Biomedical Engineering Reference
In-Depth Information
with LTBP is not required for latency [31] and it is obviously
not desirable for the latent therapeutic cytokine to be
tethered to the ECM. It has previously been demonstrated
that endogenous LTBP from Chinese hamster ovary cells
does interact with recombinant LAP expressed in these cells
[7]. To prevent this interaction, the cysteine at position 33 in
the LAP is mutated to a serine residue to ensure that the
latent cytokine does not interact with LTBP during expres-
sion. This mutation also prevents the formation of spurious
disulfide bonds between this cysteine and any cysteines that
may be present in the therapeutic cytokine of interest, or
indeed, between LAP molecules. The MMP cleavage site is
flanked at either side by a flexible glycine/serine linker to
provide a hydrophilic area that will facilitate access to
enzymatic attack.
at these sites has also been demonstrated by treatment of
recombinant LAP-TGF- b with neuraminidase [41]. The
degree of sialylation of native latent TGF- b has a critical
bearing on the pharmacokinetics of this molecule. Recom-
binant LAP-TGF- b that is under-sialylated has a half-life
over 100-times shorter than fully sialylated protein [43]. The
nature and extent of glycosylation also determines the tissue
distribution of recombinant LAP-TGF- b , with most under-
sialylated material found in the liver, as opposed to the
sialylated protein which is found predominantly in the
circulation [43]. It was initially believed that glycosylation
of the LAP may be important in TGF- b latency, at least in
vitro [44], as enzymatic removal of the carbohydrate struc-
tures from the LAP-TGF- b complex produced active TGF- b
from the latent complex. However, subsequent studies have
shown conclusively that unglycosylated recombinant LAP
produced in inclusion bodies in E. coli and re-folded in vitro,
can associate with mature TFG- b and render it latent [45].
The role of these carbohydrate moieties of LAP in main-
taining latency in vivo has not been investigated.
As one of the advantages of the latent cytokine technol-
ogy is that the half-life of the latent form is significantly
longer than that of the active cytokine, producing poorly
glycosylated latent cytokines actually defeats the purpose of
this approach. Hence, the choice of expression system is of
the utmost importance. In our experience, Chinese hamster
ovary (CHO) cells, grown either as adherent cells or in
suspension, are more adept at producing recombinant pro-
teins that are sufficiently sialylated than other mammalian
cell lines such as human embryonic kidney (HEK) cells [46].
16.4.2 Expression of LAP-Fusion Proteins
Expression of the recombinant fusion protein should be
performed in mammalian cells because of their ability to
correctly fold, assemble, glycosylate, and secrete the protein.
This is particularly important for the production of these latent
cytokines, given that the structure of the LAP used to confer
latency. The LAP is a homodimer formed by two disulfide
bonds formed between two cysteine residues (Cys 223 and
Cys 225 ; Figure 16.4) on eachmonomer. The formation of both
of these disulfide bonds is necessary for the dimerization of
the LAP monomers and consequent latency. Substitution of
one or both of these cysteine residues with serine results in
loss of latency [32]. If dimerization does not occur, the LAP
shell is not formed; the cytokine is not surrounded by the LAP,
and is biologically active. Hence, the correct formation of
these disulfide bonds, and consequent dimerization of the
LAP-cytokine molecules is absolutely critical for the produc-
tion of latent, rather than active, cytokine.
The nature and extent of glycosylation of the recombinant
protein are also essential for the production of these latent
cytokines and is one of the main reasons why expression of
these proteins in nonmammalian systems such as Escher-
ichia coli or insect cell lines is not advisable. Under-glyco-
sylation is known to affect the folding and assembly of many
proteins [33,34] and can affect the yield of correctly folded
protein quite dramatically [35]. Indeed, the carbohydrate
structures of LAP have been shown to be important in the
secretion of latent TGF- b from cells [36,37,38]. However, a
more pressing reason for ensuring the correct levels of
glycosylation of the recombinant protein is that this has
profound implications for the half-life of the protein in vivo.
Recombinant proteins that are incorrectly or under-glyco-
sylated are cleared from the circulation much more quickly
than correctly glycosylated proteins [39,40]. LAP contains
three N-linked glycosylation sites, all three of which are
glycosylated [41] and two of which contain mannose-6-
phosphate residues [42]. The presence of sialic acid residues
16.4.3 Purification of LAP-Fusion Proteins
As is the case for the purification of any recombinant protein
from the supernatant of cell cultures, there are a number of
approaches that may be taken for the purification of recom-
binant cytokines. Latent cytokines can be purified from cell
culture supernatants using conventional chromatography
techniques, all of which are based on using the LAP part
of the molecule, rather than the cytokine part of the mole-
cule. Although other chromatographic methods may be
available depending on the cytokine of choice, for example,
IFN- b can be purified using blue sepharose, any such
method would depend on the ability of the cytokine to
access the chromatography medium in the presence of the
LAP. To date, methods of purification based on the cytokine
part of the molecule have not been tested.
There are a number of methods that can be used to purify
latent cytokines based on the LAP part of the molecule.
Affinity chromatography using a commercially available
antibody against human LAP immobilized on sepharose
beads was used very effectively by Adams et al. [7].
However, this protocol is costly and is not very easily
scaled up to deal with large volumes of supernatants.
Search WWH ::




Custom Search