`
`
`
`
`
`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 IPR2018-01711
`Patent 9,884,907
`_____________________
`
`
`PATENT OWNER’S SURREPLY
`
`
`
`
`
`Mail Stop "PATENT BOARD"
`Patent Trial and Appeal Board
`U.S. Patent and Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`
`
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`
`2. 
`
`3. 
`
`B. 
`
`TABLE OF CONTENTS
`
`I. 
`Introduction ...................................................................................................... 1 
`Claim construction ........................................................................................... 3 
`II. 
`III.  Lilly cannot argue safety is irrelevant when it relied on “lower
`toxicity” of humanized antibodies as a reason to modify. .............................. 4 
`IV.  Lilly’s obviousness inquiry fails...................................................................... 6 
`A. 
`Lilly
`repeatedly
`(but
`incorrectly) extrapolates small
`molecule data to anti-CGRP antibodies and dismisses
`pharmacological differences between receptor and ligand
`antagonism ............................................................................................. 6 
`Lilly failed to demonstrate that humanized anti-CGRP
`antibodies would have been safe in humans, and have had
`therapeutic advantages for treating migraine. ....................................... 8 
`1. 
`Lilly ignores myriad teachings showing CGRP’s
`vasoprotective role. .................................................................... 8 
`Risk concerns for treating migraine with long-acting
`anti-CGRP antibodies are not “unfounded.”.............................. 9 
`Any alleged safety of blocking the CGRP pathway
`with small molecules,
`receptor antagonists, or
`aptamers is irrelevant to the safety of anti-CGRP
`antibodies. ................................................................................ 11 
`Lilly improperly dismisses safety concerns with
`antagonizing CGRP in migraineurs. ........................................ 15 
`The record does not demonstrate a reasonable expectation
`that a full-length anti-CGRP antibody would treat migraine. ............. 17 
`1. 
`A full-length anti-CGRP antibody would not have
`been expected to access the site of action. ............................... 18 
`The uncertainties in the field—which Lilly improperly
`ignores or dismisses—further diminish a reasonable
`expectation of success. ............................................................. 21 
`Lilly’s attempt to distinguish Novartis fails. ................................................. 24 
`V. 
`VI.  Lilly never articulated which prior art antibody a POSA would have
`humanized and administered to humans in order to arrive at the
`claimed methods. ........................................................................................... 25 
`
`C. 
`
`4. 
`
`2. 
`
`- i-
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`VII.  Lilly misreads Teva’s secondary indicia arguments, and has not
`rebutted the presumption of nexus................................................................. 26 
`VIII.  Conclusion ..................................................................................................... 29 
`
`
`
`- 2 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`I.
`
`Introduction
`
`Teva’s Patent Owner Response (“POR”) exposed multiple infirmities that
`
`defeat Lilly’s obviousness case. Lilly’s principal references—Olesen and Tan—
`
`would not have motivated a POSA to treat migraine with anti-CGRP antibodies.
`
`Olesen describes a migraine study with BIBN4096BS, which, contrary to Lilly’s
`
`allegations, cannot be extended beyond small-molecule receptor antagonists. Tan is
`
`even further removed from the claimed methods: it is a basic research paper
`
`attempting to “prob[e] the role of CGRP as an endogenous vasodilator” in rats,
`
`reporting that a full-length anti-CGRP antibody failed to show immunoblockade in
`
`vivo.
`
`Unable to overcome these and other fatal defects, Lilly instead argues that
`
`the claims “do not require clinical efficacy and do not mention safety,” and Lilly
`
`does not have the burden of showing that “the claimed methods would be clinically
`
`effective and safe.” Reply, 3. Lilly is wrong. First, the claims must be construed to
`
`require therapeutic efficacy. Second, it was Lilly who relied on an alleged “lower
`
`toxicity” of humanized antibodies as a reason to combine the art; Lilly cannot now
`
`retreat from the relevance of safety considerations on Reply. Teva rebutted Lilly
`
`by, inter alia, showing that Lilly failed to demonstrate efficacy and to fully
`
`consider safety.
`
`Lilly’s hindsight-driven argument selectively cherry-picks references Lilly
`
`- 1 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`believes support its arguments, while ignoring references that undermine them.
`
`Even worse, on cross, Lilly’s expert Dr. Charles and his replacement, Dr.
`
`Balthasar1, distanced themselves from unfavorable portions of Lilly’s own
`
`references. And they refused to consider teachings that highlight safety concerns
`
`associated with long-term inhibition of CGRP.
`
`On Reply, Lilly pivots from its initial rationale, arguing instead that Lilly
`
`does not have to consider safety. But Lilly cannot re-craft its challenge to attempt
`
`to rehabilitate its Petition. Henny Penny Corp. v. Frymaster LLC, No. 18-1596
`
`(Fed. Cir. 2019).
`
` Lilly’s new arguments that Covell’s carcass studies support its speculation
`
`that Tan’s antibody would eventually reach its site of action given more time fail,
`
`because “assignment of a site … of antibody localization was not possible.”2
`
`Similarly, Lilly’s speculation about “increased dose” goes against safety concerns
`
`regarding long-term CGRP ligand antagonism, which would remove CGRP’s
`
`protective role during ischemic events, where the risk of stroke and heart attacks
`
`are elevated.
`
`
`1 On Reply, Lilly proffered testimony from new expert Dr. Balthasar to
`
`repair Dr. Charles’ discredited opinions. POR, 3-4.
`
`2 Emphasis added throughout unless otherwise noted.
`
`- 2 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`Lilly also fails to rebut Teva’s strong showing of numerous indicia of non-
`
`obviousness, which support confirming the challenged claims. In short, Teva
`
`demonstrates that Lilly’s Petition fails to prove obviousness of the claimed
`
`methods. Lilly’s Reply does not salvage its Petition.
`
`
`
`Finally, in light of Arthrex v. Smith & Nephew, No. 2019 WL 5616010 (Fed.
`
`Cir. Oct. 31, 2019), the Board should stay this proceeding until the Arthrex
`
`mandate issues. Until then, the Board is operating unconstitutionally. Fed. R. App.
`
`P. 41(c). Also, the Arthrex remedy is constitutionally insufficient. Cf. Polaris
`
`Innovations v. Kingston Tech., No. 2018-1768, Dkt. No. 90, 2 (Fed. Cir. Nov. 8,
`
`2019).
`
`II. Claim construction
`Teva disagrees with Lilly’s argument that the term “treating” does not
`
`require “achievement of any clinical result.” Reply, 2. As Dr, Ferrari explained, a
`
`physician’s goal in treating patients for migraine is to obtain “beneficial or desired
`
`clinical results.” EX2269, ¶19; POR, 11. Dr. Charles also agrees that, as a treating
`
`physician, he would not be “satisfied with a therapeutic that did not result in a
`
`clinical change in a patient.” EX2336, 41:19-42:2. Thus, the challenged claims
`
`require at least a reasonable expectation that treatment with anti-CGRP antibodies
`
`would have a beneficial clinical result in an individual. POR, 10-11.
`
`The specification defines “effective amount” as “an amount sufficient to
`
`- 3 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`
`effect beneficial or desired results,” such as “eliminating or reducing the risk,
`
`lessening the severity, or delaying the outset of the disease,” “reducing pain
`
`intensity, duration, or frequency of headache attack, and decreasing one or more
`
`symptoms resulting from headache,” etc. EX1001, 19:3-22; POR, 11; EX2269,
`
`¶21; EX2266, ¶54. The Board agrees. Paper 12, 11-13. The claims, therefore,
`
`require clinical results.
`
`Lilly misreads “effective amount” to include amounts “sufficient to reduce
`
`biochemical or histological symptoms … without requiring any clinical result.”
`
`Reply, 3. Yet, the specification unambiguously ties biochemical or histological
`
`symptoms to clinical results, as explained above. EX1001, 19:3-22. Dr. Charles
`
`agreed that a physician treating migraine would want “to administer an effective
`
`amount that will help you achieve that clinical response.” EX2336, 66:1-5. This
`
`admission undercuts Dr. Charles’ testimony that “biochemical or histological
`
`changes” merely requires “inhibiting cAMP activation,” which admittedly “can
`
`change without any effect on symptoms of a disease.” EX1340, ¶¶7-8; EX2336,
`
`57:13-15.
`
`III. Lilly cannot argue safety is irrelevant when it relied on “lower toxicity”
`of humanized antibodies as a reason to modify.
`
`Lilly admits that a POSA would have considered side-effects in developing
`
`human treatment methods with antibodies: the Petition argued that a POSA would
`
`- 4 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`have chosen anti-CGRP antibodies to treat migraine “to avoid the known side
`
`effects of existing small-molecule migraine drugs,” and that humanized antibodies
`
`avoid “serum sickness and other harmful effects.” Petition, 30-33; POR, 24-25.
`
`Facing the art-recognized cardiovascular and cerebrovascular safety
`
`concerns associated with administering anti-CGRP antibodies to humans—which
`
`Lilly’s Petition ignored—Lilly downplays safety considerations on Reply. Reply,
`
`3; POR, 24-28; EX2269, ¶¶113-149, 153; EX2266, ¶¶57, 117-128. But Lilly’s
`
`Petition rationale, which requires reviewing the record and claimed therapeutic
`
`methods through the lens of clinical safety, controls. Henny Penny, No. 18-1596, 9
`
`(“[A]n IPR petitioner may not raise in reply ‘an entirely new rationale’ for why a
`
`claim would have been obvious.”). The Board should also reject Lilly’s belated
`
`attempts to fill holes in its Petition with new art and expert testimony.
`
`Lilly’s Reply argument that Teva’s own specification fails to
`
`“acknowledg[e] []or provide[] solutions for any of Teva’s alleged concerns” as
`
`“persuasive evidence [safety] concerns were not real,” is unavailing. Reply, 1, 3.
`
`Settled law mandates assessment of obviousness from a POSA’s point of view, not
`
`the patentee’s: “one should not go about determining obviousness under [section]
`
`103 by inquiring into what patentees (i.e., inventors) would have known or would
`
`likely have done, faced with the revelations of references.” Standard Oil Co. v.
`
`American Cyanamid Co., 774 F.2d 448, 454 (Fed. Cir. 1985); see also Kimberly-
`
`- 5 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`Clark Corp. v. Johnson & Johnson, 745 F.2d 1437, 1454 (Fed. Cir. 1984).
`
`Thus, the Board must consider art-recognized safety concerns of CGRP-
`
`antagonism in its obviousness assessment.
`
`
`
`IV. Lilly’s obviousness inquiry fails
`A. Lilly repeatedly (but incorrectly) extrapolates small molecule data
`to anti-CGRP antibodies and dismisses pharmacological
`differences between receptor and ligand antagonism
`
`Lilly’s obviousness argument rests on Olesen’s alleged “proof of concept”
`
`with the small-molecule receptor antagonist BIBN4096BS, which Lilly extends “to
`
`other CGRP antagonists.” Petition, 25-26; Reply, 4-5; EX1025, 1104, 1108-1109.
`
`But as Teva previously explained, nothing in Olesen (or any cited art) supports
`
`Lilly’s broad reading of “CGRP antagonists.” POR, 12-14; EX2266, ¶¶60-63;
`
`EX2269, ¶98.
`
`Lilly’s Reply attempts to pin Teva’s expert’s references to “CGRP
`
`antagonists” as allegedly touting “blockade of the CGRP pathway” and a “new
`
`emergent treatment principle” for migraine. Reply, 5; EX1332, 443; EX1290, 657.
`
`Putting aside that EX1332 was published years after November 14, 2005, and
`
`cannot inform what a POSA would have concluded from Olesen in 2005, EX1332,
`
`EX1290, and EX1297 undeniably discuss only BIBN4096BS as a “CGRP
`
`antagonist.” EX1332, 443; EX1290, 657; EX1297, S119. Thus, nothing in the
`
`record supports Lilly’s extension of “CGRP antagonists” beyond BIBN4096BS.
`
`- 6 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`Lilly also continues to ignore material differences between receptor and
`
`
`
`ligand antagonism. POR, 20-24; EX2266, ¶¶64-76. And Dr. Foord’s testimony that
`
`receptor and ligand antagonism are not “alternative” because of CGRP receptor
`
`reserve remains sound and effectively unrebutted. EX2266, ¶75; EX1300, 69:18-
`
`22; EX2339, 112:21-114:9. Lilly instead mischaracterizes Dr. Foord’s testimony
`
`on receptor reserve, arguing that he allegedly “conceded” that “27% CGRP
`
`receptor occupancy is required to generate a half-maximal response.” Reply, 19-
`
`20; EX2266, ¶¶41-45, 69-76. This is misleading: Dr. Foord explained that 27%
`
`receptor occupancy was observed in arteries containing lower receptor reserve, not
`
`in the relevant microvasculature environment, which has a high receptor reserve.
`
`EX2266, ¶71; EX2065, 1070-1071, Figure 2; EX1096, 559; EX2003, 915;
`
`EX2219, 57; EX2339, 112:8-19. And while Dr. Charles argued that “normalizing
`
`elevated or inappropriate levels of CGRP” is all that is needed to treat migraine,
`
`Dr. Charles did not provide any evidence that such “normalization” could be
`
`achieved with anti-CGRP antibodies, or that it would result in a safe and effective
`
`migraine treatment. Reply, 19 (emphasis in the original); EX1340, ¶¶114-118.
`
`Lilly’s Reply also did not cure its failure to consider cross-reactivity
`
`concerns in the Petition. Reply, 19-20; POR, 21. On Reply, Dr. Charles belatedly
`
`testified—without explanation or support—that “hypothetical and unsupported
`
`concerns about ligand-receptor cross-binding would not have deterred
`
`- 7 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`development of a humanized anti-CGRP antagonist antibody.” EX1340, ¶120;
`
`Reply, 19-20. He misses the point: one cannot equate receptor and ligand
`
`antagonism without considering the differences, which he failed to do. EX2266,
`
`
`
`¶68.
`
`B.
`
`Lilly failed to demonstrate that humanized anti-CGRP antibodies
`would have been safe in humans, and have had therapeutic
`advantages for treating migraine.
`
`Lilly’s Petition argued that a POSA “would have chosen antibodies to avoid
`
`the known side effects of existing small-molecule migraine drugs.” Petition, 30,
`
`EX1016, ¶135. But Lilly’s Dr. Balthasar undermined any broad assertion that
`
`antibodies are necessarily safer than small-molecule drugs. EX2337, 52:8-19
`
`(conceding target-related toxicity and death from a therapeutic antibody); EX1057,
`
`1348. Furthermore, Dr. Charles admitted: “we’re always concerned about safety
`
`issues, including stroke, in any circumstance where we’re administering therapy.”
`
`EX2338, 81:3-5. Yet, the Petition did not address how to overcome safety concerns
`
`associated with increased incidence of stroke in migraineurs or long-term depletion
`
`of CGRP. The Reply fails to fill this hole.
`
`1.
`
`Lilly ignores myriad teachings showing CGRP’s
`vasoprotective role.
`
`Teva demonstrated the importance of CGRP for vascular health as of 2005.
`
`POR, 24-28. Lilly is wrong to dismiss CGRP’s potent cerebro- and
`
`- 8 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`cardioprotective roles. Id.
`
`
`
`By 2005, the art recognized CGRP’s “pivotal role” in the “physiology and
`
`pathophysiology of cardiovascular regulation,” including protecting tissues during
`
`ischemia. EX2003, 923; EX2269, ¶113; POR, 24-25; EX2338, 25:8-26:17, 29:16-
`
`30:25, 35:1-36:21; EX2340, 53-54; EX2341, 246. Lilly’s EX1284 demonstrates
`
`this: CGRP reduced infarct size in an ischemia rat model by up to 89%, while
`
`BIBN4096BS blocked “[t]he cardioprotective effect of CGRP,” concluding:
`
`“CGRP is a very potent myocardial protective substance.” EX1284, 591-593,
`
`Figure 3; EX2338, 21:17-22:4. Lilly’s papers omitted this unfavorable information,
`
`and its expert refused to even acknowledge it as “germane” on cross. EX2338,
`
`20:1-21:3. Lilly’s refusal to consider CGRP’s vasoprotective role dooms its
`
`obviousness challenge.
`
`2.
`
`Risk concerns for treating migraine with long-acting anti-
`CGRP antibodies are not “unfounded.”
`
`Lilly does not contest that anti-CGRP antibodies increase blood pressure in
`
`animals. Reply, 14-16; EX1022, 568, Figure 23; EX2084, H687; EX1033, 102,
`
`
`3 Lilly newly-introduced the Tan “thesis” (EX1287) on Reply, though none
`
`of its experts cite it. Reply, 6, 16. Beyond untimeliness (see Paper 38), Lilly has
`
`not established that EX1287 was publicly available or qualifies as prior art. Lilly
`
`- 9 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`
`Table 24; EX2269, ¶128; EX2266, ¶¶122, 125; POR, 26-27. Instead, Lilly attempts
`
`to undermine these findings by arguing that the blood pressure increases were
`
`“mild” and “transient,” and would not have been a concern. Reply, 14. Lilly
`
`wrongfully discounts these findings; according to Dr. Ferrari’s unrebutted
`
`testimony, even “mild” or “transient” increases in BP can have a significant effect:
`
`“mortality from a myocardial infarction (MI) or cerebrovascular accidents doubles
`
`for each 20-mm Hg increase in systolic blood pressure (BP) above 115 mm Hg.”
`
`EX2127, S383; EX2269, ¶131. Further, none of Lilly’s antibody-related references
`
`determine the consequences of obliterating CGRP-mediated vasodilation during
`
`an ischemic attack—the real concern in the field in 2005. EX2269, ¶¶132-136,
`
`159; POR, 8.
`
`
`also failed to establish that a POSA would ignore the weight of the prior art and
`
`lack of suitability of a full length anti-CGRP antibody for human therapeutic use.
`
`4 Lilly relies on Wong’s allusion to safety (Reply, 15-16), but Dr. Balthasar
`
`admitted he could not “draw direct reference from the sentence [referring to MAP
`
`and heart rate] alone to which study it refers to.” EX2339, 158:17-20.
`
`- 10 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`3.
`
`Any alleged safety of blocking the CGRP pathway with
`small molecules, receptor antagonists, or aptamers is
`irrelevant to the safety of anti-CGRP antibodies.
`Lilly incorrectly argues that “the CGRP pathway5 could be safely
`
`antagonized to treat migraine” because “more recent” art (see graphic below)
`
`shows that no safety concern existed. Reply, 12, 16-17. Further, Lilly’s small
`
`molecules, receptor antagonists, and aptamer art does not inform about safety of
`
`anti-CGRP antibodies. Reply, 13-14; EX1283, Abstract; EX2152, 165; EX1284,
`
`Abstract; EX1285, Abstract; EX1318, Abstract; EX1263, Abstract.
`
`
`5 Lilly disguises its reliance on receptor antagonist art by using the phrases
`
`“CGRP pathway” or “CGRP antagonist.” Reply, 2, 4-5, 7-8, 12-15, 17.
`
`- 11 -
`
`

`

`Case IPR2018-01711
`
`Case IPR2018-01711
`Patent No. 9,884,907
`Patent No. 9,884,907
`
`Ex. 2079
`
`- CGRP “Improves myocardial contractility
`in patients with congestive heart failure"
`
`- Evaluated circadian variation in plasma
`concentrations of CGRP
`
`- CGRP “delays the onset of myocardial
`ischaemia during treadmill exercise"
`
`- comic-37)
`
`- mama-37)
`
`|:|= Teva Exhibit
`!= Lilly Exhibit
`|:|= Studies showing CGRP vasoprotecfive activity
`I:I= Short-term CGRP receptor antagonist studies
`- = Short-term aptamer study
`
`- CGRP “may be acting as a local mediat-
`in the pathophysiology of [acute myocardial
`infarction]”
`
`- CGRP levels were “reduced” after
`subarachnoid haemorrhage (SAH)
`
`
`
`
`
`cardiovascular episodes."
`
`- Age-related changes in CGRP levels “may
`contribute to the increased incidence of
`
`- CGRP“may reverse significantlythe W
`vasoconstriction seen after SAH"
`
`- CGRP(8-37)
`
`the myocardium during ischemia.”
`
`- CGRP is possibly involved in “protection of
`
`
`
`
`
`- CGRP(8-37)
`
`- BIBN4DQGBS
`
`CGRP protects against “hypertension-
`induced" organ damage and from
`"ischemiai’reperfusion injury"
`
`- BIBN4OQGBS
`
`
`
`- CGRP plays a “pivotal role" in the
`
`"physiology and pathophysiology of
`cardiovascular regulation."
`
`
`
`
`
`
`- 12 -
`
`-12-
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`
`BIBN4096BS: BIBN4096BS is not probative of anti-CGRP antibody safety.
`
`BIBN4096BS has a “significantly” shorter half-life than an IgG, which circulates
`
`in the body for weeks. POR, 29-30; EX2272, ¶¶54-56, 83; EX2269, ¶¶50, 100,
`
`104; EX2266, ¶65. Any adverse effects from BIBN4096BS would be cleared
`
`within 24 hours after administration. POR, 29-30; EX2338, 86:13-87:9; EX2266,
`
`¶¶65, 120; EX2269, ¶50; EX2272, ¶¶55-56, 83; EX1042, 645; Petition, 30;
`
`EX1059, 143, Figure 4.16; EX2338, 85:15-86:12. Moreover, each BIBN4096BS
`
`study was short-term, and would not have informed a POSA of the long-term risks,
`
`expected with only a single administration of antibody. POR, 30; EX2269, ¶123;
`
`EX2266, ¶131; EX1025, 1104; EX1042, 647; EX2019, Abstract; EX1290, 657;
`
`EX1297, S119; EX2338, 83:22-88:12.
`
`Further, the BIBN4096BS studies were all done in “healthy volunteers,”
`
`which would not assess whether BIBN4096BS would “block rescue mechanisms
`
`in the times of ischemia.” EX1303, 87:5-17, 89:21-90:3; EX2269, ¶¶113-144;
`
`EX2193, 222; EX2157, 533; EX1025, 1108; EX1042, 647; EX2338, 83:22-84:13;
`
`POR, 7-8, 24-25.
`
`Dr. Charles ignored Olesen’s warning against relying on its BIBN4096BS
`
`study for cardiovascular safety: “our data base was too small for us to assess
`
`cardiovascular safety.” EX1025, 1109; POR, 29; EX2269, ¶50; EX1340, ¶83. Dr.
`
`Balthasar, however, confirmed that Olesen’s statement is consistent with “what I
`
`- 13 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`would expect from a clinical investigation” because one “can’t make
`
`extrapolations off of—beyond the sample size that’s present.” EX2339, 119:5-
`
`120:4. Finally, it is of no consequence that an increase in blood pressure observed
`
`with “other CGRP-pathway antagonists” “did not deter their development.” Reply,
`
`15. A POSA would not have equated short-term receptor antagonism with long-
`
`term ligand antagonism. POR, 20-24, 29-31.
`
`Triptans: Triptans have a considerably shorter half-life than IgGs. EX1282,
`
`7; EX2338, 99:1-4; EX1340, ¶19; EX2272, ¶¶54-59; Petition, 29-30; EX1059,
`
`143, Figure 4.16; EX2266, ¶¶65, 120; POR, 29-30. Thus, any adverse effects of
`
`triptans are eliminated from the body within hours. Triptans also operate by a
`
`different mechanism from anti-CGRP antibodies: triptans agonize 5-HT receptors.
`
`EX1040, 180-181; EX2269, ¶49; EX2338, 99:20-100:1; EX1303, 23:22-24:18.
`
`Despite mechanism-of-action and pharmacokinetic differences, Dr. Charles
`
`analogizes antibody-induced BP increases to similar increases observed with
`
`triptans, arguing (without explanation) that such transient changes were not of
`
`concern. EX1340, ¶93. His analogy fails to explain the basis for equating the two
`
`classes of molecules.
`
`Aptamers: Aptamers have a short half-life—“hours to days”—and would
`
`not have been informative about safety of long-acting antibodies. EX1309,
`
`Abstract; EX2338, 114:6-115:5. Aptamers have characteristics that are similar to
`
`- 14 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`small-molecules and are not antibody “analogs.” EX1309, Abstract; Reply 14.
`
`Further, the record lacks evidence that aptamers would have been safe in humans.
`
`The only in vivo study in the record with aptamers is an abstract that lacks any
`
`experimental detail for its mouse model. EX1240, Abstract F022; EX2338, 115:19-
`
`117:7, 118:9-120:24.
`
`*
`
`*
`
`*
`
`
`
`In sum, the record lacks any evidence that treating migraine with anti-CGRP
`
`antibodies in humans would have been safe.
`
`4.
`
`Lilly improperly dismisses safety concerns with
`antagonizing CGRP in migraineurs.
`
`The Reply dismisses migraine’s comorbidity with stroke and hypertension as
`
`“irrelevant” because “claims at issue do not distinguish between target patient
`
`populations.” Reply, 18. But it was Lilly—not Teva—who focused on migraine,
`
`making safety considerations in migraine patients relevant. Petition, 24-33;
`
`EX2192, 73:6-74:11.
`
`As Dr. Ferrari explains, it was known that CGRP played a key role in
`
`preventing escalation of mild ischemic events into full-blown infarctions
`
`(stroke/heart attack). EX2269, ¶¶104-105, 119-124; POR, 8; EX2338, 80:1-82:20;
`
`EX2343, 1123. Lilly attempts to minimize CGRP’s importance by asserting “the
`
`absolute risk” of stroke and myocardial ischemia in young women is “low.” Reply,
`
`- 15 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`18; EX1340, ¶¶107-109. But a patient need not have a stroke or myocardial
`
`infarction for the concern over CGRP antagonism to be pertinent. A POSA would
`
`have been concerned with “common” ischemic episodes, such as transient
`
`ischemic attacks (TIAs) and angina,6 expecting that long-term loss of CGRP’s
`
`protective effect could lead to the development of more serious events. EX2269,
`
`¶¶108-110, 114, 119-124.
`
`TIAs are a known stroke precursor in about 5,000-12,500 patients annually.
`
`EX2144, 1665; EX2269, ¶108; EX2338, 65:12-68:6; EX2342, 2901. Any increase
`
`in the incidence or severity of TIAs would have been a serious concern. Dr.
`
`Charles offered no rebuttal to Dr. Ferrari’s testimony that CGRP inhibition would
`
`worsen common ischemic episodes in migraineurs. Dr. Charles’ only argument is
`
`that “anti-migraine treatments could be contraindicated in patients with particular
`
`risk factors, as had been done with sumatriptan and ergots.” Reply, 18; EX1340,
`
`¶¶112-113. Yet, he has not demonstrated that a contraindication solution, like that
`
`for sumatriptan and ergots, would be sufficient for anti-CGRP antibodies, given
`
`different mechanisms of action of these molecules.
`
`
`6 Dr. Charles wrongly refused to acknowledge that “angina” is an ischemic
`
`event. EX2338, 55:9-11; EX2269, ¶109; EX2147, S332.
`
`- 16 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`C. The record does not demonstrate a reasonable expectation that a
`full-length anti-CGRP antibody would treat migraine.
`
`
`
`Tan confirmed: engagement of CGRP at the synaptic cleft is “a prerequisite
`
`for immunoblockade.” EX2266, ¶102; EX1022, 571, 565-566, Abstract; POR, 36.
`
`Lilly failed to prove that a POSA would have had a reasonable expectation that an
`
`anti-CGRP antibody can distribute into the synaptic cleft—and thus treat
`
`migraine—when Tan’s IgG did not. POR, 3; EX1022, 568; EX2266, ¶¶82-91;
`
`EX2269, ¶¶137-138. None of Lilly’s additional references (EX1048-1050)
`
`demonstrate antibody’s ability to enter synaptic clefts during migraine, as
`
`explained below.
`
`Lilly is forced to rely on Tan’s speculation: “[w]ith repeated administration,
`
`IgG should eventually ... achieve the sufficiently high concentrations required for
`
`immunoblockade” in Tan’s assay. Petition, 45; EX1022, 571. This speculation is
`
`unsubstantiated and “provides no more than hope,” which “is not enough to create
`
`a reasonable expectation of success [for a treatment method] in a highly
`
`unpredictable art.” Osi Pharmaceuticals, LLC v. Apotex Inc., 2018-1925, 18 (Fed.
`
`Cir. 2019). This is especially true here, given art-known uncertainties related to
`
`central versus peripheral site of action of migraine drugs and the validity of CGRP
`
`as a migraine biomarker.
`
`- 17 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`1.
`
`A full-length anti-CGRP antibody would not have been
`expected to access the site of action.
`
`Nothing in the record demonstrates that IgGs would distribute into the
`
`synaptic cleft—the relevant site of action. Reply, 10. Tan’s C4.19 failed to engage
`
`in the synaptic cleft in the in vivo assay. EX2266, ¶134; EX1022, 571, 565,
`
`Abstract; POR, 35-36. Lilly’s Dr. Vasserot agreed with Teva’s experts on this
`
`shortcoming and the unreliability of Tan’s results. EX2191, 118:12-119:1; POR, 3-
`
`4.
`
`Unable to demonstrate synaptic cleft distribution of C4.19 in vivo, Lilly
`
`defaults to Tan 1994’s data with isolated vas deferens. Reply, 10; EX1339, ¶20.
`
`But Tan 1994’s “tissue bath” is not an “equivalent” to an in vivo study, as Dr.
`
`Balthasar admits, and does not represent an antibody’s ability to penetrate multiple
`
`biological compartments in a complex system. EX2339, 77:20-78:3; EX1021, 705.
`
`Further, Dr. Foord’s unrebutted testimony demonstrates: “the vas deferens is not
`
`symbolic of a synapse” because it is an “electrically conductive tissue,” not
`
`“dependent of the work of separate synapses.” EX1343, 61:17-62:4. Thus, Tan
`
`1994 does not demonstrate IgG access to the synaptic cleft in vivo.
`
`Lilly’s reliance on Louis and Dockray is unavailing. There, the antibodies
`
`- 18 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`“leaked” into the interstitial space7 due to “plasma extravasation.” EX1022, 571;
`
`
`
`EX1048, Abstract; EX1050, Abstract; EX1049, 259; EX2339, 42:15-17; 57:12-15;
`
`62:13-22. But such “leakage” does not occur during migraine. EX2339, 34:18-
`
`35:6, 73:1-8, 13-17; EX2299, Abstract. And as Dr. Balthasar admitted, “antibodies
`
`can transport out of leaky vasculature more efficiently than non-leaky vasculature.”
`
`EX2339, 26:17-22; EX2345, 550-552. Thus, EX1048-1050 say nothing about
`
`IgG’s access to synaptic clefts during migraine. And, while Dr. Balthasar purports
`
`to disagree with Dr. Foord’s conclusion that C4.19 did not reach the synaptic cleft,
`
`he admitted that he was unqualified to do so. EX2339, 82:12-85:1. Thus, the record
`
`does not support Lilly’s allegation that C4.19 “clearly diffuses into the synaptic
`
`cleft.” Reply, 10.
`
`Faced with Tan’s C4.19’s failure to engage at the site of action in vivo,
`
`Lilly’s Petition speculated, without support, that “slow distribution to the site of
`
`immunoblockade could be overcome by (1) increasing the dose, (2) chronic
`
`administration, and/or (3) active immunization.” Petition, 42-43. Lilly now—
`
`belatedly—relies on Dr. Balthasar’s assertion that Covell’s “carcass” experiments
`
`
`7 The interstitial space is a space “between cells, within tissue.” EX2339,
`
`84:2-3. The synaptic cleft represents a different biological compartment within this
`
`space. EX2266, ¶90; POR, 37.
`
`- 19 -
`
`

`

`
`
`Case IPR2018-01711
`Patent No. 9,884,907
`
`show that “full-length antibodies were expected to distribute from general
`
`circulation into the interstitial space, so long as they are given sufficient time.”
`
`Reply, 10-11; EX1339, ¶¶24-29; EX1247, 3972. Covell does not support this
`
`conclusion.
`
`Covell presents no data relating to an antibody’s penetration of, or binding
`
`CGRP in, the synaptic cleft. EX2339, 93:19-95:6; 95:10-12; EX1022, 566;
`
`EX2266, ¶90; POR, 37. Further, the art shows that in a carcass, “assignment of a
`
`site, or sites, of antibody localization was not possible.” EX2279, 3045; EX2266,
`
`¶89; EX2339, 92:14-16. Dr. Balthasar admitted that movement of antibodies is a
`
`“random process,” and one needs to “consider a number of factors” in determining
`
`the amount of time required to achieve “concentration [] of interest” in the site of
`
`action. EX2337, 64:10-65:5. Lilly considered none of these factors. Thus, Lilly
`
`provides no credible evidence that IgG would distribute into the interstitial space—
`
`much less the synaptic cleft—with “longer distribution times and/or higher doses.”
`
`Reply, 11.
`
`Lilly attempts to discredit Dr. Foord’s explanation that C4.19 failed to enter
`
`the synaptic cleft because of steric constraints, arguing that the referenced
`
`antibodies were too large and the referenced synapses were too small. Reply, 11-
`
`12. But even if an antibody could enter the synaptic cleft based purely on the
`
`dimensions of each—which Teva does not concede—Lilly has not proven that the
`
`- 20 -
`
`

`

`Case IPR2018-01711
`Patent No. 9,884,907
`
`antibody would even reach the cleft in the first place. And to the extent that Lilly
`
`argues that “Teva followed Tan’s instructions in its patent examples” (Reply, 11,
`
`emphasis in original), Lilly has not demonstrated that Tan prompted the inventors
`
`to use humanized anti-CGRP antibodies to treat migraine; regardless, what Teva
`
`did is irrelevant