3D distribution of endogenous skin componentsSpatial distributions of endogenous molecular components and active compounds were determined by 3D ToF–SIMS analysis of tape strip samples collected at different depths into SC (represented by consecutive tape strips acquired from the same skin area). Each tape strip sample comprise (ideally) a single corneocyte layer embedded in a thin layer of skin lipids, transferred onto a tape substrate by attaching and detaching the tape from a skin surface37 (Fig. 1a). Thus, initial ToF–SIMS analysis of the tape strip probes specifically the lipid matrix located on top of the corneocytes, due to the high surface specificity of this measurement (< 1 nm)38. Upon sputter erosion, the thin lipid layer is removed, allowing for analysis of the corneocyte bodies, followed by gradual removal of also the corneocyte bodies, thereby exposing the underlying lipid layer and tape substrate for analysis (Fig. 1a). Repeated 2D ToF–SIMS imaging measurements during the entire sputter erosion process provides 3D information (with the sample depth represented by sputter time), including (i) depth distributions of specific molecular species from depth profiles (Fig. 1b), (ii) 2D distributions of specific molecular species at different depths into the lipid/corneocyte layer, represented by ion images acquired after different sputter times (Fig. 1c), and (iii) molecular composition of separate lipid and corneocyte compartments, respectively, from mass spectra of selected 3D regions (Supplementary Fig. S1).Figure 13D ToF–SIMS and SEM analysis of tape strip sample. (a) Schematic representation of the measurement principle, including tape stripping, ToF–SIMS imaging analysis of the top lipid surface and 3D analysis by combined 2D analysis and sputter etching. The analysis area is smaller than the sputter area to ensure homogeneous etching rates. (b) Depth profiles of ions representing proteins, tape and skin lipids (C24:0, C18:1 and cholesteryl sulfate). A sputter time of 82 s corresponds approximately to 1015 ions/cm2. (c) ToF–SIMS ion images of the tape strip sample prior to (t = 0 s) and after (t = 180 s) 3D ToF–SIMS analysis, showing proteins in red, tape in blue and skin lipids (C24:0 + C26:0) in green. (d) SEM micrographs of corneocytes on the tape surface prior to (t = 0) and after (t = 156 s) 3D ToF–SIMS analysis (different areas). Note the residual proteins at the corneocyte periphery after 3D ToF–SIMS analysis. See Table 1 for ions used to represent the different components in (b,c). Scale bars in (c,d) are 10 µm.The different compartments of the tape strip sample are clearly discernible in the depth profiles (Fig. 1b), where the outer lipid phase is represented by initially high, but rapidly decreasing, intensities of the skin lipids (sputter time 0–5 s) and the underlying corneocyte region is represented by an extended, flat maximum of the protein intensity, accompanied by very low lipid intensities (20–46 s). At the end of the corneocyte region (at around 50 s), a simultaneous decrease in protein intensity and increase in tape intensity signals the start of a regime characterized by a gradually decreasing fraction of the sample surface being covered by corneocyte material, and an increasing fraction of the underlying tape substrate being exposed. Magnification of the vertical scale reveals maxima in the skin lipid profiles (Fig. 1b, right panel) at about 90–100 s for C24:0 and cholesteryl sulfate and at 60–70 s for C18:1, consistent with detection of the thin lipid matrix layer between the corneocytes and the tape substrate. The shifted maximum for C18:1, as compared to C24:0 and cholesteryl sulfate, is consistent with results of a previous study, in which the C18:1 distribution was found to be associated with cholesteryl oleate32. It should be pointed out, however, that it is not known whether these distributions are affected by the fact that ex vivo skin was used, which in contrast to in vivo skin is in equilibrium with regards to 3D molecular distributions.ToF–SIMS ion images and scanning electron microscopy (SEM) micrographs acquired after 3D ToF–SIMS analysis (Fig. 1c,d, Supplementary Fig. S2) demonstrates the presence of protein residues on the tape surface, even after very long sputter times. The protein residues are particularly prominent at the periphery region of the corneocytes, whereas the tape is exposed at the corneocyte centers, indicating complete removal of the corneocyte material. In addition to the relatively thick protein residues at the corneocyte periphery, thinner protein fibers (possibly bundles of keratin fibers) are also observed, occasionally crossing the corneocyte centers (Supplementary Fig. S2). Given a constant and homogeneous sputter erosion rate across the analysis area, the protein residues can be assumed to reflect thicker regions of protein material in the deposited corneocytes on the tape strip sample. Indeed, SEM and ToF–SIMS images acquired prior to sputter erosion (Fig. 1c,d) demonstrate a variable topography of the corneocyte surfaces and elevated protein structures at the corneocyte edges. The strong topography of the corneocyte surfaces may be rationalized by the fact that the corneocytes are completely dehydrated and largely collapsed on the tape surface in the vacuum environment during analysis. Furthermore, the apparent protein accumulation at the corneocyte periphery is consistent with observations of elevated concentrations of corneodesmosomes at the corneocyte periphery, as well as possible morphological variations of the cornified envelope12. In the depth profiles, the inhomogeneous thickness of the corneocytes is reflected in the gradual changes of the protein and tape intensities after conclusion of the protein plateau at 90–100 s and the broad maxima of the lipid intensities representing the lipid layer between the corneocytes and the tape (Fig. 1b), as variable sputter times are needed to remove the corneocyte material and expose the underlying lipid matrix and tape, across the analysis area.3D distribution of active compoundsThe four actives, caffeine, JAD, 2-MR and oxybenzone (Fig. 2a), were detected with high specificity in the tape strip samples using peaks corresponding to molecular ions (or, for caffeine, nearly molecular ions), which are clearly distinguished from the skin background in the mass spectra (Fig. 2b, Supplementary Fig. S3). The measured depth profiles of caffeine, 2-MR and oxybenzone revealed features that are very similar to the protein depth profiles, i.e., an initially increasing intensity to a maximum plateau followed by a gradual decrease, thus indicating localization primarily to the corneocytes for these actives (Fig. 2c). In contrast, JAD displayed a profile comprising a peak at the initial, lipid matrix-associated regime, followed by a flat or slowly increasing intensity in the corneocyte region and a final decrease corresponding to the gradual replacement of corneocyte residues with tape on the sample surface. Thus, the JAD profile provides clear evidence for localization of this active to both the lipid matrix and the corneocyte bodies.Figure 2Spatial distributions of caffeine, JAD, 2-MR and oxybenzone in SC. (a) Selected properties of the four actives, (b) negative ToF–SIMS spectra of the ions used to monitor the four actives, from analyses of tape strips #2, #4, #10 and from tape strip #2 of a skin sample not treated with the actives (“Blank”). Note that the spectra are individually normalized to show the peak intensities of the actives in relation to the skin background. (c) Depth profiles of the four actives in tape strip #2 (red curves). Note the different depth profile for JAD, displaying a peak in the lipid matrix region, as compared to the other three actives, which show depth profiles with similar features as the protein. The black symbols show the background signal obtained from the “Blank” sample. (d) Relative quantification of actives concentration versus depth in SC, i.e. TS2, TS4 and TS10, as determined from the ToF–SIMS depth profiles. Displayed intensities are mean values and error bars correspond to +/− 1 standard deviation (N = 3). (e) Quantification of active concentrations as determined by LC–MS/MS for tape strips from the same skin sample as analysed by ToF–SIMS. Bars display mean values with ± 1 sd (samples in duplicate).Whereas the depth profiles in Fig. 2c were acquired from the second consecutive tape strip (TS2), similar shapes of the depth profiles were obtained from TS4 and TS10, indicating that the partitioning between lipid matrix and corneocytes remains essentially unchanged with increasing depths into SC for all four actives (Supplementary Fig. S4). Furthermore, the relative concentrations of the active compounds were determined from the depth profiles of TS2, TS4 and TS10 by normalizing the accumulated signal intensities of the actives to those of the proteins (see “Methods” section), providing estimates of the depth distributions into SC. For all four actives, the results show considerable concentration reductions with increasing depth into SC (Fig. 2d). However, the depth distributions are less steep for caffeine and JAD, with smaller relative decrease between TS2 and TS4, as compared to 2-MR and oxybenzone. For verification and quantification, the concentrations of caffeine, 2-MR and oxybenzone in the intermediate tape strips from the same skin sample were determined by liquid chromatography coupled with tandem mass spectrometry (LC–MS/MS) (Fig. 2e). The results are in good agreement with the depth distributions obtained from the ToF–SIMS data (Fig. 2d) and, in addition, consistent with a previous study showing the classical decreasing profile for JAD as a function of depth in the SC (despite a different vehicle was used compared to the present work)19.A quantitative representation of the partitioning between lipid matrix and corneocytes was obtained for caffeine and JAD from the intensity ratio of the detected signal in the top lipid region of the depth profile (0–7 s) to that in the top layer also including the corneocyte region (0–62 s), see Fig. 3a. Here, the intensity ratio of the protein depth profiles serves as a reference corresponding to complete localization to the corneocyte, whereas higher values represent increasing fractions of the active being localized to the lipid matrix. Due to possible matrix effects, however, the intensity ratios between lipid matrix and corneocytes cannot be considered as quantitative concentration ratios, but they can be used to compare different samples/measurements. Roughly constant values of this “lipid phase fraction” were observed for JAD in TS2, TS4 and TS10, at about twice the values for proteins, thus confirming significant JAD localization to both lipid matrix and corneocytes and that the lipid/corneocyte partitioning is roughly unchanged as a function of depth into SC. For caffeine, the fraction that is detected in the lipid region of the depth profile is similar to that for proteins, consistent with localization mainly to the corneocytes. A possible trend towards increasing localization to the lipid region with increasing depth into SC is apparent for caffeine, but the significance of this trend is unclear.Figure 33D distributions of caffeine and JAD in lipid matrix versus corneocytes. (a) Fraction of actives detected in the top lipid region relative to the combined top lipid and corneocyte regions, estimated from the depth profiles as the ratio of measured intensities in the lipid region (t = 0–7 s) to total intensities in the lipid and corneocyte region (t = 0–62 s). The protein results represent the expected ratio for corneocyte localization, whereas higher values correspond to increasing fractional localization to the lipid matrix. Mean values and error bars corresponding to +/− 1 standard deviation (N = 3). (b) Depth profiles (from TS2) of JAD (red), protein (blue) and C18:1 (green), together with a fit to the JAD profile (black) given by a weighted sum of the protein profile and C18:1. (c) Depth profile of caffeine (same measurement as in (b)), where the black line is a fit using the protein depth profile only. (d) Ion images acquired from tape strip sample (TS2) after t = 60, 90 and 120 s of sputter erosion, showing proteins, tape, C18:1 and overlay images of caffeine (red) and JAD (green). Note the correlation between the distributions of C18:1 and JAD, whereas the distribution of caffeine is homogeneous across the protein-covered parts of the surface.A correlation was observed between JAD and the C18:1 fatty acid, in both the depth profiles and 2D ion images (Fig. 3b–d, Supplementary Figs. S5, S6), indicating a spatial association between these components in the skin structure. Firstly, it was found that the JAD depth profile in several cases could be largely reproduced by a weighted sum of the protein and C18:1 depth profiles (Fig. 3b), thus indicating a correlation of JAD with both C18:1 and proteins in the depth direction. In contrast, the caffeine profile fits well with the protein profile only (Fig. 3c). Secondly, ion images acquired in the depth region around the bottom of the corneocyte layer show the appearance of a thin lipid layer containing C18:1 between the corneocytes and the tape substrate and, furthermore, that elevated intensities of JAD are colocalized with C18:1 in this layer (Fig. 3d, Supplementary Figs. S5, S6). In contrast, caffeine displays a homogeneous distribution in areas with remaining corneocyte residues, i.e., areas with high protein intensities, consistent with localization mainly to the corneocytes. For JAD, the elevated intensities associated to C18:1 are superimposed on a homogeneous distribution in areas with remaining proteins/corneocytes (particularly evident in Supplementary Fig. S6), as expected for JAD localization to both the lipid matrix and corneocytes.Morphological details of the corneocyte structureDetailed inspection of ion images from the corneocyte/tape interface reveals additional details of the corneocyte structure and morphology (Fig. 3d, Supplementary Figs. S2, S5 and S6). At a sputter time of 60 s, the protein image is bright and largely homogeneous in the area covered by corneocytes, whereas the tape image is dark in this area, indicating that corneocyte material still covers the surface and that the sputter erosion has not yet reached the tape substrate. At 90 s, the underlying tape surface has partially appeared at the center of most corneocytes, accompanied by correspondingly darker areas in the protein image, indicating partial removal of the corneocyte material from the cell centers. At 120 s of sputter erosion, more of the proteins have been removed and more of the tape substrate is exposed at the centers of the corneocytes. However, the corneocyte periphery areas are still covered by protein residues, and thin fibers (about 1–2 µm diameter) crossing the corneocyte centers are clearly visible in the protein image. The presence of a thin C18:1-containing lipid layer below the corneocytes is evident in the C18:1 images, showing the appearance of bright areas at the corneocyte centers in parallel with corresponding dark and bright areas in the protein and tape images, respectively. With increasing sputter erosion, also the lipid layer is gradually removed, leaving the tape increasingly exposed on the surface. Furthermore, in the ion images acquired after 90 s sputter erosion, the presence of subcircular structures, one in each corneocyte is evident (Fig. 3d). The origin of these structures is as yet unknown, but the increased tape signal (and decreased protein signal) in these structures at t = 90 s indicate a reduced protein thickness, possibly caused by an elevated water content of this structure in the hydrated state.Corneocyte double layersIn most cases, the tape strip samples comprised a single corneocyte layer transferred onto the tape substrate. However, the 3D ToF–SIMS analysis occasionally revealed evidence for double corneocyte layers deposited on the tape substrate (Fig. 4). Firstly, the corneocyte region in the protein depth profiles (i.e., the constant plateau) is extended to about twice the sputter time for a single corneocyte layer, with a corresponding delayed onset of the increase in tape intensity (Fig. 4a). Given a constant sputter erosion rate, these depth profiles are consistent with sputter erosion through two consecutive corneocyte layers. In addition, the C24:0 lipid profile shows a clear maximum at the same sputter time as for a single corneocyte layer, i.e., at a sample depth corresponding to a lipid layer between two corneocyte layers, and then another (weaker) maximum at approximately twice the sputter time of the first, representing the lipid layer below the bottom corneocyte layer. The second C24:0 maximum is preceded by the onset of the increasing tape intensity, thus consistent with the lipid layer between the second corneocyte layer and the tape substrate. The C18:1 profile similarly shows two maxima, which both are equally shifted to shorter sputter times, consistent with previous results32. The caffeine profile surprisingly indicates a gradually decreasing concentration upon sputter erosion into and through the second corneocyte layer (Fig. 4b); an observation that, however, requires further verification and investigation.Figure 4ToF–SIMS data of double corneocyte layer in tape strip sample. (a) Depth profiles of proteins (black), tape (green), C24:0 (blue) and C18:1 (red) from two measurements of the same tape strip sample, attributed to a single corneocyte layer (crosses) and a double corneocyte layer (circles). Note the matching intensity maxima of the two C24:0 profiles at t ≈ 95 s, and the presence of another maximum for the double layer profile at approximately twice the sputter time, t ≈ 180 s. (b) Caffeine depth profiles of the same measurements as in (a), attributed to single (red crosses) and double (blue circles) corneocyte layers. (c) SEM micrograph of tape strip sample after 180 s sputter erosion, i.e., well over the sputter time needed to remove most of a single corneocyte layer. (d,e) SEM image as in (c) with superimposed ToF–SIMS images, showing proteins in red, tape in blue and SC lipids (C24:0 + C26:0) in green, of the same area (d) prior to sputter erosion (t = 0 s), and (e) after sputter erosion (t = 180 s).The occasional double corneocyte layers can also be observed in ion images acquired after increasing sputter erosion times (Fig. 4e, Supplementary Fig. S7). Superimposed SEM and ion images acquired after 180 s sputter erosion, i.e. long after removal of the first corneocyte layer and just at the onset of tape exposure below the second corneocyte layer (according to the depth profile in Fig. 4a), shows a distinct area with largely continuous protein residues (only minor tape spots exposed), whereas the adjacent area is mainly bare tape and isolated corneocyte edge structures (Fig. 4c,e). Ion images of the same analysis area acquired prior to start of the sputter erosion show the presence of lipid-covered corneocytes over a large part of the tape strip surface (Fig. 4d), thus indicating that the area with remaining protein residues after 180 s sputter erosion corresponds to a double corneocyte layer, whereas part of the adjacent area with mainly bare tape was originally covered by a single corneocyte layer. These observations indicate that the “ideal” view of tape stripping as a layer-by-layer process may not always be correct, with important consequences for quantification measurements based on this sampling method. Usually, quantification of the amount of SC removed by tape stripping is done by double weighing39 or alternatively by contrast image method40. In the absence of such measurements, the amount of SC removed by tape stripping is unknown and the resulting distribution profile within SC has to be considered with attention.
