Advertisement

REAL-TIME INFRARED Spectroscopic measurement of Natural Moisturising Factor

  • John Chittock
    Correspondence
    Corresponding author: Sheffield Dermatology Research, Department of Infection & Immunity & Cardiovascular Disease, Faculty of Medicine, Dentistry & Health, The University of Sheffield Medical School, Beech Hill Road, Sheffield S10 2RX, UK. Tel: +44 (0)114 21 59539.
    Affiliations
    Sheffield Dermatology Research, The University of Sheffield Medical School, UK
    Search for articles by this author
  • Michael J. Cork
    Affiliations
    Sheffield Dermatology Research, The University of Sheffield Medical School, UK

    The Paediatric Dermatology Clinic, Sheffield Children's Hospital, UK
    Search for articles by this author
  • Simon G. Danby
    Affiliations
    Sheffield Dermatology Research, The University of Sheffield Medical School, UK
    Search for articles by this author
Open AccessPublished:November 08, 2022DOI:https://doi.org/10.1016/j.jid.2022.10.005

      Abbreviations:

      Atopic Dermatitis (AD), Natural Moisturising Factors (NMF), Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR-FTIR)

      LETTER

      The quantification of Natural Moisturising Factor (NMF) is of value to scientists and clinicians with an interest in dry skin disorders such as atopic dermatitis (AD). Hygroscopic amino acids and their derivatives, originating from Filaggrin (FLG) catabolism, represent a predominant component of NMF maintaining the physical permeability barrier of the skin; (
      • Scott I.R.
      • Harding C.R.
      • Barrett J.G.
      Histidine-rich protein of the keratohyalin granules. Source of the free amino acids, urocanic acid and pyrrolidone carboxylic acid in the stratum corneum.
      ) that when deficient are synonymous with xerosis and greater AD severity (
      • Horii I.
      • Nakayama Y.
      • Obata M.
      • Tagami H.
      Stratum corneum hydration and amino acid content in xerotic skin.
      , Nouwen et al., 2019). Tape stripping with chromatography is a fully quantitative assessment of NMF ex vivo but time consuming and labour intensive in larger cohorts. In this pilot study, we tested a portable, hand-held Attenuated Total Reflectance - Fourier Transform Infrared (ATR-FTIR) spectrometer as an alternative in vivo measure of NMF, by modelling chemometric absorption using a single, quantitative composite value obtained by established ex vivo laboratory assay. The spectroscopic model was verified by examining known scenarios of reduced NMF abundance in the skin such as the established FLG pathophysiology in AD.
      A cohort of 26 participants with healthy skin (n=15) or a history of AD (n=11; 2 with mild active disease) were recruited and completed the single study visit. Written, informed consent was obtained and approval granted by the University of Sheffield Research Ethics Committee (uREC ref: 021945). Four sampling data points across the antecubital fossa and forearm were split equally between model calibration and validation, each consisting of four baseline ATR-FTIR measurements performed contiguous to 3 serial tape strips collected in duplicate (n=6) for laboratory NMF analysis (Supplementary Figure S1). On average, the three predominant components of NMF analysed ex vivo – total free amino acid pool (fAA), pyrrolidone carboxylic acid (PCA) and urocanic acid (UCA) - were significantly reduced in the AD group compared to healthy skin (Supplementary Table S1). Transepidermal water loss (TEWL) and capacitance measurements to assess barrier function, were similar between groups, and the proportion of FLG loss-of-function (LOF) variants were 20% and 36% respectively. A plot of ex vivo versus in vivo modelled NMF is presented by Figure 1a and b. Using a six-factor predictive model, the coefficient of determination for calibration (R2=0.73) and validation (R2=0.70) sampling data points indicate satisfactory accuracy and precision (±0.35μmoles mg-1) denoted by the root mean square error of calibration (RMSEC). A plot of model loading – the strength of association between wavenumber absorption and latent factor (Figure 1d) – corresponded to an in vitro NMF profile collected by the same spectrometer (Fig 1e). Similar outputs were obtained by Amide I (1640cm-1) and II (1540cm-1) normalisation (Supplementary Table S2). To verify the spectroscopic technique, NMF was modelled before and after bathing the antecubital fossa in an independent cohort of volunteers (n=5, Supplementary Figure S2a); with on average, a 67% reduction in NMF induced by a 20-minute water soak (p=0.0036). Furthermore, the main study cohort (n=26) was stratified two ways (healthy skin / AD and wild type / FLG LOF mutation carriers) and modelled NMF abundance compared, to assess the inherited and acquired FLG defect (
      • Kezic S.
      • O'Regan G.M.
      • Yau N.
      • Sandilands A.
      • Chen H.
      • Campbell L.E.
      • et al.
      Levels of filaggrin degradation products are influenced by both filaggrin genotype and atopic dermatitis severity.
      ). In all scenarios, similar changes in absorbance were observed that matched an in vitro NMF spectrum (Supplementary Figure S2b-h). These regions were indicative of the carboxylate (-COO-) asymmetric or symmetric stretch (1540/1400 cm-1) and methylene group (CH2) vibrations around 1480cm-1 (
      • Takada S.
      • Naito S.
      • Sonoda J.
      • Miyauchi Y.
      Noninvasive In Vivo Measurement of Natural Moisturizing Factor Content in Stratum Corneum of Human Skin by Attenuated Total Reflection Infrared Spectroscopy.
      ). Interestingly, 1340cm-1 is assigned to the hydroxyl group (C-)OH bending mode of serine, an abundant amino acid within the SC (
      • Nakagawa N.
      • Naito S.
      • Yakumaru M.
      • Sakai S.
      Hydrating effect of potassium lactate is caused by increasing the interaction between water molecules and the serine residue of the stratum corneum protein.
      ). At the antecubital fossa but not the forearm (Supplementary Figure S3), in vivo modelled NMF was on average 0.64μmol mg-1 lower in AD compared to healthy skin; with all values measured in FLG LOF mutation carriers being below the wild type mean, regardless of AD history (Figure 2a and b). This discrimination between groups was supported by a Receiver Operating Characteristic area under the curve (AUC) of 0.81 and 0.83 respectively (Figure 2c and d).
      Figure thumbnail gr1
      Figure 1PLS chemometric modelling of surface NMF in the mid infrared spectral region. (a) Plot of ex vivo quantified versus in vivo ATR-FTIR modelled NMF (the sum of fAA, PCA and UCA) for calibration and (b) validation data sets (see for further details). R2 = coefficient of determination denoting linear regression model goodness-of-fit. Respective residual plots inset. Individuals with active AD are shaded red. (c) Image of the spectrometer used. (d) Loading plot associating wavenumber absorption to latent factor (6 in total, colour coded) with (e) an in vitro NMF absorbance spectrum presented for reference.
      Figure thumbnail gr2
      Figure 2In vivo modelled NMF discriminates AD and FLG null genotype from controls. Cohort stratification (n=26) to compare mean in vivo NMF at the antecubital fossa between (a) healthy skin / AD and (b) wild type (WT) / FLG LOF mutation carriers. Only the model validation data points are presented (see for further details). Individuals with active AD are shaded red. p values denote the result of an unpaired student’s t test. A Receiver Operating Characteristic (ROC) curve of modelled NMF is presented below the corresponding graph (c) AD/Healthy: area under the curve=0.81, 95% CI, 0.63-0.99, p=0.008; (d) FLG/WT: area under the curve=0.83, 95% CI, 0.66-0.99, p=0.01).
      This methodology was limited to the discrimination of subclinical AD and requires further validation due to the absence of more active disease. Compare this to in vivo Raman Spectroscopy (RS) where more comprehensive classification of FLG genotype by NMF (AUC 0.94) was reported in adults with moderate-severe disease (
      • O'Regan G.M.
      • Kemperman P.M.
      • Sandilands A.
      • Chen H.
      • Campbell L.E.
      • Kroboth K.
      • et al.
      Raman profiles of the stratum corneum define 3 filaggrin genotype-determined atopic dermatitis endophenotypes.
      ). Our work is ongoing to replicate this ATR-FTIR methodology in a more diverse AD cohort of greater severity, but a key limitation may be the shallow sampling depth (approximately 1.5μm) of the evanescent wave (
      • Brancaleon L.
      • Bamberg M.P.
      • Sakamaki T.
      • Kollias N.
      Attenuated total reflection-Fourier transform infrared spectroscopy as a possible method to investigate biophysical parameters of stratum corneum in vivo.
      ) whereas RS permits composite NMF profiling across the full SC depth without the requirement of tape stripping. This surface constraint may also render the current ATR-FTIR methodology susceptible to patients washing or applying topical treatments prior to measurements in the clinic. On the flip side, it can be argued that ATR-FTIR is comparatively the more affordable technology that permits the rapid sampling of multiple anatomical sites, with a real-time NMF output displayed by the device.
      In summary, we provide preliminary evidence to suggest that measurement of NMF in vivo using ATR-FTIR is robust and comparable to an established ex vivo technique. Considering the portable device used with no sample preparation required, this methodology has the potential to offer new opportunities of clinical research where laboratory access is not feasible. The technology has many potential uses; knowledge of FLG variant status may predict a patient’s response to emollients (
      • Danby S.G.
      • Andrew P.V.
      • Taylor R.N.
      • Kay L.J.
      • Chittock J.
      • Pinnock A.
      • et al.
      Different types of emollient cream exhibit diverse physiological effects on the skin barrier in adults with atopic dermatitis.
      ) or systemic immunosuppressives (
      • Roekevisch E.
      • Leeflang M.M.G.
      • Schram M.E.
      • Campbell L.E.
      • Irwin McLean W.H.
      • Kezic S.
      • et al.
      Patients with atopic dermatitis with filaggrin loss-of-function mutations show good but lower responses to immunosuppressive treatment.
      ) indicative of future personalised treatment stratergies. Our data is also suggestive of attenuated NMF in patients generally clear of symptoms (
      • Engebretsen K.A.
      • Bandier J.
      • Kezic S.
      • Riethmuller C.
      • Heegaard N.H.H.
      • Carlsen B.C.
      • et al.
      Concentration of filaggrin monomers, its metabolites and corneocyte surface texture in individuals with a history of atopic dermatitis and controls.
      ). Tracking this defect longitudinally with further novel measures of subclinical inflammation, may be of clinical value for monitoring remission following treatment of active disease (
      • Byers R.A.
      • Maiti R.
      • Danby S.G.
      • Pang E.J.
      • Mitchell B.
      • Carre M.J.
      • et al.
      Sub-clinical assessment of atopic dermatitis severity using angiographic optical coherence tomography.
      ). Another use is in predisposed neonates who possess a skin barrier defect long before the onset of clinical AD (
      • Horimukai K.
      • Morita K.
      • Narita M.
      • Kondo M.
      • Kabashima S.
      • Inoue E.
      • et al.
      Transepidermal water loss measurement during infancy can predict the subsequent development of atopic dermatitis regardless of filaggrin mutations.
      ). There is evidence to suggest that low NMF associates with skin barrier breakdown at birth (
      • Chittock J.
      • Cooke A.
      • Lavender T.
      • Brown K.
      • Wigley A.
      • Victor S.
      • et al.
      Development of stratum corneum chymotrypsin-like protease activity and natural moisturizing factors from birth to 4 weeks of age compared with adults.
      ). Therefore, as presented here in adults with unaffected skin, the hypothesis that NMF abundance may also be discriminative in neonates and be predictive of AD onset either alone, or in conjunction with other biomarkers, is an intriguing proposition yet to be determined.

      DATA AVAILABILITY STATEMENT

      No large datasets were generated by this study.

      ORCiDS

      John Chittock: 0000-0002-1595-7441
      Michael J. Cork: 0000-0003-4428-2428
      Simon G. Danby: 0000-0001-7363-140X

      CONFLICTS OF INTEREST STATEMENT

      All authors have none to declare

      AUTHOR CONTRIBUTIONS

      Conceptualization: JC, MJC, SD; Data Curation: MJC, SD; Formal analysis: JC, SD; Funding Acquisition: MJC, SD; Investigation: JC; Methodology: JC, SD; Project Administration: JC, SD; Resources: MJC, SD; Supervision: MJC, SD; Writing-original draft preparation: JC; Writing-review and editing: JC, MJC, SD.

      Uncited reference

      • Nouwen A.E.M.
      • Karadavut D.
      • Pasmans S.
      • Elbert N.J.
      • Bos L.D.N.
      • Nijsten T.E.C.
      • et al.
      Natural Moisturizing Factor as a clinical marker in atopic dermatitis.
      .

      ACKNOWLEDGEMENTS

      This study was funded by the LEO Foundation awards LF16062 and LF18005. We are very grateful to our volunteers for their participation, without which, the study would not be possible. Many thanks to Rob Hanson of the Department of Chemistry for his assistance with HPLC. We would also like to thank Leung Tang, Graham Miller and Alexandra Harvey at Agilent Technologies for sharing expertise in ATR-FTIR analysis and providing access to Agilent’s proprietary prototype 3B2P sampling interface. Figures were produced using Biorender.com and GraphPad prism 9.

      Supplementary Material

      Table S1Study cohort characteristics. Mean ±SD presented. 1Averaged across all sampling data points per person (see Supplementary Materials and Methods); 2Whole body EASI score averaged from two individuals with active AD; 3Carrying at least one FLG LOF allele; 4Cumulative mass of SC removed by tape stripping determined by densitometry averaged across all sampling data points; +ex vivo laboratory quantification of fAA: free amino acids; PCA: pyrrolidone carboxylic acid and UCA: urocanic acid from tape strips. NMF is the sum of these three components. p values denote the result of an unpaired student’s t test to compare groups.
      HealthyADp value
      n1511
      Age (years)37 ±1436 ±13-
      Sex (%female)6645-
      1TEWL (g/m2/hr)13.17 ±3.0514.52 ±4.210.37
      1Capacitance (units)33.71 ±6.8329.48 ±7.120.14
      2EASI score-2.53 ±0.39-
      3FLG LOF (%)3/15 (20)4/11 (36)-
      1+NMF (μmoles mg-1)1.28 ±0.670.77 ±0.250.02
      1+fAA (μmoles mg-1)1.05 ±0.530.66 ±0.220.04
      1+PCA (μmoles mg-1)0.18 ±0.110.08 ±0.030.01
      1+UCA (μmoles mg-1)0.05 ±0.030.03 ±0.010.03
      4SC mass (mg-1)0.47 ±0.080.45 ±0.050.63
      Table S2Model outputs using alternative Amide normalisation modes. FA: forearm; CF: antecubital fossa. Asterisks denote the result of an unpaired student’s t test to compare groups (Healthy/AD and WT/FLG LOF mutation carriers).*p=,0.05.
      1640cm-1 normalisation1540cm-1 normalisation
      HealthyADWTFLGHealthyADWTFLG
      Calibration (R2)0.720.71
      Validation (R2)0.720.71
      in vivo NMF FA site 1 (μmol mg-1)0.970.700.910.710.970.710.900.77
      in vivo NMF CF site 2 (μmol mg-1)1.380.74**1.270.66*1.370.73**1.250.71*

      SUPPLEMENTARY TABLE LEGENDS

      Table S1: Study cohort characteristics. Mean ±SD presented. 1Averaged across all sampling data points per person (see Supplementary Materials and Methods); 2Whole body EASI score averaged from two individuals with active AD; 3Carrying at least one FLG LOF allele; 4Cumulative mass of SC removed by tape stripping determined by densitometry averaged across all sampling data points; +ex vivo laboratory quantification of fAA: free amino acids; PCA: pyrrolidone carboxylic acid and UCA: urocanic acid from tape strips. NMF is the sum of these three components. p values denote the result of an unpaired student’s t test to compare groups.
      Figure thumbnail fx1
      Figure S1Overview of the model build. TS: tape stripping; RCF/LCF: Right/left antecubital fossa; RFA/LFA: Right/left forearm. ATR-FTIR: Attenuated Total Reflectance - Fourier Transform Infrared Spectroscopy.
      Figure S2: Correlating spectral regions with NMF abundance. (a) in vivo modelled NMF before (T0) and after (H20) soaking the antecubital fossa with water (20 minutes) in an additional cohort of five healthy volunteers. A significant reduction in NMF was observed using a paired student’s t test (c) Mean ATR-FTIR spectra and (d) mean difference spectra (T0-H20) showing the changes in absorbance following the water soak. (e) Mean ATR-FTIR spectra and (f) mean difference spectra (n=26) at the antecubital fossa obtained from healthy (blue line) and AD subjects (red dotted line). (g) Mean ATR-FTIR spectra and (h) mean difference spectra (n=26) at the antecubital fossa obtained from wild type (blue) and FLG LOF mutation carriers (red dotted line). Consistent changes in absorption were observed around (1) 1580cm-1 (2) 1480cm-1 (3) 1400cm-1 and (4) 1340cm-1 that correspond to an in vitro spectrum of NMF (b).
      Figure S3: Evaluation of in vivo modelled surface NMF at the forearm. No significant difference in mean NMF was found at the forearm using an unpaired student’s t test for (a) healthy compared to AD and (b) wild type compared to FLG LOF mutation carriers. Corresponding mean ATR-FTIR spectra shown below each graph. Blue line = healthy (left) and wild type (WT-right). Red dotted line = AD (left) and FLG LOF mutation carriers (right). Please refer to Figure 1 for key.

      SUPPLEMENTARY MATERIALS & METHODS

      Participants

      A cohort study was designed to compare surface NMF levels between volunteers with either AD or healthy skin using a portable ATR-FTIR device for in vivo spectroscopic quantification. Volunteers were recruited from the local community of the city of Sheffield, UK between November 2017 and October 2018. A diagnosis of AD was made using the UK working party criteria (
      • Williams H.C.
      • Burney P.G.
      • Pembroke A.C.
      • Hay R.J.
      The U.K. Working Party's Diagnostic Criteria for Atopic Dermatitis. III. Independent hospital validation.
      ). Healthy volunteers had no history of skin disease. An additional cohort of five healthy volunteers was recruited to investigate the effect of a short water soak (20 minutes, 1ml distilled water warmed to 37ºC contained by an open chamber) on NMF. All volunteers were asked not to apply any topical products or shower the morning of the study visit.

      Skin assessments

      All skin assessments were performed during a single visit to our dedicated, climate-controlled skin barrier suite located at the University of Sheffield. Room conditions were maintained at 20±2°C and 38-50% relative humidity. The volar aspect of the forearm and the antecubital fossa were the designated study sites. The Eczema Area and Severity Index (EASI) score was employed as a measure of AD severity. Transepidermal Water Loss (TEWL) was assessed using an AquaFlux AF200 closed chamber condensing device (Biox Systems Ltd, London, UK). Skin capacitance was measured using a Corneometer CM825 probe (CK electronic GmbH, Cologne, Germany). Volunteers acclimatised to the room conditions for 20 minutes prior to assessment.

      Infrared Spectroscopy

      A portable 4300 handheld Fourier Transform Infrared (FTIR) spectrometer with mercury cadmium telluride detector (Agilent Technologies, Santa Clara, USA) was equipped with a 3-bounce / 2-pass diamond Attenuated Total Reflectance (ATR) accessory to collect absorption spectra at the skin surface in the mid infrared region from 32 scans at 4cm-1 resolution. The area of the sampling probe that contacts the skin is approximately 79mm2.

      NMF laboratory analysis ex vivo by tape stripping

      Adapted from a published assay, (
      • Takada S.
      • Naito S.
      • Sonoda J.
      • Miyauchi Y.
      Noninvasive In Vivo Measurement of Natural Moisturizing Factor Content in Stratum Corneum of Human Skin by Attenuated Total Reflection Infrared Spectroscopy.
      ) SC collected by tape stripping the skin surface (3 serial 22mm tape strips in duplicate per sampling data point, see Figure S3) was cut and pooled in 750μl methanol. Samples were then subjected to an ultrasonic bath (20 mins) agitated at 4°C (20 mins) filtered using a 0.2μm syringe filter and dried. Distilled water (200μl) was used to resuspend samples before analysis. Isocratic elution of pyrrolidine carboxylic acid (PCA peak at 210nm) and urocanic acid (UCA peak at 270nm) was performed in 0.1M phosphate buffer (pH 2.75) containing 1% acetonitrile using a Shimadzu HPLC system (Shimadzu, Kyoto, Japan) equipped with Synergi Hydro RP column (Phenomenex, Macclesfield, UK). 25μl of sample was injected in duplicate. The same extract was used to quantify the total free amino acid pool (fAA) by o-phthalaldehyde derivatization in triplicate (
      • Nakagawa N.
      • Sakai S.
      • Matsumoto M.
      • Yamada K.
      • Nagano M.
      • Yuki T.
      • et al.
      Relationship between NMF (lactate and potassium) content and the physical properties of the stratum corneum in healthy subjects.
      ). Quantification of each NMF component was achieved by standard curve interpolation. The sum of all components (fAA pool, PCA and UCA) was calculated and normalised relative to the amount of SC removed by tape stripping (
      • Voegeli R.
      • Rawlings A.V.
      • Doppler S.
      • Heiland J.
      • Schreier T.
      Profiling of serine protease activities in human stratum corneum and detection of a stratum corneum tryptase-like enzyme.
      ) to produce a single composite measure of NMF.

      FLG genotyping

      Genomic DNA was extracted from buccal swabs using the QIAamp DNA mini kit (Qiagen, Hilden, Germany). The four common European mutations were screened by Taqman (R501X and 2282del4) or sequencing (R2447X and S3247X) using established primer and probe sets (
      • Sandilands A.
      • Terron-Kwiatkowski A.
      • Hull P.R.
      • O'Regan G.M.
      • Clayton T.H.
      • Watson R.M.
      • et al.
      Comprehensive analysis of the gene encoding filaggrin uncovers prevalent and rare mutations in ichthyosis vulgaris and atopic eczema.
      ).

      Chemometric modelling

      To evidence regions of IR absorption by NMF components in vitro, chemicals were purchased from Sigma (Merck Life Science UK Ltd., Dorset, UK) dissolved in water at the following mol%: Serine 31%; Glycine 16%; PCA 13%; Histidine 8%; Citrulline 6%; Ornithine 6%; Threonine 6%; UCA 4%; Arginine 3%; Alanine 3%) and analysed using the same spectrometer. For the in vivo quantification of NMF by ATR-FTIR, Partial Least Squares (PLS) regression modelling using the chemometrics software package Microlab Expert (Agilent Technologies, Santa Clara, USA) was employed to calibrate infrared absorption across the fingerprint spectral region (1090-1653cm-1) using a single, composite, quantitative measure of NMF (see NMF laboratory analysis ex vivo by tape stripping above). For each volunteer, four sampling data points consisting of ATR-FTIR performed prior to tape stripping were entered into the modelling, split equally between calibration and validation sets (Supplementary Figure S1). Four spectral repeats were averaged for each individual sampling data point. Prior to modelling all spectra were normalised relative to Amide III at 1245cm-1 (
      • Zhang G.
      • Moore D.J.
      • Mendelsohn R.
      • Flach C.R.
      Vibrational microspectroscopy and imaging of molecular composition and structure during human corneocyte maturation.
      ).

      Statistical analysis

      Study data was collated in Excel and all tests performed using GraphPad prism 9 (San Diego, California, USA). Group means were compared using a student’s t test. The coefficient of determination assessed the linear regression model fit of ex vivo and in vivo NMF. Discrimination of AD phenotype and FLG LOF genotype by in vivo modelled NMF was explored using binary logistic regression and Receiver Operating Characteristic (ROC) curve.

      REFERENCES

        • Brancaleon L.
        • Bamberg M.P.
        • Sakamaki T.
        • Kollias N.
        Attenuated total reflection-Fourier transform infrared spectroscopy as a possible method to investigate biophysical parameters of stratum corneum in vivo.
        J Invest Dermatol. 2001; 116: 380-386
        • Byers R.A.
        • Maiti R.
        • Danby S.G.
        • Pang E.J.
        • Mitchell B.
        • Carre M.J.
        • et al.
        Sub-clinical assessment of atopic dermatitis severity using angiographic optical coherence tomography.
        Biomed Opt Express. 2018; 9: 2001-2017
        • Chittock J.
        • Cooke A.
        • Lavender T.
        • Brown K.
        • Wigley A.
        • Victor S.
        • et al.
        Development of stratum corneum chymotrypsin-like protease activity and natural moisturizing factors from birth to 4 weeks of age compared with adults.
        Br J Dermatol. 2016; 175: 713-720
        • Danby S.G.
        • Andrew P.V.
        • Taylor R.N.
        • Kay L.J.
        • Chittock J.
        • Pinnock A.
        • et al.
        Different types of emollient cream exhibit diverse physiological effects on the skin barrier in adults with atopic dermatitis.
        Clin Exp Dermatol. 2022; https://doi.org/10.1111/ced.15141
        • Engebretsen K.A.
        • Bandier J.
        • Kezic S.
        • Riethmuller C.
        • Heegaard N.H.H.
        • Carlsen B.C.
        • et al.
        Concentration of filaggrin monomers, its metabolites and corneocyte surface texture in individuals with a history of atopic dermatitis and controls.
        J Eur Acad Dermatol Venereol. 2018; 32: 796-804
        • Horii I.
        • Nakayama Y.
        • Obata M.
        • Tagami H.
        Stratum corneum hydration and amino acid content in xerotic skin.
        Br J Dermatol. 1989; 121: 587-592
        • Horimukai K.
        • Morita K.
        • Narita M.
        • Kondo M.
        • Kabashima S.
        • Inoue E.
        • et al.
        Transepidermal water loss measurement during infancy can predict the subsequent development of atopic dermatitis regardless of filaggrin mutations.
        Allergol Int. 2016; 65: 103-108
        • Kezic S.
        • O'Regan G.M.
        • Yau N.
        • Sandilands A.
        • Chen H.
        • Campbell L.E.
        • et al.
        Levels of filaggrin degradation products are influenced by both filaggrin genotype and atopic dermatitis severity.
        Allergy. 2011; 66: 934-940
        • Nakagawa N.
        • Naito S.
        • Yakumaru M.
        • Sakai S.
        Hydrating effect of potassium lactate is caused by increasing the interaction between water molecules and the serine residue of the stratum corneum protein.
        Exp Dermatol. 2011; 20: 826-831
        • Nouwen A.E.M.
        • Karadavut D.
        • Pasmans S.
        • Elbert N.J.
        • Bos L.D.N.
        • Nijsten T.E.C.
        • et al.
        Natural Moisturizing Factor as a clinical marker in atopic dermatitis.
        Allergy. 2020; 75: 188-190
        • O'Regan G.M.
        • Kemperman P.M.
        • Sandilands A.
        • Chen H.
        • Campbell L.E.
        • Kroboth K.
        • et al.
        Raman profiles of the stratum corneum define 3 filaggrin genotype-determined atopic dermatitis endophenotypes.
        J Allergy Clin Immunol. 2010; 126: 574-580
        • Roekevisch E.
        • Leeflang M.M.G.
        • Schram M.E.
        • Campbell L.E.
        • Irwin McLean W.H.
        • Kezic S.
        • et al.
        Patients with atopic dermatitis with filaggrin loss-of-function mutations show good but lower responses to immunosuppressive treatment.
        Br J Dermatol. 2017; 177: 1745-1746
        • Scott I.R.
        • Harding C.R.
        • Barrett J.G.
        Histidine-rich protein of the keratohyalin granules. Source of the free amino acids, urocanic acid and pyrrolidone carboxylic acid in the stratum corneum.
        Biochimica et biophysica acta. 1982; 719: 110-117
        • Takada S.
        • Naito S.
        • Sonoda J.
        • Miyauchi Y.
        Noninvasive In Vivo Measurement of Natural Moisturizing Factor Content in Stratum Corneum of Human Skin by Attenuated Total Reflection Infrared Spectroscopy.
        Applied Spectroscopy. 2012; 66: 26-32