Compatibility of a Protein Topical Gel with Wound
Dressings
JUNYAN A. JI,1 OLEG BORISOV,2 ERIKA INGHAM,3 VICTOR LING,2 Y. JOHN WANG1
1
Late Stage Pharmaceutical and Device Development, Process Research and Development, Genentech, Inc., 1 DNA Way,
South San Francisco, California 94080
2
Protein Analytical Chemistry, Process Research and Development, Genentech, Inc., 1 DNA Way, South San Francisco,
California 94080
3
Early Stage Pharmaceutical Development, Process Research and Development, Genentech, Inc., 1 DNA Way, South San
Francisco, California 94080
Received 20 March 2008; revised 1 May 2008; accepted 6 May 2008
Published online 11 July 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21465
ABSTRACT: The compatibility between several dressing materials and a recombinant
human vascular endothelial growth factor (rhVEGF) topical methylcellulose gel for-
mulation was investigated. The dressings being studied were Adaptic1, Non-stick
Dressing, Conformant 21, Opsite1TM and TegaporeTM. The criteria to select a compa-
tible dressing include protein stability, absence of leachables from the dressing, and
ability to retain gel on wound. An LC MS method with sample treatment using cellulase
was developed to determine protein oxidation in gel formulations. Results showed that
rhVEGF was significantly oxidized by Adaptic dressing in 24 h. Protein oxidation was
likely due to the peroxides, as determined by FOX assay, released into the protein
solution from the dressing. Furthermore, Adaptic dressing caused protein adsorption
loss, formation of high MW protein adducts, and released leachables as determined by
RP-HPLC, LC MS, and SEC. No protein oxidation or loss was observed after exposure to
the other four alternative dressings. However, unknown leachables were detected in the
presence of Opsite and Non-stick Dressing. The pore sizes of the Conformant 2 and Non-
stick dressings were too large to hold the topical gel within the wound area, making them
unsuitable for patient use. No rhVEGF bioactivity loss was observed in the presence of
Tegapore. In conclusion, Tegapore was considered suitable for the rhVEGF topical gel.
ß 2008 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 98:595 605, 2009
Keywords: compatibility; rhVEGF; topical gel formulation; dressing; protein oxida-
tion; leachables
INTRODUCTION wound and therapeutic treatment to additional
materials. Similar to the importance of evaluating
The selection of wound dressing is an important compatibility between drug and intravenous
factor for quality care and wound healing. Wound infusion set when given by that route, one must
dressings provide passive support for wound study the compatibility between the drug and
healing by creating an environment that is clean, dressing material when applied topically with
moist, and protected,1,2 but they also expose the dressing material. Previous studies have exam-
ined the cellular response to the presence of
dressing materials by testing the effects of the
Correspondence to: Junyan A. Ji (Telephone: [phone];
extracts from different dressing materials on the
Fax: [phone]; E-mail: [e-mail])
survival and proliferation of keratinocyte,3 and
Journal of Pharmaceutical Sciences, Vol. 98, 595 605 (2009)
the bio-compatibilities of materials used in wound
ß 2008 Wiley-Liss, Inc. and the American Pharmacists Association
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009 595
596 JI ET AL.
management by determining cell growth rate in liquid formulation contains 5 mM sodium succi-
fibroblast cultures containing dressing material.4 nate buffer at pH 5.0, 0.01% polysorbate 20, 10.4%
These studies demonstrated that testing the trehalose dehydrate and 0.2 mg/mL, 1.8 or 5.0 mg/
dressings in vitro and using the appropriate mL rhVEGF. The gel formulation was made by
assays were critical before clinical application. mixing the liquid formulation with 4.7% methyl-
Growth factors have been reported to participate cellulose, which resulted in a final concentration
in the process of wound healing, including of 0.2 mg/mL rhVEGF in 5 mM sodium suc-
epidermal growth factor (EGF), platelet derived cinate buffer at pH 5.0, 3% methylcellulose with
growth factor (PDGF), fibroblast growth factor trace amount of polysorbate 20 and trehalose
(FGF), transforming growth factor (TGF-b1), dihydrate. Methylcellulose (USP grade) was pur-
insulin-like growth factor (IGF-1), human growth chased from Dow Chemical Company (Midland,
hormone (hGH) and granulocyte-macrophage MI). Acetonitrile (HPLC grade) was obtained from
colony-stimulating factor (GM-CSF).5,6 The medi- Burdick & Jackson (Muskegon, MI). Formic
cated dressings which were applied directly to the acid and trifluroacetic acid were purchased from
wound site were reviewed by Boateng et al.7 Those Pierce (Rockford, IL). t-Butyl hydroperoxide
examples implied the compatibility of growth (t-BHP) (70% in water, analytical grade), potas-
factors with the dressing materials such as sium chloride and potassium phosphate were
collagen, alginate, hyaluronic acid, etc. Finetti purchased from Sigma Aldrich Co. (St. Louis,
and Farina8 studied the compatibility of the rhb- MO). Cellulase was purchased from Worthington
FGF solution with different dressings using an Biochemical Corporation (Lakewood, NJ). The
HPLC method. However, the compatibility of dressings tested in the studies were products
dressing materials with a concomitantly adminis- available on the market, and in detail: Adaptic1
tered wound healing drug formulation, for exam- Dressing (300 300, J&J Wound Management UK,
ple, growth factor gel was not sufficiently studied. Lot #405303), Non-stick Dressing (400 300, J&J
Recently, a recombinant human vascular Hospital Products for Home Care, Lot # 2625A),
endothelial growth factor (rhVEGF) topical gel Conformant 21 WoundVeil (400 400, Smith &
formulation was developed using methylcellulose Nephew, Lot #509774), Opsite1TM (400 5.500,
as the gelling agent, for the potential treatment of Smith & Nephew, Lot #0534), and TegaporeTM
wound healing of chronic foot ulcers in diabetic (300 400, 3M, Lot # 2009-01DA).
patients.9 The purpose of this study was to
investigate the compatibility between dressing
materials and the rhVEGF topical gel formula-
Methods
tion. The criteria to select a compatible dressing
Sample preparation
include protein stability, absence of leachables
from the dressing material, and ability to hold gel
Five pieces of the 1 cm 1 cm dressing were cut
form on wound. Protein stability was monitored
and placed in a single glass vial containing 1 mL of
by LC MS, reversed-phase chromatography, and
0.2 mg/mL rhVEGF liquid formulation or 1 g of
size-exclusion chromatography. The dressings
0.2 mg/mL rhVEGF gel formulation. The samples
evaluated in this report were Adaptic1
were covered in aluminum foil and incubated for 3
(J&J Wound Management), Non-stick Dressing
and 24 h at room temperature. For the Non-stick
(J&J Hospital Products for Home Care),
Dressing, the cotton was peeled off to avoid its
Conformant 21 (Smith & Nephew), Opsite1TM
absorption of the protein solution and only the
(Smith & Nephew), and TegaporeTM (3M). Based
outer layer was used in testing. The supernatants
on the results, a suitable dressing material for use
of the samples at time 0, 3, and 24 h were sampled
with the rhVEGF topical gel was identified.
for analysis. A liquid or gel protein formulation
sample not exposed to the wound dressing was run
in parallel as the control.
MATERIAL AND METHODS
Reverse-phase HPLC (RP-HPLC)
Materials
RP-HPLC was carried out on an Agilent 1100
Recombinant vascular endothelial growth factor series instrument using a Zorbax 300SB-C8
liquid and gel formulation samples were supplied column (4.6 mm 150 mm, 3.5 mm). The solvent
by Genentech, Inc., South San Francisco, CA. The A was 0.1% trifluoroacetic acid (TFA) in H2O and
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009 DOI 10.1002/jps
COMPATIBILITY OF PROTEIN GEL WITH DRESSING 597
the solvent B was 0.08% TFA in acetonitrile. The 1258C; desolvation temperature: 2008C; and scan
samples were analyzed with a linear gradient range: 800 3800 amu.
from 0% B to 60% B at a flow rate of 1 mL/min in
60 min. The column temperature was set at 608C. Ferrous oxidation with xylenol orange (FOX) assay
The UV detection was set at 214 nm. The assay
A colorimetric assay was used to determine the
variation of the % main peak area was within
peroxide levels of dressing exposed solutions.
2%.
Samples were diluted 1:10 in a working solution
containing 250 mM ammonium ferrous (Fe II)
Size-exclusion HPLC (SEC) sulfate, 25 mM H2SO4, 100 mM sorbitol and
125 mM xylenol orange. The samples were left at
SEC was carried out on an Agilent 1100 series
ambient temperature for a minimum of 30 min
instrument using a TSK2000SWXL column
and the absorbance at 560 nm was measured
(7.8 mm 300 mm). The mobile Phase solvent
using an HP 8453 UV spectrophotometer with a
was 0.2 M potassium phosphate and 0.25 M
1 cm path length cell. The concentration of H2O2
potassium chloride at pH 6.8. The samples were
ranging from 0.06 to 2 ppm in the samples was
analyzed with an isocratic gradient at a flow rate
then interpolated from a standard curve. One
of 0.5 mL/min in 30 min. The column temperature
millimolar of H2O2 equals to 34 ppm. This method
was set at ambient temperature. The UV detec-
was adapted from the PeroXOquant Quantitative
tion was set at 280 nm.
Peroxide Assay Kits (Pierce, Rockford, IL). The
assay variation was within 1 mM.
UV visible spectrometry
Photomicrography of the dressings
UV vis spectra were obtained using a 1 cm path
length cell on an HP 8453 spectrophotometer. Photomicrographs were taken using a Canon EOS
Placebo not exposed to any dressing was used as a camera with a MPE 65 mm macro lens to measure
blank. UV scan ranged from 190 to 600 nm. The the pore sizes of the dressings. The scale is shown
extinction coefficient is 0.37 L/(g cm) at 280 nm. in the pictures.
Potency assay
LC MS
The bioactivity of protein was measured by
Gel samples were prepared by adding 10 mL
human umbilical vein endothelial cell-based
of cellulase solution (20 mg/mL in 5 mM sodium
(HUVEC) assay in which the ability of the protein
succinate buffer, pH 5.0) to 600 mL of the rhVEGF
to stimulate cell proliferation was tested.
gels. Samples were vortexed/mixed for 2 3 min to
facilitate methylcellulose digestion. After being
liquefied, the samples were diluted to a final
RESULTS AND DISCUSSION
concentration of 0.02 mg/mL rhVEGF with 0.1%
formic acid aqueous solution. Liquid formulation
Analytical Method Development
was diluted with 5 mM sodium succinate buffer
solution to a final rhVEGF concentration of rhVEGF was expressed in E. coli and purified as a
0.05 mg/mL for injection. covalent homodimer. Each monomer is composed
RP-HPLC was carried out on an Agilent 1100 of 165 amino acids and the molecular weight of the
series instrument using a Zorbax 300SB-C8 protein is ca. 38 kDa. Each monomer contains six
column (2.1 mm 150 mm, 3.5 mm). The solvent methionine residues located at positions 3, 18, 55,
A was 0.1% formic acid in H2O and the solvent B 78, 81, and 94 in the kinase domain receptor
was 0.1% formic acid in acetonitrile. The samples (KDR) receptor binding domain (residues 1 110).
were separated with a linear gradient from 15% B Met3, Met18, and Met55 are most susceptible to
to 90% B at a flow rate of 0.2 mL/min in 20 min. oxidation as these residues are with the largest,
The column temperature was set at 708C. exposed surface area. Met78, Met81, and Met94
Micromass qTOF1 (Waters Corp., Milford, MA) are buried within the hydrophobic regions of the
was operated in a positive electrospray ionization molecule.10,11 The positions of Met18, Met78, and
mode with the following operating parameters. Met 81 are proximal to the region responsible for
Source spray needle voltage: 3200 V; declustering binding to the KDR receptor and thus, the
cone voltage: 45 V; source block temperature: formation of metsulfoxide could have an effect
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009
598 JI ET AL.
on binding due to the change in polarity and
subsequently affect its bioactivity.12 Therefore,
rhVEGF oxidation was the main focus of stability
concern in this study. In addition, size-exclusion
chromatography was used to determine the
potential aggregate formation.
RP-HPLC with UV detection is a method of
choice for monitoring protein oxidation, but the
high viscosity of the cellulose gel samples makes
direct injection to an RP-HPLC column impossi-
ble. One approach for testing a viscous sample is to
dilute it extensively to reduce the viscosity. This is
not practical for the rhVEGF molecule, since it has
a very low extinction coefficient of 0.37 L/(g cm)
Figure 1. Comparison of LC MS total ion chro-
due in part to the lack of tryptophan residue and
matograms (TICs) for the protein liquid (gray line)
low in aromatic residues in its primary structure.
and gel (dark line) formulations. Peaks at 14 16 min
Extensive dilution of the rhVEGF gel samples
from gel sample were due to the interference from
resulted in a low protein concentration that was
cellulase.
below UV detection limit.
To reduce the viscosity of the gel formulation,
cellulase was considered. Cellulase is an enzyme
complex which breaks down cellulose to beta- formulations of rhVEGF. In liquid formulation,
glucose in an optimal pH range of 4.2 5.2. The rhVEGF eluted at 11 min. In the gel formulation
enzymatic mechanism was extensively studied.13 after cellulase treatment, the protein peak at
The viscosity of the cellulose gel is expected to about 11 min could also be clearly observed;
decrease significantly due to the break-down of although there was a high broad background due
the cellulose s polysaccharide backbone. However, to partially digested cellulose as well as extra
it was also observed that cellulase could cause peaks at 14 16 min which were from the
nonspecific cleavage of the protein backbone (data interference of cellulase. Figure 2 compares the
not shown) due to the impurities in the cellulase deconvoluted mass spectra of the placebo gel
purchased.14 To minimize or eliminate protein spiked with the liquid rhVEGF before and after
cleavage but still reduce sample viscosity, our digestion with cellulase. The liquid rhVEGF
approach was to use a high ratio of cellulase to formulation was used as the reference control.
protein and less reaction time. This approach
resulted in the viscosity of the cellulose gel being
reduced rapidly, within a period of 2 3 min, for LC
injection while integrity of the protein structure
was minimally affected because of its shorter
exposure time to the cellulase enzyme.
RP-HPLC analysis requires reproducible chro-
matographic separation and baseline resolution of
all species of interest for reliable quantitative
results. Interferences from cellulose and its
degraded products made it impossible to accu-
rately quantify protein variants using peak area
integration. In our work, LC MS method was
employed in order to overcome this problem. Mass
spectrometric detection not only can provide high
sensitivity for the analysis, it also serves as
another dimensional separation of the protein
Figure 2. Deconvoluted mass spectra showing mea-
species of interest, regardless of chromatographic
sured molecular weight of the unmodified and oxidized
interferences.15
protein obtained during LC MS of a liquid reference,
Figure 1 compares the total ion chromatograms
liquid reference spiked and mixed into a placebo gel
obtained during LC MS analyses of gel and liquid before and after cellulase digestion.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009 DOI 10.1002/jps
COMPATIBILITY OF PROTEIN GEL WITH DRESSING 599
The observed mass of rhVEGF at 38300 Da Adaptic Dressing
matches its theoretical mass. The species at
Adaptic dressing is a primary dressing that is
38316 Da is its mono-oxidized form. In this
indicated for dry to highly exuding wounds where
analysis, levels of protein oxidation were mea-
adherence of dressing and exudate is to be prevent-
sured using LC MS by determining the percen-
ed (http://www.jnjgateway.com/ or http://www.
tage of the peak height of the mono-oxidized form
jnjgateway.com/home.jhtml?locźtreng&pageź
when the main peak was normalized to 100%.
viewcontent&contentidź09008b9880b6b5ce&par-
Protein oxidation levels for all samples in Figure 2
entidź0900). It can be used for burns, grafts,
were comparable with a general variation of the
venous ulcers, pressure ulcers, eczema, surgical
quantification of protein oxidation using LC MS
incisions, lacerations, etc. It also protects regen-
being about 10%.16 This shows that protein
erating tissue and minimizes patient pain at
oxidation was not promoted by samples prepara-
dressing changes. Wound exudate can easily pass
tion and cellulase treatment procedure as
through the Adaptic dressing to the secondary
described above.
absorbent dressing, typically gauze. Therefore, it
To further examine the validity of using LC
was initially considered to be used for the rhVEGF
MS in determination of protein oxidation,
topical gel product for diabetic foot ulcers.
rhVEGF was subjected to a known oxidizing
Although rhVEGF for wound healing was
agent, t-BHP, and the changes in LC MS
developed in the methylcellulose gel, the compat-
spectra were investigated. Figure 3 shows that
ibility of dressings with the protein was at first
protein oxidation levels in the gel formulation
tested in liquid formulation. This was done
increased with increased amounts of added
because the viscosity of gels created difficulties
t-BHP. The rhVEGF oxidation level of nonspiked
in developing chromatographic assays. Protein
gel control was 45%. The initial rhVEGF oxida-
oxidation level in gel formulation was verified by
tion level was caused by methylcellulose during
LC MS afterward. The excipients in the liquid
preparation of the gel. It has been reported
formulation are identical to those in the gel
that cellulose derivatives contains residual per-
formulation except the gelling agent, methylcel-
oxide.17 When protein gels were spiked with
lulose, was omitted. The rhVEGF dose was set at
t-BHP to reach concentrations of 1, 5, and 10 ppm
100 mL gel per 1 cm2 wound area. For a size of
and kept overnight at room temperature,
10 cm2 wound area, this would require 1000 mL gel
protein mono-oxidized levels were measured to
or liquid. Based on that, the experiment was
be 55%, 64%, and 72%, respectively. The inves-
designed using five pieces of the 1 cm 1 cm
tigation of LC MS detection showed it was able
dressing, total 10 cm2 on both sides. The rhVEGF
to detect the changes of the protein oxidation
gel product was designed for daily application;
levels.
stability up to 24 h at ambient temperature was
used.
Figure 4 shows the reversed-phase chromato-
grams of rhVEGF solutions after exposure to the
Adaptic dressing for 3 and 24 h. Compared to the
Figure 4. Reverse-phase chromatograms of rhVEGF
Figure 3. Comparison of the oxidation levels of (solid lines) solution after exposure to the Adaptic dres-
rhVEGF gels spiked with 1, 5, and 10 ppm t-BHP. sing for 0 h (solid), 3 h (dash), and 24 h (dash dot).
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009
600 JI ET AL.
t ź 0 sample, the main peak of rhVEGF decreased the Adaptic dressing, and reached up to 71% after
while the front-running peak increased after the 24 h of exposure time. This is consistent with the
exposure to the Adaptic dressing. At 24 h, the observation of the increased front-running peak in
front-running peak height was almost the same as reversed-phase chromatograms in Figure 4. For-
that of the main peak. It has been determined that mation of higher molecular weight (HMW) species
the front-running peak primarily contained oxi- clustered around 39200 Da was also observed by
dized rhVEGF form (data not shown). However, LC MS as shown in Figure 5b. Occurrence of
since some oxidized species were found to coelute these species was observed already after 3 h of
with the main peak, an accurate quantification exposure time to Adaptic dressing (data not
based on a reversed-phase method was difficult. shown) and became even more pronounced after
As mentioned earlier, mass spectrometry can 24 h. These HMW species are possibly rhVEGF
serve as the second dimensional separation to adducts formed upon exposure to the Adaptic
resolve the protein species of interest, regardless dressing, but the identity of these species was not
of chromatographic interferences, and provide determined since it was outside the scope of this
quantitation information. LC MS analysis as investigation.
seen in Figure 5a shows that protein oxidation The causes for rhVEGF oxidation upon its
occurred. Oxidation levels increased with exposure to the Adaptic dressing were sought.
increased incubation time of the protein with Adaptic dressing is made of knitted cellulose
acetate and impregnated with a specially for-
mulated petrolatum emulsion designed to prevent
dressing adherence and protect the wound (http://
www.jnjgateway.com/ or http://www.jnjgateway.
com/home.jhtml?locźtreng&pageźviewcontent
&contentidź09008b9880b6b5ce&parentidź0900).
The word   emulsion  implies the presence of
surfactants meaning that Adaptic dressing likely
contains such common surfactants as polysorbate.
Low level of residue peroxide in surfactant has
been reported to induce protein oxidation.18 In
addition, the dressing is rendered sterile by
gamma radiation which would potentially cause
the formation of free radicals and increase residual
peroxide in polysorbate.
In order to determine the amount of peroxide
released in the rhVEGF solution from the Adaptic
dressing, the FOX assay was used. Table 1 shows
the peroxide levels measured in the super-
natant of liquid formulation when exposed to
the Adaptic dressing. After just 1 h exposure to the
Adaptic dressing, there was 31 mM peroxide
(ź1.06 ppm H2O2) in the placebo and 24 mM
(ź0.82 ppm H2O2) in the rhVEGF solution. The
amount of peroxide remained at comparable levels
through 24 h. This level of peroxide was sig-
nificant, considering that each came from five
pieces of 1 cm 1 cm dressing, and thus strongly
indicating that oxidation of rhVEGF was most
Figure 5. (a) Effect of exposure of the rhVEGF solu-
likely caused by the peroxide in the petrolatum
tions to the Adaptic dressing by comparing protein
emulsion coated on the Adaptic dressing.
mono-oxidation levels ((&) exposed for 24 h; (~)
The results discussed above were conducted in
exposed for 3 h; (&) unexposed for 24 h; (~) unexposed
the rhVEGF liquid formulation. The gel formula-
for 3 h. (b) Deconvoluted mass spectra of rhVEGF in
tion was then tested by the LC MS method to see
liquid formulation exposed to Adaptic dressing, show-
whether it would generate comparable results.
ing formation of higher molecular weight adducts after
24 h. Figure 6 compares deconvoluted mass spectra of
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009 DOI 10.1002/jps
COMPATIBILITY OF PROTEIN GEL WITH DRESSING 601
Table 1. Peroxide Level (mM) in the Solution Exposed to the Adaptic Dressing
Measured by FOX Assay
Time (h)
0 1 3 6 24
0.2 mg/mL rhVEGF solution unexposed (control) 3 3 2 3 2
Placebo (formulation buffer) exposed 2 31 27 46 30
0.2 mg/mL rhVEGF solution exposed 3 24 23 23 24
gel samples exposed to the dressing for 3 and 24 h observed in the SEC analysis (Fig. 8) in which the
with the unexposed controls. In the control main peak area decreased with increased expo-
samples, rhVEGF oxidation remained unchanged sure time to the dressing. The level of the rhVEGF
between 3 and 24 h of exposure time. In contrast, main peak area loss is comparable with that
the samples exposed to the Adaptic dressing had determined by the reversed phase chromatogra-
rhVEGF oxidation with significant increase of phy as shown in Figure 7. Both chromatographic
mono- and doubly oxidized protein species. Even assays indicated that protein loss occurred,
more significant oxidation was observed after 24 h possibly due to adsorption by the dressing
of exposure to the dressing, with almost quanti- material.
tative loss of nonoxidized protein and pronounced In addition to protein oxidation and protein loss,
formation of multiply oxidized species. further analysis of the samples by size exclusion
In Figure 4, it was also noticed that total peak chromatography in Figure 8 shows that the peak
areas of the rhVEGF in the reverse-phase area of high MW (early eluting) species increased
chromatograms decreased significantly at 24 h with the exposure time. These species are unlikely
when protein solutions were exposed to the to be protein, as they appeared even in the absence
Adaptic dressing. The rhVEGF main peak loss of protein. They are likely to be polymeric
over time based on reverse-phase chromatogra- leachables released from the Adaptic dressing.
phy is shown in Figure 7 and followed a linear
relationship. The same phenomenon was also
Alternative Dressings
Since rhVEGF degradation occurred after expo-
sure to the Adaptic dressing, an alternative
dressing needed to be identified. Dressings that
Figure 6. Deconvoluted mass spectra of rhVEGF gel Figure 7. rhVEGF protein loss in the protein solu-
formulation exposed to Adaptic dressing at 3 h (c) and tion when exposed to the Adaptic dressing determined
24 h (d). (a) and (b) are controls at 3 and 24 h. by rp-HPLC.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009
602 JI ET AL.
Figure 8. Size-exclusion chromatograms of placebo
(dash dot line) and rhVEGF solutions (solid lines) after
Figure 10. Size-exclusion chromatograms of the
exposure to the Adaptic dressing for 0, 3, and 24 h.
rhVEGF solutions after exposure to the four alternative
dressings for 24 h.
do not contain   emulsion  which was likely to
release peroxide were studied. The alternative
dressings chosen for this work were Moreover, no significant changes of protein oxida-
tion were observed by LC MS method as shown in
Non-stick Dressing: made of porous high den- Figure 11. Oxidation levels remained unchanged
sity polyethylene net film with Rayon inner upon exposure to the dressings for up to 24 h.
layer. Deconvoluted mass spectra for all the alternative
Conformant 21 WoundVeil: made of poly- dressing components did not show any significant
ethylene coarse mesh. formation of higher MW protein adducts (data not
Opsite1TM: made of polyurethane film. shown). Finally, the FOX assay results in Table 2
TegaporeTM: made of nylon fine mesh. indicated zero or below detection levels of peroxide
in the protein solutions exposed to these dressings
for 24 h, in contrast to the formation of peroxides
These dressings are semipermeable mesh or
from Adaptic dressing (see Tab. 1).
film dressings with different vapor permeability,
adhesiveness, conformability and extensibility.19 Although all four alternative dressings did not
cause rhVEGF oxidation or higher MW protein
As shown in Figures 9 and 10, rhVEGF did not
adducts, the total ion chromatograms from
show increased oxidation or protein loss by
LC MS analysis showed some differences. Un-
RP-HPLC and SEC, respectively, after 24 h
known species, likely polymer or plasticizer, were
exposure to these dressings. Specifically, there
observed in some of the exposed samples as shown
was no change in height of the front-running
in Figure 12. Of the four dressings being studied,
peaks in the reverse-phase chromatograms and
peak areas did not decrease. The high MW peak
was also absent in the size exclusion chromato-
grams in contrast to Adaptic dressing study.
Figure 9. Reversed-phase chromatograms of rhVEGF Figure 11. Comparison of the protein mono-oxida-
solution after exposure to the four alternative dressings tion levels in the solutions exposed to the four alter-
for 24 h. native dressings after 24 h.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009 DOI 10.1002/jps
COMPATIBILITY OF PROTEIN GEL WITH DRESSING 603
Table 2. Peroxide Levels (mM) in 0.2 mg/mL rhVEGF
Solutions Exposed to the Non-stick Dressing (J&J
Hospital Products for Home Care), Conformant 2
(Smith & Nephew), Opsite (Smith & Nephew) and
Tegapore (3M) Determined by FOX Assay
Sample Time (h) Buffer Protein Solution
No exposure 0 6 6
No exposure 24  4
Non-stick 24 1 7
Conformant 2 24 3 4
Tegapore 24 0 2
Opsite 24 4 2
 : not tested.
Non-stick and Opsite dressing showed an increase
Figure 13. UV vis spectra of the 1.8 mg/mL rhVEGF
in the levels of these compounds as a function of
solutions exposed to the dressings for 24 h. The rhVEGF
the exposure time (Fig. 12a and b). For the Opsite
solution unexposed was used as the control (24 h con-
dressing, UV vis spectra (Fig. 13) also show the
trol: solid line; Conformant 2: dash line; Tegapore: Dot
unknown species in dressing exposed buffer.
line; Non-stick dressing: dash dot dot line; Opsite: dash
These compounds were not further identified in dot); UV vis spectrum of the formulation buffer exposed
to the Opsite dressing for 24 h is also shown in the figure
(short dash line).
this study, but presumably are polymers leached
from the dressing. The UV vis spectra of the
rhVEGF solutions exposed to the other three
dressings, given in Figure 13, also show that
the protein absorbance at 275 nm was the same as
the control sample, supporting that no obvious
protein loss occurred in the samples exposed to
these dressings as described in the results of RP-
HPLC and SEC analyses.
There are further criteria for wound healing
dressing selection beyond the need for drug
compatibility. Because the rhVEGF topical gel
product is for the potential treatment of foot ulcers
for diabetic patients, it is important that the
dressing can hold the gels within the wound area
with minimal seepage when slight pressure is
applied. The viscosity of the rhVEGF topical gel
product is about 3000 cP. All the dressings
investigated herein are semipermeable film dres-
sings of which pore size determines this char-
acteristic. Photomicrographs of the dressings
were taken to measure the pore size of their
network (Fig. 14). The pore sizes of Conformant 2,
Non-stick Dressing and Tegapore were deter-
mined to be ca. 600, 200, and 80 mm, respectively.
Opsite dressing was not tested due to the
leachables described above. The pore sizes of
Figure 12. Total ion chromatograms of the rhVEGF
the Conformant 2 and Non-stick dressings were
solutions incubated with the dressings for 3 h (a) and
24 h (b). too large as evidenced by the gel product seeping
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009
604 JI ET AL.
Figure 14. Photomicrographs of the dressings to measure the pore sizes of their
network. The sizes of & in Conformant 2, Non-stick and Tegapore are 0.2, 0.2, and
0.08 mm, respectively.
through the net of both dressings. Tegapore CONCLUSIONS
displayed the least amount of loss of the topical
gel from seepage and had the smallest pore size. Under the conditions studied, Adaptic dressing
Bioactivity of the rhVEGF solution exposed to caused rhVEGF oxidation, protein adsorption
the Tegapore dressing for 24 h was tested. The loss, formation of high MW protein adducts, and
specific activity was determined as 82% whereas released polymeric leachables. Oxidation of pro-
that of the unexposed protein solution was 91%. tein by the Adaptic dressing was likely due to the
Since the relative error for the potency assay peroxides released into the protein solution from
experiment is about 20%, the results suggest the dressing. No oxidation or protein loss was
that the bioactivity of the protein was not affected observed after exposure to the four alternative
by this dressing. dressings. However, unknown leachables were
Table 3 summarizes the test results of the detected in the presence of Opsite and Non-stick
compatibility study between the five dressing Dressing. Furthermore, the pore sizes of the
materials and the rhVEGF topical methylcellu- Conformant 2 and Non-stick dressings were too
lose gel formulation. large to hold the topical gel within the wound
Table 3. Summary of the Comparability Study Results of the Dressings with rhVEGF
rhVEGF Protein Level of Polymeric Pore
Dressings Oxidation Adsorption Peroxidea (mM) Leachables Size
Adaptic1 Yes Yes 30 Yes 
Opsite1TM No No 4 Yes 
Non-stick No No 1 Yes 200 mm
Conformant 21 No No 3 No 600 mm
TegaporeTM No No 0 No 80 mm
 : not tested.
a
At 24 h.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009 DOI 10.1002/jps
COMPATIBILITY OF PROTEIN GEL WITH DRESSING 605
area, making them unsuitable for patient use. 7. Boateng JS, Matthews KH, Stevens HNE, Ecc-
leston GM. 2007. Wound healing dressings and
Finally, Tegapore would be suitable for the topical
drug delivery systems: A review. J Pharm Sci
rhVEGF topical gel due to its compatibility with
97:2892 2923.
the protein, absence of polymeric leachables, and
8. Finetti G, Farina M. 1992. Recombinant human
the appropriate pore sizes. Most importantly, no
basic-fibroblastic growth factor: Different medical
rhVEGF degradation and bioactivity loss were
dressings for clinical application in wound healing.
observed in the presence of Tegapore.
Il Farmaco 47:967 978.
9. Breen TJ, Bunting S, Samba CP. 2006. Wound
healing. WO 2006/138468.
ACKNOWLEDGMENTS
10. Vries CD, Escobedo JA, Ueno H, Houck KA,
Ferrara N, Williams LT. 1992. The fms-like tyr-
The authors wish to thank Dr. Ann Daugherty for
osine kinase, a receptor for vascular endothelial
information on methylcellulose gel, Kathleen growth factor. Science 255:989 991.
Kosewic and Dr. Max Tejada for HUVEC assay 11. Waltenberger J, Claesson-Welsh L, Siegbahn A,
Shibuya M, Heldin CH. 1994. Different signal tra-
support, Dr. Steve Shire and Matthew Field for
nsduction properties of KDR and Flt1, two recep-
assay discussion, Al Stern for photomicrographs,
tors for vascular endothelial growth factor. J Biol
Paul Kwon, MD for recommending dressing mate-
Chem 269:26988 26995.
rials, and Dr. Sherry Martin-Moe for guidance
12. Duenas E, Keck R, DeVos A, Jones AJ, Cleland
and encouragement.
JL. 2001. Comparison between light induced and
chemically induced oxidation of rhVEGF. Pharm
Res 18:1455 1460.
REFERENCES
13. Worthington Enzyme Manual, 1993 Edition.
14. Dienes D, Börjesson J, Hägglund P, Tjerneld F,
1. Queen D, Orsted H, Sanada H, Sussman G. 2004. Lidén G, T
eczey K, Stålbrand H. 2007. Identifica-
A dressing history. Int Wound J 1:59 77. tion of a trypsin-like serine protease from Tricho-
2. Doughty D. 2005. Dressings and more: Guidelines derma reesei QM9414. Enzyme Microb Technol
for topical wound management. Nurs Clin North Am 40:1087 1094.
40:217 231. 15. Houde D, Kauppinen P, Mhatre R, Lyubarskaya
3. Paddle-Ledinek JE, Nasa Z, Cleland HJ. 2006. Y. 2006. Determination of protein oxidation
Effect of different wound dressings on cell viability by mass spectrometry and method transfer to
and proliferation. Plas Reconstr Surg 117:110S quality control. J Chromatogr A 1123:189 198.
118S. 16. Silva JC, Gorenstein MV, Li G-Z, Vissers JPC,
4. Turner TD, Spyratou O, Schmidt R. 1989. Biocom- Geromanos S. 2006. Absolute quantification of pro-
patibility of wound management products: Standar- teins by LCMS. J Mol Cellular Proteom 5:144 156.
dization of and determination of cell growth rate in 17. Wasylaschuk WR, Harmon PA, Wagner G,
L929 fibroblast cultures. J Pharm Pharmacol 41: Harman AB, Templeton AC, Xu H, Reed RA.
775 780. 2007. Evaluation of hydroperoxides in common phar-
5. Greenhalgh DG. 1996. The role of growth factors in maceutical excipients. J Pharm Sci 96:106 116.
wound healing. J Trauma Inj Infect Crit Care 41: 18. Ha E, Wang W, Wang YJ. 2002. Peroxide forma-
159 167. tion in polysorbate 80 and protein stability.
6. Mann A, Niekisch K, Schirmacher P, Blessing M. J Pharm Sci 91:2252 2264.
2006. Granulocyte-macrophage colony-stimulating 19. Thomas S, Loveless P. 1988. Comparative review
factor is essential for normal wound healing. of the properties six semipermeable film dressing.
J Investig Dermatol Symp Proc 11:87 92. Pharm J 240:785 787.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 2, FEBRUARY 2009