Anal. Chem. 1990, 62, 674-677
674
In Vitro Percutaneous Absorptiometry by Simultaneous Measurement Using the Photoacoustic Method and Absorbance Ryuichi Takamoto, Ryujiro Namba, Okitsugu Nakata, a n d Tsuguo Sawada*J
Shiseido Co., Ltd., Toxicological & Analytical Research Center, 1050 Nippa-cho, Kohoku-ku, Yokohama 223, Japan
A novel In vltro percutaneous absorptiometry utlllzlng a
portable open-ended photoacwstk (PA) cell as the longitudlnal dMuslon cell was developed. With this system It was feaslble to measure the reductlon of drug applied to a membrane by the PA method and the amount of drug penetratlng the membrane to the Wuslon cell by absorbance, slmuhaneously In real tlme. A shlkonin olntment prepared in a hydrocarbon veMcle was used as the model sample. The In vitro percutaneous abeorptkmetry was performed by means of a physiological sallne solution and the skin of a hairless mouse. As a result, after the lag tlme the absorbance lncreased In proportlon to tkne, whereas the PA slgnal reduced In proportlon to the square root of tlme. As the signal obtained by the PA method corresponds to the amount of drug released from the okrtment,a good correlation wlih Hlguchl's theory is attained. &"My,these redts suggested the usefulness of this novel percutaneous absorptlometry technlque whlch utlllzes the PA method.
INTRODUCTION ,Establishment of an appropriate test method is indispensable for research and development in percutaneous absorptiometry. Various methods of percutaneous absorptiometry have been reported, although needless to say that most such methods employed by numerous researchers, as well as those used generally, are in vitro methods using a diffusion cell. In this method, the donor side and the receiver side of the diffusion cell must be filled with physiological saline solution or a certain solvent depending on the purpose of the test. This in vitro test is easily operable and the results are reproducible. Additionally, the results can be analyzed fairly easily by using Higuchi's theory (1). For example, the amount of drug released from such a suspension-type ointment can be calculated by this diffusion theory. If the release is less than 30%, the equation is given by
Q
= (D(2Co - C,)C,t)1/2
(1) where Q is the amount of drug released per unit area of application, D is the diffusion coefficient of drug in the ointment, Co is the initial concentration of drug in the ointment, C, is the solubility of the drug in the ointment, and t is the time after application. This equation can also be described as
Q = kt1i2
( k = constant) (2) According to these remarkably simple relationships, the amount of drug released from such a suspension-type ointment is shown to be proportional to the square root of time. We came up with an idea that if by means of a longitudinal cell, which more closely reflects the clinical conditions as compared to the conventional parallel cell, we could both sensitively and simultaneously a t real time measure the rePresent address: Department of Industrial Chemistry, Faculty of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku Tokyo 113, Japan. 0003-2700/90/0362-0674$02.50/0
duction of drug on the donor and its increase on the receiver side, we could obtain all the information related to the vehicle, the drug, and its penetration in a single sample run. Thus, we developed a novel in vitro test method capable of simultaneously measuring the reduction of drug on the donor side by the PA method and the amount of drug penetrating through the skin by absorbance, in real time. A laser beam was employed as the light source. The portable open-ended PA cell that we developed was used for the PA method. Applications of percutaneous absorptiometry by the PA method using the open-ended PA cell have been reported by Kolmel et al. (2,3) and its usefulness has been demonstrated. We targeted the development of a hypersensitive in vitro test method using a laser beam as the light source. Furthermore, by applying the PA method and absorbance simultaneously to the longitudinal diffusion cell, we could expect that the cutaneous permeability of the drug and the releasing rate of the drug from the vehicle can be analyzed in closer relation to actual clinical conditions. Thus, the newly developed in vitro test method for measuring PA signal and absorbance simultaneously was applied to the membrane transmission test by using an artificial membrane and in vitro percutaneous absorptiometry using the skin of a hairless mouse, to study its validity. EXPERIMENTAL SECTION The in vitro percutaneous absorptiometer for simultaneously measuring the PA signal and absorbance was developed, as shown in Figures 1 and 2. The argon laser beam (wavelength, 488 nm; Spectra Physics, Model-164) used as the radiation source was intensely modulated with an A.O. modulator (Intra Action Corp., AOM-40) at 3.3 kHz (Function Synthesizer, NF, 1925) and was separated into two spectral directions, for both PA and absorbance measurements. The intensity of the radiation s o m e was adjusted so as to give an output of 10 mW (below the maximum permissible exposure ( W E )on human skin) from the optical fiber (Mitsubishi Rayon, SK-40, 1 mm in diameter x 1 m) at the PA cell. The portable open-ended P A cell consisting mainly of brass was processed as a stick and chromium plated to enhance the scattering of light. Furthermore, it was equipped with an optical fiber to allow freedom of measurement, and the microphone could be moved to control the volume inside the cell. The capacity and the thickness of the air layer inside the cell were 0.1 cm3and 0.2 cm, respectively. The pipe connecting the sample chamber with the microphone chamber was 12 cm long. These values were chosen so that the cell would resonate at several kilohertz, a frequency which is hardly influenced by neither the surface potential of skin nor environmental noises. Before measurement, the microphone chamber of the cell was shifted around to control the capacity of the cell to maximize the amplification of PA signals by resonance. The obtained PA signals were analyzed by a lock-in amplifier (NF, 5610A) and thereafter recorded on a chart recorder (Rikadenki Kogyo, NP-0393). The laser beam for measuring the absorbance was further separated into two directions,one for the quartz longitudinal diffusion cell (Corp.Atock) capable of measuring absorbance and the other to the photocell (United Detector Technology, PIN-1ODP) as the reference signal (lo).The measured signal (Z) and the reference signal (lo)were introduced into a differential amplifier (NF, P-61), and after calculation of the absorbance (-log (Z/lo))with a computer (Nippon Denki, PC9801-VX),the relation between PA signal and absorbance as 0 1990 American Chemical Society
ANALYTICAL CHEMISTRY. VOL. 62. NO. 7. APRIL 1. 1990 675
.
0
20
10
Concentration o f Shixonin
["9/5m+l
Figure 3. Calibration curve of shikonin ointment by open-ended PA cell (fl SD shown in error ban). Magnetic Stirrer + n ' o u a r t z
Longitudinal DiffUSlO" C e l l
--
'01
Flgure 1. Illustration of diffusion cell applied in simultaneous measurement of the PA method and abswbance.
0
I
O s e 1 0 Concentration or Sntuonin IDm1
Figure 4. Calibration curve of shikonin by absorbance method.
Flgure 2. Diagram of the in vitro percutaneous absorptiometry by simultaneous measurement of PA signal and absorbance. a function of time was determined. The major noise factors that affected the measurement of absorbance were scattered beams of laser and room light. By performing the experiments in a dark room and also by covering the photocell with an aluminum sheet to avoid light scattering, it became posaible to transmit only the laser beam to the receiver side for optimum detection. The laser beam was emitted onto a point that was found not to be affected by stirring the solution. During measurement by the open-ended PA cell, a spike noise was found to generate when the rotor contacted the quartz cell. Therefore, the shape of the rotor wa8 modified 80 as to give a stable rotation without touching the quartz cell. Consequently, a stable signal of less than 10% variation was achieved. Shikonin ((S)-5,8-Dihydroxy-2-(l-hydroxy-4-methyl-3-pentenyl)-1,4-naphthalenedione;C16H1605,Wako Pure Chemical, Reagent class, >95%), a wound healing drug, which absorbs 488-nm light, was chosen a8 the model sample. Ointments containing 0.1-1%, 2%, and 3% of shikonin in a hydrocarbon base (Taisho Pharmaceutical, Plastibase) were prepared. Shikooin was dissolved in isopropyl almhol (IPA) and mixed with the base, and the solvent was evaporated under vacuum. A porous silicone membrane (Fuji System) and the skin (defatted) of a hairless mouse (male m o w of SKH/HR-1 strain, 6 weeks old), were used. IPA and physiological saline solution were employed as solvents for the receiver side of the diffusion cell. RESULTS AND DISCUSSION Precision and Sensitivity of the Simultaneous PA/ Absorbance Measuring System. The precision of detection of PA/absorbauce measurements for shikonin was investigated. Shikonin ointments (0.1-1.0%) were applied to a silicone membrane at an amount of 2 mg and a 5 mm diameter circle (0.2 cm2), respectively, and were used as standard samples for the PA measurements. To measure the ahsor-
Tlme rm7n.l
ngue 5. StabllHy of a novel in vitro pemnaneous absupb'omeby: (a) sample, 3% Shikonin ointment: solvent. nil: (b,c) sample, vehicle: solvent. physiological saline solution
bance, shikonin solutions (0.1-10 ppm in IPA) were applied to the receiver side of the diffusion cell. In Figures 3 and 4 are shown the results of the calibration curves for the PA/ absorbance measurements. The PA signal obtained by an open-ended PA cell, gave a signal to noise (S/N) ratio of 46 and a correlation coefficient of 0.991 for shikonin (2-20 pg/5 mm diameter) with +1 standard deviation (SD). The correlation coefficient for the absorbance measurements was 0.998 for shikonin at 0.1-10 ppm. The next experiment was performed by applying only the base material to the silicone membrane, and the receiver side of the diffusion cell was filled with physiological saline solution. As shown in Figure 5, PA signal and the absorbance were found to be relatively constant. No bubbles were observed between the membrane and the receiver, until the end of measurement. Furthermore, the
676
ANALYTICAL CHEMISTRY, VOL. 62, NO. 7, APRIL 1, 1990
. . . . PA signal L.'."'':.' 0 .
1 0 2 0
3 0 4 0
*
5 0
e
.
6 0
7 0
c
i_* 1 0
2 0
1 5
2 5
T i m e [hour]
Figure 8. In vitro percutaneous absorptiometry using hairless mouse skin.
Time rmin.1 Flgure 6. Permeability in vitro experiment with artificial membrane.
c 2 0
5
1 .J
5
5. 0
Lag T i m e ( 7 . 3 h )
10
15
10
25
T i m e [hour]
Flgure 9. Time course of shikonin passing through the membrane determined from the absorbance. 0
10
20
30
40
5 0 6 0 70 Time [min.]
Flgure 7. Comparison of absorbance (amount of shikonin passing through the membrane) and PA signal (amount of shikonin released from the ointment.
effect of shikonin upon laser beam emission was studied, the result of which is also shown in Figure 5. When PA signal was measured with an argon laser of 488 nm of the 3% shikonin ointment as the sample applied on a silicone membrane, the PA signal was found to be constant after repeated radiation of the beam for 200 min, implying that shikonin is not decomposed under this condition. Furthermore, the same result was obtained under radiation for up to 25 h.
Permeability Experiment with Artificial Membrane. PA signal and absorbance were measured simultaneously on 1.3 mg of 1% shikonin ointment (equivalent to 13 pg of shikonin) applied (5 mm diameter) on a silicone membrane, using IPA as the diffusion cell solvent. As seen from Figure 6, the PA signal decreased as the absorbance increased. Since shikonin dissolves well in IPA in this system, the diffusion rate of shikonin in the membrane is very fast. It was observed that after a lag time of about 30 s, the absorbance increased, and 60 min later, only the vehicle remained on the membrane. The relationship between the amount of shikonin determined from the calibration graphs (Figures 3 and 4) and time is shown in Figure 7 . The amount of shikonin at 60 min, as calculated from absorbance, was 13.1 pg, which was identical with the amount initially applied. Therefore, 100% of the initially applied shikonin can be considered to have transmitted the membrane. After about 30 min, the graph gradually loses linearity, which is possibly due to the reduction of concentration gradient at the donor and receiver sides of the cell. Time course of PA signal was assayed assuming that
the difference between the initial PA signal ( 8 0 ) and that t hours later (Qt),namely (Qo - Qt), corresponds to the amount of shikonin released from the ointment after t hours. Accordingly, it can be considered as the release of shikonin from the ointment, as a function of time. From these results, it seems that the time course of PA signal is interpretable to absorbance, or vice versa. It is considered that the difference between the PA curve and the absorbance curve is mainly due to the variation (20%,f l SD) of PA calibration curve (Figure 3). Since the diffusion coefficient of shikonin is extremely high in this system, not much time difference is expected between the decrease of shikonin above the membrane and its increase below the membrane. Hence, the results obtained are considered fairly reasonable. Furthermore, the time course of PA signal was shown to indicate the reduction of shikonin above the membrane, and it was found that the transmission of drug through the membrane can be evaluated by the PA method. In Vitro Percutaneous Absorptiometry Using Hairless Mouse Skin. A series of in vitro percutaneous absorptiometries was performed on the skin of hairless mouse. Within a 5 mm diameter circle on the skin of a hairless mouse (defatted), 6 mg of 3% shikonin ointment (equivalent to 180 pg of shikonin) was applied. Employing physiological saline solution as the diffusion cell solvent, the absorbance and PA signal were measured simultaneously. The result is shown in Figure 8. An increase in absorbance and decrease in PA signal were also observed in the hairless mouse skin system, resembling a living body. Fluctuations in the absorbance are considered to be caused by floating matter and water-soluble matter deriving from the skin. Quantification of shikonin as calculated from absorbance is shown in Figure 9. The amount of shikonin passing through the skin started to increase proportionally to time after the 7 h lag time, and it was found that 3.4% of the initially applied shikonin had transmitted
ANALYTICAL CHEMISTRY, VOL. 62, NO. 7, APRIL 1, 1990
.
a 4.01
s
0 -;3.01
a. '
.
O
i
L
3.0 4.0 5.0 T i m e Ihour14 Flgure 10. Correlation between PA signal (Oo 0,)and square root of time.
0
1.0
2.0
-
the skin after 25 h. The rather longer lag time observed here is probably due to temperature (4), which in this experiment was ambient (20 "C), whereas the in vitro percutaneous absorptiometry experiments were undergone at 37 O C , with respect to body temperature. After 25 h, hardly any bubbles were observed between the skin and the receiver. Further, the PA signal (Qo- Q t ) , corresponding to the release of shikonin from the ointment, increased in proportion to the square root of time as shown in Figure 10. This result correlates favorably with Higuchi's theory. After about 15 h, the graph gradually loses linearity, which is possibly deriving from the reduction of partition coefficient (skin/vehicle) of shikonin due to the concentration of the receiver side approaching a saturated state. Consequently, we can expect this in vitro percutaneous absorptiometry system, whereby the PA signal and the absorbance is measured simultaneously, to be capable of analyzing both the drug-releasing process and its transmission process through the skin, real time, in a system resembling actual clinical conditions.
CONCLUSION A novel in vitro percutaneous absorptiometry system, capable of measuring the reduction of the drug above the skin and the amount permeating through the skin simultaneously in real time, was developed upon application of the PA signal and absorbance measurement utilizing a longitudinal diffusion cell. Experiments performed on model shikonin ointments confirmed the following two points. First of all, in the system
677
using silicone membrane with high diffusion coefficient of drug, the time course of PA signal and absorbance showed a similar trend. In a practical in vitro percutaneous permeability experiment using the skin of a hairless mouse, the PA signal showed a favorable correlation with Higuchi's theory. These results suggested the usefulness of this newly developed in vitro percutaneous absorptiometry, enabling us to measure the PA signal and absorbance simultaneously. A practical problem, yet unsolved with regard to using a laser beam as the light source, is the limitation in wavelength despite its high sensitivity, which restricts the number of drugs amenable to this technique. However, by oscillating multiple Ar+ laser beams and using a nonlinear optical crystal wavelength conversion unit (Ascal, UVA-4), CW-UV lights of several wavelengths are attainable. By oscillating UV light with this method, we have recently carried out a series of experiments on various drugs. This newly developed in vitro percutaneous absorptiometry system can be further modified by introducing a thermoregulating device with which to control the temperature of the diffusion cell solvent. By doing so, we should be able to perform various studies, such as the releasing process of drugs and vehicle, under conditions resembling those of actual clinical situations. From this study it became clear that the membrane transmission phenomenon of a drug is measurable by means of a portable open-ended PA cell. Applications of this PA cell to the measurement of human percutaneous absorption in vivo are under way.
ACKNOWLEDGMENT The authors thank Fujihiro Kanda (Toxicological & Analytical Research Center, Shiseido Co., Ltd.) for help in critical reading of this paper. Registry No. Shikonin, 517-89-5.
LITERATURE CITED (1) Higuchi, T. J . Pharm. Sci. 1981, 50, 874-875. (2) Kolmel, K.; Sennhenn, B.; Giese, K. J . SOC. Cosmet. Chem. 1986, 37, 375-305. (3) Giese, K.; Nicolaus, A.; Sennhenn, B.; Kolmel, K. Can. J . fhys. 1988, 64, 1139-1142. (4) Fritsch, W. C.; Stoughton, R. B. J . Invest. Dermatol. 1963, 4 7 , 307-312.
RECEIVED for review August 22, 1989. Accepted December 18, 1989.
