A High Throughput Method to Predict Skin Penetration and Screen Topical Formulations

Mar 1, 2013 | Contact Author | By: Konstantin Tsinman, PhD, Pion Inc.; B:aacute;lint Sinkó, PhD, SinkoLAB Scientific Bt
Your message has been sent.
(click to close)
Contact the Author
This item has been saved to your library.
View My Library
(click to close)
Save to My Library
Title: A High Throughput Method to Predict Skin Penetration and Screen Topical Formulations
skin permeabilityx high throughputx transdermal penetrationx artificial membranex transdermal formulationsx
  • Article
  • Media
  • Keywords/Abstract
  • Related Material

Keywords: skin permeability | high throughput | transdermal penetration | artificial membrane | transdermal formulations

Abstract: This work studies the applicability of a 96-well-based skin-mimetic artificial membrane permeability model to differentiate between topical pharmaceutical and cosmetic formulations. Results are compared with data obtained from in vitro Franz cell permeability measurements and reveal the applicability of this method to assess cosmetic formulation permeation and speed the discovery process.

View citation for this article

K Tsinman and B Sinkó, A high throughput method to predict skin penetration and screen topical formulations, Cosm & Toil 128(3) 192-199 (Mar 2013)

Market Data

  • Smoking is a major contributor to skin discoloration, ECM breakdown, deep wrinkling, premature skin aging, poor wound-healing and the formation of abnormal skin growths.
  • Telangiectasias—small, dilated blood vessels near the skin’s surface—are a response to the body's need for more oxygen, stifled by smoking.
  • Exposure to the pollutants present in cigarettes accelerates the degradation of the ECM.
view full article

Related Material

Market Intelligence

In past decades, the understanding of mechanisms for skin penetration has significantly improved.1 Studies have identified the stratum corneum (SC), the outermost layer of the skin, as the main barrier for penetration of pharmaceutical and cosmetic ingredients,2 and due to its unique composition—i.e., about 50% ceramides, 35% free fatty acids and 15% cholesterol—the SC differs from other biological membranes.3 To facilitate the development of transdermal delivery for the compounds of interest, estimations of skin penetration rates and comparisons of dermatological formulations have become crucial to pharmaceutical and cosmetics research.4

Several methods have been developed for the prediction of skin penetration. Although the most useful data is obtained from in vivo studies in humans, such measurements are expensive, labor-intense and slow to perform. This makes their application in lead selection and optimization impractical. In vivo measurements using rodents5 are well-known to overestimate the penetration rate since compounds can permeate down the hair follicle,6 and although pig ear skin is also used for in vitro tests based on its structural equivalence to human skin,7 the most generally accepted in vitro permeation models are based on human skin as the membrane.

Static and flow-through Franz diffusion cell systems are most commonly used8, 9 for the implementation of these in vitro models; however, measurements based on them are only partially standardized and a broad range of experimental setup and equipment conditions can cause significant intra- and inter-laboratory variations in permeability results.10–13 The Franz cell apparatus is also inefficient when a large number of active compounds and/or their formulations must be tested in parallel.

Considering these limitations, the authors developed a parallel artificial membrane permeability assay (PAMPA), which was introduced14 in the pharmaceutical drug discovery process as a low-cost method for high throughput measurements to predict the passive membrane permeability of molecules. This method uses a 96-well microtitre “sandwich” plate assembly to form 96, two-chamber cells used as donor/receiver compartments for permeability studies (see Figure 1).

Since the first iteration, more targeted PAMPA models have been developed, e.g., for the prediction of gastrointestinal absorption,15–17 modeling the blood-brain barrier,18 and recently, for predicting transdermal penetration.19 The latest transdermal model of the PAMPA, the Skin PAMPA, incorporates a mixture of synthetic analogs of ceramides, cholesterol and free fatty acids in a ratio mimicking the composition of the stratum corneum lipid matrix.19 Such a skin-mimetic composition was lacking in the earlier artificial skin models.20 The latest model was also optimized to correlate with highly standardized Franz diffusion cell permeability measurements through heat separated human epidermis.

The present work evaluates the ability of the Skin PAMPA19 to differentiate between topical formulations and compares the results with data obtained from in vitro Franz cell permeability measurements common in the pharmaceutical and cosmetic industries. The aim was to demonstrate the ability of this model to be easily standardized and provide high reproducibility yet remain more cost-effective and less labor-intensive than other current methods.

Materials and Methods

Chemicalsa: A novel silicone-based anhydrous gel formulation containing 5% ibuprofen (Formula A) and a viscous fluidlike test formulation containing silicone and an acrylic co-polymer (Formula B) were provided for the studyb. A commercial benchmark 5% ibuprofen formulation (Formula C) along with three available over-the-counter 1% formulations of diclofenacc–e also were used. Universal bufferf was adjusted to the required pH with 0.5 M NaOH, and spectroscopic- grade dimethyl sulfoxide (DMSO) also was obtainedg.

Method: A system designed for high-throughput permeability and membrane retention measurementsh was used for liquid handling, UV data collection and permeability data analysis. The 96-well PAMPA sandwichesj used are pre-coated with a skin-mimetic artificial membrane containing a mixture of certramide, i.e., synthesized ceramidelike analog,21 cholesterol and free fatty acids. These sandwiches are pre-loaded with stirring disks. UV platesk also were obtained for the study, and a stirring devicem from the company was used. Before the assay was performed, the membranes were hydrated overnight by immersion of the top (acceptor) compartment of the assembly into ionic strength pH 7.4-adjusted buffer. Hydration of the membrane ensures that it poses both hydrophobic and hydrophilic properties, which are essential to mimicking the stratum corneum barrier. Before forming the sandwich, the bottom (donor) plate was prefilled with a 180-μL aqueous solution of test compounds indomethacin, warfarin, piroxicam, niflumic acid, progesterone, carbamazepine, chlorpromazine and verapamil, dissolved in the bufferf at various pH values and containing 0.5–1% v/v DMSO. The acceptor plate was filled with 200 μL of the bufferf at pH 7.4. The resultant sandwich was incubated with stirring at room temperature for 5 hr.

After the permeation time, the PAMPA sandwich was separated and 150 μL of both the donor and acceptor compartments were transferred to UV plates. UV absorption (220–500 nm) was measured with a UV-Vis plate reader driven by softwareh. The model compound concentrations in the buffer were chosen according to their solubility and UV detection limits (50–300 μM). The data was measured with at least six replicates on each plate, and the experiment of every plate was repeated three to five times within a six month period. Permeability values were calculated taking into account mass balance as described previously:15

Eq. 1   

where Pe is the effective permeability coefficient (cm/s); A is the filter area (0.3 cm2); VD and VA are the volumes in the donor (0.18 cm3) and acceptor phase (0.2 cm3); t is the incubation time (s); ΤLAG is the steady-state time(s) estimated empirically; CD(t) is the concentration (mol cm-3) of the compound in the donor phase at time t; CD(0) is the concentration (mol cm-3) of the compound in the donor phase at time 0; R is the membrane retention factor, determined by:

Eq. 2   

and ra is the sink asymmetry ratio (gradient-pH-induced), defined as follows:

Eq. 3    

Concentrations of test samples present in the receiver compartment were measured by comparing the UV absorbance of the compounds measured in the acceptor wells after incubation with the UV absorbances of references at known concentrations. Permeability values could be estimated using the following formula:

Eq. 4    

Formulations: As noted, this PAMPA method was modified to study topical formulations, i.e. creams, gels, ointments, etc. In this study, formulations of ibuprofen and diclofenac were used. First, hydration of the PAMPA membrane was carried out as described. Formulations were then placed in the top compartment (30 ± 5 mg per well) of the 2-chamber PAMPA sandwich while the bottom compartment was filled with 180 µL pH 7.4 bufferf, serving as the receiver chamber (see Figure 2). Stirring in each well of the bottom (receiver) compartment was carried out by the previously mentioned devicem.The concentration of ibuprofen and diclofenac penetrating from formulations to the receiver compartment of the PAMPA sandwich was measured similarly to the method described above. Permeability values were not calculated in the case of formulations.

Franz cell measurements: In vitro Franz cell permeability experiments were conducted with the test formulations for 8 hr at 32°C. Heat-separated epidermis22–25 from two human donors was used. The receptor fluid was phosphate buffered saline of pH 7.4. Approximately 15 mg of the test formulations were homogeneously (visually) spread over a 0.63-cm2 permeation area. A commercial ibuprofen benchmark was evaluated at the same time. Triplicate cells were used for each formulation, and the ibuprofen concentration was measured using ultra-performance liquid chromatography.26

The cumulative amount of drug permeated through the skin was plotted as a function of time. The steady state of diffusion can be identified by the linearity of the concentration-time curve. The flux, J, corresponds to the slope, mol hr-1, of the curve. The permeability coefficient (Kp in cm/hr) can be calculated using the following expression:

Eq. 5    

in which A represents the actual diffusion surface area (0.63 cm2) and C0 (mol cm-3) corresponds to the initial concentration in the donor phase during the experiment.

Results and Discussion

Stability of pre-coated membranes: A set of pre-coated sandwich plate assemblies was stored at ambient temperature (23 ± 3°C) for a period of six months. Permeability for a set of eight test compounds, representing a broad spectrum of lipophilicity at pH 6.5, was measured after one week, one month, three months and six months following the preparation of the plates. Figure 3 shows the averaged effective permeability (Pe in 10-6 cm/sec) and standard deviations for the set. The standard deviation for the measurements was within 0.2 logarithm permeability units, which is consistent with historic data from other PAMPA models.16 This study demonstrated that Skin PAMPA assemblies can be stored for up to six months at room temperature without compromising permeability measurements.

Topical formulations: The ability to demonstrate that this method can also be applied for comparative testing of topical formulations was one of the main goals of this work. Non-steroidal anti-inflammatory drug ibuprofen was therefore selected as a model compound for the study, as described earlier. The concentrations of ibuprofen permeating into the receiver compartment were measured at 30 min, 60 min and 150 min by means of a UV/VIS plate reader. Figure 4 compares the UV spectra for ibuprofen in the receiver compartment after having penetrated through the PAMPA barrier from the different formulations after 60 min. The spectral influence of the formulations was subtracted by using blank wells with formulations having no active pharmaceutical ingredient (API).

Even a quick visual comparison indicates that Formula A provided the highest cumulative amount of ibuprofen, compared to the other formulations or to the slurry. Quantitative analysis indicated that the amount of ibuprofen permeated from a silicon-based formulation (Formula A) was about three times the amount that permeated from the commercial benchmark at corresponding time points. These results were compared to the data for Formulas A and C measured using Franz cells and human epidermis as a barrier.26 Figure 5 demonstrates that Skin PAMPA results for ranking order of Formula A and Formula C agree with the conclusions obtained from ibuprofen flux measurements through human epidermis.

Figure 6 illustrates the ability of Skin PAMPA to differentiate between three commercially available formulations of diclofenac. These examples of successful application of Skin PAMPA to study the permeability of API from topical formulations indicate that this method could also be applied to investigate the penetration of cosmetic actives from formulations.


The Skin PAMPA test model described was developed for the prediction of skin penetration in the early stages of drug discovery. Such a model could also be used for topical formulation testing in cosmetic research. Although the artificial membranes include no active biological elements such as skin metabolism, the system is skin-mimetic since it contains components similar to human SC, i.e., the penetration rate-limiting barriers in skin. The standardization potential and high-throughput nature of this approach can be a valuable, cost-effective complement to Franz cell permeability experiments for early skin penetration prediction, topical formulation ranking, and for understanding the effect of each formulation component on penetration phenomena.


Send e-mail to ktsinman@pion-inc.com.

  1. J Hadgraft, Skin deep, Eur J Pharm Biopharm 58 291–299 (2004)
  2. J Hadgraft and ME Lane, Skin permeation: The years of enlightenment, Int J Pharm Sci 305 2–12 (2005)
  3. MEMJ Meuwissen, J Janssen, C Cullander, HE Junginger and JA Bouwstra, A cross-section device to improve visualization of fluorescent probe penetration into the skin by confocal laser scanning microscopy, Pharm Res 5 352–356 (1998)
  4. JA Bouwstra, PL Honeywell-Nguyen, GS Gooris and M Ponec, Structure of the skin and its modulation by vesicular formulations, Prog In Lip Res 42 1–36 (2003)
  5. RP Zendzian, Dermal absorption of pesticides in the rat, Am Ind Hyg Ass J 61 473–483 (2000)
  6. H Todo et al, Permeation pathway of macromolecules and nanospheres through skin, Biol Pharm Bull 33 1394–1399 (2010)
  7. C Salerno, AM Carlucci and C Bregni, Study of in vitro drug release and percutaneous absorption of fluconazole from topical dosage forms, AAPS PharmSciTech 11 986–993 (2010)
  8. TJ Franz, Percutaneous absorption. On the relevance of in vitro data, J Invest Dermatol 64 190–195 (1975)
  9. TJ Franz, The finite dose technique as a valid in vitro model for the study of percutaneous absorption, Curr Probl Dermatol 7 58–68 (1978)
  10. G Khan, Y Frum, O Sarheed, GM Eccleston and VM Meidan, Assessment of drug permeability distribution in two different model skins, Int J Pharm 303 81–87 (2005)
  11. Y Frum, GM Eccleston and VM Meidan, Evidence that drug flux through synthetic membranes is associated with normally distributed permeability coefficients, Eur J Pharm Biopharm 67 434–439 (2007)
  12. R Chilcott et al, Inter- and intralaboratory variation of in vitro diffusion cell measurements: An international multicenter study using quasi-standardized methods and materials, J Pharm Sci 94 632–638 (2005)
  13. S Ng, JJ Rouse, FD Sanderson, V Meidan and GM Eccleston, Validation of a static Franz diffusion cell system for in vitro permeation studies, AAPS PharmSciTech 11 1432–1441 (2010)
  14. M Kansy, F Senner and K Gubernator, Physicochemical high throughput screening: Parallel artificial membrane permeability assay in the description of passive absorption processes, J Med Chem 41 1007–1010 (1998)
  15. A Avdeef, Absorption and Drug Development, ch 7, Permeability, John Wiley & Sons, Hoboken, NJ USA (2003)
  16. A Avdeef et al, PAMPA: Critical factors for better prediction of absorption, J Pharm Sci 96 2893–2909 (2007)
  17. A Avdeef and O Tsinman, PAMPA—A drug absorption in vitro model: 13. Chemical selectivity due to membrane hydrogen bonding: In combo comparison of HDM-, DOPC- and DS-PAMPA models, Eur J Pharm Sci 28 43–50 (2006)
  18. O Tsinman, K Tsinman, N Sun and A Avdeef, Physicochemical selectivity of the BBB microenvironment governing passive diffusion-matching with a porcine brain lipid extract artificial membrane permeability model, Pharm Res 28 337–363 (2011)
  19. B Sinkó at al, Skin PAMPA: A new method for fast prediction of skin penetration, Eur J Pharm Sci, 45(5) 698–707 (2012).
  20. G Ottaviani, S Martel and PA Carrupt, Parallel artificial membrane permeability assay: A new membrane for the fast prediction of passive human skin permeability, J Med Chem 49 3948–3954 (2006)
  21. B Sinkó, M Pálfi, Sz Béni, J Kökösi and K Takács-Novák, Synthesis and characterization of long-chain tartaric acid diamides as novel ceramide-like compounds, Molec 15 824–833 (2010)
  22. S Prodduturi et al, Transdermal delivery of fentanyl from matrix and reservoir systems: Effect of heat and compromised skin, J Pharm Sci 99(5) 2357 (2010)
  23. D Bucks, Skin penetration studies: Rationale and design of studies to screen and identify promising topical formulations, Drug Delivery Tech 8(8) 44–47 (2008)
  24. RL Bronaugh, RF Stewart and M Simon, Methods for in-vitro percutaneous absorption studies. VII. Use of excised human skin, J Pharm Sci 75 1094 (1986)
  25. RJ Scheuplein, Mechanism of percutaneous absorption. I. Routes of penetration and the influence of solubility, J Invest Dermatol 45 334–346 (1965)
  26. K Tsinman et al, Application of skin PAMPA to differentiate between topical pharmaceutical formulations of ibuprofen, poster presented at AAPS 2012 Annual Meeting and Exposition, Chicago IL, (Oct 14–18, 2012)

This content is adapted from an article in GCI Magazine. The original version can be found here.



Figure 1. Schematic view of the PAMPA sandwich

Figure 1. Schematic view of the PAMPA sandwich

This method uses a 96-well microtitre “sandwich” plate assembly to form 96, two-chamber cells used as donor/receiver compartments for permeability studies.

Figure 2. Adapting the Skin PAMPA method to formulation testing

Figure 2. Adapting the Skin PAMPA method to formulation testing

Formulations were placed in the top compartment (30 ± 5 mg per well) of the 2-chamber PAMPA sandwich while the bottom compartment was filled with 180 µL pH 7.4 buffer, serving as the receiver chambe.

Figure 3. Permeability values and standard deviations for the eight test compounds

Figure 3. Permeability values and standard deviations for the eight test compounds

Averaged effective permeability (Pe in 10-6 cm/sec) and standard deviations for the set; the standard deviation for the measurements was within 0.2 logarithm permeability units, which is consistent with historic data from other PAMPA models.

Figure 4. UV spectrum of ibuprofen in different receiver compartments of the skin PAMPA assembly

Figure 4. UV spectrum of ibuprofen in different receiver compartments of the skin PAMPA assembly

UV spectrum of ibuprofen in different wells of the receiver compartment of the skin PAMPA plate assembly after 60 min of incubation; top left, top right, bottom left and bottom right graphs correspond to Formulas A, B, C and slurry formulations in donor wells (n ≥ 12 for each form). Note: the same absorbance scale for all forms eases the comparison.

Figure 5. Concentration of ibuprofen measured

Figure 5. Concentration of ibuprofen measured

Concentration of ibuprofen measured at 30 min and 60 min time points in the PAMPA assay (left Y-axis), and averaged flux measured at 1 hr time point of Franz cell permeability assay using human epidermis (right Y-axis)

Figure 6. Amount of diclofenac per unit area that permeated

Figure 6. Amount of diclofenac per unit area that permeated

Amount of diclofenac per unit area permeating from the test formulas at 20 min, 40 min and 60 min; Formula A included diclofenac c; Formula B, diclofenac d; and Formula C, diclofenac e (see Footnotes for c, d, e); 1% w/v of diclofenac was used for comparison.

Footnotes [Tsinman 128(3)]

a All chemical compounds used for the described studies were purchased from Sigma-Aldrich.
b The silicone based anhydrous gel formulation was provided by Dow Corning Corp.
c Voltaren is a product of Novartis.
d Diclofenac-ratiopharm is a product of SkyePharma.
e Flector is a product of Pfizer.
f The Prisma HT buffer used is a product of Pion Inc.
g Spectroscopic grade DMSO was obtained from Burdick and Jackson.
h The PAMPA Evolution system was developed by Pion Inc.
j The 96-well Stirwell PAMPA sandwiches are manufactured by Pion Inc.
k The UV plates were from Greiner Bio-one(UV-star micro plate, clear, flat bottom, half area)
m The Gut‑Box stirring device is manufactured by Pion Inc.

Next image >