The benefits and pitfalls of drug manufacturing IP (T 2543/22)

The Board of Appeal decision in T 2543/22 relates to a manufacturing method for preparing a therapeutic peptide. The Board of Appeal found that whilst various methods were known, the skilled person would not have had a reasonable expectation of success in applying them to produce the specific claimed peptide. The case highlights some interesting aspects of pharma IP strategy, particularly as related to manufacturing IP and whether this is best protected with patents or trade secrets. 

Benefits and pitfalls of drug manufacturing IP

Patents to the drug substance itself will always be the preferred IP strategy for protecting a therapeutic drug product. However, there are some cases where product IP is not possible, for example because the product is well known or difficult to define, or has limited useful patent term if clinical development has taken a long time. In these circumstances, manufacturing IP is another way of protecting return of investment in the drug. However, manufacturing patents bring with them a number of risks and potential pitfalls. 

In many cases, manufacturing IP may be easy for any potential competitor, generic or biosimilar company to design around. It is challenging for generics and biosimilars to design around a drug substance patent, given that regulatory approvals are tied to the drug substance itself. By contrast regulatory approvals will generally not specify that any one particular manufacturing process must be used to make the drug. Generic and biosimilar companies are thus free to develop their own manufacturing processes to avoid any manufacturing IP owned by the innovator company. 

Drug manufacturing

Added to the problem of the potential design-around by competitors is the issue of the limited term of protection afforded by patents. There are a number of drug products on the market for which the innovator has retained exclusively far longer than would have been possible with a patent, solely through the process of manufacturing trade secrets. One notable example is the hormonal replacement therapy (HRT) Premarin. Pfizer has retained market exclusivity for Premarin since its introduction in 1942 solely through retaining secrecy of its complex manufacturing process. If a patent had been filed for the manufacturing process it would have likely long expired, leaving generics to copy the method disclosed in the patent. Therefore, if manufacturing a drug is particularly hard, trade secrets may be a better approach for retaining market share without competition as they are not limited to a 20 year term like patents. 

Another problem with manufacturing patents is the challenge of detecting and proving infringement. Just as the innovator of a drug can keep their manufacturing process secret, so too can the competitors (see also IPKat on how this issue also applies to AI-assisted drug discovery). It may not be until litigation discovery that infringement can be proven. Without access to the competitor's facilities or detailed manufacturing records, gathering evidence of infringement or a manufacturing process could be extremely difficult. 

Amgen's therapeutic peptide manufacturing IP 

The patent at issue was EP3126373, owned by Amgen, covering a manufacturing method for Amgen's peptide drug AMG 416, marketed as Parsabiv for use in decreasing parathyroid hormone levels to treat various conditions. The manufacturing patent expires 4 years after the SPC for Parsabiv, based on the composition of matter (CoM) patent (EP2459208). Interestingly, unlike the opposition for Amgen's other patents relating to Parsabiv, the oppposition in this case was filed anonymously. The appeal was from the Opposition Division's decision to reject the opposition and maintain the patent as granted.

AMG 416 contains a disulfide bond between a D-cysteine and L-cysteine residue. The key feature of the claimed manufacturing method for AMG 416 was the formation of this disulfide bond between the specific cysteine residues. The claimed method particularly involved using an intermediate peptide containing a 2-pyridinesulfenyl (SPy) activating group on the D-cysteine residue. 

Inventive step of manufacturing 

The main issue on appeal was whether the use of SPy intermediate was obvious. The closest prior art identified for inventive step by the Board of Appeal was the composition of matter (CoM) case for AMG 416 (WO 2011/014707). CoM patent disclosed the structure and sequence of AMG 416, but only provided general information about peptide synthesis methods. 

On appeal, the Opponent argued that the invention was obvious because forming disulfide bonds using the methods described in the patent activation was well-known in peptide chemistry. The Opponent submitted that it would thus have been obvious to use these methods to prepare the intermediate specified in the claim and use this intermediate to prepare AMG 416. The Patentee argued in response that whilst individual elements or the method were known, the skilled person faced multiple choices in designing a synthetic route to AMG 416. 

The Board of Appeal agreed with the Patentee, and found that the skilled person would have needed to divert from the explicit teaching in the closet prior art to perform the entire synthesis process (r.6.3.1). The Board of Appeal concluded that starting from the closest prior art, "the skilled person would not have had a reasonable expectation of producing AMG 416 in an alternative way " (r. 6.4). The Board of Appeal concluded that the claimed method was not obvious in view of the closest prior art in combination with either common general knowledge or any of the other documents cited. The appeal was therefore dismissed and the patent maintained as granted. 

Final thoughts

Amgen was therefore successful in retaining its manufacturing patent for Parsabiv. However, the question remains whether generics will need to use the claimed process for manufacturing their own versions of Parsabiv, and whether it will be clear whether or not they have done so. It is notable that the opposition in this case was filed anonymously, indicating that whoever challenged the patent does not want to alert Amgen to the possibility that they are thinking of using the method. There is also the question of how easy the patent would be to design around. The patent indicates that the claimed invention improves the yield, which may be critical to commercial viability. However, there may be other methods for improving yield. Finally, it is interesting that Amgen chose to patent their manufacturing process in this case instead of preserving the method a trade secret. Possibly Amgen took that view that, whilst the method should be considered inventive by EPO standards, it was not so non-obvious that other companies would not think of using it themselves.  

Further reading

More on pharma IP strategy: 

• IP strategy for AI-assisted drug discovery (Oct 2024)
• Functional claims for pharmaceutical formulations: Validity versus enforcement (T 2130/22) (Jan 2025)
• Plausibility as a moving target: Phase III clinical trial results sink second medical use patent (T 0816/22) (Jan 2025)
• Insilico Medicine: Lessons in IP strategy from a front-runner in AI-drug discovery (Feb 2025)