Vicky Clark

Principal, VC Legal

Drafting agreements is often described as more of an art than a science. There is certainly plenty of scope for creativity when drafting patent licences in the biotech field. The bottom line for the licensor is, however, that it receives its royalty. Given the risk of future disputes relating to royalties and the business impact that such disputes can have, it is essential to be aware of common pitfalls which may provide the licensee with the means of avoiding royalty payments.

There are at least two parties in any negotiation, each with a different perspective and different interests. This article is written from the perspective of the licensor. It examines some of the issues which may arise when negotiating royalty provisions and suggests ways in which to minimise the risk for the licensor of future disputes over royalty payments.

Definition of Licensed Products
Most patent licence agreements are drafted such that royalties are only payable in respect of ‘Licensed Products’ supplied by the licensee. Hence the way in which Licensed Products are defined has a fundamental impact on the royalties that will be payable to the licensor in due course. Where a product is at an advanced stage of development, it may be possible to specifically identify the product on which royalties will be payable (by use, for example, of the licensee’s name or code for the product). Where a licence is granted at an early stage of development, and no commercial product as such yet exists, it is necessary to define the products on which royalties are due in a generic way.

It is not uncommon for licensees to insist that the generic definition of Licensed Products is explicitly linked to infringement of valid claims of the licensed patents (a Licensed Product, on which royalties are payable, would be one which infringes a valid claim of one of the licensed patents in the territory in which it is supplied by the licensee). A definition of this kind raises plenty of scope for later disputes over royalty payments when a commercial product finally reaches the market.

Does the Product Infringe?
The question of whether a product infringes a patent is rarely straightforward. Cases of literal infringement are few and far between, particularly in relation to biotech patents. Even if a case appears at first glance to be one of literal infringement, the defendant can usually identify some ambiguity in the claims that are alleged to be infringed and this can be used as a basis to resist payment of royalties. The dispute between Celltech Chiroscience and Medimmune Inc, which is the subject of a recent Court of Appeal decision, is a case in point.

Celltech brought proceedings in the United Kingdom against Medimmune to recover royalties which Celltech claims are due to it under a licence granted to Medimmune in respect of Celltech’s ‘Adair patent’. The Adair patent covers certain humanised antibodies. Medimmune has developed a humanised antibody for the treatment of respiratory syncitial virus, which it sells under the trade name ‘Synagis’.

Under the licence agreement between Celltech and Medimmune, royalties of 2 per cent of net receipts were payable on all antibodies which, but for the licence, would infringe a valid claim of the Adair patent. When Medimmune began to sell Synagis in the United States, Celltech sought to recover its royalty under the licence agreement. Medimmune refused to pay, claiming that the manufacture and sale of Synagis in the United States does not infringe the US version of the Adair patent.

Due to the dispute resolution provisions in the licence agreement, Celltech issued proceedings against Medimmune in the United Kingdom. The critical issue which the UK courts had to decide (by way of a preliminary issue) was whether the sale of Synagis in the United States infringes the US version of the Adair patent. This may sound a strange matter to be referred to the UK courts but the dispute resolution provisions in the agreement were clear and the UK courts are familiar with dealing with questions of foreign law.

Celltech were not able to present a straightforward case of infringement. The Adair patent relates to humanised antibodies which are part human and part mouse. In such antibodies, the mouse residues are called ‘donor elements’ and the human residues are called ‘acceptor elements’. Celltech identified that in order to be effective, not only must the part of the antibody that binds to antigen (the complimentary determining regions or ‘CDRs’) be donor elements, but certain other residues must also be murine derived.

One of these non-CDR, murine-derived or donor elements is at the heart of the dispute between Celltech and Medimmune. Using the ‘Kabat’ numbering system, the US Adair patent specifies that the amino acid at position 23 must be a donor element. The amino acid at position 23 of the Medimmune antibody is an acceptor (human derived) element. Celltech claims that, while there was no literal infringement of the Adair patent, Synergis infringes the Adair patent under the US ‘doctrine of equivalents’ as the residue used by Medimmune at position 23 is a conservative substitution3 for the amino acid specified in the Adair patent.

The doctrine of equivalents requires an analysis of the patent much like the purposive approach used by the UK courts. The analysis was complicated by defences relied upon by Medimmune, which would not have been available under English law, namely amendment estoppel and argument estoppel. Basically, the limitation in the US Adair patent relating to the nature of the residue at position 23 was introduced by Celltech during the prosecution of the US patent in order to overcome disclosures in a prior art document cited against the patent. As a result of the principles of amendment and argument estoppel which are available under US law, the UK court essentially held that the Adair patent had to be construed strictly and that Celltech could not rely on its equivalence argument. Hence despite the fact that the residue used by Medimmune in the Synergis antibody is a conservative substitution for the residue specified in the Adair patent, the UK court held (both at first instance and at the Court of Appeal) that Synergis does not infringe the US Adair patent and, consequently, that no royalties are payable by Medimmune on sales of the product in the United States.

Putting this in the context of the above discussion, Celltech might have reasonably assumed when Medimmune entered into the licence agreement that Medimmune accepted that its product infringed the Adair patent. Hence Celltech may have expected in due course to receive a royalty from Medimmune. In reality, Medimmune may have always known that it would ultimately contest infringement and resist paying royalties. By entering into the licence Medimmune could, however, have avoided distracting infringement claims at an early stage of product development.

This is not an unusual scenario. By entering into licence agreements, licensees can buy themselves valuable breathing space from infringement proceedings without making any significant commitment to the licensor. What is more, as and when the licensee makes it clear that it does not intend to pay royalties and alleges non-infringement, depending on how the licence is drafted, the licensor may be tied into the dispute resolution mechanism set out in the licence, rather than being able to choose the most appropriate forum to litigate the issue of infringement.

So how can the licensor avoid this ‘infringement trap’? The licensor should try and obtain an express acknowledgement from the licensee in the licence agreement that royalties will be payable on the product under development by the licensee. This does not necessarily involve the licensee expressly acknowledging infringement (which can be a sensitive issue for many licensees) but could state that, irrespective of the issue of infringement, the licensee accepts that royalties will be due on product ‘X’ (defined as the product under development by the licensee).

An alternative is to uncouple the definition of Licensed Products from the issue of infringement. This would involve defining Licensed Products as any product the development, manufacture or supply of which utilises any of the ‘licensed technology’. Licensed technology would be defined as the technology disclosed in the patents and, if appropriate, the know-how disclosed to the licensee. Know-how can often prove important in negotiating provisions of this nature, as where the licensee has obtained the benefit of confidential know-how, over and above the information disclosed in the licensed patents, the licensor has good justification for uncoupling the issue of royalties from patent infringement.

An argument often raised by the licensee in relation to infringement is that it is anti-competitive for royalties to be payable on any products which do not infringe a valid patent. The existence of valuable know-how can side-step this issue. In addition, there are good arguments to support the view that, when it comes to royalties, the European Commission will allow the parties to an agreement wide discretion in deciding the issue of compensation. In the situations to which this discussion relates, where infringement is arguable and the licensee has benefited from the licensed technology, so long as a patent is in force in the territory in which the product is supplied, it should not be anti-competitive for the parties to acknowledge in advance the issue of infringement.

Valid Claims
Linked to the issue of infringement is the provision, often sought by licensees, that royalties are only payable on products sold in territories in which a ‘valid claim’ exists. The requirement of validity provides the licensee with another potential get-out, as arguments can almost always be raised that a given patent or claim of a patent is invalid. Hence a licensee, seeking to avoid royalty payments, merely has to assert that the claims alleged to be infringed are not valid. The licensor invariably then has little choice but to invoke the dispute resolution mechanism in the licence agreement to resolve the dispute. Note that, as with infringement, this robs the licensor of the decision as to which is the most appropriate forum to decide the issue of validity.

To avoid this potential dispute, the licensor should not accept the natural meaning of ‘valid claim’. Instead the licensor should seek to expressly define in the agreement a valid claim as one which is in force and which has not been held to be invalid by any patent office or court of competent jurisdiction. The licensor may also require that the decision of invalidity must not be final (i.e. a decision from which there is no appeal or which has not in fact been appealed). Using this approach, the licensee would be free to challenge the validity of any of the licensed patents in the normal way but would remain bound to pay royalties in respect of a patent until it was finally found to be invalid.

Other Considerations
If the licensee will not budge on the issues of infringement/validity of the claims, the licensor, being aware of the risk of future disputes, may seek to structure the compensation for the licence in ways that do not rely entirely on royalty income. For example, the licensor could negotiate an initial lump sum licence fee and high milestone payments.

A minimum royalty provision can also help to alleviate the problem by ensuring that, irrespective of the issues of infringement and validity, the licensor will receive a fixed sum every quarter.

The licensor can also ensure that there are deterrents in the agreement against challenging infringement and/or validity. For example, the licensor could reserve the right to terminate the agreement if the licensee contests infringement or the validity of any of the claims of the licensed patents. This is entirely compatible with the provisions of the existing Technology Transfer Block Exemption and the proposed block exception that is due to replace it in May 2004. Without the licence in place, the licensor would then be free to sue the licensee for infringement in the most appropiate territory.

Royalty Stacking
Cambridge Antibody Technology (‘CAT’) announced recently that Abbott Laboratories, one of its major licensees, is contesting amounts due to CAT under the licence agreement between Abbot and CAT relating to Humira (Abbott’s recently approved monoclonal antibody for the treatment of arthritis). From CAT’s announcement it appears that royalty-stacking provisions are at the heart of the dispute between CAT and Abbot.

The concept of royalty stacking arises from the risk that multiple patents may affect a single product. Such a risk is said to be particularly high in the biotech field which is dominated by patent filings.

Take, for example, a gene therapy product. The most obvious patent to affect such a product is a patent that covers the DNA sequence for the gene. For an effective product, however, promoter sequences and an appropriate vector may be required. Each of these may also be covered by third-party patents. In addition, tools may have been used to develop the product. How was the DNA sequence identified? Were patented techniques used for this purpose? If so, do the patents relating to such techniques have reach-through claims that could affect the final gene therapy product?

Royalty stacking arises when, in order to take a product to market, the developer of the product takes licences from all of the owners of the patents which affect the final product. When the royalty payments are added together, the licensee may find itself with a non-profitable product. Hence it has become quite usual for licensees to insist on including anti-stacking provisions in licence agreements.

A typical anti-stacking provision states that the royalty rate payable to the licensor will be reduced if the licensee is obliged to enter into licences with third parties in relation to the product. It is not difficult to see how such a provision can lead to a disparity between the expectations of the licensor as to the royalty it will receive from the licensee and the actual royalty the licensee is contractually obliged to pay.

Notice
The first point to make is that to avoid such a disparity in expectations, the licensor should ensure that it is notified immediately if the licensee enters into a further licence that might affect the royalty received by the licensor. At least the licensor will then be able to avoid the sudden disappointment when the product comes to market and it finds that its royalty income is going to be far less than expected.

Control
Ideally, if the licensor accepts the concept of anti-stacking, it should retain some control over third-party licences entered into by the developer of the product. A scenario may arise in which the licensee believes that a third-party licence is required but the licensor disagrees. When the issue of infringement is ambiguous, the licensee may be keener to obtain the comfort of a licence than the licensor. As far as the licensee is concerned, the anti-stacking provisions will ensure that the money all comes out of the same pot, whereas the licensor is faced with a hit to its royalty income. One way to address this problem is to require the agreement of the parties to third-party licences and, in the absence of agreement, to refer the matter to an independent expert for determination as to whether a licence is required.

If the parties decide that a third-party licence is required, it is not unusual for the licensor that is affected by the anti-stacking provision to actually have conduct of the negotiations. That way the licensor can seek to minimise the impact of the third-party licence on its own royalty stream.

Floors
Another point to consider is putting a floor on the extent to which the royalty rate can be reduced in response to stacking. A royalty may start at, say, 5 per cent and be reduced in response to stacking but no lower than, say, 3 per cent. The licensor should also ensure that its own royalty rate is reduced only on condition that all other licensees accept similar anti-stacking provisions and the royalty rates of all licensees are reduced on a pro-rata basis in response to a third-party licence.

The minimum floor operates in a similar way to a minimum royalty which, as when dealing with the infringement and validity issues discussed above, can be a very useful tool to protect the income stream of the licensor in the event of disputes over stacking.

Research Tool Patents
Finally, it is not unreasonable for the licensor to make a distinction in relation to anti-stacking between research tool patents and product patents. The licensor may take the view that where the licensee has to enter into third-party licences in relation to patents which affect the actual product, anti-stacking measures may be appropriate. Where research tool patents are concerned, the licensor may feel that such patents are the responsibility of the licensee and should not affect the licensor’s royalty in any way.

Combination Products

It is not unusual to see provisions in patent licenses dealing with ‘combination products’. In the pharmaceutical field, a typical example of a combination product would be where an active ingredient is sold with some kind of mechanical device for administering it, such as an inhaler. The licence agreement may deal with such a product by stating that royalties will not be payable on the total value of the combination product. Hence an apportionment is required and royalties are only payable on the value attributable to the active ingredient, not the value attributable to the inhaler.

Unless provisions of this nature are carefully drafted, there is potential for future disputes as it is not always entirely clear where one product ends and another begins. Take, for example, a single pill or capsule that contains two actives. Is this one product or two? Consider also biotech products where the delivery device is not mechanical, such as an inhaler, but biological, such as a virus. Are the active and the virus two separate products?

It is possible that provisions relating to combination products could be invoked by the licensee in relation to biotech products to achieve anti-stacking by the back door. It is advisable, therefore, to ensure that any provisions relating to combination products set out examples of the type of scenarios that such provisions are intended to address. This ensures that in the future there can be no stretching of the interpretation of such clauses to cover situations that were never intended.

Conclusion
In view of the damage to the business of biotech companies that can be caused by royalty disputes, it is important for such companies to be aware of and to minimise the areas of ambiguity that can arise when drafting royalty provisions. It is, of course, impossible to avoid such disputes entirely, but it is possible to minimise the risks and to make such disputes as unattractive as possible to the licensee.

• Celltech Chiroscience Ltd v Medimmune Inc., [2003] EWCA Civ 1008.
• Each of the parties is entitled to call an expert in the foreign law in question (US patent law in this case) and the court then reaches its own conclusions based on the representations made to it.
• A conservative substitution is one in which an amino acid is replaced by another very similar amino acid, which substitution has little or no effect on the activity of the protein.
• Note that the case only resolved the issue of antibodies sold in the United States, not other jurisdictions. Celltech has issued separate proceedings against Medimmune in respect of sales of Synergis in Germany which Celltech claims infringe the German version of the Adair patent.
• Note that this is an entirely hypothetical consideration of the parties’ positions and may or may not equate to the positions adopted by the parties. The author has no personal knowledge of the parties’ intentions.
• In the Celltech case it may suit the company that all proceedings have to be brought in the UK courts but it is possible that due to procedural differences it would be advantageous to litigate the issue of infringement of foreign patents in the foreign jurisdiction in question.
• Hence the prevailing view that reach-through royalties are enforceable even though the product on which royalties are payable does not infringe the licensed patents.
• The position may be different in other jurisdictions and the US anti-trust laws may, for example, be stricter than the position adopted by the EU Commission on this point.
• Commission Regulation (EC) No. 240/96 of 31 January 1996 on the application of Article 85(3) of the Treaty to certain categories of technology transfer agreements, OJ 1996 L31, 9 February 1996, at 2 to 13.

To make any comments on this article, or to ask a question of the author, please contact the publisher. If you would like to submit an article please subscribe to our PL Intelligence service.

The opinions expressed in the articles published in this section do not necessarily reflect those of Pharmalicensing or UTEK Corporation. No actions including proposals to or agreements with other companies should be taken by any reader without obtaining specific business or legal advice. Neither the publisher nor the authors accept any liability for any actions or activities undertaken by any reader or other third party as a consequence of these articles or for any errors or omissions therein.