By Kevin E. Noonan —

Genentech announced last week that the U.S. Patent and Trademark Office had determined it would issue a re-examination certificate in Re-examination Control No. 90/007,542 involving U.S. Patent No. 6,331,415 ("the Cabilly II patent").  While a major victory for Genentech, it represents a bitter defeat for Medimmune, Genentech's licensee, as well as other biotechnology companies affected by Genentech's exclusive rights to basic recombinant DNA technology.

The Cabilly II patent claims fundamental methods for making recombinant cells expressing both an immunoglobulin light chain and heavy chain; these methods are useful for making any of several genetically-engineered antibody molecules.  Claim 1 of the patent, which will issue in the re-examination certificate unchanged from the claim in the '415 patent, reads as follows:

The Cabilly II patent issued on December 18, 2001, on a application filed June 10, 1988.  That application claimed priority as a continuation from U.S. Application No. 06/483,457, filed April 8, 1983, and issued as U.S. Patent No. 4,816,567 on March 28, 1989.

Prosecution of the Cabilly II patent was delayed by an interference declared between the Cabilly II application and U.S. Patent No. 4,816,397 to Boss.  The Board entered judgment against Cabilly on August 13, 1998.  Genentech filed suit under 35 U.S.C. § 146 that was not resolved in its favor until March 18, 2001.  The first of two ex parte re-examination requests (which were merged) was filed (by counsel for Medimmune) on May 13, 2005, and the re-examination ordered on July 7, 2005.  This re-examination was merged with re-examination Control No. 90/007,859 on June 6, 2006.

The following is a timeline of the procedural history of this patent:

All claims of the Cabilly II patent were rejected for obviousness-type double-patenting over parent U.S. Patent No. 4,816,567 ("the Cabilly I patent"); claim 1 of this patent reads as follows:

The significance of such a determination would be that the Cabilly II patent would expire 17 years after the grant date of Cabilly I (March 28, 2006) rather than 17 years after its own grant date (December 18, 2018).

The Office asserted the obviousness-type double patenting rejections against the three independent claims of the Cabilly II patent (1, 15, and 21) based solely on the parent Cabilly I patent; several of the remaining claims were rejected on obviousness-type double patenting grounds on the combination of the Cabilly I patent and other art, including U.S. Patent No. 4,399,216 to Axel et al., and several scientific publications, including one to Rice and Baltimore.  Genentech countered with numerous declarations from biotechnology luminaries, including Douglas Rice, the co-author of the Rice and Baltimore paper; Steven McKnight (linker-scanning technique); Arthur Riggs (DNA methylation); Michael Botchan (genetic mapping using Southern blot hybridization); and Sidney Altman (1991 Nobel Prize winner for ribozymes), among others.  Key points of contention in the re-examination was whether the Cabilly I patent encompassed expression of both an immunoglobulin light chain and an immunoglobulin heavy chain in its recitation of "DNA sequence encoding a chimeric immunoglobulin heavy or light chain," i.e., whether "or" should be interpreted to mean "in the alternative" or to mean "and/or" (the "logical or").  At various times during the re-examination this was cast as a "genus/species" question, but ultimately the issue was whether it would have been obvious to take the teachings of Cabilly I, interpreted to teach in the alternative, and express both immunoglobulin light chain and heavy chain molecules in the same cell, as well as whether the skilled worker in 1983 would have had a reasonable expectation of success that these co-expressed molecules would form a functional antibody, as required in the Cabilly II claims ("an immunoglobulin molecule or an immunologically functional immunoglobulin fragment").

Genentech amended claim 21, which will issue in the re-examination certificate reading:

In the Notice of Intent to Issue a Re-examination Certificate, the Patent Office opined that the Cabilly I patent was limited to expression of immunoglobulin light chains and heavy chains in separate cells; the Office determined that the Kaplan reference (EP 0 044 722) was limited to the same disclosure (relying on the Harris, McKnight, Botchan, Rice, and Colman declarations).  Relying on the same declaratory evidence, the Office determined that the Axel patent did not teach co-expression of two "foreign" (i.e., exogenous) DNA sequences.  Further relying on the same declarations, the Office determined that the Rice and Baltimore reference (1982, Proc. Natl. Acad. Sci. U.S.A. 79: 7862-65) did not teach production of a functional immunoglobulin comprising an exogenous light chain and endogenous heavy chain; the Office came to the same conclusion regarding the Ochi reference (1983, Nature 302: 340-42).  The Dallas reference (WO 82/03088) was limited to recombinant E. coli transformed with exogenous E. coli genes and did not teach methods for producing "multiple eukaryotic proteins from a single cell host."  The Office determined that the Moore patent (U.S. Patent No. 5,840,545) was limited to producing rFv (recombinant single-chain immunoglobulin molecules) comprising heavy and light chain variable regions in separate cells.  Studies regarding production of immunoglobulin in Xenopus oocyte cells (Deacon and Valle, 1976, Biochem. Soc. Trans. 4: 818-20) were not directed towards producing recombinant immunoglobulins in eukaryotic cells according to the Notice.  The Builder reference (U.S. Patent No. 4,511,502) was limited to recovering expressed recombinant polypeptide from bacterial cells, but did not specifically teach recombinant immunoglobulin molecules.  Finally, the Accolla reference (1980, Proc. Natl. Acad. Sci. U.S.A. 77: 563-66) described methods for making immunoglobulins against carcinoembryonic antigen (CEA), but not recombinant versions thereof.

The Notice summarizes the reasoning it reconsidered its earlier reject
ion of the Cabilly II claims:

The course of the Cabilly II re-examination is informative for several reasons that are relevant in the broader context of U.S. patent policy.  A careful reading of the Office Actions and the evidence adduced by Genentech reveals the difficulties attendant in evaluating, particularly so long after the fact, what the skilled worker would have known at the time the Cabilly invention was made (over 25 years ago).  Genentech benefitted from having declarants who were actively involved in the science of recombinant DNA technology and immunology at that time and who could opine on their specific knowledge of what they and their colleagues were doing as well as the state of the art.  Although the timeframe is long, it is not totally irrelevant under the modern patent term regime of 20 years from a first filing date:  an applicant who files a provisional application and receives patent term adjustment for PTO delay and patent term extension for regulatory delay might be able to obtain a term that is substantially equivalent.

This raises a separate issue relevant to proposals in the last Congress (almost sure to resurface in this Congress) about post-grant review.  U.S. re-examination (both ex parte and the more recently-established inter partes versions) was intended to provide an avenue short of litigation for challenging patents that may have been improvidently granted.  Similar regimes exist in other countries, most particularly in the European Patent Convention.  One major difference between the U.S. systems and those found abroad is that oppositions in Europe and elsewhere are limited and cannot be brought at any time during the patent term.  While this has the effect of precluding later-accused infringers or competitors who are not diligent from pursuing the administrative avenue of opposition, patents can still be invalidated in Europe by nullity actions (albeit with the added expense and inconvenience of having to be pursued country-by-country).

The advantage of limiting the time for filing oppositions (in Europe, nine months after publication of the granted patent) is that it imposes the equivalent of a "statute of repose" on challengers, giving the patentee the benefit of being secure in its patent rights.  In the U.S., no such repose exists, and in view of the expense (and risk) of patent litigation perhaps there cannot be.  But the recent Supreme Court decision in Medimmune, Inc. v. Genentech, Inc. (on this same Cabilly patent) has expanded the limits of subject matter jurisdiction for declaratory judgment actions to establish invalidity.  So perhaps it might be wise to at least consider, during any future Congressional hearings for patent reform including yet another post-grant review system, whether there is an advantage in granting patentees repose after some sufficient time for granted patents to be challenged.  The Cabilly II re-examination illustrates the advantages that would be gained in having determinations like obviousness and obviousness-type double patenting decided more contemporaneously with the events that resulted in the claimed invention.