FUSION PROTEINS WITH TOXIC ACTIVITY - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 252

typically require both a free N- or C-terminus for activity
and a flexible linker connecting the enzyme to the targeting
domain. This has to be taken into account when designing
the fusion protein [46]. Potency could be enhanced
5-40-fold by combining two RNAse molecules with a single
antibody moiety [47]. The inefficient production and, some-
times, low potency could delay the introduction of immu-
noRNAses [48]. A complete summary of the current state of
the art of ImmunoRNAses is given in Chapter 22.

cell-targeting motifs to form a novel class of human immu-
notoxins [56].

17.4.4 Antibody-Directed Enzyme Prodrug Therapy

For decades, small molecules have been used in chemo-
therapy. However, these highly potent drugs cause massive,
sometimes dose limiting, side effects and tumor cells
develop resistance. Therefore, strategies are desired that
limit the distribution of the drug but not the toxicity.
Specificity both for the drug and the targeted cell is needed.
This goal can be fulfilled by combining the specific activity
of an enzyme to activate a nontoxic small molecule with the
high target specificity of an antibody. So, a new approach
was born [57], and called ADEPT [58]. This approach
consists of two elements: an antibody-enzyme fusion and
a prodrug. ADEPT uncouples binding from killing to
increase specificity. Both the large fusion protein and the
small prodrug alone are nontoxic, but they generate cyto-
toxicity when co-localized. Nowadays no longer full anti-
bodies are used; a single scFv or a target specific ligand is
sufficient to enable binding to a surface receptor. Mono-
valency is even preferred, since cross-linking of receptors by
a bivalent binder could lead to internalization, thus removing
the required enzymatic activity from the surface.
The whole procedure comprises three steps. First, the
antibody-coupled enzyme is injected. After completion of
administration, unbound enzyme fusion proteins have to be
removed from the organism to focus enzymatic activity only
on targeted cells. This can be achieved either with small
constructs with a short half-life or through phagocytosis if
the fusion protein is mannosylated [59]. The clearance
should be fast to have almost maximum concentration of
the tumor bound enzyme. After removal of unbound enzyme
finally the prodrug is added. This prodrug is metabolized by
the enzyme, thus generating a high local concentration of the
active drug molecule in proximity of malignant cells.
ADEPT is the only internalization independent enzyme
strategy [60].
As additional level of safety, preferably nonhuman
enzymes such as cytosine deaminase and b -lactamase are
used to avoid unspecific drug activation by endogenous
enzymatic activity. But these foreign proteins can sometimes
cause an immune response. In the case of b -lactamase,
mutagenesis reduced immunogenicity fivefold [61]. If
enzymes of human origin such as b -glucuronidase are
utilized, a prodrug formulation must be selected that cannot
reach the intracellular enzyme. Alternatively, the substrate
specificity of a human enzyme can be modified by muta-
genesis to introduce prodrugs that cannot be metabolized by
the native enzyme. Besides the short half-life of unbound
enzyme fusions, the generated drug should also have a short
circulation time to avoid diffusion to healthy tissue. The
diffusion radius defines the bystander effect that kills even

17.4.3 Proteases

Other mechanisms to disturb cellular functions rely on
proteolytic activities. Some proteases are key enzymes
that mediate apoptosis, such as the members of the caspase
family. In particular, a constitutively active caspase 3 has
been fused to various targeting moieties such as a single-
chain anti-HER2 antibody for instance. To enable the escape
from the endosome, this molecule is also equipped with a
translocation domain based on Pseudomonas exotoxin A.
Jurkat cells expressing and secreting this construct remained
healthy, whereas co-cultivated HER2-positive cells such as
SKBR-3 or SKOV-3 were selectively killed [49]. Another
construct containing caspase 3 was directed against immune
cells expressing the interleukin-2 receptor (IL-2R). The
IL-2-caspase 3 fusion protein eliminated regulatory T-cells
responsible for autoimmunity in a diabetes mouse model
[50]. A second example was the successful treatment of
experimental inflammatory colitis in mouse with the same
construct [51].
The lamin A cleaving, apoptotic function of caspase 6
was utilized to treat HER2 positive osteosarcoma lung
metastasis in mice. As before, domain II of Pseudomonas
exotoxin A was included again, and the therapy was per-
formed by genetic transduction [52]. The DNA construct
coding for constitutively active caspase 6 fused to a single-
chain anti-HER2 antibody was injected into BALB/c
athymic mice bearing human breast SK-BR-3 tumors.
The survival of all

treated animals was significantly

prolonged [53].
Another gene therapeutic example for targeted cell kill-
ing is the granzyme B (GrB) fusion protein directed against
HER2. This protease can initiate apoptosis through multiple
mechanisms. As described before, the same single-chain
antibody and translocation domain was used. After trans-
fection, the expression of this DNA construct selectively
destroyed HER2-positive tumor cells in nude mice. Contin-
uous secretion of the fusion protein is responsible for a
substantially prolonged survival of these animals [54]. GrB
serves also as the cell-killing moiety for a single-chain anti-
melanoma antibody. Melanoma cells were selectively killed
through apoptosis when exposed to the fusion protein [55].
Owing to the low immunogenicity of human enzymes, Gr B
and other proteases have now been combined with many