Pfizer Inc. et al. v. Teva Pharms. USA et al., No. 2012-1576 (Fed. Cir.)
Teva and several other generic pharmaceutical companies each submitted an ANDA seeking approval to market a generic version of Lyrica, for treating seizures and pain. In a unanimous decision last Thursday, the Federal Circuit affirmed the judgment of the U.S. District Court for the District of Delaware by finding claim 2 of U.S. Patent Nos. 6,197,819, covering the active ingredient of Lyrica, pregabalin, infringed and not invalid for lack of enablement, written description, or obviousness.
Pfizer and Northwestern University initally asserted four Orange Book-listed patents against the ANDA filers, but the case ultimately hinged on claim 2 of the '819 patent, which reads quite simply:
2. 4-amino-3-(2-methylpropyl) butanoic acid, or a pharmaceutically acceptable salt thereof.
The compound of claim 2 contains a single chiral center, and may exist, for example, in enantiomerically pure (R)- or (S)- forms or a racemic form, the latter containing equal quantities of both enantiomers. Lyrica and the proposed generic products each contain pregabalin, which is only the S-enantiomer of 4-amino-3-(2-methylpropyl) butanoic acid. Before the district court, the defendants asserted that claim 2 of the '819 patent should be interpreted to cover only the racemic form of 4-amino-3-(2-methylpropyl) butanoic acid (also "3-isobutyl-gamma-aminobutyric acid" or "3-isobutylGABA"). If so construed, the proposed products would not literally infringe the claim. Further, the defendants alleged that claim 2 is invalid for failing the enablement and written description requirements of 35 U.S.C. 112(a) and for obviousness under 35 U.S.C. 103 in view of three asserted prior art references.
The district court's Markman order construed claim 2 of the '819 patent to encompass all stereochemical forms of 3-isobutylGABA. The parties stipulated to infringment by the defendants' proposed products if claim 2 was found valid and enforceable, as interpreted by the district court. This appeal followed the district court's finding that claim 2 of the '819 patent was enforceable and not invalid for the reasons offered by the defendants. On appeal, the defendants asserted that the district court erred in claim construction and its findings with respect to enablement, written description, and obviousness. The Federal Circuit affirmed the district court on all issues.
With respect to claim construction and infringement, the Federal Circuit declined to read limitations into claim 2 that were not present in the claim text. After noting that the defendants' expert admitted that claim 2 covers "3-isobutylGABA in any isomeric form," the court stated that the '819 patent made clear that the patent applicants were capable of indicating when individual enantiomers and when racemic forms of compounds were intended, by using the appropriate (R,S)- or (R)- or (S)- prefixes. Nor could the court find any clear disavowal in the '819 patent that would indicate that a lack of stereochemical indication was intended to limit the disclosure to only the racemic form. In view of the lack of any stereochemical limitation within claim 2, the court upheld the district court's interpretation and finding of infringement of claim 2 of the '819 patent.
On enablement, the ANDA filers asserted that if claim 2 were interpreted to cover all compositions of 3-isobutylGABA, regardless of enantiomeric form, then the specification failed to "teach a skilled artisan how to prepare every conceivable mixture of 3-isobutylGABA’s enantiomer." However, the defendants failed to convince the court that such mixtures were anything other than routine to one of ordinary skill in the art. The court noted that the '819 patent disclosed the method for synthesizing the compound and states that the compound's "enantiomers may be prepared or isolated by methods already well known in the art." Citing In re Hogan, 559 F.2d 595, 606 (CCPA 1977), the court declined to require an applicant to disclose "a detailed recipe for preparing every conceivable permutation of the compound they invented to be entitled to a claim covering that compound."
For written description, Appellant Sun Pharmaceuticals (only) focused on a lack of disclosure by the '819 patent regarding isolation of the individual enantiomers. Sun proposed that since later application filings by the patentee indicated that enantiomer separation for 3-isobutylGABA may have been less than routine, and that the inventors admitted that at the time of filing of their original applications, they had not yet actually separated the 3-isobutylGABA enantiomers, then the applicants were not in possession of the broader claim scope. However, the court found Sun's focus on the enantiomers to be unpersuasive. Rather, the court focused on whether "relevant identifying characteristics" were provided such that "persons of ordinary skill in the art  recognize that the inventor invented what is claimed." Here, the court noted that disclosure of the claimed structure (in the absence of stereochemical definition), in vitro and in vivo data for the compound, and a method of synthesizing the compound satisfied the statutory requirements.
Finally, the ANDA applicants argued that the district court erred in finding claim 2 of the '819 patent not obvious in view of three proffered prior art references. The references were alleged to disclose that 3-isopropylGABA and other homologous compounds may have anticonvulsant activity, that 3-isobutylGABA would have been expected to have similar activity because of its close structural simlarity to the former, and that the anti-convulsant gabapentin, another 3-substituted gamma-amino butyric acid (GABA) derivative, provided motivation to try other substitutions as the 3-position of GABA.
The court applied the two-part test for obviousness of a chemical compound outlined in Eisai and Takeda by reviewing for (1) whether the asserted prior art compound would have been selected as a lead compound for further modification and (2) whether the prior art would have taught a skilled artisan to make the "specific molecular modifications" necessary to the selected lead compound to yield the claimed compound with a reasonable expectation of success.
As to the first prong, the court reiterated that more than "mere structural similarity" is necessary for a prior art compound to serve as a lead compound. Rather, other properties, such as chemical activity or potency, inform the lead compound selection. Here, the court was unmoved by the offer of either 3-isopropylGABA or gabapentin as a lead compound. The court found insufficient details in the cited art regarding properties of either compound that would have lead to their selection for further research. With respect to gabapentin, the prior art was vague as to whether the compound was in the prior art at all, and if so, whether any of its relevant properties, such as anticonvulsant activities in the context of other anticonvulsants known at the time would have highlighted the compound for further modification. For 3-isopropylGABA, the court noted that the prior art failed to teach its "mechanisms of action, including whether it has anticonvulsive properties."
Moving to the second prong, the court noted that even if either compound were selected as a lead compound, the proffered art failed to teach the precise modifications necessary to yield 3-isobutylGABA. Rather, the court noted that the cited art merely suggest "trillions" of compounds "without calling out alkyl groups in particular or singling out isobutyl specifically." The court concluded by agreeing with the district court's finding that the appellees had convincingly established the complicated and unpredictable nature of anticonvulsant field at the time of filing, thereby precluding any reasonable expectation of success in achieving anticonvulsant in 3-isobutylGABA "even if Appellants were able to establish that the prior art taught the substitution of isobutyl at GABA’s 3-position."