Note: Descriptions are shown in the official language in which they were submitted.
CA 02450795 2003-12-19
WO 96!40784 PCT/US96/09980
lVdETHOI) OF SOL~.T~yLiZII~TG, P~T>I~ING, A1~TD REFOLDII~TG PIROTEIN
Technical Field of the Invention
The invention :relates to methods useful for refolding, solubilizing,
formulating
and purifying proteins. These methods are particularly usefi~l for proteins
that; have been
engineered by genetic x ecombination at~d produced irl bacterial, yeast or
other cells in a
form that has a non-native tertiary structure.
Background of the Invention .
To understand fully the entire process of g~rne expression, it is as
irrvportant to
understand the proves>~ for the folding of the peptide chain into a
biologically active
protein as it is important to understand the synthesis of the primary
sequt;nce. The
biological activities of proteins depend not only on their amino acid
sequences but also
on the discrete conformations of the proteins concerned, and slight
disturbances to the
conformational integrioy of a protein can destroy its activity. Tsou et ~xl
(1988)
Biochemistry 27:1809- l 812.
Under the proper conditions, the in vitr~ refolding of purified, denatured
proteins
to achieve the native secondary and tertiary structure is a spontaneous
process. To
avoid formation of stable, but undesired, structures, it is necessary to use
the tertiary
interactions (which are formed late in folding) with their high degree of
selectivity power
to select and further stabilize those early local structures that are on the
correct folding
pathway. Thus, the finite, but very low, stability of local structures could
be the kinetic
"proofreading" mechanism of protein folding. The activated state of folding
with the
highest energy is a distorted form of the native protein, and the slowest,
rate-Limiting
step of unfolding and refc'Iding appears to be close to the native state in
learns of ordered
structure. In addition, the refolding of many proteins is not completely
reversible in
vitro, and reactivation yields of less than 100% are frequently observed,
which holds true
in particular for experiments at high protein concentration, and competing
aggregation
of unfolded or partially refolded protein molecules may be the major reason
for a
CA 02450795 2003-12-19
i~VO 96/4074 PC'i'/TJS96/099~0
_2_
lowered reversibility, as described in Fischer and Schn~id, (1990)
~iocherttisiry 29:2205-
2212.
In the case of suciently Large protein rnolec~ales, the nascent polypept:ide
chain
acquires its native three-dimensional structure by the modular assembly of
rnicro-
domains. Variables including temperature, and cosolvents such as
polyols,'.~rea, and
guanidiniuchloride, have been tested to dete~ine their role in stabilizing and
destabilizing protevl conformations. The action of eos~olvents may be the
result of direct
binding or the alterations of the physical properties of water, as described
in Jaenicke et
cal. ( 1991 ) Biocher,~is3 0 ( 13 ):3147-3 I 61.
Experimental observations of how unfolded proteins uefold to their native
three-
dimensional structures contrast with many popular theories of protein folding
mechanisms. Under conditions which allow for refolding, unfolded protein
rr'olecules
rapidly equilibrate between different conformafiions ~~rior to complete
refolding. The
rapid prefolding equilibrium favors certain compact conformations that have
somewhat
lower free energies than the other unfolded conformations. The rate-limiting
step occurs
late in the pathway and involves a high-energy, distorted form of thve native
conformation. There appears to be a single transition through which
essentially all
molecules fold, as described in Creighton et al. (1!88) Pa~oc. Nat. ~cac~
S'ci. I,S~4
85:5082-5086.
Various methods of refolding of purified, recombinantly produced proteins have
been used. For example., the protease encoded by th,e human immunodeficier~cy
virus
type I -I) can be produced in ~"sche~°ichia coli, yielding inclusion
bodies harboring
the recombinant 1'3:IV-I protease as described by Hiu:i ~~ al. (1993) J
~'t°ot. Chem.l2:
volumes of cold 0.1 IvI sodium acetate, pI-I 5.5, contaiong 5~/o ethylene
glycol and 10%
glycerol. Exclusion of glycerol and ethylene glycol ledl to gradual loss of
protein due to
precipitation. About 85 ang of correctly folded I-IIV-a protease per liter
of.E toll cell
culture was obtained by this method, and the enzyme ha.d a high specific
activity.
CA 02450795 2003-12-19
-3-
Another example of refolding a recombinant protein is the isolation and
refolding of H-ras from inclusion bodies of E. coli as described by DeLoskey
et al.,
(1994) Arch. Biochetn. and Biophys. 311:72-78. In this study, protein
concentration,
temperature, and the presence of 10% glycerol were varied during refolding.
The
yield of correctly folded Ii-ras was highest when the protein was refolded at
concentrations less than or equal 0.1 mg/ml and was independent of the
presence of
10% glycerol. The yield was slightly higher at 4° than at 25°C.
The refolding of Tissue Factor Pathway Inhibitor (also known variously as
Lipoprotein-Associated Coagulation Inhibitor (LACI), Extrinsic Pathway
Inhibitor
(EPI) and Tissue Factor Inhibitor (EFI) and hereinafter referred to as "TFPI")
produced in a bacterial expression system has been described by Gustaf'son et
al.,
(1994) Protein Expression and Purification 5: 23~-241. In this study, high
level
expression of TFPI in recombinant E. coil resulted in the accumulation of TFPI
in
inclusion bodies. Active protein was produced 'by solubiization of the
inclusion
bodies in 8M urea, and purification of the fill-length molecule was achieved.
by canon
exchange chromatograp=by and renaturation in 6M urea. The refolded mixture was
then fractionated to yield a purifaed nonglycosylated TFPI possessing in vitro
biological activity as measure in the Prothombin clotting time assay
comparable to
TFPI purified from mammalian cells.
A non-glycosyla~4ed form of TFPI has also been produced and isolated from
Escherichia coli (E. coil) cells as disclosed in Lt.S. Patent No. 5,212,091.
The
invention described in IJ.S. Patent No. 5,212,091 subjected the inclusion
bodies
containing TFPI to sulfitolysis to form TFPI-S-sulfonate, purified TFPI-S-
sulfonate
by anion exchange chromatography, refolded TFPI- S-sulfonate by disulfide
exchange
using cysteine and purified active TFPI by cation exchange chromatography. The
form of TFPI described in U.S. Patent No. 5,212,092 has been shown to be
active in
the inhibition of bovine f~.ctor Xa and in the inhibition of human tissue
factor-induced
coagulation in plasma. In some assays, the E. coil- produced TFPI has been
shown to
be more active than native TFPI derived from SK hepato~na cells. However, TFPI
produced in E. coli cells is modified in ways that increase heterogeneity of
the
protein.
CA 02450795 2003-12-19
WO 96/40784 PCT/US96109980
_Q_
A need exists in the art of refolding recombinantly produced proteins to
increase
the amount of correctly folded TFPI during the refolding process. A need also
exists for
increasing the solubility ~of TFI'I. Presently the yields of recombinantly
produced I
have been lower than desirable, and a need exists in the art of producing
correctly folded
TFPI. See for example Gustafuson et al. ( 1994) P'rotetn Expression and
Pur~cation .5:
233-241.
TFPI inhibits the coagulation cascade in at Ieas~t two ways: preventing
formation
of factor VIIa/tissue factor complex and by binding to the active site of
factor Xa. The
primary sequence of TF'PI, deduced from cT3NA sequence, indicates that the
protein
contains three Kunitz-type enzyme inhibitor domains. The first of these
domains is
required for the inhibition. of the factor VIIaltissue facts~r complex. The
second Kunitz-
type domain is needed for the inhibition of factor Xa. ~'he function of the
third Kunitz-
type domain is unknowns. 'fFPI has no known enzymatic activity and is thought
to
inhibit its protease targets in a stoichiometric manner; namely, binding of
one TFPI
Kunitz-type domain to tl"'e active site of one protease molecule. The carboxy-
ternainal
end of TFPI is believed to have a role in cell surface localization via
heparin binding and
by interaction with pho5pholipid. 'I°FPI is also kruown as Lipoprotein
Associated
Coagulation Inhibitor (I~ACI), Tissue Factor Inhibitor (TFI:), and Extrinsic
Pathway
Inhibitor (EPI).
Mature TFPI is 276 amino acids in length with a negatively charged amino
terminal end and a positively charged carboxy-terminal end. 'T'FPI contains 1S
cysteine
residues and forms 9 disulphide bridges v~rhen correctliy folded. The primary
sequence
also contains three Asn-X-Ser/Thr N-linked glycosylation consensus sites, the
asparagine residues located at positions 145, 195 and 256. The carbohydrate
component of mature I is appro~cimately 30~f~ of the mass of the protein.
however,
data from proteolytic mapping and mass spectral data imply that the
carbohydrate
moieties are heterogeneous. TFPI is also found to be phosphorylated at the
serine
residue in position 2 of the protein to varying degrees. The phosphorylation
does not
appear to affect TFPI function.
TFPI has been isolated from human piasra~a and from human tissue culture cells
including HepG2, Chang liver and SK hepatoma cells. Recombinant TFPI has been
CA 02450795 2003-12-19
W~ 96/40784 PCTNS96/09980
-~-
expressed in mouse C 127 cells, baby hamster kidney cells, Chinese hamster
ovary cells
and human SK hepatoma cells. Recombinant TFPI fi°om the mouse C 12.'~
cells has been
shown in animal models to inhibit tissue-factor induced coagulation.
A non-glycosylated form of recombinant TF'PI has been produced and isolated
fromEscherichia coli (~:. coli) cells as disclosed in U_S. Pat. No. 5,212,091.
This form
of TFPI has been shown to be active in the inhibition of bovine factor Xa and
in the
inhibition of human tissue factor-induced coagulation in plasma. Methods have
also
been disclosed for purification of TFPI from yeast cell culture medium, such
as in
Petersen et al, J.Biol.Chem. 1i~:13344-13351 (1993).
Recently, another protein with a high degree of structural identity to TFPI
has
been identified. Sprecher et al, Proc. Nat. Acad. Sri., USA 91:3353-3357
(1994). The
predicted secondar~r structure of this protein, called TIWI-2, is virtually
identical to TFPI
with 3 Kunitz-type domains, 9 cysteine-cysteine linkages, are acidic amino
terminus and
a basic carboxy-terminal tail. The three Kunitz-type domains of TFPI-2 exhibit
43%,
35% and 53% primary sequence identity with TFPI ICunitz-type domains 1, 2, and
3,
respectively. Recombinant TFPI-2 strongly inhibit:. the arnidolytic activity
of factor
~lIla/tissue factor. By contrast, I-2 is a weak inhibitor of factor Xa
ainidolytic
activity.
TFPI has been shown to prevent mortality in a lethal Escherichra Golf (E.
cola)
septic shock baboon model. Creasey et al, J. Clin. Invest. 91:2850-2860
(1993).
Ad tion of TFPI a9: 6 mglkg body weight shortly after infusion of a lethal
dose of
~ coli resulted in survival in all five TFl?I-treated animals with significant
improvement
in quality of life compared with a mean survival time fir the eve control
animals of 39.9
hours. The administration of TFPI also resulted in significant attenuation of
the
coagulation response, off' various measures of cell ir'jtary and significant
reduction in
pathology normally observed in E coli sepsis target organs, including kidneys"
adrenal
glands, and lungs.
Due to its clot-inhibiting properties, TI~PI may also be used to prevent
thrombosis during microvascular surgery. For example, U.S. 5,276,015 discloses
the
use of TFPI in a method for reducing thrombogenicity of microvascular
anastomoses
CA 02450795 2003-12-19
-6-
wherein TFPI is administered at the site of the microvascular anastomoses
contemporaneously with microvascular reconstruction.
TFPI is a hyd~-oph~bic protein and as such, has very limited solubility in
S aqueous solutions. This limited solubility has made the preparation of
pharmaceutically acceptable formulations of TFPI difficult to manufacture,
especially
for clinical indications which may benefit from administration of high doses
of TFPI.
Thus, a need exists in the art for pharmaceutically acceptable compositions
containing
concentrations of TFPI which can be administered to patients in acceptable
amounts.
Brief Description of the Drawings
Figure 1 is a coomassie stained SDS-PAGE analysis of TFPI peak fractions
from the Phenyl SepharoseTMHIC refolding procedure.
Figure 2 is a plot of the recovery of native TFPI from the HIC column.
Figure 3 is a plot of the recovery of native TFPI from a second HIC column.
Figure 4 is the amino acid sequence of TFPI.
Figure 5 shows the solubility of TFPI at different pH conditions. About 10
mg/mL TFPI in 2M urea was dialyzed against 20 mM acetate, phosphate, citrate,
glycine, L-glutamate and succinate in 1 SO mM NaC 1. The concentration of
remaining soluble TFPI after dialysis was measured by IJ~~ absorbance after
filtering
out the precipitates through 0.22 mm filter units.
Figure 6 shows the solubility of TFPI as a fiznction of concentration of
citrate
in the presence of 10 mM Na phosphate at pH f. TFPI solubiity increases with
increasing concentration of citrate.
Figure 7 shows the solubility of'TFPI as a function of concentration of NaCl.
TFPI solubiity increases with increasing salt concentration, indicating salt
promotes
solubility of TFPI.
Figure 8 shows effect of pH on the stability of TFPI prepared in 10 mM Na
phosphate, 150 mIVI NaCl and 0.005% (wlv) polysorbate-80. Stability samples
containing 1~0 mg/mL TFPI were incubated at 40°C for 20 days. Kinetic
rate
constant for the remaining soluble TFPI was analy~.ed by following decrease of
the
main peak on canon exchange chromatograms.
CA 02450795 2003-12-19
W~ 96/40784 P~:TIUS96109980
_'~_
Figure 9 shows the percentage of remaining soluble TFPI measured by canon
exchange HPLC (A) and remaining active TFPI b;y prothrombin time assay (E) as
a
function of phosphate concentration. The formulation contains 150 mglmL TFPI
prepared in 150 mM l~ZaC1 and 0.005 ~o (w/v) polysorbate-80 at pH 7 with
varying
concentrations of phosphate.
Figure 10 shows loss of soluble TFPI at 40 ° C measured by both
cation-
exchange HPLC (triangle) and prothrombin time assay {circle) for 0.5 mg/rnL
TFPI
formulated in 10 ml~I hla citrate, pH 6 and 1~0 mM ~TaCI.
Figure 11 shows loss of soluble TFPI at 40 ° C measured by both ,
cation-
exchange HPLC (open symbol) and prothrombin time assay (closed symbol) for 0.5
mglmL TFPI formulated in 10 mM ImTa phosphate, pH 6 and either 150 mM ~TaCl
{triangle) or 500 mM I~aCI (circle).
Figure 12 shows loss of soluble TFPI at ~0°C measured by both cation-
exchange HPLC (open symbol) and prothrombin dme assay (closed symbol) for 0.5
mglmL TFPI formulated in 10 mM Na acetate and pH 5.5 containing 150 mM IolaCl
(triangle) or S %'o (w/v) sucrose (square) or 4.5 J m'u~nitol (circle).
Figure 13 shows two non-reducing SDS gels for TFPI formulation s<~mples at
pH 4 to 9 stored at 40°(r for 0 and 20 days.
Figure i4 shows the time course of a polyphosphate-facilitated rhTFPI refold
monitored using SDS PAGE.
Figure 15 shows the absorbance at 2g0 nm during the loading and elution of
the S-Sepharose HP column used to purify rhTFPI from a polyphosphate-
facilitated
refold.
Figure 16 shows SDS PAGE analysis of fracaions collected during elution of
the S-Sepharose HP column used to purify rhTFPI from a polyphosphate-
facilitated
refold. Figure 1 ~' shows the absorbance at 280 nrn during the loading and
elution of the Q-Sepharose IiP column used to purify rhTFPI from a S-Sepharose
pool prepared from a polyphosphate-facilitated refold.
CA 02450795 2003-12-19
WO 96/40784 PC.'TfIJS96/09980
-g-
Figure 18 shows SDS PAGE analysis of fractions collected during elution of
the Q-Sepharose HP column used to purify rhTFPI from a S-Sepharose pool
prepared
from a polyphosphate-facilitated refold.
Figure 19 shows the time course of a polyethyleneimine-facilitated rhTFPI
refold monitored using SDS PAGE.
Figure 20 shows the absorbance at 280 nm during the loading and elution of
the S-Sepharose HP ~ column used to puri;Fy rhTFPI firom a
polyethyleneimine-facilitated refold.
Figure 21 shows SDS PAGE analysis of fractions collected during elution of
the S-Sepharose HP column used to purify rh1 FPI from a polyethyleneimine
facilitated refold.
Figure 22 shows the absorbance at 280 nm duriatg the loading and elution of
the t'~Sepharose HP column used to purify rhTFPI from a S-Sepharose pool
prepared
from a polyethyleneimine-facilitated refold.
Figure 23 shows. SDS PAGE analysis of fractions collected during elution of
the Q-Sepharose HP column used to purify rhTFPI from a S-Sepharose pool
prepared
from a polyethyleneimine-facilitated refold.
Figure 24 shows the canon exchange HPLC analysis of a 0.4.
polyphosphate-facilitate~~ rhTFPI refold in the absence of urea.
Figure 25 shows results of ration exchange HPLC analysis of an evaluation of
different levels of cysteine on a rhTFPI refold in 0.4 l polyphosphate, SO
ml~i Tris in
the absence of urea.
Figure 26 shows the effect of polyphosphate chain length on the course of a
polyphosphate facilitated refold of rhTFPI inclusion bodies as monitored b;~
canon
exchange HPLC.
_ Figure 27 shows the effect of concentration of polyphosphate (Glass I-I) on
the
refolding of rhTFPI fro~r~ inclusion bodies as monitcored by radon exchange
IEiPLC.
Figure 28 shows the ration exchange HPLC analysis of polyethyleneimine
and polyphosphate-facilitated refolding of purified and reduced rhTFPI.
CA 02450795 2003-12-19
_g_
Summary of the Invention
It is an object of an aspect of the present invention to describe a method of
S refolding protein.
It is another object of an aspect of the invention to provide aqueous
formulations of TFPI.
It is another object of an aspect of the invention to provide methods for
modifying a protein's solubility using charged polymers.
It is still another object of an aspect of the present invention to describe a
method of refolding '~: FPI including the steps of adding charged polymers to
a
solution of denatured T:1~'PI prior to allowing the TF'PI to refold
Additionally, it is another object of an aspect of the invention to describe a
method of refolding TFPI including the step of immobilizing charged polymers
on a
1 ~ column and passing a solution of denatured TFPI through the column and
eluting the
refolded TFPI after the ~~efolding has occurred.
It has now been found that solubility of TFPI is strongly dependent on pH and,
surprisingly, that polyanions such as citrate, isocitrate, and sulfate have
profound
solubilizing effects on rCFPI. This finding is surprising in light of the
hydrophobic
nature of TFPI and the hydrophilic character of these counterions. Thus,
citrate,
isocitrate, sulfate as well as other solubilizers described hereinbelow can be
used to
produce pharmaceutically acceptable compositions having TFPI concentrations
sufficient for administration to patients. It has also been shown that other
organic
molecules can act as secondary solubilizers. These secondary solubilizers
include
PEG, sucrose, mannitol, and sorbitol.
The invention re~.ates to pharmaceutically acceptable compositions wherein
TFPI is present in a concentration of more than 0.2 mg/mL solubilizing agents.
The
solubilizing agents may be acetate ion, sodium chloride, citrate ion,
isocitrate ion,
glycine, glutamate, succinate ion, histidine, imidazole and sodium dodecyl
sulfate
(SDS) as well as charged polymers. In some compositions, TFPI may be present
in
concentrations of more than 1 mg/mL and more than 10 mg/mL. The composition
may also have one or more secondary solubilizers. The secondary solubilizer or
solubilizers may be polyethylene glycol (PEG), sucrose, mannitol, or sorbitol.
CA 02450795 2003-12-19
-10-
Finally, the composition may also contain sodium phosphate at a concentration
greater than 20 mM.
Although the solubility of TFPI is quite low between pH 5 and 10, it has been
found that L-arginine can increase the solubility by a factor of 100. The
solubility is
very dependent on the concentration of arginine, as 300 mM is about 30 times
more
effective than 200 mM. Urea also is quite effective in solubilization of TFPI.
Further, it has been found that aggregation of TFPI appears to be the major
degradation route at neutral and basic pH conditions and that fragmentation
occurs at
acidic pH conditions.
It has also been :found that active TFPI monomers can be separated away from
TFPI oligomers that are produced during the process of folding recombinant
TFPI
produced E. coil. Some misfolded/modified mc~nomeric forms of TFPI are also
removed during this process: The separation employs hydrophobic interaction
I5 chromatography. The oligomeric species of TFPI bind more tightly to a
hydrophobic
resin than does the active TFPI monomers. Resins such as PharmaciaTM octyl
sepharose and ToyopearlTM butyl 650-M have been successful. The process is
earned
out in the presence of high salt, such as 1 M ammonium sulfate or 0.5 M sodium
citrate.
In accordance with one aspect of the present invention, there is provided a
method of refolding an improperly folded or denatured protein comprising the
step of
adding charged polymers to a solution comprising the protein prior to allowing
the
protein to refold.
In accordance with another aspect of the present invention, there is provided
a
method of refolding TFPI comprising the step of adding a charged polymer to a
solution comprising improperly folded or denatured TFPI prior to allowing the
TFPI
to refold.
In accordance with a further aspect of the present invention, there is
provided
a method of refolding TFPI comprising the step of immobilizing polymers of
sulfated
polysaccharides on a column and passing a solution of denatured TFPI through
the
column and eluting the refolded TFPI after the refolding has occurred.
CA 02450795 2003-12-19
10a
Detailed Description of the Invention
It is a discovery of the present inventors that polyionic polymers such as
polyethylene imine and polyphosphate, can modify the ionic interactions within
proteins. The masking of certain areas of high charge density within proteins
using
polyions can have nun ;emus effects. Proteins whose solubility is reduced
through the
infra- and/or inter- molecular neutralization of oppositely charged areas can
have their
solubility improved by masking one of the charged regions with polycations or
polyanions. Barriers to conformational flexibility and specific attractive or
repulsive
forces which interfere with the refolding process can be modulated as well.
Proteins
which require strong denaturants such as urea or guanidine hydrochloride to
solubilize
and maintain solubiity during purification operations can be solubilized and
processed
20
30
effectively using polyions.
CA 02450795 2003-12-19
~4'O 96/40784 ~'~'C1U~96/099~0
_1l_
Many proteins lacking a clear region of charge localization in the primary
sequence can still demonstrate areas of charge localization due to their
secondary
structure. Thus many proteins can have their solubility, refolding,
and,purification
characteristics modified through interaction with charged polymeric templates.
The
nature of the modification will depend on the specific protein structure, the
chain
length, charge, and charge density of the ionic polymer.
We have characterized refolding of pure TFPI in a guanidine or a urea based
refolding buffer and the results indicate that refolding efficiencies and
I~netl.cs can be
significantly improved by the addition of charged polymers including heparin,
dextrin
sulfate, polyethyleneimine (PEI) and polyphosphates. These polymers increase
TFPI
solubility and enhance refolding through ionic interactions with either the hT-
terminus
or the C-terminus. In addition to the polymer additives, refolding pure TFPI
requires
a cysteinelcystine redox buffer where the refolding reaction can be compleGrd
within
4g hours. Refolding yields are a strong function of gI-I, redox concentration,
and
polyri~er additives; however refolding efficiencies as high as 60 ~ can be
achieved for
pure TFPI under optimum refolding conditions.
It has been found by the inventors of this invention that addition of
glycosaminoglycans or sulfated polysaccharides such as, for example, heparin
and
dextrin sulfate, to a solution containing a denatured protein prior to
refolding
increases the amount of correctly folded, active protein where the protein is
capable
of binding to the glycosaminoglycan or sulfated polysaccharide and is
subjected to
renaturing conditions.
Recombinant I3I~.~ technology has allowed the high level expression of many
proteins that could not normally be isolated from natural sources in any
appreciable
quantities. In E. cmla and several other expression systems, the protein is
fr~"quently
expressed in an inactive, denatured state where the primary amino acid uence
is
correct, but the secondary and tertiary structure and any cysteine disulfide
bonds are
not present. The denatured protein present in an inclusion body may be in such
a
CA 02450795 2003-12-19
WO 96/40784 PC'TltTS96109980
-I2-
conformation that charged residues of different parts of the amino acid
bacl~bone that
are not normally in contact are able to interact and form strong ionic bonds
between
positively charged and negatively charged amino acid residues. The formation
of
these ionic bonds may limit the hydration that must occur to effect
dissolution of the
inclusion body. '.fhe protein in an inclusion body rr~ay also be complexed
with other
cellular components such as membrane components and nucleic acid which may
also
limit the access of solvent (water) to charged and normally hydrated residues.
Also
found in the unfolded state is that hydrophobic residues which are normally
found
buried in the interior of a protein are more exposed t~ the polar aqueous
environment.
Such occurrences may work to prevent the dissolution of inclusion bodies in
solvents
other than strong chaot~~opic agents such as urea oI° guanidine or
detergent.=. such as
SI7S.
Charged polymers preferably in aqueous solution can interfere with and
disrupt the undesirable ionic interactions that occur within a polypeptide
chain as
found in an inclusion body or other environaalent. "y"he charged polymers ntay
help
dISrUpt the undesirable ionic interactions and facilitate solvation of ionic
and polar
residues, promoting dissolution without the need for strong chaotropes or
detergents.
The charge, charge density, and molecular weight (chai:n length) of the
charged
polymer may vary depending on the specific protein. Suitable polymers include:
sulfated polysaccharides, heparins, dextran sulfates, agaropectins, carboxylic
acid
polysaccharides,alginic acids, carboxymethyl celluloses, polyinorganics,
polyphosphates, polyarrunoacids, polyaspartates, polygl~Itamates,
polyhistidines,
polyorganics, polysaccharides,DEAE l~extrans, polyorganic amines,
polyethyleneinimes, polyethyleneinime celluloses, polyamines, polyamino acids,
polylysines, and polyarginines.
Proteins with pI greater than 7 may benefit more from interactions with
negatively charged polymers, as these proteins will have a positive charge at
pH 7.
Proteins with pIs below 7 may interact more strongly with positively charged
polymers at neutral pI3. ~~hanging the solution pH will modify the total
charge and
charge distribution of any protein, and is another variable to be evaluated.
CA 02450795 2003-12-19
WO 96!40984 PCT/US96/09980
_13_
Be
Recombinant I~1~IA technology has allowed the high level expression of rryany
proteins that could not normally be isolated from natural sources in any
appreciable
quantities. In E, coli and several other expression systems, the protein is
:frequently
expressed in an inacti~re, denatured st~.te where the primary amino acid
sequence is
correct, but the secondary and tertiary structure and any cysteine disulfide
bonds are
not present. The denatured protein must be refolded to the proper active
conformation, which often requires overcoming significant energy barriers
imposed
by ionic attraction and repulsion, restrictions on bond rotation, and other
types of
conformationally induced stresses. Specific ionic attraction between opposite
charges
and/or repulsion between like charges can sever~eby Iimit the refolding
pathways
available to the denatured protein and reduce the efficiency of the refolding
process.
Some proteins leave specific areas of charge whose interactions may limit
1~ conformational flexibility or promote aggregation. I~iany proteins lacking
a clear
region of charge localization in the primary sequence can still demonstrate
areas of
charge localization due to their secondary structure found in refolding
intermediates,
misfolds and improperly folded protein. Proteins which are particularly
suitable
according to the present invention include but are not limited to TFPI, TFPI
muteins,
TFPI-2, tissue plasminngen activator, BST, PST. While it is believed that
these
methods will be suitable for use with proteins generically, those which are
most
suitable are those which are improperly folded, aggregated, oligomerized, or
inactive.
These will likely be proteins which have at least one highly charged domain,
and
possibly more, which can interact. In the case of TFPI, as well as other
proteins,
two oppositely charged domains interact with each other to prohibit proper
folding
and to cause oligomerization and aggregation. Proteins having many disulfide
bonds
will also be most likely to benef t from the present ~rnethods. Preferably the
protein
will have at least 2 and more preferably the protein will have at least 4 or 6
disulfide
bonds.
CA 02450795 2003-12-19
w~ 96/40'84 PCTNS96/09980
l~ _
~harge~l polyma:rs can be used to modify the charge, charge density, and
reduce or eliminate ionically mediated limitations to confarmation that may
arise in
the unfolded state. The juxtaposition of charged groups that are not normally
in
proximity may have result in dead-end refolding pathways from which the
refolding
process may never recover.
The introduction of extra positive or negative charges through
the
complexation with charged polymers may allow refolding to
proceed more facilely
for several reasons: iarst different types of charge distributions
may better
accommodate the refolding process; second, the addition
of the charged polymer may
enhance the solubility of the unfolded protein, reducing
or eliminating the need for
chaotropes which have a negative effect on protein conformation.
Because of the frequently unique structure associated with
most proteins the
charged polymer that demonstrates preferred characteristics
may vary. Evaluation of
the isoelectric pI3 (pI) of the protein can serve as a s
' g point. At neutral pIi a
1 ~ protein with a pI less than 7 will possess a net negative
charge, and will thus be more
likely to bind a positively charged polymer. The protein
with a pI greater than 7 will
possess a net positive charge at neutral pI~ , and will
trove a stronger tendency to bind
a negatively charged polymer. However, it is well established
that charges are
unevenly distributed aroa~nd a protein, and significant
charge localization can occur.
The possibility of localized concentrations of charges reduce
the ability to predict
which type of charged polymer may be most effective for
any application.
Theoretically, for any protein with, a specific distribution
of interacting charges and
conformational requirear~ents, there exists a charged polymer
of appropriate
compositions in terms of molecular weight, charge, and charge
distribution which
would maximize refolding efficiency. Other variables, such
as pI3 and solvent ionic
strength, would also be evaluated. Initial screening would
involve polyethyleneimine,
I~EAE dextran, dextran sulfate, and polyphosphate at several
different concentrations
and molecular weights. ~Jork with rhTFPI has demonstrated
the significant impact
that polyphosphate chain length and concentration can have
on the course of the TFPI
refolding reaction. Relatively short chain length (n~5)
produces high levels of
CA 02450795 2003-12-19
W~ 96/4Q784 P~'~'/t3S96/09980
_ 1~ _
aggregate. The optimal polyphosphate chain length for refolding rhTFPI was
approximately 25 re ting units. Longer chain length polyphosphates (n= 75)
also
produced more aggregate and less properly folded rr~onomer.
Proteins consist of chains of amino acids, thc; exact composition and sequence
of which constitutes one of the primary structure determinants of the protein:
The
secondary determinant of protein structure is the result of the conformational
guidance that the individual amino acid bonds have on protein conformation.
T'hirdiy,
the specific amino acid sequence directs the formation of tertiary structure,
such as
~i-sheets and «- helices. The three dimensional nature of protein
conformatiion often
brings into proximity amino acid residues that are not normally close to each
other
based on the direct sequence of the polypeptide chain. °rhe functional
farm of a
protein is generally a modestly stable conformation held together by a
combination of
cysteine disulfide bonds, ionic bonds, and hydirophobic and Van de:r Waals
interactions.
In general, protein solubility can be related tc> the nuumber of charged and
t~ a
CA 02450795 2003-12-19
WO 96/40784 PCT/US96/09980
- 1~ -
Complexation with charged polymers with relatively high charge density
represents one approach to increasing the charge density of any protein. ~
protein
with a small number of positively charged residues lysine or argininel can be
complexed with a negatively charged polymer such as polyphosphate. Sonne of
the
negatively charged groups of the polymer will interact with the positively
charged
groups present in the protein. The remaining charg~;d groups on the polymer
will be
free to interact with the solvent, in most cases water, and effectively
increase the
charge density and s~lvation of the protein. .Alternatively, a positively
charged
polymer such as polyethyleneimine can be used 'to complex with the negatively
charged residues of the protein. In some cases both. types of charged polymers
may
work equally effectively, in other cases, one charge type may be more
effective than
others. The effectiveness of any particular charged polymer, will depend on
protein
amino acid composition, protein amino acid distribution, protein
confoi°mation,
charged polymer charge density, charged polymer chain length, solution pI~,
and
other variables. However, it is likely that for any given protein, a
complementary
charged polymer that will bind to the protein and essentially increase the
charge
density of protein ran be found that will improve the solubility
characteristics of that
protein in aqueous medium.
Definitiflns
The term "processing" as used herein refers to the steps involved in the
purification and preparation of pharmaceutically-acceptable amounts of
proteins.
Processing may include one or more steps such solubilization, refolding,
chromatographic separation, precipitation, and formulation.
The term "charged polymer" and "charged polymeric template" refer to any
compound composed of a backbone of repeating stn~ctural units linked in linear
or
non linear fashion, some of which repeating units contain positively or
negatively
charged chemical groups. The repeating structural units ;may be
polysaccharide,
hydrocarbon, organic, or inorganic in nature. The repeating units may range
from n
= 2 to n=several million
CA 02450795 2003-12-19
_ 17-
The term "positively charged polymer" as used herein refers to polymers
containing chemical groups which carry, can carry, or can be modified to carry
a
positive charge such as ammonium, alkyl ammonium, dialkylammonium, trialkyl
ammonium, and quaternary ammonium.
The term "negatively charged polymer" as used herein refers to polymers
containing chemical g~°oups which carry, can carry, or can be modified
to carry a
negative charge such as derivatives of phosphoric and other phosphorous
containing
acids, sulfuric and other sulfur containing acids, nitrate and other nitrogen
containing
acids, formic and other carboxylic acids
The term "polyethyleneimine" as used herein refers to polymers consisting of
repeating units of ethylene imine (H3h1+-(CH2-CH2-hTH2+)x-CH2-CH2-NH3+).
The molecular weight can vary from 5,000 to greater than 50,000.
The term "polyphosphate" as used herein refers to polymers consisting of
repeating units of orthophosphate linked in a phospho anhydride linkage. The
number
of repeating units can range from 2 (pyrophosphate) to several thousand.
Polyphosphate is frequently referred to as sodium hexametaphosphate (SHMP).
Other common names include Grahams salt, CalgonTM, phosphate glass, sodium
tetrametaphosphate, and Glass HT~.
The term "refold" as used herein refers to the renaturation of protein.
Typically, the goal of refolding is to produce a protein having a higher level
of
activity than the protein would have if produced without a refolding step. A
folded
protein molecule is most stable in the conformation that has the least free
energy.
Most water soluble proteins fold so that most of the hydrophobic amino acids
are in
the interior part of the molecule, away from the water. The weak bonds that
hold a
protein together can be disrupted by a number of treatments that cause a
polypeptide
to unfold, i.e. denature. A folded protein is the product of several types of
interactions between the amino acids themselves amd their environment,
including
ionic bonds, Van der ~laals interactions, hydrogen bonds, disulfide bonds and
covalent bonds.
CA 02450795 2003-12-19
WO 96/40784 pC'fIUS96/09980
_18_
The term xdenature" as used herein refers to the treatment of a protein or
polypeptide in which results in the disruption of the ionic and covalent bonds
and the
fan der Waals interactions which exist in the molecule in its native or
nenatured
state. Denaturation of a protein can be accomplished,, for example, by
treatment vvtith
8 M urea, reducing agents such as mercaptoethanol, heat, pH, temperature and
other
chemicals. Reagents such as Il IvI urea disrupt both the hydrogen and
hydrophobic
bonds, and if mercaptoethanol is also added, the disulfide bridges (S-S) which
are
formed between cysteines are reduced to two -S-1-3: groups. Refolding of
proteins
which contain disulfide iznkages in their native or refolded state may also
involve the
oxidation of the -S-:F3 groups present on cysteine residues for the protein to
reform the
disulfide bonds.
The term "glycosaminoglycan" as used herein refers to polysaccharides
containing alternating residues of uronic acid and hexosamine and usually
contain
sulfate. The binding of a protein in a refolding reaction as described herein
to a
glycosaminoglycan is through ionic interactions.
The term "dextran sulfate" as used herein refers a polyanionic derivative of
dextaart, ranging in molecular weight from 8, to SO~,OnO daltons. Dex~rans are
polymers of glucose in which glucose residues are joined by aI,6 linkages~
The term °heparin" as used herein refers to 2 glucoaminoglycans or
heparinoids which are based on a repeating disaccharide (-4DGlcA(p)~31;
4GIcNAca1-
)n but are subject to extensive modification ai~er assembly. Heparin is stored
with
histamine in mast cell granules and is thus found in mast connective tissues.
In general
heparins have shorter chains than heparin.
The term "HIG" as used herein refers to hydrophobic interaction chromatography
which employs a hydrophobic interaction between the column and the molecule of
interest to separate the sulfated polysaccharides and other contaminants from
the
refolded product.A.
Negatively charged polymers include sulfated polysaccharides, such as
heparins,
dextran sulfates, and aga;iopectins, as well as carbo:~ylic acid
polysaccharides such
asalginic acids and carboxyr~ethyl celluloses. Polyinorganics such as
polyphosphates are
CA 02450795 2003-12-19
W~ 9614074 ~'CTl1JS96/099~0
1g _
also included. Polyamino acids such as polyasparatate, polyglutamate, and
polyhistidine
can also be used.
Positively charg~°d polymers include polysaccharides such as I3EAE
dextran,
polyorgnic amines, such as polyethyleneimines, polyethyleneimine celluloses,
and
polyamines, as well as ttie polyamino acids, polylysine and polyarginine.
Combinations
of polymers may be used, of either charge polarity. In addition, amphoteric co-
polymers
may also be used.
As used herein, "TFPT" refers to mature Tissue Factor Pathway Inhibitor. As
noted above, TFPI is also known in the art as Lipoprotein Associated
Coagulation
Inhibitor (LACI), Extrinsic Pathway Inhibitor (FPI) and Tissue Factor
Inhibitor (TFI).
I~iuteins of TFPI which :retain the biological activity of TFPI are
encompassed in this
definition. Further, TFPI which has been slightly modified for production in
bacterial
cells is encompassed in the definition as well. For example, a TFPI analog has
an alanine
residue at the amino-terminal end of the TFPI polypeptide has been produced in
Escherichaa c~la. See IJ.,~. 5,212,081.
As used herein, "pharmaceutically acceptable composition" refers to a
composition that does not negate or reduce the biological activity of
formulated TFPI,
and that does not have any adverse biological effects when formulated TFPI is
administered to a patient.
As used herein, "patient" encompasses human and veterinary patients.
As used herein, the; term "solubilizer" refers to salts, ions, carbohydrates,
amino
acids and other organic molecules which, when present in solution, increase
the
solubility of TFPI above 0.2 mg/mL. Solubilizers may also raise the
concentrations of
TFPI above 1 mgfmL and above 10 mg/mL. It should be noted that solubilizers
may act
as stabilizing agents. Stabilizing agents preserve the unit activity of I in
storage and
may act by preventing formation of aggregates, or by preventing degradation of
the
TFPI molecule (e.8. by acid catalyzed reactions).
As used herein, the term "secondary solubilizers" refers to organic salts,
ions,
carbohydrates, amino acids and other organic molecules which, when present in
solution
with a solubilizer, further increase the solubility of TFPI. Secondary
solubilizers may
CA 02450795 2003-12-19
-20-
have other effects as wall. For example, secondary stabilizers may be useful
in
adjusting tonicity (e.g. to aotonicity).
The amino acid sequence of TFPI is disclosed in U.S,. Patent No. 5,106,833 and
Figure 4. Muteins of TFfI and TFPI-2 are disclosed in ZJ.S. Patent No.
6,10:3,500. As
described in Il.S. Patent No.6,103,500, a=nuteins of TFPI and TFPI-2, with
single or
multiple point mutations, and chimeric molecules of 'hFPI and TFPI-2 can be
prepared.
For instance, the lysine rE;sidue in the P 1 site of the first Kunitz-type
domaia~ of TFPI
may be replaced with arginine. Muteins, cont<~ining one to five amino acid
substitutions, may be prepared by appropriate mutagenesis of the sequence of
the
recombinant cloning vehicle encoding TFPI or TFPI-2. Techniques for
miztagenesis
include, without limitation, site specific mutagenesis. Site specific
mutagenesis can be
carried out using army number of procedures known in the art. These techniques
are
described by Smith (1985) Annual Review of C;eneti~°~, 19:423, and
~nodifi~cations of
some of the techniques ar° described in METHODS IN EN:~YMOLOCp~', 1
'_i4, part E,
(eds.) Wu and tlrossman (1987), chapters 1'~, 18, lCe, and 20. A preferred
procedure
when using site specific mutagenesis is a modification of the Clapped Duplex
site
directed mutagenesis method. The general procedure: is described by Kramer, et
al., in
chapter 17 of the Methods in Enzymology, above. ~f~,rrlother technique for
generating
point mutations in a nucleic acid sequence is overlapping PCR. The procedure
for
using overlapping PCR to generate point mutations is described by Higuchi in
Chapter
22 of PCR PROTOCOLS: A C~L1IDE TO METTIODS AND APPLICATIONS, (eds.)
Innis, i~elfand, Sninsky and White (Academic Press, 1.980).
Alternatively, hybrid proteins containing the first Kunitz-type domain .of
TFPI-2
and the second and third hunitz-type domains of TFF'I could be produced. One
skilled
in the art of DNA cloning in possession of the DNA wracoding TFPI and TFPI-2
would
be able to prepare suitable DNA molecules for production of such a chirneric
protein
using known cloning procedures. Alternatively, synthetic DNA molecules
encoding
part or all of each Kunitz--type domain and peptide sequences linking the
Kunitz-type
domains can be prepared. As a further alternative, the: overlapping PCR
technique may
CA 02450795 2003-12-19
-21 -
also be used to prepare DNA encoding chimeric molecules containing TFPI and
TFPI-2 sequences.
TFPI can be prepared in yeast expression systems as described in U"S. Patent
No. 6,103,500. Methods have also been disclosed for purification of TFPI from
yeast cell culture medium, such as in Petersen et al., 3.Biol.Chem. i x:13344-
13351
X1993). In these cases, recombinant TFPI is secreted from the yeast cell. TFPI
recovered in such protocols is also frequently heterogeneous due to N-terminal
modification, proteolyti~ degradation, and variable glycosylation. Therefore,
a need
exists in the art to produce mature TFPI that is authentic ~i.e. having the
correct N-
terminal amino acid sequence), full-length and homogeneous.
TFPI can be produced in E. coli as described in ~J.S. Patent No. 5,212,091
which discloses a method of producing TFPI by expression of a non-glycosylated
form of TFPI in an E. cc~li host.
In one aspect of the invention recombinantly produced proteins which have
the ability to bind polymers of sulfated polysaccharides such as, for example,
heparin or dextran sulfhte are refolded. The invention provides a method that
facilitates refolding of a denatured recombinantly produced protein product
using
polymers of sulfated polysaccharides which act as a templates for the
refolding
protein. Without being limited to any particular theory, the inventors believe
that
the interactions betwee~l the refolding protein and the polymeric template may
minimize aggregation of the refolding intermediates and provide an environment
for
the protein to refold to its native conformation. The polymer acting as a
template
may bind a domain or region of protein to stabilize the intermediate and allow
further folding to occur without aggregation. The protein aggregates, if
fonmed, are
generally less active tha~a non-aggregated refolded protein, and generally
result in a
reduced overall yield of active refolded protein. The NaCI concentration of
the
refolding conditions is considered important and is selected to achieve the
maximum
efficiency of refolding by maximizing the interaction between the template
polymer
and the refolding proteins. For example, it has been found by the inventors
that
approximately 0.2 M concentration of NaCI or louver promotes binding of the C-
terminal and/or the third Kunitz domain of TFPI to heparin or other sulfated
polysaccharide polymer. The binding of polymer to the intermediate is presumed
to
CA 02450795 2003-12-19
-22-
facilitate the solubility of the intermediate and provide an environment for
tree rest of the
protein to refold by reducing aggregation of the refolding intermediates.
General Methods
TFPI may be prepared by recombinant methods as disclosed in U.S. 5,212,091.
Briefly, TFPI is expressed in Eschew-ichirx coil cells and the inclusion
bodies containing
TFPI are isolated from. the rest of the cellular material. The inclusion
bodies are
subj ected to sulfitolysis, purified using ion exchange clhromatography,
:refolded by
disulfide interchange reacaion and the refolded, active TFPI purified by
ration exchange
chromatography. TFPI may also be produced in yeast as disclosed in LJ.S.
Patent No.
6,103,500.
T FPI activity may be tested bgT the prothrombin time assay (PTT assays).
Bioactivity of TFPI was measured by the prothrombin clotting time using a
model RA4
Coag-A-MateTU from Organon Teknika Corporation (Oklahoma City, OK). TFPI
samples were first dilutes: to 9 to 24 ug/mL with a 'fBSA buffer (50 mM Tris,
100 mM
NaCl, 1 mg/mL BSA, p1-1 7.5). Then 10 uL of ~arify 1'T'~ (pooled normal
~>lasma from
~rganon Teknika Corp.) was mixed with 90 uL of diluted TFPI samples in a
sample tray
and warmed to 37 C in the instrument. Finally SimplastinTM Excel
(Thrornboplastin
from ~rganon Teknika Corp.) was added to start the clotting. The time delay in
clotting
due to anticoagulant activity of TFPI was measured and converted into TFPI
concentration in the measured samples by comparison to a TFPI standard curve.
The amount of soluble TFPI may also be quantified by measuring the area of the
main peak on a ration exchange chromatogram. HI'LC analysis of TFPI samples
was
performed using a Waters 620 LC system (Waters Corporation, Milford, MA)
equipped
with as Water 717 plus heater/cooler autosampler . I7~ata acquisition was
processed by a
Turbochrom system from Perkin-Elmer.
The ration exchange (IEX) method used a pharmacia Mono S FIFZ 5/5 glass
column. The column was equilibrated in 80% buffer A (20 mM sodium acetate
trihydrate:acetonitrile solution (70:3 0 v/v) at pH 5.4) and 20% buffer B (20
mM sodium
acetate trihydrate - 1.0 Ir~1 ammonium chloride:acetonitrile solution (70:30
vlv) at pH
5.4). After a sample was i~ajected, a gradient was applied to elute the TFPI
at a flow rate
CA 02450795 2003-12-19
w~ 96140754 PC'°r//~JJS96/09980
~23-
All reagents are ~;r.S.P. or A.C.S. grade. Suppliers include J.T. Eaker and
Sigma
Co. (St. Louis, C3).
°The present invention will now be illustrated by reference to the
:following
examples, which set ford, certain embodiments. I-Iowever, it should be noted
:hat these
embodiments are illustr~.tivc and are not to be construed as restricting the
invention in
any way.
E hIPLES
Example 1 ~ Refolding denatured I
'The following ex~~mple describes the making of stock solutions, the I3Itr
column
preparation, the initial recovery and purification ofTFfI prior to refolding,
the refolding
of I, and the recovery of active I.
The TFPI stock was prepared fr~m refractile bodies resulting ):corn the
expression of recombinant I in bacteria. The refi-a.ctile bodies were
solubili,~ed at 10
mglml in 8 urea, 50 r~ Tris pl~i 8.5 containing 10 a 1.7TT, and this solution
was
clarified by centrifugation at 10,000 x g for 10 minutes.
The column prel>aration for the initial purification of the solubilize7:'~I
was
prepared with S-Sepharose beads nixed in 7.5 ure<~, 10 Tris and 10 ~ sodium
phosphate (pI-I 6.5) conta.ning 5 I)TT and 1 Fi7TA. The solubilized '. I at a
concentration of 5 mgirr~l was then run over the S-Sepharose column and eluted
with a
sodium chloride gradierdt of 0 to 1 NI. The purified TFPI had an absorbency at
wavelength 280 nm of 3.2 (which is equivalent to 4.1 mgJml using an extinction
coecient of 0.78).
The dextran sulfave stock consisted of dextrar;~ sulfatc of molecular weight
8000
daltons available from Sil;ma, item number D-4911, made up at 50 mg/ml (6.2'.>
in
50 Tris (pI~ 8.8) in 0.1 Ie~: sodium chloride, and stored at -20 degrees
c~:ntigrade
between uses.
CA 02450795 2003-12-19
-24-
The heparin stock, if heparin was used to conduct the refolding, was of
molecular weight 600t) to 30,000 daltons, (with an average molecular weight of
18,000 daltons) prepared as a sodium salt available from Sigma Co. (St. Louis,
MO),
item number H-3x93, made up at 60 mg/ml (3.33 arzM) in 50 mM Tris (pT~:8.8) in
0.I
sodium chloride, and stored at -20°centigrade between uses.
To the S-Sepharose purified TFPI either dextran sulfate stock solution or
heparin stock solution ~~an be added. Dextran or heparin was added to TFPI
under
denaturing conditions in 6 to 8 M urea. With 4°C reagents, the
denaturing solution
containing TFPI was diluted to 3 M urea, 50 mM Tris (pH 8.8), 0.2 M sodium
chloride, and 0.5 mg/rnl TFPI, and to a final dextran sulfate c~ncentration of
0.6
mg/ml (75 MM) or a fr;ial heparin concentration of 1.5 mg/ml (83 ~uM),
depending on
which was used to facilitate the refolding. Cystine was added to the refolding
sohation to a final concentration equal to the anal DT'T concentration. The
refolding
solution was incubated at 4°C with gentle agitation for from 4 to 6
days, preferably 5
days.
As an illustration of this procedure the following is a detail of a protocol
for
refolding a 5 ml solution of TFPI in dextran sulfate or heparin.
To 610 p1 of TFPI stock either 60 ~l of dextran sulfate with 65 p1 of SO mM
Tris (pH 8.8) in 0.1 M NaCl, or 125 pI of heparin stock solution with 50 mM
Tris
(pH 8.8) or O.I M NaCl was added. The refolding ;>olution was mixed and
.allowed to
incubate 10 minutes on ice. Next, 4.2 ml of refolding buffer containing 2.5 M
urea,
50 mM Tris (pH 8.8) and 165 mM sodium chloride was added to the refolding
solution and mixed. Finally, 61 ~l of 50 mM Cy:>tine made up in 120 m~~I
sodium
hydroxide was added and the total solution was incubated at 4°C with
gentle agitation
for 4 days. The free sulfhydryl content was checl~ed with Ellman's reagent
(also
called DTNB). Idoacetamide was added, to 20 mM, made up at I M in I00% ethanol
for storage at .20°C.
The hydrophabic interaction column (HIC) was prepared from Butyl-650M
Tosohaas ToyopearlTM rE;sin particle size 40-90, pai°t # 014702. The
butyl resin was
washed in 3 M urea, I M ammonium sulfate, 50 mM Tris, 10 mM sodium phosphate,
pH 6.5 and resuspended at a 50% slurry.
CA 02450795 2003-12-19
w~ 96/40784 ~CTII~JS96A09980
_2S_
The refolding samples, stored at -20~C remained in the standard refolding
buffer
containing 3 M urea, SO mM Tris, pH 8.8; 1-4 mM redox, 0.5 mg/ml TFPI, and 0.2-
0.6
M Na~l depending on condition. Samples refolded ,,pith dextran or heparin had
0.2 M
salt, and samples without dextran or heparin had 0.6 M Na~l.
S The following steps were performed at room temperature to effect the further
purification of the refolded TFPI. To 300 pI of refolded sample, an equal
volume of 2
~I amrnoraium sulfate, 3 M urea, 50 mM Tris, and 1 d3 mM sodium phosphate (pH
6.5)
was added. Next, 100 p1 of washed Butyl-6SOM beads was added to the diluted
refolded sample. The solution with the beads was incubated with gentle
rcPcking or
mixing for 30 minutes at room temperature. The mix was then spun in an
ependorf
centrifuge for 5 second s, and put in a rack and allowed to sit for one nunute
for the
beads to settle flat in the tube. The supernatant was aspirated carefully, so
as not to
disturb the beads.
To wash the T'FPI-bound beads, 1 ml of wash bufl'er camposed of 1 M ammonium
1S sulfate, 3 M urea, SO mla~ Tsis, I0 nxM sodium phosphate (pH 6.S) was
addsrd to the
beads to remove the remaining dextran sulfate or heparin. 'The washed mixture
was re-
spun in an ependorf centrifuge for 5 seconds, and allowed to sit for one
minute for the
beads to settle as 'before. The supernatant was removed, and the beads there
washed
with the gash buffer a final time, and spun and allowed to sit as before.
After the final
wash and settling, the supernatant was removed with a flame-pulled-tip Pasteur
pipette
very carefully.
To elute the refolded I, 300 p1 of elution bufl;er composed of 3 M urea, 0.1 M
ammonium sulfate, S0 m.IVI Tris and 10 mM sodium phosphate (pH 6.S) was added
to
the slurry of beads and rocked for more than 10 nunutes. The beads were
pe;lleted by
spinning in an ependorf centrifuge, and the supemata~t containing refolded
TFPI was
recovered. To a~roid contamination of the beads with the product, somsr of the
supernatant was left behind.
CA 02450795 2003-12-19
w~ 96/40784 ~'C°TlHJS96I09980
-26-
Example 2 - C of Dextran Sulfate Refold
The sample of TFPI was renatured at a concentration of 0.5 mg/ml TFI'I, 0.6
mgJml
Dextran sulfate, 3.0 M I:lrea, 200 NaCI and SO m~M Tris (pH ~.5). The HIfC
column
was prepared from TosoHaas Butyl beads for HIC, ~.6 mmD/100mmL, in a 1.66 ml
slurry. The flow rate was set for 1.0 ml/min. Before loading the HIC column,
the
sample was diluted 2:3 with 3.0 M Urea and 3.0 M NH4S~, at a final pH of 5.68;
2 ml
of sample was loaded. The gradient start was 33 MES/33 mM PE;S/33 mM
sodium acetate, 1.0 M NH~S04, and 3.0 M Urea, pH 6.0; the gradient end was 33
mM
MES/33 mM HEPES/33 mM sodium acetate, 3.0 e/ Urea at pH 6Ø The gradient
volume was 5.0 CV. From this column, the recover5r of native TFPI was 68%. The
results ofthis run are sl;~own in Figure 2.
A second HIC column was also run. The sample of denatured TFPI was diluted 2:3
with 3.0 M Urea, 1.5 W NH4SC1~ and two rnl were loaded. The gradient start was
33
mM MES/33 mM HEPESi33 mM sodium acetate, O.S M NH4S~~, and 3.O M Urea, pH
6.0; the gradient end was 33 MESf33 mM HEIPES/33 sodium acetate, 3.0M
Urea at pH 6Ø 'The g~°adient volume was 5.0 CV. "fhe recovery of
native ~°FPI from
this second column was 74~/°. The results of this ruri are shown in
Figure ~.
The samples were analyzed by non-reducing SDS-PAGE as illustrated in Figure
1. Correctly refolded, active I species major bard) are seen on the gel.
Example 3
About 10 mglmL TFPI in 2M urea was dialyzed against one of the following: 20
mM acetate, 20 mM phosphate, 20 mM citrate, 20 rnM glycine, 20 I,-glutamate or
20 mM succinate in 150 mM NaCI as described above. 6-10 mglml~ ~"FPI bulls
stock
was loaded into SpecIPo:r 7 dialysis tubings cutoff 3,a ). Dialysis was
carried
out either at 4; C or ambient temperature. Three changes of buffer at a
protein
solution to buffer ratio: 1 to 50-lnU, were made during c~urse of dialysis
over 12 to
24 hr time period. Aft~:r dialysis, TFPI solution was filtered by Cos 0.22
micron
filter units to separate p>recipitated TFPI from soluible TFPI. The solubility
of TFPI
was then measured by ~TVIVis absorbance assuming an absorptivity 0.68
~mg/mL)'1
cm'' at 278 nm. The solutions were prepared at v;ous pH levels by titration
with
HCl or NaOH.
CA 02450795 2003-12-19
w0 96/40784 P~'T1U896/0'9980
After completion of dialysis, the precipitates were filtered through 0.22 ,um
filter units. The concen;xation of remaining soluble m FPI after dialysis was
measured
by tJ~1 absorbance. Figure 1 shows the results of these experiments.
solubility of
TFPI increased greatly :in solutions containing 20 r~iacetate, 20 m~I
phosphate, 20
S m~I L-glutamate and ~?0 mI~ succinate at pI-I levs~ls below 7 and
particularly at or
below pI~ 4.5. solubility of TFPI was also su~~stantially increased in
solutions
containing 20 mIvT glyc~,ne above pI~ 1~. Figure 2 shows the solubility of
',fFPI as a
function of concentration of citrate ion ire the presence of 10 m11~ Na
phosphate at pH
7. TFPI solubility incre~~ses with increasing concentration of citrate. Figure
3 shows
the solubility of TFPI as a function of concentration of NaCI at p.I~ ~.U.
TFPI
solubility increases with increasing salt concentration, indicating salt.
promotes
solubility of TFPI.
The solubility o:~ TFPI was studied using a number of different solubilizers
and secondary solubilizers. Table 1 shows solubility of TFPI in varying buffer
1S solutions measured by IJ-V' absorbance after dialyzing 6 to 10 mg/mL TFPI
into these
buffer solutions.
CA 02450795 2003-12-19
'6~'~ 96/4074 PCT'/LJS96/099~0
~2~m
Tabl 2 1
Satt effect _ Solubilit
Content H c m /ml
_ _ uv
l OmM NaP04. 7 0.21
~
iOmM NaP04, 150mM NaCJ 7 0.72
20mM NaP04, 150mM NaCI 7 0.85
20mM NaPO4. 0.5M NaCI 7 6.71
20mM NaP04, 1 M NaCI ~ 7 8.24
~
_--
pE-f effect
Content N c m Imf
uv
20mM NaOAc, 150mM NaCI 3 10.27
20mM NaOAc, 150mM NaG! 3.510.25
20mM Na0Ac,150mM NaCB 4 7.54
_ 4.5i .75
20mM NaOAc, 150mM NaCi
20mM NaOAc, i 50mM NaCI 5 1 .1 5
20mM NaOAc, 150mM NaC! i 5.50.85
20mM NaP04, 150mM NaCI 5.50.89
20mM NaP04, 150mM NaCI 6 0.78
20mM NaP04. 150mM NaCI ~ 6.50.79
20mM NaP04, 150mM NaCI g5
20mM NaP04, 150mM NaCI ~ ~5 ~:
A -- 8~
20mM NaP04, 150mM NaCB 8 0.86
~
20mM NaCitrate, 150mM NaCI 4 2.1 7
20mM NaCitrate, 150mM NaCI 4.51 .19
20mM NaCitrate, 150mM NaCI . 5 1 .1
20mM NaCitrate, 150mM NaCI 5.51.84
20mM NaCitrate, 150mM NaCI 6 2.09
20mM NaCitrate, 150mM NaCI 6.S2.12
20mM NaCitrate, 150mM NaCI 7 1 .92
s
20mM GI tine, 150mM NaCI 9 0.32
~ g
20mM Glycine, 150mM NaCI 1 0.9
' 0
_ 1
20mM GI cine, 150mM NaCI 1
13.94
20mM'L-Glutamate, 150mM NaCI 4
9.07
20mM L-Glutamate, 150mM NaCI 5
1
.21
20mM Succinate, 150mM NaCI 4
8.62
20mM Succinate, 150mM NaCI 5
1.21
20mM Succinate, 150mM NaCI ~ 6
1
.07
Citrate
Content o-i
c
m
/ml
uv
lOmM NaP04, 20mM NaCitrate 7
1
.1
6
CA 02450795 2003-12-19
WO 96140984 PdCT/63S96109980
Table 1 (c0nt.)
tOmM NaP04, 50mM NaCitrate 7 5.g1
tOmM NaP04, 100mM NaCitrate 7 12.7
~
t 0mM NaP04, 200mM NaCitrate V _ ~ t 5.9
7
lOmM NaP04, 300mM NaCifrate 7 8.36
M 2+, Ca2+ and of hos hate
Content _ H c m Iml
uv
lOmM NaPO~, 7 50mM NaCI, t mM MgCl2 7 0.66
lOmM NaP04, 150mM NaCi, tOmM IVi CI2 7 1.02
lOmM NaP04, 150mM NaCI, 0.7mM _CaCl2 7 0,67
tOmM NaP04, 150mM NaCI, 1mM CaCl2. 7 0,71
tOmM NaP04, 150mM NaCt, iOmM tri hos hate 7 3,6q
tOmM NaP04, 5% PEG-400
~ 0.07
7
tOmM NaP04, tOmM EDTA _ 0.36
~
7
lOmM NaP04, 100mM Na2S04 7 5.08
lOmM NaP04, 100mM !_-as antic acid 7 0.4
lOmM NaP04, 100mM Succinic acid 7 2.33
tOmM NaP04, 100mA4 Tartaric acid _ 2.56
7
20mM NaP04, 100mM Malefic acid 7 p
2omM NaP04, t00mM Malic acid 7 1.87
lOmM NaP04, t OOmM l- lutamic acid 7 0
_tOmM NaP04, 150mM NaCI _ 7 0.25
tOmM NaP04, 100mM isocitrate ~ 7 10.83
NaOAc, NaP~4 and NaCi
Content H r. m ml
' uv
tOmM NaOxlc, t50mM NaCI 4.5t.76
lOmM NaOAc _ 4.54.89
t OmM NaOAc 5.54.95
tOmM NaOAc 6.5' S.1
t OmM NaOAc 7 5.87
lOmM NaP04, 150mM NaC! 4.50.14
tOmM NeP04 4.54.97
tOmM NaP04 5.50.79
1 OmM NaP04 6.5~ 0.09
- t
lOmM NaP04 7 0.94
~
a
50mM NaOAc 5 5.24
SmM NaOAc 5.54.59
t OmM NaOAc 5. 5:05
5
20mM NaOAc .5 5.04
5
_ _ 5.7't
SOmM NaOAc _ - I 5.5
t00mM NaOAc 5,51.4
200mM NaOAc - -__- ~ 5.5 t.32 -
CA 02450795 2003-12-19
W~ 96!40784 ~'C~'liJS96109980
~30~
Table 1 (cont.)
5mM NaOAc, 150mM NaCI ~ 5.5 ~ 0.65
lOmM NaOAc, 150mM NaC!
f 5.5 0:69
20mM NaOAc, 150mM NaC1 5.5 0.7.1
~
50mM NaOAc, 150mM NaCI 5.5 0.91
Fi dro hob j
icchain !en ih
_ ! H c
Content ! mo/ml
uV
lOmM NaP04, 50mM Formic acid 0 7 0,12
lOmM NaP04, 50mM Acetic acid 7 I 0.16
lOmM NaP04, 50mM Pro anoic ae:id 7 0.16
_
lOmM NaPO4. 50mM Butanoic a_cid_ ~ 7 0.13
0
lOmM NaP04, 50mM Pentanoic acid ~ 0.14
~ 7
lOmM NaP04, 50mM Fiexanoic said ~ ~
~ ~ 0.1
7 I! 1
t
~thBfS
Content ! pf~l c (mg/ml~ uv ~
I
20mM NaOAc, 3% Mannitot, 2% Sucrose, 5% PEG-400~ 4 19.9
20mM Na Citrate, 3J Mannitol, 2/z~ Sucrose, ~ 6.5 0.72
5% PEG-400
20mM NaCitrate, 159mM NaCl, S.o PEG-X00 8.5 2.18
20mM NaOAc, 150mM NaCI, 5% PEG-~&00 ~4 t 9.8
20mM Na Citrate. 130mM NaCl, 1~% Gfycine, 0.25% ~ _
Tween-80. ' .48
5/ PEG-400
_ _
20mN~ Na Citrate, 130mM NaCi, 1 '/ Glycine, ~~ 6.5 ~ 1 .32
0.25% Tween-80
_~ Solubi8ity
Content ) pEi c (mg/m!),uv
j
Smt~ NaAc_etate ; 5.5 8.9
I
5mM NaAcetate, e% Sucrose _
t 5.5 1 1
~SmM NaAcetale, 0.01 /~ Polvsorbate-80 ~ 5.5 7
~
CA 02450795 2003-12-19
W~ 96!40?84 ~CT/~IS96109980
-31 -
Table 1 (cost.)
5mM NaAcetate, 8% Sucrose, 0.01 % 5 _ 1 2
Polysorbate-80
lOmM NaAcetate ~ 7.6
5.5
1 omM NaAceiate, 8! Sucrose ~ 5.5 1 0
1 OmM NaAcetate, 8% Sucrose, 0.01 ; 5.5 12.1
% Polysorbate-80
5mM NaAcetate, 5% Sorhitol 5.5 7.8
5mM NaAcetate, 4.5% Mannitol l 5.5 9.2
5mM Histidine 6 5.5
5mM Histidine 6.5 1
5mM NaCitrate I 5.5 0.1
5mM NaCitrate I 6 0.1
5mM NaCitrate 6.5 0.1
~
5mM NaSuccinate - 5.5 0.6
5mM NaSuccinate 6 0.3
5rnM NaSuccinate 6.5 0.2
t OmM Imidazole j 6.5 2.5, 10.8
t OmM Imidazole , 3 0.8
iOmM Imidazole, 8/~ Sucrose _~ 6.5 12.2
5mM NaAcetate ~ 6 8.2
lOmM Imidazcte, 5mM NaAcetate i 6.5 12.8
t OmM NaCiirate i 6 0.2
1 OOmM NaCitrate ~ 6 8.1
100mM NaCitrate ~ 7 9.3
tOmM Naphosphate, 260mM Na2S04 6 9.1
lOmM NaPhosphate, 100mM NaCitrate 8 8.8
~
lOmM NaCiirate, 1 % L- lutamic acid f 6 4.6
lOmM NaCitrate, 2! l_-I sine 6 1.1
iOmM NaCitrate, 0.5% L-as antic acid ~ 6 0.4
lOmM NaCitrate, 0.1 % Phos hate lass '7 5.9
lOmM Tris, 100mM NaCitrate 8 8.5
1 OmM NaGitrate, 1 M Gi cine 6 0.3
lOmM NaCitrate, 300mM GI cine ~& 0.3
lOmM NaCitrate, 280mM Glycerol 6 0.3
lOmM NaCitrate, 0.5M NH4 2504 6 8.3
lOmM NaCitrate, l2omM NH4 2504 6 8.8
lOmM NaCitrate, 260mM Na2S04 6 9.4
lOmM NaP04, 0.1% Phos hate lass 7 15.8
lOmM NaCitrate, 0.1% SOS 6 11.2
lOmM NaCitrate, 0.02! SOS 6 7.8
lOmM NaAcetate, 8% PEG-400 5.5 13.~
lOmM NaAcetate,.1~50raM NaCI, 8% PEG-4005.5 _0.6
lOmM NaAcetate, 8% PEG-400 6 16.2
lOmM NaCitrate, 8% PEG-400 - ~ 0.2
CA 02450795 2003-12-19
W~ 96/40784 PC'T/I,TS96/09980
~'he stability of '1'FPI stored at various pH conditions was tested. '1"FPI
was
prepared by dialysis as above fn 10 mM IeTa phosphate, 150 mM lVaCl and 0.005%
(w/o) polysorbate-80. Stability samples containing 150 mg/mL TFPI were
incubated
at 40iC for 20 days. Kinetic rate constant for the remaining soluble 7::~?PI
was
analyzed by following decr of the main on ration exchange chromatogra>rcs.
As can be seen in Figure ~, the decay rate constant increases at pH above 6.0,
indicates yore aggregation at higher pIi conditions.
TFPI was also fE~rrnulated at a concentration. of 150 mg/mL in 150 nnM NaCI
and 0.005% (wlv) polysorbate-80 at pII 7 with varying concentrations of
phosphate.
~°igure SA shows the percentage of remaining soluble T'FPI measured by
the ration
exchange HPLC. IncrE;asing concentrations of phosphate ion in solution
resulted in
higher levels of soluble TFPI remaining after incubation at 40°C.
Higher levels of
phosphate ion also resulted in higher levels of active "iCFPI as assayed by
the
prothrombin time assay. 'T'hese results are shown in F°igua°e
5B.
Stability of TFP:i at a concentration of 0.5 mglmL and formulated in 10 mM
I~Ta citrate, pH 6 and i50 mM l~aC1 was also tested at 40°C over a 40
day period.
As seen in Figub, ca~.ion-exchange ~IPLC (triangle) shows the presence of
soluble
TFPI at levels greater ~;han 60 % initial, even after' the 40 day incubation.
In like
manner, the prothrombin time assay (circle) shows the presence of active TFPI
at
levels greater than 60 % initial, even after the 0 day incubation.
Figure 7 shows loss of soluble T'FPI at 40°C measured by bond cation-
exchange HPLC (open symbol) and prothrombin time assay (closed symbol) for 0.5
mglmL ~"FPI formulated in 10 mM Na phosphate, p~I 6 and either 1~0 mM NaCI
(triangle) or 500 mM bT6tCl (circle).
Figure 8 shows loss of soluble 1'FPI at 40°C measured by both cation-
exchange HPLC (open symbol) and prothrombin tune assay (closed symbol) for 0.5
mgfmi, TFPI formulated in 10 mM IVa acetate and pI°I 5.5 containing 150
mhTaCl
(triangle) or 8 % (wlv) sucrose (square) or 4.5 %~ (wl~r) mannitol (circle).
CA 02450795 2003-12-19
VV~ 96!40784 F°CT/~1S96/09980
- 3~
Figure 9 shows two non-reducing SIBS gels for TFPI formulation samples in
mM NaPO~, 150 mM NaCl, and 0.005 % polysorbate-80 at pH 4 to pH 9 stored at
40~C for 0 days (lowers and z0 days (upper. No Ioss on TFPI is seen at 0 days.
However, at 20 days cleavage fragments of TFPI may be seen at the Iower pH
range
5 (i.e. pH 4 and pH 5). without being bound t~ a pa °culax theory, it
is believed that
these fragments may result from an acid catalyzed reaction,.
Finally, Table 2 shows the half life of remaining soluble TFPI at
X10°C for
various formulations. 0.5 mgt TFPI was formulated in these formulation
° conditions and incubated at 40°C. Samples were withdrawn at
predetermined time
10 intervals and loss of soluble and active TFPI were examined by the IEX-
H:PLC and
the PT assay. Half life for remaining soluble TFPI was then calculated by
performing a single exponential fitting to the IEX-HPL~ and PT assay results.
Elution of TFPI in displacement mode from chromatography resins using
polyionic compounds.
TFPI is first bound to a resin in a low salt buffer. Next a buffer containing
the polyionic compound used to elute TFPI in displacement mode, is pum through
the column. This compound binds stronger to the resin than TFPY and elisplaces
TFPI. For a pasitiveiy charged resin (anion exchanger) a negatively charged
compound is used and for a negatively charged resin (canon exchanger) a
p~sitively
charged compound is used.
Partially purified 1"FPI was used as starting material. TFPI, in 6 urea, 20
Tris, pH 8.0 was lo;~ded onto a column packed with an anion exchange resin, Q
Sepharose HP, to 20 mgfmL resin. After loading, the column was washed with 6 M
urea, ~ 20 mM Tris, pH 9Ø P1 was eluted d 10 mgiml of (lass H
(polyphosphate) in 6 M urea, 10 mM Tris, pH 9Ø
Elution of TFPI from chromatography resin ia~ aqueous buffer using polyionic
compounds.
CA 02450795 2003-12-19
WO 96140°784 PC'><YfJS96/09980
-34~
Table 2
tll2 (da at 40~
0.5 m ml TFPI formulated in: iEX-HPLC PT assa
~
mM Na Acetate, 150 mM NaCI~H _'10.5 s 17.2
5.5 0 ~
10 mM Na Citrate, 150 mM NaCI, 12.2 24.4
p1-! 5.5
i0 mM Na Acetate, 8/~ v~lv) Sucrose,~H48.2 42.2
5.5
10 mM Na Acetate, 4.5/~ Mannitoi, 47.'7 46.C
pH 5.5
i0 mM Na Succinate, 150 mM NaCI, 7.8 7 1 .0
pH 6.0
10 mM Na Citrate, 150 mM NaCf, 13.0 18.8.
H 6.0
10 mAA Na Phosphate, 150 mM NaCI, 7.8 1 1 .2
H 6.0
10 mM Na.Phosphate, 500 mM NaCI, 52.2 58.9
pH 6.0
[ 10 mM Na CBtrate, 150 mM NaCI, 10.0 14.8
pH 6.5 r
CA 02450795 2003-12-19
-35-
For a positively charged resin, a positively charged compound is used and for
a negatively charged resin, a negatively charged compound is used.
TFPI, in 3.5. M urea, I mg/ml polyphosphate, ~0 mM Tris, pH 5.9 was loaded
onto a canon exchange resin, SP Sepharose HP. After loading the column was
washed with a non-urea containing buffer, 10 mglmi polyphosphate, 10 rriM
sodium
phosphate, pH 5Ø TFI'I was eluted in the same buffer at pH 7.5, without
urea.
EXampIe 7
Selective eluticln of TFPI from ion exchange resins using polyionic
compounds.
Because of thw charged ends of TFPT, oppositely charged polyionic
compounds can bind to these ends. When the polyionic compound has a higher
strength of binding to T FPI than does the resin, the TFPI may be selectively
eluted
from the chromatography resin.
TFPI, in 3.5. M urea, 1 mg/ml polyphosphate, 50 mM Tris, pH 5.9 was loaded
onto a canon exchange ~°esin, SP Sepharose HP. After loading, the
column is Washed
with 6 M urea, Img/ml polyphosphate, 10 mM sodLium phosphate, pH 5.9. TFPI was
eluted in a 25 column volume gradient up to 20 mg/ml of polyphosphate. 7CFPI
starts
to elute at about 2-3 mg%ml of polyphosphate.
Example 8
Neutralization of polyionic compounds prior to chromatographic separation of
TFPI. TFPI can interact with charged polymers. This interaction may prevent
binding and purification to chromatographic resins. By neutralizing the
charged
polymer with an oppositely charged polymer, TFPI may bind to the resin.
In a buffer containing polyphosphate (Glass H), TFPI does not bind to Express
Ion STM (Whatman) and no purification is achieved. By mixing PEI into the
column
load, TFPI now binds to the resin and TFPI can be purified.
CA 02450795 2003-12-19
~~ 96/40784 PC'T/L3~9o6/09980
-3~-
Refolding arad purification of recornbinan,t human PI (rh °I) using
polyphosphate ~Cilass Il:) Facilitated Refolding process.
Inclusion bodies containing about ~0 g of rh pI were thawed by
r°emoving
S the containers from the -20°C freezer and incubating them in a cold
room at 4-1°C
overhead stirrer. The ~:ontents were mixed for apF~roximately 15 minutes, and
then
the absorbance of the solution is measured at 2$0 nm. If the absorbance is
greater
than the mixture was diluted with sufficient dissolution buffer to obtain an
absorbance
at 280 nm of I.0-1.1. 'The solution was incubated with gentle agitation for 15-
30
minutes, and then sufficient cysteine was added to give a cysteine
concentration of
0.1 . The solid I,~-cysteine was dissolved in approximately 50 ml of purified
water and added to the rc;fold mixture. The p~I was checked and adjus to pI~
10m2
if necessary. The refold mixture was incubated 'with gentle agitation fo,r 96-
120
hours.
after approximately 96 h, the refolding process was terminat by adjusting
the pI-I of the refold mixture to pIi 5.9 using glacial acetic acid.
Stin°ing was
continued for 90 minutes and the pI-I checked. More acid was added, if noes to
adjust the pFi to ~.9 +~- 0.1. ~ two-step fil tion process was a to remove the
°c t formed during previous steps and prepare tl~e acidified refold
mixture
for SP-Sepharose chromatographyo First the adidigied refold mixture i.s passed
through A Cuno 60LP .depth filter falter housing rr~odel 8ZP1) using a
peristaltic
_ pump ('~ - 3/a inch inne~° diameter silicon tubing).
The filter system was washed with 8-10 I, of deionized b Ivi ur before use.
The filtrate was collec in a 1 I, polyethylene . clc pressure was
maintained at a constant 20 PSI. Initial flow rate for a new filter was
approximately
5-6 L per minute. Filters were replaced when the :~~ow ra.te dropped below 1
I, per
CA 02450795 2003-12-19
W~ 96/40784 F~~°t°/L3S9~6/4D99~0
_37-
minute in order to maintain the back pressure at 20 PSI. The second stage of
the
filtration used a 0.45 micron filter c °dge (Sartorius Sartobran pI3 ~r
equivalent)
with a peristaltic pumpir~h system. After filtration, tif~e pfd was checked,
and adjusted
to p~I 5.9 if neces
The acidified, fi:~tered refold was loaded onto the equilibra SP S~epharose
column at a flow rate of approximately 80.0 mlf~nin. Flour rate was adjusted
to
accommodate overnight loading of the acidified f lte:red refold mixture. e:
column
eras equilibrated in 6 I~ urea, 20 mA~ sodium phosphate buffer pI~ 5.9 prior
to
loading. After loading, the column is gashed with 2 ~F of 6 Iii urea, 0.3 rvI
l~laCl,
20 msodium phosphate buffer, pIi :~.9 prior to the gradient elution st~:p. The
column flow rate was increased to 190-2 mllmin for the wash step and all
subsequent steps (fin velocity =- 47 c~n/hr~. °I'he product was eluted
from the
column using a linear s~dt gradient from 0.3 to 0. ii I~IaCI in 6 11~ urea, 20
m
250210). The total volc~me of the gradient was 71.0 liters or 13.0 C~. prhe
p~T ~f
the gradient buffers was 5.92 (+l- .02). Fractions are evaluated qualitativeay
using
SIBS PAGE and poaleti ba on the content of the correctly folded SC-59735
relative
to other misfolds and impurities. After ling the process s is referred to as
the 1.
'I°he pI-I of the 1 was next adjusted to pH S.0 with 2.S 1V a~i. The S
pool was concentrated 2-3 fold to approximately :~ L, using an Arnicon IBC-
10I.
uI tration unit con ° ° ~g an Arr~icon ~NI10 spiral cartridge
(10, NT.~J. cut-off
membrane). After conc~;ntration, the concentrated S pool was diafaltered
against 7
volumes of 6 urea, 20 mTris-~iCl buffer, ~a~i 5Ø a diafltraation was
considered complete vrhe.n the conductivity of the r~etentate was below 2 rnS.
The
diltered concentrate was drain from the ultrafnitration unit and the unit was
CA 02450795 2003-12-19
W~ 96/40784 PC'r/I1~9G/09980
_38-
washed with approximately 1 L of diafiltration buffer. The was is combined
with the
concentrate to form the ~-load.
An Amicon column (7.0 cm diaaa~eter) was packed with approximatel~r 700 ml
of Q-Sepharose high performance medium (Pharrr~acia Q-Sepharose ). The
column was packed with 20% ethanol at 20 si. T'he bed height after paclang was
approximately 18 cm. The column was equilibrated with 5 CF of 6 M urea; 0,02 M
TrisIHCl buffer, pH 8. The target for protein loading is 8-10 mg ~rotein/ml Q
Sepharose resin. The Q load was applied to the column at a flow rate 30-35
ml/min
(5Q cm/hr): After loading, the column was washed with approximately 5 Chi' of
6 M
, 20 mM TrisIHCI buffer, pH 8.0, or until the absorbance at 280 nm returned to
baseline. The product was eluted using a sodium chloride gradient from 0-0.15
M
I~TaCl in 6 M urea, 20 n~M Tris/I3C1 buffer, pH 8.C) over 25 column volumes.
The
first seven column volumes were collected as a single fraction, followed by 30
fractions of 0.25 column volume each.
Fractions are roc.~tanely analyzed by reducing and non-reducing SDS-PAGE
and size exclusion chromatography. Fractions are pooled based on aggregate
content
( < ~ % by SEC LC Method MSL 13929) and qualitative evaluation by SDS PAGE
to assess purity. The fractions are stored frozen at -20~C until pooled.
Acceptable Q Sepharose fractions were pooled, and the pH of the Frool was
adjusted to 7.2 using 2 M HCI. The pool was then concentrated approximately 5
fold
in an Amicon DC-1 a~ltraf~ltration system contaiir~ing a SlYl Amicon I'NI-10
cartridge (I0,000 MVVCCI spiral cartridge membrane). The concentrated Q 1'001
was
then diafiltered against seven column volumes of 2 M urea, 0.15 M NaCI; 20 mM
°um phosphate buffer, pH 7.. Following ultrafiltration, the solution
drained
from the ultrafiltration system. Approximately 100 ml of 2 M urea, 0.15 M
NaCI,
20 mM sodium phosphate buffer, pH 7.2 was circniated thr~ugh the
ultr~nltration
system for approximately 5 min. The rinse solution was combined with the
original
concentrate and the solutaon was filtered through a 0.4~ micron vacuum alter
unit
(Nalgene).
CA 02450795 2003-12-19
VVO 96J40784 PCTII~JS96J09980
_3g_
Refolding and Purification of rhTFPI using Polyethyleneimine (PEI)
Facilitated Refolding Process.
Inclusion bodies containing about 40 g of rhTFPI were thawed by removing
the containers from the -20°C fzeezer and incubating them in a cold
room at 4-10°C
for approximately 96 hours. The thawed inclusion bodies were then disperst;d
with a
high shear mixer t~ reduce the clumping that occurs during freezing. 'The
inclusion
body slurry urea vigorously blended for approximately 1 minute using a
polytron
homogenizer (Brinkman model PT45l80) or until the inclusion bodies were then
added to 40 L of 6 M urea 100 mM TrislHCl buffer pH 9.8 containing 300 mM
NaCl and 0.4 g!L PEI contained in a 100 L polyethylene tank equipped with an
overhead stirrer. The z:uxture was vigorously stirred for 20-30 min. a pH was
monitored and adjusted to pH 9.8 as neces ,Che absorbance of the dissolved
inclusion body mixture ~~as measured at 280 nm, and if the absorbance was
greater
IS than 2.1, the sample was diluted with 10 liters of the dissolution buffer
described
above to obtain an A280 value of 2.0-2.1. Gentle agitation was continued for
another
15-30 minutes. Next, the dissolved inclusion body solution was diluted with an
equal
volume of 1.0 M urea, :300 mM NaCI solution. Finally; L-cysteine was added to
give a final concentration;: of 0.25 mM. The solid L-~cysteia~e was dissolved
in 50 ml
of WFI and added as a solution to the diluted refa~ld, The pH was checked and
adjusted, if necessary. ~'he refold continued with gentle mixing for 96-120
hours
with periodic checks of thc~ pH, and adjustment to pH 9.8, if' necessary. The
progress
of the refold was monitored by Mon-S ration exchange and prothrombin dine
assays.
After approximately 96 h, the refolding process was terminated by adjusting
the pH of the refold to pH 5.9 using glacial acetic acid. Stirring was
contin,aed for
90 minutes and the pH choked. More acid was added, if necessary to adjust the
pH
to 5.9 !- 0.1.
A two-step ~ltraticsn process was used to remove the particulates that formed
during previous steps and prepare the acidified refold for SP-Sepharose HP
chanmatograph. First, the acidified refold is passed through a Cuno 60LP depth
filter
CA 02450795 2003-12-19
(filter housing model 8ZP1P) using a peristaltic pump (1/4 - 3/8 inch inner
diameter
silicon tubing).
The filter system was washed with 8-10 L of deionized 6 M urea before use.
5 The filtrate was collected in a 100 L polyethylene tank. Back pressure was
maintained at a constant 20 PSI. initial flow rate for a new filter was
approximately
5-6 L per minute. Filters were replaced when the flow rate dropped below 1 L
per
minute lIl Order to maintain the back pressure at 2.0 PSI. The second stage of
the
filtration used a 0.45 micron filter cartridge (Sartorius Sartobran pH or
equivalent)
10 with a peristaltic pumping system. After f Itration, the pH was checked,
and adjusted
to pH 5.9, if necessary.
The acidified, filtered refold was loaded onto the equilibrated SP Sepharose
HP column at a flow rate of approximately 80.0 mllmm. Flow rate was adjusted
to
accommodate overnight loading of the acidified filtered refold. The column was
then
15 washed with 5.5 column volumes of 6 M urea, 0.311~I NaCI, 20 mM sodium
phosphate
buffer, pH 5.9. The column flow rate was increased to 190-200 ml/mm for the
wash
step and all subsequent steps (linear velocity = ~ ~7 cm/hr). The product was
eluted
from the column using a linear salt gradient from 0.3 to 0.5 M NaC 1 in 6 M
urea, 20
mM sodium phosphate buffer, pH 5.9. The gradient was formed by delivering fi M
20 urea, 0.5 M NaCl, 20 mM sodium phosphate buffer into 6 M urea, 0.3 M NaCl,
20
rnM sodium phosphate buffer into 6 M urea, 0.3 M I~laCl, 20 mM sodium
phosphate
buffer. Limit buffer was pumped with a MasterflexT~ pump (model 7553-20) with
a
MasterflexT'~ head (model 7015.2I) at a flow rate of approximately 100 mi/mm
with
vigorous mixing using a ParatrolTM A mixer from Parametrics (model 250210).
The
25 total volume of the gradient was 71.0 liters or 13.0 Ctl. The pH of the
gradient
buffers was 5.92 (+/-.02).
Fraction collection was started when the column inlet conductivity reached
28.0 - 28.5 mS/crn as measured by the in-line Radiometer conductivity meter.
Forty
500 ml fractions (0.1 C ~I) were collected. A Pharma.cia Frac-300 fraction
collector
30 was used with numbered, 500 ml polypropylene bottles. When the fraction
collection
was stopped, the remainder of the gradient was collected as a pool.
CA 02450795 2003-12-19
w~ 96/40784 PC~°IUS96I09980
-41 -
Column fractions were assayed by A280, size exclusion ~iPLC, and in
addition, for informational purposes, SIBS PAGE, reverse phase I~PLC, and P'I'
assays. Fractions were pooled if they met the pooling criteria of containing
20~ of
less aggregate as determined by the in process SEC HPLC. Pooled SP Sepharose
fractions are referred to as the S Pool.
The pH of the S-pool was next adjusted to pI3 8.0 with 2.5 N Na~H. The S
Pool was concentrated 2-3 fold to approximately 2 L using an Amicon i~C-10L
ultrafiltration unit containing an Amicon 1'MI0 spiral cartridge (10,000 N..
cut-off
membrane). After concentration, the concentrated S Pool was diafiltered
against 7
volumes of 6 M urea, 20 mM Tris-~Cl buffer, pH 8Ø The diafiltration was
considered complete when the conductivity of the retentate was below 2 mS. The
diafiltered concentrate was drained from the ultrafltration unit and the unit
was
washed with approximately 1 L of diai'~ltration buffet°. The was is
combined with the
concentrate to form the (~-load.
An Atnicon column (7.0o cm diameter) was packed with approximately 700
ml of Q-Sepharose high performance medium (Pharmacia Q-Sepharose I-IP). The
column was packed in 20%~ ethanol at 20 psi. Ttae bed height after packing was
approximately 18 c.m The column was equilibrated with 5 CV of 6 M urea, 0.02 M
TrisBHCI buffer, pI~ 8. The target for protein loading is 8-10 mg protein/ml Q
Sepharose resin. The Q load was applied t~ the column at a flow rate 30-35
ml/min
(50 cmlhr). After loading, the column was washed with approximately 5 C~ of 6
M
urea, 20 mM Tris/~ICl buffer, pH 8.0, or until the absorbance at 280 nm
returned to
baseline. The product was eluted using a sodium chloride gradient from .15 M
NaCI in 6 M urea, 20 mM TrisB~ICI buffer, p~I 8.0 over 25 column volumes. 'The
first seven column volumes were collected as a - single fraction, followed by
30
fractions of 0.25 column volume each.
Fractions are routinely analyzed by reducing and non-reducing SI7S-PAGE
and size exclusion chromatography. Fractions are pooled based on aggregate
content
(5 % by SEC HPLC) and qualitative evaluation by SIBS PAtiE to assess purity:
The
fractions are stored frozen at -20~C until pooled.
CA 02450795 2003-12-19
WO 96/40784 PC'FNS96/09980
- 42 -
The ~-Sepharose fractions to be pooled were thawed by incubation at 2-
8°C,
pooled, and the pH of the pool was adjusted to 7.2 using 2 IvIHCI. The pool
was
then concentrated approximately 5 fold in an Amicon DC-1 ultrafiltration
system
containing a S1Y1 Amicon -10 cartridge (10,000 M1~VC~ spiral cartridge
membrane). The concentrated Pool was then diafrltered against seven column
volumes of 2 M urea, 0.15 M NaCI, 20 mM sodium phosphate buffer, pH 7.2.
Following ultrafiltration, the solution was drained from the ultrafiltration
system.
Approximately 100 ml of 2 M urea, 0.15 M NaCI, 20 mM sodium phosphate buffer,
pH 7.2 was circulated through the ultrafiltration system for approximately 5
min.
'The rinse solution was combined with the original concent~°ate and
filtered through a
0.45 micron vacuum flier unit (Nalgene).
Solubilization, refolding, and purification of rhTFPI from inclusion bodies
using polyphosphate in the absence of chaotropes such as urea (GDS 5327089,92)
About 2 g of rhTFPI (43 m/ inclusion body slurry containing 4~ mg/ml
rhTFPI) was dissolved with mixing in 4 L of 50 mM; 'I'ris buffer, pH 10.5
containing
4 g/1 polyphosphate (Glass Ii, FMC Corporation) 2-8°C. Sufficient
cysteine and
cystine was added to male the solutions 0.1 mM and 0.05 mM respectively. The
pH
was maintained at pH 10.5 with 1 N NaOH. The refold solution was incubated at
2-
8 ° C with gentle mixing for 72-96 h.
column volumes of 0.4 ~O Glass H, 20 mM sodium phosphate pH6 buffer. The
column was eluted using a linear pH gradient Pram 0.4 9o Glass H, 20 mM sodium
phosphate buffer pH 6 to 0.4 3o Glass H, 50 mM 'T'ris pH 8 buffer. Fractions
were
collected and analyzed by SDS PAGE. Relatively pure rhTFPI could be refolded
and
purified in this manner.
CA 02450795 2003-12-19
VVO 961~07~4 PC"Ti'~JS96/09980
- 43
improved solubility of rhTFPi in water by :formation of a complex between
TFPI and polyphosphate (GDS 5327046-47)
About 10 g of purified rhTFPi in about 1 litsmr of 2 IvI urea, 12S mr4 sodium
S chloride, 20 sodiulr phosphate pI~ 7.4 buffer was thawed by incubation at 2-
S ° G
for 1g-36 h. Sufficient dry urea was added to make the solution 6114 in urea.
The
solution was then filtered through a 0.2 micron filter. Five g of
polyphosphate glass
(Glass Ii, FItIC) was dissolved in SO rnl of 6 r4 urea, adjusted to p 7 with 1
I~
removed from the ultrafiltration unit. The ultrafiltration unit was washed
with about
1S0 ml of purified water and the was added to the protein concentrate. The
fanal
15 protein concentrate contained almost 10 g of protein in 400 ml of water
(about 24
mg/rnl protein). The normal solublllty of rhTFPI ire water is less than O.S
mg/ml.
I3se of cationic polymers for removal of ~'. cn~ri contaminants front TFPI ~Il
lysates and refractile bodies.
20 The use of cationic polymers to precipitate a.nd remove E coli con inants
- from crude I interanediates (lysates, refractile bodies) can signifi tly
improve
subsequence process operations (refolding, chronnatography etc.) A random
screening of cationic polymers identified candidates which selectively
precipitate
bacterial contaminants while TFPI remains in solution. Specifically, tz
polymer
2S 624 precipita substantial amounts of bacterial contaminants, while leaving
I in
solution in an aqueous environment.
Solubilized TFPi refractile bodies (in 3.S ltd guanidine hydrochloride, 2
sodium chloride, SO mlvl S, SO mIVI dithiothrnitol, pI-I7.1) was the starting
material used for a polyrl~er screening experiment. This material was diluted
10 fold
30 into a 0.S%~ solution of various polymers. The prE:cipitates from this
experiment
CA 02450795 2003-12-19
W~ 9614084 PCT1US96I09980
-44-
were analyzed by SDS-PAGE for the presence of TFPI. Betz polymer 624
precipitated substantial amounts of contaminants, no TFPi, and resulted in a
clear
aqueous solution.
Fyam=1_34 .
The use of aqueous two phase extraction with a polyethylene glycol (PEG),
polyphosphate, urea system offers processing advantages fcir TFPI
purification.
Typical aqueous two phase systems consist of two polymer systems (e.~., PEG
and
dextran) or a polymer and salt (e.~., PEG and sulfate). The system described
here
has advantages in that the polyphosphate chain length can be optimized for the
separation, is inexpensive and is s ific in removing problematic contaminants
from
TFPI refractile bodies known to interfere with refolding and chromatography
(native
polyphosphate and associated divalent metals).
TFPI refractile bodies were solubilized in '7 M urea, 10 mM CAPS, 1 l
monothioglycerol pIilO. Polyphosphate and PEG of different chain lengths were
added to form two phases. Upon phases separation, the TFPI partitioned into
the
PEG rich upper phase, leaving the polyphosphates and associated contaminants
in the
lower phase. Separation is effected by both PEG and polyphosphate chain length
and
can be optimized by varying both.
Charged polymer facilitated refolding of rccombinant tissue plasminogen
activator (t-PA) from ~ crab inclusion bodies
Five grams (wet weight) of inclusion bodies containing about 2 grams of
recombinant tissue plasminogen activator are added to about 1 liter of 0.5 9'
Glass gI,
50 mM Tras buffer pkI 10.8 containing I mM reduced glutathione (GS~I) and 0.2
mM glutthione disulfide (GSSG). The mixture is thoroughly blended using a
polytron (l3rinkman) horr~ogenizer for 2-3 minutes to thoroughly disperse the
inclusion bodies. The mixture is incubated with mixing using an overhead
stirrer for
15 minutes while the pI~ is maintained at 10.5-10.9 using I N Na~PI. The
mixture is
then gently mixed for 48-~2 hours at 2-BoC.
Examp
CA 02450795 2003-12-19
-45-
Charged polyrme~- facilitated refolding of bovine somatotropin from ~. coli
inclusion bodies
Ten grams (wet weight) of inclusion bodies containing 5 grams of bovine
somatotropin are added to about 1 liter of 1~/o Glass H, 5(1 mI~ Tris buffer
pH 10.5.
The mixture is thoroghly blended using a polytron (~rinkman) homogenizes for 2-
3
minutes to thoroughly disperse the inclusion bodies>. The mixture is incubated
with
mixing using an overhead stirrer for 25 minutes while the pH is maintained at
10.4
10.6 using 1 N Na~H. solid cysteine (121 mg) is added to make the reaction 1
mM
cysteine, and the refolding reaction is mixed for 48-;7? hours
The present invention has been descriir,ed with reference to specific
embodiments. However, this application is intended to cover those changes and
substitutions which may be made by those spilled in the art without departing
from
the spirit and the scope of the appended claims.
