Note: This article was first published in law360 in October 2016.
An
antibody, also known as an immunoglobulin, is a large, Y-shaped protein that is
used by the immune system to bind to antigens on pathogens. An antibody molecule has two heavy
chains and two light chains, with each chain having a constant region and a variable
region. Each of the variable regions
have Complementary Determining Regions (CDRs), which vary greatly in amino acid
sequence from one antibody to another, and are primarily responsible for the
binding affinity of the antibody. Each
light chain and heavy chain has three CDRs, with a total of six CDRs.[i]
Patents
to antibodies are valuable since pharmaceutical companies market and sell
antibodies as therapeutic drugs.
Examples of therapeutic drugs that are antibodies include HUMIRA, ERIBITUX, and AVASTIN. For this
article, a review of recently issued U.S. patents was done to formulate
strategies for prosecuting patent applications claiming an antibody.
Obtaining
a patent with a broad claim to an antibody is full of potential pitfalls. To obtain the broadest claim possible to an
antibody, rather than claiming a whole antibody with the specific sequences of
the six CDRs, many applicants often attempt to claim less than the six CDRs,
such as a heavy or light chain by itself.
This strategy of claiming less than the six CDRs is often unsuccessful,
particularly if the disclosure of the patent application is limited to a single
“intact” antibody with all the six CDRs.
In most instances the patent examiners force the applicants to make a
narrowing amendment, and claim an antibody with all six specific CDR sequences:
The experiments provided in the specification were
performed with intact antibody, i.e.,
two heavy chains and two light chains. The specification as filed provided no
demonstration that the heavy chain by itself bound vaccinia virus H3L protein,
provided no demonstration that the heavy chain variable region by itself bound
vaccinia virus H3L protein, and provided no demonstration that the heavy chain
CDRs are functional in the context of any other antibody backbone.
[ii]
In this case, additional disclosure in the specification, such as a showing
that a light chain or a heavy chain by itself bound to the antigen, would have
probably overcome the examiner’s rejection.
Of the approximately one hundred patents that we reviewed for this
article, we came across only one patent that claims a single light chain or
heavy chain (in form of an isolated nucleic acid) of an antibody by itself.[iii]
Some
applicants pursue another unsuccessful strategy, which is to claim a CDR
sequence that is broader than the actual CDR sequence disclosed in the
specification. Strategies employed by
applicants include claiming a homologous sequence to a particular CDR sequence
or claiming an altered CDR sequence, such as “wherein said antibody comprises
zero, one, or two amino acid substitutions in said set of CDRs.” The examiners
typically reject such claims to a broader CDR sequence:
[W]hile being enabling for an antibody which binds
to the extracellular domain of PDGFRβ or VEGF-A comprising all fully defined 6
CDRs (3 from the heavy chain and 3 from the light chain), [the specification] does
not reasonable [sic] provide
enablement for a PDGFR β or VEGF-A antibody comprising only CDR3 of the heavy
chain defined or “3 or fewer amino acid substitutions relative to a second set
of CDRs” or “one, or two amino acid substitutions.”
[iv] Another strategy by applicants to obtain a broad
patent is to claim a fragment of an antibody, such as by
using the claim term “antigen-binding fragment thereof” or “antigen-binding
portion thereof.” These claim terms are
allowed by the USPTO.[v]
The PTAB (Patent Trial and Appeal Board) has construed the claim term “antigen-binding
portion” in an Inter Partes Review regarding
the drug Humira® to mean fragments of the light chain variable
region and the heavy chain variable region of the claimed antibody that retain
the ability to specifically bind the specified antigen:
[A]lthough we agree that the claimed antibody must
include the light chain variable region and the heavy chain variable region of
D2E7, we do not agree that the claimed ‘an antigen binding portion’ must
include the complete light chain and heavy chain variable regions of D2E7, as
AbbVie asserts. This is particularly
true given the claim’s use of a “(TNFα) antibody, or an antigen-binding portion thereof” (emphasis added), which is
disjunctive.
[vi]
The PTAB, however, refused to construe the claim term to encompass solely the
CDRs, dismissing the petitioner’s argument that the phrase “antigen-binding
portion” encompasses an antibody fragment that can be as small as one CDR (5 to
17 amino acids).[vii] Nevertheless, the claim term “antibody or
antigen-binding fragment thereof” or similar language broadens the scope of an
antibody claim by encompassing at least only the fragments of the variable
regions that are responsible for antigen binding. Applicants should also consider dependent
claims on Fab (which includes one constant and one variable region domain) and
scFv fragments (and other fragments) of the antibody.[viii]
Other
strategies to cover a broader antibody claim were also successful. Despite being unable to claim a sequence that
is homologous to a CDR sequence, an applicant is allowed to claim a homologous
sequence to the variable region of the antibody other than the CDR regions of
the antibody. This sort of claim is
typically presented as a dependent claim that depends on an independent claim
that claims an antibody with a particular CDR sequence:
1. An anti-human TNF-α antibody or antigen-binding
fragment thereof which specifically binds to human TNF- α, wherein said
antibody or antigen-binding fragment comprises the VH CDR
polypeptide sequences of SEQ ID NO: 7, 8, and 9 and the VL CDR
polypeptide sequences of SEQ ID NO: 4, 5, and 6.
23. An anti-TNF-α antibody or antigen-binding
fragment thereof according to claim 1 that binds human TNF- α., wherein said
antibody or antigen-binding fragment comprises VH and VL
polypeptides which each are at least 90% identical to the VH chain
of the polypeptide of SEQ ID NO: 3 and the VL chain of the
polypeptide sequence of SEQ ID NO: 2.
[ix]
The claims of this recently issued patent stood apart from the others in that
it claimed in an independent claim “an
amino acid sequence that is at least 95% identical” to the whole variable
region, including the CDRs:
An antibody, comprising: a) a heavy chain variable
domain comprising an amino acid sequence that is at least 95% identical to the
amino acid sequence of SEQ ID NO. 1; and b) a light chain variable domain
comprising an amino acid sequence that is at least 95% identical to the amino
acid sequence of SEQ ID No. 2, wherein the antibody binds TNFα.
[x]
The examiner actually refused to allow this claim since it did not claim the
exact sequence of the CDRs. The
applicant appealed and the PTAB reversed the examiner since the specification
disclosed “73 amino acid substitutions, including substitutions in the CDRs:”
The Specification also discloses the location and
sequence of the three CDRs of each of the heavy and light chain variable
domains (FF 4). In addition, the
Specification discloses 73 amino acid substitutions, including substitutions in
the CDRs (FF 5). Citing to Couto, the Specification states that an “antibody
having any of these substitutions should neutralize TNFα activity as antibodies
with all of these substitutions have been shown to neutralize TNFα activity”
(FF 5) . . . .The Examiner argues, however, that “[t]here does not appear to be
an antibody in [Couto] ‘which contains 73 amino acid ... substitutions relative
to the antibody defined by SEQ ID NOS: 1 and 2.’” Instead, the Examiner argues
that Couto “discloses four antibodies (see the ‘Group F’ antibodies) in Figures
lA and B, each appearing to have a sub-set of the framework and CDR mutations
set forth in the Tables 1 and 2 of the instant specification” We are not
persuaded.
[xi]
This case highlights the importance of having disclosure in the specification
about amino acid substitutions in the CDRs when claiming an antibody with a
sequence that is broader than the actual sequence of the CDRs.
One
way to claim an antibody without claiming a particular antibody sequence is to
claim an antibody that binds to a specific antigen. Applicants typically claim an antibody to a
particular epitope of an antigen molecule.
An epitope is the part of an antigen molecule to which an antibody
attaches itself. An epitope can either
be linear (formed by amino acids next to each other in a sequence) or
conformational (formed by amino acids that are apart from each other in the
sequence but are in proximity to each other in the three dimensional structure
of the molecule). Applicants do not
typically claim an epitope, but instead claim an antibody that binds to the
epitope since claiming the epitope by itself has limited value. First, since the epitope is typically
naturally found, it cannot be patented under Section 101. Second, a pharmaceutical company is not going
to market or sell an epitope, but most likely an antibody that binds to the
epitope to treat a condition.
The
USPTO’s written description guidelines state that there is written description
support for a claim of “an isolated antibody capable of binding to antigen X”
even if “the specification does not describe an actual reduction to practice of
an antibody that binds to antigen X by reference to a deposit (e.g., deposit of a hybridoma) or by
describing an antibody in structural terms sufficient to show possession.”[xii] While the USPTO’s written description
guidelines for claiming an antibody to a specific antigen are lenient, the
Federal Circuit has taken a more demanding approach. In one case, the Federal Circuit affirmed a
jury’s verdict that the following claim to an antibody that binds to an antigen
was invalid due to lack of written description:
A
neutralizing isolated human antibody, or antigen-binding portion thereof that binds
to human IL-12 and disassociates from human IL-12 with a koff rate
constant of 1x10-2 s-1 or less, as determined by surface
plasmon resonance.
[xiii]
In holding this claim to be invalid, the Federal Circuit stated that the
specification only described one type of structurally similar antibodies and
that those antibodies were not representative of the full variety or scope of
the genus covered by the claim:
All of the antibodies described in AbbVie’s patents
were derived from Joe-9 and have VH3 type heavy chains and Lambda
type light chains. Although the described antibodies have different amino acid
sequences at the CDRs, they share 90% or more sequence similarity in the
variable regions and over 200 of those antibodies differ from Y61 by only one
amino acid. The patents describe that other VH3/Lambda antibodies
may be modified to attain IL-12 binding affinity. However, the patents do not
describe any example, or even the possibility, of fully human IL-12 antibodies
having heavy and light chains other than the VH3 and Lambda types.
In contrast, Centocor’s Stelara, which falls within
the scope of the claimed genus, differs considerably from the Joe-9 antibodies
described in AbbVie’s patents. Stelara has VH5 type heavy chains and
Kappa type light chains. The variable regions of Stelara only share a 50%
sequence similarity with the Joe-9 antibodies, which is far lower than the 90%
sequence similarity shared among the Joe-9 antibodies described in AbbVie’s
patents. Centocor’s expert testified that antibodies with 80% sequence
similarity to J695 could bind to completely different antigens, J.A. 6496–97,
thus illustrating the significant structural differences between Stelara and
the Joe-9 antibodies and the unpredictability of the field of invention.
Centocor also presented evidence of other differences between Stelara and the
Joe-9 antibodies, such as CDR length and epitope binding site.
[xiv]
The Federal Circuit’s decision is in conflict with the USPTO’s written
description guidelines that allow claiming an antibody to a specific antigen
even if the specification does not disclose a specific antibody that binds to
the particular antigen.[xv] To ensure that an antibody claim without
structural limitations will survive an invalidity challenge in the event of
litigation, applicants should consider disclosing more species of antibodies
and go beyond what the USPTO’s written description guidelines require.
The
USPTO is allowing typical antibody-epitope claims like this one:
An anti-granulysin antibody, or an
scFv or Fab fragment thereof, wherein the antibody has the ability to
neutralize an activity of granulysin, wherein the antibody binds to epitope
regions located at residues 64-67 and at residues 97-108 in the sequence of
granulysin (SEQ ID NO:79).
[xvi]
A claim to an antibody that binds to a particular epitope can be obtained even
if antibodies to the target protein that contain the epitope exist in the prior
art as long as those prior art antibodies bind to another epitope on the
antigen:
Applicant has claimed antibodies which neutralize
granulysin and which bind to a specific epitope in the protein. The prior art
discloses antibodies which bind to and neutralize granulysin (see for example
Okada et al. and Chen et al. discussed above), as well as methods of using such
antibodies to treat disorders such as GVHD (see particularly U.S. Patents
7,718,378 and 8,377,436). Notably, the '436 patent is the closest prior art and
while it contains issued method claims for treating SJS, TEN, and GVHD with
antibodies which bind the 15 kDa form of human granulysin, there is no
direction or guidance to the precise epitope within granulysin which must be
bound as is recited in the instant claims. Further, as was discussed earlier in
this action, there is more than one epitope in granulysin which when bound by
an antibody neutralizes the activity of granulysin. Thus, the instant claimed
inventions are species which lie within the genus of products and methods
disclosed in the prior art, and absent the guidance and direction of the
instant specification (such as the epitope mapping data presented in the
instant specification) there does not appear to be anything which would lead
artisans to that which is presently claimed.
[xvii]
Some Applicants claim an antibody binding to a particular epitope, and not
other epitopes:
1. A recombinant antibody to a human IL-3Rα chain,
which does not inhibit IL-3 signaling and binds to the B domain of the human
IL-3Rα chain but does not bind to the C domain of the human IL-3Rα chain,
wherein said antibody binds to a region within positions 101-203 of SEQ ID NO:
2, but does not bind to any region within positions 204-308 of SEQ ID NO: 2.
U.S. Patent 9,394,370, claim 1. The original independent claim had only the
negative limitation of “does not bind to the C domain of the human IL-3Rα chain,”
but the examiner forced the applicant to add the wherein clause to further
define the positions in the sequence where the claimed antibody did not bind.[xviii]
Negative claim limitations like these
may be useful to further define an antibody and to exclude the prior art. While the Federal Circuit has recently held
that negative claim limitations do not have a higher written description
standard,[xix]
it may be a good practice to draft the specification to have support (and even
have dependent claims) for excluding epitopes that were disclosed in the prior
art.
Another
typical antibody claim that is not limited to a particular sequence of an
antibody is a claim to a method of treating a particular condition by
administering an antibody that is specific to a particular target:
A
method for reducing incidence of or treating allodynia in an individual,
comprising administering to the individual an effective amount of an anti-CGRP
antagonist antibody, wherein the anti-CGRP antagonist antibody is
a monoclonal antibody.
[xx] These claims only
claim an antigen in broad terms and do not specify a sequence for the
antigen/epitope, or the antibody. While
these claims seem broad, in one case, the Federal Circuit
construed a similar claim (where a method of treatment with an antibody to a
target was claimed) to be limited to a particular epitope on the target,
despite the claim not reciting a particular epitope. Specifically, in a Biogen patent that was
before the Federal Circuit in 2013, the patent claimed a method of treatment
with an anti-CD20 antibody without specifying the particular epitope on the
CD20 to which the antibody binds:
A method of treating chronic lymphocytic leukemia in
a human patient, comprising administering an anti-CD20 antibody to the patient
in an amount effective to treat the chronic lymphocytic leukemia, wherein the
method does not include treatment with a radiolabeled anti-CD20 antibody.
[xxi]
The Federal Circuit agreed with the district court that the claim term
“anti-CD20 antibody” is limited to the particular epitope on CD20 that Biogen’s
antibody binds. The claim term
"anti-CD20 antibody" was construed to mean “rituximab and antibodies
that bind to the same epitope of the CD20 antigen with similar affinity and
specificity as rituximab” in view of the prosecution history in which Biogen
responded to an enablement rejection by stating that “one of skill in the art
could readily identify an antibody that binds to CD20 with similar affinity and
specificity as does Rituxan® using techniques that are well known in
the art.”[xxii]
The dissent noted that Biogen in its “response omitted the very term ‘epitope’
that the majority claims they adopted as a limitation.” Thus, the slightest statement in the
prosecution history can be used to limit the literal scope of the claim to the
particular epitope of the target for which an example is illustrated in the
specification.
In
the Biogen case, after the district court construed the term "anti-CD20
antibody," Biogen stipulated to non-infringement and did not move forward
based on infringement under the doctrine of equivalents. In Biogen a routine statement in an office
action response was relied on to limit the literal scope of the claim, much
less the scope of equivalents, which is relatively easy to forfeit in a
biotechnology case. By being overly
aggressive, such as by claiming more than the actual CDR sequences, and then
narrowing the claim during prosecution, applicants are forfeiting the scope of
equivalents. Forfeiting equivalents is
not limited to arguments/amendments before the patent office, but also under
the “disclosure-dedication doctrine.”
This doctrine is based on the general notion that “when a patent drafter
discloses but declines to claim subject matter . . . this action dedicates that
unclaimed subject matter to the public.”[xxiii] Many antibody patent applications are written
with overly broad disclosures that may limit the scope of equivalents. For example, these applications typically
have disclosure that encompasses a low percentage of homologous sequences (in
this case at least 60%):
In one embodiment, an antibody of the present
invention comprises a heavy chain, wherein the variable domain of the heavy
chain comprises three CDRs wherein the sequence of CDRH-1 has at least 60%
identity or similarity to the sequence given in SEQ ID NO:1, CDRH-2 has at
least 60% identity or similarity to the sequence given in SEQ ID NO:2 and/or
CDRH-3 has at least 60% identity or similarity to the sequence given in SEQ ID
NO:3. In another embodiment, an antibody of the present invention comprises a
heavy chain, wherein the variable domain of the heavy chain comprises three
CDRs wherein the sequence of CDRH-1 has at least 70%, 80%, 90%, 95% or 98%
identity or similarity to the sequence given in SEQ ID NO:1, CDRH-2 has at
least 70%, 80%, 90%, 95% or 98% identity or similarity to the sequence given in
SEQ ID NO:2 and/or CDRH-3 has at least 70%, 80%, 90%, 95% or 98% identity or
similarity to the sequence given in SEQ ID NO:3.
[xxiv]
Even if there is not a narrowing amendment, such a board disclosure could doom
the scope of equivalents for a claim that is limited to the actual sequence of
the CDRs.
In
summary, a review of recently issued antibody patents illustrates some
successful strategies to obtain patents with relatively broad claims to
antibodies, but many other strategies result in a narrowing amendment and
forfeiture of the scope of the equivalents.
Claiming an antibody fragment and a variable region other than the CDRs
with a particular homology seem to be successful strategies. Claiming less than a whole antibody (other
than the typical language of “an antigen binding fragment thereof”) typically
result in a narrowing amendment unless the applicant has disclosure in the
specification to argue that such claims are enabled and supported by the
specification. Claims to antibodies that
bind to particular antigens/epitopes or to administration of antibodies to treat
a condition can be used to avoid limiting a claim to a particular sequence of
an antibody, but the Federal Circuit has narrowly construed or invalidated such
claims.
[i] Each antibody has a total of 12
CDRs since each antibody has two light chains and two heavy chains.
[ii] U.S. Patent No. 9,447,172,
Office Action of Sept. 3, 2015, page 9.
[iii] U.S. patent 9,447,180, claims 1
and 2.
[iv] U.S. Patent
9,441,034, Office Action of June 8, 2015, page 2-3.
[v] See e.g. U.S. Patent 9,465,041. Claiming “An isolated antibody or a binding fragment
thereof.”
[vi] IPR2015-01514, paper No. 9, at
9.
[vii] Id.
[viii] A Fab fragment is composed of
one constant and one variable domain of each of the heavy and the light
chain. A single-chain variable fragment
(scFv) is a fusion protein of the variable regions of the heavy and light
chains connected with a short linker peptide.
[ix] U.S. Patent 9,056,905, claims 1
and 23. See also U.S. Patent 9,102,727, claims 1 and
2.
[x] U.S. Patent 9,365,644, claim
1.
[xi] Ex parte Yaohuang Ke, Appl. No. 12/425,763, Appeal No. 2013-009436,
Feb. 10, 2016.
[xii] USPTO written Description
Guidelines, March 25, 2008, version 1, page 45 <http://www.uspto.gov/sites/default/files/web/menu/written.pdf>.
[xiii] AbbVie Deutschland GMBH v. Janssen Biotech, Inc., 759 F.3d 1285,
1292 (Fed. Cir. 2014).
[xiv] Id. at 1300.
[xv]< http://jolt.law.harvard.edu/digest/patent/written-description-problems-of-the-monoclonal-antibody-patents-after-centocor-v-abbott> (Discussing Centocor Ortho Biotech, Inc. v. Abbott Labs,
636 F.3d 1341, 1344 (Fed. Cir. 2011).
[xvi] U.S. Patent 9,428,575, claim 1.
[xvii] U.S. Patent 9,428,575, Notice of
Allowance dated July 26, 2016, page 7.
[xviii] Appl. No. 13/912,978 , Amendment
of January 19, 2016.
[xix] <http://www.ipwatchdog.com/2015/12/04/negative-claim-limitations-no-higher-standard/id=63608/>
[xx] U.S. Patent No.9,328,167, claim
1. See
also U.S. Patent No. 9,428,582.
[xxi] U.S. Patent 7,682,612, Claim
1.
[xxii] Biogen Idec, Inc. v. Glaxosmithkline LLC, 713 F.3d 1090, 1094-1095
(Fed. Cir. 2013).
[xxiii] Johnson & Johnston Assocs. v. R.E. Serv., 285 F.3d 1046, 1054
(Fed. Cir. 2002) (en banc).
[xxiv] U.S. Patent 9,428,570, 8:24-45.
No comments:
Post a Comment