FUSION PROTEINS FOR HALF-LIFE EXTENSION - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 98

erythro poietin (Arane sp 1 ) that had a thr eefold increased
half-life compared to the wild-ty pe molecu le [65]. Besid es
the artific ial introdu ction of g lycosylation sites into the
protein of intere st, a similar effect can be achi eved by adding
a natu rally occur ring peptide sequence that contain s mul ti-
ple O- linked glycosl ylation sites [66] . The molecu lar mass
of the glycosyl ated peptide is 8 kDa instea d of 2.9 kDa for
the peptide without glycans. Although the natu ral pept ide is
derived from the carb oxyl terminal domain (CTP) of chori -
onic gonadot ropin (CG) ; it can als o be positioned at the
amino-te rminus. Even larger molecu les can be designed that
contain repe ats of this peptide . Since its first descr iption
CTP has been used with several ther apeutic prot eins. Fo r
instance, hGH was tested in a numb er of configur ations with
CTP. The best variant of hGH with CTP at both termin i was
fourfold more potent than the hGH alone and showed a
drastically mor e than 10-fold prolonged half-life [67] . EPO
equipped with three copies of CTP had eightf old improved
potency and threefold extended half-lif e. Intere stingly, it
was even superior to the hyperglycosylated EPO variant
Aranes p 1 [68]. The CTP concep t is explained in more detail
in Chapte r 13.
A more mas sive glycosyl ation can be obta ined usin g the
hydrox yproline (Hy p)-Glyco technol ogy. Here the recombi -
nant prot ein is expressed as fusion to the Hyp-rich repe titive
peptide (HypR P) tag that dir ects the formati on of proline
hydrox ylation, and subsequent Hyp-O -glycosy lati on. This is
a plan t-specific post translationa l modifi cation. Typically
tandem repe ats of Ser-Hyp or Ala-Hyp are used since
they lead to highly branched hete ropolysa ccharid es with
15-150 residue s, similar to the glyc osylation that can be
found in many plant gums . When testi ng this method with
IFN- a 2b, first the secr etion from plant cells was drastical ly
improved up to 1500-fol d and secon d the in vivo half-l ife in
mice was extended 13-f old [69]. Sim ilar results were
obtaine d with hGH as fusion part ner. The 10 Ser-Hyp
repeats at the C-terminu s of hGH with a corr espond ing
glycosyl ation of 25 residues in averag e incr eased the circu-
lation time more than 6-fold and the secret ion from plan t
cells 500-f old [70]. The downside of this appro ach is the
need to express in plant cells, which is not that widely
establish ed and the risk of hete rogeneous glyc osylation,
which might be a regulatory hu rdle.
A more versat ile glyc osylation that works in any eukar y-
otic organism can be obta ined using Genetic Polymers TM
that cont ain the standar d N-linked glycosylat ion motif Asn -
Xaa-Ser/Th r with Xaa being any amino acid except proline.
The rest of the peptide repeats consist s primari ly of G, N, Q
with mi nor subst itutions of A, S, T, D, E. Th e repeats can be
amino acid tripl ets up to sextets. Multiple (2-500 ) pept ide
repeats together with their correspo nding additional glyc o-
sylation wi ll obviously incr ease the hydrod ynamic radius of
any protei n molecu le fuse d to it, thus extendi ng the in v ivo
half-life. Testing this tec hnology with 155 tripl et repe ats of

NNT in conjunct ion to the carboxy terminus of G- CSF
improved the circul ation half-lif e in mice fourfol d (http://
www.aequusbi ophar ma.com). It has to be taken into
accou nt, that excessive N-linked glycosyl ation can give
rise to high mannose cont ent that might be a target for
manno se recep tors to elim inate prot eins relatively fast from
the blood stream .

6.4 AGGREG ATE FORMI NG PEPTIDE FUSION S

Other naturally occur ring pept ide repe at seque nces can be
found in gelatine and elastin. Th erefore, it was quite logical
to test these proteins als o for their ability to extend half-lif e.
The examples using ela stin-like pept ides (ELP) are q uite
numerous , and initially this sequence has primari ly been
used as tool for downs tream proce ssing becau se o f its
unusua l ability to form aggregates at high temper atures
and becomi ng a solub le monome r at lower temper atures.
This phenomeno n is cal led reversible therm al phase transi -
tion and the critica l temperatu re for the p hase transi tion is
depend ent on the leng th of the polymer. The longe r the
polymer, the highe r is the transi tion temperat ure [71]. Th e
pentamer repeat motif is VPGXG and typically 90-120
repeats are used. With the exception of proline any amino
acid can be incorporated at the X position. Currently, a
vasoac tive pept ide and GLP- 1 are in clini cal trials (http://
www.phasebio .com). Both molecu les form d epots in the
body from whi ch active molecu les are slowly dissoci ating.
Therefore, the mechanism that leads to half-life extension is
very different to other peptide repeats that increase the
hydrodynamic radius. Unfortunately, in the case of interleu-
kin-1 receptor antagonist (IL-1Ra) ELP fusion drastically
reduced the bioactivity more than 500-fold [72]. This could
be an obstacle for therapeutic applications. But recently an
interesting approach combining ELP with cell penetrating
peptides (CPP) and hyperthermia was discussed. Here the
tumor is artificially heated up above 40 C. At this tempera-
ture, ELP can aggregate, thus achieving a high local con-
centration of therapeutic agents. The combination of CPP,
ELP, and anticancer peptides in a single genetic fusion can
have a synergistic therapeutic effect [73]. A detailed descrip-
tion of the ELP technology can be found in Chapter 14.
The other natural peptide repeat gelatine has only
recently been used as fusion partner for G-CSF. Repeats
of the amino acids G, P, E, Q, N, S, and K resulted in
hydrophilic 116 residue long gelatine-like protein (GLK).
This fusion protein had a more than fivefold improved half-
life. Owing to its hydrophilic nature no, aggregates could be
found. Interestingly, the mechanism for half-life extension
relies not necessarily only on the increase of the hydro-
dynamic radius, but rather on the additive effect of shielding
G-CSF against proteolysis and electrostatic repulsion due to
its negative charge [74].