`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`MYLAN PHARMACEUTICALS INC., CELLTRION, INC., and
`APOTEX, INC.
`Petitioners
`
`v.
`REGENERON PHARMACEUTICALS, INC.,
`Patent Owner
`
`
`Inter Partes Review No.: IPR2021-008811
`U.S. Patent No. 9,254,338 B2
`
`
`
`
`EXPERT DECLARATION OF DR. MARY E. GERRITSEN, Ph.D.
`IN SUPPORT OF PETITIONER’S REPLY
`
`
`
`
`1 IPR2022-00258 and IPR2022-00298 have been joined with this proceeding.
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 1
`
`
`
`TABLE OF CONTENTS
`
`Page
`
`I.
`
`QUALIFICATIONS AND BACKGROUND. ................................................ 1
`
`A.
`
`B.
`
`C.
`
`Education and Experience. .................................................................... 1
`
`Bases for Opinions and Materials Considered. ..................................... 1
`
`Scope of Work. ...................................................................................... 2
`
`II.
`
`LEGAL STANDARDS. .................................................................................. 2
`
`III.
`
`PERSON OF ORDINARY SKILL IN THE ART. ......................................... 2
`
`IV. THE PRIOR ART DISCLOSED VEGF TRAP-EYE/AFLIBERCEPT. ........ 3
`
`A.
`
`B.
`
`C.
`
`D.
`
`E.
`
`Regeneron Informed the Patent Office that VEGF Trap-Eye
`Was Encoded by the Sequences in the Prior Art. ................................. 3
`
`The Identity of Aflibercept/VEGF Trap-Eye Was Widely
`Reported and Well-Known. ................................................................. 15
`
`VEGF Trap-Eye Is Not a Genus. ........................................................ 23
`
`The Terms VEGF Trap-Eye and Aflibercept Referred to the
`Same Molecule. ................................................................................... 30
`
`Even if VEGF Trap-Eye and Aflibercept Were Presented as
`Different “Products,” POSAs Still Understood the Active
`Ingredients To Be the Same Molecule. ............................................... 35
`
`V. KNOWLEDGE OF THE AMINO ACID SEQUENCE OF VEGF
`TRAP-EYE WOULD RESULT IN A DRUG PRODUCT THAT
`WOULD PROVIDE TREATMENT USING THE CLAIMED
`DOSING REGIMEN. .................................................................................... 38
`
`VI. REGENERON’s PRIOR ART PATENTS. ................................................... 43
`
`
`
`i
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 2
`
`
`
`
`1. My name is Dr. Mary E. Gerritsen and I have been retained by counsel
`
`for Mylan Pharmaceuticals, Inc. (“Mylan” or “Petitioner”), to provide my opinions
`
`in support of Petitioner’s Reply. I am the same Dr. Gerritsen who wrote declarations
`
`in support of Mylan’s Petition for Inter Partes Review of U.S. Patent Nos. 9,669,069
`
`B2 (“’069 patent”) and 9,254,338 B2 (“’338 patent”), instituted as IPR2021-00880
`
`and IPR2021-00881, respectively. I also have been asked to reply to the opinions
`
`and views of Patent Owner’s declarants, Lucian V. Del Priore, M.D., Ph.D., and
`
`Alexander M. Klibanov, Ph.D.
`
`I.
`
`QUALIFICATIONS AND BACKGROUND.
`A. Education and Experience.
`
`2. My qualifications, education, and experience are set forth in my
`
`previous report, Ex.1003,2 and my curriculum vitae is attached as Exhibit 1061. I
`
`incorporate both as if set forth herein.
`
`B.
`
`3.
`
`Bases for Opinions and Materials Considered.
`
`In addition to my education, knowledge of the relevant published art,
`
`training, and experience, in forming the opinions I provide in this declaration, I have
`
`also considered the exhibits cited herein and in Exhibits 2048 and 2049.
`
`
`2 Unless otherwise noted, all citations to exhibits refer to exhibits filed in IPR2021-
`00881.
`
`1
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 3
`
`
`
`Scope of Work.
`
`
`
`I have been retained by Petitioner as an expert in this matter to provide
`
`C.
`
`4.
`
`various opinions regarding the ’069 patent (IPR2021-00880 Ex.1001) and ’338
`
`patent (IPR2021-00881 Ex.1001). I receive $350 per hour for my services. No part
`
`of my compensation is dependent upon my opinions given or the outcome of this
`
`case. I have no financial interest in this matter.
`
`II. LEGAL STANDARDS.
`
`5.
`
`For my opinions in this declaration, I understand that it requires
`
`applying various legal principles. As I am not an attorney, I have been informed
`
`about various legal principles that govern my analysis. I have used my
`
`understanding of those principles in forming my opinions. I summarized my
`
`understanding of those legal principles in my previous report, Exhibit 1003, and I
`
`incorporate that understanding as if set forth herein.
`
`III. PERSON OF ORDINARY SKILL IN THE ART.
`
`6.
`
`It is my opinion that a person of ordinary skill in the art (“POSA”)
`
`would have: (1) knowledge regarding the diagnosis and treatment of angiogenic eye
`
`disorders, including the administration of therapies to treat said disorders; and (2) the
`
`ability to understand results and findings presented or published by others in the
`
`field, including the publications discussed herein. Typically, such a person would
`
`2
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 4
`
`
`
`
`have an advanced degree, such as an M.D. or Ph.D. (or equivalent, or less education
`
`but considerable professional experience in the medical, biotechnological, or
`
`pharmaceutical field), with practical academic or medical experience in: (i)
`
`developing treatments for angiogenic eye disorders, such as age-related macular
`
`degeneration (“AMD”), including through the use of VEGF antagonists, or (ii)
`
`treating of same, including through the use of VEGF antagonists.
`
`7.
`
`I understand that Patent Owner contends that the skilled artisan is an
`
`ophthalmologist with experience in treating angiogenic eye disorders, including
`
`through the use of VEGF antagonists.
`
`8. My opinions set forth in this declaration remain the same under either
`
`Patent Owner’s or Petitioner’s definition.
`
`IV. THE PRIOR ART DISCLOSED VEGF TRAP-EYE/AFLIBERCEPT.
`A. Regeneron Informed the Patent Office that VEGF Trap-Eye Was
`Encoded by the Sequences in the Prior Art.
`
`9.
`
`I have reviewed the declarations of Drs. Klibanov and Del Priore.
`
`(Ex.2049, Klibanov Decl.; Ex.2048, Del Priore Decl.). They provide opinions that
`
`amount to speculation about what a POSA could have believed about VEGF Trap-
`
`Eye and aflibercept. I have also reviewed Regeneron’s Patent Owner Response,
`
`where they too present arguments that the sequence of VEGF Trap-Eye could have
`
`been uncertain to a POSA. (Paper 38, Patent Owner Response, 24-35). I disagree.
`
`3
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 5
`
`
`
`
`I have reviewed statements that Regeneron has made to the U.S. Patent
`
`10.
`
`and Trademark Office (“USPTO”) in connection with Regeneron’s efforts to obtain
`
`patent term extension for some of their prior art patents. In those applications for
`
`patent term extension, Regeneron represented that multiple prior art Regeneron
`
`patents disclosed the exact sequence of VEGF Trap-Eye/aflibercept. Based on those
`
`statements made by Regeneron, and based upon what a POSA would have been able
`
`to discern from those prior art patents, I disagree with Regeneron and the opinions
`
`of its experts.
`
`11.
`
`In my opinion, the representations by Regeneron to the USPTO leave
`
`no doubt that the VEGF Trap-Eye/aflibercept sequences were set forth in the prior
`
`art, and set forth in the claims of the ’338 patent and ’069 patent. In light of this
`
`clear disclosure, I question the relevance of Dr. Klibanov’s and Dr. Del Priore’s
`
`additional speculation about what a POSA could have believed about the structure
`
`and sequence of VEGF Trap-Eye.
`
`12. For example, I have reviewed Regeneron’s submissions to the USPTO
`
`with respect to U.S. Patent No. 7,374,758 (“’758 patent”) and U.S. Patent No.
`
`7,070,959 (“’959 patent”), including Regeneron’s Applications for Extension of
`
`Patent Term Under 35 U.S.C. § 156. (Ex.1024, ’758 FH, 12/22/2011 PTE; Ex.1102,
`
`’959 FH, 12/22/2011 PTE).
`
`4
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 6
`
`
`
`
`13. Therein, Regeneron explains that the application is being submitted
`
`“for extension of the term of United States Letters Patent No. 7,374,758.” (Ex.1024,
`
`’758 FH, 12/22/2011 PTE, 1; see also Ex.1102, ’959 FH, 12/22/2011 PTE, 1). The
`
`’758 patent and ’959 patent have also been submitted as exhibits in this IPR. (See,
`
`e.g., Ex.1010, ’758 patent; Ex.1023, ’959 patent). The ’758 patent application was
`
`filed December 17, 2004, and the patent issued May 20, 2008. The ’959 patent
`
`application was filed May 23, 2000, and the patent issued July 4, 2006. I understand
`
`that both the ’758 patent and ’959 patent claim priority to an application filed on
`
`June 8, 1999.
`
`14. Regeneron also states that “[t]he name of the approved product is
`
`EYLEA™. The name of the active ingredient of EYLEA™ is aflibercept, also
`
`known as VEGF trap, VEGF-trap, VEGF Trap-Eye and VEGF-TRAPR1R2.”
`
`(Ex.1024, ’758 FH, 12/22/2011 PTE, 2 (emphasis added); Ex.1102, ’959 FH,
`
`12/22/2011 PTE, 2 (same)). Regeneron further explains that “[a]flibercept is a
`
`fusion protein consisting of (a) a vascular endothelial growth factor (VEGF) receptor
`
`component having immunoglobulin-like (Ig) domains consisting of an Ig domain 2
`
`of a first VEGF receptor that is human Flt1 and an Ig domain 3 of a second VEGF
`
`receptor that is human Flk1; and (b) an Fc portion of human IgG1.” (Ex.1024, ’758
`
`FH, 12/22/2011 PTE, 2; Ex.1102, ’959 FH, 12/22/2011 PTE, 2 (same)).
`
`5
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 7
`
`
`
`
`15. Regeneron also lists the inventors of the ’758 patent and ’959 patent,
`
`one of which, George D. Yancopoulos, is the lone inventor listed on the face of the
`
`’338 patent and ’069 patent. (Ex.1024, ’758 FH, 12/22/2011 PTE, 3; Ex.1102, ’959
`
`FH, 12/22/2011 PTE, 3; see also, e.g., IPR2021-00880 Ex.1001, ’338 patent;
`
`Ex.1010, ’758 patent; Ex.1023, ’959 patent).
`
`16. The ’758 patent and ’959 patent PTE applications also include a section
`
`titled “Statement Regarding Patent Claims Relative to Approved Product
`
`[§ 1.740(a)(9)].” (Ex.1024, ’758 FH, 12/22/2011 PTE, 4; Ex.1102, ’959 FH,
`
`12/22/2011 PTE, 4).
`
`17. Regeneron states that “[a]t least the following claims of the ’758 patent
`
`claim a method of using the approved product: claims 1 and 2.” (Ex.1024, ’758 FH,
`
`12/22/2011 PTE, 4). Claims 1 and 2 of the ’758 patent read as follows:
`
`6
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 8
`
`
`
`
`
`
`
`(Ex.1010, ’758 patent, 97:10-28).
`
`18.
`
`In the ’959 patent PTE Application, Regeneron states that “[a]t least the
`
`following claim of the ’959 patent claims a method of manufacturing the approved
`
`product: claim 11.” (Ex.1102, ’959 FH, 12/22/2011 PTE, 4). Claim 11, and the
`
`claim that it incorporates, claim 8, from the ’959 patent, are set forth below:
`
`7
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 9
`
`
`
`
`
`
`
`(Ex.1023, ’959 patent, 40:5-23).
`
`19. Regeneron also provided a description of the approved product. This
`
`included a recitation of portions of the EYLEA™ label. (Ex.1024, ’758 FH,
`
`12/22/2011 PTE, 5; Ex.1102, ’959 FH, 12/22/2011 PTE, 5). It also included the
`
`following statement:
`
`Aflibercept is also described in Holash et al. Proc. Natl. Acad. Sci.
`USA, August 20, 2002, Vol. 99, No. 17, pp. 11393-11398 (“Holash,”
`Attachment G) as VEGF-TrapR1R2, which has the Ig domain 2 of VEGF
`receptor 1 (VEGFR1; also known as Flt-1) fused to the Ig domain 3 of
`VEGF receptor 2 (VEGFR2; also known as Flk-1), which in turn is
`fused to the constant region (Fc) of human IgG1. See paragraph
`
`8
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 10
`
`
`
`bridging pages 11393 and 11394 and Figure 1A. Moreover, Holash et
`al. demonstrate that aflibercept is a VEGF antagonist that binds to and
`inhibits the biologic activity of human vascular endothelial growth
`factor (VEGF) in various in vitro and in vivo assay systems.
`
`
`
`
`(Ex.1024, ’758 FH, 12/22/2011 PTE, 5 (emphasis added); see also Ex.1102, ’959
`
`FH, 12/22/2011 PTE, 5). Thus, Regeneron has provided a statement to the USPTO
`
`that aflibercept is the VEGF-TrapR1R2 molecule described in Holash. As I discuss
`
`below, this is consistent with my review of Holash, as well as with Regeneron’s
`
`other publications, including those published before the filing date of the ’338 patent
`
`and ’069 patent.
`
`20.
`
`In the ’959 patent PTE Application, Regeneron also states that “Holash
`
`(Attachment G) also describes the method of producing aflibercept (VEGF-
`
`TrapR1R2) as expressing a recombinant DNA construct in Chinese hamster ovary
`
`cells (See ‘Engineering VEGF-Traps’ in the Materials and Methods section on page
`
`11393-11394.” (Ex.1102, ’959 FH, 12/22/2011 PTE, 5).
`
`21.
`
`In the ’758 patent PTE application, Regeneron also provided a detailed
`
`account, like the one in Holash, of how the molecule was constructed and which
`
`domains were included:
`
`Aflibercept includes an Ig domain 2 from VEGFR1 fused to an Ig
`domain 3 from VEGFR2 as described in Holash (See Attachment G,
`
`9
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 11
`
`
`
`
`
`paragraph bridging pages 11393 and 11394 and Figure 1A). The ’758
`patent refers to VEGFR-1 as Flt1 in column 25, line 57 and refers to
`VEGFR-2 as Flk1 in column 26, line 12. In addition, Holash discloses
`that VEGFR-1 is also known as Flt-1 and VEGFR-2 is also known as
`Flk-1. See Figure 1A of Holash. Thus, aflibercept has a VEGF receptor
`component consisting of an Ig domain 2 of a first VEGF receptor
`human Flt-1 and an Ig domain 3 of a second VEGF receptor human
`Flk-1 as defined in claim 1.
`
`
`(Ex.1024, ’758 FH, 12/22/2011 PTE, 6; see also Ex.1102, ’959 FH, 12/22/2011 PTE,
`
`6).
`
`22.
`
`In the ’758 patent PTE Application, Regeneron further states that “[t]he
`
`disclosure of the Flt1 and Flk1 components in the approved product and the
`
`construction of the expression vector used in making the active ingredient in the
`
`approved product is discussed in the ’758 patent in Example 20, column 29, lines
`
`41-56.” (Ex.1024, ’758 FH, 12/22/2011 PTE, 6).
`
`23. Regeneron further states that “the amino acid sequence of both the
`
`Flt1 and Flk1 components of the approved product are disclosed in Figures 24A-
`
`24C. Flt1 Ig domain 2 spans amino acid residues 27 through 129 and Flk1 Ig
`
`domain 3 spans amino acid residues 130 through 231 of the fusion protein.”
`
`(Ex.1024, ’758 FH, 12/22/2011 PTE, 6 (emphasis added)).
`
`10
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 12
`
`
`
`
`24. Regeneron continues, describing the Fc domain portion of aflibercept,
`
`and stating that “[t]he disclosure of the Fc multimerizing component in the actual
`
`product is discussed in Example 20, column 29, lines 41-56, and its amino acid
`
`sequence is disclosed in Figures 24A-24C, from amino acid residue 232 through
`
`458.” (Ex.1024, ’758 FH, 12/22/2011 PTE, 7 (emphasis added)).
`
`25.
`
`In the ’959 patent PTE Application, Regeneron also states that:
`
`(Ex.1102, ’959 FH, 12/22/2011 PTE, 6 (emphasis added)).
`
`26.
`
`In the ’959 patent PTE Application, Regeneron continues, explaining
`
`that:
`
`
`
`11
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 13
`
`
`
`
`
`
`
`
`(Ex.1102, ’959 FH, 12/22/2011 PTE, 6-7 (emphasis added)).
`
`
`27. Regeneron continues, stating that “[t]he complete DNA and deduced
`
`amino acid sequences of the pVEGFR1R2FcΔC1(a) chimeric molecule is set forth
`
`in FIGS. 24A-24C.” (Ex.1102, ’959 FH, 12/22/2011 PTE, 7).
`
`28. Regeneron continues, stating that:
`
`12
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 14
`
`
`
`Thefusion polypeptide as describedin the ‘959 patent was
`produced utilizing the cell culture system described in Example 21,
`column 29,lines 31-67 through column 30,lines 1-17. More
`particularly, suspension cultures of recombinant Chinese hamster ovary
`(CHO K1/E1A)cells, which constitutively express the fusion
`polypeptide, were used to manufacturethe fusion polypeptide. The
`cells were grownin bioreactors and the protein productwasisolated
`and purified by affinity and size exclusion chromatography. The process
`is providedin greaterdetail in column 29,lines 48-58. The fusion
`polypeptide was harvested from the bioreactor using the process
`described in Example 22.
`
`(Ex.1102, °959 FH, 12/22/2011 PTE,7).
`
`
`
`29. [have preparedatable in which I provide a side-by-side comparison of
`
`(1) the disclosures of the ’758 patent, Figs. 24A-24C, which Regeneronhasstated is
`
`the sequence of VEGF Trap-Eye/aflibercept, and (2) the sequences claimed in the
`
`°338 patent, claim 1:
`
`
`
`Regeneron’s PTE|’338 patent claim 1°758 patent
`
`
`
`
`
`spans amino acid residues 27 through|component comprising amino acids 27
`
`application
`
`
`
`“Aflibercept includes an Ig domain 2|“wherein the VEGF antagonist is a
`
`
`
`
`
`from VEGFR1 . . .. VEGFR-1 is also|VEGF receptor-based chimeric
`
`
`
`known as Fit-1 .
`
`.
`
`.
`
`. Flt] Ig domain 2} molecule comprising (1) a VEGFRI1
`
`13
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 15
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 15
`
`
`
`129 [of Figures 24A-24C.]” (Ex.1024,
`
`to 129 of SEQ ID NO:2.” (IPR2021-
`
`
`
`’758 FH, 12/22/2011 PTE, 6).
`
`00880 Ex.1001, 23:12-14).
`
`“Aflibercept includes . . . an Ig domain
`
`“wherein the VEGF antagonist is a
`
`3 from VEGFR2 . . . . VEGFR-2 is also
`
`VEGF
`
`receptor-based
`
`chimeric
`
`known as Flk-1 . . . . Flk1 Ig domain 3
`
`molecule comprising . . . (2) a VEGFR2
`
`spans amino acid residues 130 through
`
`component comprising amino acids
`
`231 [of Figures 24A-24C.]” (Ex.1024,
`
`130-231 of SEQ ID NO:2.” (IPR2021-
`
`’758 FH, 12/22/2011 PTE, 6).
`
`00880 Ex.1001, 23:12-16).
`
`“The disclosure of the Fc multimerizing
`
`“a
`
`multimerization
`
`component
`
`component in the actual product is
`
`comprising amino acids 232-457[ ]of
`
`discussed in Example 20, column 29,
`
`SEQ ID NO:2.”
`
` (IPR2021-00880
`
`lines 41-56, and its amino acid sequence
`
`Ex.1001, 23:16-18).
`
`is disclosed in Figures 24A-24C, from
`
`amino acid residue 232 through 458.”
`
`(Ex.1024, ’758 FH, 12/22/2011 PTE, 7).
`
`
`
`30. The ’959 patent PTE Application contains similar language to that
`
`noted above for the ’758 patent PTE Application, amounting to the same disclosure
`
`(Ex.1102, ’959 FH, 12/22/2011 PTE, 6-7), leaving no doubt that the prior art
`
`14
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 16
`
`
`
`
`sequences set forth in the ’758 patent and ’959 patent are identical to that set forth
`
`in claims 1 and 14 of the ’338 patent.
`
`31.
`
`I also note that the relevant claim language for the ’069 patent claim 1
`
`is identical, as far as I can tell. Therefore, the above analysis also applies to claim 1
`
`(and claim 12, which sets forth the nucleic acid sequence) of the ’069 patent.
`
`32. Lastly, I have reviewed exhibits that include sequence alignments
`
`comparing the sequences from the prior art ’758 patent with the amino acid and
`
`nucleotide sequences claimed in the ’338 patent and ’069 patent. (IPR2021-00880
`
`Ex.1082 and Ex.1083; IPR2021-00881 Ex.1093 and Ex.1094). I have confirmed
`
`that the sequences are the same.
`
`B.
`
`The Identity of Aflibercept/VEGF Trap-Eye Was Widely Reported
`and Well-Known.
`
`33.
`
`In my opinion, the above disclosures by Regeneron that the sequences
`
`set forth in at least the ’758 patent and ’959 patent are the sequences for VEGF Trap-
`
`Eye/aflibercept, and the alignments showing that those sequences 100% align with
`
`the sequences claimed in the ’338 patent and ’069 patent, are enough to show that
`
`the sequences are an inherent aspect of the prior art disclosures of VEGF Trap-Eye.
`
`I also understand that Regeneron’s expert conceded that “VEGF Trap-Eye has the
`
`same amino acid sequence as aflibercept.” (Ex.2049, Klibanov Decl., ¶48).
`
`However, I understand that Regeneron’s experts Dr. Klibanov and Dr. Del Priore
`15
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 17
`
`
`
`
`have provided opinions that Petitioner’s prior art does not disclose the amino acid
`
`sequence of VEGF Trap-Eye/aflibercept; which amino acid residues comprise the
`
`VEGF receptor-1 or -2 domains; which amino acid residues comprise the Fc domain;
`
`and whether or not VEGF Trap-Eye has the same amino acid sequence as aflibercept.
`
`(Ex.2049, Klibanov Decl., ¶41; Ex.2048, Del Priore Decl., ¶46).
`
`34.
`
`I disagree with Drs. Klibanov and Del Priore for the following reasons.
`
`35. First, while Petitioner’s cited references do not include a verbatim
`
`recitation of the amino acid sequence, it is clear from a review of the prior art that
`
`the identity of VEGF Trap-Eye/aflibercept was known and widely discussed in the
`
`art, including by Regeneron, and thus was part of the foundation of knowledge of a
`
`POSA. Dr. Del Priore admitted that his statement regarding the public availability
`
`of the VEGF Trap-Eye/aflibercept sequence was based only on his own experience
`
`in learning the sequence. (Ex.2048, Del Priore Decl., ¶65; Ex.1111, Del Priore Dep.
`
`Tr, 89:8-19).
`
`36. The structure and sequence of the VEGF Trap-Eye/aflibercept
`
`molecule set forth in each of Petitioner’s asserted references can be traced back to
`
`Regeneron’s original PNAS article (Ex.1004, Holash) and its early patent
`
`applications and patents (Ex.1010, Figs. 24A-C; Ex.1023, Figs. 24A-C), as well as
`
`16
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 18
`
`
`
`
`the public disclosure of the aflibercept sequence (Ex.1040, WHO Drug Info, 118-
`
`19).
`
`37. For example, in 2002 Regeneron published in PNAS its description of
`
`the molecule that Regeneron would proceed to bring into cancer and AMD clinical
`
`trials. (Ex.1004, Holash). This is the same Holash article that Regeneron
`
`represented to the Patent Office discloses the construction of VEGF Trap-
`
`Eye/aflibercept, known in Holash as VEGF TrapR1R2. (See above; see also, e.g.,
`
`Ex.1024, ’758 FH, 12/22/2011 PTE, 5; Ex.1102, ’959 FH, 12/22/2011 PTE, 5).
`
`38. Holash describes in detail the construction of VEGFR1R2 (i.e., VEGF
`
`Trap-Eye/aflibercept), including in the section of Holash spanning 11393 to 11394.
`
`Briefly, it describes the parental VEGF-Trap which contained the first three Ig
`
`domains of VEGFR1. (Ex.1004, Holash, 11393). Here is the portion of Fig. 1
`
`depicting the native VEGF receptor 1 (Flt-1) and the parental VEGF-Trap:
`
`(Ex.1004, Holash, 11394). The Regeneron authors then state that “VEGF-TrapΔB1
`
`was created by removing a highly basic 10-aa stretch from the third Ig domain” and
`
`
`
`17
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 19
`
`
`
`“VEGF-TrapΔB2 was created by removing the entire first Ig domain from VEGF-
`
`
`
`Trap ΔB1.” (Ex.1004, Holash, 11393):
`
`
`
`(Ex.1004, Holash, 11394). Finally, the authors state that “VEGF-TrapR1R2 was
`
`created by fusing the second Ig domain of VEGFR1 with the third Ig domain of
`
`VEGFR2.” (Ex.1004, Holash, 11393-94). Figure 1 also depicts the native VEGFR2
`
`(Flk-1), and the configuration of VEGF-TrapR1R2:
`
`
`
`(Ex.1004, Holash, 11394; see also, e.g., Fig. 1 caption (“VEGF-TrapR1R2 possesses
`
`the second Ig domain of VEGFR1 and the third Ig domain of VEGFR2 fused to the
`
`Fc portion of human IgG1.”)). The authors provide another description of the
`
`construction of VEGF-TrapR1R2 later in the article, with detailed justifications for
`
`each step of the engineering of the molecule. (Ex.1004, Holash, 11394-95
`
`(paragraph spanning pp. 11394 to 11395)). Dr. Del Priore was unable to identify
`
`18
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 20
`
`
`
`
`any other protein in Holash that contains the second Ig domain of VEGFR1 and the
`
`third Ig domain of VEGFR2 fused to the Fc portion of IgG1. (Ex.1111, Del Priore
`
`Dep. Tr., 84:8-87:19).
`
`39. Holash proceeds to characterize VEGFR1R2 with respect to its in vitro
`
`binding and cell proliferation activity (11395); its VEGF blockade activity (11395-
`
`96); its ability to block tumor growth in vivo (11396); and its comparison to anti-
`
`VEGF antibodies in in vivo melanoma models (11397).
`
`40.
`
`In conclusion the authors state that “[h]erein we describe the
`
`engineering of an anti-VEGF agent, termed VEGF-TrapR1R2,” which was
`
`“engineered to have minimal interactions with extracellular matrix, and this property
`
`apparently accounts for its satisfying pharmacokinetic profile.” (Ex.1004, Holash,
`
`11397). The authors further state that “[i]n this manuscript, we compare the efficacy
`
`of the VEGF-Trap to that of a monoclonal antibody to VEGFR2 in cancer models
`
`and find that far lower circulating levels of VEGF-TrapR1R2 are required for similar
`
`efficacy. Tumors treated with highest doses of the VEGF-Trap are not only stunted
`
`but also strikingly avascular.” (Ex.1004, Holash, 11397). The authors conclude that
`
`the “superior blocking and pharmacologic properties” of VEGF-TrapR1R2 “demand
`
`that
`
`it be
`
`tested
`
`in human patients suffering from diseases
`
`involving
`
`neoangiogenesis,” and that the molecule had recently undergone toxicology testing
`
`19
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 21
`
`
`
`
`in monkeys and “is currently in human clinical trials for several different types of
`
`cancer.” (Ex.1004, Holash, 11397).
`
`41.
`
`In my opinion, and contrary to the opinions of Drs. Klibanov and Del
`
`Priore, it is clear from these disclosures that in the Holash article, Regeneron was
`
`describing not only the construction of the molecule that would later be known as
`
`VEGF Trap-Eye and aflibercept (among other monikers), but also its introduction
`
`into clinical trials. Accordingly, I disagree with the suggestion by Drs. Klibanov
`
`and Del Priore that a POSA would have considered VEGF Trap-Eye to be a genus
`
`of proteins or to comprise any of the other VEGF-Trap precursors described in
`
`Holash. In my opinion, a POSA would have concluded that Regeneron was
`
`describing in its Holash publication the development of a single agent to be used as
`
`an anti-angiogenic VEGF blocker. I understand that Dr. Del Priore, in his
`
`deposition, agreed that VEGF Trap-Eye was “one of Regeneron’s family of
`
`proteins.” (Ex.1111, Del Priore Dep. Tr., 77:12-13).
`
`42. My opinion is further supported by other publications, including those
`
`authored by George Yancopoulos, the named inventor on the ’338 patent and ’069
`
`patent, as well as other Regeneron-associated authors.
`
`20
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 22
`
`
`
`
` For example, George Yancopoulos and others authored a book chapter
`
`43.
`
`entitled “Clinical Development of VEGF Trap.” (Ex.1113, Del Priore Dep. Ex. 10,
`
`Rudge 2008, 415).
`
`44. The Regeneron authors begin by noting efforts over the years at
`
`developing anti-VEGF agents, and stating that “VEGF Trap is one such protein-
`
`based agent that has been engineered to bind and sequester VEGF, as well as
`
`placental growth factor (PIGF), with high affinity.” (Ex.1113, Del Priore Dep. Ex.
`
`10, Rudge 2008, 415 (emphasis added)). The Regeneron authors continue:
`
`VEGF Trap has been shown to effectively inhibit pathological
`
`angiogenesis in numerous preclinical models of cancer and eye
`
`disease, and is now being evaluated in clinical trials in several types
`
`of cancer, as well as the ‘wet’ or neovascular form of age-related
`
`macular degeneration (AMD).
`
`(Ex.1113, Del Priore Dep. Ex. 10, Rudge 2008, 415 (emphasis added)). I read this,
`
`and in my opinion a POSA would have read this, as referring to a single agent, and
`
`that that same molecular agent was currently in clinical trials for both cancer and eye
`
`diseases (e.g., AMD).
`
`45. The Regeneron authors continue, explaining that “[t]he VEGF Trap is
`
`a soluble chimeric receptor in which key domains of VEGFR1 (domain 2) and
`
`21
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 23
`
`
`
`
`VEGFR2 (domain 3) are fused to the constant region (Fc portion) of human IgG1.”
`
`(Ex.1113, Del Priore Dep. Ex. 10, Rudge 2008, 416). For this statement, the authors
`
`cite Holash, which I discuss above. In Holash, the only molecule that matches the
`
`description in Rudge is VEGF TrapR1R2. (Ex.1004, Holash, 11394-95; Fig. 1). From
`
`this statement and from Holash itself, it is clear that Regeneron had a single agent
`
`that it was testing in both cancer and angiogenic eye disorders. I have not seen any
`
`publications or testimony from Regeneron’s experts showing, or even suggesting,
`
`that the terms aflibercept and VEGF Trap-Eye indicated a difference in, modification
`
`to, or truncation of, the amino acid sequence of the molecule.
`
`46.
`
`Indeed, Regeneron’s experts admit that they have not cited any
`
`publications showing that the use of the different terms VEGF Trap-Eye and
`
`aflibercept indicated an actual difference in the sequence or structure of the
`
`molecule, or that VEGF Trap-Eye had a different sequence than that set forth in the
`
`Challenged Claims. (Ex.1111, Del Priore Dep. Tr., 95:14-18).
`
`47. While Regeneron’s experts speculate that a POSA could believe VEGF
`
`Trap-Eye and aflibercept to have been different molecules, I have seen no evidence
`
`from the experts that any POSA actually believed this.
`
`48. Rather, the literature following Holash contained numerous articles
`
`discussing both aflibercept and VEGF Trap-Eye, all citing back to Holash and its
`
`22
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 24
`
`
`
`
`disclosure of VEGF TrapR1R2. (See, e.g., Ex.1075, Ni, 408 (citing to Wulff
`
`(Ex.1129), which disclosed preclinical studies of “VEGF TrapR1R2,” the same
`
`nomenclature used in contemporaneously-published Holash); Ex.1119, Klibanov
`
`Dep. Ex. 12, Rudge 2007, 18363, 18370 (discussing VEGF Trap, including
`
`“aflibercept” as a keyword, and citing back to Holash (ref. 20)); Ex.2080, Heier
`
`2009, 2 (discussing VEGF Trap-Eye and citing back to Holash and discussing the
`
`data presented in Holash for VEGF TrapR1R2); Ex.1120, Klibanov Dep. Ex. 13,
`
`Gomez-Manzano, 940, 945 (“a new anti-VEGF agent, VEGF Trap/aflibercept
`
`(henceforth referred to as VEGF Trap), has been developed by incorporating
`
`domains of both VEGF receptor 1 (VEGFR-1) and VEGFR-2 fused to the constant
`
`region of human immunoglobulin G1,” and citing Holash); Ex.1123, Quiram, 1009-
`
`10 (discussing VEGF Trap-Eye and its structure, and citing back to Holash)).
`
`C. VEGF Trap-Eye Is Not a Genus.
`
`49. Regeneron’s experts have sought to cast doubt on the identity of the
`
`VEGF Trap-Eye/aflibercept molecule by speculating that a POSA would have
`
`believed VEGF Trap-Eye to refer to a genus of proteins. (Ex.2049, Klibanov Decl.,
`
`¶¶67-80; Ex.2048, Del Priore Decl., ¶¶87-91). I firmly disagree.
`
`50. First, Dr. Klibanov states in his declaration that a POSA may have
`
`believed that “VEGF Trap-Eye . . . may include multiple amino acid sequences.”
`
`23
`
`Mylan Exhibit 1115
`Mylan v. Regeneron, IPR2021-00881
`Page 25
`
`
`
`
`(Ex.2049, Klibanov Decl., ¶80). However, by January 2011 Regeneron had
`
`completed phase 1, 2, and 3 clinical trials of VEGF Trap-Eye in AMD patients.
`
`(IPR2021-00880 Ex.1001, ’338 patent, Examples 1-4). No POSA would reasonably
`
`believe that Regeneron had conducted clinical trials with a “genus” of proteins
`
`comprising “multiple amino acid sequences.”
`
`51. Second, as discussed above, Regeneron publications and others
`
`consistently use both the terms VEGF Trap-Eye and aflibercept to refer back to the
`
`single molecule described in Holash, which was described as having domains from
`
`VEGFR1 and VEGFR2, and the only molecule that Holash disclosed as having
`
`progressed into human clinical trials: VEGF TrapR1R2. (See, e.g., Ex.1075, Ni, 408
`
`(citing to Wulff (Ex.1129), which disclosed preclinical studies of “VEGF TrapR1R2,”
`
`the same nomenclature used in contemporaneously-published Holash); Ex.1119,
`
`Klibanov Dep. Ex. 12, Rudge 2007, 18363, 18370 (discussing VEGF Trap,
`
`including “aflibercept” as a keyword, and citing back to Holash (ref. 20)); Ex.2080,
`
`Heier 2009, 2 (discussing VEGF Trap-Eye and citing back to Holash and discussing
`
`the data presented in Holash for VEGF TrapR1R2); Ex.1120, Klibanov Dep. Ex. 13,
`
`Gomez-Manzano, 940, 945 (“a new anti-VEGF agent, VEGF Trap/aflibercept
`
`(henceforth referred to as VEGF Trap), has been developed by incorporating
`
`domains of both VEGF receptor 1 (VEGFR-1) an