`
`
`
`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 IPR2018-01427
`Patent 8,597,649
`_____________________
`
`
`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 IPR2018-01427
`Patent No. 8,597,649
`
`
`B. 
`
`2. 
`
`3. 
`
`I. 
`II. 
`
`TABLE OF CONTENTS
`
`Introduction ...................................................................................................... 1 
`Lilly premised its case on the “therapeutic utility” of anti-CGRP
`antibodies; it should be held to this rationale. ................................................. 3 
`III.  Lilly’s new “potential therapeutic” rationale fails the obviousness
`inquiry. ............................................................................................................. 5 
`A. 
`The record does not demonstrate that a full-length anti-
`CGRP antibody would have been reasonably expected to be
`effective in vivo. .................................................................................... 6 
`Lilly failed to demonstrate that humanized anti-CGRP
`antibodies would have been safe in humans. ...................................... 10 
`1. 
`Lilly improperly ignores myriad teachings showing
`CGRP’s vasoprotective role. .................................................... 10 
`Risk concerns of antagonizing CGRP with long-acting
`antibodies were not “theoretical.” ............................................ 11 
`Any alleged safety of blocking the CGRP pathway
`with small molecules,
`receptor antagonists, or
`aptamers is irrelevant to the safety of anti-CGRP
`antibodies. ................................................................................ 13 
`Lilly improperly dismisses safety concerns with
`antagonizing CGRP in migraineurs. ........................................ 19 
`IV.  Lilly’s motivation arguments fail
`to consider pharmacological
`differences between receptor and ligand antagonism. ................................... 21 
`Lilly failed to prove that a POSA would have arrived at the claimed
`affinity (KD). .................................................................................................. 24 
`VI.  Lilly never articulated which prior art antibody a POSA would have
`humanized in order to arrive at the claimed antibodies. ................................ 24 
`VII.  Lilly misreads Teva’s secondary indicia arguments, and has not
`rebutted the presumption of nexus................................................................. 25 
`VIII.  Conclusion ..................................................................................................... 28 
`
`4. 
`
`V. 
`
`- i-
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`I.
`
`Introduction
`
`Teva’s Patent Owner Response (“POR”) exposed multiple infirmities that
`
`defeat Lilly’s obviousness case. Teva demonstrated that Lilly’s principal
`
`references—Tan 1995 and Wimalawansa—would not have motivated a POSA to
`
`develop an anti-CGRP antibody for human therapeutic use. Tan 1995, a basic
`
`research paper attempting to “prob[e] the role of CGRP as an endogenous
`
`vasodilator” in rats, reported that its full-length anti-CGRP antibody failed to show
`
`immunoblockade in a rat saphenous nerve assay. Wimalawansa—far from
`
`suggesting humanizing anti-CGRP antibodies, as Lilly argues—expressly conveys
`
`to a POSA that “[c]learly, more data from carefully designed studies are necessary
`
`before … humanized anti-CGRP monoclonal antibodies … can be evaluated as
`
`therapeutic agents.” And Lilly has not shown that such data existed prior to 2005.
`
`Unable to overcome these and other fatal defects, Lilly accuses Teva of
`
`“impermissibly reading safety and efficacy requirements into the claims.” But it
`
`was Lilly who premised its Petition on human therapeutic use. Teva rebutted
`
`Lilly’s arguments by, inter alia, showing that Lilly failed to consider safety, and
`
`fell short of demonstrating efficacy.
`
`Rather than considering the prior art as a whole, as it must, Lilly, through
`
`hindsight, selectively cherry-picks references Lilly believes support its arguments,
`
`while ignoring references that undermine them. Even worse, on cross-examination
`
`- 1 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`Dr. Charles and his replacement, Dr. Balthasar1, distanced themselves from
`
`unfavorable portions of Lilly’s own references upon which they themselves relied.
`
`Specifically, these experts refused to consider teachings that call into question the
`
`safety of long-term (as with an antibody) inhibition of CGRP, the body’s most
`
`potent vasodilator. Given that Lilly constructed its obviousness case on a
`
`
`
`therapeutic utility, the lack of safety and efficacy in its evidence cannot be ignored.
`
`On Reply, Lilly now pivots from its initial “human therapeutic use”
`
`arguments, focusing instead on the mere “potential” for therapeutic use. But Lilly
`
`cannot re-craft its challenge on Reply to attempt to rehabilitate its Petition. Henny
`
`Penny Corp. v. Frymaster LLC, No. 18-1596 (Fed. Cir. 2019).
`
`To support its new rationale, Lilly again cherry-picks isolated, out-of-
`
`context phrases to repaint the field as of 2005. For example, to support the alleged
`
`in vivo effectiveness of Tan’s C4.19 antibody, Lilly cites to Teva’s Dr. Ferrari. But
`
`Dr. Ferrari’s testimony referred to Tan’s Fab’ fragment, not full-length C4.19. And
`
`Lilly’s new arguments that carcass studies support speculation that Tan’s antibody
`
`would eventually reach its site of action given more time also fail because
`
`
`1 Teva discredited a number of Dr. Charles’ opinions, and showed him to be
`
`unqualified to offer them. POR, 3-4. On Reply, Lilly submitted the declaration of a
`
`new expert Dr. Balthasar in an effort to repair Dr. Charles’ failed opinions.
`
`- 2 -
`
`

`

`
`
`Case IPR2018-01427
`Patent No. 8,597,649
`
`“assignment of a site … of antibody localization was not possible.” 2
`
`Similarly, Lilly’s speculation about “increased dose” goes squarely 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. This is important because migraineurs were known
`
`to have increased risk of these life-threatening conditions.
`
`Lilly also fails to rebut Teva’s strong showing of numerous indicia of non-
`
`obviousness, which support confirming the challenged claims. In short, Teva’s
`
`Response demonstrates that Lilly’s Petition fails to show that the claimed
`
`humanized antibodies would have been obvious. Lilly’s Reply does not salvage its
`
`Petition.
`
`II. Lilly premised its case on the “therapeutic utility” of anti-CGRP
`antibodies; it should be held to this rationale.
`
`Lilly provided only one reason for developing a humanized anti-CGRP
`
`antibody: human therapeutic use. Petition, 13-14, 26-30, 32, 34-35. Lilly
`
`specifically argued the reason for humanizing a murine antibody was to retain “the
`
`antibody’s therapeutic utility” “in humans.” Id., 34-35. By doing so, Lilly (not
`
`Teva) read “safety and efficacy requirements into the claims.” Under a similar
`
`challenge to composition of matter claims, as here, the Board held the petitioner to
`
`
`2 Emphasis added throughout unless otherwise noted.
`
`- 3 -
`
`

`

`
`
`Case IPR2018-01427
`Patent No. 8,597,649
`
`its “therapeutic utility” motivation arguments. Phigenix v. ImmunoGen, IPR2014-
`
`00676, Paper 39, 16 (P.T.A.B. Oct. 27, 2017) (“when asserting that an ordinary
`
`artisan would have had a reason to … prepare [the claimed] immunoconjugate,
`
`Petitioner relies on the position that an ordinary artisan would have expected such
`
`an immunoconjugate to work clinically to treat tumors in humans upon reading
`
`the cited references”); see also id., 21 (“Petitioner does not persuade us … that a
`
`skilled artisan would have had a reasonable expectation of success … that [the
`
`claimed] immunoconjugate would be useful in the treatment of breast tumors in
`
`humans.”).
`
`As in Phigenix, having relied on therapeutic utility for motivation, Lilly was
`
`required to consider and demonstrate that the art evidenced efficacy and safety of
`
`the claimed antibodies for therapeutic use. Lilly recognized as much by arguing
`
`safety with respect to “immunogenicity” in humans in its reason to humanize.
`
`Petition, 35. But Lilly’s Petition stops there, saying nothing to address the myriad
`
`teachings in the art about cardiovascular and cerebrovascular safety concerns
`
`associated with long-term depletion of CGRP. Lilly also failed to prove reasonable
`
`expectation of efficacy.
`
`Recognizing gaps in its case that Teva highlighted, Lilly now seeks to retreat
`
`from its original therapeutic-based rationale. Lilly argues for the first time on
`
`Reply that a POSA would have been motivated to make “any … antibody having
`
`- 4 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`
`known functional properties.” Reply, 4 (emphasis in original). But Lilly’s rationale
`
`in the Petition controls; the Board should review the record through the lens of
`
`“therapeutic utility” of the claimed antibodies. 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.”). To the extent that Lilly is now trying to fill
`
`holes in its Petition with new art and testimony from a new expert, such belated
`
`arguments should be rejected. Id.
`
`III. Lilly’s new “potential therapeutic” rationale fails the obviousness
`inquiry.
`
`As with its original “therapeutic utility” rationale, Lilly’s new argument—
`
`motivation to humanize based on “the prospect of creating a ‘potential
`
`therapeutic’”—equally falls short of establishing obviousness. Reply, 5; citing
`
`Sanofi-Aventis U.S. LLC. v. Immunex Corp., IPR2017-01884, Paper 96, 17
`
`(P.T.A.B. Feb. 14, 2019). Sanofi-Aventis is inapposite because its mouse antibody
`
`MAb230 had properties the literature identified as being desirable characteristics in
`
`a therapeutic: a high affinity and the ability “to block both IL-4 and IL-13
`
`activity.” Id., 18. The field also identified targeting receptors—as the claimed
`
`antibody in that case did—as “more promising” than targeting their ligands
`
`(cytokines), and “several companies” were seeking to do just that. Id., 20. This
`
`utility led the Board to conclude that MAb230 was “a promising candidate from
`
`- 5 -
`
`

`

`
`
`Case IPR2018-01427
`Patent No. 8,597,649
`
`which to derive an effective therapeutic.”3 Id., 18. The record here reveals no such
`
`known utility in the art for Teva’s antibodies, and Lilly has not provided sufficient
`
`evidence to show a mouse anti-CGRP antibody to be a more “promising
`
`candidate” than the receptor antagonists the art was focused on.
`
`A. The record does not demonstrate that a full-length anti-CGRP
`antibody would have been reasonably expected to be effective in
`vivo.
`
`Nothing in the record demonstrates that full-length anti-CGRP antibodies
`
`would be expected to successfully achieve immunoblockade in vivo to suggest a
`
`“promising candidate” or an “effective therapeutic,” as in Sanofi-Aventis. As Dr.
`
`Foord explained, Tan’s C4.19 antibody failed to engage at the site of action in the
`
`in vivo saphenous nerve assay. EX2054, ¶102; EX1022, 571, 565, Abstract; POR,
`
`42. Lilly’s own Dr. Vasserot agreed with Teva’s experts on this shortcoming and
`
`the unreliability of Tan’s results. EX2191, 118:12-119:1; POR, 3. Dr. Charles—a
`
`clear outlier among the experts—testified that C4.19 “showed a 16% reduction in
`
`
`3 Also, in Sanofi-Aventis, Patent Owner stated, in response to an enablement
`
`decision: “There is no reasonable basis for concluding that antibodies that bind to
`
`‘the perfect target’ and inhibit ‘an important regulator’ would be therapeutically
`
`ineffective.” Sanofi-Aventis U.S. LLC. v. Immunex Corp. at 24. Here, no such
`
`admission has been made.
`
`- 6 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`skin blood flow” in a saphenous nerve assay. EX1002, ¶121; POR, 3. But given
`
`Dr. Charles’ gross mischaracterization of prior art in these proceedings (EX2192,
`
`182:21-183:12; 154:18-20), his opinion must be given little weight. POR, 3-4.
`
`In an attempt to find any support for in vivo effectiveness of Tan’s C4.19—
`
`
`
`and thus a reason to humanize it with a reasonable expectation of success—Lilly
`
`resorts to mischaracterizing Teva’s experts’ testimonies. For example, Lilly cites to
`
`Dr. Ferrari for his alleged admission of C4.19’s in vivo activity. Reply, 3. But Dr.
`
`Ferrari was referring to Tan’s Fab’ fragment, not the full-length antibody. EX1303,
`
`160:2-9, 162:12-22. Lilly then points to Dr. Tomlinson’s acknowledgment that he
`
`“humanized antibodies all the time” as evidence of motivation in 2005. Reply, 5.
`
`But Dr. Tomlinson was discussing his humanization activities from 2007 to 2016,
`
`not prior to 2005. EX1301, 55:1-13. Similarly disingenuous is Lilly’s allegation
`
`that “Teva’s experts conceded that a POSA would have found it appropriate to use
`
`humanized antibodies throughout drug development, including binding assays, in
`
`vitro testing, and animal studies.” Reply, 5-6. The transcripts illuminate the truth:
`
`
`
`Dr. Tomlinson stated that a POSA would have humanized an antibody
`
`to be tested only when “[g]iven sufficient motivation to do so ….
`
`which at the time wasn’t the case,” for CGRP. EX1301, 204:4-15;
`
`
`
`Dr. Ferrari admitted that he is “not an expert in [the drug
`
`development] field.” EX1303, 54:25-55:6;
`
`- 7 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`
`
`Dr. Rapoport’s testimony related to clinical trials in humans; not in
`
`vitro or animal testing. EX1304, 145:15-146:3.
`
`
`
`Ultimately, Lilly failed to prove that a POSA would have viewed Tan’s
`
`C4.19 as engaging at the site of action in vivo. Instead, 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, 45. Now that Teva has highlighted this speculation, Lilly
`
`belatedly relies on Dr. Balthasar’s assertion that “carcass” experiments show that
`
`“full-length antibodies were expected to distribute from general circulation into the
`
`interstitial space4, so long as they are given sufficient time.” Reply, 20; EX1329,
`
`¶¶24-29. But the art shows that in a carcass (a crude mix of leftover body parts),
`
`“assignment of a site, or sites, of antibody localization was not possible.” EX2270,
`
`3045; EX2273, 92:14-16. Thus, Lilly fails to show that a full-length antibody
`
`would distribute into interstitial spaces with additional time.
`
`Lilly also provides no credible evidence that, even if the antibody reaches
`
`the interstitial spaces, it would be able to access the synaptic cleft, especially in
`
`sufficient concentration for immunoblockade. It is undisputed that the “carcass”
`
`
`4 The interstitial space is a space “between cells, within tissue.” EX2273,
`
`84:2-3.
`
`- 8 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`experiment presents no data relating to an antibody’s penetration of the synaptic
`
`cleft or of an antibody’s binding to CGRP in the synaptic cleft. EX2273, 93:19-
`
`95:6; 95:10-12; EX1022, 566. Other references Lilly advances to fill this gap are
`
`equally uninformative.
`
`
`
`For example, Lilly cites to Tan 1995’s statement that a full-length antibody
`
`“clearly diffuses” to the site of action as evidence of its in vivo effectiveness.
`
`Reply, 20. But Tan 1995 based this conclusion on another study: Tan 1994.
`
`EX1022, 571. Tan 1994’s isolated vas deferens “tissue bath” is not an “equivalent”
`
`to an in vivo study, and thus is not a representation of an antibody’s ability to
`
`penetrate multiple biological compartments in a complex system, such as a human
`
`body. EX2273, 77:20-78:3; EX1021, 705. Similarly, in the Louis/Dockray
`
`experiments, the antibodies “leaked” into the interstitial space due to “plasma
`
`extravasation.” EX1022, 571; EX1048, Abstract; EX1050, Abstract; EX1049, 259;
`
`EX2273, 42:15-17; 57:12-15; 62:13-22. Such “leakage” does not occur in a
`
`neuromuscular junction or during migraine. EX2273, 34:18-35:6, 73:1-8, 13-17;
`
`EX2269, Abstract. And as Dr. Balthasar admitted, “antibodies can transport out of
`
`leaky vasculature more efficiently than non-leaky vasculature.” EX2273, 26:17-22;
`
`EX2268, 550-551. Thus, EX1021 and EX1048-1050 would have not been
`
`informative as to a full-length antibody’s ability to enter synaptic clefts during
`
`migraine. And, while Dr. Balthasar purports to disagree with Dr. Foord’s
`
`- 9 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`conclusion that C4.19 did not reach the synaptic cleft, he admitted that he was
`
`unqualified to provide such an opinion. EX2273, 82:12-83:1.
`
`None of Lilly’s cited art rehabilitates Tan 1995’s showing that the full-
`
`
`
`length antibody failed to immunoblockade, such that a POSA would have had a
`
`reasonable expectation of success. And Lilly has not explained why a POSA would
`
`have considered such an antibody as a “potential therapeutic.”
`
`B.
`
`Lilly failed to demonstrate that humanized anti-CGRP antibodies
`would have been safe in humans.
`
`Dr. Charles confirms: “[c]ertainly, we’re always concerned about safety
`
`issues, including stroke, in any circumstance where we’re administering therapy.”
`
`EX2272, 81:3-5. Yet neither the Petition nor the opening declarations (EX1002
`
`and EX1003) discussed the increased incidence of stroke in migraineurs or safety
`
`concerns with depleting CGRP’s vasoprotective role long-term, nor provided
`
`evidence of how a POSA would have overcome these safety concerns and
`
`proceeded to humanize an anti-CGRP antibody for human use. The Reply equally
`
`fails to fill this hole.
`
`1.
`
`Lilly improperly ignores myriad teachings showing CGRP’s
`vasoprotective role.
`
`Teva demonstrated the art-recognized importance of CGRP for vascular
`
`health. POR, 23-24. Lilly’s attempts to dismiss CGRP’s potent cerebro- and
`
`cardioprotective roles, as recognized by the field during the relevant time, is in
`
`- 10 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`error.
`
`
`
`As of 2005, the field believed that CGRP played a “pivotal role” in the
`
`“physiology and pathophysiology of cardiovascular regulation.” EX2003, 923;
`
`EX2141, ¶36; POR, 10. This role included protecting tissues during ischemia
`
`through, e.g., ischemic preconditioning. EX2272, 25:8-26:17, 29:16-30:25, 35:1-
`
`36:21; EX2263, 53-54; EX2264, 246.
`
`Even Lilly’s EX1284 demonstrates CGRP’s cardioprotective role: CGRP
`
`reduced infarct size in an ischemia rat model by up to 89%, while BIBN4096BS
`
`blocked “[t]he cardioprotective effect of CGRP.” EX1284, 591-592, Figure 3.
`
`Notably, Lilly omitted this unfavorable information, and its expert refused to even
`
`acknowledge it as “germane” during cross-examination. EX2272, 20:1-21:3.
`
`Importantly, the authors of EX1284 concluded that “CGRP is a very potent
`
`myocardial protective substance.” EX1284, 593; EX2272, 21:17-22:4.
`
`2.
`
`Risk concerns of antagonizing CGRP with long-acting
`antibodies were not “theoretical.”
`
`Concerns regarding CGRP ligand antagonism were not “theoretical,” as
`
`Lilly and its experts assert. Reply, 9-10. By 2005, researchers observed negative
`
`effects when administering anti-CGRP antibodies to animals. Tan 1995
`
`demonstrated that anti-CGRP antibodies increased baseline mean arterial pressure
`
`- 11 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`(MAP).5 EX1022, 568, Figure 2; EX2141, ¶51; EX2054, ¶78; POR, 25. Similarly,
`
`Tjen-A-Looi demonstrated that an anti-CGRP antibody increased pulmonary
`
`arterial pressure in rats. POR, 23; EX2084, H687; EX2054, ¶81. And Wong
`
`showed that an anti-CGRP antibody6 blocked CGRP-induced decrease in blood
`
`pressure and heart rate in rats. EX1033, 102, Table 2; EX2054, ¶81.
`
`
`
`Lilly does not contest that anti-CGRP antibodies increase blood pressure in
`
`animals. Reply, 11. Instead, Lilly attempts to undermine these in vivo examples by
`
`arguing that the blood pressure increases were “minor” and “transient” and would
`
`not have been a safety concern. Id. Lilly is wrong; according to Dr. Ferrari’s
`
`
`5 Lilly introduced a new exhibit—the Tan Thesis (EX1287)—on Reply, even
`
`though none of its experts rely upon it. Id., 11-12. Beyond untimeliness (see Paper
`
`43), Lilly has not established that EX1287 was publicly available or qualifies as
`
`prior art. In any event, it would not have prompted a POSA to ignore the greater
`
`weight of the prior art and the expectation that a full length anti-CGRP antibody
`
`would not be suitable for human therapeutic use.
`
`6 Lilly relies on Wong’s allusion to its antibody’s safety (Reply, 12), but Dr.
`
`Balthasar admitted that he could not “draw direct reference from the sentence
`
`[referring to MAP and heart rate] alone to which study it refers to.” EX2273,
`
`158:17-20.
`
`- 12 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`unrebutted testimony, “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; EX2141, ¶54. 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. EX2141, ¶46.
`
`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 argues that “more recent” art than that cited by Teva shows that no
`
`safety concern existed. Reply, 13. Lilly is incorrect. Lilly’s “more recent”
`
`references involve either BIBN4096BS or CGRP8-37 (see graphic below), neither of
`
`which is informative as to the safety of anti-CGRP antibodies, given their short
`
`half-lives and different mechanism of action (both are receptor antagonists). Reply,
`
`13-14; EX1283, Abstract; EX2152, 165; EX1284, Abstract; EX1285, Abstract;
`
`EX1318, Abstract; EX1263, Abstract.
`
`- 13 -
`
`

`

`Case IPR2018-01427
`
`Case IPR2018-01427
`Patent No. 8,597,649
`Patent No. 8,597,649
`
`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"
`
`- mama-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."
`
`
`
`
`
`
`- 14 -
`
`-14-
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`Lilly argues that “blocking the CGRP pathway7 was [] safe … for treating
`
`migraine.” Reply, 6. Devoid of any evidence of safety of anti-CGRP antibodies, all
`
`Lilly offers are references relating to small molecules (with “significantly” shorter
`
`half-lives than full length antibodies, as Dr. Charles admitted) and receptor
`
`antagonists. None of these molecules would have been probative of the safety of
`
`anti-CGRP antibodies.
`
`
`
`References discussing BIBN4096BS are not probative of the safety of an
`
`anti-CGRP antibody for multiple undisputed reasons. BIBN4096BS is a small-
`
`molecule receptor antagonist with a “significantly” shorter half-life than a full-
`
`length antibody. As such, any of its adverse effects would be cleared from the body
`
`within 24 hours after administration, in stark contrast to an antibody’s effects,
`
`which will linger for weeks. POR, 6, 22, 28-30; EX2272, 86:13-87:9.; EX2054,
`
`¶¶76, 84; EX2137, ¶¶50-55, 60-71, 74-76; EX2141, ¶22, 62, 68; EX1042, 645;
`
`Petition, 34; EX1059, 143, Figure 4.16; EX2272, 85:15-86:12. Moreover, each of
`
`Lilly’s BIBN4096BS references were short-term studies, assessing effects for 24
`
`hours or shorter, and would not have informed a POSA of the risks of an antibody
`
`
`7 Lilly artfully uses the phrase “CGRP pathway” or “CGRP antagonist” to
`
`disguise its reliance on art directed to receptor antagonists (Reply, 6, 9, 14, 24-25),
`
`whereas the challenged claims recite ligand antagonists. EX1001, 101-102.
`
`- 15 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`that circulates in the body for weeks, even after only a single administration. POR,
`
`30; EX2141, ¶22; EX2054, ¶84; EX1025, 1104; EX1042, 647; EX1290, 657;
`
`EX1297, S119; EX2272, 83:22-88:12.
`
`
`
`Further, to the extent any of Lilly’s cited studies tested cerebrovascular or
`
`cardiovascular safety of BIBN4096BS, they were all done in “healthy volunteers,”
`
`which, according to Dr. Ferrari, “are the wrong experiments to assess
`
`cardiovascular safety of any drug interacting with CGRP” because they do not
`
`study whether the antagonist would “block rescue mechanisms in the times of
`
`ischemia.” POR, 26-27; EX1303, 87:5-17, 89:21-90:3; EX2141, ¶¶48-59; EX2193,
`
`222; EX2157, 533; EX1025, 1108; EX1042, 647; EX2272, 83:22-84:13. What is
`
`more, Olesen specifically warned against relying on its study for cardiovascular
`
`safety: “our data base was too small for us to assess cardiovascular safety.”
`
`EX1025, 1109; POR, 29; EX2141, ¶22. Dr. Charles blatantly ignored this warning
`
`(EX1330, ¶34), but Dr. Balthasar confirmed that Olesen’s statement is consistent
`
`with “what I would expect from a clinical investigation” because one “can’t make
`
`extrapolations off of -- beyond the sample size that’s present.” EX2273, 119:5-
`
`120:4. Finally, it is of no consequence that an increase in MAP observed with
`
`CGRP8-37 “did not deter development of BIBN4096BS or other receptor
`
`antagonists.” Reply, 11. As described, a POSA would not have equated short-term
`
`receptor antagonism with long-term ligand antagonism. POR, 28-34.
`
`- 16 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`Lilly’s reliance on the reported safety of triptans is even less probative of
`
`
`
`anti-CGRP antibody’s safety. As with BIBN4096BS, triptans are small-molecule
`
`therapeutics with a half-life of only “about two hours.” EX1282, 7; EX2272, 99:1-
`
`4. Thus, any adverse effects of triptans are similarly eliminated from the body
`
`within hours. Triptans also operate by a mechanism that is very different from anti-
`
`CGRP antibodies: triptans agonize 5-HT receptors. EX1040, 180-181; EX2141,
`
`¶21; EX2272, 99:20-100:1. And researchers were aware of three different potential
`
`actions of triptans downstream of the 5-HT receptors, none of which were linked
`
`directly to CGRP. EX1040, 180. While researchers had observed that triptans
`
`could inhibit the release of CGRP, researchers were not discussing CGRP as being
`
`the source of the triptans’ anti-migraine activity before 2005. EX1303, 23:22-24:18
`
`EX1040, 180. Further, researchers were seeking to generate improved triptans that
`
`targeted 5-HT receptor subtypes so that they would lack vasoconstrictive
`
`activity—a goal that would not be compatible with inhibiting the activity of CGRP,
`
`a highly potent vasodilator. EX2267, ¶19; EX2272, 102:15-104:9.
`
`Dr. Charles analogizes antibody-induced BP increases to similar increases
`
`observed with triptans, arguing that such transient changes were not of concern.
`
`EX1330, ¶44. But triptans’ mechanism of action and pharmacokinetics differ from
`
`that of antibodies and Dr. Charles has not explained the basis for equating the two
`
`classes of molecules. Moreover, because triptans were used as a treatment for
`
`- 17 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`
`acute migraine, physicians would not have been concerned with triptans’ long-term
`
`effects. EX2141, ¶21; EX1031, 322; EX1040, 176. Even then, the
`
`“vasoconstrictive risk factors” and “the overall associated cardiovascular risks
`
`continue to limit their use.” EX2141, ¶21; EX2162, 443-444; EX1304, 55:1-57:25.
`
`Lilly has not explained why a POSA would have looked to triptans in making any
`
`conclusion as to the safety of anti-CGRP antibodies.
`
`Finally, Lilly’s Reply advances an entirely new theory based on “aptamers.”
`
`The Board should reject this argument as untimely. Henny Penny, No. 18-1596, 9;
`
`Paper 43, 4-5. Moreover, Lilly’s aptamer art does not suggest safety of the claimed
`
`anti-CGRP antibodies. First, as with BIBN4096BS and triptans, aptamers have a
`
`short half-life—“hours to days,” not “weeks”—and would not have been
`
`informative on the safety of long-acting antibodies. EX1309, Abstract; EX2272,
`
`114:6-115:5. Second, aptamers are not “analogs to antibodies,” as Lilly
`
`simplistically argues. Reply, 9. Aptamers “bridge the gap between small molecules
`
`and biologics” and have many characteristics that are similar to small-molecules.
`
`EX1309, Abstract. Further, there is no evidence on the record that aptamers would
`
`have been safe in humans. And the only in vivo study in the record with aptamers,
`
`Messlinger, is an abstract that lacks specifics such as controls used or experimental
`
`details for its mouse model. EX1240; EX2272, 115:19-117:7, 118:9-120:24. Dr.
`
`Charles’ blind reliance on Messlinger’s scarce disclosure merely because
`
`- 18 -
`
`

`

`
`
`Case IPR2018-01427
`Patent No. 8,597,649
`
`Messlinger was “widely known as an extremely careful scientist” once again
`
`demonstrates Dr. Charles’ bias in these proceedings. EX2272, 117:8-16, 121:12-
`
`122:6. For example, Dr. Charles “disagreed” with Susan Brain’s conclusion that
`
`CGRP played a cardioprotective role even though the Brain review provided
`
`substantially more evidence than the Messlinger abstract despite Dr. Charles’
`
`admission that Dr. Brain was “a respected person in the field.” EX2272 41:9-12,
`
`43:16-24; EX2003, 915.
`
`In sum, the record lacks any evidence that anti-CGRP antibodies would have
`
`been safe in humans. Lilly has not shown that a POSA would have found such
`
`antibodies “potentially therapeutic” in 2005.
`
`4.
`
`Lilly improperly dismisses safety concerns with
`antagonizing CGRP in migraineurs.
`
`Having failed in its Petition to consider migraine’s comorbidity with stroke
`
`and hypertension, Lilly now attempts to dismiss these concerns by arguing that
`
`they are “legally irrelevant” because “claims at issue do not distinguish between
`
`target patient populations.” Reply, 15. But it was Lilly—not Teva—who
`
`specifically focused on migraine in its Petition as part of its motivation analysis, as
`
`Dr. Charles admitted. Petition, 30-31; EX2192, 73:6-74:11. Thus, safety
`
`considerations of anti-CGRP antibodies in migraine patients are “legally relevant.”
`
`Lilly next argues that such considerations are irrelevant for most migraine
`
`- 19 -
`
`

`

`Case IPR2018-01427
`Patent No. 8,597,649
`
`patients because “the absolute risk” of stroke and myocardial ischemia in young
`
`women is “low.” Reply, 15. But this argument is nothing more than an attempt to
`
`detract from Lilly’s original failure to consider safety of the claimed antibodies in
`
`migraineurs, despite the known link between stroke and migraine. EX2272, 80:1-
`
`82:20; EX2266, 1123. In any event, Dr. Charles’ analysis on the risk to
`
`migraineurs is wrong.
`
`
`
`As Dr. Ferrari explains, it was known that CGRP played a key role in
`
`preventing escalation of mild ischemic events into full-blown infarctions, such as
`
`stroke and heart attacks. EX2141, ¶¶37-47; POR, 10. Dr. Charles attempts to
`
`minimize the importance of CGRP by pointing to the overall rate of stroke or
`
`myocardial infarction in young women w