Genentech Inc. defended two dosage patents for blockbuster breast cancer drug Herceptin against assertions of obviousness by Hospira Inc. by telling a Patent Trial and Appeal Board panel the Pfizer Inc. unit had failed to show a skilled practitioner in 1999 would have expected the drug to remain effective after three weeks.
Hospira lawyer Amanda Hollis of Kirkland & Ellis LLP told the panel the only novel invention now claimed by Genentech is the once-in-three-weeks injection timetable. But that isn’t even the primary invention claimed by the patent, which is the idea of front-loading a bigger dose in advance of smaller maintenance doses.
“Their patent’s life is on the line and this is the only thing they could come up with,” Hollis said.
But Genentech attorney Lisa J. Pirozzolo of WilmerHale countered that Hospira had failed to show that the dosing pattern is obvious given a broad array of uncertainties about targeted antibodies, at the time a fairly new therapy for cancer.
“[A person ordinarily skilled in dosing] would not risk efficacy by extending the dosing intervals [at that time],” Pirozzolo said.
Herceptin, federally approved under a weekly regimen since 1998, treats breast and gastric cancers, and it earned $2.5 billion in the U.S. in 2016 alone and $6.5 billion globally in 2015. Pfizer and other companies are closing in on producing biosimilars that could eat into that market, and are challenging various patents tied to the drug.
Right after the panel heard of Hospira’s challenge of U.S. Patent Numbers 6,627,196 and 7,371,379, it heard another in which Celltrion Inc. attacked the same patents on similar grounds, but with support from different studies and experts. Samsung Bioepis Co. Ltd. filed an inter partes review as well, which was later merged with Hospira’s IPR. Hospira also challenged two other patents tied to the drug in a case scheduled for oral arguments before the same panel on May 18.
Meanwhile, Mylan NV dropped its patent challenges as part of a licensing deal in which it will pay Genentech for the rights to produce a biosimilar.
Tuesday’s arguments revolved around whether an ordinarily skilled practitioner would have been motivated to pursue more spaced-out injections as more convenient and would have reasonably expected they would work. Genentech said Hospira had established neither, and Pirozzolo said every dosing patent challenge can’t simply rest on a vague convenience motive without evidence specific to the drug, and at that point efficacy was the far more important consideration in advancing the new drug.
But Administrative Patent Judge Chris Paulraj said he found Genentech’s argument against motive less compelling than its argument that Hospira couldn’t prove an expectation of success. Most of the discussion centered on the appropriateness of Hospira’s experts’ method of coming up with the same doses described in the patent using what was known in 1999. Hospira said it showed how obvious the conclusion was, while Genentech said it cherry-picked to work backwards from the conclusion.
Paulraj asked Hollis why Genentech wouldn’t have simply started with the three-week regimen if it was so desirably convenient and obvious. She said Genentech itself admitted it didn’t stop looking to improve the drug after the initial weekly dose was FDA-approved, showing a motive to address convenience. She also said the level of the antibody — called trastuzumab — needed in the blood to remain effective was known, and that textbook equations were enough for a skilled expert to reach the three-week conclusion.
“They have raised so much dust that’s baseless,” Hollis told the panel of a variety of efforts to discredit Hospira’s experts’ methodology.
Genentech argued that the linear model used by Hospira’s expert to reach his conclusion was undercut by the fact that the drug is non-linear: It doesn’t dissipate at a constant rate over time.
Hollis countered that among the only evidence backing the attack on its expert’s method was a chart Genentech’s expert drew just for the litigation to demonstrate how a non-linear drug could vary from linear. Further, that graph was designed solely for litigation purposes, she said, and it depicted an example of a drug’s concentration in the blood falling off a cliff — the opposite of antibodies, whose rate of dissipation tapers off over time.
Pirozzolo responded that Genentech’s expert had only drawn the graph to show linear assumptions were inaccurate for nonlinear drugs. She also said Hospira cherry-picked a high-end estimate of the drug’s durability to reach its conclusion, which risked overstating how much trastuzumab remained in the blood and hurting efficacy.
Hollis said the estimate was well within the range known from past literature at the time. She also said aside from the graph made up for litigation, the only other evidence offered undercutting one of their expert’s testimony was published seven years after the patent was filed, and involved a small-molecule drug instead of a biologic.
Both sides accused each other of introducing impermissible new arguments while contending they were only responding to explain the other’s error. Pirozzolo said the original sin in the dispute was the faulty reliance on a linear method and that Hollis overstated what was known about antibody dissipation at the time. Hollis defended Hospira’s expert’s methodology as essentially the same as what is suggested by prior studies and the patent itself.
Counsel for both sides and a spokesperson for Pfizer declined to comment. A representative for Genentech did not respond to a request for comment.
The patent at issue is U.S. Patent Number 6,627,196; and 7,371,379.
Hospira is represented by Amanda Hollis, Ben Lasky, Stefan Miller, Karen Younkins, Sarah K. Tsou, Mark McLennan and Christopher Citro of Kirkland & Ellis LLP.
Genentech is represented by David L. Cavanaugh, Owen K. Allen, Robert J. Gunther Jr., Lisa J. Pirozzolo, Kevin S. Prussia, Andrew Danford and Robert M. Galvin of WilmerHale, and Adam R. Brausa and Daralyn J. Durie of Durie Tangri LLP.
The case is Hospira Inc. v. Genentech Inc., case numbers IPR2017-00804 and IPR2017-00805, at the Patent Trial and Appeal Board.
REPRINTED WITH PERMISSION FROM THE MAY 8, 2018 EDITION OF LAW360 © 2018 PORTFOLIO MEDIA INC. ALL RIGHTS RESERVED. FURTHER DUPLICATION WITHOUT PERMISSION IS PROHIBITED. WWW.LAW360.COM