2023 Volume 71 Issue 1 Pages 31-40
Intercellular lipids fill the interstices of corneocytes and serve a barrier function. The amount of transdermal water evaporation varies depending on the packing structure of intercellular lipids, as this structure is important for maintaining barrier efficacy. This packing structure consists of a mixture of crystals (orthorhombic and hexagonal) and liquid crystals (fluid phase), and the proportion of these phases is thought to affect barrier function. However, there have been no methods to visualize the actual distribution of the domains formed by packing structure in intercellular lipids. In this study, the planar distribution of intercellular lipid structures was determined using focal plane array (FPA)-based Fourier transform (FT) IR imaging analysis of stratum corneum cell units obtained by grid stripping. The lipid composition of ceramides was revealed by electrospray ionization tandem mass spectrometry (ESI-MS/MS)-based shotgun lipidomics. The distribution of domains formed by packing structures and the lipid composition of ceramides was compared in skin with high- or low-transepidermal water loss (TEWL). The orthorhombic proportion was lower in high-TEWL skin than in low-TEWL skin. ESI-MS/MS-based shotgun lipidomics analysis showed that the alpha-hydroxyceramide content in the low- and high-TEWL groups differed regarding the distribution of fatty acid chain lengths. The evaluation of stratum corneum cell units using FPA-based FTIR imaging is an innovative technology that can visualize the distribution of domains formed by intercellular lipid-packing structures. Increased proportions of alpha-hydroxyceramide subclasses such as alpha-hydroxy-sphingosine ceramide and alpha-hydroxy-phytosphingosine ceramide were associated with a reduced proportion of the orthorhombic packing structure domain.
Intercellular lipids in the stratum corneum (SC) protect the body from allergens and pathogens and inhibit water evaporation, thereby serving as a barrier. Intercellular lipids are composed of ceramides (CERs), cholesterol, and free fatty acids (FAs).1) Intercellular lipids exert their barrier function by forming a characteristic lamellar structure2) consisting of orthorhombic and hexagonal packing structures and liquid crystals.3,4) A study involving electron diffraction measurement reported different domain sizes of orthorhombic and hexagonal packing structures.5) Additionally, the proportion and distribution of the crystalline and liquid-crystalline structures are considered to affect barrier function. A relationship between the packing structure and transepidermal water loss (TEWL) value was proposed based on the phase transition behavior of intercellular lipids.6) To assess the skin barrier, it is important to clarify the microdomain distribution within the packing structures rather than the averaged information.
In the past, X-ray diffraction and electron diffraction were used to analyze the crystal structure of the packing structure, whereas the analysis of the liquid crystal was difficult because of the influence of keratin-derived diffraction.7,8) More recently, however, the packing structure of intercellular lipids, including liquid crystals, has been analyzed by Fourier transform IR spectroscopy (FTIR). Differences in packing structure were evaluated by confirming the shift of the peak derived from CH2 symmetric stretching vibration of the alkyl chains of intercellular lipids.4,9) Imaging analysis of the packing structure of SCs has been performed in vertical SC cross sections by transmission FTIR.10) In addition, Raman spectroscopy was used to evaluate the planar distribution of packing structures in intercellular lipid models.11–13) However, the horizontal distribution of the packing structure of intercellular lipids in the SC has not been examined.
Further clarification of the distribution and proportions of domains formed by orthorhombic and hexagonal packing structures and liquid crystals in the human SC will help identify the relationship between the proportion of each constituent and TEWL, an index of barrier function. This approach will contribute to the development of more effective skin care products that improve barrier function. In this study, we used focal plane array (FPA)-based FTIR imaging technology14) to analyze the shift of the CH2 symmetric stretching vibration-derived peak around 2850 cm−1 and thus calculate the distribution and proportions of domains formed by packing structures in SC cell units. The distribution was analyzed using human SC samples from high- and low-TEWL skin. The correlation between this distribution and the lipid composition of CERs was determined using electrospray ionization tandem mass spectrometry (ESI-MS/MS)-based shotgun lipidomics. Recently, the chain length distributions of FAs and sphingoid bases of CERs have been reported.15–18) However, the relationship between these distributions and TEWL remains unclear. We therefore also analyzed CER subclasses and their carbon chain lengths.
Methanol, propan-2-ol, chloroform, acetyl chloride, and ammonium acetate were of analytical grade. Deuterated CER[NS] (NS = nonhydroxy-sphingosine) D3 (36 : 1;2, [sum of carbon atoms in long-chain bases (LCBs) and FAs]: [sum of double bonds in LCBs and FAs]; [sum of hydroxyl groups in LCBs and FAs] (cat#2201) was purchased from Matreya LLC (State College, PA, U.S.A.). N-Tricosane (cat#91447, purity > 99.5%) were purchased from Supelco (Bellefonte, PA, U.S.A.). Calcium fluoride plates were provided by Japan High Tech (Fukuoka, Japan).
Skin SamplingAll sampling from human subjects was performed with informed consent in accordance with the Declaration of Helsinki after approval by the Bioethics Committee of Interface, Inc. (Akita, Japan) (decision number IF19-037). Exclusion criteria were as follows: current or prior skin diseases, current allergic diseases, current medication use, pregnancy, and exposure to UV radiation in a context other than daily activities. Human skin SCs were obtained from 36 healthy Japanese female donors between the ages of 25 and 69 years. Subjects washed the skin surface of the cheek area with cleansing gel and creamy soap. After washing, subjects spent at least 15 min in a thermo-hydrostatic chamber at 20 °C and 50% humidity. All SC samples were prepared as described in Sadowski et al.19) The SC sampling site for tape stripping was the middle of the left cheek. To prevent any influence by factors such as washing or cosmetics application, the uppermost SC layer was removed twice with D-Squame tape (D100, CuDerm corporation, Dallas, TX, U.S.A.). Then, a third stripping was performed, and the SC sample was used for lipid quantification analysis.
Copper grids (Thin Bar Grid, Gilder Grids Ltd., Grantham, U.K.) (cat#G2200C or cat#G2650C: 200 or 600 mesh) for electron microscopy coated with acrylic adhesive (POLYTHICK, Sanyo Chemical Industries, Ltd., Kyoto, Japan) were used to obtain SC cells with intact intercellular lipids. Grids with the adhesive applied only to the copper frame portion were made by applying a POLYTHICK/ethyl acetate (1 : 3) solution, absorbing the excess solvent with filter paper, and drying. To collect SC cell samples after three stripping sessions with D-squame tape, a grid was pressed on the sampling site for 5 s and removed to store the SC cells attached to the grid. SC cells within the cheek area of six subjects, described below, were collected according to the protocol. Five SC cells were analyzed per subject.
Transepidermal Water Loss Measurements to Determine Barrier FunctionTEWL was measured using a Vapometer (SWL5001JT, Delfin Technologies Ltd., Kuopio, Finland). Three measurements were performed at the center of the left cheek, in adjacent locations to prevent overlap.
Skin ConductanceSkin conductance was measured using a SKICON-200EX (Yayoi Co. Ltd., Tokyo, Japan). Five adjacent measurements were performed in the same location as the TEWL assessment. The lowest and highest values were excluded, and the remaining three values were used for statistical analysis.
FTIR Imaging MeasurementsFrom a total of 36 subjects, the five with the highest TEWL and the five with the lowest TEWL were selected. The three subjects with the highest TEWL and lowest SC water content in vivo and the three with the lowest TEWL and the highest SC water content in vivo were then chosen for analysis. Subject Nos. 11, 37, and 45 comprised the high-TEWL group, while Subject Nos. 18, 19, and 44 comprised the low-TEWL group. The domains formed by packing structure distribution of SC intercellular lipids in the grids was measured by FPA-based FTIR imaging. The instrument used was a Hyperion3000 FTIR microscope combined with a Vertex70 Fourier transform infrared spectrometer (Bruker Optik GmbH, Ettlingen, Germany). The microscope was equipped with a 36 × Cassegrain objective and a 64 × 64 mercury cadmium telluride FPA detector with 2 µm pixel resolution. SC cell samples with the widths of 30–80 µm were analyzed in transmission mode. The samples were scanned 64 times in the spectral range of 900–3900 cm−1 with a resolution of 4 cm−1, and the average spectrum was analyzed. The domains formed by packing structure was determined according to the shift of the peak position derived from CH2 symmetric stretching vibration, which was inferred from the center of gravity of the peak.20) FT-IR data were processed and analyzed using the OPUS 7.2 software suite (Bruker Optik GmbH).
First, we validated whether the phase transition of n-tricosane, which is well known to be characterized by thermotropic phase behavior,4,21–23) could be detected using the position of the CH2 symmetric stretching band. This alkane has been found to exist in four different phases: orthorhombic, face-centered orthorhombic (FCO) or pseudo-hexagonal, hexagonal, and liquid crystal, in the temperature range of 28–51 °C with phase transition temperatures of 40.4, 45.8, and 47.6 °C, respectively.23)n-Tricosane was measured under the same conditions as for SC cells. To collect spectra from n-tricosane, a small amount of powder was placed between the calcium fluoride plates (the thickness of n-tricosane was less than 2 µm). Prepared plates were set on an automatic heating stage (10036, Linkam Sci. Instr., Tadworth, U.K.). In the temperature studies, the n-tricosane plates were heated at the rate of 1 °C/min, and spectra were collected at 1 °C increments between 28 and 51 °C. The temperature was maintained while each spectrum was collected.
Figure 1 shows a bright-field image of SC cells in the grid and an image of the incident infrared light on the sample. The distribution of the domains formed by intercellular lipid packing structure was analyzed for five SC cells per subject.
(A) Bright field image of optical micrograph of SC cells attached to the frame of a copper grid for electron microscopy. The SC cell layer was stripped from the surface of human skin by the grid stripping method, then FPA-based FTIR imaging was performed. Bar indicates 20 µm. (B) Top-view image of the FPA detector. When irradiated with infrared light and detected by a 64 × 64 FPA detector with 128 µm per side, pixel information with a resolution of 2 µm is obtained. The distribution of the domains formed by packing structure of hydrocarbon chains of intercellular lipids on SC cells was identified. The red frame indicates the analysis area. (C) Side-view image when infrared light is incident on the sample.
The nonacyl-CER content of tape-stripped SCs from the high- and low-TEWL groups was examined by ESI-MS/MS-based shotgun lipidomics.19,24,25) Acyl-CERs were excluded because the ionization methods are different, and it is therefore difficult to compare the results to those of nonacyl-CERs. Samples for CER analyses were prepared from the third strip of tape. SCs were extracted from tape strips, then analyzed and quantified as follows, using the method described in Sadowski et al.19)
Samples were injected directly into a QExactive mass spectrometer (Thermo Fisher Scientific, Waltham, MA, U.S.A.) equipped with a TriVersa NanoMate ion source (Advion Biosciences, Ithaca, NY, U.S.A.), and immediately analyzed in both positive and negative ion modes with a resolving power of Rm/z = 200 = 280000 for MS and Rm/z = 200 = 17500 for MSMS. MSMS fragmentation was performed at 35% normalized collision energy and was triggered by an inclusion list covering the corresponding MS mass ranges.26) Combining both MS and MSMS data, all CER subclasses were monitored as acetate adducts. The previously reported precursor ions and confirmatory MSMS fragments19,24,25,27) are summarized in Supplementary Table S1.
An internal standard of 14 pmol CER[NS] D3 36 : 1; 2 (18 : 1; 2, 18 : 0; 0) per sample was used in this study because CER[NS] is not natively present in skin samples. Quantification was performed by normalizing the isotopically corrected intensity of the monoisotopic peak of each native species to the isotopically corrected intensity of the monoisotopic peak of the internal standard. The amounts of CER molecular species were calculated from the ratios of characteristic MSMS fragments, as described previously.26) Nonacyl-CERs were normalized by deuterated CER[NS].
CERs are divided according to subclass and chain length.28) Supplementary Figure S1 shows that differences between the 8 CER subclasses are due to varying combinations of two types of acyl chains (non-hydroxy and alpha-hydroxy) and the acid-amide bonds of four types of sphingoid bases (dehydrosphingosine, sphingosine, phytosphingosine, and 6-hydroxyl).27,29) Quantitative data were used to calculate the total amount of nonacyl-CERs (pmol/µg (the amount of protein)) and the percent of each CER amount relative to the total nonacyl-CER amount (%). A Squamescan 850 A (Heiland Electronic GmbH, Wetzlar, Germany) was used to determine the amount of protein in SCs. The amount of protein for correction was calculated based on the regression equation optical density = 0.623x + 2.703 (R = 0.85), following the method of Voegeli et al.30)
Lipid ExtractionA tape-stripped sample containing one stripping disc was placed in a 2 mL polypropylene tube, and 900 µL of methanol containing an internal standard was added for lipid extraction. Samples were shaken at 4 °C and 1400 rpm for 1 h. The extract was then transferred to a multi-well plate and dried in a speed vacuum concentrator. The dried extracts were resuspended in an acquisition mixture of 7.5 mM ammonium acetate in chloroform : methanol : propan-2-ol (1 : 2 : 4, V : V : V). All liquid handling steps were performed using a Hamilton Robotics STARlet robotic platform equipped with Anti Droplet Control for pipetting of organic solvents.
Data AnalysisWelch’s t-test with Bonferroni’s multiple comparison was used to analyze the difference in the proportions of domains formed by packing structures between the high- and low-TEWL groups. Welch’s t-test was used to analyze skin parameters, total amount of CER, and the average data from studies of packing structures. Two-way ANOVA was performed to analyze quantitative data on subclass distribution and chain length distribution of CERs. The significance level for rejection of the null hypothesis was set at p < 0.05.
Table 1 shows the skin parameters of the six selected subjects: Subject Nos. 11, 37, and 45 had high TEWL values (high-TEWL group) and Subject Nos. 18, 19, and 44 had low TEWL values (low-TEWL group). The skin of the high-TEWL group had a low moisture value while that of the low-TEWL group had a high moisture value. The skin parameter results of all subjects are shown in Supplementary Table S2.
| Group | High-TEWL | Low-TEWL | Means ± S.D. | p-Value (High-TEWL vs. Low-TEWL) | |||||
|---|---|---|---|---|---|---|---|---|---|
| Subject No. | No. 11 | No. 37 | No. 45 | No. 18 | No. 19 | No. 44 | High-TEWL | Low-TEWL | |
| Age | 24 | 45 | 25 | 63 | 69 | 48 | 31.3 ± 11.8 | 60.0 ± 10.8 | 0.036* |
| TEWL (g/m2/h) | 15.6 ± 0.5 | 21.6 ± 0.1 | 16.2 ± 0.2 | 7.4 ± 0.3 | 7.5 ± 0.1 | 7.5 ± 0.4 | 17.8 ± 3.3 | 7.5 ± 0.1 | 0.032* |
| Skin conductance (μS) | 85.0 ± 6.2 | 47.7 ± 7.5 | 91.0 ± 7.8 | 176.7 ± 4.6 | 189.0 ± 14.4 | 156.3 ± 21.4 | 74.6 ± 23.5 | 174.0 ± 16.5 | 0.005** |
| Total amount of nonacyl-CERs (pmol/µg) | 276 | 278 | 201 | 197 | 373 | 418 | 251.4 ± 25.1 | 329.3 ± 67.5 | 0.459 |
| Mean FTIR symmetric stretching peak position (cm−1) | 2849.4 ± 1.7 | 2850.0 ± 2.1 | 2849.7 ± 0.9 | 2848.5 ± 3.0 | 2849.1 ± 0.9 | 2848.7 ± 0.7 | 2849.7 ± 0.3 | 2848.8 ± 0.3 | 0.021* |
TEWL (mean ± standard deviation (S.D.)), skin conductance (mean ± S.D.), CER quantification results, averaged packing structure information, and the results of Welch’s t-test comparing the high- and low-TEWL groups are summarized. [*(p < 0.05), ** (p < 0.01)]
Supplementary Fig. S2 shows the results for the thermotropic behavior of the position of the CH2 symmetric stretching mode of n-tricosane. These phase transitions were clearly indicated by the temperature dependence of the magnitude of the position of the CH2 symmetric stretching vibration with respect to the orthorhombic-FCO and hexagonal-liquid crystal phase transitions, except for the FCO-hexagonal phase transition. Within the orthorhombic phase (28–40 °C), the CH2 symmetric stretching mode was centered at 2848.4 cm−1. Within the FCO and hexagonal phases (41–47 °C), the CH2 symmetric stretching mode was centered at 2850.3 cm−1. In the liquid crystal phase (48–51 °C), the CH2 symmetric stretching mode reached a plateau at 2852.3 cm−1.
Analysis of the Distribution of Domains Formed by Intercellular Lipid Packing Structure Based on FPA-Based FTIR ImagingFigure 2 shows a representative example of domains formed by packing structure distribution according to FPA-based FTIR imaging of SC cell samples. Based on the shift of the peak position due to the phase transition of n-tricosane, peak position values less than 2848.4 cm−1 were identified as mainly orthorhombic packing domain (OR), those between 2848.4 and 2850.3 cm−1 were identified as mainly orthorhombic and hexagonal packing domain (OR/HEX), those between 2850.3 and 2852.3 cm−1 were identified as mainly hexagonal and liquid crystal domain (HEX/LIQ), and those larger than 2852.3 cm−1 were identified as mainly liquid crystal domain (LIQ) with high fluidity. The deep blue areas show OR, the red areas show HEX/LIQ, and the white areas show LIQ. The distribution of packing structures was not uniform, resulting in domain formation. Typical IR spectra of OR, HEX, and LIQ are shown in Supplementary Fig. S3.
The upper row indicates the high-TEWL group: (A) subject No. 11, (B) subject No. 37, and (C) subject No. 45. The lower row indicates the low-TEWL group: (D) subject No. 18., (E) subject No. 19., and (F) subject No. 44. Based on the spectral information derived from intercellular lipids, the domains formed by packing structure is mapped by color according to the position shift information of the peak derived from the CH2 symmetric stretching vibration for approximately 70 spectra detected by the FPA detector. Each domains formed by packing structure is shown according to the color bar. The white domain indicates a region larger than 2852 cm−1. Bar indicates 5 µm. OR, mainly orthorhombic packing domain; OR/HEX, mainly orthorhombic and hexagonal packing domain; HEX/LIQ, mainly hexagonal and liquid crystal domain; LIQ, mainly liquid crystal domain.
Supplementary Table S3 shows the number of pixels analyzed in five independent SC cells per subject. FPA-based FTIR imaging was carried out to determine the distribution of domains formed by intercellular lipid packing structures. Figure 3 shows the proportions of domains formed by packing structures in each group. The average FTIR symmetric stretching peak positions calculated from all the pixels in each subject are shown in Table 1. The peaks were approximately 2849.7 cm−1 in the high-TEWL group and approximately 2848.8 cm−1 in the low-TEWL group. The mean value of the FTIR symmetric stretching peak positions differed significantly between the two groups.
The top and bottom three columns represent the high-TEWL and low-TEWL groups, respectively. Proportions are shown from left to right as OR, OR/HEX, HEX/LIQ, and LIQ. Five SC cell samples were collected from each subject by grid stripping, and the proportions of domains formed by packing structures were confirmed based on information from a total of about 350 spectra. OR, mainly orthorhombic packing domain; OR/HEX, mainly orthorhombic and hexagonal packing domain; HEX/LIQ, mainly hexagonal and liquid crystal packing domain; LIQ, mainly liquid crystal domain.
Figure 4 compares the peak proportion of each domain between the two groups. The number of pixels attributed to OR, counted at wavenumbers below 2848.4 cm−1, was significantly higher in the high-TEWL group than in the low-TEWL group. The number of pixels for OR/HEX, counted at wavenumbers between 2848.4 and 2850.3 cm−1, showed no significant difference between the two groups, but this domain had the largest proportion. The low-TEWL group had an OR proportion ranging from 3–8%, whereas the high-TEWL group had an OR proportion ranging from 13–26%. The maximum proportion of liquid crystals was estimated to be about 32% in high-TEWL skin and about 10% in low-TEWL skin. The proportion of domains containing LIQ in intercellular lipids is relatively small, and the crystalline phase, OR/HEX, constitutes the major domain at 60–80%.
SC cells containing intercellular lipids were collected from the high- and low-TEWL groups and analyzed by FPA-based FTIR imaging. Peak position values below 2848.4 cm−1 are classified as OR, those between 2848.4 and 2850.3 cm−1 as OR/HEX, those between 2850.3 and 2852.3 cm−1 as HEX/LIQ, and those above 2852.3 cm−1 as LIQ. From left to right, comparisons of the OR, OR/HEX, HEX/LIQ, and LIQ proportions are shown. Welch’s t-test results showed significant differences between groups for OR proportion. Data are presented as means ± S.D., n = 3. [* (p < 0.05)]. OR, mainly orthorhombic packing domain; HEX/LIQ, mainly hexagonal and liquid crystal packing domain; LIQ, mainly liquid crystal domain.
To evaluate differences in CER composition, the quantitative results for each CER subclass are shown in Fig. 5. Two-way ANOVA of CER composition in the high- and low-TEWL groups showed no significant difference. We compared the lipid composition of each CER subclass according to chain length. CERs consist of sphingoid chains and FA chains joined by amide bonds (Supplementary Fig. S1), and the lipid compositions of these components were determined. Table 2 shows the results of two-way ANOVA of CER chain length distribution between the high- and low-TEWL groups. Significant differences in C-chain distribution were observed for CER[NP] (NP = non-hydroxy-phytosphingosine), CER[ADS] (ADS = alpha-hydroxy-dihydrosphingosine), and CER[AP] (AP = alpha-hydroxy-phytosphingosine). Although not significantly different, a trend toward more CER[AS] (AS = alpha-hydroxy-sphingosine) was observed in the high-TEWL group. Significant differences in the S-chain were observed for CER[NP], CER[ADS], and CER[AP].
ESI/MS/MS-based shotgun lipidomics was used to analyze the lipid composition of CERs in SC samples collected from the two groups by tape stripping, and the quantitative results were compared among the nonacyl-CERs. The red columns show the mean values of the three subjects in the high-TEWL group, and the blue columns show the mean values of the three subjects in the low-TEWL group. Two-way ANOVA results showed no significant differences in the distribution of the quantitative values among subclasses. Data are presented as means ± standard error (S.E.), n = 3.
| CER subclass | p-Values | |
|---|---|---|
| C-chain | S-chain | |
| CER [NDS] | 0.475 | 0.547 |
| CER [NS] | 0.824 | 0.908 |
| CER [NP] | 0.041* | 0.026* |
| CER [NH] | 0.100 | 0.196 |
| CER [ADS] | 0.005** | 0.003** |
| CER [AS] | 0.050 | 0.165 |
| CER [AP] | 0.026* | <0.001** |
| CER [AH] | 0.491 | 0.545 |
The chain length distribution of the C-chain and S-chain of the eight subclasses of nonacyl-CERs were compared between the high- and low-TEWL groups by two-way ANOVA. [*(p < 0.05), ** (p < 0.01)]
Figure 6 shows the proportions of FA chains by chain length in CER subclasses containing alpha-hydroxy groups. The proportions of C16, C24, and C26 FA chains were higher than those of other molecular species in CER[AS], and were greater in the high-TEWL group. The proportion of C24 FA chains was higher than those of other molecular species in CER[AP], and was greater in the high-TEWL group. On the other hand, CER[ADS] was less abundant than other alpha-hydroxyceramide subclasses, but it showed a broad distribution of FA chains with high proportions of C24 to C27, including odd-numbered chains, and most proportions were lower in the high-TEWL group. The proportions of the contents of S-chain species by chain length are shown in Fig. 7. In the S-chain, the proportion of C18 FA chains that were not elongated was higher than the proportions of other molecular species in CER[AS] and CER[AP], whereas the proportion of C18 FA chains was increased in the high-TEWL group compared to the low-TEWL group. Although no significant difference in CER[AH] (AH = alpha-hydroxy-6-hydroxysphingosine) was observed between the high- and low-TEWL groups, the proportion of C18 FA chains was higher in the high-TEWL group. Figure 8 shows the proportions of FA chains by chain length in CER subclasses containing non-hydroxy groups. The proportions of C24 and C26 FA chains were higher than those of other molecular species in CER[NP], and were greater in the low-TEWL group. The proportions of the contents of S-chain species by chain length are shown in Fig. 9. In the S-chain, the proportions of C16, C17, and C18 FA chains that were not elongated were greater in the low-TEWL group.
The relative abundance of each CER to the total nonacyl-CERs (%) was quantified by ESI/MS/MS-based shotgun lipidomics. (A) CER[ADS], (B) CER[AS], (C) CER[AP], (D) CER[AH]. Data are presented as means ± S.D., n = 3.
The relative abundance of each CER to the total nonacyl-CERs (%) was quantified by ESI/MS/MS-based shotgun lipidomics. (A) CER[ADS], (B) CER[AS], (C) CER[AP], (D) CER[AH]. Data are presented as means ± S.D., n = 3.
The relative abundance of each CER to the total nonacyl-CERs (%) was quantified by ESI/MS/MS-based shotgun lipidomics. (A) CER[NDS], (B) CER[NS], (C) CER[NP], (D) CER[NH]. Data are presented as means ± S.D., n = 3.
The relative abundance of each CER to the total nonacyl-CERs (%) was quantified by ESI/MS/MS-based shotgun lipidomics. (A) CER[NDS], (B) CER[NS], (C) CER[NP], (D) CER[NH]. Data are presented as means ± S.D., n = 3.
This study is first to visualize the distribution of the domains formed by packing structure of intercellular lipids in SC cell units using FPA-based FTIR imaging, and it confirmed that the proportions of OR, OR/HEX, HEX/LIQ, and LIQ differed between the high- and low-TEWL groups. This measurement method accurately captured the phase transition behavior of n-tricosane, and the phase transition temperatures shown in Supplementary Fig. S2 were estimated to be about 41 °C for orthorhombic-FCO and about 48 °C for hexagonal-liquid crystal, consistent with previously reported values.4,21–23)
Conventionally, X-ray diffraction measurements have been used to evaluate the packing structure in terms of the hexagonal/orthorhombic ratio (RHexa/Orth).31–33) However, intercellular lipids contain not only orthorhombic and hexagonal packing structures, but also liquid crystals.3,4) Our FPA-based FTIR imaging could determine the LIQ proportion.
The FPA-based FTIR imaging method used in this study is based on more than 4000 spectra of 64 × 64 FPA elements. Furthermore, by collecting samples in SC cell units by grid stripping, it was possible to evaluate the distribution of domains formed by intercellular lipid packing structures in SC monolayers, which are easily penetrated by IR and are aligned horizontally. Our results revealed that the high-TEWL group had a significantly decreased OR proportion. The proportions of HEX/LIQ and LIQ tended to be higher in the high-TEWL group. The domain size of the packing structure has been determined using electron diffraction measurements.5) Our method made it possible to visualize the planar distribution of domains formed by packing structures consisting of a heterogeneous mixture of OR, OR/HEX, HEX/LIQ, and LIQ. No consistent trend was observed in the sizes of these domains. There was a heterogeneous planar distribution of domains formed by packing structures, suggesting the existence of defects in the barrier of the microdomain.
We revealed that the proportion of OR was lower in the domains formed by packing structure distribution near the surface of the SC in skin with high TEWL, and that this decreased proportion affected barrier function.
The reason for the decreased OR proportion in the high-TEWL group was evaluated in terms of the ratio of each CER class to the total nonacyl-CERs and the ratios of FAs or sphingoid bases by chain length for each alpha-hydroxyceramide subclass to the total nonacyl-CERs. The ratio of each CER class to the total nonacyl-CERs in this study (Fig. 5) was consistent with those in earlier studies.19,29) A high correlation has been reported between TEWL and the relative abundance of CER[AS].34) We revealed that not only CER[AS] but also the proportion of CER[AP] was increased in the high-TEWL group (Fig. 5). By contrast, a decrease in CER[NP] was observed in high-TEWL skin (Fig. 5), consistent with previous reports showing a negative correlation between TEWL and the relative abundance of CER[NP].34) The characteristics of the distribution of FA chain lengths in alpha-hydroxyceramides (CER[AS], CER[AP] and CER[AH]) were consistent with those of a study by Kawana et al.15) The change in the distribution of FA chain lengths in CER[ADS], CER[AP], and CER[AS] (Table 2) suggests that alpha-hydroxyceramide influences the domains formed by packing structure proportion. Similarly, the differences in the distribution of C- and S-chain lengths in CER[NP] (Table 2) suggests that these differences affect the proportion of each phase. Figure 8 shows that the C-chain length distribution of CER[NP] shows a characteristic decrease in the long C24 and C26 hydrocarbon chains in the high-TEWL group, which is thought to be one of the factors reducing hydrophobic interactions and thus lowering barrier function. The S-chain length distribution of CER[NP], shown in Fig. 9, indicates that there are more C16, C17, and C18 hydrocarbon chains with short hydroxyl groups in the high-TEWL group, which also reduces hydrophobic interactions and can impair barrier function.
In particular, the present study showed that alpha-hydroxyceramides with abundant C24 and C26 elongated FA chains in the C-chain were increased in the high-TEWL group (Fig. 6). We previously reported that the combination of short-chain CER[ADS]-C18/16 (sum of carbon atoms in LCB/FA) and short-chain free FAs results in a packing structure with low barrier function.35,36) This is thought to be due to the reduced hydrophobic interaction between CER[ADS] and short-chain free FAs.35,36) Short-chain free FAs were increased in SCs from atopic eczema patients,37) and it is assumed that short-chain free FAs are similarly increased in high-TEWL skin. Therefore, the high proportions of C24 and C26 FA chains in the CER[AS] and CER[AP] alpha-hydroxyceramide subclasses are prone to weaken the packing structure by reducing hydrophobic interactions. The results are also consistent with a study of lipid model membranes, which reported that alpha-hydroxylation of CERs (CER[AS] and CER[AP]) decreased resistance to water loss.38) On the other hand, the finding that the distribution of S-chain lengths in CER[ADS] and CER[AP] differed between the high- and low-TEWL groups (Table 2) suggests that the difference in the distribution of S-chain lengths affects the proportion of domains formed by packing structures. Figure 7 shows that in the chain length distribution of CER[ADS], there was generally a low proportion of long C22 saturated hydrocarbon chains in the high-TEWL group. When the S-chain is dihydrosphingosine, there is no 4,5 trans double bond in the sphingoid base, and free rotation between carbon atoms is not constrained. Furthermore, because there is no additional hydroxyl group at the C4 position, the interfacial area per head group is smaller than that of phytosphingosine CER,39) facilitating the formation of a more tightly packing structure, and the long C22 saturated hydrocarbon chain is thought to contribute to maintaining barrier function. By contrast, Fig. 7 shows that in the chain length distribution of CER[AP], there were generally more S-chains with long C22 hydroxyl groups in the high-TEWL group. Compared to sphingosine CER, CER[AP] is characterized by a larger interfacial area per head group and difficulty in forming a densely packing structure,39) and the decrease in hydrophobic interaction due to the difference in alkyl chain lengths between CER and free FAs may contribute to reduced barrier function.
It was suggested that the decreased OR proportion in the high-TEWL group was caused by change in the total CER content of CER[AS], CER[AP] and CER[NP], and by the changes in the distribution of their chain lengths.
The mean subject age was higher in the high-TEWL group than in the low-TEWL group (Table 1, Supplementary Table S2). Therefore, changes in CER chain length distribution may be influenced by age. Future studies should confirm whether these changes are still observed after excluding the effect of age.
Visualizing the distribution of the domains formed by packing structure and elucidating CER subclass alterations in intercellular lipid constituents will facilitate a detailed understanding of skin barrier function. This visualization technique is different from other visualization methods for the whole face, water content,40) and TEWL.41) Our method characterizes the distribution of the domains formed by packing structure at the microscopic level, and can obtain images of the lipid molecular structure in a single corneocyte. It is expected to be able to detect changes in the lipid molecular structure after topical application of cosmetics and pharmaceuticals that demonstrate high transdermal absorption or the ability to improve barrier function.
The distribution of the domains formed by packing structure in the SC was analyzed using FPA-based FTIR imaging. This method enabled us to determine the proportions of OR, OR/HEX, HEX/LIQ, and LIQ, which has not previously been achieved with conventional averaged peak positions. We also successfully visualized the distribution of domains formed by packing structures in skin SC, and showed that these structures were heterogeneous. The OR proportion was lower in the SC of the high-TEWL group, and this is thought to be associated with the increased proportion of particular alpha-hydroxyceramides in the SC.
This work was supported by JSPS KAKENHI Grant Number 20K08699. Support with SC sampling and skin measurement was provided by Interface Co., Ltd. (Akita, Japan). Support regarding quantitative lipid analysis was provided by Lipotype Gmbh (Dresden, Germany). We would also like to acknowledge Keiji Tokuda of Gunma Industrial Technology Center (Gunma, Japan) for their valuable insights during our discussions.
The authors declare no conflict of interest.
This article contains supplementary materials.
