Filed on behalf of: Eli Lilly and Company
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`Filed: October 4, 2019
`
`______________________
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`______________________
`
`ELI LILLY AND COMPANY,
`Petitioner,
`v.
`TEVA PHARMACEUTICALS INTERNATIONAL GMBH,
`Patent Owner.
`______________________
`
`Case No. IPR2018-01712
`Patent No. 9,884,908
`______________________
`
`PETITIONER’S REPLY
`
`
`
`
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`
`B.
`
`C.
`
`TABLE OF CONTENTS
`Introduction ................................................................................................. 1
`I.
`Claim Construction ...................................................................................... 2
`II.
`III. Teva Fails to Rebut the Petition’s Strong Case of Obviousness ................... 3
`A.
`Teva’s Purported Concerns about the Blood-Brain Barrier Are
`Incorrect............................................................................................. 7
`Humanized Anti-CGRP Antibodies Were Expected To Access
`the Site of Action ..............................................................................10
`Teva’s Unfounded Safety Concerns Do Not Support
`Patentability ......................................................................................12
`1.
`Clinical Studies Demonstrated that the CGRP Pathway
`Could Be Safely Antagonized To Treat Migraine ...................12
`Long-Acting Ligand Antagonists Had Desirable Benefits
`and Did Not Raise Safety Concerns ........................................14
`The Prior Art Would Not Have Dissuaded a POSA from
`Pursuing a Humanized Anti-CGRP Antagonist Antibody
`for Treating Migraine ..............................................................16
`Teva’s Hypothetical “Spare Receptor Theory” and Ligand
`Cross-Binding Concerns Are Incorrect .............................................18
`Teva’s Binding-Affinity Arguments Are Incorrect ............................19
`E.
`Novartis Is Inapposite .......................................................................20
`F.
`IV. Teva’s Alleged Secondary Considerations Do Not Support
`Nonobviousness ..........................................................................................22
`A.
`Teva’s Evidence Is Not Commensurate with the Scope of the
`Challenged Claims ............................................................................22
`Teva’s Secondary Considerations Lack Nexus to the Claims ............24
`Teva’s Reliance on Industry Acclaim Is Misplaced ..........................24
`
`B.
`C.
`
`2.
`
`3.
`
`D.
`
`i
`
`

`

`D.
`
`IPR2018-01712
`Patent No. 9,884,908
`Teva Failed to Establish Unexpected Results or Industry
`Skepticism ........................................................................................25
`Teva’s Purported Evidence of Commercial Success, Licensing,
`and Long-Felt Need Do Not Support Patentability ............................27
`Conclusion ..................................................................................................27
`
`E.
`
`V.
`
`
`
`ii
`
`

`

`TABLE OF AUTHORITIES
`
`IPR2018-01712
`Patent No. 9,884,908
`
` Page(s)
`
`Cases
`Alcon Research, Ltd. v. Apotex Inc.,
`687 F.3d 1362 (Fed. Cir. 2012) .......................................................................... 4
`
`Allergan, Inc. v. Apotex Inc.,
`754 F.3d 952 (Fed. Cir. 2014) ...................................................................... 3, 23
`Bayer Healthcare Pharm., Inc. v. Watson Pharm., Inc.,
`713 F.3d 1369 (Fed. Cir. 2013) .................................................................. 17, 24
`Cole Kepro Int’l, LLC v. VSR Indus., Inc.,
`695 F. App’x 566 (Fed. Cir. 2017) ................................................................... 27
`Equistar Chemicals, LP v. ExxonMobil Chemical Patents Inc.,
`IPR2017-01534, Paper 64 (PTAB Dec. 7, 2018) ................................................ 4
`Genetics Inst., LLC v. Novartis Vaccines & Diagnostics, Inc.,
`655 F.3d 1291 (Fed. Cir. 2011) ........................................................................ 22
`In re Kao,
`639 F.3d 1057 (Fed. Cir. 2011) ........................................................................ 23
`In re Kubin,
`561 F.3d 1351 (Fed. Cir. 2009) .......................................................................... 4
`Novartis Pharm. Corp. v. West-Ward Pharm. Int’l Ltd.,
`923 F.3d 1051 (Fed. Cir. 2019) ........................................................................ 20
`Paint Point Med. Sys., Inc. v. Blephex, LLC,
`IPR2016-01670, Paper 44 (PTAB Feb. 28, 2018) ............................................ 26
`S. Ala. Med. Sci. Found. v. Gnosis S.P.A.,
`808 F.3d 823 (Fed. Cir. 2015) .......................................................................... 24
`
`
`
`iii
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`
`BBB
`
`CGRP
`
`FDA
`
`IPR
`
`GLOSSARY
`
`Blood-brain barrier
`
`Calcitonin gene-related peptide
`
`U.S. Food and Drug Administration
`
`Inter partes review
`
`Italicized text
`
`Emphasis added unless otherwise indicated
`
`Lilly or Petitioner Eli Lilly and Company
`
`MAb
`
`MAP
`
`POSA
`
`Teva or Patent
`Owner
`
`Monoclonal antibody
`
`Mean arterial pressure
`
`Person of ordinary skill in the art
`
`Teva Pharmaceuticals International GmbH
`
`’908 patent
`
`U.S. Patent No. 9,884,908
`
`
`
`iv
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`
`I.
`
`Introduction
`The challenged claims are broadly directed to a method of treating headache
`
`using any humanized IgG anti-CGRP antagonist antibody having properties already
`
`known in the art. To contest motivation and expectation of success, Teva argues that
`
`a POSA would not have expected to effectively and safely treat headache using such
`
`antibodies—relying on purported safety concerns, the blood-brain barrier (“BBB”),
`
`the “spare receptor” theory, and cross-binding of CGRP to ancillary receptors.
`
`But the claims do not require clinical efficacy and do not mention safety.
`
`Moreover, Teva’s patent does not acknowledge or even suggest that either the single
`
`humanized antibody or its derivatives exemplified in the specification provide
`
`solutions for any of these alleged concerns—persuasive evidence the concerns were
`
`not real. The disconnect between Teva’s critique of the prior art versus the sparse
`
`disclosure of the ’908 patent highlights a fatal deficiency in Teva’s obviousness
`
`position: improperly imposing a heightened standard for motivation and expectation
`
`of success. Indeed, the law is clear that obviousness requires only a reasonable
`
`expectation of success, not absolute predictability.
`
`The evidence confirms that Teva’s concerns are illusory. By November 2005,
`
`systemically administered CGRP antagonists had repeatedly been shown to be safe
`
`and effective, with a peripheral site of action. Time-and-again, the art encouraged
`
`making longer-acting CGRP antagonists, including humanized antibodies, for
`
`1
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`treating migraine. Antagonizing the CGRP pathway had also proven clinically
`
`effective for treating migraine, confirming a reasonable expectation of success. The
`
`challenged claims are unpatentable.
`
`II. Claim Construction
`The Board correctly construed “treating” as an “approach” for achieving a
`
`desired purpose without requiring achievement of any clinical result. Paper 11, 9-
`
`11; Pet., 21-22. Teva disagrees, but fails to demonstrate any error with the Board’s
`
`construction. Indeed, Teva and its experts disregard the patent’s express definition
`
`of “treating” as an “approach for obtaining a beneficial or desired clinical result,”
`
`which is a mere statement of intent. POR, 11-12 (further admitting that “treating” is
`
`analogous to a “physician’s goal”); Ex. 2270 ¶¶19-20; Ex. 2267 ¶53; Ex. 1343,
`
`19:25-20:8; 23:8-19. The term “treating” therefore does not require actually
`
`achieving a clinical response.
`
`The ’908 patent defines an “effective amount” to include an amount sufficient
`
`to affect “biochemical” and “histological” symptoms, which Teva again ignores.
`
`Ex. 1001, 19:3-22; POR, 12, Pet., 22-24. The Board invited testimony on whether
`
`such symptoms include stimulated cAMP formulation, but Teva provided none.
`
`Paper 11, 11-12. As Dr. Charles explains, cAMP stimulation was recognized as a
`
`direct biochemical response to elevated CGRP levels, which are characteristic of
`
`migraine. Ex. 1342 ¶¶5-10; Ex. 1018 ¶¶27-39; Ex. 1001, 26:20-28 (listing blocking
`
`2
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`cAMP activation as a property of an anti-CGRP antagonist antibody); Ex. 1303,
`
`61:15-62:12; Ex. 1343, 28:18-29:18. Thus, an “effective amount” includes amounts
`
`sufficient to reduce biochemical or histological symptoms (such as cAMP
`
`stimulation) without requiring any clinical result. See Ex. 1343, 33:17-34:6.
`
`III. Teva Fails to Rebut the Petition’s Strong Case of Obviousness
`Teva’s obviousness arguments rest on purported concerns about whether an
`
`anti-CGRP antagonist antibody would have been expected to effectively and safely
`
`treat migraine. E.g., POR, 2, 11, 25. The claims, however, do not require clinical
`
`efficacy and do not mention safety. Supra §II; Paper 11, 9-11. Consequently, Lilly
`
`does not have an “exacting burden of showing a reasonable expectation of success”
`
`that the claimed methods would be clinically effective and safe. Allergan, Inc. v.
`
`Apotex Inc., 754 F.3d 952, 962-63 (Fed. Cir. 2014).
`
`Teva’s own specification undermines its current assessment of the prior art: it
`
`neither acknowledges nor provides solutions for any of Teva’s alleged concerns.
`
`Teva’s patent contains no clinical data and no safety data whatsoever. Ex. 1303,
`
`55:7-57:19. As Teva’s expert Dr. Foord testified, the limited animal studies in Teva’s
`
`patent “will never satisfy concerns about safety and efficacy.” Ex. 1300, 173:20-
`
`175:6; see also Ex. 1303, 31:8-14, 54:12-23; infra §III.C.
`
`Similarly, Teva now argues that a POSA would have believed that an anti-
`
`CGRP antibody must cross the BBB to be clinically effective (POR, 39-44), but
`
`3
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`Teva’s patent does not address it. Ex. 1342 ¶¶54-56; Ex. 1345, 61:5-65:2; infra
`
`§III.A-B. Teva’s patent similarly fails to address the spare receptor theory or safety
`
`concerns about receptor cross-binding. Infra §III.D.
`
`The disconnect between Teva’s assessment of the prior art versus the paucity
`
`of its own disclosure illuminates the inappropriately heightened standard for
`
`motivation and expectation of success that Teva relies upon. Equistar Chemicals,
`
`LP v. ExxonMobil Chemical Patents Inc., IPR2017-01534, Paper 64, 40 (PTAB Dec.
`
`7, 2018) (holding claims obvious where “Patent Owner’s arguments demand more
`
`of the prior art than is provided by the sparse disclosure of the ’709 patent”); see
`
`also Alcon Research, Ltd. v. Apotex Inc., 687 F.3d 1362, 1369 (Fed. Cir. 2012)
`
`(affirming obviousness where purported safety concerns were not addressed in the
`
`invalidated patent); In re Kubin, 561 F.3d 1351, 1360 (Fed. Cir. 2009) (all that is
`
`required is a reasonable expectation of success).
`
`Regardless of the standard applied, the art provides express motivation and a
`
`reasonable expectation of success for treating headache with a humanized anti-
`
`CGRP antagonist antibody.
`
`Olesen established in a Phase II clinical trial that antagonizing the CGRP
`
`pathway effectively treats migraine, validating “CGRP antagonism as a new
`
`therapeutic principle” for treating migraine. Ex. 1025, 1104, 1108-1109; Ex. 1029,
`
`4
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`119(S26). Prior art researchers “expect[ed] that CGRP antagonists will be effective
`
`anti-migraine drugs.” Ex. 1024, 422. Teva’s experts agree:
`
`• Dr. Ferrari wrote in 2007 that Olesen “firmly establish[ed] blockade of
`
`the CGRP pathway as a novel and important new emerging treatment
`
`principle” for migraine. Ex. 1332, 443.
`
`• Based on Olesen, Dr. Ferrari in 2005 praised “CGRP antagonists” as
`
`“promising, new antimigraine drugs.” Ex. 1290, 657.
`
`• Based on Olesen, Dr. Rapoport in 2005 called for “[p]reventive drugs”
`
`for migraine headache that “antagonis[ed] the effect of CGRP.”
`
`Ex. 1297, S119.
`
`Despite these contemporaneous admissions, Teva attempts to limit Olesen’s
`
`teachings to small-molecule receptor antagonists. But Tan expressly discloses that
`
`targeting the CGRP ligand with antibodies and targeting CGRP receptors (e.g., with
`
`CGRP8-37 or Olesen’s clinically effective BIBN4096BS) were “alternative”
`
`approaches for antagonizing the CGRP pathway. Ex. 1022, 566, 571; see also
`
`Ex. 1040, 182 (“inhibition of CGRP or antagonism of CGRP receptors could be a
`
`viable therapeutic target for the pharmacological treatment of migraine”); Ex. 1033,
`
`95. Multiple ligand antagonists, including humanized anti-CGRP antagonist
`
`antibodies, had been disclosed for treating migraine. Pet., 28-30; Ex. 1096, 567,
`
`570; Ex. 1082, Abstract; Ex. 1240, 923.
`
`5
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`The advantageous properties of anti-CGRP antagonist antibodies were also
`
`known: Tan’s antibody specifically bound and antagonized CGRP in vitro and in
`
`vivo, including in the same animal model Teva used to support its original claims to
`
`treating migraine. Exs. 1021, 1022; Pet., 18. Although Teva argues Tan is a basic
`
`research paper having “nothing to do with humans or treatment” (POR, 4), Dr. Tan
`
`contemporaneously wrote that there is “no reason” why humanized anti-CGRP
`
`antagonist antibodies should not be developed and used for treating migraine.
`
`Ex. 1287, 247; see also Ex. 1096, 567, 570 (contemporaneously disclosing
`
`humanized anti-CGRP antagonist antibodies for treating migraine).
`
`Remarkably, Teva argues it was unclear whether CGRP levels increase during
`
`migraine. POR, 19-21. This is irrelevant, as Olesen had validated CGRP
`
`antagonism as a therapeutic target. Ex. 1342 ¶¶25-30. It is also incorrect, as
`
`numerous prior-art publications reported increased CGRP levels during migraine.
`
`Exs. 1043, 1045, 2066, 2067. Indeed, the background section of Teva’s patent
`
`acknowledges that “levels of CGRP … are elevated in patients during migraine
`
`headache.” Ex. 1001, 2:24-27 (citing Ex. 1043); see also Ex. 1332, 443 (Dr. Ferrari
`
`distinguishing Teva’s cited Tvedskov reference and publishing that CGRP levels
`
`increase in the headache phase of severe migraine).
`
`6
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`A. Teva’s Purported Concerns about the Blood-Brain Barrier Are
`Incorrect
`Teva incorrectly asserts that a POSA would have understood that anti-
`
`migraine therapeutics must cross the BBB, and thus larger molecules like anti-CGRP
`
`antagonist antibodies would not have been expected to treat migraine. POR, 39-43;
`
`Ex. 1342 ¶¶54-56; Ex. 1345, 61:5-65:2.
`
`Extensive prior-art evidence established that peripheral antagonism of the
`
`CGRP pathway treats migraine. Using the same animal model used in Teva’s patent,
`
`Petersen 2004 showed that Olesen’s BIBN4096BS selectively blocked CGRP in the
`
`periphery (i.e., in vessels without a BBB) and not in central pial arteries. Compare
`
`Ex. 1090, Abstract, with Ex. 1001, 69:50-70:56; see also Ex. 1345, 52:6-54:7;
`
`Ex. 1342 ¶¶36-42.
`
`Teva contests whether pial arteries possessed a BBB. POR, 41-42. But the
`
`prior art recognized that “pial arteries … have a [BBB].” Ex. 2068, 39. Further, the
`
`authors of Petersen 2004 and numerous others expressly recognized the peripheral
`
`site of action: “[t]he present study strongly suggests that the clinically effective
`
`migraine drug BIBN4096BS (Olesen et al., 2004) does not cross the BBB.”
`
`Ex. 1090, 703; see also Ex. 2215, 75 (“BIBN4096BS does not appear to pass the
`
`[BBB] freely”); Ex. 1031, 326 (Dr. Ferrari’s advisor, Dr. Saxena, concluded that
`
`7
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`“BIBN4096BS does not seem to penetrate the [BBB].”). Even Teva’s expert
`
`admitted it was “unlikely” BIBN4096BS crossed the BBB. Ex. 1343, 76:12-77:8.
`
`Teva also criticizes Petersen 2004 because it is an animal study. POR, 42.
`
`But Teva disregards another prior-art study, Petersen 2005, that confirmed in humans
`
`that antagonizing CGRP in the periphery prevented CGRP-induced headache—with
`
`no effect on central vasodilatation. Ex. 1333; Ex. 1342 ¶¶43-47. Petersen 2005
`
`monitored both the middle cerebral artery and cerebral blood flow using the “best
`
`available” technology, concluding that BIBN4096BS “prevents or treats headache”
`
`in an extracerebral manner, i.e., in areas “devoid of a blood-brain barrier.” Ex. 1333,
`
`211; Ex. 1345, 78:2-22.
`
`Ignoring the human data (and despite its criticism of animal studies), Teva
`
`relies principally on three animal studies, Storer, Fischer, and Levy, to argue that
`
`crossing the BBB was necessary. POR, 40-41. But these studies did not change the
`
`prior-art understanding that peripheral antagonism of the CGRP pathway treats
`
`migraine. Ex. 1342 ¶¶48-53. Indeed, Storer, Fischer, and Levy used invasive,
`
`“open-window” animal models that expose and manipulate the brain’s dura mater—
`
`criticized by Teva’s expert as non-physiologic. Ex. 2298, 699; Ex. 2310, 5878;
`
`Ex. 2307, 1172; Ex. 1345, 48:8-49:11; Ex. 1342 ¶¶48-49. Levy further suffered
`
`from other experimental limitations, including failing to evaluate any CGRP
`
`8
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`antagonist and failing to measure neuronal activities for an adequate duration.
`
`Ex. 2298, 699; Ex. 1047, 56, 59-60; Ex. 1345, 60:1-61:1; Ex. 1342 ¶¶51-52, 31-35.
`
`Regardless, even these studies support a peripheral site of action. Ex. 1342
`
`¶¶50, 53. For example, Storer and Fischer collectively observed that central and
`
`peripheral BIBN4096BS administration suppressed activity of the trigeminocervical
`
`complex, further indicating that peripheral antagonism was effective. Ex. 2307,
`
`1175-1176; Ex. 2310, Abstract; Ex. 1342 ¶50.
`
`Other prior-art migraine treatments were similarly understood to act
`
`peripherally. Ex. 1342 ¶¶18-21, 57-62; see also id., ¶¶63-66. Sumatriptan, which
`
`Teva admits inhibits CGRP release, poorly penetrates the BBB, indicating a
`
`“peripheral point of action.” Ex. 1281, S73; Ex. 1303, 23:22-24:22; Ex. 2291,
`
`Abstract; POR, 14, 41 n.11. Other triptans with better BBB penetration showed no
`
`additional clinical benefit versus sumatriptan. Ex. 2291, 2. Ergotamine and
`
`atenolol, other prior-art migraine treatments, also acted peripherally. Ex. 1334, 329;
`
`Ex. 1241, Abstract.
`
`Teva’s cited animal studies did not dissuade development of other
`
`peripherally acting migraine drugs beyond BIBN4096BS, sumatriptan, ergotamine,
`
`and atenolol. These included CGRP-binding aptamers, which showed no
`
`“propensity to traverse the blood/brain barrier.” Ex. 1310, 2244; Exs. 1082, 1240;
`
`Ex. 1342 ¶24. Dr. Messlinger, the senior author of Teva’s cited Fischer reference,
`
`9
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`was himself investigating anti-CGRP aptamers for treating migraine. Exs. 1240,
`
`2310. A POSA would have reasonably expected an anti-CGRP antagonist antibody
`
`to successfully treat headache without crossing the BBB.
`
`B. Humanized Anti-CGRP Antibodies Were Expected To Access the
`Site of Action
`Teva contends that Tan’s full-length antibody failed to engage in the synaptic
`
`cleft. POR, 3, 17, 36-38. But Tan expressly disclosed that MAb C4.19 “clearly
`
`diffuses into the synaptic cleft,” consistent with multiple experiments with
`
`polyclonal anti-CGRP antibodies establishing in vivo effectiveness. Ex. 1022, 571
`
`(citing Ex. 1021); Exs. 1048-1050; Pet., 48-49; Ex. 1341 ¶¶15-39. Tan 1994 showed
`
`that “the concentration of [full-length] antibody had reached equilibrium in the
`
`synaptic cleft after 45 min.” Ex. 1021, 709; Ex. 1341 ¶21. Teva’s pharmacologist,
`
`Dr. Foord, did not consider Tan 1994 in preparing his declaration. Ex. 1343, 53:16-
`
`20, 56:24-57:17.
`
`Further,
`
`consistent with well-understood principles of
`
`antibody
`
`pharmacokinetics and the express guidance of Tan, a POSA would have expected
`
`that longer distribution times and/or higher doses would improve distribution of full-
`
`length antibodies to the synaptic cleft, enhancing response. Ex. 1341 ¶¶24-34;
`
`Ex. 1022, 571; Ex. 1247, 3972. Tan explains that its results were “consistent with
`
`reported antibody distribution characteristics,” in which full-length antibody
`
`10
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`molecules take a longer time to distribute than their relatively smaller Fab’
`
`fragments, particularly within the short time frames tested. Ex. 1022, 571; see also
`
`Ex. 1018 ¶148. Indeed, Tan observed a more pronounced 16% response when a
`
`higher antibody dose and a longer, two-hour incubation time were employed.
`
`Ex. 1022, 569; Pet., 18-19, 49-50; Ex. 1341 ¶¶24-26.
`
`Teva contends that Tan’s full-length antibody dosing instructions for the rat
`
`saphenous nerve assay are “speculati[ve]” and unrelated to migraine (POR, 4, 37-
`
`38), but it is undisputed that Teva followed Tan’s instructions in its patent examples
`
`employing the rat saphenous model, using both larger doses and longer distribution
`
`times. Ex. 1018 ¶¶96-108. Moreover, Teva relied on results from that assay to
`
`support claims for treating migraine. Pet., 1, 50-51. Following the express teachings
`
`of the prior art, as Teva did, is not inventive.
`
`Teva’s remaining synaptic cleft arguments are also meritless. Dr. Foord, who
`
`admits he is not an antibody expert, nonetheless argues antibodies were too large to
`
`access the synaptic cleft. Ex. 2267 ¶90; Ex. 1343, 66:9-68:10, 70:4-9. His analysis
`
`was flawed: he referenced IgE antibodies rather than the claimed IgG antibodies
`
`(which are smaller); failed to evaluate the size of relevant synapses (which are
`
`larger); and ignored the mobile, three-dimensional nature of antibodies. Ex. 2267
`
`¶90; Ex. 1341 ¶¶35-39. As demonstrated by Tan and others, a POSA would have
`
`11
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`reasonably expected an anti-CGRP antagonist antibody to engage at the synaptic
`
`cleft.
`
`C. Teva’s Unfounded Safety Concerns Do Not Support Patentability
`Teva’s unfounded safety allegations (POR, 25-36) are not addressed in Teva’s
`
`patent, which contains no safety or clinical data whatsoever. Nevertheless, by
`
`November 2005, antagonizing the CGRP pathway was recognized as safe and
`
`clinically effective. Ex. 1342 §VII, ¶¶125-126.
`
`1.
`
`Clinical Studies Demonstrated that the CGRP Pathway Could
`Be Safely Antagonized To Treat Migraine
`In 2004, Olesen confirmed that antagonizing the CGRP pathway in human
`
`patients produced no serious side effects and no cardiovascular events. Ex. 1025,
`
`1109; Ex. 1342 ¶¶82-83; Pet., 26-27; Ex. 1303, 84:13-22, 73:8-18. Contradicting
`
`Teva’s argument that “Olesen says nothing about cerebrovascular safety” (POR, 30),
`
`Olesen states that BIBN4096BS “had no constrictor effect on the middle cerebral,
`
`radial, or superficial temporal artery or on regional cerebral blood flow, blood
`
`pressure, or heart rate.” Ex. 1025, 1108.
`
`Similarly, Iovino and Petersen reported “a very favourable safety profile” for
`
`BIBN4096BS in human volunteers. Ex. 1042, Abstract (reporting no “clinically
`
`relevant, drug-induced changes” in blood pressure, pulse rate, blood flow, or vital
`
`signs); Ex. 2019, Abstract; Ex. 1342 ¶¶84-87; Ex. 1303, 84:23-85:7, 87:1-11, 90:22-
`
`92:20. In 2005, in view of these studies, both of Teva’s clinician experts
`
`12
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`independently praised “CGRP antagonists” as promising, safe antimigraine drugs
`
`“without vascular side effects.” Ex. 1290, 657; Ex. 1297, S119.
`
`Olesen’s safety results were also consistent with—and even improved upon—
`
`sumatriptan. Ex. 1342 ¶¶18, 93; Ex. 1025, 1108; Ex. 1031, 326; Ex. 2010, 2561.
`
`Despite inducing “transient” blood pressure increases, sumatriptan was considered
`
`“very safe” when prescribed to appropriate migraine patients. Ex. 1342 ¶93;
`
`Ex. 1282, 1521; Ex. 1308, 1673; Ex. 1303, 211:2-9. Indeed, Teva’s clinician experts
`
`advocated daily, long-term triptan administration for migraine prevention. Ex. 1294,
`
`Abstract, 487; Ex. 1295, Abstract, 1405; Ex. 1342 ¶19.
`
`Contrary to Teva’s arguments (POR, 31-32), safety lessons from antagonizing
`
`the CGRP pathway with BIBN4096BS were pertinent for therapeutics that directly
`
`targeted CGRP, as the prior art recognized that targeting CGRP and its receptor were
`
`known, alternative techniques. Ex. 1022, 566, 571; Ex. 1018 ¶¶158-161; Ex. 1040,
`
`182; Ex. 1342 ¶¶22-24, 86; Ex. 1341 ¶¶40-44, 53-58. For example, citing
`
`Wimalawansa, researchers had developed prior art aptamers that bound CGRP for
`
`treating migraine. Ex. 1082, Abstract, 2 (ref. 19). Anti-CGRP aptamers
`
`demonstrated efficacy in a cranial model similar to Teva’s patent examples, as well
`
`as vascular safety: “[b]asal blood flow and systemic arterial pressure were
`
`unchanged.” Ex. 1240, 923; Ex. 1342 ¶80; Ex. 1341 ¶¶43, 56-58, 60; Ex. 1001,
`
`69:50-70:56. The prior art explained that “aptamers can be thought of as … analogs
`
`13
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`to antibodies,” recognizing their “long in vivo half-life” as a benefit. Ex. 1309,
`
`Abstract; Ex. 1240, 923; Ex. 1082, Abstract, 7; Ex. 1342 ¶24; Ex. 1341 ¶¶57, 60.
`
`2.
`
`Long-Acting Ligand Antagonists Had Desirable Benefits and
`Did Not Raise Safety Concerns
`Teva argues that the transient and mild increase in blood pressure observed in
`
`Tan would raise safety concerns that could be exacerbated by the long half-lives of
`
`CGRP-blocking antibodies. POR, 28-31. To the contrary, a POSA would have
`
`viewed anti-CGRP drugs with longer half-lives as desirable for preventative
`
`migraine treatments, without posing safety concerns. Ex. 1342 ¶¶11-24, 90-100;
`
`Ex. 1341 ¶¶33, 53-76; Pet., 30-31.
`
`As Dr. Balthasar explains,1 a POSA would have understood that the minor
`
`blood-pressure increase in Tan’s anesthetized rats normalized within 10 to 15
`
`minutes of administering MAb C4.19, and “had no relationship to the half-life” of
`
`the antibody. Ex. 1341 ¶¶62-66; Ex. 1022, 568. A POSA would not have viewed
`
`minor, transient blood-pressure increases as a safety concern, as similar increases
`
`observed with sumatriptan and other CGRP-pathway antagonists did not deter their
`
`
`1 In contrast to Dr. Balthasar’s antibody-pharmacology expertise, Drs. Foord and
`
`Ferrari admitted to tenuous antibody knowledge. Ex. 1300, 33:8-11; Ex. 1303,
`
`69:10-16.
`
`14
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`development or FDA approval. Ex. 1342 ¶¶90-94; Ex. 1341 ¶¶65, 72-76; Ex. 1303,
`
`25:11-17.
`
`Other references further confirm that anti-CGRP antagonist antibodies did not
`
`have any chronic blood-pressure or vascular effects. Ex. 1341 ¶¶67-71; Ex. 1342
`
`¶¶92-102. For instance, Wong conducted blood pressure testing on the same
`
`antibody (4901) evaluated in Teva’s patent and concluded that it “had no significant
`
`effect on [mean arterial pressure] and heart rate.” Ex. 1033, 101; Ex. 1001, 52:66.
`
`Similarly, Andrew confirmed that immunized animals with “high levels of
`
`circulating antibodies to rat CGRP … did not show any signs of physical or
`
`behavioural abnormality” after 10-15 weeks. Ex. 1055, 88, 93. Salmon and other
`
`researchers disabled αCGRP production entirely in knockout mice, and rather than
`
`experiencing any safety concerns, they claimed therapeutic uses of anti-CGRP
`
`antagonist antibodies for treating neurogenic inflammatory pain. Ex. 1027, ¶[0069],
`
`claims 8, 9; Ex. 1026, claim 2; Ex. 1288, Abstract. Teva disregards this compelling
`
`evidence of safety entirely.
`
`Teva incorrectly attempts to undermine Tan’s disclosures by characterizing
`
`Tan as a “basic research paper” and citing Dr. Ferrari’s purported personal
`
`knowledge of its authors. Ex. 2270 ¶147; POR, 4. But in describing his own work,
`
`Dr. Tan wrote in 1994 that there was “no reason” why humanized anti-CGRP
`
`15
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`monoclonal antibodies should not be developed and used as “therapeutic agents” for
`
`migraine and other diseases. Ex. 1287, 247.
`
`3.
`
`The Prior Art Would Not Have Dissuaded a POSA from
`Pursuing a Humanized Anti-CGRP Antagonist Antibody for
`Treating Migraine
`Ignoring more recent prior art (supra §§III.C.1-2 and below), Teva cites
`
`several early studies, largely published before Wimalawansa proposed exploring
`
`humanized anti-CGRP antibodies, to support its hypothetical safety arguments.
`
`POR, 25-28; Ex. 2270 ¶¶104-124. But none of Teva’s cited studies indicate that
`
`antagonizing endogenous CGRP—as an anti-CGRP antagonist antibody would have
`
`been expected
`
`to do—would have had any effect on cardiovascular or
`
`cerebrovascular safety. Ex. 1342 ¶¶67-74; Ex. 1303, 108:25-133:7.
`
`For example, Teva relies on early studies reporting the effects of administering
`
`exogenous CGRP (Exs. 2058, 2079, 2139), which increases CGRP levels instead of
`
`antagonizing endogenous CGRP. Ex. 1342 ¶70. Teva also relies on early studies
`
`attempting to discern CGRP’s biological effects without the benefit of a specific
`
`antagonist, making it impossible to separate the effects of CGRP from other
`
`vasopeptides. Id., ¶¶71-72; Exs. 2150, 2151, 2154, 2070, 2089, 2209. In the one
`
`study that used CGRP8-37 (a receptor antagonist) under physiological conditions, no
`
`vascular changes were observed. Ex. 2152, 165; Ex. 1342 ¶73; Ex. 1303, 111:23-
`
`119:13.
`
`16
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`Teva also ignored more recent prior-art studies demonstrating that blocking
`
`the effects of endogenous CGRP with specific antagonists does not worsen ischemic
`
`events. Ex. 1342 ¶¶75-81; Ex. 1283, 498 (“locally released CGRP does not function
`
`as a cardioprotective agent”); Ex. 1284, Abstract (CGRP-antagonism had no effect
`
`on infarct size); Ex. 1303, 134:23-136:17, 140:5-7, 142:2-7. Two 2003 publications
`
`similarly concluded that endogenous CGRP played no major role in cardiovascular
`
`regulation, including in late-stage heart failure. Ex. 1318, 76; Ex. 1285, Abstract;
`
`Ex. 1303, 142:24-143:22.
`
`Teva even overlooked a 2004 publication from Dr. Ferrari’s advisor, Dr.
`
`Saxena, studying global and regional cardio- and cerebro-vascular effects of
`
`antagonizing the CGRP pathway with BIBN4096BS. Ex. 1263, Abstract; Ex. 1303,
`
`97:10-98:2. No undesired effects were observed, even at high doses, leading the
`
`investigators to conclude that “endogenous CGRP does not play an important role in
`
`regulating systemic and regional haemodynamics.” Ex. 1263, 296; Ex. 1303, 102:9-
`
`106:19. Summarizing the pre-clinical and clinical data, Dr. Saxena advocated the
`
`use of CGRP antagonists “in patients with coronary artery disease.” Ex. 1031, 326;
`
`Ex. 1342 ¶¶88-89.
`
`Teva’s arguments regarding the risk of stroke in sub-populations of migraine
`
`patients are irrelevant. POR, 26-28; Bayer Healthcare Pharm., Inc. v. Watson
`
`Pharm., Inc., 713 F.3d 1369, 1376 (Fed. Cir. 2013) (rejecting patient sub-population
`
`17
`
`

`

`IPR2018-01712
`Patent No. 9,884,908
`arguments where the “claims at issue do not distinguish between target patient
`
`populations”). No correlation existed between migraine and stroke for two-thirds of
`
`migraine patients. Ex. 2157, 536; Ex. 1040, 177; Ex. 1303, 193:3-10; Ex. 1342
`
`¶108.
`
`Moreover, the absolute risk of stroke and myocardial ischemia in young
`
`women with migraine was very low. Ex. 2157, 535; Ex. 1303, 190:3-194:23;
`
`Ex. 1315, Abstract; Ex. 1342 ¶¶106-111. Teva also ignores that anti-migraine
`
`treatments could be contraindicated in patients with particular risk factors, as had
`
`been done with sumatriptan and ergots, and such concerns did not discourage
`
`researchers from pursuing anti-CGRP therapies in the prior art (e.g., aptamers,
`
`antibodies, and BIBN4096BS). Ex. 1342 ¶¶11-24, 112-113; Ex. 1341 ¶¶77-79;
`
`Ex. 1282, 1520; Ex. 1290, 564-565; Exs. 1026, 1027.
`
`D. Teva’s Hypothetical “Spare Receptor Theory” and Ligand Cross-
`Binding Concerns Are Incorrect
`Teva asserts that targeting the CGRP ligand would be undesirable based on
`
`Teva’s theory that “less than 1% of [CGRP] receptors needed to be bound by ligand
`
`to elicit a full response.” POR, 23; Ex. 1341 ¶¶45-52.
`
`The prior art contradicts Teva’s hypothetical, indicating the spare receptor
`
`theory does not apply. During cross-examination, Dr. Foord conceded that Teva’s
`
`spare-receptor reference, Sheykzade, discloses that 27% CGRP receptor o