Palmitoyl derivatives of interferon α: Potential for cutaneous delivery

Article abstract of DOI:10.1021/js980146k

Palmitoyl derivatives of interferon α2b (p-IFNα) were prepared by covalent attachment of the fatty acid to lysine residues in the protein through a reaction with N-hydroxysuccinimide palmitate ester. The p-IFNα was characterized by capillary electrophoresis (CE), mass spectrometry (MS), SDS- PAGE, and antiviral assay. Flow-through diffusion cells and human breast skins were used to measure cutaneous and percutaneous absorption. Formation of p-IFNα derivatives was demonstrated by CE to be dependent on reaction time and reagent: protein ratio. Electrospray MS of the crude p-IFNα mixture indicated three populations of IFNα derivatives with 10, 11, and 12 palmitoyl substitutions. The addition of palmitoyl residues to IFNα under the conditions described reduced the antiviral specific activity by 50%. However, the cutaneous absorption of p-IFNα was about 5-6 times greater than the parent protein. The amount of p-IFNα and IFN α in whole skin after 24 h of treatment was 2.106 ± 1.216 μg/cm² and 0.407 ± 0.108μg/cm², respectively. Approximately two times higher flux was detected for p-IFNα compared to the nonfatty acylated IFNα. The total amount of drug diffused in 24 h was also approximately two times higher for the p-IFNα. The results indicate a potential for using fatty acylated derivatives of IFN α for dermal and transdermal delivery.

Full text of DOI:10.1021/js980146k

Palmitoyl Derivatives of Interferon r: Potential for Cutaneous Delivery
†
†
†
MARIANNA FOLDVARI,* SAMUEL ATTAH-POKU, JIAPING HU, QIMIN LI, HUW HUGHES, LORNE A. BABIUK, AND
SANDRA KRUGER
Contribution from College of Pharmacy and Nutrition, and Veterinary Infectious Disease Organization, University of
Saskatchewan, Saskatoon, Saskatchewan, S7N 5C9.
Received March 30, 1998. Accepted for publication June 17, 1998.
and high molecular weight. Transdermal delivery of
Abstract 0 Palmitoyl derivatives of interferon R2b (p-IFNR) were
prepared by covalent attachment of the fatty acid to lysine residues
in the protein through a reaction with N-hydroxysuccinimide palmitate
ester. The p-IFNR was characterized by capillary electrophoresis (CE),
mass spectrometry (MS), SDS−PAGE, and antiviral assay. Flow-
through diffusion cells and human breast skins were used to measure
cutaneous and percutaneous absorption. Formation of p-IFNR
derivatives was demonstrated by CE to be dependent on reaction
time and reagent: protein ratio. Electrospray MS of the crude p-IFNR
mixture indicated three populations of IFNR derivatives with 10, 11,
and 12 palmitoyl substitutions. The addition of palmitoyl residues to
IFNR under the conditions described reduced the antiviral specific
activity by 50%. However, the cutaneous absorption of p-IFNR was
about 5−6 times greater than the parent protein. The amount of
p-IFNR and IFN R in whole skin after 24 h of treatment was 2.106 ±
peptides is challenging, but nevertheless a very worthwhile
effort. The various approaches to delivering peptides
through the skin have utilized methods which temporarily
compromise the integrity or physicochemical characteristics
of the skin, including the use of penetration enhancers such
as N-alkylazacycloheptanones (Azone) for desglycinamide
4
arginine vasopressin (DGAVP) or for DEγE by Bodde et
3
al. The nonionic surfactant, n-decyl methyl sulfoxide was
shown to enhance the penetration of Leu-enkephalin
through hairless mouse skin in vitro.5 Iontophoresis was
employed by some researchers in the past few years for
6
,7
the enhancement of delivery of short peptides.
Most of the above studies demonstrated enhanced pen-
etration of the different peptides through excised animal
skin in vitro. While these studies demonstrate the feasibil-
ity of the transdermal route for peptide delivery, there is
an almost total lack of knowledge concerning the delivery
of sufficient quantities of a protein drug required for the
desired pharmacological effect.
Interferon R2b (recombinant) is a 19.3 kDa molecular
weight polypeptide consisting of 165 amino acids. Inter-
feron R shows multiple biological effects including antiviral,
antiproliferative, and immunomodulatory. The mechanism
of action is through binding to specific cell surface recep-
tors. The binding induces protein kinase and 2′5′-oligoad-
enylate synthetase.8 These enzymes can inhibit protein
synthesis in the cell and therefore can prevent a virus from
replicating.9
2
2
1
.216 µg/cm and 0.407 ± 0.108 µg/cm , respectively. Approximately
two times higher flux was detected for p-IFNR compared to the nonfatty
acylated IFNR. The total amount of drug diffused in 24 h was also
approximately two times higher for the p-IFNR. The results indicate
a potential for using fatty acylated derivatives of IFN R for dermal
and transdermal delivery.
Introduction
The routine administration of protein and peptide drugs
is hindered by the lack of a reliable and convenient mode
of delivery. The oral route is often impractical due to the
digestion of proteins in the gastrointestinal tract. Thus
parenteral administration is generally the only alternative,
although the frequent injections required, due to the short
half-life of peptides (usually minutes), can decrease patient
compliance.
Most topical applications of IFNR2 are intended for local
treatment of cutaneous or genital herpes or condylomata
(
acuminata) and not for systemic effects. Very little absorp-
tion was found through either the nasal mucosa or the skin
unless high concentration of surfactants or other additives
were used.1
0,11
The delivery of IFNR or â from various
topical dosage forms such as gels and creams was tested
in clinical experiments with negative or minor positive
Newer initiatives are being explored to deliver protein
drugs by other “nonconventional” routes such as nasal,
results.1
2-16
The reason for the lack of efficacy was
pulmonary, rectal, vaginal, ocular, and transdermal.1
,2
proposed to be the inadequate delivery of drug to the target
sites.
Encouraged by the relative success with smaller drug
molecules using penetration enhancers, a number of re-
search groups have investigated the transdermal route in
more detail. The advantage of the transdermal route over
the mucosal routes is that the skin has lower proteolytic
activity compared to the mucosa; therefore, metabolism
during transit can be minimized and the bioavailability
The main goal of this investigation was to evaluate a
novel concept for the delivery of IFN R into or through the
skin using a hydrophobic molecular carrier. In this paper
we report the preparation and evaluation of palmitoyl
derivatives of IFNR (p-IFNR) for improved dermal delivery.
3
improved if enough drug can penetrate. Bodde et al. for
example demonstrated that des-enkephalin-γ-endorphin
Materials and Methods
(DEγE) had a long, 4.4 h, half-life in fresh whole skin
P r ep a r a tion of P a lm itoyl Der iva tives of IF NrsPalmitoyl
derivatives of IFNR2b (generously provided by Schering-Plough
Research Institute, Kenilworth, NJ ) were synthesized using the
active ester of N-hydroxysuccinimide and palmitic acids (C16). In
step one the synthesis of N-hydroxysuccinimide ester of palmitic
acid (NHS-P) was carried out. Briefly, equal molar amounts of
palmitic acid (Sigma Chemical Co., St. Louis, MO) and N-
hydroxysuccinimide (Sigma) were mixed together in ethyl acetate,
compared to a plasma half-life of only a few minutes.
Proteins and peptides may exhibit very low permeability
through the skin due to properties such as hydrophilicity
*
To whom correspondence should be addressed. Phone: (306) 966-
6
338; Fax: (306) 966-6377; email: foldvari@duke.usask.ca.
†
Veterinary Infectious Disease.
©
1998, American Chemical Society and
S0022-3549(98)00146-4 CCC: $15.00
Published on Web 09/10/1998
Journal of Pharmaceutical Sciences / 1203
American Pharmaceutical Association
Vol. 87, No. 10, October 1998

and then dicyclohexylcarbodiimide (DCC) was added. The mixture
was stirred overnight at 4 °C. Dicyclohexylurea (DCU) was filtered
out, and NHS-P was recrystallized from the filtrate by the addition
acid precipitation. The final preparations of 125I-IFNR and 125I-
7
p-IFNR had specific activities of 2.05 × 10 cpm/µg and 1.94 ×
7
10 cpm/µg protein, respectively. The iodinated IFN and p-IFNR
1
of ethanol at 4 °C. H NMR studies on NHS-P confirmed the
were examined by PAGE for intactness, and the protein concen-
tration was determined by densitometry.
1
expected structure. The high field H NMR was recorded on a
Bruker AMX 300 spectrometer. There was no impurity detectable
Deter m in a tion of An tivir a l ActivitysAntiviral activity of
palmitoyl derivatives of IFNR was determined by the cytopathic
effect inhibition assay using Georgia Bovine Kidney (GBK) cells
1
in the NMR spectrum. H NMR (CDCl3, 300 MHz) δ: 0.86 (t, J )
7
Hz, 3H), 1.75 9m, 2H), 2.6 (t, J ) 7 Hz, 2H), 1.38 (m, 2H), 1.28
1
8
br s, 22H), 2.85 (s, 4H). Step two involved the preparation of
palmitoyl IFN R as follows. NHS-P was dissolved in dimethyl-
formamide (DMF) or dimethyl sulfoxide (DMSO) and added at 1:1
to 25:1 molar ratio to the PBS buffer (7.5 mM Na2HPO4, 2.5 mM
NaH2PO4, 141.2 mM NaCl) containing IFNR at pH 7.2. The
mixture was kept at room temperature for 3 h with occasional
gentle agitation. After the reaction, DMF or DMSO was removed
under vacuum in a lyophilizer (Labconco) at <-50 °C, and the
residue was redissolved in sterile distilled water. The p-IFNR was
separated from free fatty acids by chromatography on Sephadex
G-25 column (Pharmacia, Uppsala, Sweden) using PBS. The
fractions were tested by polyacrylamide gel electrophoresis (PAGE)
and silver staining. The protein was quantitated by densitometry.
The fractions containing protein were pooled, freeze-dried, and
reconstituted with sterile distilled water before use.
and vesicular stomatitis virus as the challenge virus.
We are
currently using GBK cells in our antiviral assays with success,
since the GBK cells are sensitive to human IFNR. The reference
8
standard was IFNR-2b, specific activity 2.24 × 10 IU/mg (Scher-
ing-Plough).
In Vit r o Cu t a n eou s a n d P er cu t a n eou s Ab sor p t ion of
p -IF NrsThe rate of diffusion of p-IFNR across full thickness
human breast skin (freshly obtained from plastic surgery and kept
at -20 °C until used within 1 week) was investigated using Teflon,
Flow-Thru Diffusion Cells (Crown Glass Co. Inc. Somerville, NJ ),
2
which have a surface area for diffusion of 0.32 cm . The diffusion
cells are designed such that fluid may be continuously pumped
through them in order to maintain sink conditions. PBS, pH of
7.2, maintained at 37 °C, was used as the perfusion fluid. The
diffusion cells were mounted in a PosiBloc Diffusion Cell Heater
(Crown Glass Co. Inc., Somerville, NJ ), maintained at 32 °C by
water circulated through a circulating water bath. Each cell was
connected to a fraction collector. The flow rate was 3 mL per hour.
Each experiment was conducted for a continuous period of 24 h.
The test preparations [0.1 mL of solution (PBS buffer) or 0.1 g of
methylcellulose 1500 cP (2.5%) gel hydrated with PBS] labeled
with 1 I-p-IFNR, were instilled into the cells at the beginning of
each experiment. The drug concentration in the preparations was
P olya cr yla m id e Gel E lect r op h or esis of IF NrsPoly-
acrylamide gel electrophoresis (PAGE) in the presence of sodium
dodecyl sulfate (SDS) was carried out in a Mini-Protean II (BioRad,
Mississauga, ON) apparatus. The gel consisted of a running gel
containing 14% (w/v) acrylamide and stacking gel containing 5%
acrylamide. The gel thickness was 1.0 mm. The electrophoresis
buffer was 25 mM Tris, 192 mM glycine, 0.01% (w/v) SDS, pH
25
8
.6. Electrophoresis was carried out at 200 V constant voltage.
6
The electrophoresis was conducted for 45 min. After electrophore-
sis, the gels were silver stained to detect the protein.
20 × 10 IU (89.3 µg) of p-IFNR/g or mL product. The average
2
amount of drug applied was 20.7 µg/cm skin surface area. The
Ca p illa r y Electr op h or esissCE studies were performed using
P/ACE System 5500 (Beckman, Fullerton, CA) with diode array
detector and System Gold Software. Free-zone electrophoresis was
carried out using an uncoated capillary (57 cmM × 75 µm) at 23
quantity of p-IFNR in the collected fractions was determined by γ
counting.
After 24 h, the skin was removed from the diffusion cell and
rinsed thoroughly with cold (4 °C) PBS (3 × 15 mL), and the skin
was blotted with tissue paper. The skin surface was swabbed with
a cotton tip applicator dipped into PBS containing 0.5% Tween
80 two times to remove surface bound drug. Care was taken not
to disturb the stratum corneum. The skin was carefully folded
(epidermal sides together) to avoid contamination of dermal side
and placed into glass tubes. The radioactivity associated with the
skin was determined by γ counting and was considered to be the
“whole skin” counts. The skin was then stripped 10 times with a
Scotch tape, and the radioactivity associated with each strip was
determined separately. The skin after the stripping was counted
again in a clean tube to obtain the counts associated with the
viable layers of the skin (epidermis, dermis, and subcutaneous
tissue). The skin stripping technique was validated by sectioning
the paraffin embedded stripped skin to observe the complete
removal of the stratum corneum in the light microscope (results
not shown). Trichloroacetic acid (TCA) precipitation was used to
determine free and bound iodine label in percutaneous fractions
and skin homogenate prepared from treated skin samples. TCA
was added to each sample to 5% w/v concentration and was
incubated at 4 °C overnight. The supernatants and pellets were
analyzed by γ counting after centrifugation in a Beckman Mi-
crofuge at 14000 rpm for 15 min. The experiments with TCA
precipitation from skin homogenates (after tape stripping) and
fractions showed that 40-50% of radioactivity was precipitated
from both IFNR- and p-IFNR-treated samples.
°
C, 20 kV with a 5 s pressure injection. The running buffer was
0
8
.6% w/v sodium borate (Na2B4O7‚10H2O) and 0.5% boric acid, pH
.75. The detector was used at 200-300 nm. Prior to use the
capillary was washed with NaOH (0.1 M) for 10 min and for 1
min between each run.
Ma ss Sp ectr om etr ysFour microliters of IFNR-2b (25 µg/mL
in water) or 6 µL of p-IFNR prepared at 25:1 reagent:protein ratio
(400 µg/mL in water) was injected.
The mass spectrometer used in this study was a Micromass VG
Quattro II tandem quadruple mass spectrometer (VG Organic,
Tudor Road, Altrincham, UK) operated in conjunction with VG
MassLynx Software. The ionization interface used was the
atmospheric pressure electrospray (ES) source operated in positive
ion mode. The mobile phase (0.2% HCOOH, 49.9% H2O, and
4
9.9% ACN, v/v) was delivered at a flow rate of 20 µL/min by a
Model 140A dual syringe solvent delivery system (Applied Bio-
systems Ltd., Foster City, California). The sampling cone voltage
was optimized at 45 V. Mass spectra were acquired at mass unit
resolution. The molecular weights of the proteins were automati-
cally produced from multiply charged electrospray spectra by the
MaxEnt algorithm using the maximum entropy method.
Iod in a tion of p -IF NrsIodination of IFNR and p-IFNR was
1
7
carried out using the lactoperoxidase method.
Briefly, 2 mCi
1
25
Na I obtained from Amersham (Oakville, ON), was neutralized
by adding 3 volumes of 0.03 N HCl, and the total volume was made
up to 25 µL with 0.2 M sodium phosphate buffer pH 7.2. The
following were added to this mixture: 50 µL of Enzymobeads (Bio-
Rad), 15 µL of freshly made 2% â-D-glucose in 0.1 M sodium
phosphate buffer, pH 7.2, 10 µL of IFN (approximately 10 µg of
protein). The reaction mixture was incubated for 20 min at room
temperature. The reaction was stopped by adding 25 µL of 1 M
sodium azide and incubating for 15 min. Finally 125 µL of
saturated l-tyrosine in PBS was added, and the mixture was
transferred onto a Sephadex G25 column. Fractions containing
the protein were pooled. As another method, the iodination
mixture was transferred onto Bio-Spin columns (exclusion limit
Results and Discussion
F a tty Acyla tion of IF NRsFatty acylation of IFN R
utilizes the formation of an amide bond which would be
fairly stable in dosage form development and in the
biological environment. The synthetic steps for the prepa-
ration of the palmitoyl derivatives of IFNR are shown in
Figure 1. The first step involved the preparation of NHS-
P, which in turn was reacted with IFNR in a solvent. We
found that DMSO was a better medium for the reaction
than DMF for the preservation of the antiviral activity of
IFNR. The synthetic procedure described above provides
nonspecific acylation, therefore it is expected that some of
6
000) (Bio-Rad), and the iodinated protein was recovered by a brief
low speed centrifugation. To remove any possible residue of
unbound iodine, the protein preparation was dialyzed overnight
against 1 mM sodium iodide in PBS. This procedure removed
practically all acid soluble iodine as determined by trichloroacetic
1
204 / Journal of Pharmaceutical Sciences
Vol. 87, No. 10, October 1998

O
O
O
O
DCC
R–COOH + HO
N
R
O
N
+ DCU
active ester (isolated)
+
NH2
IFN-alpha
pH 7.2
O
C
R
N
H
IFN-alpha
Figure 1sSynthetic process for the fatty acylation of IFN R2. R−COOH )
palmitic acid
Figure 3sSDS−PAGE analysis of p-IFN R prepared under various conditions.
Lane 1, Bio-Rad molecular weight standard; lane 2, IFN R incubated in DMF
medium; lane 3, p-IFN R prepared in DMF medium; lane 4, IFN R incubated
in DMSO medium; lane 5, p-IFN R − prepared in DMSO medium; lane 6,
IFN R standard 100 ng; lane 7, IFN R standard 50 ng.
Figure 2sChromatographic profile of p-IFN R on Sephadex G-25 column
and SDS−PAGE pattern of the corresponding chromatographed fractions after
silver staining. Lane 1, Bio-Rad molecular weight standard; lane 2, IFN R
standard 100 ng, lanes 3−9, fractions from Sephadex column 2−5 mL.
Figure 4sStructure of IFN R-2b (as deduced from the nucleotide sequence
of cDNA) (taken from Intron A Product Monograph, Schering Canada Inc.,
1
991).
the lysine ꢀ-amino groups and the terminal amino group
on the protein will be acylated. The p-IFNR preparation
after gel filtration may be a mixture which contains IFN
R species acylated to various degrees. For the purpose of
the current study, however, the different fractions (i.e.
monopalmitate, dipalmitate, etc.) were not separated or
purified (“crude p-IFNR”). The SDS-PAGE profile of
p-IFNR separated on Sephadex G-25 column is shown in
Figure 2. The final yield of recovery of p-IFNR was
dependent on the starting concentration. At 25 µg “batch”
size it was 50.2% and at 100 µg it was 84.0% as determined
by the densitometry of p-IFNR bands of the column
fractions.
Figure 3 shows the results of the PAGE analysis in which
the IFNR was palmitoylated in media containing DMF or
DMSO (lanes 3 and 5, respectively). It was found that the
total yield of p-IFNR was 15-20% higher in media contain-
ing DMSO compared to DMF. Identical treatment except
for the addition of palmitic acid was used to evaluate the
effect of solvents on the protein (lanes 2 and 4, Figure 3).
No apparent negative effect was found. Close investiga-
tions of the PAGE bands indicated an about 6-10%
increase in molecular weight of IFNR due to acylation.
IFNR is a hydrophilic protein with 9 lysine amino acids
(at amino acid positions 31, 49, 70, 83, 112, 121, 131, 134,
and 164) available for potential covalent attachment of
fatty acids in addition to the amino terminal. The amino
acid sequence of IFNR-2b is shown in Figure 4. Although
IFNR has two disulfide bonds between residues 1 and 19
and residues 29 and 138, only the latter disulfide bond
is critical for maximal antiviral activity.20 Circular dichro-
ism studies indicated the presence of 45-70% R helix
content and the lack of â sheet configuration.21
1
9
Previous data on consensus IFNR indicate that there are
three structurally distinct domains (10-35, 78-107, and
Journal of Pharmaceutical Sciences / 1205
Vol. 87, No. 10, October 1998

Figure 5s(A) Electropherogram of lipid-modified derivatives of IFNR: effect of incubation time. (B) Electropherogram of lipid-modified derivatives of IFNR: effect
of protein:reagent ratio.
1
23-166) that are important for activity (see regions in
Ca p illa r y Electr op h or esissCE analysis indicated that
after reaction with NHS-P for 3 h, the migration time of
IFNR changed from 5.1 to 8.5 min. More detailed inves-
tigations indicated that the degree of palmitoylation ap-
peared to be time dependent (Figure 5A). After 2 and 18
min of incubation, the migration time changed to 7 and
7.8 min, respectively, with smaller changes in migration
time up to 1 h. After 1 h reaction there was no further
change in migration time.
The effect of protein: NHS-P reagent ratio on palmitoyla-
tion was also investigated (Figure 5B). At low ratios, more
heterogeneous population of lipid-modified derivatives may
form as indicated by the broader peaks with lower mobility.
At a ratio of 1:10 or higher (e.g., 1:25 routinely used in our
studies) a reproducible population of p-IFNR was obtained
with an electrophoretic mobility of 9.5 min (Figure 5B
insert).
2
2
italics, Figure 4).
It was expected that more than one
acyl group would be incorporated per protein molecule since
IFNR has nine lysine amino acids plus the terminal
cysteine. The three-dimensional structure for IFNR was
recently constructed by computer modeling for the primary
amino acid sequence of consensus IFNR.23 With the aid of
this model it may be possible to predict which lysine
residues are conformationally available for binding. Ac-
cording to the model accessible regions within the molecule
are domains 29-35, 78-95, and 123-140. The two most
important receptor binding domains are 29-35 and 123-
1
9
40 (see underlined regions, Figure 4), while domain 78-
23
5 may be an alternative active site. This suggests that
at least four lysine residues (31, 83, 131, 134) within these
regions plus the terminal amino acid may bind with fatty
acids.
1
206 / Journal of Pharmaceutical Sciences
Vol. 87, No. 10, October 1998

Table 1sMass Spectrometric Analysis of Palmitoyl-IFN r
Table 2sCutaneous and Percutaneous Absorption of IFNr and p-IFN
r through Human Breast Skin in Vitro
peak
no.
MS peak
mass
calcd
peak mass
no. of palmitoyl
group linked
relative
error %
(A) Cutaneous Delivery
whole skin µg/ stratum corneum µg/ viable layers µg/
1
2
3
21644
21888
22120
21640
21879
22117
10
11
12
0.018
0.041
0.014
2
2
2
preparation
cm n ) 6
cm n ) 6
cm n ) 6
IFN R gel
0.407 ± 0.108
0.202 ± 0.075
0.227 ± 0.145
0.228 ± 0.091
(0.98 ± 0.39%)
1.881 ± 1.160
(10.29 ± 6.35%)
(1.76 ± 0.47%)
p-IFN R gel 2.106 ± 1.216
(11.53 ± 6.66)
Ma ss Sp ectr om etr ysMass spectrometric results re-
vealed three major peaks in the p-IFNR sample with
masses of 21644, 21884, and 22120 which corresponds to
(B) Percutaneous Delivery
parameters
125I-IFN R
125I-p-IFN R
1
0, 11, and 12 palmitoyl groups linked to IFNR (Table 1).
The parent IFNR had a molecular mass of 19256, consistent
with the calculated individual amino acid composition. The
palmitoyl groups are expected to be linked to nine lysine
residues and the terminal cysteine under these conditions
2
steady-state flux (ng/cm /h)a
1.47
1.65 × 10
6.85 × 10
23.8 ± 17.4
2.71
3.03 × 10
5.45 × 10
42.7 ± 25.70
b
-5
-5
permeability coefficient (cm/h)
2
c
-12
-12
diffusion coefficient (cm /s)
total amount diffused in 24 h: (ng/
2
(
pH 7.2). Two further linking sites may be arginine
residues.
An tivir a l ActivitysSince our interest is to develop
cm ) n ) 6
a
Determined by regression analysis of the linear portion of the average
b
cumulative amount of drug diffused (Q) vs time (t) curve. Permeability
novel delivery systems for the efficient delivery of IFNR in
the treatment of human papillomavirus infections (condylo-
mata acuminata), we focused particularly on the changes
in antiviral properties of this drug. The “crude p-IFNR”
preparation was evaluated for antiviral activity as a
preliminary measure of whether the acylated derivatives
retained the original activity (Figure 3). The results
showed a complete loss of activity when the reaction was
carried out in a medium containing DMF and a 50%
preservation of activity in DMSO. At this point it is not
certain whether the partial loss of antiviral activity is due
to “overacylation”. The separation of the crude fraction into
single acyl-protein fractions will contribute to the clarifica-
tion of this question. The optimum degree of fatty acylation
will have to be addressed. Although direct comparison is
not possible, attempts have been made to prepare palmitoyl
derivatives of insulin in order to improve its lipophilicity.24
It was found that the biological activity (as measured by
the hypoglycemic effect in male rats) of the mono and
dipalmitoyl derivatives decreased by 12 and 52%, respec-
tively, compared to the parent compound, and the tripalmi-
toyl derivative had no hypoglycemic effect. At the same
time, immunoreactivity decreased by 50 and 70%, respec-
tively, for the mono and dipalmitoyl derivative, which may
be therapeutically advantageous.
coefficient (P) was calculated from Fick’s first law: (dQ/dt)SS ) JSS ) P∆C;
where P ) K D/h [J − steady-state flux; ∆C − concentration difference
p
SS
p
between donor and receiver compartments; K − partition coefficient between
2
skin and the preparation]. c Diffusion coefficient was calculated from D ) h /
L; where h − thickness of the stratum corneum (0.001 cm); L − lagtime (s).
6
The skin absorption of p-IFNR and IFNR was tested in a
gel type pharmaceutical formulation. The cutaneous ab-
sorption of p-IFNR was about 5-6 times greater than the
IFNR (Table 2A). The amount of p-IFNR and IFNR in
whole skin after 24 h of treatment was 2.106 ( 1.216 µg/
2
2
cm and 0.407 ( 0.108 µg/cm , respectively. This repre-
sents 11.53 ( 6.66 and 1.76 ( 0.47% of total drug applied,
respectively. In the viable skin layers the difference
between p-IFNR and the parent protein was 8-10-fold,
2
1.881 ( 1.160 µg/cm (10.29 ( 6.35%) and 0.228 ( 0.091
2
µg/cm (0.98 ( 0.39%).
It is worth mentioning that both the synthesis and the
cutaneous absorption of p-IFNR were found to be very
reproducible.
The calculated percutaneous absorption parameters for
the solutions of p-IFNR and IFNR are shown in Table 2B.
Approximately two times higher flux was detected for
p-IFNR compared to the nonfatty acylated IFNR. The total
amount of drug diffused in 24 h was also about two times
higher for the p-IFNR.
Interferons have numerous functional effects including
antiviral and antitumor action, immunoregulatory action,
macrophage activation, enhancement of cytotoxicity of
lymphocytes, hormonal interactions, influence on subse-
quent production of other cytokines, oncogene regulation,
and cellular differentiation. It is known that some cytokine
functions do not involve receptor binding and can act
directly on intercellular signaling pathways.2 It was
shown that substances which increase cellular permeability
may actually enhance certain activities which are not
associated with receptor binding. Therefore, partial loss
of antiviral activity does not exclude the possibility that
some other functions of IFNR are unchanged or perhaps
increased. Also, partial loss of antiviral activity may not
be prohibitive if the degree of cutaneous penetration
enhancement is substantially higher.
The evaluation of the effect of IFNR from various topical
vehicles (gel, petrolatum, solution)12 and from liposomes
was reported previously. These results indicated only a
marginal reduction of the number of lesions, the total lesion
area, and the lesion virus titer by the gel and petrolatum
26,27
5
6
formulations at 300 × 10 U/g IFNR concentration when
used four times a day for 3-5 days starting 24 h prior to
virus inoculation in a cutaneous herpes model in guinea
pigs. When treatment was initiated 24 h after virus
inoculation, no reduction in the number of lesions could
be observed. The reduction in any of the parameters tested
was not statistically significant compared to the vehicle
only control regardless whether treatment started before
or after virus inoculation.12
Cu ta n eou s Absor p tion of p -IF NrsThe main purpose
of these experiments was to determine whether the in-
creased lipophilicity of the protein or the covalently at-
tached fatty acyl residues would enhance the penetration
of IFNR into or through the skin. Similarly to IFNR,
p-IFNR was successfully iodinated by the lactoperoxidase
method and used for the skin absorption experiments. The
results showed that both the cutaneous and percutaneous
absorption of p-IFNR was greater than the original protein.
The liposomal delivery of IFNR was reported by Weiner
et al.26 also in the cutaneous herpes simplex guinea pig
model. When IFNR was entrapped into certain type of
liposomes (“skin lipid” liposomes made by the dehydration-
rehydration method) significant reduction of lesion scores
could be observed.
Studies in patients diagnosed for having condylomata
acuminata (genital warts) also confirmed that liposome
encapsulation of IFNR enhances delivery into the wart
Journal of Pharmaceutical Sciences / 1207
Vol. 87, No. 10, October 1998

lesions.27 Due to the enhanced delivery, improved clinical
effect could be observed.
11. Farr, B. M.; Gwaltney, J . M.; Adams, K. F.; Hayden, F. G.
Intranasal interferon-R2 for prevention of natural rhinovirus
colds. Antimicrob. Agents Chemother. 1984, 26, 31-34.
On the basis of the literature data and our previous
results,27 it is fairly certain that the lack of efficacy is
related to a delivery problem. The elimination of the virus
from the lesions is prerequisite for preventing recurrences
after apparent successful treatment. The maximization of
the efficiency of delivery of IFNR will be necessary in order
to achieve this.
1
2. Freeman, D. J .; McKeough, M. B.; Spruance, S. L. Recom-
binant human interferon RA treatment of an experimental
cutaneous herpes simplex virus type 1 infection of guinea
pigs. J . Interferon Res. 1987, 7, 213-222.
3. Keay, S.; Teng, N.; Eisenberg, M.; Story, B.; Sellers, P. W.;
Merigan, T. C. Topical interferon for treating condyloma
acuminata in women. J . Infectious Dis. 1988, 158(5), 934-
1
9
39.
1
1
4. Glezerman, M.; Cohen, V. ; Shoham, J .; et al. Placebo-
controlled trial of topical interferon in labial and genital
herpes. Lancet 1988, 1, 150-152.
Conclusions
5. Kobrinskii, G. D.; Melnikov, V. R.; Kulakov, V. N.; Lvov, N.
D.; Bolotin, I. M.; Barinskii. I. F. Treatment of experimental
genital herpes with liposomal interferon. Biomed. Sci. 1991,
Our results indicate that fatty acylation may increase
the cutaneous and percutaneous delivery of IFN R, and
therefore there are good possibilities for therapeutic im-
provement. There are a number of questions which we
hope to answer in the future to fully characterize the
acylation sites, including the effect of acylation at different
sites with different fatty acids on antiviral activity, the
metabolic fate of the new derivatives in the skin, receptor
binding ability, and subsequent cellular activation.
2
, 29-32.
1
6. Hamuryudan, V.; Yurdakul, S.; Rosenkaimer, F.; Yazici, H.
Inefficacy of topical alpha interferon in the treatment of oral
ulcers of Bechet’s Syndrome: a randomized, double blind
trial. Br. J . Rheumatol. 1991, 30, 395-396.
1
1
7. Sarkar, F. H.; Gupta, S. L. Radiolabeling of human leukocyte
1
25
and immune interferons with
I and lactoperoxidase.
Methods Enzymol. 1986, 119, 263-267.
8. Ohmann, B.; Gilchrist, J . E.; Babiuk, L. A. Effect of recom-
binant DNA-produced bovine IFN alpha (BoIFN-) on the
interaction between bovine alveolar macrophages and bovine
herpesvirus type-I. J . Gen. Virol. 1984, 65, 1485-95.
References and Notes
. Hayakawa, E.; Chien, D. S.; Inagaki, K.; Yamamoto, A.;
Wang, W. Lee, V. H. L. Conjuctival penetration of insulin
and peptide drugs in the albino rabbit. Pharm. Res. 1992, 9,
1
9. Wetzel, R. Assignment of the disulphide bonds of leukocyte
1
2
interferon. Nature 1981, 289, 606-607.
20. Morehead, H.; J ohnston, P. D.; Wetzel, R. Roles of the 29-
138 disulfide bond of subtype A of human a interferon in its
antiviral activity and conformational stability. Biochemistry
1984, 23, 2500-2507.
7
69-775.
. Richardson, J . L.; Illum, L.; Thomas, N. W. Vaginal absorp-
tion of insulin in the rat: Effect of penetration enhancers
on insulin uptake and mucosal histology. Pharm. Res. 1992,
2
1. Bewley, T. A.; Levine, H. L.; Wetzel; R. Structural features
of human leukocyte interferon A as determined by circular
dichroism spectroscopy Int. J . Pept. Protein Res. 1982, 20,
1
9 878-883.
3
4
. Bodde, H. E.; Verhoef, J . C.; Ponec, M. Transdermal peptide
delivery. Biochem. Soc. Trans. 1989, 1, 943-945.
9
3-96.
2
2
2
2. Fish, E. N.; Banerjee, K.; Stebbing, N. The role of three
domains in the biological activity of human interferon-R. J .
Interferon Res. 1989, 9, 97-114.
. Hoogstraate, A. J .; Verhoef, J . C.; Ijzerman, A. P.; Spies, F.;
Bodde, H. E. Kinetics, ultrastructural aspects and modeling
of transdermal peptide flux enhancement by N-alkylazacy-
cloheptanones. Int. J . Pharm. 1991, 76, 37-47.
3. Korn, A. P.; Rose, D. R.; Fish, E. N. Three-dimensional model
of a human interferon-R consensus sequence. J . Interferon
Res. 1994, 14, 1-9.
4. Hashimoto, M.; Takada, K.; Kiso, Y.; Muranishi, S. Synthesis
of palmitoyl derivatives of insulin and their biological activi-
ties. Pharm. Res. 1989, 6, 171-176.
5
6
7
. Choi, H. K.; Flynn, G. L.; Amidon, G. L. Transdermal delivery
of bioactive peptides: The effect of n-decylmethyl sulfoxide,
pH and inhibitors on enkephalin metabolism and transport.
Pharm. Res. 1990, 7, 1099-1106.
. Morimoto, K.; Iwakura, Y.; Miyazaki, M.; Nakatani, E.
Effects of proteolytic enzyme inhibitors of enhancement of
transdermal iontophoretic delivery of vasopressin and an
analogue in rats. Int. J . Pharm. 1992, 81, 119-125.
. Kumar, S.; Char, H.; Patel, S.; et al. In vivo transdermal
iontophoretic delivery of growth hormone releasing factor
GRF (1-44) in hairless guinea pigs. J . Controlled Release
25. Baron, S.; Tyring, S. K.; Fleishman, J r., W. R.; Coppenhaver,
D. H.; Niesel, D. W.; Klimpel, G. R.; Stanton, G. J ,; Hughes,
T. K. Interferons: Mechanisms of action and clinical applica-
tions. J AMA 1991, 266, 1375-1383.
26. Weiner, N.; Williams, N.; Birch, G.; Ramachandran, C.
Shipman, J r., C.; Flynn, G. Topical delivery of liposomaly
encapsulated interferon evaluated in a cutaneous Herpes
guinea pig model. Antimicr. Agents Chemother. 1989, 33,
1217-1221.
1
992, 18, 213-220.
8
9
. Clemens, M. J . Molecular mechanism of interferon action.
Br. J . Clin. Pract. 1988, 42, 5-12 1988.
. Pestka, S.; Langer, J . A.; Zoon, K. C.; Samuel, C. E..
Interferons and their actions. Annu. Rev. Biochem. 1987, 56,
2
7. Foldvari, M.; Moreland, A. Clinical Observations with topical
liposome-encapsulated interferon alpha for the treatment of
genital papillomavirus infections. J . Liposome Res. 1997, 7,
7
27-777.
1
0. Baglioni, C.; Phipps, R. J .. Nasal absorption of interferon:
1
15-126.
Enhancement by surfactant agents. J . Interferon Res.1990,
1
0, 497-504.
J S980146K
1
208 / Journal of Pharmaceutical Sciences
Vol. 87, No. 10, October 1998