N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

www.n r j ou rna l . com

Dietary sericin enhances epidermal levels ofglucosylceramides and ceramides with up-regulatingprotein expressions of glucosylceramide synthase,β-glucocerebrosidase and acidic sphingomyelinase inNC/Nga mice

Hyunae Kim, Jongsun Lee, Yunhi Cho⁎

Department of Medical Nutrition, Graduate School of East-West Medical Science, Kyung Hee University, Yongin-si, Gyeonggi-do,Republic of Korea

A R T I C L E I N F O

Abbreviations: AD, atopic dermatitis; aSMceramidase; GlcCer, glucosylceramide; GlcCesynthase, sphingomyelin synthase; SPT, seri⁎ Corresponding author.Department ofMedica

Gyeonggi-do, 446–701, Republic of Korea. TelE-mail address: choyunhi@khu.ac.kr (Y. C

0271-5317/$ – see front matter © 2012 Elsevihttp://dx.doi.org/10.1016/j.nutres.2012.09.006

A B S T R A C T

Article history:Received 8 May 2012Revised 14 September 2012Accepted 17 September 2012

We have previously reported that dietary sericin improves epidermal dryness with theincreased total Ceramide (Cer) in NC/Nga mice, an animal model of atopic dermatitis (AD). Inthis study, we hypothesized that the increased level of total Cer induced by dietary sericinwould be related to the altered metabolism of glucosylceramide (GlcCer) and sphingomyelin(SM),major precursors ofCer generation.NC/Ngamicewere feda control diet (groupCA: atopiccontrol) or dietswith 1% silk protein, either sericin (group S) or fibroin (group F) for 10weeks. Inthe epidermis of group CA, total Cer (including Cer1, 2, 3/4 and 6) and all GlcCer species werereduced; these levels in group Swere increased to levels similar to or higher than in thenormalcontrol group of BALB/c mice (group C). In addition, the protein expressions, but not mRNAexpressions, of GlcCer synthase, β-glucocerebrosidase, and acidic sphingomyelinase, enzymesfor GlcCer synthesis, GlcCer and SMhydrolysis, respectively, were highly increased in group S.The epidermal levels of total Cer (including Cer2, 3/4, and 6) and all GlcCer species and of theseenzyme proteins in group F were lower than in group S. Notably, alterations in total SM, SM1,SM3, and SM synthase 1, which were increased in group CA, were not significant betweengroups S and F. Cer5 and SM2 were not altered among groups. Dietary sericin enhanced theepidermal levels of all GlcCer and most Cer species with up-regulating protein expressions ofGlcCer synthase, β-glucocerebrosidase, and acidic sphingomyelinase.

© 2012 Elsevier Inc. All rights reserved.

Keywords:SericinCeramideGlucosylceramideSphingomyelinEpidermisNC/Nga mice

1. Introduction

Ceramide (Cer), combined with cholesterol and free fattyacids, forms the extracellular lamellar membrane structure of

ase, acidic sphingomyelr synthase, glucosylceramne palmitoyl transferase.l Nutrition, Graduate Sch.: +82 31 201 3817; fax: +8ho).

er Inc. All rights reserved

the epidermal barrier. Cer bears the structural moieties ofamide-linked non-hydroxy acids, α-hydroxy acids, ω-hydroxyacid and ester-linked fatty acids on sphingoid bases, whichare thought to play a role inmaintaining the lamellar integrity

inase; β-GlcCer'ase, β- glucocerebrosidase; Cer, ceramide; CDase,ide synthase; SM, sphingomyelin; SMase, sphingomyelinase; SM

ool of East–westMedical Science, Kyung Hee University, Yongin-si,2 31 201 3715.

.

Table 1 – Ingredient composition of the experimentaldiets fed to mice (gram per kilogram diet)

Ingredient Experimental diets 1

C CA S F

Casein 230 230 220 220Sericin - - 10 -Fibroin - - - 10L-cystine 3 3 3 3Corn oil 100 100 100 100Cellulose 50 50 50 50Vitamin mix2 10 10 10 10Mineral mix 3 35 35 35 35Sucrose 200 200 200 200Corn starch 372 372 372 372

1 Group C, BALB/c mice fed a control diet; Groups CA, S and F, NC/Nga mice fed a control diet (group CA) or diets supplemented witheither 1% sericin (group S) or fibroin (group F).2 Vitamin mix composition, AIN-93 vitamin mix #310025 (DytesInc, Bethlehem, PA, USA): niacin 3 g/kg, calcium pantothenate1.6 g/kg, pyridoxine HCl 0.06 g/kg, thiamine HCl 0.6 g/kg, riboflavin0.6 g/kg, folic acid 0.2 g/kg, biotin 0.2 g/kg, vitamin E acetate(500 IU/g) 15 g/kg, vitamin B12 (0.1 %) 2.5 g/kg, vitamin A palmitate(500000 U/g) 0.8 g/kg, vitamin D3 (400000 IU/g) 0.25 g/kg, vitamin K1/

957N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

of the epidermal barrier. Furthermore, the distinct composi-tions of these structural moieties are composed of 9 hetero-geneous Cer species (Cer1-9) in the epidermis [1].

During the differentiation processes of the epidermis, Ceris synthesized de novo with the enzymatic condensation ofserine and palmitoyl-Co A by serine palmitoyl transferase(SPT) [1]. The newly synthesized Cer is promptly modified atthe 1-hydroxy position to either glucosylceramide (GlcCer) byGlcCer synthase or to sphingomyelin (SM) by SM synthase.Although various species of Cer are generated from these 2precursors during the final stages of epidermal differentiation[1-3], all 9 Cer species (Cer1-9) are generated mainly fromGlcCer hydrolysis by β-glucocerebrosidase (β-GlcCer'ase) [1,2].Cer2 or Cer5 are generated in part from SM hydrolysis bysphingomyelinase (SMase) [1,3]. Cer ultimately undergoesdegradation by ceramidase (CDase) into sphingosine andfatty acids [1]. Depletion of Cer has been reported in skincondition that involve barrier defects, such as AD [4-6].Moreover, the marked decrease of GlcCer and SM, coupledwith alterations of Cer metabolizing enzymes and of thedegradative enzyme, CDase, has been frequently reported inthe epidermis of AD [4-7].

In our search for a dietary source that enhances theepidermal level of Cer, our attention has been drawn to silkprotein. Silk consists of 2 types of proteins, fibroin and sericin.In silk textile processing, sericin, which envelops fibroin withsuccessive sticky layers, is mostly removed, and the fibrousprotein fibroin is purified. Although fibroin is reported to be auseful biomaterial for skin health [8,9], our previous studiesdemonstrated that dietary sericin improves epidermal dry-ness [10], a major symptom of AD [11], in parallel withincreased epidermal levels of total Cer [12] in NC/Ngamice, ananimal model of AD [7,13]. However, dietary sericin andfibroin both inhibit mRNA and protein expressions of epider-mal SPT, and not of CDase [12], indicating that dietary sericindoes not enhance de novo Cer synthesis or inhibit Cerdegradation. Based on our previous studies, we hypothesizedthat the increased level of total Cer induced with dietarysericin would be related to the altered metabolism of GlcCerand SM. In this study, we examined the dietary effect ofsilk proteins on epidermal levels of individual species ofCer, GlcCer and SM and of GlcCer synthase, SM synthase,β-GlcCer'ase and SMase. NC/Ngamicewere fed 1% silk protein,either sericin or fibroin, and the epidermal levels of individualCer, GlcCer and SM species were determined. mRNA and/orprotein expressions of GlcCer synthase, SM synthase,β-GlcCer'ase, and SMase were also determined.

dextrose mix (10 mg/g) 7.5 g/kg and sucrose 967.23 g/kg.3 Salt mix composition : AIN-93G salt mix #210025 (Dytes Inc,Bethlehem, PA, USA): calcium carbonate 357 g/kg, potassiumphosphate (monobasic)196 g/kg, potassium citrate H2O 70.78 g/kg,sodium chloride 74 g/kg, potassium sulfate 46.6 g/kg, magnesiumoxide 24 g/kg, ferric citrate U.S.P 6.06 g/kg, zinc carbonate 1.65 g/kg,manganous carbonate 0.63 g/kg, cupric carbonate 0.3 g/kg,potassium iodate 0.01 g/kg, sodium selenate 0.01025 g/kg,ammonium paramolybdate 4H2O 0.00795 g/kg, sodiummetasilicate 9H2O 1.45 g/kg, chromium potassium sulfate 12H2O0.275 g/kg, lithium chloride 0.0714 g/kg, boric acid 0.0815 g/kg,sodium fluoride 0.0635 g/kg, nickel carbonate 0.0318 g/kg,ammonium vanadate 0.066 g/kg, and sucrose finely powderedsucrose 221.026 g/kg.

2. Methods and materials

2.1. Animals and diets

Five-week-old male BALB/c mice (n = 10) and 5-week-old maleNC/Nga mice (n = 30) were purchased from SLC Japan(Shizuoka, Japan). After a 1-week adaptation period, the NC/Nga mice were assigned to three groups of 10 mice each: anatopic control group (group CA) with a control diet, and groupsS and F with diets supplemented with 1.0% powdered extracts

of sericin (S) or fibroin (F), respectively. The mice were fed theexperimental diets for 10 weeks. A normal control group ofBALB/c mice was fed a control diet for 10 weeks (group C). Theingredient composition of the experimental diets is shown inTable 1. The diets of groups C and CA included no silk protein.The diet of group S included 10 g/kg sericin. The diet of group Fincluded 10 g/kg fibroin. The preparation and the molecularweights of the sericin and fibroin powders and the amino acid(AA) compositions of casein, sericin and fibroin have beendescribed previously [10,12].

During the 10-week feeding period, all mice were main-tained under conventional laboratory conditions without airfiltration to induce AD as described previously [7,13]. Themicewere housed under conditions of controlled temperature(22°C-24°C), humidity (55%-60%) and light (lights on from07:00 to 19:00). Food intakes and body weights of all groupswere monitored weekly over the 10-week feeding period.Animal care and handling conformed to the guidelinesprovided by the Animal Care and Use Review Committee ofKyung Hee University. At the end of week 10, all mice wereeuthanized by pentobarbital sodium (at a dose of 185 mg/kgIV). Epidermal strips were removed after overnight incubationof whole skin in an ice-cold 1:1 mixture of dispase II (2.4 U/mL,

Table 2 – Altered mRNA levels of β-glucocerebrosidaseand acidic sphingomyelinase in the epidermis of mice

Experimental groups 1

C CA S F

β-GlcCer'ase(% control:β-GlcCer'ase/GAPDH) 2

100 ± 5.0a 34.7 ± 7.0c 58.2 ± 4.2b 53.4 ± 6.5b

aSMase(% control:aSMase/GAPDH)

100 ± 4.6a 44.2 ± 3.2c 71.4 ± 3.0b 64.3 ± 3.7b

Values aremeans ± SEM (n = 10). Values without a common letter inthe same row are significantly different (P < .05) using 1-wayANOVA and Duncan's multiple comparison test.1 Group C, BALB/c mice fed a control diet; Groups CA, S and F, NC/Nga mice fed a control diet (group CA) or diets supplemented witheither 1% sericin (group S) or fibroin (group F).2 Total RNA in the epidermis of groups C, CA, S and F was isolated.Real-time PCR was performed for β-GlcCer'ase and aSMasemRNAs. The signal intensities from real-time PCR analysis werequantified and normalized, first to the corresponding values forthe GAPDH internal controls, and then to that of normal controlgroup (group C).

958 N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

Roche, Indianapolis, IN, USA) and RPMI (Roswell Park Memo-rial Institute)–1640 medium supplemented with 10% fetalbovine serum at 4°C as described previously [10,12].

2.2. Lipid analysis

Epidermal strips were homogenized and extracted withchloroform (CHCl3): methanol (MeOH) (2:1, v/v) to obtain thetotal lipids. The fractionations of Cer, GlcCer, and SM wereachieved by high-performance thin-layer chromatography(HPTLC) on a Silica gel 60 plate (Merck, Darmstadt, Germany)with solvent systems as described previously [14]. Specifically,Cer and GlcCer were fractionated first using CHCl3: MeOH:water (40:10:1, v/v/v) to 2 cm and again to 5 cm, followed byCHCl3: MeOH: acetic acid (47:2:05, v/v/v) to 8.5 cm and finallyn-hexane: diethyl ether: acetic acid (30:15:0.5, v/v/v) to the top;SM was fractionated using CHCl3: MeOH: acetic acid: water(25:15:4:2, v/v/v).The fractionations containing individualspecies of Cer, GlcCer and SM that had co-migrated withtheir respective standards (Cer1, Cer2, Cer3/4, Cer5, Cer6/7;GlcCer-A, GlcCer-B/C/D; SM1, SM2, SM3; Avanti Polar Lipids,Inc, Alabaster, AL, USA) were scanned by a TLC III scanner(DigiStore2; CAMAG, Muttenz, Switzerland). These fraction-ations were scanned after being treated with cupric acetate-phosphoric acid and heated to 160°C for 15minutes [14]. Basedon the calibration curves of the various concentrations of eachrespective standard, the levels of the Cer, GlcCer and SMspecies were expressed as μg/μg protein. The protein concen-tration of the epidermal homogenates was measured by amodified Lowry method [15].

2.3. Western blotting analysis of ceramide metabolizingenzyme proteins

Equal amounts of protein from the epidermis were separatedon 10% sodium dodecyl sulfate–polyacrylamide gels and then

blotted to a nitrocellulose membrane [10,12]. The membraneswere incubated with primary antibodies against GlcCersynthase (38 kd) (sc-50511), SM synthase 1 (49 kd) (sc-67097),β-GlcCer'ase (60 kd) (G4171), acidic SMase (aSMase) (55-90 kd)(sc-9817), or keratin 5 (62 kd) (ab53121) (GlcCer synthase, SMsynthase, and aSMase antibodies: Santa Cruz Biotechnology,Inc, Santa Cruz, CA, USA; β-GlcCer'ase antibody: Sigma-Aldrich Co, St Louis, MO, USA; keratin 5: Abcam plc,Cambridge, MA, USA), followed by incubation with IgG-HRPsecondary antibodies. With an enhanced ECL detectionsystem (GE Healthcare, Buckinghamshire, UK), the signalintensities were quantified by densitometry and normalizedto both the corresponding values of the keratin 5 internalcontrol and the signals observed in the normal control group(group C).

2.4. Quantitative real-time polymerase chain reaction ofceramide metabolizing enzyme mRNAs

Total RNA of epidermis was extracted by using Trizol reagent(Gibco, New York, NY, USA). Isolated RNA was quantified andelectrophoresed in 1.2% agarose gels to assess the samplequality as described previously [16].

Reverse transcription of isolated total RNA (5 μg) wasperformed using a SuperScript III first-strand synthesissystem (Invitrogen, Carlsbad, CA, USA) according to themanufacturer's instructions. Quantitative real-time polymer-ase chain reaction (PCR) was performed in 384-well platesusing Thermal Cycler Dice Real Time System TP850 (TakaraBio Inc, Shiga, Japan). Each 25-μL reactionmixture consisted of12.5 μL of SYBR Premix Ex Taq (TaKaRa), 2 μL of diluted cDNA,and 0.5 μL of 10 μM forward and reverse primers forβ-GlcCer'ase, aSMase or GAPDH. The primer sequences wereas follows:β-GlcCer'ase, 5′-GTG ACT TCT CCA TCC GTG TCT-3′(forward), 5′-CGT AGG TTC ATT CTC CGC TGT-3′ (reverse) [17];aSMase, 5′-TGG CTC TAT GAA GCG ATG GC-3′ (forward),5′-TTG AGA GAG ATG AGG CGG AGA C-3′ (reverse) [18]; andGAPDH, 5′-CCA TGG AGA AGG CTG GGG-3′ (forward), 5′-CAAAGT TGT CAT GGA TGA CC-3′) (reverse) [19]. Thermal cyclingwas initiated at 95°C for 10 seconds to activate the polymer-ase, followed by 40 cycles of 95° for 5 seconds, 60°C for 30seconds, and 72°C for 30 seconds. The relative expressions ofβ-GlcCer'ase and aSMase mRNAs were normalized first to thelevel of an internal control gene (GAPDH) and then to the levelin group C.

2.5. Statistical analyses

From our previous studies [12], a sample size of 7 wasdetermined to provide >80% power to detect a difference of40% with a SD of 25%and 2-tailed α = .05 in the epidermal levelof total Cer at the end of feeding period between groups. Asample size of 10 mice per group for epidermal levels of Cer,GlcCer, SM, β-GlcCer'ase, and aSMase (Table 2: Figs. 1 and 3),and 7 mice per group for epidermal levels of GlcCer synthaseand SM synthase 1 (Fig. 2) sufficiently powered the analyses.Results are expressed as means ± SEM. All data were analyzedwith 1-way analysis of variance (ANOVA) using SPSS forWindows (SPSS 13.0, SPSS Inc, Chicago, IL, USA) [20], and the

959N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

differences among the groups (C, CA, S, and F) weredetermined by Duncan's multiple comparison test. Differ-ences with P < .05 were considered significant.

3. Results

3.1. Epidermal levels of Cer species

With no significant differences in food intake and bodyweights among groups (data not shown), the epidermal levelof total Cer in group CA was significantly lower than in groupC (Fig. 1A). In groups S and F, the levels of total Cer werehigher than in group CA and similar to the levels in group C.However, the increased level of total Cer in group S was higher

Fig. 1 – Altered levels of ceramide, glucosylceramide and sphing(group C) and NC/Nga mice fed a control diet (group CA) or diets s10 weeks. Total lipids were fractionated into total Cer (A), total Gfurther separated into Cer1-6, GlcCer-A/B/CD or SM1-3 by HPTLC.in individual lipid species are significantly different (P < .05) usin

than in group F, which was consistent with our previousstudies [12].

Further fractionation into individual Cer species revealedthat Cer1-6 were separated, and Cer7-9 were absent in allgroups of BALB/c and NC/Nga mice, as reported previously [7].In the epidermis of group C, Cer2 was the most abundant andcomprised 81.8% of total Cer, as reported previously in amurine model [1,3]. Cer1 is the most hydrophobic Cer species,with amide linked ω-hydroxy acids and ester-linked fattyacids on sphingosine bases. Cer1 and themore hydrophilic Cerspecies, such as Cer3/4, Cer5 and Cer6 [1], accounted for 19.2%of the total Cer in group C. In group CA, the level of Cer2 wassignificantly lower than in group C; moreover, the levels ofCer1, Cer3/4 and Cer6 were significantly or modestly lowerthan in group C.

omyelin in the epidermis of BALB/c mice fed a control dietupplemented with 1% sericin (group S) or fibroin (group F) forlcCer (B) and total SM (C), of which individual species wereValues are means ± SEM (n = 10). Means with different lettersg 1-way ANOVA and Duncan's multiple comparison test.

Fig. 2 – Altered protein expressions of glucosylceramide synthase and sphingomyelin synthase 1 in the epidermis of BALB/cmice fed a control diet (group C) and NC/Nga mice fed a control diet (group CA) or diets supplemented with 1% sericin (group S)or fibroin (group F) for 10 weeks. (A) Representative expressions of glucosylceramide synthase (GlcCer synthase) andsphingomyelin synthase 1 (SM synthase 1) in the epidermis of groups. (B) The signal intensities frommultiple experiments of Awere quantified, and integrated areas were normalized, first to the corresponding value of the keratin 5 internal control andthen to the signal observed in the group C. Values are means ± SEM (n = 7). Means with different letters are significantlydifferent (P < .05) using 1-way ANOVA and Duncan's multiple comparison test.

960 N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

After dietary supplementation of sericin, the level of Cer2and of minor Cer (Cer1, Cer3/4, and Cer6) in group S wassignificantly higher than in group CA. Furthermore, theselevels increased to values similar to or higher than in group C,reflecting the increased level of total Cer. Although theincreased level of Cer1 in group F was similar to those ingroup S, the levels of Cer2 (themajor Cer species in mice) [1,3],Cer3/4 and Cer6 were significantly or modestly lower than ingroup S. The epidermal level of Cer5 did not differ among allgroups regardless of AD induction or dietary supplementationwith sericin or fibroin.

3.2. Epidermal levels of glucosylceramide andsphingomyelin species

The epidermal level of total GlcCer in group CA was far lessthan in group C (Fig. 1B). In contrast, the level of total GlcCer ingroup Swas higher than in group CA andwas similar to that in

group C, as indicated by the increased levels of total Cer, Cer1,Cer2 and Cer3/4 in Fig. 1A. In the further fractionation intoGlcCer-A (acylGlcCer, the most hydrophobic component) andGlcCer-B/C/D (the more hydrophilic component) [1,2], theincreased levels of GlcCer-A and GlcCer-B/C/D in group S werealso similar to those in group C. However, the increase of totalGlcCer and of all GlcCer species in group F was modestly lessthan in group S.

SM could be separated into three major sub-fractionationsby HPTLCwith distinctive amide-linked (N-acyl) fatty acid (FA)compositions: long-chain FA (SM1; C22-26, the most hydro-phobic SM), shorter-chain FA (SM2; primarily C16) or short-chain α-hydroxy FA (SM3; C16-18, the most hydrophilic SM)[1,3]. Notably, AD induction in NC/Nga mice increased thelevel of total SM, SM1, and SM3 in group CA, whereas thealterations of these levels in groups S and F were not apparent(Fig. 1C).The epidermal level of SM2 did not differ among anyof the groups.

961N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

3.3. Protein expressions of glucosylceramide synthase andsphingomyelin synthase 1 in the epidermis

Of various SM synthases, SM synthase 1 is located in the Golgiapparatus, the cellular compartment responsible for de novoSM synthesis [21]. To understand the mechanisms underlyingthe altered levels of GlcCer and SM, protein expressions ofGlcCer synthase and SM synthase 1 were determined (Fig. 2).The protein expression of GlcCer synthase in group CA wassignificantly less than that in group C. This finding indicatesthat AD induction inhibits the glucosyl-transfer metabolismof newly synthesized Cer into GlcCer. Although the proteinexpressions of GlcCer synthase in groups S and F were higherthan in group CA, the expression level in group Fwas less thanin group S. The increased protein expression of GlcCersynthase in group S was similar to that in group C, a patternsimilar to that of the increased levels of total GlcCer and allGlcCer species.

The protein expression of SM synthase 1 in group CAincreased up to almost 2-fold higher relative to group C,

Fig. 3 –Altered protein expressions of β-glucocerebrosidase and acontrol diet (group C) and NC/Nga mice fed a control diet (group C(group F) for 10 weeks. (A) Representative expressions of β-GlcCintensities from multiple experiments of A were quantified, andvalue of the keratin 5 internal control and then to the signal obsewith different letters are significantly different (P < .05) using 1-w

indicating that the increase of SM1 and SM3 observed with ADinduction is caused by a high protein expression of SMsynthase 1. The protein expression of SM synthase 1 in groupsS and Fwas lower than in groupCA, but higher than in groupC.

3.4. Protein and mRNA expressions ofβ-glucocerebrosidase and acidic sphingomyelinasein the epidermis

Distinct from other SMases, the aSMase alteration is morefrequently related tomaintaining the epidermal barrier [1,6,7].To further determine the contributions of GlcCer and SMhydrolysis to the altered generation of Cer, the protein andmRNA expressions of β-GlcCer'ase and aSMase were deter-mined (Fig. 3 and Table 2). In the epidermis of group CA,the protein expressions of β-GlcCer'ase and aSMase weresignificantly less than those in group C (Fig. 3). This resultsuggests that the low expressions of both β-GlcCer'ase andaSMase proteins are metabolic features related to thedecreased generation of Cer (Fig. 1A) or the accumulation of

cidic sphingomyelinase in the epidermis of BALB/cmice fed aA) or diets supplemented with 1% sericin (group S) or fibroiner'ase and aSMase in the epidermis of groups. (B) The signalintegrated areas were normalized, first to the correspondingrved in the group C. Values are means ± SEM (n = 10). Meansay ANOVA and Duncan's multiple comparison test.

962 N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

SM (Fig. 1C) in the epidermis of group CA. In group S, theprotein expressions of β-GlcCer'ase and aSMase were higherthan those in group CA and were further increased relative togroup C. In group F, the protein expressions of β-GlcCer'aseand aSMase were increased to levels similar to those in groupC but were far less than in group S.

The mRNA expressions of β-GlcCer'ase and aSMase ingroup CA were significantly less than in group C (Table 2),which was similar to the β-GlcCer'ase and aSMase proteinexpression patterns (Fig. 3). Although the protein expressionsof β-GlcCer'ase and aSMase in group F were increased tolevels similar to those in group C and were further enhancedin group S, the mRNA expressions of β-GlcCer'ase andaSMase in both groups S and F were less than in group C.Future studies are needed to determine whether this findingindicates that the translation of these mRNAs or the stabilityof these proteins are increased by dietary fibroin and/or bydietary sericin.

4. Discussion

Although Cer plays a crucial role in maintaining the lamellarintegrity of the epidermal barrier [1], little is known about thedietary modulation of Cer metabolism in AD, a typicalepidermal barrier defect disease. The epidermal level of totalCer, including Cer1, Cer2, Cer3/4 and Cer6, was decreased withAD induction in NC/Ngamice. Although the decreased level oftotal Cer was consistent, the decreased levels of Cer1, Cer2,Cer3/4 and Cer6 in group CA do not agree with prior reports, inwhich Cer1 and Cer3 are decreased in AD-induced NC/Ngamice [7]. This discrepancy is likely due to the period of ADinduction; in contrast to a 6- or 8-week induction [7], a longer10-week period was employed in the present study thatseemed to further reduce the levels of other Cer species, suchas Cer2, Cer4, and Cer6.

The decreased level of Cer species in group CA couldbe related to the accumulation of GlcCer and/or SM. Theaccumulation of these 2 precursor lipids of Cer generation[1-3] has been reported in the epidermal barrier defect ofGaucher or Niemann-Pick disease [22,23]. However, in AD-induced NC/Nga mice, the reduced levels of all GlcCerspecies (GlcCer-A/B/C/D) paralleled the reduced levels ofmost Cer species with low expressions of GlcCer synthaseand β-GlcCer'ase proteins. Conversely, SM seemed to accu-mulate along with a high expression of SM synthase 1 and alow expression of aSMase. In addition to β-GlcCer'ase oraSMase deficiencies seen in Gaucher or Niemann-Pickdisease [22,23], AD induction (group CA) seemed to furtherreduce GlcCer synthase and conversely enhance SMsynthase 1, thereby resulting in reduced GlcCer and Cerspecies and SM accumulation. In several reports with ADpatients, the reduced levels of Cer or GlcCer are reported toparallel the decreased SMase or de novo Cer syntheticactivities [1,4,6,]. Alternatively, unusual expression of SM orGlcCer deacylase, which metabolizes SM to sphingosylphosphorylcholine or GlcCer to glucosyl sphingosine, couldalso explain the decreased levels of Cer, GlcCer and SM inthe lesional epidermis of patients with AD [24,25]. However,the expression of deacylase has not been reported in mice

[7]. Apart from the epidermis of AD patients [1,4], these data,along with our previous studies in which AD inductionincreases the protein expressions of SPT and CDase in thedry epidermis of NC/Nga mice [10,12], suggest that althoughde novo synthesis of Cer and SM could be enhanced to acertain extent, AD induction inhibits the glucosyl-transfer ofnewly synthesized Cer into GlcCer and further decreases thegeneration of most Cer species in NC/Nga mice. Thisdecrease was seen along with decreased GlcCer and SMhydrolysis and enhanced Cer degradation in NC/Nga mice.

In the samemanner, the recovered levels of Cer and GlcCerspecies in group S could be explained by the increasedexpressions of GlcCer synthase, β-GlcCer'ase and aSMase.The levels of total Cer in group S, including Cer1, Cer2, Cer3/4and Cer6, were similar or higher than in group C, which couldreflect the depletion of GlcCer or SM. However, the levels oftotal GlcCer and all GlcCer species were similar to the levels ofgroup C. Furthermore, the protein expressions of β-GlcCer'aseand aSMase in group S were significantly increased up toalmost 3-fold higher than in group C. In addition, dietarysericin enhances protein expression of GlcCer synthasesimilar to the level of group C. These data, together with ourprevious studies in which dietary sericin inhibit proteinexpressions of epidermal SPT, and not of CDase [12], indicatethat although de novo Cer synthesis is inhibited, dietarysericin enhanced glucosyl-transfer of newly synthesized Cerinto GlcCer, and increased GlcCer and SM hydrolysis for Cergeneration; thereby, epidermal dryness in AD inducedNC/Ngamice is recovered [10]. Dietary sericin has also been shown todecrease the expression of SM synthase 1, which is increasedin group CA, to a certain degree. However, the alteration oftotal SM, SM1 and SM3 in group S was not apparent, indicatingthat the stimulatory effect of sericin on Cer generation is notlikely to due to de novo SM synthesis.

Because both Cer2 and Cer5 are reported to be generated inpart from SM1 and SM3 [1,3], the altered level of Cer2, a majorCer species in the epidermis [1,3], could be explained by thealtered levels of SM1 and SM3. However, all Cer species,including Cer2 and Cer5, could be alternatively generated fromGlcCer hydrolysis [1,2], which appears to suffice for maintain-ing the epidermal barrier [26]; furthermore, Cer5 levels did notdiffer among groups. Therefore, the altered levels of Cer2 ingroups CA and S did not seem to be related to the SM1 and SM3alteration. Because Cer5 or SM2 are minor species comprising<5% of either the total Cer or SM [1,3], the altered levels werenot enough to exert a significant difference. Regardless of theindividual Cer, GlcCer, and SM species, AD induction inhibitedepidermal levels of GlcCer and Cer and enhanced the SMaccumulation. In other words, dietary sericin supplementa-tion appears to enhance the levels of all Cer and GlcCerspecies in general, but not their specific species.

Sericin has low digestibility to pepsin and pancreatin dueto its protease-resistant property in vitro [27]. Undigestedsericin functions as a dietary fiber in the intestine, therebyincreasing the fecal excretion of cholesterol and triglycerol oraltering systemic lipid profiles [28], which in turn may thenaffect ceramide metabolism in the epidermis. However, theaddition of a low% of sericin into casein does not inhibitpepsin and pancreatin digestion [27]. Furthermore, wheneither peptide (prepared by protease treatment in vitro with

963N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

a molecular weight confirmed around at 5410) or AA forms ofsericin are supplemented in diet as the same percentage ofprotein form of sericin (molecular weight 65300) (group S), theepidermal level of Cer is far less than in group CA [12]. Thisfinding suggests that the increased levels of Cer species ingroup S could not be explained solely by an altered systemiclipid profile related to the low digestibility of sericin [27,28].Alternatively, dietary sericin increases the serum level ofadiponectin [29], which indirectly mediates a decrease ofhepatic lipogenesis [30]. In this study, protein expressionsof β-GlcCer'ase, aSMase and GlcCer synthase in the epidermisof group S were significantly higher than in group CA. Thesedata suggest that ingested sericin or partially digested sericinpeptide fragments, whose AA sequences might be distinctfrom those prepared by protease treatment in vitro [12], couldindirectly promote systemic responses to alter the expressionand/or activity of target enzymes (such as SPT, β-GlcCer'ase,GlcCer synthase or aSMase) in Cer metabolism. The systemicresponses of dietary sericin could be due to complexmechanisms that remain to be elucidated in future studies.

In contrast to dietary sericin, the beneficial effects ofdietary fibroin (molecular weight 101700) [12] on the epider-mal levels of GlcCer and Cer specieswere not apparent despitethe increase in protein expression of β-GlcCer'ase and aSMasethat was similar to the level of group C. Sericin contains highlevels of serine (30.4% of total AAs), aspartate (19.1%) andglycine (12.2%). In contrast, only 12.1% of total AAs in fibroinare serine, with the predominant AAs being glycine (44.4%)and alanine (29.3%) [10,12]. Although glycine and serine can beinterconverted by a reversible reaction in mammalian tissue[31], ingested fibroin or partially digested fibroin peptidefragments seem to have distinct AA compositions fromthose of ingested sericin or partially digested sericin peptidefragments. These differences might result in varying indirectsystemic responses or biological activities [32,33], such as lessapparent increases in epidermal levels of GlcCer synthase,β-GlcCer'ase and aSMase proteins in group F relative to groupS. Dietary fibroin also has also been shown to decrease theexpression of SM synthase 1, which is increased in group CA,to a similar degree as did dietary sericin. Furthermore, thealteration of total SM, SM1 and SM3 in group F was similar tothat in group S, indicating again that the de novo SM synthesisis not likely themajor targetmetabolism for the enhanced Cergeneration that occurs with dietary sericin.

Our investigation had some limitations. We did notdetermine the altered activities of GlcCer synthase, SMsynthase 1, β-GlcCer'ase and aSMase because of the shortageof epidermal samples. In addition, the mRNA or proteinexpressions of ceramide synthase, another key enzyme of denovo Cer synthesis, was not determined. Despite thoselimitations, the extensive investigation of altered Cer, GlcCerand SM metabolism in the epidermis of sericin fed NC/Ngamice, an animal model of AD [7,13] provides valuableadvances of a possible mechanism to understand how dietarysericin could improve epidermal dryness, a major symptom ofAD [11].

Based on the results of this study, we accept our hypothesisand conclude that dietary sericin increases epidermal levels ofall GlcCer species, a major repository of all Cer [1,2] and furtherenhances the generation of most Cer species along with up-

regulated expressions of GlcCer synthase, β-GlcCer'ase andaSMase proteins. Sericin can be used as a dietary supplementa-tion to improve the skinbarrier function inAD, but future studiesare required to investigate these effects in human subjects.

Acknowledgment

This study was supported by a grant of the National ResearchFoundation (2010–0016554) in Republic of Korea. The authorsdeclare no conflicts of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.nutres.2012.09.006.

R E F E R E N C E S

[1] Holleran WM, Takagi Y, Uchida Y. Epidermal sphingolipids:metabolism, function, and roles in skin disorders. FEBS Lett2006;580:5456–66.

[2] Hamanaka S, Hara M, Nishio H, Otsuka F, Suzuki A, Uchida Y.Human epidermal glucosylceramides are major precursors ofstratum corneum ceramides. J Invest Dermatol 2002;119:416–23.

[3] Uchida Y, Hara M, Nishio H, Sidransky E, Inoue S, Otsuka F,et al. Epidermal sphingomyelins are precursors for selectedstratum corneum ceramides. J Lipid Res 2000;41:2071–82.

[4] Macheleidt O, Kaiser HW, Sandhoff K. Deficiency of epidermalprotein-bound omega-hydroxyceramides in atopicdermatitis. J Invest Dermatol 2002;119:166–73.

[5] Arikawa J, Ishibashi M, Kawashima M, Takagi Y, Ichikawa Y,Imokawa G. Decreased levels of sphingosine, a naturalantimicrobial agent, may be associated with vulnerability ofthe stratum corneum from patients with atopic dermatitis tocolonization by Staphylococcus aureus. J Invest Dermatol2002;119:433–9.

[6] Jensen JM, Fölster-Holst R, Baranowsky A, Schunck M,Winoto-Morbach S, Neumann C, et al. Impairedsphingomyelinase activity and epidermal differentiation inatopic dermatitis. J Invest Dermatol 2004;122:1423–31.

[7] Aioi A, Tonogaito H, Suto H, Hamada K, Ra C, Ogawa H, et al.Impairment of skin barrier function in NC/Nga Tnd mice as apossible model for atopic dermatitis. Br J Dermatol 2001;144:12–8.

[8] Inouye K, Kurokawa M, Nishikawa S, Tsukada M. Use ofBombyx mori silk fibroin as a substratum for cultivation ofanimal cells. J Biochem Biophys Methods 1998;37:159–64.

[9] Mason R. Fabrics for atopic dermatitis. J Fam Health Care2008;18:63–5.

[10] Kim H, Lim YJ, Park JH, Cho Y. Dietary silk protein, sericin,improves epidermal hydration with increased levels offilaggrins and free amino acids in NC/Nga mice. Br J Nutr2012:1–10 [ahead of print in January 16th, http://dx.doi.org/10.1017/S0007114511007306].

[11] Leung DY. Atopic dermatitis: new insights and opportunitiesfor therapeutic intervention. J Allergy Clin Immunol 2000;105:860–76.

[12] Park KH, Choi YS, Kim H, Lee KG, Yeo JH, Jung DH, et al.Dietary effect of silk protein on ceramide synthesis and theexpression of ceramide metabolic enzymes in the epidermisof NC/Nga mice. J Korean Soc Food Sci Nutr 2007;36:554–62.

964 N U T R I T I O N R E S E A R C H 3 2 ( 2 0 1 2 ) 9 5 6 – 9 6 4

[13] Suto H, Matsuda H, Mitsuishi K, Hira K, Uchida T, Unno T,et al. NC/Nga mice: a mouse model for atopic dermatitis. IntArch Allergy Immunol 1999;120(Suppl. 1):70–5.

[14] Kim H, Kim J, Park J, Kim SH, Uchida Y, Holleran WM, et al.Water extract of gromwell (Lithospermum erythrorhizon)enhances migration of human keratinocytes and dermalfibroblasts with increased lipid synthesis in an in vitro woundscratch model. Skin Pharmacol Physiol 2012;25:57–64.

[15] Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Proteinmeasurement with the Folin phenol reagent. J Biol Chem1951;193:265–75.

[16] Kim H, Oh I, Park KH, Kim NM, Do JH, Cho Y. Stimulatoryeffect of dietary red ginseng on epidermal hydration andceramide levels in ultraviolet-irradiated hairless mice. J MedFood 2009;12:746–54.

[17] Takagi Y,NakagawaH, YaginumaT, TakemaY, ImokawaG.Anaccumulation of glucosylceramide in the stratumcorneumdueto attenuated activity of beta-glucocerebrosidase is associatedwith the early phase of UVB-induced alteration in cutaneousbarrier function. Arch Dermatol Res 2005;297:18–25.

[18] Zhang Y, Duan RD. Boswellic acid inhibits expression of acidsphingomyelinase in intestinal cells. LipidsHealthDis 2009;8:51.

[19] Kim J, KimY, Seo D, Kim S, Lee S, ChoY. Oral supplementationof Lithospermum erythrorhizon prevents the development ofatopic dermatitis with reducing ceramide degradation in theepidermis of NC/Nga mice. Phytother Res 2009;23:1250–6.

[20] Norušis MJ. SPSS 13.0 statistical procedures companion.Prentice Hall: Upper Saddle River; 2005.

[21] Fukasawa M, Nishijima M, Hanada K. Genetic evidence forATP-dependent endoplasmic reticulum-to-Golgi apparatustrafficking of ceramide for sphingomyelin synthesis inChinese hamster ovary cells. J Cell Biol 1999;144:673–85.

[22] HolleranWM,ElGinnis,MenonGK,Grundmann JU, FartaschM,McKinney CE, et al. Consequences of beta-glucocerebrosidasedeficiency in epidermis. Ultrastructure and permeabilitybarrier alterations in Gaucher disease. J Clin Invest 1994;93:1756–64.

[23] Schmuth M, Man MQ, Weber F, Gao W, Feingold KR, Fritsch P,et al. Permeability barrier disorder in Niemann-Pick disease:sphingomyelin-ceramide processing required for normalbarrier homeostasis. J Invest Dermatol 2000;115:459–66.

[24] Ishibashi M, Arikawa J, Okamoto R, Kawashima M, Takagi Y,Ohguchi K, et al. Abnormal expression of the novel epidermalenzyme, glucosylceramide deacylase, and the accumulationof its enzymatic reaction product, glucosylsphingosine, in theskin of patients with atopic dermatitis. Lab Invest 2003;83:397–408.

[25] Hara J, Higuchi K, Okamoto R, Kawashima M, Imokawa G.High-expression of sphingomyelin deacylase is an importantdeterminant of ceramide deficiency leading to barrierdisruption in atopic dermatitis. J Invest Dermatol 2000;115:406–13.

[26] Chujor CSN, Feingold KR, Elias PM, Holleran WM.Glucosylceramide synthase activity in murine epidermis:quantitation, localization, regulation, and requirement forbarrier homeostasis. J Lipid Res 1998;39:277–85.

[27] Sasaki M, Yamada H, Kato N. A resistant protein, sericin,improves atropine-induced constipation in rats. Food SciTechnol Res 2000;6:280–3.

[28] Seo CW, Um IC, Rico CW, Kang MY. Antihyperlipidemic andbody fat-lowering effects of silk proteins with differentfibroin/sericin compositions in mice fed with high fat diet.J Agric Food Chem 2011;59:4192–7.

[29] Okazaki Y, Kakehi S, Xu Y, Tsujimoto K, Sasaki M, Ogawa H,et al. Consumption of sericin reduces serum lipids,ameliorates glucose tolerance and elevates serumadiponectin in rats fed a high-fat diet. Biosci BiotechnolBiochem 2010;74:1534–8.

[30] Awazawa M, Ueki K, Inabe K, Yamauchi T, Kaneko K, OkazakiY, et al. Adiponectin suppresses hepatic SREBP1c expressionin an AdipoR1/LKB1/AMPK dependent pathway. BiochemBiophys Res Commun 2009;382:51–6.

[31] Fallon HJ, Davis JL, Goyer RA. Effect of protein intake on tissueamino acid levels and the enzymes of serine biosynthesis inthe rat. J Nutr 1968;96:220–6.

[32] Aramwit P, Damrongsakkul S, Kanokpanont S, Srichana T.Properties and antityrosinase activity of sericin from variousextraction methods. Biotechnol Appl Biochem 2010;55:91–8.

[33] Aramwit P, Kanokpanont S, De-Eknamkul W, Kamei K,Srichana T. The effect of sericin with variable amino-acidcontent from different silk strains on the production ofcollagen and nitric oxide. J Biomater Sci 2009;20:1295–306.