Antibody Molecules, Genetic Engineering of
341
advantage in certain applications such
as tumor targeting for detection and/or
therapeutic purposes. In fact, scFv frag-
ments display a combination of rapid,
high-level tumor targeting with concomi-
tant clearance from normal tissues and
circulation that have made radiolabeled
scFvs important tools for detection and
treatment of cancer metastasis in both pre-
clinical models and patients. An example
of a successful scFv in the clinic is MFE-23.
This scFv is speci±c for carcinoembry-
onic antigen (CEA), a glycoprotein that is
highly expressed in colorectal adenocarci-
nomas. MFE-23, expressed in bacteria, has
been used in two clinical trials: a gamma
camera imaging trial using
123
I-MFE-23
and a radioimmunoguided surgery trial
using
125
I-MFE-23, in which tumor de-
posits are detected by a handheld probe
during surgery. Both trials showed that
MFE-23 is safe and effective in localizing
tumor deposits in patients with cancer.
Despite their successful use in some
tumor-targeting studies, a signi±cant lim-
itation of using scFvs for targeting
in vivo
is their monovalent binding to antigen.
Intact antibodies have signi±cant avidity
as a result of the presence of two antigen
binding sites. To address this problem, an-
tibody engineers have used the monomeric
scFv as a building block for larger engi-
neered fragments. One approach consists
of producing dimers of scFv by incorpo-
rating a carboxy-terminal cysteine residue
so that a disul±de bridge forms, yielding
(scFv)
2
fragments. Alternately, the single-
chain concept has been extended by using
an additional linker peptide to join the two
scFv molecules in tandem. Another ap-
proach for the production of scFv dimers
results from the observation that the use of
a very short linker peptide to connect the
antibody variable regions caused the for-
mation of ‘‘cross-paired’’ dimers, in which
the V
L
of one molecule associates with the
V
H
of a second and the V
L
of the second
molecule associates with the V
H
of the
±rst (Fig. 6). These noncovalent dimers,
also known as
diabodies
,arecapab
leofb
i
-
valent binding to antigen. scFv dimers and
diabodies have a molecular weight simi-
lar to that of the antibody Fab fragments
(55–60 kDa) but contain two antigen bind-
ing sites. Diabodies show signi±cant im-
provement in tumor targeting compared
to monovalent scFv. scFv fragments can
also be fused to an immunoglobulin C
H
3
domain, resulting in a self-assembling bi-
valent ‘‘minibody’’ (Fig. 6).
To further increase the avidity of anti-
body fragments, several laboratories have
generated fragments with an increased va-
lence. One strategy to develop fragments
with increased valence has been to ex-
tend the diabody approach by decreasing
the length of the interdomain linker pep-
tide, which may result in the formation
of tribodies and tetrabodies. It is also
possible to obtain larger bivalent or multi-
valent fragments by fusing scFvs to protein
domains normally involved in protein as-
sociation such as helix bundles or leucine
zippers. An alternative approach for pro-
ducing multivalent fragments has been
the fusion of scFvs to the bacterial pro-
tein streptavidin. Since streptavidin is a
tetramer composed of four noncovalently
linked monomers, four scFvs assemble
to form a tetrameric structure with four
antigen binding sites. A practical example
of this technology was the development
of Rh-speci±c scFv fused to streptavidin.
This fusion protein named scFv::strep was
able to directly agglutinate antigen-positive
red blood cells (a reaction that is impos-
sible to achieve by using antigen-speci±c
monomeric scFv or IgG), suggesting the
potential use of scFv::strep as a blood-
typing reagent.
previous page 341 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online next page 343 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online Home Toggle text on/off