Effect of Diffusive Direction across the Skin on the Penetration Profile of Chemicals in Vitro

Abstract

Skin has various types of transporters and is a biochemically active organ. These aspects of skin influence the distribution of chemicals in skin and their elimination from skin. The biochemical and histological variations of the skin must be taken into account when conducting transdermal penetration research. Here we used hairless mouse skin to investigate the percutaneous absorption of chemicals in vitro from the stratum corneum (SC) side to the viable skin (VS) side (forward direction) and from the VS side to the SC side (backward direction). We examined the effects of molecular weight, lipophilicity (Log K_o/w), electric charge, and the molecular structure of penetrants. The penetration flux of verapamil hydrochloride (VRP) for the backward direction was 3.2 times larger than that for the forward direction. The flux values of benzoic acid (BA) and para-hydroxybenzoic acid (pHBA) for the forward direction were 2.1 and 4.6 times larger than those for the backward direction, respectively. This directional difference was caused by the active transporter for VRP, the histological distribution of BA solubility, and the intermolecular hydrogen bonding between pHBA and skin tissue in the stripped skin. Across intact skin, in contrast, there was no difference in the skin penetration profile between the forward direction and backward directions.

Skin consists of two different layers, the stratum corneum (SC, the lipophilic layer) and viable skin (VS, the hydrophilic layer). The diffusion-partitioning model based on Fick’s law of diffusion is widely applied to the mathematical model of transdermal penetration.¹⁾ Penetration parameters such as diffusion and partition coefficients used in this model can be determined from in vitro penetration experiments to predict the skin penetration flux. Each layer of the skin assumes an inert and monolithic membrane for penetration. However, skin, especially VS, has biochemical activity. Skin enzymes are distributed around the basement layer of the epidermis²⁾ and their distribution influences the penetration of drugs.³⁾

Active transporters are present in the skin and may affect the percutaneous absorption of penetrants.⁴⁾ One of the xenobiotic transporters, P-glycoprotein (P-gp), is expressed not only in human skin but in hairless mouse skin as well⁴⁾ and is locally observed in dermis and hypodermis by immunohistochemical analysis.⁵⁾ The other transporters, oligopeptide transporter, monocarboxylate transporter, organic anion transporting polypeptide, and organic cation transporter, were also expressed in the skin.^4,6,7) A Zn transporter, Znt5/Slc30a5, participates in the mast cell-mediated delayed-type allergic response and may be involved in allergic contact dermatitis.⁸⁾ These transporters may influence the dermatokinetic distribution and elimination of chemicals in order to maintain the internal environment.

The biochemical and histological variations of the skin should be considered in transdermal penetration research. To investigate the effect of skin’s histological and biochemical homogeneity on the skin penetration of chemicals, we used hairless mouse skin in vitro to determine the percutaneous absorption from the SC side to the VS side (the forward direction) and from the VS side to the SC side (the backward direction). The model chemicals were selected with respect to their molecular weight, lipophilicity (Log K_o/w), electric charge, and molecular structure. We also studied the percutaneous absorption of both stripped skin and intact skin, and we discuss the importance of the diffusive direction of chemicals in transdermal therapeutic systems.

MATERIALS AND METHODS

Materials

Verapamil hydrochloride (VRP) was purchased from Sigma (St. Louis, MO, U.S.A.), and benzoic acid (BA), propranolol (PRP), p-hydroxybenzoic acid (pHBA), indomethacin (IM), caffeine (CF), progesterone (PS), testosterone (TS), prednisolone (PN), cyanocobalamin (VB₁₂), allethrin (AT) and dexamethasone (DM) were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Salicylic acid (SA) was purchased from Kanto Chemicals (Tokyo, Japan). Other chemicals were special reagent-grade and purchased from Wako Pure Chemical Industries, Ltd. The physicochemical properties of the chemicals are summarized in Table 1; some properties were obtained from the literature.⁹⁾

Table 1. Physicochemical Properties of the Chemicals Used

Drug	Molecular weight	pKa	Log KO/W	Electric charge at pH 7.4	Concentration in solvent (µg/mL)
BA	122	4.19	1.87	−1	3.38×104±5.27×102
SA	138	2.97	2.26	−1	6.20×104±4.67×103
pHBA	138	4.57	1.58	−1	2.47×104±3.24×102
CF	194	10.4	−0.07	+1	2.96×104±1.87×102
TS	288	−	3.32	0	5.91×102±4.93×10
AT	302	−	4.96	0	1.20×102±3.56
PS	315	−	3.87	0	1.58×102±1.43×10
IM	358	4.17	4.27	−1	4.43×102±9.11
PN	360	−	1.42	0	2.13×103±6.15×10
DM	392	−	0.72a)	0	1.20×103±7.05×102
VRP	491	8.6a)	5.69a)	+1	500
VB12	1355	−	−4.30	0	1000

a) Values are from Kasim et al. 2004. BA; benzoic acid, SA; salicylic acid, pHBA; p-hydroxybenzoic acid, CF; caffeine, TS; testosterone, AT; allethrin, PS; progesterone, IM; indomethacin, PN; prednisolone, DM; dexamethasone, VRP; verapamil hydrochloride and VB₁₂; vitamin B₁₂.

Penetration Experiments

Abdominal intact and stripped skin were excised from female hairless mice (Kud : Hr⁻ strain, 7 weeks old, Kyudo Co., Saga, Japan). The SC of the stripped skin was removed completely by adhesive tape-stripping the skin 20 times (CF-24F, Nichiban Co., Tokyo, Japan). Skin samples were mounted on side-by-side permeation cells (KH-5P, Vidrex Co., Fukuoka, Japan). The receptor cell was filled with 5 mL of fresh solvent and the donor cell was filled with the chemical solution. The solvent was 40% polyethylene glycol (PEG) 400 solution for BA (saturated concentration, C_S), SA (C_S), pHBA (C_S), TS (C_S), AT (C_S), PS (C_S), PN (C_S) and DM (C_S) and phosphate buffer at pH 7.4 for CF (C_S), IM (C_S), VRP (500 µg/mL) and VB₁₂ (1000 µg/mL). Concentration of model chemicals in solvent was also summarized in Table 1. In the penetration experiments for BA, SA and pHBA, we also used buffer that adjusted the pH at 7.4 (phosphate buffer), 5.2 (phosphate buffer), 4.2 (citric-phosphate buffer), 3.0 (citric-phosphate buffer), and 2.6 (citric-phosphate buffer) saturated with the chemicals. At predetermined time intervals, 200 µL of receptor solution was withdrawn for an HPLC assay. The steady-state penetration flux was evaluated from the linear portion of the plot of the cumulative amount penetrated against time. All animal experiments were conducted in accord with our institutional guidelines and were approved by the Animal Care and Use Committee, Kyushu Institute of Technology.

BA Concentration in the Skin

The stripped skin sample was cut into pieces of 2.2×1.8−cm pieces, and each skin sample was embedded into Tissue-Tek OCT compound (Sakura Finetechnical Co., Tokyo, Japan) at −20°C. The skin samples were sliced into 20-µm sections, and each two sliced skin sections from the surface of each skin sample was immersed into BA solution with buffer (pH 7.4, 4.2 and 2.6) and then equilibrated at 37°C for 24 h. The BA concentration in the solution was adjusted at 50 µg/mL. Equilibrated solution was centrifuged at 15,000 rpm (MRX-152, Tomy Seiko Co., Tokyo, Japan) for 15 min, and the BA concentration in the supernatant was assayed by HPLC. We then calculated the skin/buffer partition coefficient K_skin/buffer using the BA amount in each sliced skin section and the BA concentration in the solution.

HPLC Analysis

All model chemicals were assayed in an HPLC assembled LC 10A system (Shimadzu Co., Kyoto, Japan) with four HPLC columns: a SUMIPACK-ODS A-212 (6×150 mm, Sumitomo Chemical Co., Tokyo, Japan) for BA and CF, a CAPCELL Pak C18 UG120 (4.6×250 mm, Shimadzu GLC, Osaka, Japan) for SA and pHBA, a TSK gel ODS 80Ts (4.6×150 mm, Tosoh Co., Tokyo, Japan) for VRP, DM, IM, PS, TS and PN, and an L-column ODS (4.6×150 mm, Chemical Evaluation and Research Institute, Tokyo, Japan) for AT. The column oven temperature was maintained at 40°C. The other assay conditions are summarized in Table 2.

Table 2. Analytical Conditions of the Model Chemicals Used

	Wave length (nm)	Mobile phase	Flow rate (mL/min)
BA	220	35% Acetonitrile/65% (0.048 M NaH2PO4–0.041 M H3PO4)	1.0
SA	270	25% Methanol/75% 0.1 M KH2PO4 (pH 7.0)	1.0
pHBA	270	10% Methanol/90% 0.1 M KH2PO4 (pH 7.0)	0.8
CF	263	30% Methanol/70% distilled water	1.0
TS	254	65% Methanol/35% distilled water	1.0
AT	225	80% Acetonitrile/20% distilled water	1.0
PS	254	70% Methanol/30% distilled water	1.0
IM	254	70% Methanol /30% 0.05% phosphate aqueous solution	1.0
PN	254	50% Methanol/50% distilled water	1.0
DM	254	58% Methanol/42% distilled water	1.0
VRP	260	50% Methanol/50% 50 mM NaH2PO4 (pH 2.2)	1.0
VB12	280	12% Acetonitrile/88% 10 mM phosphate buffer (pH 7.4)	1.0

Abbreviations are indicated in the footnote of Table 1.

RESULTS AND DISCUSSION

Directional Difference of Skin Penetration

We carried out the skin penetration experiments of the chemicals VRP, BA, AT, DM, IM, CF, PS, TS, PN, and VB₁₂, which have different molecular weights, lipophilicity, and electric charge, across stripped skin. The directional differences in the skin penetration profiles are shown in Fig. 1. The cumulative amount penetrated of model chemicals is influenced by their physicochemical properties and chemicals with high molecular weight, high lipophilicity, and nonelectric charge have the lower penetration across stripped skin.¹⁰⁾ These results in Fig. 1 are agreed with the VRP profile from the backward direction is higher than that from the forward direction (Fig. 1a), and the BA profile shows the reverse (Fig. 1b). The skin penetrations of AT and DT show the same profile between the forward and backward directions (Figs. 1c, 1d) and the other chemicals IM, CF, PS, TS, PN, and VB₁₂ also have the same results (data not shown). It has been reported that the percutaneous absorption of rhodamine 123, flurbiprofen, and itaconazole also has a directional difference across the stripped skin.^4,5) These findings may indicate that the molecular weight and lipophilicity are not responsible for the directional difference of skin penetration within the range of the present experiments.

Fig. 1. Penetration Profiles of VRP (a), BA (b), AT (c) and DM (d) across Stripped Skin of Hairless Mice

The forward and backward directions represent the skin penetration from the SC side to the dermis side and from the dermis side to the SC side, respectively. Each data point is the mean±S.D. of three experiments.

Effect of Transporter on Skin Penetration

VRP is known to be a substrate of organic cation transport system^11,12) that is represented by organic cation transporters (OCT; SLC22A1, 2 and 3) and of P-gp (ABCB1) that is a member of ATP-binding cassette family.¹³⁾ Organic cation transport system acts as an influx transport for essential nutrients on the cell membrane and, on the other hand, P-gp acts as an efflux transport for xenobiotic substances. Ito et al has reported that P-gp enhanced the penetration flux for the forward direction to eliminate its substrates from skin to blood.⁵⁾ Thus, our results in Figs. 1a and 1b may indicate that the percutaneous absorption of VRP and BA is mediated by the influx transporter and the efflux transporter, respectively. Thus, we investigated co-penetration experiments of VRP and BA with propranolol (PRP), the substrate of both organic cation transport system and P-gp, and we added PRP at the same concentration to each donor solution (500 µg/mL for VRP and 30 mg/mL for BA) to measure the skin penetration flux. The penetration profiles of VRP with PRP for the forward direction agreed with that for the backward direction (Fig. 2a), whereas the forward profile of BA was different from its backward profile (Fig. 2b). Thus, PRP acts as a competitive inhibitor only for the backward direction of VRP and, on the other hand, it reduced the cumulative amount of BA to a half in both directions (Figs. 1b, 2b) because a chemical potential of BA in the donor solution relatively decreased. Therefore, the directional difference of the VRP penetration profile was caused mainly by the transport of organic cation transport system in the skin and transporters did not influence the penetration profile of BA.

Fig. 2. Effect of PRP on the Skin Penetration Profiles of VRP (a) and BA (b) across Stripped Skin from Hairless Mice

Each data point is the mean±S.D. of three experiments.

Effect of a pH Gradient throughout Stripped Skin

We conducted penetration experiments with BA with a pH gradient between the donor solution (pH 4.2, BA saturated in 40% PEG400) and the receptor solution (pH 9.0, 40% PEG400). The natural pH values of the SC surface and the deeper layer in vivo are about 5.0 and 7.4, respectively.¹⁴⁾ In vitro, the pH values are easily changed by solution contacting the skin surface.¹⁵⁾ We also conducted an experiment using the same pH values for the donor and receptor solutions. The pH values selected were 2.6, 4.2 (pK_a of BA) and 7.4. The steady-state flux is shown in Fig. 3. There was no significant difference in the penetration flux between the forward and backward directions under the same pH. The steady-state fluxes of BA at pH 4.2 were the highest values for the forward and backward directions. These findings show that the pH gradient throughout stripped skin is important for the directional difference of BA penetration, and they may indicate that there is an interaction between BA and skin tissue at pH 4.2.

Fig. 3. Effect of the pH Gradient on the Steady-State Penetration Flux of BA

PEG has a pH gradient between the donor (pH 4.2) and receptor (pH 9.0) solution, and the pH of the buffer was adjusted to be the same in the donor and receptor solutions. The columns indicate the mean±S.D. of three experiments.

Skin consists of SC and VS (the viable epidermis and dermis). The basement layer of the viable epidermis (the hydrophilic layer) becomes the SC (the lipophilic layer) after dehydrating from the cell and, thus, there may be a lipophilicity gradient throughout the viable epidermis. The pH of the solution changes the lipophilicity of BA. The pH gradient may thus cause the skin distribution of BA solubility. We investigated the skin/buffer partition coefficient K_skin/buffer at the pH values of 7.4, 4.0 and 2.6 (Fig. 4). K_skin/buffer had a constant value throughout the stripped skin at pH 7.4 and, in contrast, decreased with the increase in the distance from the skin surface at pH 4.0 and 2.6. At the depth of 40 µm, the values of K_skin/buffer at pH 4.0 were 4.3 and 2.1 times larger than those at pH 7.4 and pH 2.6, respectively. Therefore, the high BA solubility around the surface of viable skin (ca. 40 µm) at pH 4.2 may cause the directional difference in skin penetration.

Fig. 4. Skin Distribution of the Skin/Buffer Partition Coefficient K_skin/buffer of BA

* p<0.05, by t-test. The columns indicate the mean±S.D. of three experiments.

Effect of Molecular Structure

We also studied the effect of the molecular structure of the chemicals on the directional difference of BA penetration. We selected SA and pHBA, with structures similar to that of BA; SA and pHBA have a hydroxyl group on the ortho- and para- to carboxyl group of BA, respectively. There was no directional difference in the SA penetration flux (Fig. 5a). The pHBA penetration flux, in contrast, was larger for the forward direction than for the backward direction when the pH of the donor solution was different from the pH of receptor solution (Fig. 5b). SA tends to form intramolecular hydrogen bonding between the hydroxyl group and carboxyl group and stabilizes in solution. pHBA and BA form an intermolecular hydrogen bonding between the skin components lipid and protein. The interaction of pHBA and BA with skin tissue may cause the directional difference.

Fig. 5. Steady-State Flux of SA (a) and pHBA (b) across Stripped Skin

The skin penetration of SA showed no difference between the forward and backward directions but pHBA showed a directional difference in PEG and pH 5.2/7.4 solution with a pH gradient. * p<0.05, by t-test. The columns indicate the mean±S.D. of three experiments.

Penetration across Intact Skin

Transdermal therapeutic systems (TTSs) usually focus on enhancing the SC penetration delivered drugs across intact skin. To research the mechanism underlying the directional differences in the skin penetration, we investigated the skin penetration experiments with stripped skin without SC from hairless mice, and then we examined the skin penetration profile of VRP and BA across intact skin with SC. Figure 6 shows no difference in these chemicals’ penetration profiles between the forward and backward directions. This indicates that the intact skin has no directional difference for chemical penetration. SC is a rate-limiting barrier for transdermal absorption because the ratio of the SC diffusion coefficient to that of the VS is generally less than 1/100. SC protects the internal environment and also maintains the pH gradient in skin when acid or alkaline solution is in contact with the skin surface.¹³⁾ Our findings revealed that the directional difference in chemical penetration at intact skin was able to ignore. Therefore, we concluded that intact skin with SC could be used in the diffusion-partitioning model based on Fick’s law of diffusion on TTS research. However, the thickness of the SC varies with age, anatomical site and disease’s conditions of each patient affecting the skin penetration flux of drugs from a TTS device.¹⁶⁾ We need to pay attention to assume the penetration profiles of drugs on clinical application.

Fig. 6. Penetration Profiles of VRP (a) and BA (b) across Intact Skin from Hairless Mice

Each data point is the mean±S.D. of three experiments.

CONCLUSION

We investigated the percutaneous absorption of 12 chemicals in the forward direction and backward direction in hairless mouse skin in vitro. We found that VRP, BA and pHBA penetration across the stripped skin had a directional difference. The penetration flux of VRP for the backward direction was 3.2 times larger than that for the forward direction. The flux values of BA and pHBA for the forward direction were 2.1 and 4.6 times larger than those for the backward direction, respectively. This directional difference is caused by the organic cation transport system for VRP, the skin distribution of BA solubility, and the intermolecular hydrogen bonding between pHBA/BA and skin tissue in the stripped skin. There was no significant difference in the skin penetration profile of chemicals across intact skin for both the forward direction and backward direction. Therefore, intact skin can be used in the diffusion-partitioning model based on Fick’s law of diffusion on TTS research.

REFERENCES

• 1)  Tojo K, Chiang CC, Chien YW. Drug permeation across the skin: effect of penetrant hydrophilicity. J. Pharm. Sci., 76, 123–126 (1987).
• 2)  Hikima T, Maibach HI. Distribution of hydrolytic activity catalyzes the biotransformation of prednisolone 21-acetate in human skin. Skin Pharmacol. Appl. Skin Physiol., 14, 196–202 (2001).
• 3)  Tojo K, Yamada K, Hikima T. Diffusion and metabolism of prednisolone farnesylate in viable skin of the hairless mouse. Pharm. Res., 11, 393–397 (1994).
• 4)  Li Q, Tsuji H, Kato Y, Sai Y, Kubo Y, Tsuji A. Characterization of the transdermal transport of flurbiprofen and indomethacin. J. Control. Release, 110, 542–556 (2006).
• 5)  Ito K, Nguyen HT, Kato Y, Wakayama T, Kubo Y, Iseki S, Tsuji A. P-Glycoprotein (Abcb1) is involved in absorptive drug transport in skin. J. Control. Release, 131, 198–204 (2008).
• 6)  Schiffer R, Neis M, Höller D, Rodríguez F, Geier A, Gartung C, Lammert F, Dreuw A, Zwadlo-Klarwasser G, Merk H, Jugert F, M Baron J. Active influx transport is mediated by members of the organic anion transporting polypeptide family in human epidermal keratinocytes. J. Invest. Dermatol., 120, 285–291 (2003).
• 7)  Ito K, Kato Y, Tsuji H, Nguyen HT, Kubo Y, Tsuji A. Involvement of organic anion transport system in transdermal absorption of flurbiprofen. J. Control. Release, 124, 60–68 (2007).
• 8)  Nishida K, Hasegawa A, Nakae S, Oboki K, Saito H, Yamasaki S, Hirano T. Zinc transporter Znt5/Slc30a5 is required for the mast cell-mediated delayed-type allergic reaction but not the immediate-type reaction. J. Exp. Med., 206, 1351–1364 (2009).
• 9)  Kasim NA, Whitehouse M, Ramachandran C, Bermejo M, Lennernäs H, Hussain AS, Junginger HE, Stavchansky SA, Midha KK, Shah VP, Amidon GL. Molecular properties of WHO essential drugs and provisional biopharmaceutical classification. Mol. Pharm., 1, 85–96 (2004).
• 10)  Nielsen JB, Nielsen F, Sørensen JA. In vitro percutaneous penetration of five pesticides—effects of molecular weight and solubility characteristics. Ann. Occup. Hyg., 48, 697–705 (2004).
• 11)  Kubo Y, Kusagawa Y, Tachikawa M, Akanuma S, Hosoya K. Involvement of a novel organic cation transporter in verapamil transport across the inner blood–retinal barrier. Pharm. Res., 30, 847–856 (2013).
• 12)  Kubo Y, Shimizu Y, Kusagawa Y, Akanuma S, Hosoya K. Propranolol transport across the inner blood–retinal barrier: potential involvement of a novel organic cation transporter. J. Pharm. Sci., 102, 3332–3342 (2013).
• 13)  Bansal T, Mishra G, Jaggi M, Khar RK, Talegaonkar S. Effect of P-glycoprotein inhibitor, verapamil, on oral bioavailability and pharmacokinetics of irinotecan in rats. Eur. J. Pharm. Sci., 36, 580–590 (2009).
• 14)  Turner NG, Cullander C, Guy RH. Determination of the pH gradient across the stratum corneum. J. Investig. Dermatol. Symp. Proc., 3, 110–113 (1998).
• 15)  Wagner H, Kostka KH, Lehr CM, Schaefer UF. pH profiles in human skin: influence of two in vitro test systems for drug delivery testing. Eur. J. Pharm. Biopharm., 55, 57–65 (2003).
• 16)  Tojo K, Hikima T. Bioequivalence of marketed transdermal delivery systems for tulobuterol. Biol. Pharm. Bull., 30, 1576–1579 (2007).