Promosome v BioNTech
UPC_CFI_846/2024 UPC_CFI_485/2025 UPC_CFI_535/2025, LD Munich, Decision of 7 July 20261
The LD Munich has decided that EP 2401365, belonging to The Scripps Research Institute and licensed to Promosome LLC, is invalid. Furthermore, the LD Munich considered substantively the issue of infringement notwithstanding that the patent was held invalid, and also decided that the patent was not infringed by BioNTech’s Comirnaty product (a mRNA vaccine against COVID).
The defendants suggested that the UPC would not have competence for an infringement that was exclusively carried out before 1 June 2023. This was relevant as the original Comirnaty was sold only before 1 June 2023. However, in line with Court of Appeal caselaw2, the LD Munich held that the UPC is competent over alleged infringements before 1 June 2023.
In addition, the defendants challenged the standing to sue of Promosome LLC as exclusive licensee pursuant to Article 47(2) UPCA. The allegation that the licence agreement was invalidly executed was rejected following extensive evidence from the claimant about the signatories of the document. Moreover, the argument that the agreement was not an exclusive licence was rejected – an exclusive licence can be limited in terms of geographical scope, field of use or duration.
The claimed invention was to a method of improving full-length protein expression efficiency by mutating secondary initiation codons located within the polynucleotide coding sequence in a manner such that the amino acid sequence of the protein remains unaltered. The secondary initiation codons are sequences that can act as a start codon occurring within the coding sequence, resulting in expression of a part protein and reducing the expression of the full protein. The claim required that the mutation results in a decrease in initiation of protein synthesis at the one or more secondary initiation codons, resulting in a reduction of ribosomal diversion away from the primary initiation codon, thereby increasing expression efficiency of the full-length protein initiated at the primary initiation codon.
Accordingly, the court was of the view that the mutation had to be made at a site that did actually act as a secondary initiation codon; it was not sufficient that the site could be a secondary initiation codon because it had a sequence that could act as a start codon.
It was under this functional understanding of the claimed invention that the court decided that infringement had not been made out. It was undisputed that in the manufacture of the Comirnaty product the wild-type sequence was substantially mutated to produce the Comirnaty mRNA sequence. This included many potential secondary initiation codons. However, the defendants argued that the changes were made as a part of routine codon optimisation (where a codon is changed to another codon corresponding to the same amino acid, but for which more matching complementary tRNA is present in the host organism). The court considered that the evidence did not establish that any of the potential start codons did actually act as a secondary initiation codon.
The court held that the patent lacked novelty over two prior art documents.
D2 (Matsuda et al) describes mutation of a second initiation codon in a bicistronic mRNA derived from Turnip Yellow Mosaic Virus (TYMV) genomic RNA. A bicistronic nucleic acid encodes two distinct proteins, each of which can be translated using different (but overlapping) open reading frames for translation and each of which has its own initiation codon. The counterdefendants argued that this should be regarded as a second primary initiation codon, and not a secondary initiation codon according to the claims. However, the court rejected this argument, holding that the bicistronic scenario fell within the scope of the claims.
D1 discloses the claimed subject matter, but the main argument was that the disclosure was not enabled. The court accepted that a non-enabled disclosure would be not be novelty destroying, but concluded that all of the features of the method claimed in claim 1 of the Patent are disclosed in D1 in an enabling way.
A number of auxiliary requests were presented but none of these was found patentable either. One amendment (AR4) additionally specified further mutations of one or more miRNA binding sites – this technique was known from D9 and D10. The court considered but did not decide whether the EPO “partial problem” approach should be applied in this situation. In any case the court considered that the alleged invention constituted the mere aggregation or juxtaposition of two known techniques and saw no suggestion in the patent that the two techniques were in any way functionally connected or interdependent. Accordingly, this formulation of the invention was held to lack inventive step.
The patent was held invalid and ordered to be revoked. Unusually, the UPC nevertheless conducted a full analysis of the infringement situation based on the patent claims as granted, and concluded that in any case infringement had not been substantiated.