Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02389386 2002-04-29
WO 01/34651 PCT/GB00/04279
1
ANTIBODIES BINDING A NON NATURALLY OCCURRING ENANTIOMER (L-BIOTIN) AND THEIR
USE AS
TARGETING AGENTS
Field of the Invention
This invention relates to targeting agents, and in particular to antibodies
and to
their use in targeting anti-tumour agents in vivo.
Background to the Invention
The targeting of therapeutic agents to particular sites in vivo, is well
known. In
particular, it is very desirable to target anti-cancer agents to a tumour
site, to increase
the concentration of the agent at the site and thereby improve its
effectiveness in
neutralising the tumour. Many agents that target tumours rely on the
specificity of
monoclonal antibodies for delivering the diagnostic or therapeutic agent to
the target
site. One approach has been to use a radionuclide-antibody conjugate which
localises
at a target tissue where the radionuclide may exert its cytotoxic effect.
There are problems with the utility of radionuclide-antibody conjugates, for
example poor penetration of the conjugates to the target site owing to the
high
molecular weight of the conjugate. In addition, for the conjugate to be
therapeutically
effective, it must be given time to localise at the target site. The
radionuclide is
therefore present in the body for prolonged periods, and this results in
undesirable
toxicity at non-target sites.
US-A-5630996 describes the use of antibody-streptavidin conjugates to target
a radionuclide-labelled biotin. Streptavidin has high affinity for biotin and
is able to
localise the radionuclide at the target site through the biotin-streptavidin
interaction.
However, streptavidin is a protein that is immunogenic in humans, and
consequently
may not be suitable for repeated, long-term therapeutic use.
Summary of the Invention
The present invention is based on the realisation that L-biotin, i.e. any of
the
three non-naturally-occurring enantiomers of biotin, is a suitable ligand for
antibody-
targeted therapy. More generally, this is true for any endogenous chiral
compound, so
that any non-naturally-occurring enantiomer (NNOE) can be used, according to
this
invention. Since neither an anti-NNOE antibody nor the NNOE itself will
interact with
endogenous/naturally-occurring enantiomer (NOE) or the receptors or carrier
proteins
for NOEs, the NNOE can be used to target therapeutic agents to a desired locus
in
vivo. By way of example, L-biotin is biologically inactive and therefore, in
general,
naturally occurring molecules which bind to the NOE or active enantiomer, D-
biotin, will
not disrupt the targeting of the agent.
CA 02389386 2002-04-29
WO 01/34651 PCT/GB00/04279
2
According to a further aspect of the invention, the NNOE is covalently or
otherwise linked to a therapeutic or diagnostic agent, and used to target the
agent to
a site in vivo. The agent may be a chemotherapeutic agent, e.g. a
radiolabelled
compound.
Typically, the NNOE will target a site in vivo by interaction with an antibody
localised at the target site. Accordingly, a further aspect of the invention
is an antibody
having affinity for the NNOE and not the NOE. The antibody may be a
bifunctional
antibody, of which one functionality is as described above and the other
involves, for
example, affinity for a tumour.
Alternatively, a novel conjugate comprises an antibody and the NNOE. The
conjugate may be administered to a patient to localise the NNOE at a target
site.
Subsequently, usually after the administration of a clearing agent, an
antibody-
therapeutic agent conjugate is administered, the antibody having
affinityforthe NNOE,
to localise the agent at the target site.
1 S In another novel conjugate, the NNOE molecule is linked to an enzyme that
can
be used to convert a suitable prodrug into an active cytotoxic form.
Description of the Invention
By way of illustration only, the invention will now be described for a
specific
NNOE, i.e. L-biotin. This is an example of a preferred NNOE for use in the
invention,
i.e. is a relatively small, non-toxic organic molecule that exhibits neutral
biodistribution
in the body, is water/serum-soluble, and is readily excreted by the kidney. It
is therefore
a suitable molecule for use in targeting therapeutic or diagnostic agents in
vivo.
Antibodies for use in the present invention may be produced using conventional
techniques, for example, hybridoma synthesis, recombinant DNA techniques,
phage
display or other library technology. The antibodies may be derived from any
species,
including rodent, although it is preferred that the antibodies are derived
from mammals
other than rodents, e.g. sheep, goats or cows, and have higher affinity.
Typically, the
antibodies will have an affinity of at least 10'° I/mol, preferably at
least 10" I/mol, more
preferably at least 10'2 I/mol and most preferably at least 10'3 I/mol, e.g.
up to 10'5 or
10'6 I/mol, for the respective ligands.
A preferred embodiment of the invention is a bifunctional antibody. A
bispecific
antibody has affinity for the NNOE rather than the NOE, and may also have
affinity for
a target site. As part of a multi-step procedure, in which the antibody is
administered
and allowed to localise before the active agent is given, a bispecific
antibody allows the
use of more potent doses of the active agent, with reduced systemic side-
effects. A
CA 02389386 2002-04-29
WO 01/34651 PCT/GB00/04279
3
bifunctional or other antibody according to the invention may be a whole
antibody or
may be an engineered antibody or a fragment, e.g. f(ab)z. In a further
embodiment, the
antibody may comprise two single chain fv fragments. The preparation of
bifunctional
sFvs is well known. For example, Carter etal., Current Opinion in
Biotechnology (1997)
S 8:449-454, discloses the production of bifunctional sFvs using a phage
display library.
The antibodies may be fully human. Alternatively, antibodies for use in the
invention may be modified by recombinant DNA techniques to "humanise" them,
making them less immunogenic when administered to a patient. A humanised
bifunctional antibody should comprise at least the hypervariable region from
both a
monoclonal antibody having affinity for target antigen, and a monoclonal
antibody
having affinity for the NNOE. The remainder of the antibody variable region
may be of
human immunoglobulin. A higher proportion of human immunoglobulin may be
present
in a whole antibody or a fragment, e.g. F(ab')2. When a single-chain Fv
fragment is
used, the fragment may comprise hypervariable regions as described above and,
optionally, the variable framework from human immunoglobulin.
In addition to having affinity for the NNOE, the bifunctional antibody will
also
have affinity for a particular target site. Typically, the target site will be
a tumour and
the antibody will have affinity for a tumour-associated antigen. An example of
a
tumour-associated antigen is the carcinoembryonic antigen (CEA) which is found
on
colorectal tumours and other adeno-carcinomas.
In a preferred embodiment, the NNOE is covalently or otherwise linked to a
therapeutic or diagnostic agent. The agent may, for example, be a
radionuclide.
Administration of the NNOE-radionuclide conjugate will localise the
radionuclide at a
tumour site through binding to the antibody, to exert a cytotoxic effect on
the tumour.
Radionuclides having a cytotoxic effect are well known. A preferred
radionuclide that
may be used in the invention is a radioisotope of phosphorus, yttrium or
iodine, e.g. I'23,
I'24 and I'25 which may be used for diagnostic purposes, and I'3', l~°
and P32 which may
be used in therapeutics. Another suitable agent is a cytotoxic drug, e.g.
ricin or
calicheamycin.
In an alternative embodiment, the NNOE is linked (conjugated) to an enzyme.
The enzyme is capable of converting a suitable prodrug into an active
cytotoxic form.
The term "prodrug" is used herein to define an inactive form of a drug which
may be
cleaved by enzymic action to release the therapeutically-active form. Suitable
enzyme-
prodrug systems are known to those skilled in the art and include
carboxypeptidases
and modified mustard gas derivatives. The NNOE targets the enzyme to the
target site
CA 02389386 2002-04-29
WO 01/34651 PCT/GB00/04279
4
and subsequent administration of the prodrug will result in the active agent
being
localised at the target.
Methods for attaching therapeutic agents to the NNOE will be apparent to those
skilled in the art. For example, procedures for conjugating D-biotin to drugs
are known,
and can be adopted for the purposes of this invention.
For use in the invention, the antibody and the NNOE-cytotoxic agent may be
formulated in a kit, e.g. comprising the two components separately packaged or
in
separate containers. Each component may be formulated with a suitable carrier
or
excipient, examples of which are well known, depending on the route of
administration,
e.g. oral or intravenous.
These two components will usually be administered sequentially. The
respective times and durations of administration, and the effective amount of
each may
readily be determined by the skilled person, and will depend on typical
factors such as
the location, severity and spread of the tumour, the condition of the subject
etc. It is of
course a feature of this invention that the amount of cytotoxic agent that is
required will
be less than in the absence of the antibody.
In a preferred embodiment, a bispecific antibody is administered, then a
clearing
agent and then the effector molecule. This sequence, and suitable clearing
agents, will
be familiar to those skilled in the art, e.g. with reference to ADEPT
technology; see, for
example, Drugs of the Future 1996, 21(2): 167-181, incorporated herein by
reference.
The following Example illustrates the invention. In particular, it
demonstrates the
specificity of an anti D-biotin antibody for D-biotin as compared to L-biotin.
Examale
Nunc Maxisorb ELISA plates (Nunc, Denmark) were coated with cytochrome
C-D-bt. Cytochrome C-D-bt (C2022, Sigma Chemical, Poole, UK) was diluted to
200
ng/ml in PBS at pH 7.2 and 100 NI of this solution added to each well of the
plate and
incubated for 1 hr at 37°C. Prior to the addition of test samples, the
plates were
blocked with low fat milk protein. For blocking, the plates were washed 3
times in PBS
with 0.01 % Tween 20, and 200 NI of 0.2% low fat milk protein was added to
each well
and incubated at 37°C for 30 minutes.
Samples of the high affinity sheep monoclonal antibody to D-biotin (1D10, KS
Biomedix, UK) were prepared for assay of anti D-biotin activity as follows:
100 NI aliquots of the purified monoclonal antibody at a concentration of 20
ng/ml in 1 % BSA in PBS at pH 7.2 were pre-incubated with various
concentrations of
CA 02389386 2002-04-29
WO 01/34651 PCT/GB00/04279
either D-biotin or a racemic mixture of D,L-biotin at 37°C for 1 hr
prior to addition to the
assay plate.
After washing the assay plates as above, the pre-incubated samples were then
added to the blocked ELISA plate wells and incubated for 1 hr at 37°C.
5 After the sample incubation, the plates were washed again and a 100 NI of
secondary conjugate was added for 1 hr at 37°C. The conjugate used was
a donkey-
anti-sheep antibody conjugated to alkaline phosphatase (A5187, Sigma Chemical,
Poole, UK) diluted at 1:5,000 in 1% BSA in PBS at pH 7.2.
After conjugate incubation, the plates were washed as above and 100 NI of
PNPP substrate (N-2770, Sigma Chemical, Poole, UK) added to each well and
incubated for 30 minutes at 37°C. The plates were then read at 405 nm
in a microtitre
plate reader.
The results were as follows:
OD D-biotin ~ D,L-biotin
5 NM biotin(s) 0.144 1.139
1.25 NM biotin(s) 0.278 1.639
0.3125 NM biotin(s) 0.576 1.841
zero biotins 2.132 2.213
It can be seen that the pure D-biotin is more effective at inhibiting the
binding
of the anti D-biotin monoclonal antibody to immobilised cytochrome C-D-biotin
than the
racemic mixture that contains both D and L biotins. This confirms that the
binding of
the antibody to L-biotin is reduced or absent in solution.