FUSION PROTEINS WITH A TARGETING FUNCTION - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 373

that utilizes the presence of scavenger receptor class B type I
(SR-BI) on liver cells. This receptor binds high-density
lipoproteins (HDLs) with their main component apo-
lipoprotein A-I (ApoA-I). Combining ApoA-I with IFN- a
in a fusion protein solves two issues simultaneously. It
provides liver targeting and prolongs the half-life of the
relatively small IFN- a . The circulation of this bifunctional
protein is shorter than an albumin fusion protein, but ApoA-
I-IFN- a is more potent and less toxic than IFN- a alone [30].
A more detailed description of this concept can be found in
Chapter 29.

corresponding dimer, but still 2.3-fold less potent than the
original IFN- b . However, the presence of the Fc-part
improved the half-life threefold [36]. Besides IFN- b ,
IFN- a also was successfully tested as monomeric Fc-fusion
protein for pulmonary delivery [37]. Monomeric Fc-fusion
proteins are discussed in Chapter 7.
In some instances, lung epithelia might not be accessible
from the airway. For example, in cystic fibrosis (CF) a thick
mucus layer covers the lung surface. In this case, delivering
therapeutic proteins is almost impossible by inhalation.
Therefore, a novel strategy was designed that reaches the
lung epithelia through the polymeric immunoglobulin recep-
tor (pIgR). This receptor is present on the basolateral surface
of epithelial cells. There it internalizes and transfers poly-
meric antibodies (IgA and IgM) to the apical surface of the
cell where the antibodies are released to the lumen. This
transfer is essentially irreversible because the secretory
component (SC) to which the polymeric Igs are attached
to is cleaved from the receptor [38].
Many CF patients suffer from bacterial infections, to
which the body reacts with the release of neutrophil elastase
(NE), a protease that destroys Pseudomonas, but also elastin
in the airways. Usually, NE is inhibited by a 1 -antitrypsin
(A1AT), but the natural secretion is insufficient to counteract
the massive NE release in CF. Therefore, A1ATwas fused to
a single-chain Fv antibody directed against the SC of the
pIgR. In a human tracheal xenograft model, which resem-
bles human lung epithelia on nude mice, it could be dem-
onstrated that the SC-specifc Fv-A1AT fusion protein was
successfully transferred from the blood stream to the apical
lumen of lung epithelia. It was found that the fusion protein
was sevenfold more enriched in the lumen than A1AT
without the receptor-specific Fv. In general, the transfer
efficiency was more dependent on the expression level of
pIgR than on the concentration of the fusion protein [39].
Immunoglobulin-based transport in the lung is summarized
in Figure 25.2.
The upper airways and the nasal epithelia as entry point for
viral infections have been utilized as target for a novel concept
to protect from influenza. Influenza A and B viruses bind with
their hemagglutinin-neuraminidase to surface sialic acids.
The prototypic N-acetylneuraminic acid (Neu5Ac) is usually
connected to the penultimate galactose residues either by
a (2,3)- or a (2,6)-linkages. Particularly, the Neu5Ac- a (2,6)-
Gal that can be found on respiratory epithelia is the preferred
receptor for human viruses. It has been demonstrated that the
absence of this specific glycosylation drastically reduces the
infection efficiency and protects from influenza viruses.
Therefore, a fusion protein was designed that contained an
exoglycosidase connected to a cell surface anchoring
sequence. As broad-spectrum glycosidase, the catalytic
domain of a sialidase from Actinomyces viscosus was
selected, whereas the targeting domain was derived from
the human protein amphiregulin. This 20 amino acid long

25.2.3 Lung

The lung as organ with direct contact to the environment
could be a preferable area for targeted therapies. In the past
years, inhalable insulin was in the public focus [31]. Addi-
tionally, studies of pulmonary uptake of somatropin became
available [32]. But both molecules are rather small, and not
as complicated to deliver as larger proteins.
Similar to transport across the BBB, the TfR has also
been evaluated for the epithelial transcytosis of large bio-
molecules in the lung. In a first attempt, granulocyte colony-
stimulating factor (G-CSF) was conjugated to Tf, and a
threefold higher apical-to-basolateral transcytosis was
demonstrated in primary cultured rat alveolar epithelial
cell monolayers for the conjugate, compared to G-CSF
alone [33].
The best-studied approach for pulmonary delivery of
proteins relies on the presence of the FcRn in lung epithelia.
This receptor can also be found in the kidney and intestine.
There it is responsible for the uptake of maternal immuno-
globulin from the milk and the transfer to the bloodstream.
However, the receptor is still present in adult humans, too.
On the basis of this mechanism, the FcRn can be used as
entry point for pulmonary delivery. The FcRn is expressed
ubiquitously in alveolar epithelia cells, but is higher
expressed in the upper and central airways, the bronchi.
Most of the Fc-fusion proteins that so far have been tested
for inhaled administration were monomers with one of the
heavy chains without a therapeutic protein attached to it.
Monomeric erythropoietin (EPO) Fc fusion had a twofold
better affinity to the FcRn and a 30% longer half-life than the
dimeric variant. Compared to the original EPO, monomeric
EPO-Fc achieved a much higher systemic concentration
when administered at equimolar concentrations in primates
[34]. The better distribution of the monomer might be an
effect of smaller size and better transport, less steric hin-
drance, and lower tendency to adhere to mucosal surfaces
owing to only one glycosylated EPO molecule. In a Phase I
clinical study, an EPO-Fc dimer reached 70% bioavailability
in the central lung regions [35]. Recently, an interferon- b
(IFN- b ) monomeric Fc fusion was tested for inhalation in
primates. The monomer was fourfold more active than the