Frequency and thermal effects on the enhancement of transdermal transport by sonophoresis

Journal of Controlled Release 88 (2003) 85–94 www.elsevier.com / locate / jconrel

Frequency and thermal effects on the enhancement of transdermal transport by sonophoresis a a ˜ Delgado-Charro a , Gustavo Merino , Yogeshvar N. Kalia , M. Begona b,c Russell O. Potts , Richard H. Guy a , * a

School of Pharmacy, University of Geneva, 30 quai E. Ansermet, CH-1211 Geneva 4, Switzerland b Centre International de Recherche et d’ Enseignement, ‘ Pharmapeptides’ Archamps, France c Cygnus, Inc., Redwood City, CA, USA Received 23 July 2002; accepted 20 November 2002

Abstract The aims of this work were: (i) to examine the role of ultrasound (US) frequency and intensity on the transport of glucose and mannitol across porcine skin in vitro, (ii) to quantify the energy delivered to the skin during application of low-frequency sonophoresis, and (iii) to ‘deconvolute’ the thermal effect, induced by US application to the skin, to the enhanced permeability of the cutaneous barrier. Low- (20 kHz) and high-frequency (10 MHz) sonophoresis were first compared. Only low frequency US resulted in significantly increased permeation. Low-frequency, US-induced enhancement of mannitol transport was symmetric; that is, mannitol flux was the same when ‘delivered’ or ‘extracted’ from a donor solution (in both cases, the US probe was present on the surface side of the skin). Calorimetry was used to quantify the US energy delivered by the sonicator. Subsequently, the US-enhanced transdermal transport of mannitol, during which a significant (and US intensity-dependent) temperature increase occurred, was compared to that provoked, in the absence of sonophoresis, by a comparable thermal effect. Only 25% of this enhancement was attributable to the increased temperature induced by US. It follows that another mechanism, most probably cavitation, is principally responsible for the lowered skin barrier function observed.  2002 Elsevier Science B.V. All rights reserved. Keywords: Sonophoresis; Stratum corneum; Skin; Transdermal delivery

1. Introduction The efficiency of the stratum corneum (SC) as a barrier to molecular transport has led to the development of a variety of non-invasive technologies either *Corresponding author. Tel.: 133-4-5031-5021; fax: 133-45095-2832. E-mail address: [email protected] (R.H. Guy).

to increase the potential applications of transdermal drug delivery or to enable the facile extraction of molecules for monitoring and / or diagnosis. Ultrasound has been shown to enhance the transdermal permeability of various drugs, including insulin [1,2] and, recently, the use of ‘reverse sonophoresis’ for glucose monitoring has also received attention [3]. However, the mechanisms of action of sonophoresis remain incompletely defined, and have been various-

0168-3659 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved. doi:10.1016 / S0168-3659(02)00464-9

86

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

ly suggested [1] to involve either local heating, stratum corneum lipid disorganization, and / or cavitation, i.e. the formation and oscillation of small gaseous inclusions. Among the important unknowns are the optimum US frequency and pressure amplitude necessary to maximize molecular transport across the skin. The thermal effect of US on the skin results from the transfer and conversion of mechanical energy generated by the vibration of a piezoelectric crystal in the sonophoresis probe [4]. Absorption by the skin of this energy causes a temperature increase, which is directly related to the intensity of the sound wave [5]. Prediction of the actual temperature increase produced by a particular sonophoretic profile is difficult, however, without a precise knowledge of the acoustic absorption coefficients, and of the conduction and convection properties, of the tissues involved. Furthermore, experimentally, there have been few studies quantifying US-induced temperature increases in the skin (such as considering the manner in which thermal effects depend on the distance of the transducer from the tissue surface [5]). Recently, most investigations have focused upon the use of US at a frequency of about 20 kHz, at a 10% duty cycle (0.1 on, 0.9 off), for periods from minutes to a few hours [1,6]. While the results from this work implicate cavitational effects as a principal mechanism, the role of the accompanying thermal effect has not been deduced. Given that skin permeability can increase significantly with temperature (for example, the absorption of estradiol doubled when the temperature was increased by 10 8C [7]), and that phase transitions of the intercellular lipids of the stratum corneum (SC) can occur at temperatures close to physiological [8,9], it is clearly possible that thermal changes can contribute to sonophoreticallyenhanced transdermal transport. The first objective of this work was to examine in vitro the extent to which high- and low-frequency sonophoresis enhanced glucose transport across the skin at different applied intensities. Glucose was chosen because of the presently intense interest in developing noninvasive technologies for monitoring blood sugar levels in diabetics [3,10]. Due to the fact that enzymatic biotransformation of glucosylceramides by b-glucocerebrosidase [11] leads to an endogenous glucose ‘depot’ within the skin and

because exogenous glucose can also be metabolized during its transdermal passage [10], an unequivocal interpretation of sonophoretically-enhanced glucose transport is complicated. For this reason, more detailed experiments were therefore carried out with mannitol, a sugar isomeric with glucose, but one not found endogenously in the skin and one not subject to enzyme attack. The second objective of the research reported here was to ‘decouple’ the thermal contribution of US-enhanced skin permeability from other mechanisms of action. As a component of this effort, it was necessary furthermore to evaluate quantitatively the energy delivered to the skin during sonophoresis.

2. Material and methods

2.1. Materials Glucose, mannitol, glucose dehydrogenase (from Cryptococcus uniguttulatus) and the constituents of phosphate-buffered saline were purchased from Sigma–Aldrich (Saint Quentin Fallavier, France). NADP was obtained from ICN Pharmaceuticals (Orsay, France). [ 14 C]Mannitol (specific activity 56.0 mCi / mmol) was from Amersham Pharmacia Biotech (Orsay, France). De-ionized water (resistivity.18 MV cm) was used to prepare all solutions.

2.2. Skin Full-thickness skin from porcine ears (S.O.D.E.X.A., Annecy, France) was used in all experiments. The tissue was excised from the outer part of the ear and separated from the underlying cartilage with a scalpel. The skin was cleaned under cold running water, cut into pieces of |11 cm 2 , and then stored frozen at 220 8C until use. The storage period was never longer than 1 month and was typically much shorter. Just prior to an experiment, the skin was thawed at room temperature over a 1-h period.

2.3. Equipment 2.3.1. High frequency sonophoresis High frequency US was produced from a signal generator (Hewlett-Packard HP 8116A, Palo Alto,

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

CA, USA), connected to an oscilloscope (Chauvin Annoux, CA 902, 20 MHz, Paris), a power amplifier (325 LA, ENI Co, Santa Clara, CA, USA), and a 10 MHz transducer (Sofranel, Zurich, Switzerland). The US was transmitted to the skin in a pulsed mode, i.e. a 50% on, 50% off duty cycle at intensities of 0.2 and 2 W/ cm 2 .

2.3.2. Low frequency sonophoresis Low frequency US at 20 kHz was delivered from a commercially available sonicator (VCX 400, Sonics and Materials, Danbury, USA). Signal intensities of 10, 25 and 70%, for the glucose experiments, and 70%, in the case of mannitol experiments, were applied every second using a duty cycle of 0.1 s on, 0.9 s off, to minimize excessive temperature increases [1]. 2.4. Experimental protocol 2.4.1. Diffusion cell Transport experiments were carried out in custommade, vertical, glass diffusion cells. The skin membrane separated the upper and low compartments, with the SC always facing the upper chamber in which the US probe was positioned (Fig. 1). A

Fig. 1. Schematic diagram of the experimental set-up used for US treatment of the skin. Transport of solutes was studied both from the lower chamber to the upper (‘extraction’) and in the opposite sense (‘delivery’). The active cross-sectional area was 9.4 cm 2 . The ultrasound probe was positioned at a distance of |0.5 cm from the stratum corneum (SC) surface.

87

Teflon sheet was placed at the bottom of the lower compartment to absorb ultrasound and thereby reduce multiple reflections in the diffusion cell [7]. The distance between the US transducer and the surface of the skin was approximately 0.5 cm.

2.4.2. Glucose experiments High- and low-frequency ultrasound were applied at operating frequencies of 10 MHz and 20 kHz, respectively. The upper chamber contained 11 ml of phosphate-buffered saline (pH 7.4). The lower, subdermal chamber (i.e. dermal side of the skin) contained 38 ml of either 5 or 25 mM glucose and was stirred continuously throughout the experiment. US was applied for 2 h, after which the upper chamber, ‘receptor’, phase was removed and analyzed for ‘extracted’ glucose. An equal volume of fresh buffer solution was then introduced into the upper compartment, and passive glucose extraction across the US-pretreated skin was measured during a further 2 h, to assess the degree to which sonophoresis had perturbed skin barrier function. For comparative and control proposes, glucose transport was also measured: (i) under identical experimental conditions but without US application, and (ii) in exactly the same way with US application but with no exogenous glucose introduced into the lower compartment (i.e. to determine the contribution of a glucose ‘depot’ within the skin itself to the USinduced extraction flux). 2.4.3. Mannitol experiments Only low-frequency US was applied (20 kHz). However, transport of mannitol was followed in both directions: (i) ‘extraction’ from the lower chamber into the upper, and (ii) ‘delivery’ in the opposite sense (upper to lower). In the former case, as for glucose, the upper chamber was filled with phosphate-buffered saline (pH 7.4), while the lower compartment contained either 5 or 25 mM mannitol solution (spiked with 5 mCi D-[1- 14 C]mannitol). In the latter case, mannitol was incorporated into the upper compartment solution and its ‘delivery’ across the skin was followed. The ‘donor’ solution comprised 11 ml of either 5 or 25 mM mannitol (spiked with 5 mCi D-[1- 14 C]mannitol). The subdermal chamber was filled with PBS at pH 7.4. In both ‘extraction’ and ‘delivery’ experiments,

88

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

the appropriate controls were carried out: (a) passive transport without US, and (b) passive transport for 2 h following the US treatment. For all transport experiments (glucose and mannitol), a minimum of six replicates was carried out.

2.5. Analysis The amount of glucose transported across the skin was determined using the glucose dehydrogenaseNADP coupled assay [12,13]. The reaction between the enzyme, cofactor and b-glucose produced NADPH, which was quantitatively assayed using fluorescence spectroscopy. Mannitol transport was deduced from the 14 Cradioactivity quantified in the ‘receptor’ chamber (upper or lower for ‘extraction’ and ‘delivery’, respectively) by liquid scintillation counting. Each sample was mixed with 10 ml of scintillation cocktail (Ultima Gold XR, Packard Bioscience Co, Rungis, France), and counted in a liquid scintillation counter (Model LS 6500, Beckman Instruments Inc., Gagny, France).

2.6. Quantification of US intensity The US intensity, or the power delivered, by the commercially available, low frequency sonicator is typically reported as a percentage of its maximal efficiency. To relate this efficiency to units of power (W/ cm 2 ), a calorimetric method was employed. In this approach, the sonicator probe was immersed in 500 ml of deionized water under adiabatic conditions; that is, this volume was held within an insulated, jacketed system. The sonicator was then activated for a period of 8 h at a 10% duty cycle (0.1 on, 0.9 off), corresponding to 48 min of total US application, as a function of various intensity settings in the range of 5–70% of the apparatus’ maximal efficiency. The temperature increase (DT ) in the water volume was recorded with a digital thermometer allowing the US intensity (I) delivered to be calculated from the following equation: mCp I 5 ]](DT ) A where m is the mass of water of specific heat Cp , and

A is the ultrasound probe area. These measurements were made in triplicate.

2.7. Contribution of thermal effects to USenhanced transport A passive control experiment, without US or the temperature increase induced by its application, was first carried out. The diffusion cell was assembled with the skin separating the upper and lower compartments. The former was filled with a 25 mM aqueous solution of radiolabelled mannitol in phosphate-buffered saline at pH 7.4 (PBS), while the latter was charged with PBS alone, and was magnetically stirred. Passive diffusion was allowed to proceed for 2 h, at the end of which an aliquot of the receptor solution was assayed as before, so as to permit the amount of mannitol transported to be deduced. The next step was to determine the manner in which US application caused the temperature of the donor solution and skin to change over the 2-h period of experiment. With this in mind, a thermistor was introduced into the upper compartment and positioned in the mannitol solution (as described above, but without 14 C-radioactivity) in close proximity to the skin surface. The receptor chamber was again filled with PBS. The US probe was then activated at 20 kHz, at a duty cycle of 0.1 on, 0.9 off, at each of three nominal intensities (10, 25 and 70%) for a period of 2 h. During this time, the temperature detected by the thermistor was recorded every 5 min for 30 min, and then somewhat less frequently for the next 90 min. Given that the largest temperature increase was induced, as expected, at the nominal intensity of 70%, it was decided to measure mannitol transport across the skin under the influence of sonophoresis applied under these conditions. The composition of the donor and receptor phases and the method of mannitol detection were as for the passive experiment. To separate out the thermal effect of US, in a final experiment, the temperature change over 2 h induced by US at the nominal intensity of 70% was reproduced by setting the diffusion cell on a heaterstirrer plate and manually adjusting the heating control in a manner appropriate to mimic the US

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

effect. Subsequently, mannitol delivery during this 2-h heating protocol was determined in exactly the same way as described before. All experiments, whether those measuring temperature changes, or those determining mannitol delivery, were repeated at least five times. Differences between the amounts of mannitol delivered across the skin in 2 h following US application, by increased temperature alone, and via passive diffusion were evaluated statistically by ANOVA.

89

cantly different from the amount of endogenous sugar released from the skin during a 2-h period. As has been previously reported in the literature, therefore, the enhancement induced by high-frequency sonophoresis is at best modest and sometimes imperceptible [1,14].

3.2. Low frequency glucose sonophoresis

The effect of high frequency US, both during and immediately post-treatment, on the outward transport of glucose across the skin is shown in Fig. 2. At the intensities used, no significant impact of sonophoresis was observed. Indeed, the glucose detected in the upper chamber was never signifi-

In contrast, and in agreement with previous reports [3,15,16], US applied at low-frequency was able to provoke an increase in glucose ‘extraction’ across the skin (Fig. 3). At the highest intensity considered (70%), the amount of glucose transported both during, and subsequent to, a 2-h period of treatment at 20 kHz was significantly (P,0.01) elevated (compare the y-axis scales in Figs. 2 and 3). At lower intensities (10%, 25%), however, the ‘extraction’ of glucose, both during and after sonophoresis, was indistinguishable from the release of the endogenous sugar. The fact that only the highest intensity was able to provoke glucose trans-

Fig. 2. The effect of high-frequency US on the outward transport of glucose across the skin. The impact of 2 h of US at 10 MHz applied at either 0.2 or 2 W/ cm 2 (open bars) is compared to the amounts diffusing post-sonophoresis in 2 h (filled bars), when a 5 mM solution was present in the lower compartment. As a control, the ‘release’ of glucose from the skin when US was applied with no exogenous sugar present subdermally (hatched bars) is also shown.

Fig. 3. The effect of low frequency US on the outward transport of glucose across the skin. Glucose transport during 2 h of US at 20 kHz, applied at different intensities, is compared to the amounts diffusing post-sonophoresis in 2 h when either a 5 or 25 mM solution was present in the lower compartment. As controls, (a) the passive transport of glucose without US, and (b) the ‘release’ of glucose from the skin when US was applied with no exogenous sugar present subdermally, are also shown.

3. Results and discussion

3.1. High frequency glucose sonophoresis

90

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

port implies that cavitational effects, i.e. the oscillation and collapse of bubbles in close proximity to the skin barrier, may be involved in the mechanism of enhancement. It has been shown that such cavitation is inversely proportional to US frequency and directly correlated with intensity [2]. As an example, in another study examining the US-enhanced extraction of glucose, it was shown that the skin’s electrical conductivity (a useful marker of skin barrier function) was most dramatically increased by low frequency US applied at its higher intensities (e.g. 14 W cm 22 ) [6]. More recently, it has been shown that this enhancement of skin conductivity correlates well with the amplitude of broadband noise, again suggesting the important role of transient cavitation in low frequency sonophoresis [17]. The data in Fig. 3 reveal two additional features worthy of discussion. First, there is considerable variability in the results. In part, and certainly of importance at the lower intensities, this may be due to the variation in the endogenous depot of glucose within the skin. In addition, as US acts on the skin barrier itself (as opposed to iontophoresis which acts on the transporting molecules and ions), it follows that the effects of sonophoresis will depend upon the ‘quality’ of the barrier which is subjected to the treatment, i.e. barriers which are inherently more permeable pre-treatment will be more susceptible to the physical perturbation of US, and vice versa. This may explain why the more successful reports of US-extraction of glucose across the skin include the use of a surfactant, chemical enhancer to better ‘normalize’ the increased transport effects observed [3,15]. The second, striking observation from Fig. 3 is that the ‘extraction’ of glucose is not concentrationdependent. At the lower intensities, this lack of sensitivity is not too surprising given that the amounts extracted are not distinguishable from the appropriate controls. For the highest intensity, at which more glucose is able to cross the skin’s barrier, the absence of a clear concentration dependence may imply a role for cutaneous metabolism; indeed, it has been previously shown that the biotransformation of glucose to (for example) lactate and pyruvate is possible during an in vitro experiment [18]. To prove such a hypothesis, however, clearly requires further investigation.

3.3. Low frequency mannitol sonophoresis To simplify interpretation of the results of this study, subsequent experiments employed the nonmetabolizable sugar mannitol, a compound isomeric with glucose but obviously distinguishable as a nonendogenous marker and readily available with a radiolabel to facilitate its detection. Attention was focused on US applied at 20 kHz at the highest intensity (70%) used. The results obtained in this second phase of the investigation are summarized in Table 1 and Fig. 4. Initially, two important controls were carried out. On the one hand, the passive transport of mannitol was first established and was confirmed to be very low. Obviously, unlike the situation with glucose, there was no ‘release’ of endogenous radiolabeled mannitol from the skin. In light of the subsequent experiments to be undertaken, the passive diffusion of mannitol from solutions at 5 and 25 mM was studied in both the ‘extraction’ (lower compartment to upper) and ‘delivery’ (upper to lower) directions. The permeation rates were so small that no significant differences between the amounts transported in 2 h could be distinguished (Table 1). Table 1 Mannitol transport (nmol) across the skin during 2 h of lowfrequency US and during the 2 h immediately following sonophoresis, as a function of concentration in the ‘driving’ solution Experimental conditions

‘Driving’ mannitol concentration

n

5 mM

25 mM

663 862

1062 563

70635 64628

316642 250685

8, 9 8, 9

Post-US ‘extraction’ Post-US ‘delivery’

147646 156646

5456152 4086263

8, 9 8, 9

Stripped skin ‘delivery’

n.d.

Passive ‘extraction’ Passive ‘delivery’ US ‘extraction’ US ‘delivery’

17306572

6 6

6

Results from both ‘extraction’ and ‘delivery’ experiments are presented together with the corresponding controls. As a point of reference, mannitol transport across skin from which the stratum corneum (SC) had been completely stripped is also shown. Data represent the mean6S.D.; n, number of replicates; n.d., not determined.

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

Fig. 4. The effect of low-frequency US on the transdermal transport of mannitol across the skin. US was applied for 2 h at a frequency of 20 kHz and at an intensity of 70%. Mannitol was present in either the lower compartment of the diffusion cell (‘extraction’ experiments) or in the upper chamber (‘delivery’ experiments) at either 5 or 25 mM. The passive controls are shown for comparison.

On the other hand, the transport of mannitol (this time, at a single concentration, in one direction only) was measured across skin from which the stratum corneum (SC) has been completely removed by tapestripping. The purpose of this experiment was to determine the upper limit of the enhancement possible by US, assuming that there can be no greater barrier disruption induced by sonophoresis than the total ablation of the SC. Compared to an intact barrier, mannitol transport was increased by between two and three orders of magnitude. The first experiment considering the effects of low-frequency US on mannitol transport examined the amounts extracted during a 2-h period of sonophoresis, and during the 2 h immediately posttreatment, when the sugar ‘driving’ concentration was either 5 or 25 mM. In contrast to glucose, there was a significant concentration-dependent response when the mannitol concentration was increased fivefold (Fig. 4, Table 1). In addition, post-US, mannitol transport continued at an elevated rate implying a sustained effect of sonophoresis on the skin’s barrier. The level of effect can be appreciated by comparing the amount of mannitol diffusing through skin without SC with that which permeates in the 2 h immediately post-US treatment. It would

91

appear that US reduces skin barrier function by 25–35%, i.e. an important perturbation. As stated previously, current thought points towards cavitational effects as the principal mechanism by which low-frequency US decreases skin barrier function [17]. Electron microscopy evidence supports the hypothesis that US can disrupt either corneocytes and / or lipid bilayers within the SC, allowing perhaps the formation of a continuous aqueous phase through which the transdermal transport of hydrophilic solutes can occur [19]. Recently, attempts have even been made to estimate the dimensions of such an aqueous channel pathway across the barrier [20]. It should also be noted that US may lead to the loss of SC intercellular lipid and to some significant local heating. To further substantiate that the action of lowfrequency US is on the skin per se rather than imparting an additional driving force to the permeant beyond the concentration gradient (as is the case for iontophoresis, for example), the mannitol ‘extraction’ experiments were compared with the results from ‘delivery’ studies, i.e. in which the mannitol was placed in the upper (SC side) compartment of the diffusion cell. The results (Table 1, Fig. 4) reveal that the permeation enhancement is symmetric and is not significantly dependent upon the direction of transport. Equally, post-US treatment, the skin’s permeability remains elevated in a manner completely similar to that observed before (Table 1, Fig. 3). Comparable results were reported by Tang et al. after low-frequency US pre-treatment of the skin in vitro; however, some recovery of skin conductivity was observed after application of a similar protocol in vivo [21].

3.4. Quantification of US intensity The data from the experiments evaluating the energy delivered from the US probe are summarized in Fig. 5, which plots the calculated US intensity (I) and the observed temperature increase (DT ) in the water volume as a function of the nominal percentage of maximal sonicator output. It is observed that the energy delivered and DT increases were certainly proportional to, although not completely linear with, the nominal intensity settings of the apparatus. At the four nominal settings examined (5,

92

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

Fig. 5. Calculated US intensity (I) and observed increase in temperature (DT ) in a water volume as a function of the nominal percentage of maximal sonicator output.

10, 25 and 70%), the average US intensities were 2.7, 4.5, 8.1 and 15 W cm 22 , respectively. These values agree reasonably well with those previously reported [6]. Clearly, at US intensities above 8 W cm 22 , significant increases in the temperature of the vehicle in contact with the skin can be anticipated.

3.5. Contribution of thermal effects to USenhanced transport Fig. 6 illustrates how the 2-h application of US at 20 kHz, at a duty cycle of 0.1 on, 0.9 off, and at

Fig. 6. Temperature increases, as a function of time, induced by sonophoresis at three different intensities (10, 25, 70%), corresponding to 4.5, 8.1 and 15 W cm 22 . Each data point represents the mean6S.D. of six measurements.

three different intensity levels (10, 25 and 70% corresponding to 4.5, 8.1 and 15 W cm 22 ), increased the temperature of the donor solution in contact with the skin in the sonophoresis diffusion cell. The two higher intensities resulted in temperature increases of |10 and 20 8C, on average, at the end of a 2-h treatment. It is noteworthy that one-half of these temperature increments are achieved within only 10 min of US application. This situation is quite distinct from that in which high-frequency US (10 MHz) is similarly applied and which elicits no detectable changes in temperature (,2 8C) over 2 h, or more [22]. To separate out the thermal contribution of US application to the enhancement of transdermal delivery, it was necessary to simulate the heating profile generated by the sonicator in 2 h without activation of the probe. This was achieved by manually regulating the temperature of the diffusion cell so as to mimic the heating effect of US. At the highest intensity used (15 W cm 22 ), this proved to be possible and allowed a true ‘control’ of the sonophoresis experiment (Fig. 7). The mannitol transport results are summarized in Fig. 8. Clearly, sonophoresis at 20 kHz (10% duty cycle), and at an intensity of 15 W cm 22 , significantly enhances skin permeability. The amount of mannitol ‘delivered’ in 2 h across the barrier increased from 6.264.3 to 2116105 nmol, an average enhancement of 35-fold. However, the temperature

Fig. 7. Simulation of the thermal effect of low-frequency sonophoresis (15 W cm 22 ) by simple heating of the diffusion cell (see text for details). Each data point represents the mean6S.D. of six measurements.

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94

Fig. 8. Transdermal mannitol delivery (mean6S.D.; n55) across the skin in 2 h (a) passively, (b) with application of US at 20 kHz (0.1 on, 0.9 off) and 15 W cm 22 , and (c) when heated (in the absence of US) in a manner identical to the thermal profile induced by sonophoresis.

‘control’ experiment revealed that |25% of this improvement could be attributed solely to a thermal effect (which was clearly and significantly different from the passive, isothermal control). The literature supports the observation that increasing temperature leads to an increment in skin permeability [9]. Not only are there references (and generalizations such as that suggesting a doubling of permeability for every 10 8C increase in temperature) reporting the phenomenon, there are also more mechanistic studies which clearly relate enhanced percutaneous transport to specific biophysical events, including phase transitions within the intercellular lipid domains of the SC [8,9,23]. Indeed, in human skin, such transitions have been seen within the range covered by the US-induced thermal effect [8]. Nevertheless, it is apparent that the enhancement effect of US goes well beyond that provoked by a simple temperature increase (even one as significant as that observed here at the highest sonophoretic intensity). An alternative mechanism must therefore be invoked, and the impact of cavitational effects has been widely discussed and supported by other investigations [1,6,17,24]. The US intensity is obviously an important parameter with respect to the resulting permeabilization of the barrier [6,25] and the results seen here are consistent with the dramatic manner in which skin conductivity was increased by the application of sonophoresis at 14 W cm 22 (equally at 20 kHz and 10% duty cycle). It may be reasonable to postulate a synergistic effect between the cavitation phenomenon and the

93

increased temperature—the latter, perhaps, ‘softening up’ (either by disordering or facilitating extraction of, for example) the lipids for disruption by the US energy. Experimental strategies, in which heating effects are dissipated rapidly or US application times are kept short, may shed light on this point. It is worth noting that some of the largest effects of US have been seen when short-term applications have been used in conjunction with surfactant pre-treatment of the skin [26–28] (an alternative approach to ‘soften up’ the barrier, perhaps?). Such mechanistic questions remain important, of course, as the longterm commercial viability of sonophoresis is examined, and a more detailed understanding of the impact of US on the barrier per se represents a focus for ongoing research. In summary, this work confirms the much greater efficiency of low-frequency sonophoresis as a method to enhance transdermal permeability. The effect of US at 20 kHz and 70% intensity is symmetrical (i.e. increasing solute transport equally in both directions) and, under the conditions of these experiments, results in an enhancement of transport which is approximately 25–35% of that achieved by completely removing the SC. Only about one-fourth of this enhancement was attributable to the increased temperature induced by US. Hence, it follows, at least for the model permeant used here, that another mechanism, most probably cavitation, is principally responsible for the lowered skin barrier function observed.

Acknowledgements We thank Cygnus, Inc. and the Swiss National Science Foundation for financial support. Conversations with Peretz Glikfeld and Professor S. Mitragotri have provided valuable insight.

References [1] S. Mitragotri, D. Blanckschtein, R. Langer, Transdermal drug delivery using low frequency sonophoresis, Pharm. Res. 13 (1996) 411–420. [2] K. Tachibana, Transdermal delivery of insulin to alloxan-

94

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11] [12]

[13]

[14]

G. Merino et al. / Journal of Controlled Release 88 (2003) 85–94 diabetic rabbits by ultrasound exposure, Pharm. Res. 7 (1992) 952–954. J. Kost, S. Mitragotri, R.A. Gabbay, M. Pishko, R. Langer, Transdermal monitoring of glucose and other analytes using ultrasound, Nat. Med. 6 (2000) 347–350. A.D. Benwell, S.H.P. Bly, Sources and applications of ultrasound, in: M.H. Repacholi, M. Grandolfo, A. Rindi (Eds.), Ultrasound: Medical Application, Biological Effects and Hazard Potentials, Plenum Press, New York, 1987, pp. 29–47. G.R. Ter Haar, Biological effects of ultrasound in clinical applications, in: K.S. Suslick (Ed.), Ultrasound its Chemical, Physical, and Biological Effect, VCH, New York, 1988, pp. 305–320. S. Mitragotri, J. Farrell, H. Tang, T. Terahara, J. Kost, R. Langer, Determination of threshold energy dose for ultrasound-induced transdermal drug transport, J. Controlled Release 63 (2000) 41–52. S. Mitragotri, D.A. Edwards, D. Blanckschtein, R. Langer, A mechanistic study of ultrasonically-enhanced transdermal drug delivery, J. Pharm. Sci. 84 (1995) 694–706. C.L. Gay, R.H. Guy, G.M. Golden, V.H. Mak, M.L. Francoeur, Characterization of low-temperature (i.e., ,65 degrees C) lipid transitions in human stratum corneum, J. Invest. Dermatol. 103 (1994) 233–239. R.O. Potts, Physical characterization of the stratum corneum: the relationship of mechanical and barrier properties to lipid and protein structure, in: J. Hadgraft, R.H. Guy (Eds.), Transdermal Drug Delivery. Developmental Issues and Research Initiatives, Marcel Dekker, New York, 1989, pp. 23–57. G. Rao, R.H. Guy, P. Glikfeld, W.R. LaCourse, L. Leung, J. Tamada, R.O. Potts, N. Azimi, Reverse iontophoresis: noninvasive glucose monitoring in vivo in humans, Pharm. Res. 12 (1995) 1869–1873. P.W. Wertz, The nature of the epidermal barrier: biochemical aspects, Adv. Drug Deliv. Rev. 18 (1998) 283–294. J. Ziegenhorn, U. Neumman, A. Hagen, W. Bablock, K. Stinshof, Kinetic enzymatic method for automated determination of glucose in blood serum, J. Clin. Chem. Clin. Biochem. 15 (1977) 13–19. ¨ A.F. Buckmann, R. Schmid, A different acceptance of adenine-modified NADP1 by NAD(P)1-dependent glucose dehydrogenase from Bacillus megatherium and strictly NADP1-dependent glucose dehydrogenase from Cryptococcus uniguttulatus, Biotechnol. Appl. Biochem. 14 (1991) 104–113. D. Monti, R. Giannellin, M.G. Trivella, M.F. Saettone, Ultrasonically-enhanced transdermal drug delivery, in:

[15]

[16]

[17]

[18] [19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

Proceedings 2nd World Meeting APFI /APV, 1998, pp. 905– 906. S. Mitragotri, M. Coleman, J. Kost, R. Langer, Transdermal extraction of analytes using low-frequency ultrasound, Pharm. Res. 17 (2000) 466–470. S. Mitragotri, M. Coleman, J. Kost, R. Langer, Analysis of ultrasonically extracted interstitial fluid as a predictor of blood glucose levels, J. Appl. Physiol. 89 (2000) 961–966. A. Tezel, A. Sens, S. Mitragori, Investigations of the role of cavitation in low-frequency sonophoresis using acoustic spectroscopy, J. Pharm. Sci. 91 (2002) 444–453. J.A. Johnson, R.M. Fusaro, The role of the skin in carbohydrate metabolism, Adv. Metab. Disord. 6 (1972) 1–55. L. Machet, N. Cochelin, F. Patat, B. Arbeille, M.C. Machet, G. Lorette, L. Vigilant, In vitro phonophoresis of mannitol, oestradiol and hydrocortisone across human and hairless mouse skin, Int. J. Pharm. 165 (1998) 169–174. J. Wu, J. Chappelow, J. Yang, L. Weimann, Defects generated in human stratum corneum specimens by ultrasound, Ultrasound Med. Biol. 24 (1998) 705–710. H. Tang, D. Blankschtein, R. Langer, Effects of low-frequency ultrasound on the transdermal permeation of mannitol: Comparative studies with in vivo and in vitro skin, J. Pharm. Sci. 91 (2002) 1776–1794. G. Merino, Y.N. Kalia, M.B. Delgado-Charro, R.H. Guy, Mechanism of enhanced transdermal transport by low-frequency sonophoresis. The role of local heating, in: Proceedings of the 27th International Symposium on Controlled Release of Bioactive Materials, 2000, pp. 942–943. R.O. Potts, M.L. Francoeur, Lipid biophysics of water loss through the skin, Proc. Natl. Acad. Sci. USA 87 (1990) 3871–3873. S. Mitragotri, D. Blanckschtein, R. Langer, Ultrasound-mediated transdermal protein delivery, Science 269 (1995) 850–853. J. Liu, T.N. Lewis, M.R. Prausnitz, Non-invasive assessment and control of ultrasound-mediated membrane permeabilization, Pharm. Res. 15 (1998) 918–924. S. Mitragotri, D. Ray, J. Farrell, H. Tang, B. Yu, J. Kost, D. Blankschtein, R. Langer, Synergistic effect of low-frequency ultrasound and sodium lauryl sulfate on transdermal transport, J. Pharm. Sci. 89 (2000) 892–900. J. Kost, S. Mitragotri, R.A. Gabbay, M. Pishko, R. Langer, Transdermal monitoring of glucose and other analytes using ultrasound, Nat. Med. 6 (2000) 347–350. S. Mitragotri, J. Kost, Transdermal delivery of heparin and low-molecular weight heparin using low-frequency ultrasound, Pharm. Res. 18 (2001) 1151–1156.