essay?
We choose a writer who is an expert in your area of law and qualified to at least the academic level you select. Most of our writers hold a master’s degree or PhD, have teaching experience, or both.
Send us your question and, if you wish, any marking criteria. We’ll write an essay to your instructions.
That claim covered a transgenic mouse producing hybrid antibodies with human variable regions and mouse constant regions. Mouse variable-region gene segments were to be replaced in situ by the corresponding human segments in both the heavy-chain and light-chain antibody genes. The courts below held that this meant any or all of those segments, so the claim covered a range rather than one mouse. At one end was a mouse with very few human segments; at the other, a mouse carrying the whole human variable region. A typical human heavy-chain variable region has about 125 different V segments. That construction was not appealed.
The scientific problem
By the priority date, antibodies were recognised as useful in treating human disease, and mice were already used to develop them. Humans tend to reject mouse antibodies. If fully human antibody genes are implanted instead, the mouse’s immune response is weakened. The mouse becomes, in the language of the case, immunologically sick and a poor platform for discovering antibodies.
The inventive idea was a hybrid structure called a Reverse Chimeric Locus: the mouse retains its own constant-region genes, while human variable-region DNA is inserted. Hybrid antibodies can later have their mouse constant regions replaced with human ones before use in people. The greater the number of human variable segments, the wider the variety of antibodies the mouse was understood to be able to produce. That variety matters because a wider range of antibody-producing cells improves the chance of finding one that binds a particular disease-related antigen.
What the disclosure enabled
Henry Carr J found that the patents, together with existing common general knowledge, did not enable any mouse in the claimed range to be made. The Court of Appeal disagreed on the facts. It held that a skilled team could make mice containing only a small subset of the human V segments — on Kymab’s estimate, between two and six — and an unspecified number of human D and J segments. Transplantation of the whole human variable region was achieved later, but only through further inventive processes outside the patents and the prior art. There was no appeal against that finding.
The legal question was therefore whether a product patent passes the sufficiency test when it enables only some products in the claim, but the invention would contribute to the usefulness of every product in the range if and when it could be made.
Sufficiency across a relevant range
Sufficiency, in article 83 of the European Patent Convention and section 72(1)(c) of the Patents Act 1977, requires the invention to be disclosed clearly and completely enough for a skilled person to perform it. Lord Briggs, with whom Lord Reed, Lord Hodge and Lord Sales agreed, explained the underlying bargain. The patentee receives a time-limited monopoly in return for disclosure that lets the public work the invention after the monopoly expires. For a product claim, that means making the product. The skilled person may use the patent, existing knowledge and reasonable experimentation, but is not expected to invent.
The disclosure must enable substantially all the types of product within the claim. A tiny or irrelevant shortfall does not matter. Lord Briggs contrasted mice of different colours or tail lengths: failure to teach how to make a short-tailed mouse would not invalidate the claim, because tail length does not affect the antibodies the mouse exists to produce. A range is relevant where the variable defining it significantly affects the product’s value or usefulness for its purpose.
The quantity of human variable-region DNA was relevant. At the priority date it was understood to affect the mouse’s usefulness for therapeutic antibody discovery, and the patents did not teach otherwise. Keeping the mouse constant region addressed immunological sickness across the whole range, but immunological health was a means of producing useful antibodies, not an end in itself. The most valuable mice were those containing more, ultimately all, of the human variable region — the very mice the patents did not enable.
Why the Court of Appeal’s approach failed
The Court of Appeal accepted the general rule, but treated the Reverse Chimeric Locus as a groundbreaking principle of general application. Every mouse in the range would share its benefit once that mouse could be made. Limiting protection to mice that could be made immediately would, in a fast-moving field, give a short-lived reward which did not match the contribution to the art. It was enough that at least one type could be made at the priority date.
The majority held that this was neither the existing law nor a legitimate development of it. The contribution of a product claim is the product which the skilled person can make, not the bare idea. A general principle supports a wide claim only if it is reasonably likely to enable the whole relevant range to be made. It is not enough that every product would deliver the intended benefit if and when it could be made, however valuable the idea.
Lord Briggs distinguished cases where a disclosed principle would work, without further invention, with input materials not yet available. Here the human variable region had already been mapped. It could be inserted into mice, but with human constant-region genes, causing immunological sickness. The gap was the inability to create a Reverse Chimeric Locus containing more than a very small part of the human variable region. That was a shortcoming in the invention. Mice containing the whole region, described by Lord Briggs as the gold standard, depended on further and different inventions made by each company years after the priority date.
The majority accepted that this may, with hindsight, give the inventor of a groundbreaking idea a scant reward. Sufficiency across the claim was nevertheless settled law and part of the balance between patentee and public. Watering it down would favour patentees beyond the Convention and beyond proper judicial development in one Convention state.
The dissent
Lady Black would have dismissed the appeal. She did not accept that the Court of Appeal had misstated or misapplied the law. She regarded the claim as a principle of general application: every mouse within it uses the Reverse Chimeric Locus and gains the solution to immunological sickness, regardless of how much human material it contains. On that view, the invention was enabled across its scope, and protection across the range matched the patents’ technical contribution.
Verdict: The Supreme Court allowed Kymab’s appeal by four votes to one. The claims in issue were invalid for insufficiency: the disclosure did not enable substantially the whole relevant range of mice at the priority date. Lady Black dissented.
Source: Regeneron Pharmaceuticals Inc v Kymab Ltd [2020] UKSC 27
