- within International Law topic(s)
Patentees generally want to encompass both their core invention and a small amount of white space around it. In some fields, this is relatively straightforward. However, in biologics, applicants at the EPO are often forced into claims that recite a specific sequence ID which, arguably, does not encompass similar sequences. This has been the case irrespective of whether the product is an antibody, other protein therapy, nucleic acid, or a combination of the above. In the recent Board of Appeal Decision T 0137/24, the Board has confirmed that functional language can be used as a tool for obtaining breadth in how sequences are defined.
Sequence Breadth is Commercially Valuable
Unlike small molecules, for which generics are typically looking to produce identical compounds to those approved, biosimilars – which, as the name suggests, are similar but not identical – represent a significant threat to market share. Therefore, claims reciting a single protein (e.g., antibody) or nucleic acid sequence may not encompass an otherwise potentially infringing product.
This is also true for the eventual authorized product: during development, sequence modifications may be identified that achieve the same functional result. In addition, seemingly small modifications might increase stability and other important factors that assist in formulation or storage of the therapeutic.
There is also redundancy in our genetic code: in many instances, single nucleic acid modifications will not alter the amino acid that is encoded. Similarly, amino acids are grouped into structurally similar subsets, and conservative replacements within these groups are less likely to disrupt protein function.
For these reasons, small structural modifications that can still result in the same function may be considered to fall outside of the scope of a claim that recites a specific sequence. Yet, it is generally acknowledged by patent offices and courts that protection for such similar molecules should be permitted.
The Challenges With Claiming Sequence Similarity
In the biologics space, claiming sequence similarity is therefore highly desirable and widely attempted. For instance, rather than simply claiming “a polypeptide comprising SEQ ID NO: 1” an application may instead specify “a polypeptide comprising a sequence having at least 75%, 80%, 85%, 90%, 95%, 99%, or 100% identity with SEQ ID NO: 1.”
However, the EPO often objects to this language under the grounds of inventive step and/or sufficiency.
Under inventive step, a typical objection is that the technical effect cannot be acknowledged across the claim scope, because a sequence sharing only, for example, 75%, sequence identity with the exemplified sequence would not be expected to act in the same manner and is unlikely to solve the same technical problem. Based on this, examining divisions will often formulate the objective technical problem as the provision of an alternative. For example, rather than formulating the problem to be solved as an improvement over the closest prior art, it will be formulated as the mere provision of an alternative, which lowers the bar, making a finding of obviousness more likely.
The invention must also be sufficiently clear and complete for it to be carried out by the skilled person (this is called “sufficiency” at the EPO, and is similar to the written description requirement at the USPTO). Examining divisions may object that claims reciting broad sequence identity place an undue burden on the skilled person to identify working embodiments across the claim scope.
Addressing Objections to Sequence Identity Language
Increasing the percentage identity required can sometimes address the examining division’s objections. However, this is rarely successful in situations where there are non-working embodiments having high sequence identity. In such cases, an alternative strategy is typically needed.
One that is often attempted to obtain claims that permit a broader sequence definition is to include functional features within the claim. For example, following on from our earlier example, a claim may recite “a polypeptide comprising a sequence having at least 75% identity with SEQ ID NO: 1, wherein the polypeptide inhibits X”. However, examining divisions at the EPO may still raise the inventive step and/or sufficiency objections discussed above.
The Technical Board of Appeal’s (3.3.08) decision in T 0137/24 clearly set out that functional language can be an effective means to exclude non-working embodiments from the claim scope, and that this strategy can address those inventive step and sufficiency objections.
The patent in question was directed to genetically modified yeast cells for producing a cannabinoid or a cannabinoid derivative, where the yeast encodes a specified enzyme polypeptide. Claim 1 on appeal required the enzyme polypeptide to have at least 90% sequence identity to either of two specified sequences. The Board also noted that the claim required enzymatic activity.
The Board concluded that because the claim required enzymatic activity (i.e., functional language), it did not cover non-working embodiments that do not achieve the specified function. Therefore, functional language can address the inventive step objection that the claimed sequences do not achieve any purported function across their claim scope, because the claim only encompasses sequences that achieve the function. For instance, the example provided above would be limited to polypeptides that (a) share at least 75% sequence identity with SEQ ID NO: 1, and (b) inhibit X: any polypeptides that did not inhibit X would not fall within the claim scope.
The appellant-opponents argued that identifying whether sequence changes have an impact on the enzymatic activity represents an undue burden for the skilled person. Nevertheless, the Board found that the sufficiency requirement was met, and there is no undue burden because the patent provided techniques for testing the enzymatic activity. More generally, the Board set out that if testing the functional feature “only require[s] the use of techniques that are routine in the technical field,” then the skilled person is not faced with an undue burden, and the sufficiency requirement is met. What constitutes routine techniques will be case-specific, but this provides a clear strategy to address sufficiency objections to sequence identity language. For instance, evidence of a widely known lab-bench test for identifying inhibition of X would typically result in a finding that the language provided in our example is sufficient.
The opposed patent in T 0137/24 recited 90% sequence identity, but the Board’s reasoning was not limited to this percentage: it may be possible to apply the same reasoning to obtain broader sequence definitions by adding functional language.
Asserting Patents Against Biosimilars in National Courts and at the UPC
Even if the sequence definition in your claims was limited during EPO proceedings it may still be possible to assert infringement: the doctrine of equivalents may provide an opportunity to allege that molecules that do not fall within the literal scope of the claims infringe by equivalence.
A clear test for the doctrine of equivalents was set out by the UK Supreme court in 2017 (Actavis v Eli Lilly [2017] UKSC 48), and a number of Local Division decisions have shown that a doctrine of equivalence also exists at the UPC (see, for example, Plant-e v Arkyne, which we have previously reported on here).
Concluding Remarks
When drafting applications where a broader sequence definition may be of value, introducing functional features in dependent claims provides a clear strategy for reducing the likelihood that the sequence identity will be limited to 100%. In addition, it is always advisable to ensure the specification sets out how any functional features would be assessed to defend against an insufficiency attack.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
[View Source]
