When inventors are wrong, a certificate of correction can set things right—if the specification is robust
Ryan Chirnomas | January 7, 2016
Cubist Pharmaceuticals, Inc. v. Hospira, Inc.
November 12, 2015
Precedential Opinion by Bryson, joined by Wallace and Hughes
In prosecution of a claim reciting an antibiotic, a certificate of correction to correct an error in a structural diagram of the antibiotic was not invalid. This was because the specification adequately described the intended subject matter, and it was clear to those skilled in the art that the inventors were in possession of the corrected structure. As such, the certificate of correction did not alter the scope of the claims.
Cubist owns five patents at issue, including RE39,071 (the ‘071 patent), and sells a daptomycin antibiotic product marketed as Cubicin. Hospira filed an ANDA challenging the validity of the five patents. The district court held that four of the patents were invalid and not infringed, but held that the ‘071 patent was valid and infringed despite Hospira’s challenge to a certificate of correction.
The reissue ‘071 patent issued in 2006, and includes claims that recite a combination of formula 1, formula 2 and formula 3. However, the inventors later discovered that there was an error in formula 3, reproduced below:
Formula 3 is described in the specification in three ways. First, it is described as an A-21978C cyclic peptide prepared from antibiotics described in another patent. That other patent explains that the antibiotics are produced by fermentation of the bacterium Streptomyces roseosporus. Second, formula 3 is described by the code name LY146032, which was assigned by Eli Lilly and known in the art to correspond to daptomycin. Finally, the structure above was used, where RN is decanoyl.
However, formula 3 was incorrect in that L-Asn (L-asparagine), circled above, should have instead been D-Asn, to represent the D stereoisomer. It was widely believed at the time of filing of both the original application and the reissue application that daptomycin includes L-Asn. It was only discovered later that daptomycin actually includes D-Asn, not L-Asn. Thus, in 2007, Cubist requested a certificate of correction by explaining this to the examiner, and the request was granted.
In litigation, Hospira argued that the certificate of correction was invalid, and the claims should be interpreted such that formula 3 includes L-Asn. Under such an interpretation, formula 3 recites a different and much less potent antibiotic than daptomycin, and Hospira avoids infringement. However, the district court agreed with Cubist and held that the USPTO merely corrected an error in the diagram without changing the scope of the claims, since the specification makes it clear that formula 3 is in fact daptomycin.
The CAFC agreed with Cubist and the district court, and held that the USPTO merely corrected the diagram, noting that a chemical structure is “simply a means of describing a compound, and is not the invention itself.” The CAFC stressed that the entire specification must be considered, and that formula 3 is not only defined by the diagram, but is also described by other portions of the specification.
In particular, the specification refers to a co-pending application, which explains that daptomycin is obtained by fermentation of Streptomyces roseosporus. That fermentation process only results in daptomycin (including D-Asn), and cannot result in a variant with L-Asn. In fact, the variant with L-Asn can only be produced synthetically. Furthermore, evidence showed that the code name LY146032 used in the specification refers only to daptomycin, and not the variant with L-Asn.
Hospira relied upon Bayer v. Dow Agrosciences. In that case, the inventors claimed a DNA sequence to an enzyme described functionally as having monooxygenase activity, without reciting structural features. However, it was discovered during prosecution that the enzyme in question actually had dioxygenase activity, but this was never corrected. The CAFC rejected Bayer’s argument that “monooxygenase” should not have its ordinary meaning, and should instead mean “monooxygenase” or “dioxygenase.” In the present case, the CAFC held that Bayer does not control, since Bayer was seeking broad functional claim construction which was not supported by the specification and was inconsistent with the accepted scientific meaning.
Therefore, the CAFC held that formula 3 in the claims refers to daptomycin, both before and after the certificate of correction. Hospira did not meet the high evidentiary burden of clear and convincing evidence required to invalidate a certificate of correction.
Next, Hospira argued that even if the certificate of correction was granted, the claims then are invalid for failing to comply with the written description requirement. Hospira relied upon In re Wallach, where the inventors were only shown to be in possession of about 5% of the DNA encoding a protein having particular characteristics, but sought to claim the DNA encoding the entire protein.
However, the CAFC rejected this argument by pointing out that Cubist only claimed what they produced, and they identified the produced material in several ways, such as the disclosure of the formation product of Streptomyces roseosporus and the LY146032 code name. Unlike in Wallach, the specification disclosed “relevant identifying characteristics that distinguished daptomycin from other compounds” and thus demonstrated that the inventors had possession of the claimed subject matter. As such, the claims comply with the written description requirement.
Finally, Hospira argued that the claims of RE39,071 are invalid because they violate the recapture rule. In prosecution of the parent, a claim was presented which did not require formulas 1 and 2 but did require formula 3. This claim was rejected three times, but only based on indefiniteness, not based on prior art. The claim was eventually cancelled.
The CAFC cite In re Mostafazadeh and In re Clement to reiterate that the recapture rule only applies if the reissue claims are broader than the original claims and the broader aspect relates to subject matter which was surrendered during prosecution to overcome a prior art rejection. Since the claim at issue was significantly narrower than the original claim due to inclusion of formula 1 and formula 2, and since the original claim was not cancelled to overcome a prior art rejection, the recapture rule does not apply. Therefore, for the above reasons, the ‘071 patent is not invalid, and Hospira infringes.
Furthermore, the CAFC held that the district court was correct in holding that Hospira did not infringe the other four patents, and that these patents were invalid. However, that issue is beyond the scope of this article.
Although typically non-controversial, a certificate of correction may have a huge impact to save a patent. It is possible that if a certificate of correction had not been filed, it would have been very difficult for Cubist to obtain the desired claim construction in litigation. Thus, errors should be corrected as soon as possible, particularly where they are due to an initial misunderstanding of the science.
It appears that the CAFC will be far more forgiving of corrections to structural features of an invention than corrections of functional features of an invention. In view of this, in emerging technologies, it is best to try to claim the subject matter both functionally and structurally if possible.
If working in emerging technologies, it may be beneficial to describe the subject matter in multiple ways in a robust specification. Thus, if there is an error in one manner of description which is used in the claims, the other descriptions may help to clarify.