October 2010
Selection of and Motivation to Modify a Lead Compound Follows from the Possession of Useful Properties, Not Mere Structural Similarity
| Judges: Lourie (author), Friedman, Linn |
| [Appealed from D.N.J., Judge Martini] |
In Daiichi Sankyo Co. v. Matrix Laboratories, Ltd., No. 09-1511 (Fed. Cir. Sept. 9, 2010), the Federal Circuit affirmed the district court’s ruling that the patent at issue was not invalid as obvious under 35 U.S.C. § 103. Defendants Matrix Laboratories, Ltd., Mylan Inc., Mylan Laboratories, Inc., and Mylan Pharmaceuticals, Inc. (collectively “Mylan”) failed to establish a prima facie case for the selection of a lead compound and, even if selected, failed to establish a motivation to modify the lead compound to obtain the claimed compound.
Daiichi Sankyo Company, Ltd. and Daiichi Sankyo, Inc. (collectively “Daiichi”) own U.S. Patent No. 5,616,599 (“the ’599 patent”), which claims compounds and their use as angiotensin receptor blockers (“ARBs”) for treating high blood pressure. Claim 13 covers the compound olmesartan medoxomil, an FDA-approved drug and the active ingredient in Benicar®, Benicar HCT®, and Azor®.
The invention of olmesartan medoxomil follows from the development of small-molecule ARBs beginning in the late 1970s. Years later, scientists at E.I. du Pont de Nemours and Company (“DuPont”) invented the first orally active ARB, losartan. Like the earlier compounds, losartan contained an imidazole ring core, but DuPont modified the 1- and 5-position substituents found in a lead compound. In its patent covering losartan, DuPont disclosed more than 400 related compounds as well as binding affinity data for more than half of the compounds. The data were used by chemists to correlate chemical structure with activity (“structural-activity relationships,” or SARs), which could be used to develop improved compounds.
More than twenty pharmaceutical companies launched research programs for ARBs; Daiichi’s program resulted in the synthesis of olmesartan medoxomil. Olmesartan shares the imidazole ring, 1-position biphenyltetrazole substituent, and 2-position alkyl group of losartan. Olmesartan differs from losartan, however, at both the 4- and 5-positions.
At the 4-position, olmesartan replaces losartan’s lipophilic chlorine atom with a hydrophilic hydroxyisopropyl group. The vast majority of compounds disclosed in DuPont’s patent contained lipophilic groups at the 4-position. One compound disclosed in the DuPont patent with a hydrophilic group at the 4-position is a regioisomer of losartan, in which the 4- and 5-position substituents were exchanged.
At the 5-position, olmesartan replaced losartan’s hydroxymethyl group with a masked carboxy group, carboxy medoxomil. In the body, the medoxomil group is removed to yield the carboxylic acid group. Losartan’s hydroxymethyl group is also metabolized to a carboxylic acid in the body.
Some second-generation ARBs, which are prior art to olmesartan, replaced losartan’s 4-position chlorine atom with other lipophilic groups, such as alkyl or perfluoroalkyl groups. The court determined that ARBs disclosed in another patent to DuPont, U.S. Patent No. 5,137,902 (“the ’902 patent”), are the closest structurally to olmesartan. Example 6 of the ’902 patent differs only by the replacement of a hydroxy group with hydrogen. Other second-generation ARBs vary more significantly by using, for example, a different core ring than imidazole.
Mylan filed several ANDAs and served Paragraph IV certifications challenging the validity of the ’599 patent. Daiichi filed suit for patent infringement. The parties stipulated to infringement of claim 13, but the case proceeded on Mylan’s counterclaim of invalidity for obviousness. The district court ruled after a bench trial that the ’599 patent was not invalid as obvious, and Mylan appealed.
Mylan first challenged the lower court’s finding that one of skill in the art would not have chosen the ARBs in DuPont’s ’902 patent as lead compounds. The Federal Circuit, however, affirmed the ruling that Mylan failed to show that one of ordinary skill in the art would have selected the ’902 patent’s ARBs as lead compounds. The Court accepted as true that the ’902 compounds represented a continuation of the disclosure and the data found in the earlier DuPont patent, and recognized that the compounds exhibited about 2- to 4-fold higher activity than the most active compounds in DuPont’s previous disclosure. However, other second-generation ARBs exhibited even greater activity and had been more thoroughly studied than the DuPont ARBs. Thus, the Court found that a skilled artisan would have selected the other secondary ARBs, which exhibited 7-, 100-, and 180-fold higher activity to be a lead compound, not the ’902 compounds.
“[P]roving a reason to select a compound as a lead compound depends on more than just structural similarity, but also knowledge in the art of the functional properties and limitations of the prior art compounds.” Slip op. at 14.
The Court did not accept Mylan’s argument that because the ’902 compounds have the closest structure in the prior art, that should be dispositive for finding them to be lead compounds. Instead, the Court emphasized that “it is the possession of promising useful properties in a lead compound that motivates a chemist to make structurally similar compounds.” Slip op. at 14. To choose the ’902 compounds as lead compounds would suffer from hindsight bias. The state of the art at the time of the invention must be the basis for finding motivation to select and then modify a lead compound. The Court stated that “[p]otent and promising activity in the prior art trumps mere structural relationships.” Id. at 15. The oral activity, binding activity, and selectivity data, for example, among the other second-generation compounds would motivate their selection as lead compounds over the ’902 compounds. A court is not required to find a single, best lead compound. Here, several compounds were selected as leads, and the ’902 compounds were not included in that set. The lower court did not commit clear error in reaching this finding.
Mylan next challenged the finding that one would not be motivated to modify the ’902 patent’s compounds at the 4- and 5-positions where the first DuPont patent specifically taught a hydroxyalkyl group at the 4-position, and the art taught medoxomil as a prodrug providing improved oral activity. The Federal Circuit, however, again affirmed the lower court’s finding that even if the ’902 compounds were selected as lead compounds, one of skill in the art would not be motivated to modify them to obtain olmesartan medoxomil.
The Federal Circuit first explained that the prior art as a whole taught away from the use of a hydrophilic substituent at the 4-position of the imidazole ring. The SAR data and the use of lipophilic groups at this position in the other second-generation compounds would teach away from modifying the ’902 compounds’ lipophilic alkyl groups to the hydrophilic hydroxyisopropyl group of olmesartan. The DuPont data revealed a clear preference for lipophilic groups at the 4-position. Three subseries analyzing the binding affinity as a function of the 4-position substituent confirmed the preference for having a lipophilic group. “Thus, the compounds in the prior art, including [Mylan’s] proposed lead compounds, favor lipophilic 4-substituents rather than the 4-hydrophilic group of olmesartan medoxomil.” Id. at 17.
Regioisomers that transpose the 4- and 5-position substituents and DuPont’s second-generation ARBs all demonstrated the preference for lipophilicity at the 4-position. No other second-generation ARB besides olmesartan had a hydrophilic group at this position. Mylan argued that the motivation to modify was nonetheless found in one of the DuPont example compounds having a hydrophilic group at the
4-position. The Court found, however, that the SAR data contradicted this, and that Mylan’s argument relied on selecting the ’902 compounds, which improved on losartan by using even more lipophilic groups at the 4-position, only to reject that very feature to obtain olmesartan medoxomil.
