Controlledand modulatedrelease of basic fibroblastgrowthfactor Elazer R. Edelman Biomedical Engineering Center, Harvard-Massachusetts lnstm!te of Technology Divxiwn of Health Soences and Technology, Cambridge, MA 02 139 and Department of Medicine, Cardiovascular Dwision. Brigham and Women’s Hospital and Harvard Medical School, Boston~ MA 02 1 15, USA
Edith Mathiowitzand Robert Langer Department of Chemical Engineering, Massachusetts Institute of Technology. Cambridge, MA 02 139, USA (Received 23 October 1990: revised 19 December 1990; accepted 2 1 December 1990)
Deparfments of Biological Chemistry, Molecular Pharmacology and Surgrcal Research, Children’s Hospital and Harvard Medical School, Boston, MA 02 115, USA
Basic fibroblast growth factor has mult~variate effects in stimulating cell growth and the processes that surround tissue repair. Pathophysiologic studies have been hampered by the stability of the compound. Though very potent, basic fibroblast growth factor is rapidly degraded when injected or ingested. Controlled release of basic fibroblast growth factor would allow for examination of the chronic effects of this compound. Conventional matrix polymer-based release devices were fabricated and basic fibroblast growth factor released in a sustained fashion, but 99% of basic fibroblast growth factor mitogenic activity was lost. The source of these losses was identified and preventative measures examined. Preservation and stabilization of basic fibroblast growth factor was accomplished by binding the factor to heparinSepharose beads. This permitterl#olonged storage, repeated handling, and the encapsulation of basic fibroblast growth factor witbin a microspherical controlled-release device using a naturally occurring polymer material, alginate. Encapsulation was accomplished with 77% efficiency and 87.5 f 12% of the basic fibroblast growth factor was released in a biologically active form. Release activation and regulation was achieved when cleavage of the basic fibroblast growth factor-heparin bonds was enhanced (e.g. by enzymatic bond cleavage with heparinase). Kinetic profiles were identified for a variety of experimental conditions and the effects of the controlled release of basic fibroblast growth factor on BALBc/3T3 fibroblasts examined. Keyw0rd.s: Controlled release, fibroblasts, heparin
Polypeptide growth factors are an important class of agent that control cell growth and activity. These compounds have in viva half-lives in the order of seconds to minutes and are occasionally unstable in solution’_ At the same time they are capable of exerting their influence at concentrations within the nanogram and picogram rangezm4.Controlled studies of their action on cells and on laboratory animals have been hampered, because it is difficult to administer these compounds over long periods of time in bioactive form at physiological doses. Previous efforts at incorporating growth factors within a sustained release vehicle have been only partially successful. The bulk of these attempts have focused on the use of matrix-type delivery systems, using solvent casting, with hydrophobic polymer materials such as Correspondence 0 1991
to Professor
Butter-worth-Hememann
R. Langer.
ethylene-vinyl acetate copolymer (EVAc). These materials have been well characterized for matrix-type controlled release systems, can be demonstrated to provide reliable release kinetics, and are relatively inert. Epidermal growth factor incorporated within EVAc matrices was released with predictable kinetics and intact cell-stimulating activity5. Similarly, controlled release of transforming growth factor-b was achieved with this polymer system and reversible inhibition of mammal gland growth was noted6. Basic fibroblast growth factor (bFGF) in culture is a mitogen for a number of cell types such as fibroblasts, endothelial cells and smooth muscle cells2-4. It also promotes differentiation and survival of neuronal cells’ and is a potent angiogenesis factor in viva*. Because of these properties, bFGF may be useful in enhancing wound healing and damaged tissue repair, and may beof potential importance in
Ltd. 0142-9612/91/070619-08 ~;omateria/s
799 1, Vol 12 September
679
Controlled and modulated release of bFGF: E.R. Edelman et al.
research on naturally-occurring injury, such as in atherogenesis’, ischemia and infarction. A number of attempts at sustaining the release of bFGF have been reported. For example, bFGF was bound within collagen gels and induction of angiogenesis was demonstratedg. Unfortunately such systems have unpredictable release characteristics. In this paper, we present a system that has been developed and characterized for the controlled and modulated release of bFGF. By taking advantage of heparin’s ability to stabilize bFGF, we were able to bind the growth factor to a Sepharose carrier from which it was released slowly over time. The carrier, laden with bFGF, was then embedded within microspheres of alginate gel to create a single controlledrelease device for the growth factor. Kinetics profiles of bFGF release were identified for a variety of experimental conditions and the effects on biological activity were evaluated. Manipulation of the heparin-bFGF bonds was achieved enzymatically and ionically and enabled us to activate and regulate bFGF release as desired.
mitogenic assay evaluates the relative level of incorporation of tritiated thymidine into the DNA of BALBc/ 3T3 cells after the addition of aliquots from given sample solutions’3. One unit of growth factor activity was defined as the amount of growth factor in 250~1 that was required to stimulate halfmaximal DNA synthesis. The number of units in a sample was calculated by comparing cell stimulation for each sample to this value. Radiolabelled bFGF was detected using a y counter (Beckman Industries). The total number of units present in samples was extrapolated from the initial ratio of unlabelled to labelled bFGF and the efficiency of the y counter, (i.e. disintegrations per minute/unit of bFGF). In virtually all cases, radiolabelled results were compared with the kinetics obtained from the thymidine incorporation assay. The ratio of the number of units of bFGF, determined from the thymidine incorporation mitogenic assay to the number of units detected radioactively, served as an index of retention of biological activity.
MATERIALS
Controlled
Bovine serum albumin (BSA, m.w. 68 800). heparin (m.w. 12 000 - 15 000). dichloromethane, myoglobin (16 000 m.w.) and sodium alginate were obtained from the Sigma Chemical Co. (St Louis, MO, USA). Ethylene-vinyl acetate copolymer (EVAc, 40% vinyl acetate, ELVAX-40P) was obtained from DuPont Chemical Co. (Wilmington, DE, USA) and washed according to procedures described elsewhere”, tritiated thymidine from New England Nuclear (Boston, MA, USA), and heparin-Sepharose beads from Pharmacia LKB Biotechnology Inc. (Piscataway, NJ, USA). Human recombinant bFGF corresponding to the 146 amino acid form of the growth factor2-4 was obtained from Takeda Industries (Osaka, Japan). The growth factor was stored as a concentrated solution of 1 mg/ml at -80°C. Stock solutions were made after an initial 1: 100 dilution in 0.15 M saline and stored at -20°C to be used after further 1 :200 dilution as the final concentration in all experiments. lodinated bFGF was prepared using a modification of the method of Bolton and Hunter (gift of Dr S. Doctorow)“. The specific radioactivity of the ‘251-bFGF was 85 nCi/ng. at a concentration of Aliquots were stored at -20°C 1.2 pg/ml. These aliquots were thawed, mixed with the unlabelled bFGF as liquids in known ratios and stored in polyethylene tubes coated with BSA (50 mg/ml, 2 ml). The albumin prevented adhesion of bFGF to the walls of the vials and increased overall bFGF recovery after 4 wk of storage 8.7 fold above the recovery of similar amounts and concentrations of bFGF stored in untreated vials. bFGF tends to adhere avidly to polystyrene and polypropylene even with BSA coating, and vials/tubes made of these materials were avoided. The enzyme heparinase, heparin lyase, EC 4.2.2.7, an a&linked sulphated polysaccharide with a mol wt of approximately 42 900, was kindly supplied by Drs ReeseComfort and Lechband after production and purification according to Yang et al.‘*.
METHODS bFGF assay The biological activity of bFGF in sample solutions was measured by quantifying the ability of these solutions to stimulate DNA synthesis in tissue culture cells. This
620
Biomaterials
199 1. Vol 12 September
release
of bFGF from EVAc matrices
bFGF (4 ~1 at 40 pg/ml) and ‘251-bFGF (2 ~1 at 1.2 pg/ml) were combined with 500 mg of BSA in 1 ml of 0.1 5 M NaCI. The combined solution was lyophilized to a fine powder and added to 10 ml of a 10% (w/v) solution of dichloromethanedissolved EVAc. This mixture was poured into a precooled glass mould, surrounded by crushed dry ice. This provided for instantaneous freezing, producing a sheet of EVAc with BSA-bFGF dispersed homogeneously”. 14. The sheet was transferred to 4°C with a cooled spatula for 2 d and then placed under house vacuum for two additional days. Smaller pieces were cut from this large sheet and release kinetics at room temperature were determined by following BSA and bFGF levels as the saline buffer was changed on a regular basis over 2-3 wk. The spectrophotometric absorption (220 or 280 nm) of aliquots from the NaCl into which the matrices had been released was used to determine the concentrations of BSA; bFGF levels were followed using both y counting and induction of thymidine incorporation in BALBc/3T3 cells.
bFGF inactivation
and denaturation
bFGF adheres to glass and plastic with great avidity. It is possible that part of the loss in observed growth-factor activity with encapsulation was the result of the physical loss of growth factor within the storage container. ‘251-bFGF was therefore placed in 3 ml polyethylene vials (Eppendor-f), stored at 4°C and removed after 0 h, 2 h, 3 d and 74 d. A second potential source of loss in growth factor is degradation or denaturation of the compound during prolonged storage and in the steps involved in the fabrication of a controlled-release device. Device formulation often involves operations on both the incorporated substance and the incorporating material. This might involve vortexing or lyophilization of the substance or melting, pressing or dissolving the encapsulating material. These processes often require organic solvents or extremes of temperature, pressure, pH or ionic strength. Hence, the biological activity of bFGF was measured before and after gentle vortexing for 15-30s; exposure to fumes of the organic solvent dichloromethane; incubation at 4 and 50°C for 1 h; or immersion of the vial containing the growth factor within liquid nitrogen for 5-l 0 min.
Controlled
Effect
of BSA on bFGF stability
BSA and other similar glass and plastic proteins.
compounds
to prevent
To determine
whether
used to coat adherence
similar treatment
used to store the bFGF would coated with a solution
have been
the non-specific improve
of BSA (125
recovery,
mg/ml),
the bFGF.
of
Modulation of bFGF release microspheres One
problem
with
the
device fabrication.
As the alginate
was
compared
to vials that had been coated with similar volumes
of 0.1 5 M NaCI.
barrier to diffusion, buffers
and the enzyme was
studied.
bFGF
has a well-defined
characterized
affinity
and purified
for heparin
by passage
immobilized
heparinzm4. Gospodarowicz
shown
the addition
that
prevents
of heparin
in temperature.
Heparin
added to the bFGF solutions before
at 50°C
or 5-l
potential
stable
Aliquots
of bFGF
solid
have
This
1 h with
changes
was, therefore,
to dichloromethane
carrier
for the
(1.2 pg/ml)
examined
liquid
range
from the specific
beads.
on a scheduled
basis.
The
to the beads
amount
potential
of four
media
samples
M NaCl
bFGF to
taken
point.
Known
quantities
Incubated were
of labelled
with
constructed
alginate
(1.2%
beads through
by dropping
of calcium instantaneously
calcium
a mixed
chloride
solution. A 25-gauge obtained
(1.5%
as the mixture
cross-linked
by limiting
hardenrng
chloride
efficiency
of encapsulated
Spheres
envelopes
was determined growth
were
were in the
stripped
from the heparin-Sepharose were
NaCI. The solution
beads residing was changed
with the unencapsulated compared
have been released
assays
for microspheres
6 ,L/Iof bFGF and 2 ~1 of lz51-bFGF heparin-Sepharose
the bFGF was
beads with 3.0 M NaCl
and mitogenic
defined
by comparing
factor with that initially
By way of confirmation,
and both the radioactive kinetics
w/v).
a hardening
to 5 min immersion
in solution.
assay.
of sodium
entered the hardening
sphere
present
assay were
solution
bath.
Encapsulation
Release
were
Microspheres
needle was used to set the microsphere
similarly
the amount
bFGF
beads.
a needle into a beaker containing
solution
and
unlabelled
bound to 125 within
500
on a regular
according
ml) in an identical
BSA
absorbance
at
were constructed
mg/2
ml)
fashion
at 220
examined
in 2 ml of saline
basis.
alginate
(125
,ul of
pg/ml.
or
of
myoglobin
to devices
fabricated
spectrophotometerically
and 280
nm and myoglobin served as additional
controls
above.
for the experiments
described
Statistics All data are expressed
as the mean
t standard
error of the
t test and an analysis of variance
used to compare
bFGF release Figure
I
different
(ANOVA)
groups.
As predicted
approximately after
cumulative from
factor
BALBc/
3T3 cells in culture.
in matrix
and
of bFGF
activity was
was
101
When
made
2 h
t- 0.5%
of the activity
iodinated
bFGF 1.9%
and remained proved
of the of
involved
to
after
were
as the in
that,
in
present. this
to the walls
after the solution was removed. an
immediate
Biomateuals
percentage
polyethylene
was no longer
of the bFGF adhered be
to
assayed, of
the same determination
storage
demonstrated
diluted
instantaneously
maintained,
+ 0.2%.
43.1
i
activity
< 1% stimulation
of bFGF
was
was
20.9
was
at lOOO,~~g/rnl
mitogenicity
M
to
had been released
the steps and procedures
basis and, as
~1 of 0.15
and
The source of these losses was
and inactivation
solutions
recovery
of the matrix
biological
was lost, as there
by examining
and
fabrication.
Denaturation When
factor
all of the
released
release
EVAc matrices
size of the BSA carrier’0,‘4,
of the growth
Virtually
investigated
the shape
and particle
37%
2 wk.
physical
of bFGF from solvent-cast
the loading
biological
mg of the
the
activity
over time. from
from EVAc matrices
shows
biological
containing
incorporation
with
release at 4 10 nm. These mrcrospheres
50 ng/ml
of bFGF observed
to the thymidine
on a regular
(100
performed.
beads, results from the radioactive to the amount
Thus, at 4°C.
RESULTS
in NaCI) with bFGF bound heparin-Sepharose
formed size
and
heparin-Sepharose
low
incubated
were
residing
of calcium
mean. A Student’s
bound bFGF in
enhanced.
from 0 to 200
retention
beads
alone,
were
Encapsulation of heparin-Sepharose calcium alginate microspheres
from
heparinase
by monitoring
at
whereupon
constructed
ranging
bioactivity
Microspheres
with at 37°C.
storage
beads
for bFGF. BSA release was followed
release of
and compared
10 PI aliquot
at each time
and 37°C
mg/2
activation
and 6 ,LIIof ‘251-bFGF and 200
at concentratrons
release
the micro-
is desrred
were
to
was coincorporated
prolonged
release
microspheres
(0.1 5 M NaCI) changed
of
for
mg of heparin-Sepharose
4 and
for
extrapolated
of radioactive
allow
mixed
and the
in 0.15
bFGF
followed in 0.1 5 M
bFGF within
can be raised and release
alginate
250
stimulate
was
released
heparinase
until enhanced
both
the beads and was changed
at each time period was obtained
from the release
microcapsules
16 ~1 of bFGF (40 ,ug/ml)
a
The percentage
factor from the beads was followed covered
kinetics
at 4°C and maximal
might
the temperature
within
activity is temperature-sensitive
factor.
incubated
activity of the bFGF. Subsequent
~1). The solution
the mitogenic
were
and then
feature
heparinase
as
growth
inactivation
temperatures
with
fumes,
bound to the beads was determined bound
to predictably
release
experiment,
The enzyme’s
calcium
to liquid nitrogen.
were
over a concentration
of units of bFGF
present
in pH or
activity assayed
the heparin-Sepharose
number
of bFGF
(20pg/ml)
beads
2~1 of lz51-bFGF
initial lz51-bFGF
(500
and Cheng” to solutions
0 min exposure
Heparin-Sepharose
growth
relative
and mitogenic
1 h exposure
and after
incubation
with
with
losses In activity that accompany
elevation
and has been
over columns
with
offers a minrmal
can be depleted
with the heparin-Sepharose-bound spheres.
mrcrospherical
commences
envelope
heparinase bFGF
for identical
NaCI. In another
alginate
in 2 ml of 3.0 M NaCl and compared
from microcapsules
on bFGF stability
et al.
The ability of high salt-concentration
release
data obtained
is that release
the microspheres
hours after fabrication.
E R. Edelman
use of the alginate
vials were
of bFGF
of bFGF
from calcium
system
used to store
of heparin
release
controlled-release
dry, then
Effect
modulated
of the vials
let to stand and
Recovery
and
phenomenon
199
1. Vol
2 h
vials, Use of hiatus,
of the vials
Adhesion as
12 September
only
also an
621
Controlled and modulated release of bFGF: E.R. Edelman et al.
10’
lo” IO’
&-.-
?I
’r
,o‘?
6 .
.
0
I..
I..
1..
.2io”
14
7
TIME
(days)
Figure 1 bFGF release from solvent case EVAc matrices was followed using a ‘25/ radiolabel (0) and compared to the mitogenic potential of bFGF released from similar matrices (e), as evidenced by the induction of DNA synthesis within BALBc/3T3 cells. Data are displayed in both absolute units of bFGF (left ordinate) and the percentage of total incorporated bFGF (right ordinate) on a semilog scale and represents the average release from four matrices with each assay performed in duplicate. Error bars are smaller than data symbols.
additional 4.5% adhered over the next 74 d. BSA-coated vials, in contrast, had only a 2% adherence over the first 2 h, but a significant loss in activity after prolonged storage; only 37.5 f 1 .l% of the initial growth factor activity was recovered when bFGF was stored in BSA-coated vials stored for 4 wk at 4°C. In addition, when solutions were diluted immediately from the stock and then exposed to the conditions that might be observed with any of the various standard encapsulation techniques, significant losses of bFGF activity were observed. In some cases, these losses surpassed those associated with prolonged storage. After bFGF solutions were exposed to the fumes of the organic solvent dichloromethane, placed in direct contact with liquid nitrogen or incubated at 5O”C, biological activity was diminished by 94.1 k 0.2, 93.6 f 0.3, 90.2 f 0.3%. respectively (Table 7). Interestingly, despite the potential for denaturation with physical shaking, 30 s of gentle vortexing did not significantly reduce the biological activity of bFGF. Only 9% of the biological activity of the growth factor was lost with this procedure.
Preservation
of bFGF stability with heparin
Encapsulation of heparin-Sepharose-bound calcium alginate microspheres
Table 1 Recovery of bFGF after exposure to prolonged storage, extremes of temperature or organic solvents. Heparin is partially protective of the temperature affects but does not alter the inactivation of the growth factor by dichloromethane fumes Recoven/ (%)
80
.
F . ..o
l’
.
O
Storage 4°C 4 wk 50°C N,(I) CH,CI,(g)
622
Biomaterials
Hepann
5.4 9.8 6.4 5.9
21.3 19.6 24.1 4.1
t t f t
1.6 0.3 0.3 0.2
199 1, Vol 12 September
* + f f
1.9 2.0 4.0 0.5
0
0
o
o”
‘0 JO
I,0
20 0 a ,.,.,.).“.““““‘I”““/ 0 7
14
TIME Control
bFGF in
When the heparin-Sepharose beads were entrapped within calcium alginate microcapsules, 26% of the initial growth
60.
We confirmed the finding that heparin potentiated the activity of bFGF, stabilized the effect of heat on bFGF15 and found that it extended to colder temperatures as well (Table I). Heparin added to the bFGF solutions at a concentration of 20 pg/ml potentiated bFGF activity, increasing the mitogenic
Exposure
potential to 123.8 f 9.1% of the original material. However, when the same amount of heparin was added to bFGF that had previously been stored in polyethylene vials at 4°C for 4 wk, biological activity was not augmented and remained at 4.3 f 0.6% of the initial activity. Heparin was protective of the losses with storage and environmental extremes when added to solutions of bFGF before storage, direct exposure to liquid nitrogen or incubation at 50°C. Biological activity increased 3.9, 3.8 and 2.0 fold, respectively, above values obtained without heparin (Table 1). However, heparin did not prevent losses from exposure to dichloromethane fumes, as 95.9 + 0.5% of the growth factor was inactivated. Heparin immobilized to a Sepharose carrier provided a convenient substrate for handling the growth factor. bFGF was bound avidly and without saturation after incubation for 1 h with heparin-Sepharose beads. Over the range of concentrations used, we observed fairly constant binding of 77.0 + 1.3% of the initial growth factor to the beads. As the amount of heparin immobilized to the Sepharose beads was very much greater than the amount of bFGF, bFGF binding never saturated. At the highest bFGF concentration, approximately 176 units (35 ng) of growthfactoradhered to each mg of bead. Binding was confirmed in two independent ways. When the heparin-Sepharose beads bound with bFGF were placed in 300 ml of 3.0 M NaCI, 63% of the bound growth factor was removed and detected in the buffer solution within the first 30 min; the remainder leached off over the next 8 h. Similarly, when the same beads were placed in saline containing heparinase over a range from 0 to 200 units/ml, bFGF was removed from the beads with dosedependent kinetics and maintenance of 88.3 f 4.7% of growth factor activity. When the beads were placed in 0.15 M NaCI, sustained release of bFGF was prolonged for over 2 wk (Figure 2). More importantly, in contrast to the near total loss of biological activity with release from EVAc matrices (Figure 7), 97.5 + 13.7% of the bFGF released from the beads had retained its biological activity (Figure 2).
21
28
(days)
Figure 2 bFGF release from 125 mg of heparin-Sepharose beads resting in 0.15 M NaCI; this is plotted as cumulative percentage ofinitial encapsulated growth factor against time. Physical determination with radiolabelled bFGF (OJ is contrasted with assay of mitogenic ability I@), using incorporation of tritiated thymidine. The first two points on this curve represent data obtained at 30 and 60 min. respectively.
Controlled
and
modulated
release
of bFGF:
E.R. Edelman
et al.
. . . . 1
.
.
. 0 i
a 7
I 4
TIME
1
2
28
I 4
7
TIME
(days) Figure
. . .
both
The
4
release
from
bFGF
are presented
error bars represent
0
0
(days)
sensitivity
and heparinase,
(m). Data
0
temperature
heparin-Sepharose
of
at 3 7°C (+)
and 4°C
as the cumulative
fold
(Figure
when
200~g/ml presence
mitogenic
1
14
I
TIME Cumulative
displayed
against
microspheres
time
containmg
Sepharose-bound initial
bFGF
(a)
,ug/mlo
the
Data
potential active
with
with
incorporation
“‘1
standard
factor
radiolabel
as the
against
(0)
of tritlated
at 37°C
alginate
from
and
(6)
activity cells
The kinetics
as the burst of activity
retarded
(Figure
3a)
activity.
without
The
release (0)
ionic
heparinase
of
DISCUSSION
reducing
for
culture.
were remarkably
at each point in time, with 87.6 retained
its bioactivity.
intact for the duration and in extended
i
2.0%
The alginate
The
beyond
strength
release
To increase
of
radio-
similar (P = NS)
of the bFGF having remained
under all conditions
successful
release
Figure
3
demonstrates
increased
that
the total amount
at 2 wk from 2 5 to 85%. at physiological
the
capsules
were
with
4°C cumulative
produced
heparinase
heparinase
of
bFGF. heparinase
from the microspheres
enzyme
the temperature
and stored. were
fabricated
when
and released
release from the beads was
Edith Mathiowitzand Robert Langer Department of Chemical Engineering, Massachusetts Institute of Technology. Cambridge, MA 02 139, USA (Received 23 October 1990: revised 19 December 1990; accepted 2 1 December 1990)
Deparfments of Biological Chemistry, Molecular Pharmacology and Surgrcal Research, Children’s Hospital and Harvard Medical School, Boston, MA 02 115, USA
Basic fibroblast growth factor has mult~variate effects in stimulating cell growth and the processes that surround tissue repair. Pathophysiologic studies have been hampered by the stability of the compound. Though very potent, basic fibroblast growth factor is rapidly degraded when injected or ingested. Controlled release of basic fibroblast growth factor would allow for examination of the chronic effects of this compound. Conventional matrix polymer-based release devices were fabricated and basic fibroblast growth factor released in a sustained fashion, but 99% of basic fibroblast growth factor mitogenic activity was lost. The source of these losses was identified and preventative measures examined. Preservation and stabilization of basic fibroblast growth factor was accomplished by binding the factor to heparinSepharose beads. This permitterl#olonged storage, repeated handling, and the encapsulation of basic fibroblast growth factor witbin a microspherical controlled-release device using a naturally occurring polymer material, alginate. Encapsulation was accomplished with 77% efficiency and 87.5 f 12% of the basic fibroblast growth factor was released in a biologically active form. Release activation and regulation was achieved when cleavage of the basic fibroblast growth factor-heparin bonds was enhanced (e.g. by enzymatic bond cleavage with heparinase). Kinetic profiles were identified for a variety of experimental conditions and the effects of the controlled release of basic fibroblast growth factor on BALBc/3T3 fibroblasts examined. Keyw0rd.s: Controlled release, fibroblasts, heparin
Polypeptide growth factors are an important class of agent that control cell growth and activity. These compounds have in viva half-lives in the order of seconds to minutes and are occasionally unstable in solution’_ At the same time they are capable of exerting their influence at concentrations within the nanogram and picogram rangezm4.Controlled studies of their action on cells and on laboratory animals have been hampered, because it is difficult to administer these compounds over long periods of time in bioactive form at physiological doses. Previous efforts at incorporating growth factors within a sustained release vehicle have been only partially successful. The bulk of these attempts have focused on the use of matrix-type delivery systems, using solvent casting, with hydrophobic polymer materials such as Correspondence 0 1991
to Professor
Butter-worth-Hememann
R. Langer.
ethylene-vinyl acetate copolymer (EVAc). These materials have been well characterized for matrix-type controlled release systems, can be demonstrated to provide reliable release kinetics, and are relatively inert. Epidermal growth factor incorporated within EVAc matrices was released with predictable kinetics and intact cell-stimulating activity5. Similarly, controlled release of transforming growth factor-b was achieved with this polymer system and reversible inhibition of mammal gland growth was noted6. Basic fibroblast growth factor (bFGF) in culture is a mitogen for a number of cell types such as fibroblasts, endothelial cells and smooth muscle cells2-4. It also promotes differentiation and survival of neuronal cells’ and is a potent angiogenesis factor in viva*. Because of these properties, bFGF may be useful in enhancing wound healing and damaged tissue repair, and may beof potential importance in
Ltd. 0142-9612/91/070619-08 ~;omateria/s
799 1, Vol 12 September
679
Controlled and modulated release of bFGF: E.R. Edelman et al.
research on naturally-occurring injury, such as in atherogenesis’, ischemia and infarction. A number of attempts at sustaining the release of bFGF have been reported. For example, bFGF was bound within collagen gels and induction of angiogenesis was demonstratedg. Unfortunately such systems have unpredictable release characteristics. In this paper, we present a system that has been developed and characterized for the controlled and modulated release of bFGF. By taking advantage of heparin’s ability to stabilize bFGF, we were able to bind the growth factor to a Sepharose carrier from which it was released slowly over time. The carrier, laden with bFGF, was then embedded within microspheres of alginate gel to create a single controlledrelease device for the growth factor. Kinetics profiles of bFGF release were identified for a variety of experimental conditions and the effects on biological activity were evaluated. Manipulation of the heparin-bFGF bonds was achieved enzymatically and ionically and enabled us to activate and regulate bFGF release as desired.
mitogenic assay evaluates the relative level of incorporation of tritiated thymidine into the DNA of BALBc/ 3T3 cells after the addition of aliquots from given sample solutions’3. One unit of growth factor activity was defined as the amount of growth factor in 250~1 that was required to stimulate halfmaximal DNA synthesis. The number of units in a sample was calculated by comparing cell stimulation for each sample to this value. Radiolabelled bFGF was detected using a y counter (Beckman Industries). The total number of units present in samples was extrapolated from the initial ratio of unlabelled to labelled bFGF and the efficiency of the y counter, (i.e. disintegrations per minute/unit of bFGF). In virtually all cases, radiolabelled results were compared with the kinetics obtained from the thymidine incorporation assay. The ratio of the number of units of bFGF, determined from the thymidine incorporation mitogenic assay to the number of units detected radioactively, served as an index of retention of biological activity.
MATERIALS
Controlled
Bovine serum albumin (BSA, m.w. 68 800). heparin (m.w. 12 000 - 15 000). dichloromethane, myoglobin (16 000 m.w.) and sodium alginate were obtained from the Sigma Chemical Co. (St Louis, MO, USA). Ethylene-vinyl acetate copolymer (EVAc, 40% vinyl acetate, ELVAX-40P) was obtained from DuPont Chemical Co. (Wilmington, DE, USA) and washed according to procedures described elsewhere”, tritiated thymidine from New England Nuclear (Boston, MA, USA), and heparin-Sepharose beads from Pharmacia LKB Biotechnology Inc. (Piscataway, NJ, USA). Human recombinant bFGF corresponding to the 146 amino acid form of the growth factor2-4 was obtained from Takeda Industries (Osaka, Japan). The growth factor was stored as a concentrated solution of 1 mg/ml at -80°C. Stock solutions were made after an initial 1: 100 dilution in 0.15 M saline and stored at -20°C to be used after further 1 :200 dilution as the final concentration in all experiments. lodinated bFGF was prepared using a modification of the method of Bolton and Hunter (gift of Dr S. Doctorow)“. The specific radioactivity of the ‘251-bFGF was 85 nCi/ng. at a concentration of Aliquots were stored at -20°C 1.2 pg/ml. These aliquots were thawed, mixed with the unlabelled bFGF as liquids in known ratios and stored in polyethylene tubes coated with BSA (50 mg/ml, 2 ml). The albumin prevented adhesion of bFGF to the walls of the vials and increased overall bFGF recovery after 4 wk of storage 8.7 fold above the recovery of similar amounts and concentrations of bFGF stored in untreated vials. bFGF tends to adhere avidly to polystyrene and polypropylene even with BSA coating, and vials/tubes made of these materials were avoided. The enzyme heparinase, heparin lyase, EC 4.2.2.7, an a&linked sulphated polysaccharide with a mol wt of approximately 42 900, was kindly supplied by Drs ReeseComfort and Lechband after production and purification according to Yang et al.‘*.
METHODS bFGF assay The biological activity of bFGF in sample solutions was measured by quantifying the ability of these solutions to stimulate DNA synthesis in tissue culture cells. This
620
Biomaterials
199 1. Vol 12 September
release
of bFGF from EVAc matrices
bFGF (4 ~1 at 40 pg/ml) and ‘251-bFGF (2 ~1 at 1.2 pg/ml) were combined with 500 mg of BSA in 1 ml of 0.1 5 M NaCI. The combined solution was lyophilized to a fine powder and added to 10 ml of a 10% (w/v) solution of dichloromethanedissolved EVAc. This mixture was poured into a precooled glass mould, surrounded by crushed dry ice. This provided for instantaneous freezing, producing a sheet of EVAc with BSA-bFGF dispersed homogeneously”. 14. The sheet was transferred to 4°C with a cooled spatula for 2 d and then placed under house vacuum for two additional days. Smaller pieces were cut from this large sheet and release kinetics at room temperature were determined by following BSA and bFGF levels as the saline buffer was changed on a regular basis over 2-3 wk. The spectrophotometric absorption (220 or 280 nm) of aliquots from the NaCl into which the matrices had been released was used to determine the concentrations of BSA; bFGF levels were followed using both y counting and induction of thymidine incorporation in BALBc/3T3 cells.
bFGF inactivation
and denaturation
bFGF adheres to glass and plastic with great avidity. It is possible that part of the loss in observed growth-factor activity with encapsulation was the result of the physical loss of growth factor within the storage container. ‘251-bFGF was therefore placed in 3 ml polyethylene vials (Eppendor-f), stored at 4°C and removed after 0 h, 2 h, 3 d and 74 d. A second potential source of loss in growth factor is degradation or denaturation of the compound during prolonged storage and in the steps involved in the fabrication of a controlled-release device. Device formulation often involves operations on both the incorporated substance and the incorporating material. This might involve vortexing or lyophilization of the substance or melting, pressing or dissolving the encapsulating material. These processes often require organic solvents or extremes of temperature, pressure, pH or ionic strength. Hence, the biological activity of bFGF was measured before and after gentle vortexing for 15-30s; exposure to fumes of the organic solvent dichloromethane; incubation at 4 and 50°C for 1 h; or immersion of the vial containing the growth factor within liquid nitrogen for 5-l 0 min.
Controlled
Effect
of BSA on bFGF stability
BSA and other similar glass and plastic proteins.
compounds
to prevent
To determine
whether
used to coat adherence
similar treatment
used to store the bFGF would coated with a solution
have been
the non-specific improve
of BSA (125
recovery,
mg/ml),
the bFGF.
of
Modulation of bFGF release microspheres One
problem
with
the
device fabrication.
As the alginate
was
compared
to vials that had been coated with similar volumes
of 0.1 5 M NaCI.
barrier to diffusion, buffers
and the enzyme was
studied.
bFGF
has a well-defined
characterized
affinity
and purified
for heparin
by passage
immobilized
heparinzm4. Gospodarowicz
shown
the addition
that
prevents
of heparin
in temperature.
Heparin
added to the bFGF solutions before
at 50°C
or 5-l
potential
stable
Aliquots
of bFGF
solid
have
This
1 h with
changes
was, therefore,
to dichloromethane
carrier
for the
(1.2 pg/ml)
examined
liquid
range
from the specific
beads.
on a scheduled
basis.
The
to the beads
amount
potential
of four
media
samples
M NaCl
bFGF to
taken
point.
Known
quantities
Incubated were
of labelled
with
constructed
alginate
(1.2%
beads through
by dropping
of calcium instantaneously
calcium
a mixed
chloride
solution. A 25-gauge obtained
(1.5%
as the mixture
cross-linked
by limiting
hardenrng
chloride
efficiency
of encapsulated
Spheres
envelopes
was determined growth
were
were in the
stripped
from the heparin-Sepharose were
NaCI. The solution
beads residing was changed
with the unencapsulated compared
have been released
assays
for microspheres
6 ,L/Iof bFGF and 2 ~1 of lz51-bFGF heparin-Sepharose
the bFGF was
beads with 3.0 M NaCl
and mitogenic
defined
by comparing
factor with that initially
By way of confirmation,
and both the radioactive kinetics
w/v).
a hardening
to 5 min immersion
in solution.
assay.
of sodium
entered the hardening
sphere
present
assay were
solution
bath.
Encapsulation
Release
were
Microspheres
needle was used to set the microsphere
similarly
the amount
bFGF
beads.
a needle into a beaker containing
solution
and
unlabelled
bound to 125 within
500
on a regular
according
ml) in an identical
BSA
absorbance
at
were constructed
mg/2
ml)
fashion
at 220
examined
in 2 ml of saline
basis.
alginate
(125
,ul of
pg/ml.
or
of
myoglobin
to devices
fabricated
spectrophotometerically
and 280
nm and myoglobin served as additional
controls
above.
for the experiments
described
Statistics All data are expressed
as the mean
t standard
error of the
t test and an analysis of variance
used to compare
bFGF release Figure
I
different
(ANOVA)
groups.
As predicted
approximately after
cumulative from
factor
BALBc/
3T3 cells in culture.
in matrix
and
of bFGF
activity was
was
101
When
made
2 h
t- 0.5%
of the activity
iodinated
bFGF 1.9%
and remained proved
of the of
involved
to
after
were
as the in
that,
in
present. this
to the walls
after the solution was removed. an
immediate
Biomateuals
percentage
polyethylene
was no longer
of the bFGF adhered be
to
assayed, of
the same determination
storage
demonstrated
diluted
instantaneously
maintained,
+ 0.2%.
43.1
i
activity
< 1% stimulation
of bFGF
was
was
20.9
was
at lOOO,~~g/rnl
mitogenicity
M
to
had been released
the steps and procedures
basis and, as
~1 of 0.15
and
The source of these losses was
and inactivation
solutions
recovery
of the matrix
biological
was lost, as there
by examining
and
fabrication.
Denaturation When
factor
all of the
released
release
EVAc matrices
size of the BSA carrier’0,‘4,
of the growth
Virtually
investigated
the shape
and particle
37%
2 wk.
physical
of bFGF from solvent-cast
the loading
biological
mg of the
the
activity
over time. from
from EVAc matrices
shows
biological
containing
incorporation
with
release at 4 10 nm. These mrcrospheres
50 ng/ml
of bFGF observed
to the thymidine
on a regular
(100
performed.
beads, results from the radioactive to the amount
Thus, at 4°C.
RESULTS
in NaCI) with bFGF bound heparin-Sepharose
formed size
and
heparin-Sepharose
low
incubated
were
residing
of calcium
mean. A Student’s
bound bFGF in
enhanced.
from 0 to 200
retention
beads
alone,
were
Encapsulation of heparin-Sepharose calcium alginate microspheres
from
heparinase
by monitoring
at
whereupon
constructed
ranging
bioactivity
Microspheres
with at 37°C.
storage
beads
for bFGF. BSA release was followed
release of
and compared
10 PI aliquot
at each time
and 37°C
mg/2
activation
and 6 ,LIIof ‘251-bFGF and 200
at concentratrons
release
the micro-
is desrred
were
to
was coincorporated
prolonged
release
microspheres
(0.1 5 M NaCI) changed
of
for
mg of heparin-Sepharose
4 and
for
extrapolated
of radioactive
allow
mixed
and the
in 0.15
bFGF
followed in 0.1 5 M
bFGF within
can be raised and release
alginate
250
stimulate
was
released
heparinase
until enhanced
both
the beads and was changed
at each time period was obtained
from the release
microcapsules
16 ~1 of bFGF (40 ,ug/ml)
a
The percentage
factor from the beads was followed covered
kinetics
at 4°C and maximal
might
the temperature
within
activity is temperature-sensitive
factor.
incubated
activity of the bFGF. Subsequent
~1). The solution
the mitogenic
were
and then
feature
heparinase
as
growth
inactivation
temperatures
with
fumes,
bound to the beads was determined bound
to predictably
release
experiment,
The enzyme’s
calcium
to liquid nitrogen.
were
over a concentration
of units of bFGF
present
in pH or
activity assayed
the heparin-Sepharose
number
of bFGF
(20pg/ml)
beads
2~1 of lz51-bFGF
initial lz51-bFGF
(500
and Cheng” to solutions
0 min exposure
Heparin-Sepharose
growth
relative
and mitogenic
1 h exposure
and after
incubation
with
with
losses In activity that accompany
elevation
and has been
over columns
with
offers a minrmal
can be depleted
with the heparin-Sepharose-bound spheres.
mrcrospherical
commences
envelope
heparinase bFGF
for identical
NaCI. In another
alginate
in 2 ml of 3.0 M NaCl and compared
from microcapsules
on bFGF stability
et al.
The ability of high salt-concentration
release
data obtained
is that release
the microspheres
hours after fabrication.
E R. Edelman
use of the alginate
vials were
of bFGF
of bFGF
from calcium
system
used to store
of heparin
release
controlled-release
dry, then
Effect
modulated
of the vials
let to stand and
Recovery
and
phenomenon
199
1. Vol
2 h
vials, Use of hiatus,
of the vials
Adhesion as
12 September
only
also an
621
Controlled and modulated release of bFGF: E.R. Edelman et al.
10’
lo” IO’
&-.-
?I
’r
,o‘?
6 .
.
0
I..
I..
1..
.2io”
14
7
TIME
(days)
Figure 1 bFGF release from solvent case EVAc matrices was followed using a ‘25/ radiolabel (0) and compared to the mitogenic potential of bFGF released from similar matrices (e), as evidenced by the induction of DNA synthesis within BALBc/3T3 cells. Data are displayed in both absolute units of bFGF (left ordinate) and the percentage of total incorporated bFGF (right ordinate) on a semilog scale and represents the average release from four matrices with each assay performed in duplicate. Error bars are smaller than data symbols.
additional 4.5% adhered over the next 74 d. BSA-coated vials, in contrast, had only a 2% adherence over the first 2 h, but a significant loss in activity after prolonged storage; only 37.5 f 1 .l% of the initial growth factor activity was recovered when bFGF was stored in BSA-coated vials stored for 4 wk at 4°C. In addition, when solutions were diluted immediately from the stock and then exposed to the conditions that might be observed with any of the various standard encapsulation techniques, significant losses of bFGF activity were observed. In some cases, these losses surpassed those associated with prolonged storage. After bFGF solutions were exposed to the fumes of the organic solvent dichloromethane, placed in direct contact with liquid nitrogen or incubated at 5O”C, biological activity was diminished by 94.1 k 0.2, 93.6 f 0.3, 90.2 f 0.3%. respectively (Table 7). Interestingly, despite the potential for denaturation with physical shaking, 30 s of gentle vortexing did not significantly reduce the biological activity of bFGF. Only 9% of the biological activity of the growth factor was lost with this procedure.
Preservation
of bFGF stability with heparin
Encapsulation of heparin-Sepharose-bound calcium alginate microspheres
Table 1 Recovery of bFGF after exposure to prolonged storage, extremes of temperature or organic solvents. Heparin is partially protective of the temperature affects but does not alter the inactivation of the growth factor by dichloromethane fumes Recoven/ (%)
80
.
F . ..o
l’
.
O
Storage 4°C 4 wk 50°C N,(I) CH,CI,(g)
622
Biomaterials
Hepann
5.4 9.8 6.4 5.9
21.3 19.6 24.1 4.1
t t f t
1.6 0.3 0.3 0.2
199 1, Vol 12 September
* + f f
1.9 2.0 4.0 0.5
0
0
o
o”
‘0 JO
I,0
20 0 a ,.,.,.).“.““““‘I”““/ 0 7
14
TIME Control
bFGF in
When the heparin-Sepharose beads were entrapped within calcium alginate microcapsules, 26% of the initial growth
60.
We confirmed the finding that heparin potentiated the activity of bFGF, stabilized the effect of heat on bFGF15 and found that it extended to colder temperatures as well (Table I). Heparin added to the bFGF solutions at a concentration of 20 pg/ml potentiated bFGF activity, increasing the mitogenic
Exposure
potential to 123.8 f 9.1% of the original material. However, when the same amount of heparin was added to bFGF that had previously been stored in polyethylene vials at 4°C for 4 wk, biological activity was not augmented and remained at 4.3 f 0.6% of the initial activity. Heparin was protective of the losses with storage and environmental extremes when added to solutions of bFGF before storage, direct exposure to liquid nitrogen or incubation at 50°C. Biological activity increased 3.9, 3.8 and 2.0 fold, respectively, above values obtained without heparin (Table 1). However, heparin did not prevent losses from exposure to dichloromethane fumes, as 95.9 + 0.5% of the growth factor was inactivated. Heparin immobilized to a Sepharose carrier provided a convenient substrate for handling the growth factor. bFGF was bound avidly and without saturation after incubation for 1 h with heparin-Sepharose beads. Over the range of concentrations used, we observed fairly constant binding of 77.0 + 1.3% of the initial growth factor to the beads. As the amount of heparin immobilized to the Sepharose beads was very much greater than the amount of bFGF, bFGF binding never saturated. At the highest bFGF concentration, approximately 176 units (35 ng) of growthfactoradhered to each mg of bead. Binding was confirmed in two independent ways. When the heparin-Sepharose beads bound with bFGF were placed in 300 ml of 3.0 M NaCI, 63% of the bound growth factor was removed and detected in the buffer solution within the first 30 min; the remainder leached off over the next 8 h. Similarly, when the same beads were placed in saline containing heparinase over a range from 0 to 200 units/ml, bFGF was removed from the beads with dosedependent kinetics and maintenance of 88.3 f 4.7% of growth factor activity. When the beads were placed in 0.15 M NaCI, sustained release of bFGF was prolonged for over 2 wk (Figure 2). More importantly, in contrast to the near total loss of biological activity with release from EVAc matrices (Figure 7), 97.5 + 13.7% of the bFGF released from the beads had retained its biological activity (Figure 2).
21
28
(days)
Figure 2 bFGF release from 125 mg of heparin-Sepharose beads resting in 0.15 M NaCI; this is plotted as cumulative percentage ofinitial encapsulated growth factor against time. Physical determination with radiolabelled bFGF (OJ is contrasted with assay of mitogenic ability I@), using incorporation of tritiated thymidine. The first two points on this curve represent data obtained at 30 and 60 min. respectively.
Controlled
and
modulated
release
of bFGF:
E.R. Edelman
et al.
. . . . 1
.
.
. 0 i
a 7
I 4
TIME
1
2
28
I 4
7
TIME
(days) Figure
. . .
both
The
4
release
from
bFGF
are presented
error bars represent
0
0
(days)
sensitivity
and heparinase,
(m). Data
0
temperature
heparin-Sepharose
of
at 3 7°C (+)
and 4°C
as the cumulative
fold
(Figure
when
200~g/ml presence
mitogenic
1
14
I
TIME Cumulative
displayed
against
microspheres
time
containmg
Sepharose-bound initial
bFGF
(a)
,ug/mlo
the
Data
potential active
with
with
incorporation
“‘1
standard
factor
radiolabel
as the
against
(0)
of tritlated
at 37°C
alginate
from
and
(6)
activity cells
The kinetics
as the burst of activity
retarded
(Figure
3a)
activity.
without
The
release (0)
ionic
heparinase
of
DISCUSSION
reducing
for
culture.
were remarkably
at each point in time, with 87.6 retained
its bioactivity.
intact for the duration and in extended
i
2.0%
The alginate
The
beyond
strength
release
To increase
of
radio-
similar (P = NS)
of the bFGF having remained
under all conditions
successful
release
Figure
3
demonstrates
increased
that
the total amount
at 2 wk from 2 5 to 85%. at physiological
the
capsules
were
with
4°C cumulative
produced
heparinase
heparinase
of
bFGF. heparinase
from the microspheres
enzyme
the temperature
and stored. were
fabricated
when
and released
release from the beads was
