Can Kite Pharma reverse its IPR loss challenging Juno’s CAR-T patent?

Zachary Silbersher

On June 5, 2018, the Federal Circuit will hear oral argument on Kite Pharma’s ($GILD) appeal of its IPR loss related to Sloan-Kettering’s and Juno Therapeutic’s ($CELG) CAR-T patent.  What issues will it be important to listen for in the panel’s questions?

This appeal arises from a petition for inter partes review (IPR2015-01719) filed by Kite against a patent assigned to Sloan, namely, U.S. Patent 7,446,190.  The appeal could bear directly upon the pending case between Juno Therapeutics and Kite Pharma, filed in September 2017 (17-7639, C.D. Cal.).  In that case, Juno is suing Kite for infringing the ‘190 patent through sales of its Yescarta™ (KTE-C19) drug.

The ‘190 patent claims a CAR therapy that incorporates three specific elements, scFv, CD28, and CD3ζ.  Each of these elements had previously been incorporated into earlier CAR therapies.  Indeed, prior CAR-T therapies had included all three elements (scFV-CD28- CD3ζ.)  The question on appeal, and previously at issue in the IPR, is whether this combination was obvious.

A chimeric antigen receptor, otherwise known as “CAR” or chimeric T cell receptor, is an anti-cancer technology that reprograms T-cells in a patient’s immune system to attack and kill tumor cells.  The ‘190 patent does not purport to invent CAR technology; rather, it was generally previously known in the art.  Rather, certain CAR therapies were known to be deficient because they did not achieve full T cell activation and sustained immune response.  This problem was known as “anergy.”  The ‘190 patent, therefore, purports to have invented a solution to the anergy problem. 

T cells can naturally target and kill cancer cells.  On their own, however, they do not effectively combat cancer because they are rare and easily suppressed.  Thus, CAR-T technology endeavors to reprogram T-cells to more effectively do what they already can, namely, kill tumor cells.  The scFv element (single chain variable fragment) is a portion of an antibody that allows CAR-T to bind to a specific target.  The CD3ζ element send signals once the CAR-T cell is bound to the target to kill the targeted tumor cell.  It was known, however, that adding a co-stimulatory signal, CD28, could improve T cell activation, and thus the efficacy of CAR-T therapies.

Kite Pharma relied primarily on two prior art references in its attempt to invalidate the ‘190 patent: Krause and Finney.  Krause taught a CAR-T with scFV-CD28, which used the same version of CD28 claimed in the ‘190 patent.  Finney taught a CAR-T with scFV-CD28- CD3ζ.  Though the CD3ζ element was the same as claimed in the ‘190 patent, the CD28 version was different.  The following illustration from Kite Pharma’s opening brief illustrates the prior art compared to the ‘190 patent.  (See Case: 17-1647, Dkt. 23, p. 25, Federal Circuit).

During the IPR, Kite Pharma started with the Krause reference (scFv-CD28), and simply argued that it would have been obvious from Finney to add CD3ζ.  Kite showed that the prior art taught CD3ζ was required to provide a primary signal for cell activation and target killing.

The Board, however, rejected Kite’s theory.  The Board acknowledged that the prior art taught all of the claimed elements, even if the prior art did not teach them all together, i.e., with the precise CD28 version combined with scFV and CD3ζ.  Yet, the Board found that Kite had failed to show that it would have been obvious to keep Krause’s CD28 sequence in lieu of the CD28 sequence taught in Finney.  Kite argues that by doing so, the Board committed error.  This is the central question at issue on this appeal.

On appeal, Kite argues that the Board found that the art taught a CAR-T that was preferable to that claimed in the ‘190 patent, namely, the scFV-CD28- CD3ζ construct that used Finney’s version of CD28 rather than Krause’s version of CD28.  Kite, therefore, argues that this was error.  On the law, Kite is likely correct.  A patent can be obvious for disclosing a solution to a problem, even if that patented solution was neither the best nor most preferable one at the time.  In other words, just because there were better or preferred solutions to a problem does not necessarily mean that a patent claiming a less preferred solution is still not invalid.  Even less preferred solutions can still be obvious.

Yet, while Kite may be correct on that legal principle, it may still lose on the facts.  Sloan argues that Kite ignores the real reasoning of the Board’s decision as well as the facts behind it.  The Board did not reject Kite’s IPR because it found there were preferred CAR-T constructs to that claimed in the ‘190 patent.  Rather, it rejected Kite’s invalidity theory because the prior art actually taught away from the invention claimed in the ‘190 patent.

The ‘190 patent claims a CAR-T construct with a specific CD28 sequence.  That sequence includes a portion of CD28’s extracellular domain, which itself includes the “MYPPPY” amino acid motif.  This motif was known to be CD28’s binding site to its natural ligands.  Thus, the ‘190 patent claims a CAR-T construct that has two binding sites—the selected cancerous target as well as a second binding site, the MYPPPY motif.

The reason this is important is because dual-signal CAR-T constructs, like that claimed in the ‘190 patent, were known to be potentially more effective, but also more dangerous because of that.  Dual CAR-T constructs can be very powerful, killing tumor cells and persisting.  Yet, there was a known risk to be associated with that.  They could also kill “off target” healthy cells.  They could essentially turn into an army of kills that destroy cancer cells, but also unintended, non-cancerous cells. 

This is why inclusion of the MYPPPY motif in the ‘190 patent’s recital of CD28 is important.  Sloan presented evidence that scientists were concerned that including the “MYPPPY” motif, which was known to be CD28’s binding to its natural ligands, might turn the CAR-T construct into a killer of cancer cells, but also a killer of healthy cells as well.

Thus, Sloan’s retort is a powerful one.  That said, Kite manages to throw some shade on Sloan’s defense of the Board’s decision to uphold its patent.  Most importantly, Kite points out that the prior art did not specifically teach away from using the CD28 element claimed in Krause, with the MYPPPY motif.  Kite argues that the evidence of teaching away was speculative, not included in the primary references, and perhaps most importantly, it primarily came after the ‘190 patent.  Overall, teaching away evidence typically must be pretty clear and unambiguous.  It will be important to listen to whether the Federal Circuit panel picks up on this issue, and questions Sloan about this during oral argument.

Therefore, this appeal boils down essentially to a single issue – was it obvious for the ‘190 patent to include the claimed CD28 element containing the MYPPPY motif?  On the one hand, Kite sets up a clean argument that makes sense from a patent law perspective.  Starting with Krause (scFv-CD28), it would have been obvious to add the CD3ζ from Finney.  Yet, the Board rejected this argument, and Kite argues that it did so because it found that using Finney’s version of CD28 would have been preferred.  If those were the facts, then Kite would likely prevail. 

But, it is far from clear that those are the facts.  Sloan rather argues that there is more to the story.  The Board did not find that another solution to that claimed in the ‘190 patent was preferred.  Rather, it found that the art taught away from the solution claimed in the ‘190 patent.  This is because the CD28 version claimed in the ‘190 patent includes the MYPPPY motif.  That motif risked the CAR-T construct binding to ligands other than the targeted cancer cells.  Because scientists were concerned that could cause more damage than it was worth, the Board found that the construct claimed in the ‘190 patent was not, in fact, obvious.

The oral argument is scheduled for June 5.  The panel’s questioning will provide important insight into how the Court may be leaning.  It will be especially important to listen for two issues: (1) whether the panel questions Kite Pharma on why it does not view the prior art teaching away from using the MYPPPY motif; and (2) whether the panel questions Sloan on whether the prior art really did teach away from using the MYPPPY motif, or if that was really after-the-fact speculation.