`
`
`
`
`
`IN THE UNITED STATES PATENT AND TRADEMARK OFFICE
`_____________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_____________________
`
`
`ELI LILLY AND COMPANY,
` Petitioner
`
`v.
`
`TEVA PHARMACEUTICALS INTERNATIONAL GMBH,
`Patent Owner.
`
`_____________________
`
`Case IPR2018-01710
`Patent 8,586,045 B2
`_____________________
`
`TEVA PHARMACEUTICALS INTERNATIONAL
`GMBH’S PATENT OWNER RESPONSE
`
`
`
`
`
`Mail Stop “PATENT BOARD”
`Patent Trial and Appeal Board
`U.S. Patent and Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`

`

`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`
`TABLE OF CONTENTS
`
`I.
`
`Introduction ..................................................................................................... 1
`
`II. Background................................................................................................... 7
`
`A.
`
`B.
`
`Calcitonin Gene-Related Peptide (“CGRP”) ........................................ 7
`
`By 2005, there was no consensus as to the pathophysiology of
`migraine ................................................................................................. 8
`
`C.
`
`The ’045 patent and its commercial embodiments ............................... 8
`
`III. Claim construction ........................................................................................ 10
`
`A.
`
`B.
`
`“reducing incidence of or treating” .....................................................10
`
`"effective amount" ...............................................................................10
`
`IV. The challenged claims are not prima facie obvious ..................................... 11
`
`A.
`
`B.
`
`C.
`
`Lilly mischaracterizes what the prior art discloses .............................11
`
`There was uncertainty in the field regarding CGRP being a
`biomarker for migraine—a fact that Lilly ignores ..............................17
`
`Lilly incorrectly extrapolates Olesen’s small-molecule receptor
`antagonist results to any CGRP antagonist, in particular, an
`anti-CGRP antibody ............................................................................20
`
`1.
`
`2.
`
`3.
`
`Olesen’s “proof of concept” study would not have given
`a POSA a reasonable expectation that an anti-CGRP
`antibody would be effective in treating migraine .................... 20
`
`There would not have been an expectation of any
`therapeutic advantages for using an anti-CGRP antibody
`to treat migraine ....................................................................... 24
`
`Tan’s data negates any reasonable expectation that a full-
`length antibody would safely treat migraine ........................... 35
`
`- i -
`
`

`

`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`
`A POSA also would not have expected a full-length
`antibody to be efficacious because the field in 2005 was
`moving toward a central site of action for anti-migraine
`drugs ......................................................................................... 38
`
`None of the additional art Lilly cites provides a
`reasonable expectation that an anti-CGRP antibody can
`be safely used for treating migraine in humans ....................... 42
`
`4.
`
`5.
`
`D.
`
`E.
`
`Federal Circuit precedent demands a finding of non-
`obviousness .........................................................................................45
`
`Lilly failed to address a motivation to humanize Tan’s Fab’
`fragment ...............................................................................................47
`
`F.
`
`Lilly failed to prove that a POSA would have arrived at the
`affinity (KD) to CGRP, as claimed in claims 4 and 20 ........................50
`Strong objective evidence compels finding non-obviousness of the
`challenged claims .......................................................................................... 55
`
`V.
`
`A.
`
`B.
`
`C.
`
`D.
`
`E.
`
`F.
`
`The challenged claims have a presumption of nexus to the
`objective indicia of nonobviousness ...................................................55
`
`The claimed methods have received industry-wide acclaim ..............56
`
`The claimed methods satisfied a long-felt, unmet need ......................58
`
`The claimed methods achieved unexpected results ............................60
`
`Humanized anti-CGRP antibodies faced industry skepticism ............61
`
`Commercial success reinforces the non-obviousness of the
`claimed invention ................................................................................62
`
`G. AlderBio’s decision to take a royalty-bearing license to Teva’s
`patents supports nonobviousness ........................................................63
`
`H.
`
`Lilly’s simultaneous invention argument is not supported by the
`facts or the law ....................................................................................64
`
`VI. Lilly has failed to carry its burden so the Board must find for Patent
`Owner ............................................................................................................ 65
`
`- ii -
`
`

`

`Teva Pharmaceuticals International GmbH timely submits this Patent Owner
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`
`
`Response to the Petition for Inter Partes Review of U.S. Patent No. 8,586,045 filed
`
`by Eli Lilly and Company. Teva’s Response is supported by the expert declarations
`
`of Drs. Michel Ferrari, Ian Tomlinson, Steven Foord, Alan Rapoport, and Robert
`
`Stoner, and fact declaration of Jaume Pons. EX2268, ¶¶4-11; EX2271, ¶¶4-9;
`
`EX2265, ¶¶1-13; EX2262, ¶¶4-11; EX2274, ¶¶1-4; EX2331.
`
`I.
`
`Introduction
`
`The challenged claims recite novel methods of treating vasomotor
`
`symptoms, including migraine, using humanized anti-Calcitonin Gene-Related
`
`Peptide (“CGRP”) antagonist1 antibodies. Patent Owner Teva’s discovery was a
`
`breakthrough, representing the first time that anyone, anywhere developed a
`
`humanized anti-CGRP antibody for human therapeutic use. As a result, the
`
`approved use of Teva’s Ajovy® (fremanezumab-vfrm)—a commercial
`
`embodiment of the challenged patent claims—was the first and only approved use
`
`of an anti-CGRP drug for treating migraine.2 Confronted with the pioneering
`
`nature of Teva’s invention, Lilly resorts to hindsight to build its faulty obviousness
`
`case.
`
`
`1 Antagonism is inhibition of a physiological function. EX2265, ¶24.
`
`2 Ajovy® is indicated for the preventive treatment of migraine. EX2168, 1.
`
`- 1 -
`
`

`

`In deciding to institute trial, the Board accepted Lilly’s allegations that a
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`
`
`skilled artisan (“POSA”3) would have been motivated to treat migraine using a
`
`humanized anti-CGRP antibody based on Olesen’s data related to a small-molecule
`
`CGRP receptor antagonist, BIBN4096BS, and Tan—a basic laboratory research
`
`paper attempting to “prob[e] the role of CGRP as an endogenous vasodilator” in
`
`rats. EX1022, Abstract. But institution was based on a limited record reflecting
`
`Lilly’s one-sided rendition of the facts. As shown in this Response, Lilly’s
`
`arguments suffer from multiple flaws and missing steps that defeat its obviousness
`
`case. Upon consideration of the full record, the Board should find that substantial
`
`evidence exists to revisit and reverse its preliminary determination, which was
`
`based on a limited record. See Apotex Inc. et. al. v. Novartis AG, IPR2017-00854,
`
`Paper 109, 19, 23-24 (PTAB July 11, 2018). The full record demonstrates that a
`
`POSA in November 2005 would not have viewed an anti-CGRP antibody as
`
`sufficiently safe or effective for reducing incidence of or treating migraine.
`
`Lilly cherry-picks aspects of the art, ignoring important uncertainties and
`
`complexities in the field that negate obviousness. Lilly starts by arguing that CGRP
`
`was proven to be involved in migraine, citing Goadsby’s papers (EX1043,
`
`EX1044), while ignoring contradictory data that arose after Goadsby and before
`
`
`3 Teva adopts the Board’s definition of a POSA. See Decision, 7-9.
`
`- 2 -
`
`

`

`
`the patent filing showing no correlation between CGRP blood levels and migraine
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`(EX2309). Lilly’s expert Dr. Charles knew about such data, yet admittedly chose
`
`to not address them. And this is just one instance of Lilly’s and Dr. Charles’
`
`repeated “selective” presentation of the art.
`
`Lilly and its experts also ignore other pertinent prior art teachings that would
`
`have led a POSA to doubt whether an anti-CGRP antibody could be used to
`
`successfully treat migraine. At the outset, Lilly fails to explain why a POSA would
`
`have had a reasonable expectation that an anti-CGRP antibody can reach its site of
`
`action (“the synaptic cleft” 4), when Tan’s did not. EX1022, 568; EX2265, ¶¶82-
`
`91; EX2268, ¶¶137-138. Whereas the Board credited Tan as “encourag[ing]” use
`
`of a full-length antibody such as C4.19 (Decision, 32), during cross-examination,
`
`Lilly’s expert Dr. Vasserot admitted that key results with Tan’s full-length
`
`antibody C4.19 “cannot be statistically evaluated” and that he would take these
`
`
`4 Tan uses the term “synaptic cleft” to refer to CGRP’s site of action. In the
`
`context of the rat saphenous nerve assay, the term refers to a communicating cell-
`
`cell junction (here, a neuromuscular junction) that allows signals to pass from one
`
`cell to another (here, from a nerve cell to a muscle cell). The signal is carried by a
`
`diffusible neurotransmitter or neuropeptide. EX2047, G:34; EX2048, 98; EX2268,
`
`¶¶44-45.
`
`- 3 -
`
`

`

`
`results “with caution” and “would need to repeat” the assay. EX2191, 118:12-
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`119:1. Dr. Foord agrees, explaining there is no “encouragement,” only doubt,
`
`because of these shortcomings in Tan. EX2265, ¶¶81-91.
`
`The only expert to disagree is the one least qualified to opine on this matter:
`
`Dr. Charles; a treating physician with admittedly no experience in developing
`
`therapeutic antibodies. EX2192, 85:12-86:10. But Dr. Charles’ contentions are
`
`premised entirely on misreading cited references. For example, Dr. Charles states
`
`that Tan “provides guidance to a POSA on how to use full-length antibodies” and
`
`that “a POSA would have been motivated to follow Tan’s recommendations
`
`because they are consistent with how a POSA would have wanted to reduce
`
`incidence of or chronically treat migraine ….” EX1014, ¶¶134, 136. These are
`
`unsupported mischaracterizations: Tan has nothing to do with humans or treatment
`
`of any conditions whatsoever, as Lilly’s experts admit. EX2191, 122:16-123:15;
`
`EX2192, 154:21-156:9. It is a basic research paper, which unambiguously reported
`
`that full-length anti-CGRP antibodies failed to show immunoblockade in a rat
`
`saphenous nerve assay. EX2265, ¶¶15, 56, 84-85; EX2271, ¶¶10-11, 75, 90-91, 95;
`
`EX2268, ¶¶137-138.
`
`But even if Lilly was right that Tan somehow suggests efficacy of anti-
`
`CGRP antibodies to treat migraine, Lilly’s Petition still ignores the vital question
`
`of whether a POSA would have expected those antibodies to be a safe therapeutic
`
`- 4 -
`
`

`

`
`in humans. Lilly has not supported its broad statement that “targeting the ligand—
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`CGRP—as opposed to one of its receptors, had several therapeutic advantages.”
`
`Petition, 28. The evidence of record would not have led a POSA to conclude that
`
`an anti-CGRP antibody would have been less toxic than a small-molecule receptor
`
`antagonist, as Lilly posits. Rather, a POSA would have expected an unacceptable
`
`level of cardiovascular and cerebrovascular toxicity in migraine patients when
`
`targeting the ligand. EX2265, ¶¶112-133; EX2268, ¶¶119-136; EX2157, 533;
`
`EX2086, 15; EX2088, 63; EX2158, S24-S25.
`
`As Teva’s experts explain, a POSA would have expected anti-CGRP
`
`antibodies to eliminate CGRP’s vasoprotective role, potentially leading to
`
`significant cardiovascular and cerebrovascular consequences during, e.g., ischemic
`
`attacks. EX2268, ¶¶12, 103-143; EX2265, ¶¶112-133. Tan itself illustrates this
`
`concern by reporting an increase in blood pressure in experimental rats upon
`
`removing CGRP’s vasoprotective role. EX1022, Figs. 2-3; EX2265, ¶¶123-125;
`
`EX2268, ¶128. And Dr. Charles admitted that patients presenting with migraine
`
`exhibit a higher incidence of cardiovascular and cerebrovascular co-morbidities,
`
`e.g., risk of stroke, heart attack, and hypertension. EX2192, 116:13-21. In such
`
`patients, sustained vasoconstrictive effects of an anti-CGRP antibody, given its
`
`long half-life, could be catastrophic. EX2268, ¶¶132-136; EX2265, ¶¶118-127.
`
`While telling, Lilly’s silence on this point should not be surprising because the
`
`- 5 -
`
`

`

`
`facts firmly undercut Lilly’s assertion that a POSA would have pursued using an
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`anti-CGRP antibody in humans despite the known safety risks.
`
`The Board relied on Dr. Charles to find that “a POSA may well have
`
`understood Olesen’s teachings more broadly than simply the use of receptor
`
`antagonists” for treating migraine. Decision, 26. But as this Response establishes,
`
`Dr. Charles’ testimony lacks credibility and should be given minimal weight. And
`
`in any event, as Dr. Foord explains, a POSA would not have understood Olesen’s
`
`“CGRP antagonists” to extend beyond small-molecule receptor antagonists.
`
`EX2265, ¶¶60-63. Moreover, contrary to Lilly’s suggestion, Olesen would not
`
`have assuaged concerns over safety. EX2268, ¶50; EX2265, ¶¶64-76, 131. Olesen
`
`admits they did not assess cardiovascular safety. EX1025, 1109. And, as a small
`
`molecule, BIBN4096BS’s half-life is hours, whereas an anti-CGRP antibody’s is
`
`about a month, which changes the safety profiles of these drugs. EX2265, ¶65;
`
`EX2271, ¶¶69-84; EX2268, ¶¶139, 145. There are also material differences
`
`between receptor and ligand antagonism that would have further diminished the
`
`value a POSA would have drawn from Olesen vis-à-vis cardiovascular and
`
`cerebrovascular safety of anti-CGRP antibodies. EX2265, ¶¶66-81; EX2271, ¶¶85-
`
`88, 109; EX2268, ¶¶144-147. Lilly’s reliance on small-molecule receptor
`
`antagonism as a proxy for ligand antagonism with an antibody is fraught with
`
`holes, as explained by Drs. Foord and Tomlinson. EX2265, ¶¶64-81; EX2271,
`
`- 6 -
`
`

`

`
`¶¶81-84.
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`Finally, multiple objective indicia provide real-world, corroborating
`
`evidence of nonobviousness. That Teva’s invention is not obvious couldn’t be
`
`summed up any better than Lilly’s own expert’s statement, made outside the
`
`context of this case, in 2018, well over a decade after Teva’s invention: “The
`
`notion that we would be using antibodies to treat migraine is really quite a radical
`
`concept.” EX2053, 235. If Teva’s invention were obvious in 2005, as Lilly
`
`contends, Dr. Charles would not be calling it “radical” in 2018.
`
`II. Background
`A. Calcitonin Gene-Related Peptide (“CGRP”)
`CGRP is a 37-amino acid neurotransmitter belonging to a family of peptides
`
`that include calcitonin, adrenomedullin, and amylin. EX2265, ¶¶25-31; EX1022,
`
`569-570; EX2003, 904-905. CGRP interacts with receptors distributed throughout
`
`the body. EX1031, 317, 318, 320; EX1096, 539; EX2265, ¶25; EX2003, 904.
`
`CGRP is expressed as two isoforms, α- and β-CGRP, which have similar activities
`
`but different distribution. EX2265, ¶¶25, 28; EX2003, 905; EX2008, 153.
`
`CGRP is “the most potent vasodilating substance known.” EX2003, 905-
`
`906; EX2265, ¶26. As a potent vasodilator, CGRP plays a “pivotal role” in the
`
`“physiology and pathophysiology of cardiovascular regulation.” EX2003, 923;
`
`
`5 Emphasis added throughout unless otherwise noted.
`
`- 7 -
`
`

`

`
`EX2265, ¶26; EX2003, 923. By 2005, CGRP was known to be involved in long-
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`term regulation of resting blood pressure, protection against development of
`
`hypertension, and preventing escalation of mild ischemic events into full-blown
`
`infarctions, like stroke and heart attacks. EX2271, ¶¶73-74, 78-79; EX2268,
`
`¶¶113-118; EX2065, ¶¶16, 118-119; EX2058, Abstract; EX2150, 503; EX2164,
`
`174; EX2151; EX2079, Abstract; EX2003, 919; EX2009, 614-615; EX1035, 293;
`
`EX2019, 139.
`
`B.
`
`By 2005, there was no consensus as to the pathophysiology of
`migraine
`
`Migraine, a highly prevalent, debilitating disorder, affects about 11% of the
`
`population of Western countries. EX2286, 2. According to Dr. Ferrari, the causes
`
`and mechanisms of action of migraine continue to remain the subject of
`
`considerable debate even today. EX2268, ¶¶25-38, 46-48. As of November 2005,
`
`there were three principle theories with regard to the pathogenesis of migraine—
`
`the vascular, central neuronal, and neurogenic inflammation theories. EX2289,
`
`373-74; EX2295, 280-281; EX2308, Abstract, 743; EX2291, 1-2; EX2292, 387;
`
`EX2293, S239; EX2294, Abstract; EX2295, 283; EX1089, 258; EX1040, 179;
`
`EX1097, 518. But as Dr. Ferrari explains, in 2005, no consensus existed in the
`
`field as to these theories. EX2268, ¶38. EX2289, 374.
`
`C. The ’045 patent and its commercial embodiments
`To the extent that the field was investigating disrupting CGRP pathway for
`
`- 8 -
`
`

`

`
`migraine treatment before 2005, it focused on investigating antagonizing CGRP
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`receptors, not antagonizing the CGRP ligand itself. EX2265, ¶¶105-108; EX2268,
`
`¶¶48, 50-53; EX1243; 588; EX2236, Abstract; EX2239, Abstract; EX2240,
`
`Abstract; EX2300; EX1025, 1105; EX2004, E331; EX1024, 420; EX2009, 617;
`
`EX2012, 769; EX2014; Abstract; EX2071, Abstract. For example, in describing
`
`the ideal “CGRP antagonist” for treating migraine, Wimalawansa stated: the
`
`compound “should be a peptide mimetic of simple structure specific to
`
`cardiovascular CGRP receptors,” and “the antagonist must be extremely specific to
`
`the CGRP receptors located in cerebral arteries to avoid potential deleterious side
`
`effects caused by blocking other vascular and nonvascular CGRP receptors.”
`
`EX1096, 568. Indeed, BIBN4096BS—the only clinically-tested “CGRP
`
`antagonist” for treating migraine by 2005—was a small-molecule receptor
`
`antagonist. EX1025, Abstract. Further, the art specifically cited adverse events as
`
`the reason for focusing on tissue-specific receptors (as opposed to ligand
`
`antagonists): “[s]ome of the potential side effects of CGRP antagonists could
`
`possibly be minimized by being designed specifically to act on only one subtype of
`
`receptor.” EX1096, 567.
`
`Against this backdrop, and in contrast to conventional thinking at the time,
`
`the ’045 patent inventors developed methods for treating migraine using
`
`humanized monoclonal anti-CGRP antagonist antibodies, pivoting away from
`
`- 9 -
`
`

`

`
`CGRP receptor antagonism and small-molecule- (or even peptide- or aptamer-)
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`based therapeutics. These inventors demonstrated, for the first time, the therapeutic
`
`value of a humanized monoclonal anti-CGRP antagonist antibody by showing
`
`activity in a therapeutically-relevant headache model. EX1001, 68:59-69:38.
`
`Treating migraine with Teva’s Ajovy® and Lilly’s Emgality® falls within the
`
`challenged claims, and each has satisfied a long-felt need, enjoyed industry-wide
`
`acclaim, and captured a significant portion of the migraine market within just a few
`
`months of approval. See §V.
`
`III. Claim construction
` “reducing incidence of or treating”
`A.
`The Board found that a method of “reducing incidence of or treating”
`
`“reflects administering the anti-CGRP antagonist antibody based on a reasonable
`
`expectation that such administration may likely cause such a reduction in incidence
`
`in that particular individual.” Decision, 10-11 (citing EX1001, 17:60–65). Lilly’s
`
`expert agrees. EX1014, ¶103. Teva does not dispute this construction. EX2268,
`
`¶¶19-20; EX2265, ¶¶52-55. Therefore, “a method for reducing incidence of or
`
`treating” requires a reasonable expectation that the method will be therapeutically
`
`effective.
`
`"effective amount"
`
`B.
`Teva agrees with the Board that the term “effective amount” means “an
`
`- 10 -
`
`

`

`
`amount sufficient to effect beneficial or desired results” because the patent
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`provides a clear definition of that term. Decision, 11-12; EX1001, 18:38-19:3;
`
`EX2268, ¶¶21; EX2265, ¶54; Martek Biosciences Corp. v. Nutrinova, Inc., 579
`
`F.3d 1363, 1380 (Fed. Cir. 2009).
`
`ARGUMENT
`
`It is well-settled that “[a] patent composed of several elements is not proved
`
`obvious merely by demonstrating that each of its elements was, independently,
`
`known in the prior art.” KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007).
`
`Rather, obviousness requires proof that a POSA would have had (1) reason to
`
`combine the references to solve a real-world problem by making the claimed
`
`invention; and (2) a reasonable expectation of success in doing so. Institut Pasteur
`
`v. Focarino, 738 F.3d 1337, 1346 (Fed. Cir. 2013).
`
`
`
`Lilly has failed to establish either prong of obviousness.
`
`IV.
`
` The challenged claims are not prima facie obvious
`As this Response shows, the prior art completely discredits Lilly’s alleged
`
`expectations and motivations to use an anti-CGRP antibody to treat migraine. The
`
`Board should confirm all challenged claims.
`
`A. Lilly mischaracterizes what the prior art discloses
`Lilly vastly overstates what a POSA would have understood about treating
`
`migraine with anti-CGRP antibodies from Olesen and Tan. As explained below,
`
`- 11 -
`
`

`

`
`neither reference, nor the art as a whole, provided a POSA a reason to treat
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`migraine with an anti-CGRP antagonist antibody. EX2265, ¶¶14-23, 46-111;
`
`EX2271, ¶¶85-106; EX2268, ¶¶96-102, 137-141, 148-153.
`
`(a) Nothing in Olesen suggests Lilly’s broad reading of
`“CGRP antagonists”
`
`Lilly’s argument that “a POSA reading Olesen would have extended its
`
`teachings to other CGRP antagonists” (Petition, 26) is meritless, finding no support
`
`in Olesen or the art as a whole. As Dr. Foord confirms, Olesen does not suggest
`
`that “CGRP antagonists” extends to any antagonist beyond small-molecule
`
`receptor antagonists. EX2265, ¶¶60-63. Olesen investigated only BIBN4096BS, a
`
`small-molecule receptor antagonist. EX1026, Abstract. While Olesen mentions
`
`triptans, it distinguishes them from “CGRP antagonists” as “selective agonists of
`
`serotonin.” EX1025, 1105. Given this limited disclosure, as Dr. Foord explains,
`
`Olesen’s “CGRP antagonists” should not be read broadly to encompass CGRP
`
`ligand antagonists. EX2265, ¶63.
`
`Other record evidence also fails to support Lilly’s broad reading of “CGRP
`
`antagonists.” The Board relied on Olesen 2004 (EX1029) and Arulmozhi
`
`(EX1040) in its preliminary determination that a POSA would have understood
`
`“Olesen’s teachings more broadly than simply the use of receptor antagonists.”
`
`Decision, 26. But, as Dr. Ferrari explains, Olesen 2004 also relates only to
`
`- 12 -
`
`

`

`
`BIBN4096BS. EX1029, Abstract; EX2268, ¶¶98. Similarly, Arulmozhi does not
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`extend to CGRP ligand antagonism. Arulmozhi is a review paper that cites to
`
`Edvinsson (EX2009) for “inhibition of CGRP … could be a viable therapeutic
`
`target for the pharmacological treatment of migraine.” EX2268, ¶98; EX1040, 182.
`
`But in discussing CGRP antagonism, Edvinsson only refers to two classes of
`
`compounds: CGRP receptor antagonists and triptans (5HT receptor inhibitors that
`
`inhibit CGRP release). EX2268, ¶98; EX2009, 617-618.
`
`Given Olesen’s silence as to CGRP antagonists besides small-molecule
`
`receptor antagonists, Lilly has to rely on Dr. Charles’ unsupported opinions to
`
`argue its broader reading. But Dr. Charles repeatedly demonstrated his inability to
`
`provide consistent, credible testimony. For example, when asked how a POSA
`
`would have understood the following sentence: “Several compounds have been
`
`found to selectively inhibit the CGRP receptor, such as small molecular non-
`
`peptide compounds, peptides and antibodies,” (EX1026, 7:5-12), Dr. Charles
`
`testified that it “is referring to anti-CGRP antibodies that bind the ligand.”
`
`EX2277, 55:11-17. Notably, previously Dr. Charles admitted that “there’s nothing
`
`specifically discussing binding the ligand” in this section of EX1026. EX2192,
`
`180:5-7.
`
`Next, Dr. Charles claims that Wimalawansa’s statement that “[c]learly more
`
`data from carefully designed studies are necessary before any definitive
`
`- 13 -
`
`

`

`
`conclusions can be reached and before CGRP antagonists, humanized anti-CGRP
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`monoclonal antibodies, or both can be evaluated as therapeutic agents in humans”
`
`somehow “does provide data.” EX2277, 59:1-2. Dr. Charles also claims that
`
`“Wimalawansa stated that humanized anti-CGRP antibodies ‘should’ be developed
`
`and used.” EX1014, ¶65. But when cross-examined, Dr. Charles admitted that he
`
`was “paraphrasing,” and could not provide a coherent basis for such a paraphrase,
`
`other than to evasively state that it was based on his “reading of the paper.”
`
`EX2192, 132:2-4, 133:11-12.
`
`Dr. Charles’ disingenuous testimony is further exemplified by his distorted
`
`reading of Tan. When asked: in relation to what does Tan’s mAb have “inherent
`
`advantages,” Dr. Charles replied “in relation to small molecules.” EX2277, 90:14-
`
`21. This is despite (i) Dr. Charles’ admission that this section in Tan contains “no
`
`specific mention of small molecules,” and (ii) that the sentence regarding “inherent
`
`advantages” directly follows a discussion of polyclonal antibodies. EX1022, 572;
`
`EX2265, ¶¶92-94. These examples of Dr. Charles’ untenable misrepresentations of
`
`the art demonstrate that his opinion as to what a POSA would have understood
`
`from Olesen deserves no weight.
`
`(b) Tan provided no “guidance” about reducing
`incidence of or treating migraine with anti-CGRP
`antibodies in humans
`
`Tan is a basic science pharmacology paper that draws no therapeutic or
`
`- 14 -
`
`

`

`
`clinical conclusions. EX2265, ¶¶82-84; EX2268, ¶¶137-141; EX2271, ¶¶90-91.
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`The authors intended simply “to investigate immunoblockade as an alternative
`
`strategy for probing the role of CGRP as a vasodilator in vivo.” EX1022, Abstract.
`
`Thus, Tan was trying to elucidate CGRP’s role in vasodilation in an experimental
`
`animal; not whether an anti-CGRP antibody could be safely developed for treating
`
`migraine. EX2265, ¶83; EX2268, ¶138. Dr. Charles grossly mischaracterized Tan
`
`by arguing it “provides guidance… on how to use full-length antibodies” for
`
`treating migraine. EX1014, ¶134; EX2265, ¶¶86-91. But nothing in Tan has
`
`anything to do with humans, treatments, migraine, or dosing. Indeed, Lilly’s expert
`
`Dr. Vasserot admitted that Tan is silent as to “specific human diseases” or
`
`“efficacy of administering an anti-CGRP antagonist antibody to humans to treat a
`
`disease.” EX2191, 122:16-123:13. And Dr. Charles could not show otherwise.
`
`EX2192, 182:21-183:12. The evidence simply does not support Lilly’s assertion
`
`that Tan provides “guidance” on how to use anti-CGRP antibodies to treat
`
`migraine.
`
`(c) Lilly’s experts internally disagree on what the cited
`art discloses
`
`Given the liberties Dr. Charles takes in interpreting the prior art, it is
`
`unsurprising that Lilly’s other expert, Dr. Vasserot, disagrees with his opinions.
`
`For example, Dr. Charles asserts that “a 16% block in increased blood flow was
`
`- 15 -
`
`

`

`
`observed” in Tan’s saphenous nerve assay with the full-length C4.19 antibody.
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`EX1014, ¶144. But he bases his opinion on Tan’s ambiguous statement that further
`
`nerve stimulation with longer incubation time of a full-length anti-CGRP
`
`antagonist antibody “produced an AUC which was slightly smaller compared with
`
`baseline stimulation, but not by more than 16% (n=2).” EX1022, 569. As Teva’s
`
`pharmacology expert, Dr. Foord, explains, a POSA would have understood that
`
`this result is not statistically significant given the small sample size (n=2), and
`
`would have considered the result to be negative (i.e., not effective). EX2265, ¶85.
`
`Tan acknowledged as much. EX1022, 571 (“the results could not be evaluated
`
`statistically due to the inadequate sample sizes (n=2).”). Dr. Tomlinson agrees.
`
`EX2271, ¶91. More notably, even Lilly’s other expert, Dr. Vasserot, agrees.
`
`EX2191, 118:12-119:1. According to Dr. Vasserot, a POSA would not have drawn
`
`any meaningful conclusion from Tan’s “not more than 16% (n=2)” statement;
`
`instead, Dr. Vasserot testified that these results are “something that [he] would take
`
`with caution and would need to repeat.” Id., 118:21-119:1.
`
`Dr. Vasserot’s conclusions regarding Tan in his deposition do not align with
`
`the motivations to humanize he gives in his declaration. EX1015, ¶¶87-92, 94,
`
`101-102. But during a deposition, it came to light that Dr. Vasserot’s declaration
`
`conclusions on the state of the art for CGRP antibodies were based on erroneous
`
`facts Lilly’s attorneys asked him to assume. For example, Lilly’s attorneys asked
`
`- 16 -
`
`

`

`
`Dr. Vasserot to presuppose key facts like “several research groups had already
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`recognized the therapeutic potential of anti-CGRP antagonist antibodies for
`
`migraine;” that “several research groups proposed using monoclonal anti-CGRP
`
`antagonist antibodies to treat migraine and other diseases;” and that the “prior art
`
`had expressly proposed humanized anti-CGRP antagonist antibodies for treating
`
`migraine and other human diseases.” EX2278, 17:2-10, 18:5-13, 19:17-22. When
`
`not being told what to say by his attorneys6, Dr. Vasserot correctly acknowledged
`
`that Tan offered little insight on the utility of an anti-CGRP antibody for treatment
`
`in humans. Id., 122:16-123:15. This is a fundamental gap in Lilly’s argument,
`
`which is premised on the notion that a POSA would want to humanize a murine
`
`anti-CGRP antibody to use it to treat migraine. Petition, 33.
`
`Dr. Charles’ and Lilly’s overly-broad interpretation of the cited references
`
`are misleading. Dr. Charles’ declaration with respect to Olesen and Tan therefore
`
`deserves little to no weight.
`
`B.
`
`There was uncertainty in the field regarding CGRP being a
`biomarker for migraine—a fact that Lilly ignores
`
`Lilly relies heavily on the supposedly “understood” fact that “levels of
`
`CGRP … are significantly elevated in migraine patients compared to those without
`
`6 Dr. Vasserot’s unsupported testimony should get no weight. 37 C.F.R.
`
`§ 42.65(a).
`
`- 17 -
`
`

`

`
`migraine,” citing to articles by Goadsby from the early 1990’s. Petition, 11;
`
`Case IPR2018-01710
`Patent No. 8,586,045 B2
`
`EX2014, ¶28; EX1043, 185; EX1044, 48, 52-52. But Lilly provides an incomplete
`
`picture of the art as of 2005.
`
`Lilly did not consider a 2005 article by Tvedskov and Olesen that challenged
`
`the validity of Goadsby’s findings. EX2277, 23:11-17. But this article is highly
`
`pertinent to these proceedings because Tvedskov showed that “previous findings of
`
`increased CGRP level in … venous blood could not be confirmed.” EX2309,
`
`Abstract, 565, Table 3, 566-67, Figures 2, 3. Tvedskov recognized that Goadsby’s
`
`finding regarding the increased levels of CGRP during a migraine attack “has not
`
`been convincingly reproduced.” Id., Abstract. And Tvedskov’s own data showed
`
`“[n]o difference between CGRP in external jugular blood and peripheral blood was
`
`found during attack.” Id., 563. What’s more, “[m]igraine attack duration was found
`
`not to have relation to CGRP level,” and “no relation was found between jugular
`
`venous CGRP level and headache characteristics, headache intensity, [and]
`
`migraine frequency … .” Id., 564. Tvedskov ultimately concluded that “CGRP
`
`cannot be used as a biomarker of migraine” and “increased circulating CGRP level
`
`is not a validating criterion for human or animal experimental migraine models.”
`
`Id., 567. As Dr. Ferrari explains, “[i]n view of Tvedskov’s results … a POSA
`
`would have had doubts about CGRP’s status as a biomarker of migraine by
`
`November 14, 2005.” EX2268, ¶¶57, 60.
`
`- 18 -
`
`

`

`These fin