Abstract
The stratum corneum, the outermost layer of the epidermis, is the most important skin barrier against exogenous physical and chemical effects, in addition to protecting against dehydration. Ceramides are integral parts of the intercellular lipid lamellae of the stratum corneum and play an important role in the barrier function of mammalian skin. Ceramides are sphingolipids consisting of sphingoid bases linked to fatty acids by an amide bond. Typical sphingoid bases in the skin are composed of dihydrosphingosine, sphingosine, phytosphingosine, and 6-hydroxysphingosine, and the fatty acid acyl chains are composed of non-hydroxy fatty acid, α-hydroxy fatty acid, ω-hydroxy fatty acid, and esterified ω-hydroxy fatty acid. Analytical methods, such as gas chromatography/mass spectrometry, high performance thin layer chromatography with UV detection, and liquid chromatography/mass spectrometry, have been developed for the identification and quantification of ceramides in the stratum corneum. However, only a few publications relate to the mass fragmentation patterns specific to ceramide types to determine the structure of skin ceramides. Moreover, these studies provide very limited structural information and only for some ceramides. Therefore, the aim of our study was to develop a quick and easy method of quantification of ceramides, cholesterol, and free fatty acids by high performance thin layer chromatography with ultraviolet detection. High performance thin layer chromatography with ultraviolet detection was also coupled with mass spectrometry using negative ionization by electrospray and tandem mass spectrometry (MS/MS) for identification of ceramides’ structure.
Keywords
Introduction
The stratum corneum (SC), the outermost layer of the epidermis, is the most important skin barrier against exogenous physical and chemical effects, in addition to protecting against dehydration. Penetration of chemicals through the SC is generally considered to be the key step that limits percutaneous absorption. The ‘brick and mortar’ structure of the SC consists of corneocytes (the bricks) embedded in a lipid matrix (the mortar). The lipid matrix plays a key role on the integrity of the SC to maintain skin barrier function.1,2 The main lipid classes are ceramides (CERs), free fatty acids (FFAs), and cholesterol (CHOL).3,4 CERs are integral parts of the intercellular lipid lamellae of the SC and play an important role in the barrier function in the mammalian skin. Compared with other human cells/tissues, the human SC has extremely complex CERs. These consist of fatty acid moieties and sphingoid moieties, whereby both the acyl and the sphingosine carbon chains vary in length, resulting in 342–1000 structurally different SC CERs. 5 According to the nomenclature by Motta et al., 6 the name and subclass of CERs are based on the sphingosine and acyl chains molecular structure. These include sphingosine (S), phytosphingosine (P), 6-hydroxysphingosine (H), and dihydrosphingosine (dS). The SC FFAs can be non-hydroxylated (N), α-hydroxylated (A), or ω-hydroxylated (O), of which the latter can be esterified to a fatty acid (E). Fifteen different subclasses of SC CERs have been identified so far. Among the four types of fatty acid bases, the N-type CER is the most abundant in human SC.
A number of analytical methods have been developed for the identification and quantification of CERs in the SC. Gas chromatography is rarely used for the analysis of CERs, as these lipids are non-volatile and unstable in gas phase if not derivatized. 7 By contrast, liquid chromatography/mass spectrometry (LC/MS) enables a direct analysis without derivatization. MS methodologies have been developed for the qualitative and quantitative analyses of CERs. 8 However, several important issues in the quantitation of CERs, such as ionization efficiency of subclasses (H- and N-type) and molecular species of CERs, make quantification difficult. Thin layer chromatography (TLC) is still the method of choice for separation and quantification purposes9,10 but this method does not allow the identification of the CERs except by comparison of the migration front (Rf) with known and available standards.
When analysed by MS, CERs are usually run in the positive mode.11,12 This method is very sensitive and enables structural characterization by tandem mass spectrometry (MS/MS) under low collision energy range of their molecular species as protonated and lithiated forms (i.e. [M+H]+ or [M+Li]+ ions). Although usually less sensitive, the negative mode can appear more suitable for CER analyses in certain experimental conditions, especially in the presence of acetic acid (or formic acid) during their high performance thin layer chromatography (HPTLC) separation.9,13–15 In the negative mode, CERs are detected as adduct [M+CH3CO2]− ions, which should be dissociated into [M−H]− to provide structural information. 16 Several studies reported structures of fragment ions generated from [M−H]− ions of CERs by low energy collisionally induced dissociation (CID), using quadrupole tandem MS and electrospray ionization.13,14,17,18 Recently, Shin et al. 19 suggested some fragmentation mechanisms of N-type CERs, including NS and NdS. However, only one mechanism was based on a fragmentation process driven by the charge, which should be a unique process occurring under low energy conditions where only vibrational energy in ground state is considered. However, coppered deprotonated fatty acid ions activated under low energy conditions yield coordinate copper cation reduction, yielding radical raising followed by radical promoted fragmentations. 20 Charge remote processes should be produced under very fast energy transfer conditions. Hsu and Turk 16 reported extensive work on the fragmentation of NS CER. All the proposed fragmentation pathways involve the charge-driven fragmentation (CDF) process based on CID experiments on deuterium-labelled ion species obtained by H/D exchanges. Their mechanisms were confirmed by source CID-MS 2 (‘pseudo’ MS3) experiments based on the assumption that the preliminary fragmentation process was a loss of HCHO or H2+HCHO, generating the consecutive formation of the fatty acid-related fragment ions. Our product ion spectrum of the NS CER anion did not display all the fragment ions observed by Hsu and Turk, 16 which suggests that other preferred fragmentation processes could be involved in the dissociation of NS.
The aim of our study was to develop a quick and easy method to quantify CERs, CHOL, and FFAs using high performance thin layer chromatography with UV detection (HPTLC/UV). HPTLC/UV was also coupled with MS using negative ionization by electrospray and MS/MS for identification of CER structure. Complementary to previously published gas-phase fragmentation pathways of CERs, we propose alternative fragmentation pathways. These involve the CDF process, generation of directly fatty acid-related fragment ions and long-chain base-related fragment ions via stepwise pathways involving deprotonated CER isomerization into an ion–dipole complex prior to dissociation. 21 Based on the suggested mechanisms of the NS CER, we applied the same fragmentations pathways to explain the fragment ions observed for NP, AP, and AdS CERs.
Material and methods
Chemicals
MS analyses were conducted using HPLC grade methanol from Fisher Scientific (Thermo Fisher Scientific, Illkirch, France) and ultra-pure water generated from a Milli-Q system (Millipore, Saint Quentin en Yvelines, France). CHOL, cholesteryl oleate, sodium cholesterol sulphate, squalene, and palmitic acid as FFA were purchased from Sigma Aldrich (St Quentin Fallavier, France). Cer AP and NP were obtained from Evonik (Essen, Germany). Other CERs standards were purchased from Avanti Lipids (Alabaster, USA) and Sigma Aldrich (St Quentin Fallavier, France). HPTLC solvents (chloroform, hexane, acetone, acetic acid, and methanol) were purchased from Sigma Aldrich (St Quentin Fallavier, France).
SC preparation and lipids extraction
Human skin from the abdominal region of Caucasian female donors (43 ± 7 years old) with no skin diseases was obtained from surgery. All human tissue was obtained with donor consent for research from two commercial suppliers, Rangueil hospital and L’union clinic (Toulouse, France), according to certified ethical procedures. Isolation of SC was performed directly after receiving the skin tissue. The SC was isolated from the viable part of the skin using a classical trypsin digestion method. 22 The SC was dried under silica and stored before lipids extraction. The SC lipids were extracted according to the Bligh and Dyer procedure, 23 using a series of chloroform:methanol mixtures (1:2 and 1:1 v/v), each incubated for 1 h. The extracts were collected and evaporated under a stream of nitrogen at 40℃. The obtained lipids were redissolved in an appropriate volume of chloroform:methanol (2:1 v/v) and stored at −20℃ until lipids analysis. SC from 25 donors were pooled in order to produce a reference material that covers inter-individual variability.
CERs separation and quantification by HPTLC/UV
Extracted SC lipids were quantified using HPTLC/UV. The HPTLC plates were Lichrospher HPTLC Silica gel 60 F254S (Merck KGaA, Darmstadt, Germany). Automated sample application was carried out using an Automatic TLC Sampler 4 (CAMAG, Muttenz, Switzerland). The development of the plates was also automated, using an AMD-2 apparatus (CAMAG, Muttenz, Switzerland). The AMD procedure included 11 steps and was based on the use of mixtures of chloroform, acetone, and methanol/water under acidic conditions to separate the CER classes and the other lipids (Figure 1).
Optimized gradient steps used for HPTLC. The percentage of each solvent used in AMD is shown.
Quantification of CER spots was performed after staining with copper acetate and copper sulphate in phosphoric acid and charring at 150℃ using a TLC Scanner 3 (CAMAG, Muttenz, Switzerland). Measurements were performed in absorbance mode at a wavelength of 450 nm. Integration and quantification based on peak areas were performed using CATS software (CAMAG, Muttenz, Switzerland). Quantitative results for all CERs were related to CER AS as a standard which was applied at a concentration of 50 ng/µl. The results were expressed as mean ± SD of three replicates. A volume of 1 µl of SC skin extract was spotted on the TLC plate for the quantification of CHOL and FFAs. For the quantification of CERs, 8 µl of SC skin extract was spotted on the TLC plate. Instrumental precision was measured using n = 6 replicates of the five different amounts of CER AS used for the calibration curve (60, 500, 1000, 2000, and 3000 ng/spot). Inter-day assay precision was evaluated by analysis of n = 6 replicates of freshly prepared standard solution of same concentration (60–3000 ng/spot) on three different days. The repeatability of sample application and measurement of peak area were expressed as a % CV.
Mass spectrometry
MS analysis was carried out using a TSQ Quantum triple quadrupole mass spectrometer from Thermo Scientific (Les Ulis, France) equipped with an electrospray ionization source operating in the negative mode. Parameters of ionization were set as follows: spray voltage 4 kV, sheath gas flow rate 20 arbitrary units (a.u.), auxiliary gas flow rate 15 a.u., sweep gas flow rate 8 a.u., and capillary temperature 380℃. MS/MS spectra were acquired with a collision energy of 30 eV, using argon as collision gas at a pressure of 1.5 milliTorr, a scan time of 1 s, and a Q1 peak width (FWHM) of 0.7 m/z. HPTLC spots were eluted using a TLC–MS interface (CAMAG, Muttenz, Switzerland) and using methanol at a flow rate of 100 µl/min during 2 min with a Surveyor MS pump from Thermo Scientific (Les Ulis, France).
Results and discussion
Lipid separation by HPTLC/UV
FFAs, CHOL, and 10 CERs subclasses were separated using a TLC plate. The separation of the standard lipid mixture and skin SC lipid samples was achieved using an 11-step gradient. This method is based on gradient steps optimized by Farwanah et al.,
24
with a reduced number of steps to be more time efficient, while still achieving the same level of separation as that of others.
25
We focused on FFAs and CHOL to obtain a better separation of CER subclasses.24,25 Some CERs could not be separated using this protocol, e.g. CER NdS and CER NS and CER AS and EOH (Figure 2). Therefore, to enable CER analysis by MS, acetic acid was used for preconditioning rather than the eluent phase in order to avoid sodium adducts formation during electrospray ionization.
Separation of FFAs, CHOL, and 10 CERs subclasses in a skin sample using HPTLC/UV. Each CER subclass is denoted by a number.
One advantage of the HPTLC/UV technique is that it enables the separation of all SC lipids in one run in a relatively short time (∼2 h). It also results in reproducible quantification due to compressed thin bands resulting from the altered gradient steps. An advantage of our LC/MS analysis of the spots is that it was possible to separate and profile SC CER classes and distinguish their different chain lengths. LC/MS can deliver a much more detailed profile of individual subclasses. By contrast, several of the LC/MS methods currently reported cannot analyse all CER subclasses in a single run. In addition, almost all LC/MS methods described in the literature are time consuming.9,26
Mass spectrum analysis of CERs
To improve the detection of [M−H]− instead of [M+CH3CO2]−, we applied a counter-current gas (nitrogen) into the ionization source (sweep gas) instead of increasing the cone voltage in the reduced pressure zone. Usually, applying a sweep gas flow rate during the ionization process leads to a decreased total signal. However, in our experiments, the sweep gas flow rate could be acutely adjusted to detect CER only as [M−H]−, without a significant loss of signal. Figure 3 is an example of the mass spectra obtained from spots from CERs separated by HPTLC/UV. In this example, the MS analysis is of spot 6 from Figure 2. This shows a major ion at m/z 666, which could correspond to the [M−H]− ion of a CER NP, with a total number of 42 carbon atoms. Different chain lengths can be observed in this mass spectrum, with 40, 41, 43, 44, 45, and 46 atoms of carbon, with m/z 638, 652, 680, 694, 708, and 722, respectively. To confirm the identity of this suspected CER NP, a product ion experiment (MS/MS) was carried out by selecting m/z 666 as the precursor ion. Many fragment ions were observed in the corresponding MS/MS spectrum (Figure 3(b)). Interpretation of this fragmentation pattern was achieved by comparing this with the MS/MS fragmentation mechanisms of standards of CERs.
Example of the analysis of CERs using MS. (a) Mass spectrum obtained from the HPTLC/UV spot 6 assigned to CER NP (from Figure 2) and (b) the product ion spectrum of the major ion (m/z 666) by low energy CID using quadrupole tandem mass spectrometry, with argon as collision gas and a collision energy of 30 eV.
Gas phase fragmentation of CERs
To identify CERs in SC samples, they were compared with standard reference compounds of NS, NP, AP, AdS, and EOS CERs. In accordance with published MS/MS fragmentation mechanisms, we support the use of fragmentation pathways involving the CDF process and generating an ion–dipole complex. We first interpreted the MS/MS spectrum of NS CER and, based on its proposed mechanisms (Scheme 1), we applied the same fragmentations pathways to explain the fragment ions observed for NP, AP, and AdS CERs (Table 1, Figure 4).
Product ion spectra of [M−H]− ions of standard reference standards of (a) [N(24:0)S(18:1)], (b) [N(18:0)P(18:0)], (c) [A(24:0)dS(18:1)], and (d) [A(18:0)P(18:0)] by low energy CID using quadrupole tandem mass spectrometry, with argon as collision gas and a collision energy of 30 eV. Proposed fragmentation pathways involving the CDF process of CER [N(24:0)S(18:1)]. Fragmentation of NS, NP, AP, and AdS CERs. The schemes resulting in the fragment ions are depicted in Schemes 1 to 3. CER: ceramide; n.a.: not applicable; n.d.: not detected.
Interpretation of the MS/MS spectra of standards of CERs
The low energy product ion spectrum of the CER [N(24:0)S(18:1)-H]− (Figure 4(a)) displays common fragment ions, namely [(M−H)–(H2O)] − (−18 u), [(M−H)–(HCHO)] − (−30 u), [(M−H)–(CH3OH)] − (−32 u) and [(M−H)–(H2O)–(HCHO)] − (−48 u). In addition, a major fragment ion was observed at m/z 392, corresponding to the cleavage of the C2–C3 bond (numbering of the C atoms is shown in Figure 3(a)), as previously reported. 16 Ionization can competitively take place at difference acid sites, yielding a distribution of anionic molecular species. In addition, prototropy 27 can also lead to a similar distribution from excited anions if the chains conformation allows these proton transfers. From these data, three mechanisms can be proposed, all originating from the deprotonation at the C3 hydroxyl function. One mechanism involves a [1,2] hydride transfer on C2, which can produce an ion–dipole complex, including the m/z 237 fragment ion and a non-observed m/z 410 ion, which directly eliminates H2O and yields the m/z 392 ion (Scheme 1-1). However, this mechanism implies that, following alkoxide site inducing a geminated hydride transfer to the C2 site, the C2–C3 bond is cleaved; whereas, the cleavage of C2–N is more likely. A similar second mechanism can be proposed, starting from a loss of H2O, followed by a charge delocalization, and generating an ion–dipole that decomposes into the m/z 237 and 392 anion species (Scheme 1-2). Finally, a third proposed mechanism from a [1,6] proton transfer leads to an elimination of H2O and formation of an ion–dipole complex between the m/z 392 ion and a non-observed 239 species (Scheme 1-3). The three mechanisms likely occur simultaneously, explaining the detection of both the m/z 237 and 392 fragment ions. The absence of the expected product m/z 239 anions does not necessarily mean that the corresponding molecule is not acidic enough in the gas phase to transfer a proton to its anion partner. Indeed, alcohols are less acidic than ketones, ketenes, amides, and carboxylic acids. 28 The C2–C3 bond could also be cleaved following a deprotonation at the carbonyl function. In this case, charge delocalization can either initiate a [1,2] proton transfer from allyl hydroxyl group to the bond cleaved (Scheme 1-4) or it migrates to the vinyl carbon atom of allyl alcohol, inducing double bond migration to give rise to the enol anion stabilized by delocalization (Scheme 1-5). Both fragmentation processes can explain the occurrence of the m/z 408 ion.
Another bond that can be easily cleaved is the amide bond. In such a cleavage, Vietzke et al. 13 assigned the corresponding fragment ion to an amide radical ion, which is surprising for non-aromatic compounds. 29 More reasonably, Hsu and Turk 16 proposed a mechanism starting from an unusual loss of H2. However, the deprotonated primary alcohol (at C1) can promote a nucleophilic attack of the carbonyl function to yield a tetrahedral intermediate, which seems much more probable. 19 Based on the latter suggestion, we propose a slightly modified mechanism with a [1,2] hydride transfer to explain the last step of formation of the m/z 367 ion as a fatty acid (Scheme 1-6). Its favourable formation is due to the acid character of carboxylic acids compared to the imines and alcohols.
Cleavage of the amide bond can also be initiated by a deprotonation at the α-carbon of the carbonyl group which bears mobile protons, this position being enolizable. As reported by Ann and Adams, 12 this position is a possible charge site alternative to labile protons of hydroxyl group or the amide function. Thus, a [1,5] proton transfer can generate an ion–dipole complex intermediate responsible for the competitive formation of m/z 349 and 298 (Scheme 1-7). A further loss of HCHO from this latter ion may explain the formation of the m/z 268 ion after proton transfer from the primary alcohol to alkoxide site. Alternatively, the localization of the charge yielding primary alkoxide may promote an [1,4] hydride transfer, driving the amide bond reduction to give aldehyde neutral concomitant with an [1,4] proton transfer from the allyl hydroxyl site to a nitrogen atom. This results in the formation of an ion–dipole intermediate (Scheme 1-8). The two competitive ions from the dissociation of the ion–dipole complex are not detected (i.e. m/z 293 and m/z 351) since they are rapidly dissociated via a [1,2] hydride transfer to eliminate H2O and produce the m/z 278 fragment ion; whereas, the aldehyde releases HCHO.
The last bond that can be cleaved in this type of CER is located between C2 and the nitrogen atom. A [1,2] hydride transfer easily explains the formation of the m/z 366 fragment ion (Scheme 1-9). The same hydride may also be responsible for a preliminary loss of H2O from the hydroxyl function at C3. This leads to deprotonation of the carbon atom in the alpha position of the carbonyl function, followed by a cyclization with the nitrogen atom (Scheme 1-10). Finally, the charge transfer from the m/z 263 species in the dipole–ion complex to the cyclic residue generates the m/z 366 ion.
Interpretation of the MS/MS spectrum of NP CER
The same fragmentation processes may occur during the fragmentation of the [M−H]− ion of CER [N(18:0) P(18:0)] (Figure 4(b)), explaining most of the observed fragment ions (Table 1, Scheme 2). However, due to the presence of a hydroxyl site at C4, other mechanisms can be proposed and some consecutive losses can also be observed. In contrast to the process displayed in Scheme 1-10, the transfer of the charge during the loss of H2O can occur on the hydroxyl function at C4, instead of the α-carbon of the carbonyl group. After this, an [1,3] hydride transfer can explain the formation of the m/z 282 ion by the cleavage of the bond between C2 and the nitrogen atom (Scheme 2-1). From the fragmentation process displayed in Scheme 1-6, the non-observed ion from the dipole–ion complex intermediate indicates a fast dissociation of this product ion. Indeed, it can undergo a further loss of H2O to generate the m/z 280 fragment ion (Table 1). Similarly, the non-observed ion of the dipole–ion complex in Scheme 1-3 can also produce the m/z 255 fragment ion by further loss of H2O and the m/z 225 ion by an additional loss of HCHO (Table 1). Finally, the non-observed ion of the dipole–ion complex of Scheme 2-1 is detected at m/z 326 in the product ion spectrum of the analysed CER NP (Table 1). Compared to the structure of CER NS, the bond between C3 and C4 can also be cleaved considering a fragmentation process represented in Scheme 2-2, by a [1,2] hydride transfert followed by a proton transfer to explain m/z 225 and m/z 338 (Scheme 2-2).
Proposed fragmentation pathways complementary to Scheme 1 and involving the CDF process of CER [N(18:0)P(18:0)].
Interpretation of the MS/MS spectrum of AdS and AP CERs
The same mechanisms were also considered to explain the fragmentation pattern of the CER [A(24:0)dS(18:1)] (Figure 4(c) and Table 1, Scheme 3). However, the hydroxyl function at the α-carbon of the carbonyl can be implicated in several mechanisms. Concerning the cleavage of the amide bond, an [1,4] hydride transfer after an initial loss of H2O can generate the ion observed at m/z 351 (Scheme 3-1). Ionization of this hydroxyl function on the α-carbon of the carbonyl can lead to a dipole–ion complex involving the m/z 365 and 298 species after a [1,2] hydride transfer on the carbonyl function (Scheme 3-2). Then further fragmentation of the m/z 298 ion may explain the occurrence of the m/z 268, 280, and 250 ions by loss of HCHO, H2O, and H2O+HCHO, respectively (Scheme 3-2). After the loss of water, the C2–C3 bond can be cleaved according to the mechanism proposed in Scheme 3-3 to generate a dipole–ion complex between m/z 237 and 390.
Proposed fragmentation pathways complementary to Schemes 1 and 2, and involving the CDF process of CER [A(24:0)dS(18:1)].
Since the CER AP displays a hydroxyl function at the α-carbon of the carbonyl, and a hydroxyl function at C4 (Figure 4(d)), all the previously described fragmentation processes can be applied to this compound and explain the fragment ions observed in the product ion spectrum of the standard CER [A(18:0)P(18:0)] (Scheme 2-3). Corresponding m/z ratios of all observed ions for the four CERs studied are summarized in Table 1, Scheme 3.
Analysis of the standard, CER EOS, showed the characteristic fragmentation pattern of ω-esterified CERs, with the detection of only one fragment ion corresponding to linoleic acid at m/z 279.13,17
Interpretation of MS/MS spectra of CERs detected in human skin SC
Based on these proposed mechanisms of fragmentation, the interpretation of the MS/MS spectrum of detected CERs in skin by HPTLC/UV–MS was carried out. In the example shown in Figure 3, the major fragment ion was m/z 410. This was attributed to a fragment ion of the non-hydroxylated fatty acid chain of a CER [N(24:0)P(18:0)], according to the mechanism shown in Scheme 1-1. Similarly, several fragment ions can originate from an N(24:0) chain, such as m/z 349 (Scheme 1-7), 366 (Schemes 1-9 or 1-10 or 2-1), 367 (Scheme 1-6), 392 (Scheme 1-1 or 1-2 or 1-3), 408 (Scheme 1-4), and 422 (Scheme 2-2). This assignment was confirmed by the detection of several fragment ions originating from a P(18:0) chain, such as m/z 237 (Scheme 1-1), 239 (Scheme 1-5), 255 (Scheme 1-1 or 1-3), 257 (Scheme 1-4), 281 (Scheme 1-1), 283 (Scheme 1-10), 298 (Scheme 1-6), 316 (Scheme 1-7).
The m/z 267 ion can be attributed to a fragment ion of an N(18:0) chain, according to the mechanism shown in Scheme 1-8. The assignment of a mixture of CERs [N(24:0)P(18:0)] and [N(18:0)P(24:0)] was confirmed by the detection of m/z 283 coming from an N(18:0) chain (Scheme 1-6), as well as m/z 339 (Scheme 1-1 or 1-3), 367 (Scheme 1-10), and 382 (Scheme 1-6) from a P(24:0) chain.
The origin of some ions could not be attributed to the [N(24:0)P(18:0)] or [N(18:0)P(24:0)] structures. Fragment ions m/z 424 and 436 correspond to a difference of 14 mass units, with m/z 410 and 422, respectively. These most likely originated from an N(25:0) chain of CER [N(25:0)P(17:0)], according to the same fragmentation mechanisms of m/z 410 and 422 of CER [N(24:0)P(18:0)] (Schemes 1-1 and 2-2, respectively). Similarly, m/z 438 and 450 were attributed to an N(26:0) chain of CER [N(26:0)P(14:0)], according to the same mechanisms.
CERs in human skin SC identified by MS analysis after separation and detection of separate spots using HPTLC/UV.
CER: ceramide; HPTLC/UV: high performance thin layer chromatography with UV detection.
Quantification of human skin lipids by HPTLC/UV
SC lipids mainly consist of CHOL, FFAs, and CERs; therefore, these were quantified by HPTLC/UV. Calibration was performed by applying different volumes of a reference standard (CER AS) to a TLC plate. The applied volume of standard varied between 1.2 and 60 µl corresponding to five different amounts. The calibration curves were run on the same TLC plate as the lipid extracts of the human SC and were polynomial. For the lower limit of detection (LLOD) and the lower limit of quantification (LLOQ), the data were extrapolated from three calibration curves. The LLOD was 20 ng/spot and represents a concentration of 3.3 ngµl−1, while the LLOQ was 60 ng/spot and represents a concentration of 10 ngµl−1. The measurement of the peak areas at five different concentrations of CER AS used for calibration demonstrated low values of %CV. Measured precisions were between 1.9 and 3.9% for intra-day and between 1.9 and 4.2% for inter-day, indicating a good precision and reproducibility of the method.
The proportion of FFAs, CHOL, CERs and the different classes of CERs were similar to the proportion obtained by Bouwstra et al.
3
(Figure 5). Approximately 30% of the total lipids extracted corresponded to CHOL and CHOL derivatives, 10% were fatty acids and 45% were CERs. The major class of CER detected in human SC was CER NP, which accounted for up to 31% of the total CERs. Squalene coeluted with the solvent front and thus prevented the quantification of this lipid. These results are in accordance with previous works of Farwanah et al.
24
and Opitz et al.
25
Relative amounts of human skin SC lipids (FFAs, CHOL, CERs) quantified by HPTLC/UV. The values were related to CER AS as a standard reference.
Conclusion
Hsu and Turk 16 reported extensive work on the fragmentation of NS CER based on the CDF process; however, other fragmentation pathways are also possible, and not only originating from a prior loss HCHO or H2+HCHO. We propose alternative fragmentation pathways also involving the CDF process, but generating directly fatty acid-related fragment ions and long-chain base-related fragment ions via stepwise pathways involving deprotonated CER isomerization into an ion–dipole complex prior to dissociation. 21 Interestingly, the proposed mechanisms explaining the fragmentation of NS CER could be applied to explain the fragment ions observed for NP, AP, and AdS CERs. This interpretation of MS/MS spectra enabled us to identify SC lipids. These mainly consist of CHOL, FFAs, and CERs, and they were shown to be readily quantified by HPTLC/UV. This fast quantification by HPTLC/UV can be investigated further using LC/MS analyses of SC lipids. Our proposed mechanisms interpreting the fragmentation of NS CERs and their application to the MS/MS spectra of NP, AP, and AdS CERs could also be extended to other CERs. These methods (HPTLC/UV and MS identification) can be used for the analysis of different skin models and help to interpret the variation of SC lipids, in particular CERs which are key components of the skin barrier function.
Footnotes
Acknowledgements
All MS experiments were performed on the instruments of the MetaToul-AXIOM platform, partner of the national infrastructure of metabolomics and fluxomics: MetaboHUB (MetaboHUB-ANR-11-INBS-0010).
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.