United States
IP issues in biopartnering
Charles A Weiss
Kenyon & Kenyon
Sir Winston Churchill said: “All men make mistakes, but only wise men learn from their mistakes.” This chapter discusses several well-known conflicts arising from research, collaboration and licence agreements in the field of biotechnology. Not every contract dispute stems from a mistake – even the best drafting will not stop a determined party bent on evading its duties or expanding its rights under a contract from contending that its desired course of conduct is permitted – but certain contractual provisions result in disputes too often to ignore. Remembering what has gone wrong for others can help avoid conflict in the first place and increase the chance of winning a case.
This chapter focuses on the definition of ‘product’, specifying the royalty base and handling improvements to licensed or collaboration technology.
Definition of ‘product’
Several definitions incorporating the word ‘product’ are used in almost every biotechnology agreement. Different parts may have different product-based definitions – for example, the ‘collaboration product’ to be developed jointly by the parties may be different from the full scope of ‘licensed products’, and different breadth may be given to ‘competing products’ if the agreement limits the parties from separately developing products that compete with the collaboration.
The potential for problems with the definitions of ‘product’ is illustrated by a dispute between the Wisconsin Alumni Research Foundation (WARF) and Geron Corporation arising from a 1999 licence agreement concerning stem cells. At issue was the definition of ‘research products’, which were exclusively licensed to Geron. Geron sought a more expansive application of the term and WARF filed a declaratory judgment suit to obtain a judicial resolution. Read the following quote (wherein the point of dispute is stated) and then try to identify the source of conflict from the definition: “‘Research products’ shall mean products or services that (i) are used in research as research tools which would infringe the claims of patented technology owned by Geron or which Geron has a right or license to use other than the licenced patents; and (ii) which employ, are in any way produced by the practice of, are identified using or arise out of any research involving the inventions claimed in the licenced patents or that would otherwise constitute infringement of any claims of the licenced patents. ‘Research products’ specifically excludes the materials.”
At first glance, attention would probably focus on the potential difficulty under clause (ii) to determine if a particular product was “identified using” or “arose out of” the licensed technology. But the disagreement that found its way to court came from clause (i). WARF argued that only those products that Geron had an exclusive “right or license to use” satisfied that requirement; Geron contended that the term ‘licence’ did not distinguish between exclusive and non-exclusive licences and thus embraced products that were included under any third-party licence (exclusive or non-exclusive) held by Geron. Another point of contention was whether the licences Geron might obtain in the future were included, such that a product that started out falling outside the definition could later be swept in by Geron’s unilateral, post-agreement activities. If so, WARF would be at risk if it licensed to third parties even those products that currently fell outside the definition, since Geron might later convert them to exclusive research products by acquiring a third-party licence that covered them. In its legal brief, WARF conceded that: “In a perfect world, the parties’ intent with respect to research products would have been made explicit by the addition of the word ‘exclusive’ to the phrase ‘right or license to use’.” Geron responded that this omission was precisely the point, leaving the unmodified word ‘licence’ to cover licences of any nature. The case was settled at an early stage of litigation.
What drafting lessons may be drawn from this dispute? Although the language purist may object, solutions to the ‘licence’ ambiguity alleged by WARF are either to add ‘any’ as the modifier of a word intended to be used in a general sense (ie, ‘any right or licence’), or to provide an exhaustive enumeration of possible subcategories followed by ‘or otherwise’, just in case something was omitted (ie, ‘a right or licence (under a patent, know-how or otherwise, and whether exclusive, non-exclusive or otherwise’). We can also learn from the technique used in the last sentence of the definition, which is to negate possible definitional ambiguity by expressly modifying the definition as to certain subject matter (here “the materials”, which were the patented cell lines themselves).
Defining the ‘royalty base’
Defined narrowly, the ‘royalty base’ is a monetary figure that is multiplied by a royalty rate to determine the amount of royalties due. An example is the oft-used phrase ‘net sales of licensed products’, as in ‘the licensee shall pay the licensor a royalty equal to 2 per cent of net sales of licensed products’. More broadly, the royalty base may be thought of as any activity other than milestones or other one-off events that trigger a payment obligation. Broad or narrow, it is an area ripe for ambiguity and contention.
Consider the well-publicised litigation between City of Hope National Medical Center and Genentech arising from a 1976 research and licence agreement that helped usher in the modern age of biotechnology. Under the agreement City of Hope was to perform “step-wise synthesis of DNA strands suitable for incorporation in recombinant plasmids according to the procedure generally discussed in” the now-famous article by Stanley Cohen published in Scientific American in 1975. Unlike the near-universal practice today, patent rights resulting from City of Hope’s work were to be assigned, not licensed, to Genentech.
The agreement provided, in different paragraphs, for royalties to City of Hope on sales by Genentech, sales by licensees, and damages or settlements of patent litigation. Again, try to identify the possible problem areas from the following quote: “6.01: Genentech shall pay to City of Hope a royalty of 2% of the net sales of all polypeptides sold by it or its affiliates, provided only that manufacture of polypeptide employs DNA synthesized by City of Hope under this agreement and provided to Genentech by City of Hope or replications of that DNA.
“6.02: For the period commencing with the effective date of this agreement and ending five years thereafter, Genentech will pay to City of Hope the royalty provided for in Article 6.01 hereof regardless whether Genentech has secured one or more patents on developments. Thereafter, Genentech’s royalty obligations under Article 6.01 shall be limited to the payment of royalty only in respect to such manufacture, use or sale which would infringe the claims of an issued Genentech patent on Genentech Developments...but for Genentech’s ownership of the patent.
“6.08: Should Genentech license any third party under any patent acquired by it hereunder, then Genentech shall secure from that party and pay to City of Hope the same royalty City of Hope would have received had Genentech itself carried out the licenced activity.
“7.02: Should Genentech recover from any infringer, whether by way of settlement of judgment, damages or profits for infringement of any patent secured by it under this agreement, then after deduction...of Genentech’s reasonable expenses...the balance shall be treated as net sales of Genentech for royalty purposes.”
Did Genentech have to pay royalties after the first five years on products that were covered by the patents but that did not employ City of Hope DNA? This dispute over what the court called the “DNA-use requirement” turned on whether the patent royalty set out in Article 6.02 was intended to replace the material-based royalty of Article 6.01 after five years (by which time the patents presumably would have issued), or to impose an additional condition (the issuance of patents) on a still-valid DNA use requirement of Article 6.01. Did Genentech have to pay City of Hope royalties only on sales by licensees of products covered by the patents or also on its licensees’ sale of any product included in a licence of the patents without regard to whether that product actually fell under the patents? Third, did the DNA use requirement apply to monies received by Genentech from the settlement of patent infringement cases? These last two disputes arose from the question of whether Articles 6.08 and 7.02 included as a condition of payment all requirements of Articles 6.01 and 6.02.
In particular, the second dispute – since its resolution adverse to Genentech made Genentech liable for paying royalties to City of Hope on sales by licensees that were not royalty bearing to Genentech – shows the importance of clear drafting for the royalty base.
This dispute has not yet been resolved despite two jury trials in 2001 and 2002 (the first jury could not reach a verdict), 17 days of jury deliberation in the second case and one appeal decided in October 2004 (City of Hope National Medical Center v Genentech Inc, 20 Cal Rptr 3d 234 (Ct App 2004)). The case is now pending before the California Supreme Court, with a decision unlikely before mid-2006. The stakes are high: the jury awarded City of Hope US$500 million in damages, including US$200 million in punitive damages for breach of fiduciary duty.
A lesson from this case is to make each section requiring payment based on different events internally complete and able to stand on its own (instead of referring back to different sections that may have unique provisions).
For example, Genentech’s interpretation of its royalty obligation for sales by licensees would have prevailed (probably as a matter of law without ever reaching a jury) if Article 6.08 had read: “Genentech shall pay City of Hope a royalty of 2% on net sales by its licensees of polypeptides if both (i) the manufacture, use or sale of which would infringe the claims of an issued Genentech patent on Genentech Developments but for the license granted to such licensee by Genentech, and (ii) such polypeptides employ DNA synthesised by City of Hope under this agreement and provided to Genentech by City of Hope or replications of that DNA.”
Similarly, the dispute over the DNA-use requirement on sales by Genentech could have been avoided if standalone provisions had been provided for royalties in the first five years and after the first five years: for example, Article 6.01 – “For the first five years from the effective date, Genentech shall pay”, followed by the complete royalty obligation for the initial period of time – and Article 6.02 – “After the fifth anniversary of the effective date, Genentech shall pay”, followed by the complete royalty obligation for the remaining term of the agreement. Another virtue of this approach is that when a change is made to one provision during the negotiation process, there is less chance of it having an unrecognised and unintended effect on other provisions.
Improvements
In 1985 Amgen licensed certain rights to the anti-anaemia drug erythropoietin (EPO) to Johnson & Johnson subsidiary Ortho Pharmaceutical. Amgen kept for itself the right to sell EPO for dialysis patients. All other fields, including cancer treatment, went to Ortho. EPO (sold as Procrit by Johnson & Johnson and as Epogen by Amgen) became a bestselling product. Amgen then developed a new product called Aranesp (darbepoetin alpha). Described in the approved product labelling as “closely related to” EPO, Aranesp has the same number of amino acids (165) as EPO but five substitutions in the amino acid sequence that create two additional glycosylation sites. The additional carbohydrate chains increase the molecular weight and thus extend the circulating half-life.
The Ortho/Amgen agreement narrowly defined EPO by its amino acid sequence but gave each party rights to improvements made by the other: “If Amgen, on the one hand, or Ortho...improve the product organisms, and/or the licenced know-how, or make licenced products or process improvements, all such improvements shall become part of the licenced know-how and shall be promptly transferred and/or communicated to the other party...and by the provisions hereof shall be deemed to be a part of the licenced patents or licenced know-how as the case may be and licenced to Amgen or Ortho, as the case may be.”
Other pertinent definitions included: “‘Licenced know-how’ shall mean...all data, information, technology or special ability on the part of Amgen...which is reasonably related to licenced patents and licenced products for use in the licenced field…‘Licenced products’ shall mean and include any products for use in the licenced field (i) which are within the scope of a valid licenced claim of a licenced patent; or (ii) whose use is within the scope of a valid licenced claim of a licenced patent; or (iii) which are manufactured or packaged within the scope of a valid licenced claim of a licenced patent; or (iv) which utilise any licenced know-how…‘Products’ shall mean IL-2, Hepatitis B, and EPO for all human uses in the licenced field.”
Ortho was given an exclusive licence to the licensed know-how, licensed patents and licensed products in its field (all indications other than for dialysis patients).
Ortho commenced an arbitration to obtain rights to Aranesp and prevent Amgen from selling it into Ortho’s exclusive field for EPO. As the matter was arbitrated there is no public record of the parties’ arguments, but reasonable inferences can be drawn from the information that is publicly available.
The key would seem to be whether the strict definition of EPO (by amino acid sequence) incorporated into the definition of ‘product’ (and hence into the definitions of ‘licensed know-how’ and ‘licensed products’) trumped the more general language of the improvements clause. Stated differently, if something new is related to but not strictly about EPO, can it still be an improvement that becomes part of the licensed know-how and is exclusive to Ortho in its field? Amgen won the arbitration, so it can be inferred that the arbitrators’ answer to the question was ‘no’. This loss was costly for Ortho: according to papers in a recent lawsuit filed by Ortho accusing Amgen of unlawfully tying sales of Aranesp to its white blood cell products Neupogen and Neulasta, Ortho has lost about half of its market share for Procrit to Amgen.
What is the lesson here? A licensee should not rely on an improvements clause in an exclusive licence to prevent future competition from the licensor in the licensed field. Sophisticated licensors resist conveying rights to improvements, so if such rights are obtained at all they are narrowly defined. It is not known whether Ortho believed it was getting non-compete protection when this agreement was negotiated, or whether it looked back and tried to construct an argument when Aranesp loomed on the horizon. But given the relative bargaining positions of Amgen and Johnson & Johnson in 1985, one has to wonder why Ortho did not obtain an agreement from Amgen that it would not launch a directly competing product into Ortho’s exclusive field.
Comment
Expanding on Churchill’s admonition, Admiral Hyman Rickover – father of the US nuclear submarine programme – said: “It is necessary for you to learn from others’ mistakes. You will not live long enough to make them all yourself.” Whether arising from mistakes or not, the contract problems of others provide guidance to those who study not only other people’s deals, but other people’s litigations.
