BONE-TARGETED ALKALINE PHOSPHATASE - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 420

extravesicular mineralization [53-55]. Both of those issues
have been resolved with the design of bone-targeted alkaline
phosphatase, ENB-0040, which has a half-life of 34 h in
plasma of adult mice and more than 300 h in bone tissue.
Nevertheless, currently optimized treatment requires
repeated administration of large amounts of enzyme for
long-term correction. In preclinical experiments, mice
received daily injections of ENB-0040 and in the initial
clinical trials, HPP patients received SC injections of bone-
targeted TNAP three times a week [37-46].
We have recently shown that a single injection of a
lentiviral vector harboring TNAP-D 10 , that is, soluble
TNAP linked directly to D 10 , resulted in sustained ALP
expression and phenotypic correction in Akp2 / neonatal
mice [56], including prevention of epileptic seizures and
skeletal defects and preservation of a normal life-span. In
another recent study, we demonstrated that a single IV
injection of neonatal Akp2 / mice with an adeno-associ-
ated viral vector expressing TNAP-D 10 was as effective as
the lentiviral vector treatment [57]. In both of these viral
approaches, the Fc region was not included in the construct
because neither purification of the enzyme nor an increase in
plasma half-life was required. Indeed, these viral delivery
systems utilize either the liver (lentiviral vector) or skeletal
muscle (adeno-associated viral vector) as factories of bone-
targeted TNAP for the life of the animal. These delivery
systems show promise as an alternative means of delivery of
bone-targeted TNAP that would decrease treatment fre-
quency and greatly reduce treatment cost.
Lentiviral-mediated gene transfer has proven effective
for long-term expression of transgenes in nondividing
cells. Thus, we used an HIV-1-based lentiviral vector
for our study [58]. Although HIV-1 pathogenicity is almost
negligible in the current modified version of the lentiviral
vector, insertional mutagenesis remains a major concern
when using integrating vectors [59]. To minimize the
possibility of proto-oncogene activation, our self-inacti-
vating lentiviral vector contains the insulator element from
the chicken b -globin locus [60]. Nevertheless, the onco-
genic potential of integrated lentiviral vectors in differen-
tiated tissues requires further examination in a long-term,
follow-up study. Integrating lentiviral vectors have been
used for stable gene transfer into hematopoietic cells in
ex-vivo protocols [61,62], but their application for in vivo
gene transfer is limited [63], and there are no clinical
protocols for systemic lentivirus injection. In contrast,
AAV vectors are suitable for in vivo gene delivery and
have been widely used for enzyme replacement therapies
for genetic diseases, such as lysosomal storage disorders
and hemophilia [64-67]. Our data suggest that systemic
delivery of an AAV vector is a promising and realistic
approach to treat HPP. Furthermore, using the AAV8
vector, we confirmed that TNAP-D 10 had a higher affinity
for bone than TNAP lacking the D 10 moiety both in vitro

and in vivo [57]. However, we also demonstrated that
TNAP lacking D 10 could be effective to treat HPP mice,
indicating that sustained TNAP production may be the
most critical component of the treatment, more so than
bone-targeting. This issue merits further investigation, as it
is important not only for drug design but also to elucidate
the pathophysiology of HPP and other mineralization
disorders. We must clearly determine whether it is local
or systemic concentrations of PP i that influence skeletal
mineralization, as well as medial vascular calcification, a
disease caused by upregulated TNAP levels in vascular
smooth muscle cells [35,36].
To conclude, while viral delivery represents a powerful
research tool, viral vector safety remains a concern and gene
therapy remains a distant promise. In contrast, ERT has a
very good track record of safety and efficacy and is readily
approved by regulators.

28.6 CHALLENGES IN PRODUCTION AND
DEVELOPMENT

While there are many advantages to combining distinct
functional domains into a single fusion protein, as discussed
extensively in this topic, there are also unique challenges
related to fusion protein production and purification. Issues
related to ENB-0040 purification are amply discussed in an
article by Dr. Thomas P. Loisel, Group Leader at Enobia
Pharma Inc [30]. Dr. Loisel indicated that ENB-0040 did not
fit well into a platform process, even though it possessed
three potential affinity sites and each target site had a
platform purification unit operation. One site was the anti-
body Fc region for Protein A binding, but the product was
too sensitive to low pH for typical Protein A elution condi-
tions (a pH

3.7 to inactivate
viruses). Therefore, a high pH elution process (pH

4 typically for elution and

< <

11) was
developed as an alternative. The other two affinity sites
(catalytic site and polyaspartate sequence) proved to be
unusable, as did ion exchange chromatography. The final
process developed for ENB-0040 purification added hydro-
phobic interaction chromatography and a mixed-mode resin
with hydrophobic and anion exchange characteristics along
with solvent and detergent treatment to replace low pH virus
inactivation [30].

¼

28.7 CONCLUSIONS AND FUTURE
PERSPECTIVES

To date, there is compelling evidence that ENB-0040
treatment is an efficacious, well-tolerated ERT in HPP
mice and preliminary information indicates a similar effect
in infants and juveniles with moderate to severe HPP. The
success of ERT for HPP using bone-targeted alkaline