UPC Court of Appeal overturns preliminary injunction in 10x Genomics

Introduction - a case of many firsts

The UPC's first preliminary injunction (PI) in inter partes proceedings, sought on the first day of operation of the UPC, 1 June 2023, before the grant of the first Unitary Patent to be asserted at the UPC.

NanoString was the defendant, and unfortunately for them, after a 2-day hearing, the Munich Local Division found 10x Genomic's case to be more convincing. On 19 September 2023, about 2 weeks after the hearing, via a fully reasoned decision of more than 100 pages, a PI over the whole UPC territory was granted, less than 4 months after being requested.

Unsurprisingly NanoString appealed, which led to the first hearing at the UPC's Court of Appeal (UPCA) in Luxembourg, on 18 December 2023. About 2 months later, on 26 February 2024, the UPCA issued its first substantive decision, overturning the Munich Local Division. So good news for NanoString, were it not for the fact that it filed for Chapter 11 bankruptcy in the United States a few weeks earlier. The first defendant to succumb to the powerful scope and jurisdiction of the UPC?

Key outcomes and observations

In terms of how the UPC is settling in the European patent litigation landscape and how its positions are aligning with national courts and/or the European Patent Office (EPO), there are two particularly significant outcomes concerning claim interpretation and its validity standard. We cover these first, before turning to some general observations and looking to the future.

1. A firm adoption of Art. 69 EPC for claim interpretation

The UPC will apply the principles of Art. 69 of the European Patent Convention (EPC) and the Protocol on its interpretation - equally to the assessment of the infringement and validity. The UPCA lays these principles out fully in the decision, without adjustment (at least once Headnote 2 was corrected on 11 March 3034, due to a mistaken "but" that is now an "and").

But using the same principles, thereby showing their flexibility, the UPCA arrives at a broader claim interpretation than the Munich Local Division. This marks the beginning of the divergence of the UPCA from the first instance decision.

The UPC's adopting of Art. 69 and the Protocol is no different from national court proceedings whether validity and infringement are assessed together or separate (the latter at least based on the constant case law of the German Federal Supreme court, BGH). The comparison to the EPO however – which of course hears only validity – is of greater interest, as the EPO's application of Art. 69 EPC in examination and opposition proceedings is less clear.

There is EPO caselaw saying that the description should be used to interpret the claims, but there is also caselaw emphasizing that Art. 69 EPC and its Protocol are primarily for use by the judicial organs which deal with infringement cases. According to the latter, the wording of the claims should typically be given its broadest technically sensible meaning.

The UPC's position is not unexpected, but quickly highlights not only the divergence within EPO caselaw, but also potentially between the EPO and the UPC. It would seem undesirable to have the two bodies deciding on the validity of European patents at a supranational level to be applying different claim interpretation tests. While a goal of the UPC is to harmonize and replace the need to litigate national patent courts, this raises questions over whether the UPC will behave differently to the EPO on such a fundamental issue.

The EPO may provide clarity on their position in the near future, since it has now been reported that, in a pending appeal case, a referral has just been made to the Enlarged Board of Appeal on whether the description should be used to interpret the claims. This may be an opportunity for the EPO to confirm whether it will align its future caselaw with the UPC.

2. Validity standard for PI – "more likely than not" or "predominantly probable"

Like any supra-national organisation, language has been a persistent and key feature of the UPC, and has already been heavily debated in a number of interim decisions. The language of these proceedings, including the first instance hearing was German. But the parties agreed that the appeal could be heard in English. The decision was issued in both German and English. It is pertinent therefore that potentially the most interesting issue is a point of language and its impact on understanding the validity standard applied to PI.

The key issue is the interpretation and application of R. 211.2 of the Rules of Procedure (RoP):

"In taking its decision the Court may require the applicant to provide reasonable evidence to satisfy the Court with a sufficient degree of certainty that the applicant is entitled to commence proceedings pursuant to Article 47 that the patent in question is valid and that his right is being infringed, or that such infringement is imminent."

This RoP is critical given there is no express mention of validity in the UPC Agreement, and in a slight change from first instance (which considered validity as a requisite of standing and patent ownership), the UPCA confirmed the intertwinement of validity and infringement – i.e. to be infringed, a patent must be valid. The key questions are of course: how infringed and how (in)valid must the patent be so to achieve the sufficient degree of certainty? And are the standards for each independent or related?

The UPCA explains that in the context of PI proceedings, the standard of proof should not be set too high, but neither too low. The burden of proof is upon the applicant when it comes to the entitlement to initiate proceedings and the infringement of the patent and upon the defendant, except in ex parte proceedings, for lack of validity. There is little controversy here.

The UPCA moves on to interpret the wording "sufficient degree of certainty" as used in RoP 211.2. According to the English version of the appeal decision, it requires that the court considers it at least more likely than not that the applicant is entitled to initiate proceedings and that the patent is infringed. A sufficient degree of certainty is lacking if the court considers it on the balance of probabilities to be more likely than not that the patent is not valid. Hence, "more likely than not" appears to be the criterion.

However, there is a degree of uncertainty versus the German version of the appeal decision, which uses the phrase "überwiegend wahrscheinlich". In some translations, this has reported as being "predominantly probable", which suggests a higher bar must be met – for both infringement and invalidity. Given the outcome in the appeal, the question is therefore whether NanoString won as they proved that it is "predominantly probable" or "more likely than not" that the patent is invalid, or what standard is required for the next defendant.

Given the "more likely than not" wording in the English version of the decision, and the later references in the decision saying that it must therefore be "wahrscheinlicher", hence more likely, that the patent is valid than not valid, it appears that the terms "überwiegend wahrscheinlich” as used by the court indeed was meant to be indicate "more likely than not".

In any event, contrary to the first instance court, the UPCA considered it to be more likely than not that the patent is not valid. The UPCA agreed with the court of first instance that the patent is novel, but considered it to be more likely than not that the patent will prove to lack an inventive step. As a result of the broader claim interpretation, the UPCA identified fewer distinguishing features over the prior art (only one). According to the UPCA, the patent is considered to detect a plurality of analytes in a cell or tissue sample, whereas the main piece of prior art, referred to as D6 in the judgment, is intended to detect ASMs (amplified DNA molecules) and thus a plurality of analytes, but these are not present in a cell or tissue sample. The UPCA then, expressly referring to the technically qualified judges, deemed it on the balance of probabilities more likely than not that it will prove to be obvious for the skilled person to transfer the known in vitro detection method of D6 to the detection of ASMs in the cell or tissue samples, i.e. to an in situ environment, thereby arriving at the subject-matter claimed by the patent.

In another point of distinction on substantive law (especially of the EPO), the UPCA does not clearly indicate how they assessed inventive step, in particular not following the framework of the EPO's problem-solution approach. They did identify a closest prior art, without naming D6 as such, and then considered whether the prior art contained an incentive or confirmation for thinking in the direction of the patent. But no objective technical problem was formulated. Also in national jurisdictions, the problem-solution approach is not mandatory, although it is often used as a helpful tool to avoid using an impermissible hindsight reasoning. This may be a deliberate attempt by the UPCA not to set too many influential guidelines on such fundamental issues in this early proceeding. There is no doubt as to why and on what basis they reached their decision, and the decision at times makes it sound very heavily weighted to the skilled person obviously arriving at the invention.

General observations

Further to the key issues above, we would also extracted the following points:

• A mini-trial and a do-over - as expected, UPC appeals are a do-over. Here, the UPCA identified no legal error at first instance. Rather, it reassessed the claim construction, the impact of that claim construction on the distinguishing features over the prior art and the (lack of) inventive step. In doing so, it came to a different outcome and overturned the PI. Implicitly this teaches users that the UPCA will revisit everything in detail, albeit on a preliminary basis. This also follows from the appeal hearing in Luxembourg where more issues were discussed than addressed in the appeal judgment.

• Little further guidance - unlike the first instance decision, the UPCA did not seem to view this as an opportunity to provide guidance on the methodology the UPC would apply when assessing whether or not to grant a PI. This might have helped shape UPC caselaw and avoid some uncertainties in the minds of parties hesitating to use the new court. So after having read this decision, some may be somewhat disappointed. It is only 37 pages (with the assessment less than half of it). It deals only briefly with jurisdiction and entitlement, making it difficult to interpret their considerations without having access to the parties' pleadings. On other issues, such as urgency and weighing of the parties' interests, there are no comments at all. Overall, we only know why this PI was denied and not much more about the general requirements to achieve one.

• Technically qualified judges to the fore? - it is hard not to notice the express reference in the judgment to the expertise of the technically qualified judges when the validity of the patent is assessed. This suggests that the 2 technically qualified judges (TQJs) sitting on the 5-judges panel, while there was only 1 TQJ in the first instance panel, will have had a strong say in the outcome on validity. This also affirms that the TQJs are judges and that their roles encompass more than merely assisting the legally qualified judges in making the technology more accessible. The TQJ's own technical expertise is evidently considered, which is notable for lawyers coming from jurisdictions with passive judges. The 2 TQJs in the case were Cornelis Schüller and Rainer Friedrich. According to their bios on the UPC website, Cornelis Schüller is a substitute judge at the district court of the Hague, having master degrees in biochemistry and chemical engineering and a PhD in molecular biology and Rainer Friedrich holds a doctorate degree in biochemistry. They thus both have the expertise to understand the technical issues at stake in this case in the field of biotechnology.

• Influence of expert evidence? - in a case that seemingly hinges so heavily on whether or not the patent is inventive, it is somewhat surprising - at least to the lawyers of some jurisdictions - that little reference is made to expert evidence in the judgment. In the section on inventive step, there is no reference to an expert. There is only a reference to a PRV Consulting Report of the Swedish Intellectual Property Office. It is hard to tell at this stage how much weight the UPC will put on expert evidence, but at least based on this case, it seems that the UPC may trust to a large extent in the expertise of their TQJs when it comes to the technical merits of the case.

• Appeal court slower? - it took the judges about 2 months to issue the relatively short judgment. In total, the appeal took about 5 months, whereas the first instance proceedings were done in less than 4 months. This brings the entire duration of the proceedings close to 9 months. While not especially speedy, it is faster than in many EU countries. It remains to be seen, however, how the UPCA will keep delivering on speed, given that their workload will only increase.

Looking ahead - PI in the UPC or national courts?

The key outcome in this decision is that a request for a PI will be denied, if on the balance of probabilities it is more likely than not that the patent is not valid. To compare this to different national courts in Europe:

• Netherlands: a PI will be denied if there is a "serious, non-negligible" chance that the patent will be found invalid in main proceedings. The words used to formulate the tests are obviously not the same. Arguably the bar to deny a PI in the Netherlands based on the validity of the patent is somewhat lower, meaning the UPC could be considered to be more patentee friendly. However, looking at the reasoning in the appeal judgment, it seems that the approach taken by the UPCA is not so different from the approach the Dutch courts take in practice. In both courts, there is an in-depth assessment of the validity arguments, albeit preliminary, and if the court is not sufficiently convinced that the patent is valid, it will not grant a PI. Also, in both courts, the appeal court reassesses the evidence, facts and forms its own view on validity. This makes it worthwhile for the party negatively affected by a PI decision to consider appealing it. Overall the assessment whether or not to grant a PI seems thus to be quite similar between the UPC and the Netherlands.

• Germany: the test for the validity of a patent in PI proceedings in Germany diverges slightly between the courts. As PI proceedings are decided in last instance in second instance with no possibilities of a final appeal to the German Federal Supreme Court, different case law has been developed by the appeal courts and lower courts over time. The Düsseldorf appeal court has formulated the requirement that the validity of a patent must be confirmed in contradictory proceedings before it is suitable for a PI, although quite a number of exceptions have been formulated(for example generic market entries in pharmaceutical cases or trade fair matters). The Munich appeal court initially had been more lenient, but some years ago followed the principles of the Düsseldorf appeal court. The first instance courts in Munich, however, have tried to maintain a more patentee friendly position and review in each individual case the validity of the patent. If they think that the patent likely is valid they may grant a PI subject to the other requirements for PI. This discrepancy in fact let to the CJEU referral in the Phoenix case. In practice these difference play a more limited role in pharmaceutical disputes where the appeals also apply exceptions. The topic nevertheless remains controversial, in particular as there will be no final ruling by the German Federal Supreme Court.

• Italy: in the context of Italian PI proceedings for patent infringement, two primary requirements must be met. The first is the so called "fumus boni iuris" or "likelihood of success on the merit of the case", which refers to the appearance or likelihood that the patent is valid and has been infringed by the defendant. The second is the "periculum in mora" or "danger in delay", which refers to the fact that the timings of proceedings on the merits would cause irreparable harm to the patent holder. If the defendant does not contest the validity of the patent at issue, the Court will focus solely on the alleged infringement and the patent will be presumed valid. If, conversely, the defendant challenges the validity, the Court will also assess the invalidity arguments, albeit preliminary. In order to ascertain whether the patent is valid and infringed, an Italian Specialized IP Court usually appoints a Court Technical Expert, who will have to establish whether the patent satisfies the patentability requirements challenged by the defendant and whether it is infringed. Even though we are in preliminary injunction proceedings, which are therefore characterized by summary cognition, the expert's opinion will still be sufficiently detailed so that ultimately the Court assessment will be relatively thorough. After this technical expertise, an injunction will be granted if the Court deems it more likely than not that the patent is valid. Therefore, in general, the criteria for granting a preliminary injunction appear to be quite comparable between the UPC and Italian judicial practice.

• France: in PI proceedings, the patentee must demonstrate that the patent is likely to be infringed and that the patent is not invalid prima facia. On validity, the judge will assess whether the invalidity argument are serious enough, in which case the PI would be rejected, or not serious enough (in which case the PI will be granted). In theory, the standard is lower than on the merits for both validity and likelihood of infringement. In practice the discussion will often be quite extensive, but the PI order will not take a definitive position on either question. Although the "more likely than not" criteria is not used as such by French courts, the approach taken by the UPC is not so different from the approach taken by the French courts on the assessment of validity in the context of PI proceedings.

• UK: used to being an outlier compared to mainland Europe and this looks set to continue versus the UPC. In the UK, the merits of validity and infringement barely feature in the PI assessment, with the applicant only needing to achieve an arguable case and the defendant needing to satisfy a high threshold that the patentee has "no real prospect of success" (i.e. summary judgment) to defeat a PI. This typically confines validity considerations at PI to "slam dunk" arguments or novelty or added matter, which rarely succeed. Significantly more emphasis is placed on the need for irreparable harm and balance of convenience, into which this decision not does venture (and while other UPC decisions have, they have not done so to the same extent as in the UK).

Conclusion

A key takeaway from this case is the need to be prepared to attack/defend patent validity. The extent to which the UPC framework will allow mini-trials to expand in this respect (or as to infringement) remains to be seen, in particular as to any in-depth review (or perhaps even cross-examination) of expert evidence. Notably, the first instance hearing in this case lasted longer than 1 day, whereas some of the recent interim conference decisions at first instance have remarked as to the strict adherence to 1 day. If PI proceedings were to expand and continue to achieve outcomes within 9m, this raises the typical queries as to the ensuing relevance of full proceedings. In this respect, it is notable that 10x Genomics' main proceedings are progressing slowly and its opponent has already become insolvent based on a PI decision, albeit overturned. This illustrates the power of UPC PI proceedings.

As to any deterrent for patentees seeking to utilise this power, it will also be interesting to see whether NanoString will now seek compensation, as R. 213(2) RoP appears to allow for a compensation claim based on a PI revocation (apparently independent of whether the patent is subsequently found invalid). Given it has been reported that the PI was one of the reasons cited in NanoString's filing for bankruptcy, the PI (on which NanoString has also been ordered to pay penalties for breach) seems to have caused NanoString significant damage. In any event, it is certain that this dispute between 10x Genomics and NanoString is not over, and will be closely watched by many in the future, including to see the impact of this outcome on the main infringement actions and 10x Genomic's parallel proceeding against other defendants.