Ex Parte Heywood et al

Patent Trials and Appeals Board

Jul 2, 2019

13637211 - (D) (P.T.A.B. Jul. 2, 2019)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
13/637,211 12/10/2012 Sam Philip Heywood
20306 7590 07/03/2019
MCDONNELL BOEHNEN HULBERT & BERGHOFF LLP
300 S. WACKER DRIVE
32NDFLOOR
CHICAGO, IL 60606
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www .uspto.gov
ATTORNEY DOCKET NO. CONFIRMATION NO.
12-1084-WO-US 2520
EXAMINER
ROARK, JESSICA HOPE
ART UNIT PAPER NUMBER
1643
MAIL DATE DELIVERY MODE
07/03/2019 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte SAM PHILIP HEYWOOD and
DAVID PAUL HUMPHREYS 1
Appeal2018-000747
Application 13/637,211
Technology Center 1600
Before DONALD E. ADAMS, ERIC B. GRIMES, and
FRANCISCO C. PRATS, Administrative Patent Judges.
GRIMES, Administrative Patent Judge.
DECISION ON APPEAL
This is an appeal under 35 U.S.C. § 134 involving claims to a target-
binding protein, which have been rejected as obvious. We have jurisdiction
under 35 U.S.C. § 6(b ).
We reverse.
STATEMENT OF THE CASE
The invention "relates to antibodies with two antigen binding sites in
a format with suitable stability ... to be commercially viable." Spec. 1: 4---6.
1 Appellants identify the Real Party in Interest as UCB BIOPHARMA
SPRL. Br. 2.
Appeal2018-000747
Application 13/637,211
"Multivalent antibodies are known," including "a multivalent multispecific
antibody (DVD-lg) of the type shown in Fig[.] 1." Id. at 1:8, 37-38. Figure 1
is reproduced below:
Fi_gu1·C' l
-D-
DVD-lg
Figure 1 "shows a diagrammatic represent[ation] of a DVD-lg
molecule." Id. at 2:31. The Specification states that "[t]hese formats after
expression may form soluble aggregates in liquid carriers which can be
problematic when formulating a biotherapeutic agent." Id. at 2:2-3.
The Specification discloses "a stable multivalent format, which is
thought to be capable of expression in a host and with a suitable stability
profile." Id. at 2:5-6. The disclosed multivalent format includes "a disulfide
bond between VH1 and VL1." Id. at 2:22. (VH1 and VL1 correspond to the
heavy (bolded) and light chains, respectively, of each Vb domain in Figure 1.
Cf Spec. Fig. 2.) "The disulfide bond between VH1 and VL1 seems to aid
general stability of the molecules. After expression and any purification this
increased stability may, for example manifest itself in the absence of
aggregation in liquid formulations of the antibody." Id. at 2:25-28.
2
Appeal2018-000747
Application 13/637,211
Claims 1, 4, 6, 8, 12, and 16 are on appeal. Claim 1 is illustrative and
reads as follows ( emphasis added to disputed limitation):
1. A binding protein, comprising:
a first polypeptide comprising:
VH1-X1-VH2-CH-X2,
wherein VH1 is a first heavy chain variable domain, VH2 is a
second heavy chain variable domain, CH is a constant domain, X1
represents a flexible peptide linker that is one of SEQ ID NOs: 10 to
79, X2 represents an Fe region, and
a second polypeptide comprising:
VL1-X1-VL2-C,
wherein VL1 is a first light chain variable domain, VL2 is a
second light chain variable domain, C is a constant domain, and X1
represents a flexible peptide linker that is one of SEQ ID NOs: 10 to
79,
wherein VH1 and VL1 form a first binding domain and VH2 and
VL2 form a second binding domain; and
wherein there is a disulfide bond between VH1 and VL1.
Br. 15 (Claims Appendix).
DISCUSSION
The Examiner has rejected all of the claims on appeal under 35 U.S.C.
§ I03(a) as obvious based on Wu-Abbott,2 Webber, 3 and Wu-Medi. 4 Ans. 3.
The Examiner finds that Wu-Abbott "teaches dual variable domain
2 Wu et al., WO 2008/024188 A2, published Feb. 28, 2008. The Examiner
and Appellants refer to this reference as "Wu-Abbott," so we do as well.
3 Webber et al., Preparation and Characterization of a Disulfide-Stabilized
Fv Fragment of the Anti-Tac Antibody: Comparison with its Single-Chain
Analog, 32 MOL. IMMUNOL. 249-258 (1995).
4 Wu et al., US 2009/0155275 Al, published June 18, 2009. The Examiner
and Appellants refer to this reference as "Wu-Medi," so we do as well.
3
Appeal2018-000747
Application 13/637,211
imnmnoglobulins ('DVD-lg')." Id. at 3. The Examiner finds that the
"DVD2-lg" construct disclosed by Wu-Abbott includes a heavy chain and a
light chain with the variable domains, constant domains, and binding
domains that are recited in claim 1, as well as the recited peptide linkers and
Fe region. Id. at 3--4. The Examiner thus finds that "Wu-Abbott teaches
DVD-lg constructs comprising first and second polypeptides with the
domains as recited in claim 1, in which the linkers between the domains are
also as recited in claim 1." Id. at 4.
The Examiner finds that "Wu-Abbott does not teach including a
disulfide bond between VHI and VLI, as required by claim 1." Id. at 5
( emphasis omitted). The Examiner finds, however, that Webber "teaches
that introducing cysteine residues into both the VH and VL of an antibody
resulted in a disulfide bond between the VH and VL that stabilized the
antibody and improved yield." Id. The Examiner finds that Wu-Medi also
"teaches that in any of a variety of antibody constructs, introduction of a
disulfide bond between the variable domains on different chains can be used
to stabilize the antibody molecule." Id.
The Examiner concludes that it would have been obvious, in view of
the teachings of Webber and/or Wu-Medi, "to include one or more disulfide
bonds between the first variable domains of the construct of Wu-Abbott.
Webber teach methods for introducing such bonds ... and indicates that
including them improves stability of the constructs [and] Wu-Abbott notes
stability issues with their constructs." Id. at 6. The Examiner reasons that
[i]ntroduction of a disulfide bond between the first VH+VL pair
would provide stability to ... the lesser constrained of the two
pairs since the VH2+ VL2 pair is coupled at least partially by
the natural disulfide linkage that forms between CHI and C[].
4
Appeal2018-000747
Application 13/637,211
Thus, the ordinary artisan had both a good reason to introduce
disulfide bonds ... between the first variable domains in the
construct of Wu-Abbott and a reasonable expectation that
introduction of those disulfide bonds would predictably
improve the stability of the constructs.
Id. The Examiner thus concludes that it would have been obvious
to apply the same technique (introduction of disulfide bonds
between antibody variable domains, as taught by Webber and
Wu-MEDI) to a similar antibody construct (the bispecific,
DVD-lg type arrangement of Wu-Abbott) which the art had
already identified as susceptible to stability issues ... to
predictably improve the stability of that similar antibody
construct.
Id. at 7.
Appellants argue that the invention "arose from the instant inventors'
insight that the DVD-lg was susceptible to the dynamic process of
'breathing,' which can result in 'promiscuous intermolecular pairing of
variable regions.' See Dr. Humphreys Declaration filed April 28, 2016 ... at
,r,r 23 - 24 and 32 - 37." Br. 3. The Humphreys Declaration states that "[i]n
this process of breathing, the variable regions separate for a period of time
and then re-associate. However, ... instead of intra-molecular pairing, an
intermolecular pairing of variable regions can occur, which results in
dimerization and higher order aggregation." Humphreys Deel. ,r 24.
Br. 3.
Appellants argue that they
discovered that incorporating a disulfide bond between VH 1
and VL 1 addresses the problem of dynamic breathing process
and, hence, DVD-lg aggregation. Id. at ,r 36. Before this
discovery, there simply was no reason to link DVD-lg VHl and
VL 1 domains with a disulfide bond because there was no
perceived benefit to be gained.
5
Appeal2018-000747
Application 13/637,211
Appellants point out that "although Wu-Abbott Table 22, page 152,
indicates that there is aggregation, Wu-Abbott does not identify the
underlying cause and does not offer any solutions to the aggregation
problem." Id. at 4. Appellants argue that
the Office speculated that some aggregate formation and loss of
stability during storage in Wu-Abbott's constructs is due to the
lack of a disulfide bond between VH 1 and VL 1. But there is no
basis for this conclusion in the cited art ... , and the Office
provides no "articulated reasoning" and presents no evidence or
scientific reasoning supporting it.
Id. at 5.
Appellants argue that "Wu-MEDI ... provides only general guidance
regarding measures that may be taken to stabilize DVD-lg constructs and
certainly does not direct the ordinary artisan to insert a VH 1-VL 1 disulfide
bond as recited in the instant claims." Id. at 5---6.
Appellants also argue that "Webber discloses dsFvs and single chain
Fv molecules, which do not contain CHI and CKappa that drive
dimerization and, instead, consist solely of VH and VL domains." Id. at 7.
Appellants argue that "[t]he entire disclosure of Webber is limited to Fvs,
i.e., antibody molecules having a single pair of VL and VH domains ...
[which] are markedly different from Wu-Abbotts's DVD-lg constructs." Id.
at 8. Thus, Appellants argue, "[t]he ordinary artisan would have recognized
that ... the results observed with Fv would not be predictive of the results
with DVD-lg, and, so, the results observed with Fe [sic, Fv] would not have
given the ordinary artisan reason to insert a disulfide bond between VHI and
VLI in DVD-lg." Id. at 9.
We agree with Appellants that the Examiner has not shown that a
skilled artisan would have considered it obvious, based on Webber or
6
Appeal2018-000747
Application 13/637,211
Wu-Medi, to modify the DVD-lg construct disclosed by Wu-Abbott by
adding a disulfide bond linking the VHl and VLl domains. The Examiner
finds that a skilled artisan would have been led to increase the stability of
Wu-Abbott's constructs because "Wu-Abbott notes that there was some
aggregate formation and loss of stability during storage. E.g., Table 22, page
152." Ans. 5.
The cited table describes an exemplary "anti-IL-12/anti-IL-18 DVD-
Ig." Wu-Abbott 152, Table 22 heading. The table states, under "Stability
(freeze/thaw)": "- 5% aggregates after 2 freeze-thaw cycles, increased to
-13% after additional 10 freeze-thaw cycles. The reason for that is unsolved
(process-related, sequence-specific, or LC [light chain] lambda/kappa
hybrid)." Id., Table 22. The disclosure of Wu-Abbott therefore supports
Appellants' position that it "does not identify the underlying cause and does
not offer any solutions to the aggregation problem." Br. 4.
The Examiner cites Webber and Wu-Medi as evidence that a skilled
artisan would have considered it obvious to modify Wu-Abbott's construct
by adding a disulfide bond between VLl and VHI. However, even the
Examiner "acknowledge[s] that the teaching[s] of Wu-MEDI are general."
Ans. 13. Wu-Medi describes antibody constructs with generic "interchain
disulfide bonds," but the passages cited by the Examiner do not describe a
construct with a disulfide bond between the light chain and heavy chain of a
variable region.
Webber describes modification of Fv antibody fragments to form
disulfide-stabilized Fv (dsFv). Webber 249, abstract. Fv fragments are
"independent variable domains of the heavy and light chains non-covalently
associated in one-to-one stoichiometry." Id. "[T]hey are not usually stable
7
Appeal2018-000747
Application 13/637,211
heterodimers." Id. at 249, right col. Webber reports that "an Fv molecule ...
is stabilized by an inter-chain disulfide bond between the framework regions
of the heavy and light variable domains." Id. at 250, left col. Webber does
not describe aggregation of Fv molecules, or describe its dsFv as having a
reduced tendency to aggregate; rather, it reports that "stability to both
thermal and urea-induced denaturation was significantly higher for the dsFv
than for the single-chain analog [scFv]." Id. at 254, left col.
Webber states that "[a]n scFv ... is composed of the variable domains
from matching heavy and light chains which have been connected by a short
linker sequence." Id. at 249, right col. Webber states that scFvs "have a
tendency to aggregate," which "may be a consequence of the fact that the
heavy and light domains of scFvs can separate from one another if the non-
covalent, inter-domain forces are not strong enough to hold them in place,"
leaving them "free to interact detrimentally with nearby molecules or other
'sticky' surfaces." Id. at 249-250, bridging paragraph.
The Examiner reasons that a skilled artisan would have understood
that the disclosure of aggregation by Wu-Abbott indicated that its DVD-lg
was behaving, in this respect, like an scFv:
The ordinary artisan understood that aggregation was an
indication that a VH-VL pair was not forming a stable
association. When the VH-VL are tethered, such as by the
peptide linker in a scFv, the result of that instability is
aggregation. Given the wealth of guidance in the prior art
regarding the inclusion of a disulfide bond to stabilize VH-VL
interactions, the ordinary artisan would have applied this same
solution to the outer VHl-VLl pair [of Wu-Abbott's DVD-lg]
and reasonably expected it to reduce aggregate formation while
preserving antigen binding.
Ans. 13.
8
Appeal2018-000747
Application 13/637,211
This conclusion, however, is not supported by a preponderance of the
evidence. Wu-Abbott expressly states that "[t]he reason for [aggregation
after freezing and thawing] is unsolved" and suggests that it could be
"process-related, sequence-specific, or [ due to] LC lambda/kappa hybrid"
formation. Id., Table 22.
Wu-Abbott does not suggest that "a VH-VL pair was not forming a
stable association" in its DVD-lg constructs, as posited by the Examiner.
Ans. 13. Nor does Wu-Abbott suggest that, like Webber's scFvs, its DVD-lg
constructs aggregated because "the non-covalent, inter-domain forces are not
strong enough to hold [the heavy and light chains of the variable domain] in
place." See Webber 249, right col. For its part, Wu-Medi does not
specifically suggest a disulfide bond between variable domains of an
antibody construct.
In summary, we conclude that the Examiner has not shown, by a
preponderance of the evidence of record, that a person of ordinary skill in
the art would have had a reason to modify Wu-Abbott's DVD-lg construct
to include a disulfide bond between the heavy and light chains of the outer
variable domain, as required by the claims on appeal. We therefore reverse
the rejection of claims 1, 4, 6, 8, 12, and 16 under 35 U.S.C. § 103(a) based
on Wu-Abbott, Webber, and Wu-Medi.
REVERSED
9