- Email: [email protected]
Baseline studies on iontophoretic transport in hairless mouse skin: the effect of applied voltage drop and ph on the iontophoresis of a model weak electrolyte
Journal of Membrane Science, 49 (1990) 305-320 Elsevier Science Publishers B.V.. Amsterdam - Printed
305 in The Netherlands
BASELINE STUDIES ON IONTOPHORETIC TRANSPORT IN HAIRLESS MOUSE SKIN: THE EFFECT OF APPLIED VOLTAGE DROP AND pH ON THE IONTOPHORESIS OF A MODEL WEAK ELECTROLYTE
S.M. SIMS*
and WI.
Department (U.S.A.)
of Pharmaceutics,
(Received
HIGUCHI
March 29,1989;
301 Skaggs Hall, University
accepted
of Utah, Salt Lake City, UT 84112
in revised form September
12,1989)
Summary A physical
model for the iontophoretic
transport
of a weak electrolyte
across hairless
mouse
skin has been examined. The stratum corneum is modelled as parallel lipoidal and aqueous pore pathways for diffusion and is in series with a porous matrix representing the dermis-epidermis. It was assumed that only the undissociated species could penetrate the lipid phase while both charged and uncharged species could permeate the pore route. The applied electric field is assumed to influence only the charged species in the aqueous pore according to the Nernst-Planck theory. Experiments were done over a wide pH range using a four-electrode potentiostat system to control the voltage drop across the membrane. Butyric acid was chosen as the model weak electrolyte. Glucose was used to independently assess membrane damage and solvent flow effects. Permeability coefficients as a function of pH were determined both for butyric acid and glucose before, during and after iontophoresis. Experimental permeability coefficients semi-quantitatively followed the theoretical predictions.
Introduction Iontophoresis is a technique which may be employed to increase the transport of ions into or through skin by the application of an external electric field across the skin [ 11. It is a non-invasive technique which could potentially be used to administer a wide range of medicinal agents for local and systemic therapy [ 2-91. While there have been many clinically oriented iontophoresis studies reported in the literature, efforts needed to understand the fundamental physical chemistry involved in iontophoresis are just beginning [ l,lO-141. An understanding of the mechanism of charged and uncharged solute transport across skin, both with and without an applied electric field, is of fundamental impor*To whom correspondence
0376-7388/90/$03.50
should be addressed.
0 1990 Elsevier
Science
Publishers
B.V.
tance. A knowledge of the type of diffusion pathway, its limitations with regard to the size of the permeant molecule or ion, its charge status, and how each pathway can be manipulated by an external electric field, would aid both the understanding of iontophoresis, and the development of iontophoresis for drug delivery. By relating the iontophoretic fluxes to the properties of both membrane and drug, it may be possible to extrapolate these results beyond experimental systems to practical iontophoretic drug delivery devices. Several studies have provided a foundation for the understanding of the movement of molecules and ions in the stratum corneum, the barrier layer of skin. Scheuplein [ 151 addressed the issue that shunts such as hair follicles and sweat glands may be a route of solute diffusion. He suggested that shunts may be important for solute transport, especially during the transient period prior to steady state. However, transport through the intact stratum corneum (i.e., the transepidermal route) generally appears to be the dominant pathway of solute diffusion in stratum corneum. Various mechanisms and models have been proposed for the transport of solutes across the bulk stratum corneum. Michaels et al. [ 161 proposed a “brick and mortar” type model of the skin in which a solute could penetrate by only two routes: one required alternate passage through protein (corneocyte regions ) and lipid phases (intercellular lipid regions) and the other transit only through a continuous lipid phase. This model cannot accommodate ionic solutes which presumably diffuse only through an aqueous path. Recently, Ghanem et al. [ 171 have discussed a model in which the stratum corneum is treated as the diffusional barrier with parallel lipid and aqueous pore pathways for diffusion. This model provides a continuous aqueous route and could, therefore, permit ionic solutes to penetrate. In this paper the parallel pathway model is examined as a means of quantifying the iontophoretic transport of a weak electrolyte. In this model it is assumed that the electric field is able to enhance the movement of ions along the pore pathway and experimental data obtained with a weak electrolyte (butyric acid) are employed to test this aspect of the model. Theory
Physical model of the skin In this section a physical model is presented for the transport of a weak electrolyte across skin both with and without the influence of an applied electric field. The model is shown in Fig. 1. It consists of the stratum corneum in series with a porous matrix, the dermis-epidermis. The stratum corneum is comprised of parallel lipoidal and aqueous pore pathways for diffusion. It is assumed that (a) only non-ionized species are able to partition into and be transported in the lipoidal phase, and (b) both ionized and non-ionized species penetrate the aqueous route with essentially equal facility (i.e., with essentially the same permeability coefficient).
307 PHYSICAL
MODEL
OF THE
SKIN RECEIVER CHAMBER
LIPID POROUS
PORES
MATRIX LIPID
::t
-STRATUM
CORNEUM~OERMIS-EPIDERMIS-
Fig. 1. Schematic diagram of the physical model for the diffusion of charged and uncharged species across skin. The stratum corneum consists of parallel lipoidal and aqueous pore pathways which are in series with the porous dermis-epidermis layer.
For the steady state situation electrolyte:
we may write for the total flux (JT) of a weak
JT = APT AC
(1)
where PT is the total (or effective) permeability coefficient of a weak electrolyte; A is the area for diffusion and AC is the difference between the donor and receiver chamber permeant concentration. Also: JT = dQ/dt
(2)
where dQ/dt is the steady-state from donor to receiver chamber
slope of the amount of species transported as a function of time. Therefore:
PT = l/AAC dQ/dt
(3)
Now, for the model presented in Fig. 1, PT may be separated into contributions from the stratum corneum and the dermis-epidermis (resistance in a series): 1 -_-+6
1 PSC
1
(4)
PDE
where Psc is the permeability coefficient of stratum corneum and PDE is the permeability coefficient of the dermis-epidermis combination. Furthermore, for the situation of parallel lipoidal and aqueous pore pathways in the stratum corneum, we may write: Psc =X,P,
+pp
where PL is the permeability
(5) coefficient
for the lipoidal pathway,
Pp is the
308
permeability coefficient for the aqueous pore pathway, undissociated species. Combining eqns. (4) and (5), we have: 1 PT -Pp
1 +xsPL
and Xs the fraction
of
1 +PD,
Except for extremely lipophilic permeants, P DE is generally very large compared to PsC, i.e., the stratum corneum is usually the primary diffusional barrier. In this case, l/P nE can be neglected and eqn. (6 ) becomes: PT =Pp +xsPL
(7)
As will be seen in the present studies using butyric acid, PI)E is indeed much greater than PSC for hairless mouse skin, and therefore eqn. (7)is a good approximation of eqn. (6). Enhancement of ion transport via an applied electric potential While the total passive permeability coefficient may be described by eqn. (7)) the contribution to transport from an applied voltage drop across the skin requires further elaboration. We may assume that only the undissociated species enters the lipid domain whereas both ionized and non-ionized species may permeate the pore pathway. Further, we may assume the electric field influences only the charged species in the pore pathway. For this situation, we write the following equation for the permeability coefficient for a weak electrolyte solute in the presence of an electric field:
(8) where Ek,n is the enhancement factor for the ionized species due to the applied electric potential drop across the skin. The enhancement factor is defined as (1): Eion = Ji/Jp
(9)
where J, and Jp are the fluxes of the ionized species with and without the applied voltage, respectively. It can be shown [ 1,181 from the Nernst-Planck equation with the constant field assumption that: Ei,,,=-K;[l-exp(K)] where K is a dimensionless K= -,qF
Ay/RT
(16) constant
defined as: (11)
Here, Z, is the charge on species i, F the Faraday constant, Ay the electric potential drop, R the gas constant and T the absolute temperature. Equations (8-11)may be used to predict PT when PL, Pp and X, are known. Such calculations are shown for butyric acid in Fig. 2, where PT values are presented
309
PH
Fig. 2. Theoretical butyric acid permeability coefficient-pH profiles for passive diffusion (using eqn. 7) and for applied voltage drops of 250 and 500 mV (using eqn. 8). Pp= 1.8X lo-'cm/set; P,=1.3X10-6cm/sec;E,,,= 9.4; Ebo,= 18.7; pKaz4.84 at 37°C.
as a function of pH for passive diffusion and diffusion under applied voltages of 250 and 500 millivolts (mV) . Experimental
Materials Iontophoresis studies were conducted with radiolabeled [“Cl -1-butyric acid (specific activity 15 mCi/mmol) and t3H] -3-glucose (specific activity 13.5 Ci/ mmol) obtained from New England Nuclear Corporation (Boston, MA). Butyrate was chosen as a model anion because the fraction dissociated is pH dependent thus allowing assessment of the electric field effects on both the lipid and pore pathway permeabilities. Glucose was used as a model non-electrolyte to independently assess “membrane damage” and to obtain insight into solvent flow effects. Buffers ranged in pH from 2.4 to 9.5. Buffer composition [19] varied with pH and in all cases NaCl was the dominant salt species. Chemicals were reagent grade and used as received. All buffers had an ionic strength of 0.1 and were prepared in distilled, de-ionized water. Iontophoresis apparatus Skin permeabilities were measured using a four-electrode potentiostat system which maintains a constant voltage drop across a membrane mounted in a two-chamber diffusion cell. Figure 3 is a schematic diagram of the potentiostat system and its associated two-chamber diffusion cell. The potential drop across the Luggin capillary probes is maintained at the desired value by the
310 OUR-ELECTRODE POTENTIOSTAT
CE=COUNTER ELECTRODE RE =REFERENCE ELECTRODE
Fig. 3. Schematic
II/
REFERENCE
diagram of the four-electrode
potentiostat
system.
ELECTRODE
)RT
JACKET’
Fi. 4. Diagram of a half-cell for diffusion and its associated
Luggin capillary.
potentiostat which drives the appropriate current through the cell via the counter electrodes. This system has been described in detail by Srinivasan et al. [ 1]who also compared it to the conventional, two-electrode, constant current approach. Figure 4 shows one half of the two-chamber diffusion cell with its associated Luggin capillary and counter electrode. The cell is water-jacketed and main-
311
tained at 37 ’ C by circulating water (Brinkmann RM-6 model circulating water bath, American Scientific ). Each compartment has a volume of 5 ml. Iontophoresis experiments with hairless mouse skin Following cervical dislocation [ 201 of the hairless mouse (8-12 week old hairless mouse skin; SKH-HR-l), abdominal skin was excised and mounted between the two half cells with the stratum corneum facing the donor side. The two Luggin capillaries were positioned on each side with their tips very close to the membrane on either side (Fig. 4). The receiver chamber was filled with buffer of the pH to be tested. The donor solution contained the same buffer premixed with tracer levels of radiolabeled solute. The electrolyte solutions in the Luggin capillaries were the same as that in the receiver chamber. This was done to ensure that there was no interfacial potential drop between capillary solution and the cell solution, thus the applied voltage drop is the voltage drop across the membrane [ 11. Samples (100 ~1) were withdrawn from the donor side at the beginning and end of the experiment for analysis. The difference in the counts (dpm) between these two samples were always negligible. Both chambers were stirred by teflon rods with teflon propellers mounted at the bottom (Fig. 4). The teflon rods were connected to constant speed (150 r-pm) motors. The entire apparatus was maintained at 37 oC. Preliminary experiments indicated that the electric field may cause alterations in the membrane. To monitor and better understand the changes in the intrinsic transport properties caused by the electric field, each experiment was composed of three stages (Fig. 5 ) . First a passive permeability coefficient was
TIME
WR)
Fig. 5. Typical permeation
profile for a three-stage
experiment:
(m) butyric acid; ( 0 ) glucose.
312
determined in which PT = PT,O. In the second stage an electric field was applied and the total permeability coefficient due to this potential drop was determined, i.e., PT = PT,dyl Finally, a second passive permeability coefficient was determined, i.e., PT = PT,O.. A more precise description of each stage follows. Stage I: This stage involves the measurement of the initial passive permeability of total butyrate and/or glucose through hairless mouse skin. After assembly, the experimental system was allowed to equilibrate for six to seven hours and reach steady state. Five 1 ml samples were withdrawn at 30 min. intervals from the receiver chamber. Each sample was mixed with 10 ml OptiFluor (Packard Instrument Co. ) scintillation fluid and counted on a Beckman Liquid Scintillation Counter (Model LS-7500). The samples were replaced with fresh buffer. The initial passive permeability coefficient, PT,O, was calculated from eqn. (3 ). Stage II: During this stage, which lasted one hour, the iontophoretic permeability coefficient, PT,dlv,was determined for the permeant at an applied voltage drop, Av, across the skin. After the final sampling in Stage I, the cells were connected to the potentiostat (JAS Instrumental Systems, Inc., Salt Lake City, UT) and a fixed voltage drop was applied across the skin (cathode on donor side, anode on receiver side). Four 1 ml samples were taken at 15 minute intervals from the receiver side and analyzed as in Stage I. Samples were replaced with fresh buffer. PT,dWwas calculated using the last three points of Stage II (Fig. 5) from eqn. (3 ) as before. The cell current was monitored continuously and the applied voltage was confirmed via a voltmeter (Beckman Industrial Model 310). Stage III: At the end of Stage II the receiver chamber was flushed three times, then refilled with fresh buffer. After one hour, the passive permeability coefficient P,,,O. was determined as before. The total time for each experiment was, at most, 11-12 hours. Passive experiments at each pH indicated that the barrier properties of the skin remained intact during the entire experiment. Other authors [21,22] have also shown that hairless mouse skin retains its barrier function in aqueous buffer over this time range. At the end of each experiment the pH in both chambers was checked. The pH varied only 0.2-0.3 pH units in extreme cases. Results Total butyrate transport
Figure 5 presents the result of a typical three stage, dual label, permeation experiment. The best fit straight line (neglecting the first point in each stage as being part of lagtime) for each stage was used to calculate the total permeability coefficient, PT. Table 1 lists the experimentally determined permeability coefficients for total butyrate for each stage of the experiment as a function of pH and applied voltage drop. In Table 2 these data are converted to the total
313 TABLE
1
Experimental
permeability No. of expts.
PH
coefficients
for butyrate with hairless mouse skin
Permeability
P T.0
coefficient”
( x 1O-7 cm/set) P T&
P T,O'
Ar//= 250 mV 2.4
13+4.5
12kl.O
12+2.0
5 3 3 3
7.6k4.5 5.9 + 4.0 0.2 f 0.8 0.2 + 0.0 0.4kO.3
14k8.0 8.2 ? 2.7 1.2 * 1.2 1.6kO.7 2.7 & 1.8
14i3.0 8.2 ? 2.2 0.7 + 0.3 0.5 I!Y0.2 0.7kO.6
4 4
14+3.0 5.7k4.1
3 3 4
6.6f3.5 2.8kO.4 0.3 +0.1 0.2LO.l
32f13 14k5.0 17 ? 6.0
21+3.0 12 * 5.0 12 * 7.0 3.9 * 0.7
3 4
3.6 4.8 6.0 7.5 9.0 Ay/= 500 mV 2.4 3.6 4.8 5.5 6.5 7.5
5 6
9.5
0.5 I!z0.3
llk4.0 8.4 & 4.8 2.5kO.6 30*10
1.9 !C 1.0 0.4kO.l 2.6? 1.0
“Mean ?I S.D.
butyrate effective enhancement factors, E, and EP, as defined by the following equations: E, = PT,&%,o
(12)
E, =
(13)
pT,dy//pT,O’
E, represents the enhancement of butyric acid under the influence of the voltage, dy/, over the initial passive value (PT,o), while E2 is the enhancement based on PT,o, as the “control”. Also shown in Table 2 is the ratio El/E2 which is a measure of the membrane alteration (probed by using butyric acid as the permeant) caused by the applied electric field. Glucose transport Table 3 lists the experimentally determined total permeability coefficients (PO,G,P&G, Po,Gs) for glucose for hairless mouse skin at various pH values. Table 4 shows the enhancement value for glucose, EG1 and EG2, which are defined as follows: EGI
= P&,G/PO,G
(14)
EGZ
= Pdyl,G/PO,G’
(15)
The ratio EGl/EGz (Table 4) is a measure of the membrane alteration caused
314 TABLE
2
Enhancement
factors for butyric acid with hairless mouse skin Enhancement
PH
&= 250 mV 2.4 3.6 4.8 6.0 7.5 9.0 Ay/= 500 mV 2.4
factor”
1.1 kO.5 2.0 ? 0.5 l.B? 1.0 6.0 + 5.2 7.0 IfI2.6
l.OLO.1 1.0+0.3 1.1 i-o.4 1.65 1.0 3.6kO.9
l.OkO.6 2.3 ?I 1.0 1.8 + 1.0 3.9+ 1.9 2.OkO.5
9.0 & 5.2
1.4i-7.5
1.6kO.6
2.4t0.7
1.5 * 0.5
1.6kO.3
3.6 4.8 5.5
3.2 i- 1.4 2.8kO.7 4.0* 1.3
1.4 I! 0.8 1.8t0.8 2.8 I!Z0.9
2.4i0.8 1.7kO.5 1.4iO.l
6.5 7.5 9.5
22+7.1 14? 6.9 54*19
4.4+ 1.7 6.4+ 1.7 12+3.7
6.2 + 0.9 2.2 5 0.8 4.5i2.6
“Mean + S.D.
TABLE
3
Experimental PH
AMY=250 mV 3.6 4.8 6.0 7.5 9.0 AI,Y= 500 mV 3.6 4.8 5.5 6.5 7.5 9.5 “Mean + S.D.
permeability No. of expts.
coefficients
for glucose with hairless mouse skin
Experimental
permeability
coefficient
( x lop8 cm/sec)a
P 0s
P&
PO,,:,
6 3 4 4 6
4.5 + 2.2 2.1* 1.3 0.9 + 0.5 2.2 i 0.4 2.5k2.1
301?118 6.Oi-5.5
24+25 4.Ok3.0 1.5 i 1.1 3.7 + 1.7 5.2k3.1
2 3 4 4
4.9 i 3.8 4.0 I!c1.9 1.6+ 1.3 1.1 kO.4
3 4
1.2 I!I0.3 5.4k3.1
0.6 + 0.3 1.5 I! 0.6 1.8? 1.8
25i16 16? 7.0 5.6k5.1 2.5 z?z1.2 2.5kO.l 6.2 + 1.4
lOk9.0 9.8i2.2 3.8t 3.2 1.9 I! 1.0 4.9ir2.6 3.1 z!z1.3
315 TABLE
4
Enhancement
factors for glucose with hairless mouse skin Enhancement
factor’
E Gl
E G2
EG@GP
3.6 4.8
7.0 + 2.2 2.3 k 1.1
2.0 + 1.0 1.6kO.3
4.9f4.1 1.5kO.7
6.0 7.5 9.0
0.8 + 0.6 0.7 + 0.2 0.7f0.4
0.6kO.3 1.1 dIo.9 1.4 + 1.7
1.620.8 1.1 * 0.7 1.4 f 1.2
5.6? 1.2 4.3 f 1.0 3.62 1.3 2.6 I!Z1.6
3.0 + 0.9 1.6kO.5 1.4kO.5 1.5kO.4
1.9kO.l 2.9 + 1.3 2.8i 1.0 1.9 + 1.3
2.lkO.5 1.3 z!z0.7
0.6kO.2 0.2 * 0.1
4.3f3.1 5.6t2.0
PH
Av/= 250 mV
Ay/= 500 mV 3.6 4.8 5.5 6.5 7.5 9.5 “Mean 2 S.D.
by the electric field. Since it was assumed that glucose is transported only along the pore pathways (for all pH conditions), this ratio was expected to be a general measure of the irreversible alteration in the pore pathway caused by the applied voltage. The ratio Eol/EGZ was, therefore, expected to be different from the ratio EJ&, except at high pH conditions where the butyrate ion transport was expected to be much more important than the non-ionized butyric acid transport across skin. Discussion Comparison of results with theoretical prediction Considering a meaningful approach to analyze the experimental results with the model (i.e., eqns. 8-ll),one is faced with the problem of the significant membrane alterations caused by the applied voltage in most instances (Tables 2 and 4). An important question here was whether PT,o, PT,Oz,or some average should be used as the reference for calculating the enhancement (e.g., E, versus E2). In an attempt to gain some insight on the matter, the passive glucose permeability coefficients were determined as a function of the time of application of the voltage. Figure 6 shows how “membrane damage” may vary with the time of application of the voltage over a 60 minute time period. Although the variabilities are seen to be large, these results strongly suggest that membrane damage in hairless mouse skin occurs early and the effect tends
316
a I
B
01~ 1
T-rllil 20
0
TIME
OF
40
APPLIED
60
80
VOLTAGEtMIN.)
Fig. 6. Glucose EG,/EG, ratio as a function of the time voltage was applied. (0 ) pH 7.5,250 mV; ( n ) pH 7.5,500 mV; (A ) pH 9.5,500 mV.
2
4
6
8
IO
PH
Fig. 7. Experimental butyric acid PO, ( n ) and PAY/(0 ) as a function of pH. Solid lines represent the theoretical prediction. PP = 4.5 x 10-s cm/set; PL= 1.2 X 10m6 cm/set; I& =9.4; AY= 250 mV.
to plateau between 30 and 60 minutes. Because of these data, it was judged that PT,O.would be better than PT,O (or the average of the two) in comparing the experimental results with theory. Accordingly, the PT,O,values were used to deduce the PL and the Pp values for use in eqns. (7) and (8). As can be seen from Figs. 7 and 8, there is semi-quantitative agreement between experiment and theory. At low pH, the theory predicts that transport would involve predominantly the undissociated species diffusing along the lipid pathway, and therefore, there should be little or no iontophoretic enhancement. Here, there is good agreement between experiment and theory, and the
317
2
4
6
8
IO
Fig. 8. Experimental butyric acid P,. (m) and PAW(0 ) values as a function of pH. Solid lines represent the theoretical prediction. PP=4.0X 10e8 cm/set; PL= 1.2 X 10d6 cm/set; I&,,= 18.7; dv/= 500 mV.
experimental permeability coefficients approach the same asymptote with or without the applied voltage. At high pH ( > 7.4)) the pore pathway should dominate and the differences between PT,d,,,and PT,Osbecome large. Under high pH conditions, the experimental and the theoretical enhancement values are seen to agree within about a factor two or better. Membrane alteration due to the applied voltage Membrane damage has been defined as an increase in the passive permeability of a membrane [ 11. Alteration in the membrane barrier properties has been suggested by many authors. Bellantone et al. [ 231 saw a varying increase in benzoic acid flux though hairless mouse skin after termination of an applied current. Burnette and Bagniefski [lo] demonstrated that the impedance of hairless mouse skin was generally lower and Na+ flux was higher after iontophoresis. Similarly, human skin was shown to have a lower impedance and higher 3Hz0 flux after iontophoresis [24]. These studies indicate that skin undergoes some change due to iontophoresis such that the passive permeability of the skin is increased. Tables 2 and 4 present the E1/ES and the EG1/EGPratios which represent membrane damage when butyric acid and glucose, respectively, are used as the probes. The glucose probes the aqueous pore pathway over the entire pH range while butyric acid probes both the lipid and pore pathways, depending upon the pH (eqns. 7 and 8). Tables 2 and 4 show that there is generally some significant membrane alteration at both 250 and 500 mV. Because of the large variabilities in the data, it is difficult to judge (a) whether there is less damage in the lipid pathway over the pore pathway, and (b) whether the damage in
318
the pore pathway is dependent on pH. The data do indicate, however, that at high pH (especially at 500 mV) the EG1/EG2and E,/E, results correlate well; this is consistent with the expectation (eqn. 8) that, at high pH, the pore pathway will dominate in butyric acid permeation. The data also indicate that at high pH and 500 mV, membrane damage is greater than at low pH and lower voltages. This latter thought is consistent with the high baseline passive value seen in Fig. 8. Solvent flow effects due to the applied potential drop Also of importance in these experiments is the possible existence of solvent flow. Solvent flow has been suggested in several recent studies. Gangarosa et al. [ 251 demonstrated that the permeation of non-electrolytes through hairless mouse skin can be enhanced by iontophoresis due to convective flow. Pikal and Shah [ 111 suggested bulk fluid flow by electro-osmosis as the mechanism by which iontophoretic enhancement of the permeability of uncharged species occurs. Glikfeld et al. [26] reported measurable volume changes in both diffusion cell chambers when a current of 0.5 mA was applied to the system. Charged species may also be affected by solvent flow. Solvent flow is generally believed to occur in the direction of counter ion flow [ 11,251. In the present study, where the cathode is placed in the donor chamber and the anode in the receiver chamber, it was thought that glucose transport under the influence of the electric field would be retarded relative to passive glucose transport and that such information could be used to assess solvent flow during iontophoresis. Unfortunately, the data in Table 4 show considerable variability, such that a rigorous interpretation of the data is not possible. It would be instructive, however, to attempt an upper limit estimation of the solvent flow effects upon the butyrate ion transport, employing the data in Table 4. Incorporation of solvent flow effects in the Nernst-Planck equation [ 271 results in an equation similar to eqn. (11) but with a correction term for solvent flow [ 281. If an Eo2 value of 0.2 (see Table 4) is used to estimate solvent flow, one finds that the correction for the butyrate ion enhancement is lo-20%. This estimate suggests that solvent flow effects are relatively small and consistent with the views of other investigators [ 11,29,30]. Conclusions The experimentally determined pH profiles for the iontophoretic transport through hairless mouse skin of butyric acid agree semi-quantitatively with the proposed parallel lipid-aqueous pore pathway model for the transport of ions across skin. The effects of membrane damage were taken into account in the data analysis and the influence of solvent flow was assessed.
319
References V. Srinivasan, S.M. Sims, W.I. Higuchi, C.R. Behl and S. Pans, Iontophoretic transport of drugs: a constant voltage approach, in: J. Kost (Ed.), Pulsed and Self-Regulated Drug Delivery, CRC Press, Boca Raton, FL., 1990. 2 P. Tyle, Iontophoretic devices for drug delivery, Pharm. Res., 3 (1986) 318. 3 A.K. Banga and Y.W. Chien, Iontophoretic delivery of drugs: fundamentals, developments and biomedical applications, J. Controlled Release, 7 (1988) 1. 4 M. Comeau, R.B. Brummett and J. Vernon, Local anaesthesia of the ear by iontophoresis, Arch. Otolaryngol., 98 (1973) 114. 5 L.P. Gangarosa, Iontophoresis for surface local anaesthesia, J. Amer. Dent. Assoc., 88 (1974) 125. 6 K. Okabe, H. Yamaguchi and Y. Kawai, New iontophoretic transdermal administration of the Beta-blocker metoprolol, J. Controlled Release, 4 (1986) 79. 7 0. Siddiqui, Y. Sun, J.C. Liu and Y.W. Chien, Facilitated transdermal transport of insulin, J. Pharm. Sci., 76 (1987) 341. 8 J.B. Sloan and K. Soltani, Iontophoresis in dermatology, J. Amer. Acad. Dermatol., 15 (1986) 671. 9 S. Duke-Elder, Iontophoresis, in: Duke-Elder (Ed.), The Foundations of Ophthalmology, Vol. VII, C.V. Mosby Co., St. Louis, MO, 1962, p. 507. 10 R.R. Burnette and T.M. Bagniefski, Influence of constant current iontophoresis on the impedance and passive Na+ permeability of excised nude mouse skin, J. Pharm. Sci., 77 (1988) 492. 11 M.J. Pikal and S. Shah, Transport of uncharged species by iontophoresis: electroosmotic flow, Amer. Assoc. Pharm. Sci., First National Meeting, 1986, p. 795. 12 S.G. Schultz, Basic Principles of Membrane Transport, Cambridge University Press, New York, NY, 1980, p. 21. 13 J.C. Keister and G.B. Kasting, Ionic mass transport through a homogeneous membrane in the presence of uniform electric field, J. Membrane Sci., 29 (1986) 155. 14 D.E. Goldman, Potential, impedance and rectification in membranes, J. Gen. Physiol., 26 (1943) 37. 15 R.J. Scheuplein, Mechanism of percutaneous absorption, J. Invest. Dermatol., 48 (1967) 79. 16 A.S. Michaels, S.K. Chandrasekaran and J.E. Shaw, Drug permeation through human skin: theory and in vitro experimental measurement, AICHE J., 21 (1975) 985. 17 A.H. Ghanem, H. Mahmoud, W.I. Higuchi, U.D. Rohr, S. Borsadia, P. Liu, J.L. Fox and W.R. Good, The effects of ethanol on the transport of /3-estradiol and other permeants in hairless mouse skin. II. A new quantitative approach, J. Controlled Release, 6 (1987) 75. 18 T. Masada, W.I. Higuchi, V. Srinivasan,U. Rohr, J. Fox, C.R.BehlandS. Pons,Examination of iontophoretic transport of ionic drugs across skin: Baseline studies with the four-electrode system, Int. J. Pharm., 49 (1989) 57. 19 H.A. McKenzie, pH and Buffers, in: R.M.C. Dawson (Ed.), Data for Biochemical Research, Oxford University Press, Oxford, 1969, p. 500. 20 H. Durrheim, G.L. Flynn, W.I. Higuchi and C.R. Behl, Permeation of hairless mouse skin. I. Experimental methods and comparison with human epidermal permeation by alkanols, J. Pharm. Sci., 69 (1980) 781. 21 J.R. Bond and B.W. Barry, Damaging effect of acetone on the permeability barrier of hairless mouse skin compared with that of human skin, Int. J. Pharm., 41 (1988) 91. 22 N.A. Gordon, Ph.D. Dissertation, University of Michigan, Ann Arbor, MI, 1980. 23 N.H. Bellantone, S. Rim, M.L. Francoeur and B. Rasadi, Enhanced percutaneous absorption via iontophoresis. I. Evaluation of an in vitro system and transport of model compounds, Int. J. Pharm., 30 (1986) 63. 1
320 24
25
26 27 28 29
30
R.R. Burnette
and B. Ongpipattanakul,
Characterization
of the pore transport
properties
and tissue alteration of excised human skin during iontophoresis, J. Pharm. Sci., 77 (1988) 132. L.P. Gangarosa, N. Park, C.A. Wiggins and J.M. Hill, Increased penetration of non-electrolytes into mouse skin during iontophoretic water transport (iontohydrokinesis). J. Pharm. Exp. Ther., 212 (1980) 377. P. Glikfeld, C. Cullander, P.S. Hinz and R.H. Guy, A new system for in iontophoresis, Pharm. Res., 5 (1988) 443.
vitro
studies of
N. Lakshminarayanaiah, Transport Phenomena in Membranes, Academic Press, New York, NY, 1969, p. 224. V. Srinivasan and W.I. Higuchi, A model for iontophoresis incorporating the effect of convective solvent flow, Int. J. Pharm., accepted for publication. V. Srinivasan, W.I. Higuchi and M. Su, Baseline studies with the four-electrode system. I. The effect of skin permeability increase and water transport on the flux of a model uncharged solute during iontophoresis, J. Controlled Release, 10 (1989) 157. R.R. Burnette and B. Ongpipattanakul, Characterization of the permselective excised human skin during iontophoresis, J. Pharm. Sci., 76 (1987) 765.
properties
of
305 in The Netherlands
BASELINE STUDIES ON IONTOPHORETIC TRANSPORT IN HAIRLESS MOUSE SKIN: THE EFFECT OF APPLIED VOLTAGE DROP AND pH ON THE IONTOPHORESIS OF A MODEL WEAK ELECTROLYTE
S.M. SIMS*
and WI.
Department (U.S.A.)
of Pharmaceutics,
(Received
HIGUCHI
March 29,1989;
301 Skaggs Hall, University
accepted
of Utah, Salt Lake City, UT 84112
in revised form September
12,1989)
Summary A physical
model for the iontophoretic
transport
of a weak electrolyte
across hairless
mouse
skin has been examined. The stratum corneum is modelled as parallel lipoidal and aqueous pore pathways for diffusion and is in series with a porous matrix representing the dermis-epidermis. It was assumed that only the undissociated species could penetrate the lipid phase while both charged and uncharged species could permeate the pore route. The applied electric field is assumed to influence only the charged species in the aqueous pore according to the Nernst-Planck theory. Experiments were done over a wide pH range using a four-electrode potentiostat system to control the voltage drop across the membrane. Butyric acid was chosen as the model weak electrolyte. Glucose was used to independently assess membrane damage and solvent flow effects. Permeability coefficients as a function of pH were determined both for butyric acid and glucose before, during and after iontophoresis. Experimental permeability coefficients semi-quantitatively followed the theoretical predictions.
Introduction Iontophoresis is a technique which may be employed to increase the transport of ions into or through skin by the application of an external electric field across the skin [ 11. It is a non-invasive technique which could potentially be used to administer a wide range of medicinal agents for local and systemic therapy [ 2-91. While there have been many clinically oriented iontophoresis studies reported in the literature, efforts needed to understand the fundamental physical chemistry involved in iontophoresis are just beginning [ l,lO-141. An understanding of the mechanism of charged and uncharged solute transport across skin, both with and without an applied electric field, is of fundamental impor*To whom correspondence
0376-7388/90/$03.50
should be addressed.
0 1990 Elsevier
Science
Publishers
B.V.
tance. A knowledge of the type of diffusion pathway, its limitations with regard to the size of the permeant molecule or ion, its charge status, and how each pathway can be manipulated by an external electric field, would aid both the understanding of iontophoresis, and the development of iontophoresis for drug delivery. By relating the iontophoretic fluxes to the properties of both membrane and drug, it may be possible to extrapolate these results beyond experimental systems to practical iontophoretic drug delivery devices. Several studies have provided a foundation for the understanding of the movement of molecules and ions in the stratum corneum, the barrier layer of skin. Scheuplein [ 151 addressed the issue that shunts such as hair follicles and sweat glands may be a route of solute diffusion. He suggested that shunts may be important for solute transport, especially during the transient period prior to steady state. However, transport through the intact stratum corneum (i.e., the transepidermal route) generally appears to be the dominant pathway of solute diffusion in stratum corneum. Various mechanisms and models have been proposed for the transport of solutes across the bulk stratum corneum. Michaels et al. [ 161 proposed a “brick and mortar” type model of the skin in which a solute could penetrate by only two routes: one required alternate passage through protein (corneocyte regions ) and lipid phases (intercellular lipid regions) and the other transit only through a continuous lipid phase. This model cannot accommodate ionic solutes which presumably diffuse only through an aqueous path. Recently, Ghanem et al. [ 171 have discussed a model in which the stratum corneum is treated as the diffusional barrier with parallel lipid and aqueous pore pathways for diffusion. This model provides a continuous aqueous route and could, therefore, permit ionic solutes to penetrate. In this paper the parallel pathway model is examined as a means of quantifying the iontophoretic transport of a weak electrolyte. In this model it is assumed that the electric field is able to enhance the movement of ions along the pore pathway and experimental data obtained with a weak electrolyte (butyric acid) are employed to test this aspect of the model. Theory
Physical model of the skin In this section a physical model is presented for the transport of a weak electrolyte across skin both with and without the influence of an applied electric field. The model is shown in Fig. 1. It consists of the stratum corneum in series with a porous matrix, the dermis-epidermis. The stratum corneum is comprised of parallel lipoidal and aqueous pore pathways for diffusion. It is assumed that (a) only non-ionized species are able to partition into and be transported in the lipoidal phase, and (b) both ionized and non-ionized species penetrate the aqueous route with essentially equal facility (i.e., with essentially the same permeability coefficient).
307 PHYSICAL
MODEL
OF THE
SKIN RECEIVER CHAMBER
LIPID POROUS
PORES
MATRIX LIPID
::t
-STRATUM
CORNEUM~OERMIS-EPIDERMIS-
Fig. 1. Schematic diagram of the physical model for the diffusion of charged and uncharged species across skin. The stratum corneum consists of parallel lipoidal and aqueous pore pathways which are in series with the porous dermis-epidermis layer.
For the steady state situation electrolyte:
we may write for the total flux (JT) of a weak
JT = APT AC
(1)
where PT is the total (or effective) permeability coefficient of a weak electrolyte; A is the area for diffusion and AC is the difference between the donor and receiver chamber permeant concentration. Also: JT = dQ/dt
(2)
where dQ/dt is the steady-state from donor to receiver chamber
slope of the amount of species transported as a function of time. Therefore:
PT = l/AAC dQ/dt
(3)
Now, for the model presented in Fig. 1, PT may be separated into contributions from the stratum corneum and the dermis-epidermis (resistance in a series): 1 -_-+6
1 PSC
1
(4)
PDE
where Psc is the permeability coefficient of stratum corneum and PDE is the permeability coefficient of the dermis-epidermis combination. Furthermore, for the situation of parallel lipoidal and aqueous pore pathways in the stratum corneum, we may write: Psc =X,P,
+pp
where PL is the permeability
(5) coefficient
for the lipoidal pathway,
Pp is the
308
permeability coefficient for the aqueous pore pathway, undissociated species. Combining eqns. (4) and (5), we have: 1 PT -Pp
1 +xsPL
and Xs the fraction
of
1 +PD,
Except for extremely lipophilic permeants, P DE is generally very large compared to PsC, i.e., the stratum corneum is usually the primary diffusional barrier. In this case, l/P nE can be neglected and eqn. (6 ) becomes: PT =Pp +xsPL
(7)
As will be seen in the present studies using butyric acid, PI)E is indeed much greater than PSC for hairless mouse skin, and therefore eqn. (7)is a good approximation of eqn. (6). Enhancement of ion transport via an applied electric potential While the total passive permeability coefficient may be described by eqn. (7)) the contribution to transport from an applied voltage drop across the skin requires further elaboration. We may assume that only the undissociated species enters the lipid domain whereas both ionized and non-ionized species may permeate the pore pathway. Further, we may assume the electric field influences only the charged species in the pore pathway. For this situation, we write the following equation for the permeability coefficient for a weak electrolyte solute in the presence of an electric field:
(8) where Ek,n is the enhancement factor for the ionized species due to the applied electric potential drop across the skin. The enhancement factor is defined as (1): Eion = Ji/Jp
(9)
where J, and Jp are the fluxes of the ionized species with and without the applied voltage, respectively. It can be shown [ 1,181 from the Nernst-Planck equation with the constant field assumption that: Ei,,,=-K;[l-exp(K)] where K is a dimensionless K= -,qF
Ay/RT
(16) constant
defined as: (11)
Here, Z, is the charge on species i, F the Faraday constant, Ay the electric potential drop, R the gas constant and T the absolute temperature. Equations (8-11)may be used to predict PT when PL, Pp and X, are known. Such calculations are shown for butyric acid in Fig. 2, where PT values are presented
309
PH
Fig. 2. Theoretical butyric acid permeability coefficient-pH profiles for passive diffusion (using eqn. 7) and for applied voltage drops of 250 and 500 mV (using eqn. 8). Pp= 1.8X lo-'cm/set; P,=1.3X10-6cm/sec;E,,,= 9.4; Ebo,= 18.7; pKaz4.84 at 37°C.
as a function of pH for passive diffusion and diffusion under applied voltages of 250 and 500 millivolts (mV) . Experimental
Materials Iontophoresis studies were conducted with radiolabeled [“Cl -1-butyric acid (specific activity 15 mCi/mmol) and t3H] -3-glucose (specific activity 13.5 Ci/ mmol) obtained from New England Nuclear Corporation (Boston, MA). Butyrate was chosen as a model anion because the fraction dissociated is pH dependent thus allowing assessment of the electric field effects on both the lipid and pore pathway permeabilities. Glucose was used as a model non-electrolyte to independently assess “membrane damage” and to obtain insight into solvent flow effects. Buffers ranged in pH from 2.4 to 9.5. Buffer composition [19] varied with pH and in all cases NaCl was the dominant salt species. Chemicals were reagent grade and used as received. All buffers had an ionic strength of 0.1 and were prepared in distilled, de-ionized water. Iontophoresis apparatus Skin permeabilities were measured using a four-electrode potentiostat system which maintains a constant voltage drop across a membrane mounted in a two-chamber diffusion cell. Figure 3 is a schematic diagram of the potentiostat system and its associated two-chamber diffusion cell. The potential drop across the Luggin capillary probes is maintained at the desired value by the
310 OUR-ELECTRODE POTENTIOSTAT
CE=COUNTER ELECTRODE RE =REFERENCE ELECTRODE
Fig. 3. Schematic
II/
REFERENCE
diagram of the four-electrode
potentiostat
system.
ELECTRODE
)RT
JACKET’
Fi. 4. Diagram of a half-cell for diffusion and its associated
Luggin capillary.
potentiostat which drives the appropriate current through the cell via the counter electrodes. This system has been described in detail by Srinivasan et al. [ 1]who also compared it to the conventional, two-electrode, constant current approach. Figure 4 shows one half of the two-chamber diffusion cell with its associated Luggin capillary and counter electrode. The cell is water-jacketed and main-
311
tained at 37 ’ C by circulating water (Brinkmann RM-6 model circulating water bath, American Scientific ). Each compartment has a volume of 5 ml. Iontophoresis experiments with hairless mouse skin Following cervical dislocation [ 201 of the hairless mouse (8-12 week old hairless mouse skin; SKH-HR-l), abdominal skin was excised and mounted between the two half cells with the stratum corneum facing the donor side. The two Luggin capillaries were positioned on each side with their tips very close to the membrane on either side (Fig. 4). The receiver chamber was filled with buffer of the pH to be tested. The donor solution contained the same buffer premixed with tracer levels of radiolabeled solute. The electrolyte solutions in the Luggin capillaries were the same as that in the receiver chamber. This was done to ensure that there was no interfacial potential drop between capillary solution and the cell solution, thus the applied voltage drop is the voltage drop across the membrane [ 11. Samples (100 ~1) were withdrawn from the donor side at the beginning and end of the experiment for analysis. The difference in the counts (dpm) between these two samples were always negligible. Both chambers were stirred by teflon rods with teflon propellers mounted at the bottom (Fig. 4). The teflon rods were connected to constant speed (150 r-pm) motors. The entire apparatus was maintained at 37 oC. Preliminary experiments indicated that the electric field may cause alterations in the membrane. To monitor and better understand the changes in the intrinsic transport properties caused by the electric field, each experiment was composed of three stages (Fig. 5 ) . First a passive permeability coefficient was
TIME
WR)
Fig. 5. Typical permeation
profile for a three-stage
experiment:
(m) butyric acid; ( 0 ) glucose.
312
determined in which PT = PT,O. In the second stage an electric field was applied and the total permeability coefficient due to this potential drop was determined, i.e., PT = PT,dyl Finally, a second passive permeability coefficient was determined, i.e., PT = PT,O.. A more precise description of each stage follows. Stage I: This stage involves the measurement of the initial passive permeability of total butyrate and/or glucose through hairless mouse skin. After assembly, the experimental system was allowed to equilibrate for six to seven hours and reach steady state. Five 1 ml samples were withdrawn at 30 min. intervals from the receiver chamber. Each sample was mixed with 10 ml OptiFluor (Packard Instrument Co. ) scintillation fluid and counted on a Beckman Liquid Scintillation Counter (Model LS-7500). The samples were replaced with fresh buffer. The initial passive permeability coefficient, PT,O, was calculated from eqn. (3 ). Stage II: During this stage, which lasted one hour, the iontophoretic permeability coefficient, PT,dlv,was determined for the permeant at an applied voltage drop, Av, across the skin. After the final sampling in Stage I, the cells were connected to the potentiostat (JAS Instrumental Systems, Inc., Salt Lake City, UT) and a fixed voltage drop was applied across the skin (cathode on donor side, anode on receiver side). Four 1 ml samples were taken at 15 minute intervals from the receiver side and analyzed as in Stage I. Samples were replaced with fresh buffer. PT,dWwas calculated using the last three points of Stage II (Fig. 5) from eqn. (3 ) as before. The cell current was monitored continuously and the applied voltage was confirmed via a voltmeter (Beckman Industrial Model 310). Stage III: At the end of Stage II the receiver chamber was flushed three times, then refilled with fresh buffer. After one hour, the passive permeability coefficient P,,,O. was determined as before. The total time for each experiment was, at most, 11-12 hours. Passive experiments at each pH indicated that the barrier properties of the skin remained intact during the entire experiment. Other authors [21,22] have also shown that hairless mouse skin retains its barrier function in aqueous buffer over this time range. At the end of each experiment the pH in both chambers was checked. The pH varied only 0.2-0.3 pH units in extreme cases. Results Total butyrate transport
Figure 5 presents the result of a typical three stage, dual label, permeation experiment. The best fit straight line (neglecting the first point in each stage as being part of lagtime) for each stage was used to calculate the total permeability coefficient, PT. Table 1 lists the experimentally determined permeability coefficients for total butyrate for each stage of the experiment as a function of pH and applied voltage drop. In Table 2 these data are converted to the total
313 TABLE
1
Experimental
permeability No. of expts.
PH
coefficients
for butyrate with hairless mouse skin
Permeability
P T.0
coefficient”
( x 1O-7 cm/set) P T&
P T,O'
Ar//= 250 mV 2.4
13+4.5
12kl.O
12+2.0
5 3 3 3
7.6k4.5 5.9 + 4.0 0.2 f 0.8 0.2 + 0.0 0.4kO.3
14k8.0 8.2 ? 2.7 1.2 * 1.2 1.6kO.7 2.7 & 1.8
14i3.0 8.2 ? 2.2 0.7 + 0.3 0.5 I!Y0.2 0.7kO.6
4 4
14+3.0 5.7k4.1
3 3 4
6.6f3.5 2.8kO.4 0.3 +0.1 0.2LO.l
32f13 14k5.0 17 ? 6.0
21+3.0 12 * 5.0 12 * 7.0 3.9 * 0.7
3 4
3.6 4.8 6.0 7.5 9.0 Ay/= 500 mV 2.4 3.6 4.8 5.5 6.5 7.5
5 6
9.5
0.5 I!z0.3
llk4.0 8.4 & 4.8 2.5kO.6 30*10
1.9 !C 1.0 0.4kO.l 2.6? 1.0
“Mean ?I S.D.
butyrate effective enhancement factors, E, and EP, as defined by the following equations: E, = PT,&%,o
(12)
E, =
(13)
pT,dy//pT,O’
E, represents the enhancement of butyric acid under the influence of the voltage, dy/, over the initial passive value (PT,o), while E2 is the enhancement based on PT,o, as the “control”. Also shown in Table 2 is the ratio El/E2 which is a measure of the membrane alteration (probed by using butyric acid as the permeant) caused by the applied electric field. Glucose transport Table 3 lists the experimentally determined total permeability coefficients (PO,G,P&G, Po,Gs) for glucose for hairless mouse skin at various pH values. Table 4 shows the enhancement value for glucose, EG1 and EG2, which are defined as follows: EGI
= P&,G/PO,G
(14)
EGZ
= Pdyl,G/PO,G’
(15)
The ratio EGl/EGz (Table 4) is a measure of the membrane alteration caused
314 TABLE
2
Enhancement
factors for butyric acid with hairless mouse skin Enhancement
PH
&= 250 mV 2.4 3.6 4.8 6.0 7.5 9.0 Ay/= 500 mV 2.4
factor”
1.1 kO.5 2.0 ? 0.5 l.B? 1.0 6.0 + 5.2 7.0 IfI2.6
l.OLO.1 1.0+0.3 1.1 i-o.4 1.65 1.0 3.6kO.9
l.OkO.6 2.3 ?I 1.0 1.8 + 1.0 3.9+ 1.9 2.OkO.5
9.0 & 5.2
1.4i-7.5
1.6kO.6
2.4t0.7
1.5 * 0.5
1.6kO.3
3.6 4.8 5.5
3.2 i- 1.4 2.8kO.7 4.0* 1.3
1.4 I! 0.8 1.8t0.8 2.8 I!Z0.9
2.4i0.8 1.7kO.5 1.4iO.l
6.5 7.5 9.5
22+7.1 14? 6.9 54*19
4.4+ 1.7 6.4+ 1.7 12+3.7
6.2 + 0.9 2.2 5 0.8 4.5i2.6
“Mean + S.D.
TABLE
3
Experimental PH
AMY=250 mV 3.6 4.8 6.0 7.5 9.0 AI,Y= 500 mV 3.6 4.8 5.5 6.5 7.5 9.5 “Mean + S.D.
permeability No. of expts.
coefficients
for glucose with hairless mouse skin
Experimental
permeability
coefficient
( x lop8 cm/sec)a
P 0s
P&
PO,,:,
6 3 4 4 6
4.5 + 2.2 2.1* 1.3 0.9 + 0.5 2.2 i 0.4 2.5k2.1
301?118 6.Oi-5.5
24+25 4.Ok3.0 1.5 i 1.1 3.7 + 1.7 5.2k3.1
2 3 4 4
4.9 i 3.8 4.0 I!c1.9 1.6+ 1.3 1.1 kO.4
3 4
1.2 I!I0.3 5.4k3.1
0.6 + 0.3 1.5 I! 0.6 1.8? 1.8
25i16 16? 7.0 5.6k5.1 2.5 z?z1.2 2.5kO.l 6.2 + 1.4
lOk9.0 9.8i2.2 3.8t 3.2 1.9 I! 1.0 4.9ir2.6 3.1 z!z1.3
315 TABLE
4
Enhancement
factors for glucose with hairless mouse skin Enhancement
factor’
E Gl
E G2
EG@GP
3.6 4.8
7.0 + 2.2 2.3 k 1.1
2.0 + 1.0 1.6kO.3
4.9f4.1 1.5kO.7
6.0 7.5 9.0
0.8 + 0.6 0.7 + 0.2 0.7f0.4
0.6kO.3 1.1 dIo.9 1.4 + 1.7
1.620.8 1.1 * 0.7 1.4 f 1.2
5.6? 1.2 4.3 f 1.0 3.62 1.3 2.6 I!Z1.6
3.0 + 0.9 1.6kO.5 1.4kO.5 1.5kO.4
1.9kO.l 2.9 + 1.3 2.8i 1.0 1.9 + 1.3
2.lkO.5 1.3 z!z0.7
0.6kO.2 0.2 * 0.1
4.3f3.1 5.6t2.0
PH
Av/= 250 mV
Ay/= 500 mV 3.6 4.8 5.5 6.5 7.5 9.5 “Mean 2 S.D.
by the electric field. Since it was assumed that glucose is transported only along the pore pathways (for all pH conditions), this ratio was expected to be a general measure of the irreversible alteration in the pore pathway caused by the applied voltage. The ratio Eol/EGZ was, therefore, expected to be different from the ratio EJ&, except at high pH conditions where the butyrate ion transport was expected to be much more important than the non-ionized butyric acid transport across skin. Discussion Comparison of results with theoretical prediction Considering a meaningful approach to analyze the experimental results with the model (i.e., eqns. 8-ll),one is faced with the problem of the significant membrane alterations caused by the applied voltage in most instances (Tables 2 and 4). An important question here was whether PT,o, PT,Oz,or some average should be used as the reference for calculating the enhancement (e.g., E, versus E2). In an attempt to gain some insight on the matter, the passive glucose permeability coefficients were determined as a function of the time of application of the voltage. Figure 6 shows how “membrane damage” may vary with the time of application of the voltage over a 60 minute time period. Although the variabilities are seen to be large, these results strongly suggest that membrane damage in hairless mouse skin occurs early and the effect tends
316
a I
B
01~ 1
T-rllil 20
0
TIME
OF
40
APPLIED
60
80
VOLTAGEtMIN.)
Fig. 6. Glucose EG,/EG, ratio as a function of the time voltage was applied. (0 ) pH 7.5,250 mV; ( n ) pH 7.5,500 mV; (A ) pH 9.5,500 mV.
2
4
6
8
IO
PH
Fig. 7. Experimental butyric acid PO, ( n ) and PAY/(0 ) as a function of pH. Solid lines represent the theoretical prediction. PP = 4.5 x 10-s cm/set; PL= 1.2 X 10m6 cm/set; I& =9.4; AY= 250 mV.
to plateau between 30 and 60 minutes. Because of these data, it was judged that PT,O.would be better than PT,O (or the average of the two) in comparing the experimental results with theory. Accordingly, the PT,O,values were used to deduce the PL and the Pp values for use in eqns. (7) and (8). As can be seen from Figs. 7 and 8, there is semi-quantitative agreement between experiment and theory. At low pH, the theory predicts that transport would involve predominantly the undissociated species diffusing along the lipid pathway, and therefore, there should be little or no iontophoretic enhancement. Here, there is good agreement between experiment and theory, and the
317
2
4
6
8
IO
Fig. 8. Experimental butyric acid P,. (m) and PAW(0 ) values as a function of pH. Solid lines represent the theoretical prediction. PP=4.0X 10e8 cm/set; PL= 1.2 X 10d6 cm/set; I&,,= 18.7; dv/= 500 mV.
experimental permeability coefficients approach the same asymptote with or without the applied voltage. At high pH ( > 7.4)) the pore pathway should dominate and the differences between PT,d,,,and PT,Osbecome large. Under high pH conditions, the experimental and the theoretical enhancement values are seen to agree within about a factor two or better. Membrane alteration due to the applied voltage Membrane damage has been defined as an increase in the passive permeability of a membrane [ 11. Alteration in the membrane barrier properties has been suggested by many authors. Bellantone et al. [ 231 saw a varying increase in benzoic acid flux though hairless mouse skin after termination of an applied current. Burnette and Bagniefski [lo] demonstrated that the impedance of hairless mouse skin was generally lower and Na+ flux was higher after iontophoresis. Similarly, human skin was shown to have a lower impedance and higher 3Hz0 flux after iontophoresis [24]. These studies indicate that skin undergoes some change due to iontophoresis such that the passive permeability of the skin is increased. Tables 2 and 4 present the E1/ES and the EG1/EGPratios which represent membrane damage when butyric acid and glucose, respectively, are used as the probes. The glucose probes the aqueous pore pathway over the entire pH range while butyric acid probes both the lipid and pore pathways, depending upon the pH (eqns. 7 and 8). Tables 2 and 4 show that there is generally some significant membrane alteration at both 250 and 500 mV. Because of the large variabilities in the data, it is difficult to judge (a) whether there is less damage in the lipid pathway over the pore pathway, and (b) whether the damage in
318
the pore pathway is dependent on pH. The data do indicate, however, that at high pH (especially at 500 mV) the EG1/EG2and E,/E, results correlate well; this is consistent with the expectation (eqn. 8) that, at high pH, the pore pathway will dominate in butyric acid permeation. The data also indicate that at high pH and 500 mV, membrane damage is greater than at low pH and lower voltages. This latter thought is consistent with the high baseline passive value seen in Fig. 8. Solvent flow effects due to the applied potential drop Also of importance in these experiments is the possible existence of solvent flow. Solvent flow has been suggested in several recent studies. Gangarosa et al. [ 251 demonstrated that the permeation of non-electrolytes through hairless mouse skin can be enhanced by iontophoresis due to convective flow. Pikal and Shah [ 111 suggested bulk fluid flow by electro-osmosis as the mechanism by which iontophoretic enhancement of the permeability of uncharged species occurs. Glikfeld et al. [26] reported measurable volume changes in both diffusion cell chambers when a current of 0.5 mA was applied to the system. Charged species may also be affected by solvent flow. Solvent flow is generally believed to occur in the direction of counter ion flow [ 11,251. In the present study, where the cathode is placed in the donor chamber and the anode in the receiver chamber, it was thought that glucose transport under the influence of the electric field would be retarded relative to passive glucose transport and that such information could be used to assess solvent flow during iontophoresis. Unfortunately, the data in Table 4 show considerable variability, such that a rigorous interpretation of the data is not possible. It would be instructive, however, to attempt an upper limit estimation of the solvent flow effects upon the butyrate ion transport, employing the data in Table 4. Incorporation of solvent flow effects in the Nernst-Planck equation [ 271 results in an equation similar to eqn. (11) but with a correction term for solvent flow [ 281. If an Eo2 value of 0.2 (see Table 4) is used to estimate solvent flow, one finds that the correction for the butyrate ion enhancement is lo-20%. This estimate suggests that solvent flow effects are relatively small and consistent with the views of other investigators [ 11,29,30]. Conclusions The experimentally determined pH profiles for the iontophoretic transport through hairless mouse skin of butyric acid agree semi-quantitatively with the proposed parallel lipid-aqueous pore pathway model for the transport of ions across skin. The effects of membrane damage were taken into account in the data analysis and the influence of solvent flow was assessed.
319
References V. Srinivasan, S.M. Sims, W.I. Higuchi, C.R. Behl and S. Pans, Iontophoretic transport of drugs: a constant voltage approach, in: J. Kost (Ed.), Pulsed and Self-Regulated Drug Delivery, CRC Press, Boca Raton, FL., 1990. 2 P. Tyle, Iontophoretic devices for drug delivery, Pharm. Res., 3 (1986) 318. 3 A.K. Banga and Y.W. Chien, Iontophoretic delivery of drugs: fundamentals, developments and biomedical applications, J. Controlled Release, 7 (1988) 1. 4 M. Comeau, R.B. Brummett and J. Vernon, Local anaesthesia of the ear by iontophoresis, Arch. Otolaryngol., 98 (1973) 114. 5 L.P. Gangarosa, Iontophoresis for surface local anaesthesia, J. Amer. Dent. Assoc., 88 (1974) 125. 6 K. Okabe, H. Yamaguchi and Y. Kawai, New iontophoretic transdermal administration of the Beta-blocker metoprolol, J. Controlled Release, 4 (1986) 79. 7 0. Siddiqui, Y. Sun, J.C. Liu and Y.W. Chien, Facilitated transdermal transport of insulin, J. Pharm. Sci., 76 (1987) 341. 8 J.B. Sloan and K. Soltani, Iontophoresis in dermatology, J. Amer. Acad. Dermatol., 15 (1986) 671. 9 S. Duke-Elder, Iontophoresis, in: Duke-Elder (Ed.), The Foundations of Ophthalmology, Vol. VII, C.V. Mosby Co., St. Louis, MO, 1962, p. 507. 10 R.R. Burnette and T.M. Bagniefski, Influence of constant current iontophoresis on the impedance and passive Na+ permeability of excised nude mouse skin, J. Pharm. Sci., 77 (1988) 492. 11 M.J. Pikal and S. Shah, Transport of uncharged species by iontophoresis: electroosmotic flow, Amer. Assoc. Pharm. Sci., First National Meeting, 1986, p. 795. 12 S.G. Schultz, Basic Principles of Membrane Transport, Cambridge University Press, New York, NY, 1980, p. 21. 13 J.C. Keister and G.B. Kasting, Ionic mass transport through a homogeneous membrane in the presence of uniform electric field, J. Membrane Sci., 29 (1986) 155. 14 D.E. Goldman, Potential, impedance and rectification in membranes, J. Gen. Physiol., 26 (1943) 37. 15 R.J. Scheuplein, Mechanism of percutaneous absorption, J. Invest. Dermatol., 48 (1967) 79. 16 A.S. Michaels, S.K. Chandrasekaran and J.E. Shaw, Drug permeation through human skin: theory and in vitro experimental measurement, AICHE J., 21 (1975) 985. 17 A.H. Ghanem, H. Mahmoud, W.I. Higuchi, U.D. Rohr, S. Borsadia, P. Liu, J.L. Fox and W.R. Good, The effects of ethanol on the transport of /3-estradiol and other permeants in hairless mouse skin. II. A new quantitative approach, J. Controlled Release, 6 (1987) 75. 18 T. Masada, W.I. Higuchi, V. Srinivasan,U. Rohr, J. Fox, C.R.BehlandS. Pons,Examination of iontophoretic transport of ionic drugs across skin: Baseline studies with the four-electrode system, Int. J. Pharm., 49 (1989) 57. 19 H.A. McKenzie, pH and Buffers, in: R.M.C. Dawson (Ed.), Data for Biochemical Research, Oxford University Press, Oxford, 1969, p. 500. 20 H. Durrheim, G.L. Flynn, W.I. Higuchi and C.R. Behl, Permeation of hairless mouse skin. I. Experimental methods and comparison with human epidermal permeation by alkanols, J. Pharm. Sci., 69 (1980) 781. 21 J.R. Bond and B.W. Barry, Damaging effect of acetone on the permeability barrier of hairless mouse skin compared with that of human skin, Int. J. Pharm., 41 (1988) 91. 22 N.A. Gordon, Ph.D. Dissertation, University of Michigan, Ann Arbor, MI, 1980. 23 N.H. Bellantone, S. Rim, M.L. Francoeur and B. Rasadi, Enhanced percutaneous absorption via iontophoresis. I. Evaluation of an in vitro system and transport of model compounds, Int. J. Pharm., 30 (1986) 63. 1
320 24
25
26 27 28 29
30
R.R. Burnette
and B. Ongpipattanakul,
Characterization
of the pore transport
properties
and tissue alteration of excised human skin during iontophoresis, J. Pharm. Sci., 77 (1988) 132. L.P. Gangarosa, N. Park, C.A. Wiggins and J.M. Hill, Increased penetration of non-electrolytes into mouse skin during iontophoretic water transport (iontohydrokinesis). J. Pharm. Exp. Ther., 212 (1980) 377. P. Glikfeld, C. Cullander, P.S. Hinz and R.H. Guy, A new system for in iontophoresis, Pharm. Res., 5 (1988) 443.
vitro
studies of
N. Lakshminarayanaiah, Transport Phenomena in Membranes, Academic Press, New York, NY, 1969, p. 224. V. Srinivasan and W.I. Higuchi, A model for iontophoresis incorporating the effect of convective solvent flow, Int. J. Pharm., accepted for publication. V. Srinivasan, W.I. Higuchi and M. Su, Baseline studies with the four-electrode system. I. The effect of skin permeability increase and water transport on the flux of a model uncharged solute during iontophoresis, J. Controlled Release, 10 (1989) 157. R.R. Burnette and B. Ongpipattanakul, Characterization of the permselective excised human skin during iontophoresis, J. Pharm. Sci., 76 (1987) 765.
properties
of