Global menu

Our global pages

Close

Royalties under patent licence agreements: can they remain payable even if the licensed patents(s) are revoked or not practised? Lessons from Genentech Inc v Hoechst GmbH & Sanofi-Aventis Deutschland GmbH

  • United Kingdom
  • Health and life sciences

08-12-2016

Précis

In its recent decision in Genentech Inc v Hoechst GmbH & Sanofi-Aventis Deutschland GmbH (7 July 2016), the Court of Justice of the EU (“CJEU”) has ruled that an obligation on a licensee under a patent licence agreement to pay royalties on products which do not fall within the scope of the licensed patent claims, or where the obligation to pay royalties continues for the entire term of the agreement regardless of the earlier revocation of the licensed patents, does not in itself amount to a breach of EU anti-trust law (Article 101 Treaty on the Functioning of the EU) and hence is enforceable against the licensee, provided that the licensee can terminate the licence agreement on reasonable notice.

This decision largely confirms existing EU law. However, valuable lessons on the drafting of patent licences can be drawn from the context for the decision. We highlight these below.

Background

On 6 August 1992, Behringwerke AG (the predecessor of Hoechst GmbH) granted a worldwide non-exclusive licence to Genentech for the use of a human cytomegalovirus (“HCMV”) enhancer, with effect as of 1 January 1991 (the “Licence Agreement”). Enhancers are discrete segments of DNA capable of enhancing the expression of one or more functionally associated genes by upregulating transcription (the process of synthesizing RNA from a DNA template). They are used in the biotechnology sector to boost production efficiency for protein-based products. For example, by linking an enhancer to a gene encoding a biological drug, bioreactor yields of that drug from cells expressing that gene can be improved.

Genentech used the HCMV enhancer in this way in the production of its monoclonal antibody rituximab, which forms the active ingredient in its blockbuster biological medicinal product MabThera™. This is a cancer drug for the treatment of common forms of blood cancer, including non-Hodgkin's lymphoma (NHL), follicular lymphoma and chronic lymphocytic leukaemia (CLL). It is also used to treat rheumatoid arthritis and certain types of vasculitis. MabThera™ is marketed as Rituxan™ outside Europe, and had global sales in 2015 of $7.2bn.

At the time of the Licence Agreement, the HCMV enhancer technology was the subject of European Patent EP 1,731,753 (“EP ‘753”), issued on 22 April 1992, as well pending US patent applications which later issued as US 5,849,522 (on 15 December 1998) and US 6,218,140 (on 17 April 2001). The patents claim methods of using the HCMV enhancer to increase expression of a gene in a mammalian cell, isolated HCMV enhancers, plasmid DNAs comprising an ECMV enhancer operatively linked to a heterologous gene, and eukarytoic host cells transformed with such plasmids.

EP ‘753 was subsequently revoked in its entirety by the European Patent Office in opposition proceedings on 12 January 1999, leaving only the US patents in force.

The Licence Agreement

Under the Licence Agreement, Genentech undertook to pay: a one-off upfront fee (20,000 Deutschmarks); a fixed annual research fee (20,000 Deutschmarks) and a running royalty of 0.5% on the net sales of ‘Finished Products.’

The Licence Agreement defined Finished Products as “commercially marketable goods incorporating a Licensed Product …”.

‘Licensed Product’ was in turn defined as “materials (including organisms) in respect of which the manufacture, use or sale would, in the absence of this agreement, infringe one or more unexpired claims included in the rights attached to the patents under licence”.

The Licence Agreement was terminable for convenience by Genentech on two months’ written notice, and was governed by German law.

The dispute

Genentech paid the one-off fee and the annual research fee, but never paid the running royalty.

In June 2008, 14 years after the Licence Agreement was signed, Sanofi-Aventis Deutschland GmbH, a subsidiary of Hoechst, asked Genentech for information about its sales of Finished Products. This prompted Genentech to give notice of termination of the Licence Agreement, which took effect from 28 October 2008. Hoechst subsequently initiated ICC arbitration proceedings, in accordance with the dispute resolution clause in the Licence Agreement, claiming unpaid running royalties on sales of Finished Products prior to the date of termination. The seat of the arbitration was specified as Paris, with a single arbitrator.

Shortly after Hoechst initiated the arbitration, it also brought, in parallel, an action for patent infringement against Genentech in the US in respect of alleged use of the HCMV enhancer technology after the date of termination of the Licence Agreement. On the same day as Hoechst commenced those proceedings, Genentech brought proceedings for revocation of Hoechst’s US patents (US ‘122 and US ‘644). The US Courts dismissed the claims of both Hoechst and Genentech, upholding the validity of the patents but finding no infringement. Amongst other reasons, they found that Genentech did not infringe the method claims as Genentech did not practise the required step of “inserting” the isolated HCMV enhancer in a mammalian cell. It was common ground that Genentech derived the cell lines used to produce rituximab by inserting the enhancer into mammalian cells, but had done so before the US patents issued in 1998, and so that act of insertion could not amount to patent infringement. Those existing cell lines were subsequently propagated by mitosis (cell division), but this did not amount to “insertion” of the enhancer into the daughter cells.

Despite the finding of non-infringement of the US patents by the US Court, and the revocation of the European patent thirteen years earlier, the sole arbitrator hearing Hoechst’s claims in arbitration held Genentech liable to pay the unpaid running royalties up to the date of termination of the Licence Agreement, and assessed these at €108 million. He considered that Genentech’s position was based on a literal interpretation of the Licence Agreement, in particular the definition of “Licensed Product”, and he rejected this approach as a matter of German law (the governing law of the contract). He concluded instead that the commercial objectives of the parties had been to avoid the need in litigation for detailed consideration of whether the manufacture or sales of Rituxan™ in the patented territories amounted to infringement, and of whether the licensed patents were valid. He therefore concluded that the Licence Agreement should be interpreted as:

  • Applying to all US sales of Rituxan™ between the date the earliest US licensed patent issued (1998) and the termination of the Licence Agreement (2008). The full text of the arbitral award is confidential and the extracts which have been made public do not make clear whether the running royalty was also held to be due on sales of Rituxan™ in the European jurisdictions covered by EP ‘753 until termination of the Licence Agreement, though the CJEU appears to have proceeded on this basis.
  • Not requiring reimbursement of running royalties which have been paid, and not permitting the withholding of unpaid running loyalties, in the event that the licensed patents were revoked or found not to cover Rituxan™.

Genentech applied to the French courts (the supervising courts of the seat of the arbitration) to set aside the arbitral award. French arbitration law (as the law of the seat of the arbitration) contemplates such annulment applications, but only on very limited grounds – for example, breach of international public policy. Annulment proceedings do not, however, involve a review of the substantive legal issues in the award. Genentech argued that the sole arbitrator had interpreted the Licence Agreement in a way that made it incompatible with European Union anti-trust law (Article 101 TFEU), and this amounted to a breach of international public policy. The French courts asked the CJEU to rule on whether such a licence was incompatible with Article 101 TFEU; i.e., whether, for a licence agreement like that in issue, Article 101 precludes parties imposing an obligation to pay a royalty for the use of patented technology for the life of the agreement, even in the event of the revocation or non-infringement of patents protecting that technology.

As summarised above, the CJEU concluded that Article 101 TFEU did not have that effect, as long as the licensee was able to terminate the licence agreement on reasonable notice.

So what?

In light of the CJEU decision, the French annulment proceedings will almost inevitably be dismissed, leaving Genentech to satisfy the arbitral award (or possibly attempting to resist recognition and enforcement of the award in the US (or in any other jurisdictions in which they have assets), to the extent that there are arguable grounds for doing so under the New York Convention.)

Two brief observations can be made on the decision of the CJEU itself.

  • First, the CJEU decision is consistent with the Opinion of Advocate-General Wathelet, which was handed down a few months earlier. However, the Advocate-General worded his approval more narrowly than the CJEU. He limited his judgement to licence agreements where:

(1)the licensee can terminate the license agreement by giving reasonable notice;
(2)the commercial purpose of the licence agreement is to avert patent litigation;
(3)the licensee can challenge the validity or infringement of the patent during the term; and
(4)the licensee retains his freedom of action after termination (i.e. is not subject to a contractual obligation to discontinue use of the licensed technology).

  • Second, the position articulated by the CJEU is somewhat different to the position under US law, which views charging royalties under patents which have expired or been revoked as patent misuse (Brulotte v Thys Co, 379 U.S. 29 (1964), recently affirmed by the US Supreme Court in Kimble v. Marvel Entertainment, LLC, 576 U.S. __ (2015)). This is the basis for the common practice in well-drafted US-style patent & know-how licences of distinguishing between the royalty rate payable in respect of licensed products falling within a valid unexpired patent claim, and the royalty rate payable in respect of unpatented know-how. Given that the Licence Agreement covered the US territory, and in that sense the arbitral award (from a Paris-seated sole arbitrator) governs the charging of royalties within the US, it is interesting to speculate as to whether Genentech will rely on incompatibility with the US patent misuse doctrine in seeking to resist enforcement of the arbitral award in the US courts).

Some useful lessons for the drafting of patent licences can also be drawn from this salutary saga:

  • The importance of the choice of law clause. Questions of contract interpretation are resolved by the application of the governing law of the contract.
  • In this case, the application of German law appears to have allowed the sole arbitrator to put significant emphasis on the commercial purpose of the licence agreement, and to disregard the plain and ordinary meaning of the definition of “Licensed Product”. This definition was fairly standard in well drafted patent licence agreements, and clearly required a “Licensed Product” to be one that “infringe[d]… an unexpired claim”. It is questionable whether a tribunal applying English law or New York law would have felt able to depart from the plain and ordinary meaning of these words.
  • The arbitrator did so as he determined that the commercial purpose of the Licence Agreement was to avoid litigation, and in particular the need for detailed consideration of validity and infringement. What the arbitrator meant by this, and the basis for this finding, is unclear (at least, from the public reports of the arbitral award). It could be said that the commercial purpose of virtually all patent licences is to avoid litigation, at least as regards the validity of the licensed patents. This would only not be the case where the validity of the patent has previously been established in a binding decision of a court of competent jurisdiction. As regards avoiding litigation over infringement, where this is the commercial purpose of the licence agreement it would be more normal to define the Licensed Products by reference to specific products (by name or product reference number, and avoid reference to the concept of infringement in the definition.
  • Agreeing the choice of law clause is often left to the last minute in a contract negotiation. Time and budget constraints often do not allow the input of local counsel and the choice is often made ‘at the 11th hour’ or without appropriate advice. The choice of law clause is of fundamental importance. This case is a useful reminder that best practice for any material contract is always to allow time for review by local counsel before final agreement is reached.
  • Draft expressly to specify the consequences of revocation of a licensed patent. Well drafted patent licences should stipulate what will happen to royalties which have been paid, or which fell due for payment but remain unpaid, prior to revocation of a licensed patent. Are those royalties refundable? Similarly, the licence agreement should provide expressly for whether royalties continue to accrue following revocation of a licensed patent, if the licensee continues to practise the (previously patented) invention and the licence agreement has not been terminated.
  • The importance of an appropriate dispute resolution provision. The advantages of arbitration in international commercial contracts, including patent licences, are well known. They include confidentiality (subject to the choice of an appropriate seat and/or arbitral rules), the ability to appoint subject matter expert(s) as the tribunal, and, generally most importantly, ease of cross-border enforcement of arbitral awards. However, careful drafting of the arbitration clause is required. Two particular aspects are highlighted by this dispute.
  • Choice of composition of the arbitral tribunal – one or three arbitrators? It seems that the Licence Agreement provided for a sole arbitrator. This reduces costs and can allow for a more expeditious timetable. On the other hand, it means that each side will not be able to appoint a representative to the Tribunal (as is typically the case in a three-arbitrator tribunals), which is seen by some parties as an important advantage of arbitration, or at least one member of the Tribunal, although neutral to the parties, will typically nonetheless seek to ensure that its appointing party’s arguments are understood by the Tribunal.
  • The finality of arbitration is generally accepted internationally as a key policy aim, but is not always fully appreciated by contract drafters. The scope to challenge an arbitral award is strictly limited. This should be taken into account in deciding whether to provide that disputes should be resolved by arbitration. The choice of the seat of the arbitration determines the national courts which have supervisory jurisdiction over the arbitration. This in turn determines the national arbitration law which will apply to determine the grounds on which an arbitral award can be annulled by the local courts. The institutional rules stipulated by the parties as governing the arbitration are also relevant, as many (including, for example, both the ICC and LCIA Rules) expressly waive recourse which may be available under the supervisory law (for example, for arbitrations seated within the UK, under s69 of the Arbitration Act). Some countries are more liberal than others, allowing a wider range of grounds on which arbitral awards may be annulled or otherwise challenged. These factors should be prioritised when choosing the seat of the arbitration, rather than choosing solely on logistical grounds. Returning to Genentech Inc v Hoechst, French law is strict in this regard, only allowing annulment on very limited grounds. This promotes finality but, especially when combined with a sole arbitrator, leaves a losing party with few options if it believes that the sole arbitrator has made an error of law, even one which is obvious on the face of the award.

Timeline of dispute

Date Event
January 1991 Effective Date of the Licence Agreement
April 1992 EP ‘573 granted by EPO. US patent applications pending
August 1992 Licence Agreement signed

November 1997

FDA grants biologics licence application for Rituxan™
June 1998 European Medicines Agency grants centralised marketing authorisation for MabThera™/Rituxan™
December 1998 US 5,849,522 issued by US Patent & Trade Mark Office
January 1999 EP ‘573 revoked by EPO in its entirety
April 2001 US 6,218,140 issued by US Patent & Trade Mark Office
June 2008 Licensor requests statement of account of royalties from Licensee
August 2008 Licensee gives notice to terminate the Licence Agreement
October 2008

Licensor initiates arbitration proceedings

  • Licensor brings patent infringement proceedings against Licensee in US Federal District Court for the Eastern District of Texas
  • Licensee brings action for revocation of US Licensed Patents, and for declaration of non-infringement in the Northern District of California
March 2011 US District Court dismisses infringement action and revocation action
March 2012 US Court of Appeals for the Federal Circuit upholds the first-instance decision
September 2012 Sole arbitrator issues third partial award, holding Licensee liable for payment of the running royalty
December 2012 Licensee brings action before Cour d’Appel de Paris seeking annulment of third partial award
February 2013 Sole arbitrator issues final award and fourth partial award on quantum and on costs
September 2014 Cour d’Appel de Paris requests preliminary ruling from the Court of Justice of EU
July 2016 Court of Justice of EU hands down decision

For more information contact

< Go back

Print Friendly and PDF
Subscribe to e-briefings