Genentech Announces Victory in Cabilly Re-examination

    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:

A process for producing an immunoglobulin molecule or an immunologically functional immunoglobulin fragment comprising at least the variable domains of the immunoglobulin heavy and light chains, in a single host cell, comprising the steps of:
    (i) transforming said single host cell with a first DNA sequence encoding at least the variable domain of the immunoglobulin heavy chain and a second DNA sequence encoding at least the variable domain of the immunoglobulin light chain, and
    (ii) independently expressing said first DNA sequence and said second DNA sequence so that said immunoglobulin heavy and light chains are produced as separate molecules in said transformed single host cell.

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:

A method comprising
    (a) preparing a DNA sequence encoding a chimeric immunoglobulin heavy or light chain having specificity for a particular known antigen wherein a constant region is homologous to the corresponding constant region of an antibody of a first mammalian species and a variable region thereof is homologous to the variable region of an antibody derived from a second, different mammalian species;
    (b) inserting the sequence into a replicable expression vector operably linked to a suitable promoter compatible with a host cell;
    (c) transforming the host cell with the vector of (b);
    (d) culturing the host cell; and
    (e) recovering the chimeric heavy or light chain from the host cell culture.

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:

A method comprising
    a) preparing a first DNA sequence [[consisting essentially of DNA]] encoding an immunoglobulin [[consisting of an immunoglobulin]] heavy chain and a second DNA sequence encoding an immunoglobulin light chain [[or Fab region, said immunoglobulin having specificity for a particular known antigen]];
    b) inserting the DNA sequences of step a) into a replicable expression vector wherein each sequence is operably linked to a suitable promoter;
    c) transforming a prokaryotic or eukaryotic microbial host cell culture with the vector of step b);
    d) culturing the host cell so that said immunoglobulin heavy and light chains are produced as separate molecules in said transformed host cell; and
    e) recovering the immunoglobulin from the host cell culture, said immunoglobulin being capable of binding to a known antigen.

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 rejection of the Cabilly II claims:

Upon reconsideration of the declarations of Harris, McKnight, Botchan, Colman and Rice, a person of ordinary skill in the art at the time the invention was made would not have had a reasonable expectation of success modifying the Cabilly I Patent claims in accordance with the teachings of Axel, Rice, Kaplan, Builder, Accolla, Dallas, [the] Moore patent, Deacon and Valle, and Ochi references of record to produce an immunologically functional immunoglobulin molecule by independently expressing immunoglobulin heavy chain and light chain in a single transformed cell.

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.

Comments

[Humorless Democrat]

Good post Kevin.

1. I'm perplexed per a point pertaining to patent prosecution: since when can an obviousness-type double patenting rejection be made over a combination of references? I believe that such a rejection is made on the basis of the *claims* of the earlier-issued patent (and *not* the entire disclosure thereof) against the claims of the pending application - in contrast to a normal 103 rejection where the entire teaching of a combination of references can be utilized. With o.t.d.p., we're not worried about the claimed invention being obvious per se over the prior art, we're worried about the patentee extending its period of exclusivity (under the old 17-years-from-grant system; since the switch to 20 years from earliest filing, the only reason for o.t.d.p. rejections is to avoid "vexatious litigation" by patentees parceling the patents out to different parties) by claiming in a later app what it could have claimed in an ealier app. In which case, there shouldn't be resort to other prior art.

2. You seem to be in favor of l etting the patentee enjoy a quiet time with its patent after grant. How about letting the rest of world get on with using what is now old technology? This is the second recent high-profile biotech case of which I'm aware (Amgen v Roche re EPO is the other) in which the patentee was able to effectively procure 30 years or more of patent protection. Here the original app was filed in 1983 - what in the blazes are we doing giving these guys exclusivity over recombinant antibody technology until 2018?! I realize that this is a hold-over from the old, evil 17-years-from-grant-and-let-Lemelson-submarine-us regime, and that as time passes we will see fewer and fewer of these cases, but still - where is the *public's* interest in this case being considered? This is OLD technology at this point, and I'm sorry that Genentech had to wait so long for a patent, but why the heck should we suffer high prices? Better, why are we forcing manufacture and sale of products covered by this process offshore, to places like Europe that had 20-year-from-earliest-filing terms all along? Here the USA's being out of step with ROW is letting it shoot itself in the foot.

[Here2Eternity]

Take a look at the Amendment filed by Genentech leading to the Notice of Intent to Issue Reexamination Certificate. You will see in the disclosure of related matters that there is a pending Cabilly III, also pre-GATT, which just emerged victorious in yet another interference with Boss. Boss, predictably, appealed. While it is not possible to know what the claims look like, Cabilly III could be in force until nearly 2030.

Anyone have a problem with this?

[Kevin E. Noonan]

Dear Here and Humorless:

Both of you raise the issue of how long the effective term for the Cabilly and Amgen patents have turned out to be. As you both recognize, this is in part a holdover from the prior regime, of granting 17 years from issue. Accordingly, if we had not changed the patent term to 20 years from earliest filing the result would be less striking.

However, the only way to solve the problem would to enact a hard limit on patent term, no matter what. This would disproportionately impact pharma compared with other technologies, given the differences in obsolescence horizons.

Also, for a pharmaceutical we can consider the following hypothetical. If the Cabilly patent was for a pharmaceutical, and if the 20 year term had been in effect at the time the application was filed, we could have had the following result. The term, would have expired in 2003 except that it would be extended for at least 8 years because of the interference (and maybe 10 years or so if the district court proceedings are included). This brings us to an expiration date in about 2010 to 2012; if you add 5 years of patent term extension for regulatory approval delay, we are out to the patent expiring in about 2016 or so.

Now, if Congress limits the maximum term to 25 years from filing (which is the minimum in view of the maximum Hatch-Waxman term extension), then the patentee would have lost about 7 years of term (about the same amount of time as the interference). Besides providing an avenue for delaying applications, this would penalize just those inventors working in the "hottest" areas (where there may be the most competition and the most likelihood or incentive for interferences).

I agree that the results for Cabilly and Amgen are anomalous, but how much do we want to restrict patent rights for all on the basis of the few exceptions? I think it is almost impossible to find an equitable "maximum" term life, but I would be happy to hear what you think.

Thanks for the comments.

[Kevin E. Noonan]

Dear Humorless:

To get to your first point, the Office seems to have performed the analysis as if it was a 103 rejection, with the consequence being the requirement to file a terminal disclaimer rather than a determination of unpatentability. I have seen that in at lease one other case, but I don't have the cite handy (I'll let you know if I find it).

Thanks for the comment.