2.1.

Skin Samples

We collected the ears of domestic pigs at the local slaughterhouse (Gland, Switzerland) $2–3\mathrm{h}$ after the animals were sacrificed. After thoroughly washing the ears with cold water, we removed the skin from the outer sides of the ears using a scalpel, clipped the hairs using hair clippers, and dermatomed the skin to a thickness of $350–400\mu \mathrm{m}$ using a $50\text{-}\mathrm{mm}$ electric dermatome (Nouvag AG, Goldach, Switzerland). The skin was stored frozen at $-20°\mathrm{C}$ (wrapped in Parafilm and packed in Ziploc®bags) for not longer than one month prior to use.

Before use, we defrosted the samples for $30\min$ at room temperature and wiped the SC surface two to three times with cotton swabs wetted with cold $(4°\mathrm{C})$ hexane, to remove traces of sebum.

2.2.

Model Cosmetic Formulations

We used two simplified, model cosmetic formulations: Eau de toilette (EdT) and cream. EdT was prepared as 9:1 (v/v) ethanol/water solution, and the cream as water-in-oil emulsion of polyglyceryl esters, diesters, glycerol, fatty acids, and triglycerides. The two formulations contained 1% (w/v) of the volatile fragrance chemical 4-undecanolide (4U). We chose this elevated concentration (10–100 times higher than the typical concentrations of fragrance ingredients used in EdT and cream) to facilitate the spectroscopic detection of the chemical.

2.3.

Sample Preparation

Before use, we equilibrated the skin samples clamped in static Franz diffusion cells for $2\mathrm{h}$ . The receptor compartments of the cells contained phosphate buffer solution ( $150\mathrm{mM}$ , pH 7.4) thermostated to maintain a temperature of $32°\mathrm{C}$ at the skin surface. We spread evenly the model formulations on the skin surface $(5\mu \mathrm{L}∕{\mathrm{cm}}^2)$ using a Hamilton syringe and incubated the samples for 0.5, 2, or $4\mathrm{h}$ without occlusion at room temperature. At the end of the incubation period, we removed the skin samples from the diffusion cells and wiped off the excess formulation using a lint-free tissue.

2.4.

Tape-Stripping

For this procedure, we used Scotch® packaging tape (3M) cut to size. After application of a piece of the tape onto the skin surface, we rolled a $2.8\text{-}\mathrm{kg}$ metal cylinder over the tape ten times to ensure the homogeneous adhesion of the tape on the skin. We removed the tape in one rapid movement and, if necessary, repeated the stripping procedure using new pieces of tape three or five times. For the imaging experiments, we used appropriately sized pieces of tape and skin (typically, with dimensions $5\times 5\mathrm{mm}$ ) cut using a razor blade. We performed all experiments in triplicate.

2.5.

FTIR Spectroscopy and Imaging

We performed the FTIR measurements using the setup described elsewhere^{31, 32} and a modified Golden Gate accessory (Specac Ltd.). We collected conventional FTIR spectra using a single-element mercury cadmium telluride (MCT) detector, and FTIR images—using a $64\times 64$ FPA detector. The ATR imaging spectrometer is patented by Varian.⁴³ We collected the spectra with spectral resolution of $4{\mathrm{cm}}^{-1}$ in the spectral range of $3950–900{\mathrm{cm}}^{-1}$ . To obtain a reasonable S/N ratio, we co-added and averaged 64 scans. The acquisition time was $\sim 90\mathrm{s}$ .

Unless otherwise indicated, we used the spectra without smoothing or baseline correction. We corrected the spectra of SC treated with cream (shown in Fig. 4) for the absorbance of the cream itself by subtracting the spectrum of pure cream from the combined spectrum until the spectral contribution of several vibrational bands characteristic for the cream—distinct from the bands originating from the skin and located in the fingerprint region—was minimized.

Fig. 4

Conformational order of SC lipids following topical treatment. (a–c) FTIR images of the integrated absorbance of the C–O stretching band $(1080–1015{\mathrm{cm}}^{-1})$ collected from porcine skin treated with cream for $2\mathrm{h}$ followed by (a) one, (b) three and (c) five consecutive tape strips (denoted as skin 1, skin 3, and skin 5, respectively). The scale of the integrated absorbance is the same for all images. Image dimensions: $500\times 700\mu \mathrm{m}$ . (d) IR spectra collected from untreated skin (thick line), extracted and averaged from the marked areas in the IR images (thin lines; a–c are plotted top to bottom), and collected from a cream sample (dotted line). The spectra extracted from the IR images were corrected for the absorbance of cream (see Experimental section for details). All spectra are baseline corrected, normalized to the maximal intensity of the $\mathrm{C}{\mathrm{H}}_2$ asymmetric stretching band, and offset for clarity. The vertical red lines indicate the positions of the $\mathrm{C}{\mathrm{H}}_2$ asymmetric and symmetric stretching bands in the spectrum of untreated skin. The positions of these bands in all spectra are tabulated next to the spectra.

3.1.

ATR-FTIR Imaging of Stripping Tapes

Figure 1 shows typical FTIR images of the fifth consecutive tape used for tape-stripping of SC following topical application of cream; Figure 1a maps the integrated absorbance of the Amide II band $(1565–1500{\mathrm{cm}}^{-1})$ , corresponding to the lateral distribution of protein in the sample, and Figure 1b maps the integrated absorbance of the C–O stretching band $(1080–1015{\mathrm{cm}}^{-1})$ , corresponding to the lateral distribution of the cream ingredients (e.g., glycerin, isopropyl myristate, polyglyceryl esters) in the sample. The polymer tapes and the SC lipids exhibited strong but overlapping $\mathrm{C}{\mathrm{H}}_2$ modes; consequently, in our imaging experiments we could not discern and map the distribution of SC lipids on tapes. The tapes, however, did not absorb in the typical for the SC proteins Amide I and II spectral regions; thus, in all FTIR images of tapes we used the distribution of the SC proteins (illustrated by the integrated absorbance of the Amide II band) to map the presence of skin flakes on the tapes.

Fig. 1

ATR-FTIR imaging and spectroscopy of stripping tapes. FTIR images of the integrated absorbance of the Amide II band [ $1565–1500{\mathrm{cm}}^{-1}$ , (a)] and the C–O stretching band [ $1080–1015{\mathrm{cm}}^{-1}$ (b)] collected from the fifth consecutive tape used for stripping of porcine skin treated with cream for $4\mathrm{h}$ . The scales of the integrated absorbance shown next to the images use the same color coding (red for high intensity and blue for low intensity) but have different absolute values. Image dimensions: $500\times 700\mu \mathrm{m}$ . (c) IR spectra extracted from the points labeled A, B, and C in the IR images (shown respectively in red, blue, and green). The spectrum shown in black (labeled MCT) is the average spectrum collected from the imaged area using a single-element MCT detector. All spectra are baseline corrected and normalized to the maximal intensity of the Amide II band.

The FTIR maps showed presence of both skin [Fig. 1a] and cream [Fig. 1b] on the tape, thus indicating that after four-hour topical application, a considerable amount of the cream has penetrated into the SC. As expected, the lateral distribution of the cream within the SC was highly inhomogeneous, ranging from areas containing large amounts (e.g., the points labeled A and B) to areas containing none (e.g., the point labeled C).

In Fig. 1c, we compared spectra extracted from these three different spots of the ATR-FTIR images to the average spectrum collected from the entire sampled area using a single-element detector. As expected from the chemical maps, the two spectra A and B (which were extracted from areas with high local concentration of cream) showed strong absorption around $1040{\mathrm{cm}}^{-1}$ , the C–O stretching band originating from cream, while in spectrum C (which was extracted from an area containing no cream) this band was entirely missing. In spectrum MCT (the average spectrum collected from the sampled area), the C–O stretching band was extremely weak and almost invisible in the background of the C–C stretching modes originating from the skin itself; thus, if we had examined the heterogeneous tape sample using only a single-element detector, we would have arrived at the erroneous conclusion that only a minor amount of cream has penetrated into the SC down to the depth probed by the fifth tape.

This comparison demonstrates that the use of FPA instead of single-element detection greatly increases the sensitivity of detection for samples with heterogeneous chemical composition,⁴⁴ an important consideration for the application of ATR-FTIR imaging to dermatological penetration studies.

3.2.

Time-Dependent Penetration of Non-volatile Chemicals into SC

Figure 2 shows typical FTIR images of skin tape-stripped five times after topical treatment with cream for 0.5 and $4\mathrm{h}$ . The images map the lateral distribution of proteins using the integrated absorbance of the Amide II band [Figs. 2a and 2b] and of cream using the integrated absorbance of the C–O stretching band [Figs. 2c and 2d]. The maximum integrated absorbance of the Amide II found in the two samples was very similar [Figs. 2a and 2b]. The amount of cream that had penetrated into the SC clearly showed the expected time dependence: both the area and the maximal absorbance of the cream-rich regions observed after $4\mathrm{h}$ treatment [Fig. 2d] were larger than those observed after $0.5\mathrm{h}$ treatment [Fig. 2c]. Thus, the combination of ATR-FTIR imaging and tape-stripping offers a simple and direct way to follow the penetration of chemicals into SC in semiquantitative, time-dependent studies in situ, avoiding the extraction of the stripping tapes needed in the traditional practice of tape-stripping.¹⁰

Fig. 2

Time-dependent skin penetration of non-volatile chemicals. FTIR images of the integrated absorbance of the Amide II band [ $1565–1500{\mathrm{cm}}^{-1}$ , (a, b)] and the C–O stretching band [ $1080–1015{\mathrm{cm}}^{-1}$ , (c, d)] collected from porcine skin treated with cream for $0.5\mathrm{h}$ (a, c) and $4\mathrm{h}$ (b, d) followed by five consecutive tape strips. The scales of the integrated absorbance shown next to the images use the same color coding (red for high intensity and blue for low intensity); the absolute values of the scale for (a) and (b) are different from those in the scale for (c) and (d). Image dimensions: $500\times 700\mu \mathrm{m}$ .

For the purposes of this work, we expressed the depth reached by the tape-stripping procedure within SC only in terms of number of tapes used for stripping of SC and not in terms of absolute distance below the SC surface. Although the present demonstration of principle can tolerate this approximation, it is important to keep in mind that the amount of SC material removed by one tape—and thus, the depth reached within the layer by tape-stripping—is not necessarily the same for all samples and depends on the relative depth within SC. In our future studies, we will quantify the amount of SC removed by each strip using, for example, gravimetric¹⁵ or spectrophotometric⁴⁵ methods.

In principle, the approach we demonstrated here is also suitable to investigate the penetration pathways of chemicals into SC (e.g., their colocalization with the SC lipids or proteins). In the example shown in Fig. 2, the limited lateral resolution of the macro-imaging setup $(\sim 15\mu \mathrm{m})$ precluded the possibility to map separately the SC proteins and lipids; these restrictions, however, can be easily surmounted using ATR-FTIR microscopy for imaging,²² where the better lateral resolution (approximately $3–4\mu \mathrm{m}$ ) could enable the resolution of lipid-rich and protein-rich regions.

3.3.

Formulation-Dependent Retention of Volatile Chemicals into the Topmost SC

Next, we compared the amounts of the volatile fragrance chemical 4-undecanolide (4U) that have been retained within the topmost SC layers after treating the skin surface with two model cosmetic formulations, EdT and cream, each containing 1% (w/v) of 4U. Fig. 3 shows representative ATR-FTIR images of SC following topical application of the formulations for $0.5\mathrm{h}$ and tape-stripping the surface once [Fig. 3a and 3c] and five times [Figs. 3b and 3d]. The images map the integrated absorbance between 1200 and $1150{\mathrm{cm}}^{-1}$ , corresponding to the C-O-C stretching mode of 4U [indicated by arrow in Fig. 3e]. Figure 3e compares spectra extracted from the ATR-FTIR images with the spectrum of pure 4U.

Fig. 3

Formulation-dependent retention of volatile chemicals in SC. (a–d) FTIR images of the integrated absorbance of the C–O–C stretching band $(1200–1150{\mathrm{cm}}^{-1})$ characteristic for 4U, collected from porcine skin treated with 4U-containing Eau de Toilette (a, b) and cream (c, d) for $0.5\mathrm{h}$ followed by one (a, c) or five (b, d) tape strips (denoted as skin 1 and skin 5, respectively). The color scale of the integrated absorbance is the same for all images. Image dimensions: $500\times 700\mu \mathrm{m}$ . (e) IR spectra extracted from the points marked in the IR images (labeled as the image from which they were extracted) and a spectrum from pure 4U. The arrow points to the C–O–C stretching vibration of 4U that was used to construct the images in a–d. The spectra are offset for clarity.

In the sample treated with EdT, we could not detect 4U either in the topmost [Fig. 3a] or in the deeper SC layers [Fig. 3b]. This result indicated that, within the limit of detection of the instrument, the volatile fragrance chemical has evaporated completely during the period of unoccluded application, and it had not penetrated into the SC. In the sample treated with cream, we detected the presence of 4U only in the topmost SC layer [Fig. 3c] but not in the deeper SC layers [i.e., those reached after five consecutive tape strips; Fig. 3d]. This result indicated that the cream had prevented the fast evaporation of 4U from the SC surface, most probably by an occlusive action. Even though the cream itself had penetrated into the deeper SC layers [see Fig. 2d] it did not, however, assist the penetration of the fragrance to this depth.

These experiments demonstrate that our strategy is well suited to address the relative effects of different formulations and application conditions on the balance between evaporation, retention, and penetration of volatile chemicals in the SC. It is, however, important to keep in mind that the utility of the methodology is subject to the sensitivity of detection of FTIR spectroscopy, which is inferior to the sensitivity that can be achieved using, for example, fluorescence or scintillation counting.

3.4.

Influence of Topically Applied Chemicals on the Conformational Order of SC Lipids

The combination of tape-stripping and FTIR imaging is applicable also to detect the conformational changes that topically applied products induce in the lipid lamellae of SC. The images in Fig. 4 show the distribution of cream in the SC layers progressively revealed after stripping cream-treated skin with one [Fig. 4a], three [Fig. 4b], and five [Fig. 4c] tapes; to build the images we used the integrated absorbance between $1080–1015{\mathrm{cm}}^{-1}$ . In Fig. 4d, we compared spectra extracted and averaged from the areas indicated on the images (after correcting them for the absorption of the cream itself; for details, see Experimental section) with an averaged spectrum collected from the surface of the SC before the application of cream and a spectrum collected from the cream sample.

The images in Figs. 4a, 4b, 4c showed that the cream had penetrated into all studied SC layers following topical application for $2\mathrm{h}$ , with the highest absorbance observed after three tape strips. Concomitantly, the positions of the $\mathrm{C}{\mathrm{H}}_2$ symmetric and asymmetric stretching bands in the spectra of the cream-treated skin differed from those observed in the spectrum of untreated skin [Fig. 4d]. The magnitude of the blueshifts we observed correlated well with the relative amount of cream found in the samples, with the biggest shifts—of $+4$ and $+2{\mathrm{cm}}^{-1}$ for the $\mathrm{C}{\mathrm{H}}_2$ asymmetric and symmetric stretching bands, respectively—found in the SC layer revealed after three tape strips. These results suggest that lipidic components from the cream got incorporated into the lipid matrix of the SC, which initially exhibited a high degree of conformational ordering as evidenced by the positions of the $\mathrm{C}{\mathrm{H}}_2$ stretching bands in this sample, and lead to its fluidization.

Thus, these experiments demonstrate that the combination of ATR-FTIR imaging and tape-stripping is well suited to study not only the presence of exogenous chemicals in the SC, but also the depth profile of the changes that they might inflict on the organization of the native SC lipids. Previous reports have shown that transmission FTIR imaging can be used to detect changes in the secondary structure of keratins in corneocytes isolated from stripping tapes.¹⁶ The use of FTIR imaging in ATR mode might bring the additional advantage of performing such studies in situ, thus avoiding the need for isolation of the SC components from the stripping tapes.