The aged epidermal permeability barrier.
Structural, functional, and lipid biochemical
abnormalities in humans and a senescent
murine model.
R Ghadially, … , K R Feingold, P M Elias
J Clin Invest.
1995;
95(5)
:2281-2290.
https://doi.org/10.1172/JCI117919
.
Aged epidermis displays altered drug permeability, increased susceptibility to irritant
contact dermatitis, and often severe xerosis, suggesting compromise of the aged epidermal
barrier. To delineate the functional, structural, and lipid biochemical basis of epidermal
aging, we compared barrier function in young (20-30 yr) vs aged (> 80 yr) human subjects,
and in a murine model. Baseline transepidermal water loss in both aged humans and
senescent mice was subnormal. However, the aged barrier was perturbed more readily with
either acetone or tape stripping (18 +/- 2 strippings vs 31 +/- 5 strippings in aged vs young
human subjects, respectively). Moreover, after either acetone treatment or tape stripping, the
barrier recovered more slowly in aged than in young human subjects (50 and 80% recovery
at 24 and 72 h, respectively, in young subjects vs 15% recovery at 24 h in aged subjects),
followed by a further delay over the next 6 d. Similar differences in barrier recovery were
seen in senescent vs young mice. Although the total lipid content was decreased in the
stratum corneum of aged mice (approximately 30%), the distribution of ceramides (including
ceramide 1), cholesterol, and free fatty acids was unchanged. Moreover, a normal
complement of esterified, very long-chain fatty acids was present. Finally, stratum corneum
lamellar bilayers displayed normal substructure and dimensions, but were focally
decreased in number, […]
Research Article
The Aged Epidermal Permeability Barrier
Structural, Functional, and Lipid Biochemical Abnormalities in Humans and
aSenescent Murine Model
Ruby Ghadially,** Barbara E. Brown,* Sandy M. Sequeira-Martin,* Kenneth R. Feingold,** and Peter M. Elias***Department of Dermatology, University of California, San Francisco, San Francisco, California 94143; and tDerinatology Service, Department of Veterans Affairs Medical Center, San Francisco, California 94121
Abstract
Aged epidermis displays altered drug permeability, in-creased
susceptibility
to irritant contactdermatitis,and of-tensevere xerosis, suggesting compromise of the aged epi-dermal barrier. To delineate thefunctional, structural,andlipid biochemical basis of epidermal aging, we compared
barrier function in young (20-30 yr) vs aged (>80 yr) humansubjects, and inamurine model. Baseline
transepi-dermal water loss in bothagedhumans and senescent mice
was subnormal. However,the aged barrierwas perturbed
more readily with either acetone or tape stripping (18±2 strippingsvs31±5strippingsinagedvsyounghuman
sub-jects, respectively). Moreover, after either acetone treat-ment ortapestripping, the barrier recovered moreslowly
in aged thaninyounghuman subjects (50 and 80%recovery
at 24and 72h, respectively,inyoungsubjectsvs15%
recov-eryat 24 h in aged subjects), followed by afurtherdelay
over the next 6 d. Similar differences in barrier recovery
were seen in senescent vs young mice. Although the total
lipidcontentwasdecreasedinthestratumcorneumof aged mice (- 30%), the distributionof ceramides(including
cer-amide 1), cholesterol,and freefattyacids wasunchanged.
Moreover, a normal complement of esterified, very
long-chainfattyacidswaspresent.Finally,stratumcorneum la-mellar bilayers displayed normal substructure and
dimen-sions,butwerefocallydecreased innumber,with decreased secretion of lamellar body contents. Thus, assessment of
barrier function in aged epidermisunder basal conditions
ismisleading, since both barrierintegrityandbarrier repair
are markedly abnormal. These functional changes can be
attributed toaglobal deficiencyin allkeystratumcorneum
lipids, resulting in decreased lamellar bilayers in thestratum
corneuminterstices. This constellation offindingsmay
ex-plain the increased susceptibility ofintrinsically agedskin to exogenous and environmental insults. (J. Clin. Invest.
1995.95:2281-2290.) Keywords: stratumcorneum*
trans-epidermal waterloss * electronmicroscopy Iipids
Introduction
Physiological changes in intrinsically (chronologically) aged skincanproduce significant morbidityin theaging population,
Address correspondence to Ruby Ghadially, Dermatology Service
( 190),VeteransAffairs MedicalCenter, 4150 Clement Street, San Fran-cisco,CA 94121. Phone: 415-750-2091;FAX:415-751-3927.
Receivedfor publication24May1994andinrevisedformn 5 January
1995.
including altered drugpermeability, increased susceptibility to
irritant contact dermatitis, and often severe xerosis. In recent
years considerable attention has focused on functional
abnor-malities of the dermis inintrinsically and photoaged skin. The
relative lack of interestintheepidermisof theaged mayreflect thereportedlynormal thickness of the stratum corneum( 1-6), with an increase in corneocyte surface areatheonly described structuralabnormality (7). Likewise, only minorfunctional
ab-normalities have been reportedinagedepidermis. Stratum
cor-neum water content is reportedly normal (8-11) or slightly decreased (12),andonly slightdifferences in barrier function
have been reported bothin vitro (1) andinvivo (9, 13, 14). Yet, drug permeability appears to be significantly altered; for
example, the in vitro permeability of water-soluble dyes, such asfluorescein, is increased ( 15), while the permeability of lipid-soluble drugs, such as testosterone, is decreased (16). These differencesinbasaltransepidermalwaterloss(TEWL) vsdrug permeability suggest that a barrier abnormality may exist in chronologically agedepidermis, which is masked by the
seem-ingly normalmeasurements of basal TEWL.Although any dif-ferences in permeability should be explicable by changes in
stratum corneum lipid composition (17), prior studies have provided conflicting results about the lipid composition of aging
human stratumcorneum(18, 19). Theseconflicting results
re-flect data derived from sampling of surface lipids (18, 19), which maynotaccurately reflectstratum corneumlipidcontent
(20),orfromassessmentofanincomplete spectrum of epider-mal-derived lipids (19).
Since functional abnormalities in other senescent organs often only become evident when the tissue is either stressedor
required to carry out repair processes (21-23), our strategy has been to assess the barrier in intrinsically aged epidermis,
not only under basalconditions, but also when theepidermis issubjectedtostress;i.e., barrier repair after experimental abro-gation.Inyoungepidermis barrier abrogation is followed bya
repair process that includes: (a) immediate secretion of pre-formed epidermal lamellar bodies; (b) replenishment of the cytosol with newly formed lamellar bodies and further secretion of lamellar body contents; and (c) formation of lamellar bilayers in the stratum corneum interstices leading to recovery of the barriertotransepidermalwaterloss within 24h.Epidermal lipid synthesis is required to sustain this process (13) and DNA
synthesis is also regulated by barrier requirements (24). The increase in lipid synthesis is dueto increased activities of key enzymes of cholesterol synthesis, 3-hydroxy-3-methylglutaryl CoA (HMG CoA) reductase (25), ceramide synthesis, serine palmitoyl transferase (26),andfattyacidsynthesis, acetylCoA
carboxylase andfattyacidsynthetase (27).
The Journal ofClinicalInvestigation, Inc. Volume95, May 1995, 2281-2290
We therefore compared epidermal barrier function in aged (>80 yr) vsyoung(20-30 yr)human subjects aswell as in a newmurine model of skinsenescence.We describehere that
both barrier integrity (ability to withstand perturbation) and
barrier recovery after acute perturbations are altered in aging
epidermis, despite normal or subnormal basal TEWL rates. These changes in functioncanbe ascribedto abnormalities in
lipid content, which inturn, result in insufficient formation of
intercellular lamellar bilayers in senescent stratum corneum.
Methods
Human andanimal studies
Human volunteers. 21 human volunteers, 6 older than 80 yr and 15 younger than 30 yr gaveinformedconsent to participate in these studies.
Volunteers were in good general health and free from metabolic dis-eases, such asdiabetes, thyroiddisease, orhyperlipidermia, as well as
atopic dermatitis. None were taking oral corticosteroids, nonsteroidal
antiinflammatorydrugs, or oral hypoglycemicagents.
Mice. Hairless miceoutbredCrL:SKHI(hr/hr) BR (Charles River Laboratories, Wilmington, MA) were used for the senescent murine model because the othercommercially available strains that we tested
frequentlydeveloped tumors and died before the development of chro-nological senescence. Young mice were between 6 and 10 wk of age, and aged mice were 18-24 mo old at the time of study (normal life
expectancyof hairless mice is 24mo).Theagedmicewerechecked regularly for microbial diseases and tumors. Animals that showed evi-dence of eithersystemic illnessor tumordevelopmentwere notstudied.
Functional studies
Basal TEWL(human andmurine). Basal TEWL was assessed at three
sitesonthe volar forearmofthe humanvolunteers, usingan
evaporime-ter(Servomed's,Vdllingby, Sweden)and recorded ing/m2per h(28). TEWL also wasdetermined overthree different sitesin seven-young
andseven-senescentmiceusinganelectrolyticwateranalyzer(Meeco,
Inc.,Warrington, PA),andrecorded inppm/0.5cm2perh,and
subse-quentlyconverted tog/m2per h.Temperaturerangedbetween200and
22°C,relativehumiditybetween 42and49%,andatmosphericpressure between9.4-12.6mmHgatthe time ofmeasurements.
Barrier integrity. Baseline TEWLmeasurements weremadeateach of these sites on the volarforearmusing the evaporimeter, andeach site then was strippedsequentially with cellophane (Tuck;TesaTuck
Inc.,NewRochelle, NY)tape. TEWLwasassessedoverthesamesites
after eachfive tapestrippings.Barrierintegrity (the abilityof thestratum
corneum to resist barrier perturbation) is reflected bythe numberof
tapestrippingsrequiredtoattainTEWLlevelsof2 20g/m2per h. Barrier recoveryexperiments (human). A site 4cmbelow the
antecubital fossaonthe volar forearm of bothagedandyoungsubjects
wasused for recovery studies. Onearm wastreated with acetone-soaked cottonballs,while thesamesitesonthecontralateral forearmservedas
the control. We also used recovery after tape stripping as a second method to control forpossible toxic effects ofacetone. TEWL was
determined0,3, 6, 24, 48, 72,and sometimes 96 h after barrier
disrup-tion.
Barrier recoveryexperiments(murine). One flankwastreated with acetone-soakedcotton balls or withsequential tape stripping, as de-scribed above, and TEWL wasmeasured 0, 3, 6, 24,and 48 h after barrierabrogation. Skinbiopsieswereobtained ateach ofthese time
pointsfromparallelgroups ofsimilarly treated animals for the
histo-chemical studies described below, aswellasfor electronmicroscopy withreduced osmiumtetroxideasdescribedbelow.
Data shownaremean±SE. Statisticalsignificancesweredetermined
usinganunpairedStudent's ttest.
Electronmicroscopy. Shavebiopsies,obtained under local
anesthe-sia, were taken from the normal appearing, non-sun-exposed, volar
forearm orthigh skin of five of the aged (>80 yr) andeight of the
young (< 30 yr) human volunteers. Prior studies have shown that thigh andforearmdisplaysimilar barrier repair kinetics (29). Samples were fixedin half-strength Karnovsky's fixative, divided, and processed rou-tinelythrough either reduced 1.0% osmium tetroxide or 0.25% ruthe-niumtetroxide (30), followed by embeddment in an Epon-epoxy mix-ture(31).Full thicknessbiopsiesfromthe flanks ofboth aged (> 18
mo), and young(6-10wkold) hairless mice also were processed, as described above. Ultrathin sections were viewed in an electron micro-scope(1OA; Carl Zeiss, Inc., Thormwood, NY) after further contrasting inlead citrate and uranyl acetate. The quantities of lamellar bodies in the cytosol were quantitated by standard stereological methods (32), utilizing both randomly obtained sections and blinded data collection.
Histochemical localization of neutral lipids. A saturated solution of sudan red 7B in 70% ethanol was used to stain for neutral lipids.4-sm
fresh frozen cryostat sections were allowed to dry on the microscope slide and then the slides were transferred to the stainfor 10min at room temperature, followed by sequential rinsing in 70% ethanol and distilled water(33).The sections then were counterstained inhematoxylinfor 30 s at room temperature, rinsed in distilled water, and mounted for
microscopicexamination. Forfluorescence localizationof lipidsa stock solution of nile red (500
j.g/ml)
in acetone was prepared, stored at-20TC, and protectedfromlight. A fresh stainingsolution ofnile red waspreparedby the addition of 15-20ulof the stock solution per 1 ml 75% glycerol followed by brisk vortexing. Fresh frozen cryostat sections werestained,covered with aglass coverslip,andexamined after
10 min at room temperature in thedark, withafluorescentmicroscope (Ortholux II; E.Leitz,Inc., Rockleigh, NJ),using 450-nmexcitation and 500-nm emissionfrequencies (34).
Lipid content analysis. Full-thickness murine skin from both age groups was heatsplit at60°Cfor1 min to remove intactepidermis from
dermis.Nucleatedepidermal cells thenwereremoved from the stratum comeum by floating the full-thickness skin, basal sidedownward, in 0.5% trypsinin PBS for30 min, followed by vortexing. Thismethod leaves theresidual stratum corneumsheetfree ofnucleated cells (35,
36).Boththestratumcomeumwetweight,aswellasthe initialsurface
area wererecorded,and used to calculateyields.The lipidswere ex-tracted by the method ofBNligh andDyer (37), dried, weighed, and stored at-70°Cuntilanalyzed.Lipidswerefractionatedandquantified by highperformancethinlayerchromatography(HPTLC) followed by charringandscanningdensitometry, usingamodification of themethod described byPonecetal.(38). Lipidsweresolubilizedin
chloroform-methanol (2:1vol)andspottedonHPTLCplates(E.Merck, Darmstadt, Germany) usingaCAMAGLinomat IVAutospotter(CAMAG
Scien-tific Inc., Wilmington, NC). The platesweredeveloped in CAMAG
horizontal developing chambers at4°C. Neutral lipids were resolved usingtwosolvent systems.The plates firstweredeveloped three times
inpetroleumether:diethyl ether:aceticacid(80:20:1 vol)at4°C for 30
mm, 50 mm, and 70 mm, respectively, and then developed once in petroleumether aloneorupto98mm.Ceramides andglucosylceramides
were resolved using first chloroform/methanol/acetone/acetic acid (76:20:4:1 vol) for 20 mm, then chloroform/methanol/acetone (80:10:10vol)for 50 mm, andfinally
chloroform/diethylether/ethyl-acetate/methanol (76:6:20:2 vol)for98 mm. After development the plates weredippedinacharringsolutionofglacial acetic
acid/concen-trated sulfuricacid/ortho-phosphoricacid/distilledwater(1:0.2:0.2:1.9
vol)containing 1.5% wt/vol coppersulphate,drainedbriefly,and then
dried for 10minon a50°Chotplate. Plates thenweretransferredto
an 180°C oven for 15 min. After cooling to room temperature, the
plates were scanned immediately and quantitated using a automated
TLC Scanner IIequippedwithaLaboratoryDataSystemand"CATS"
software(bothfrom CAMAGScientific Inc.). Statisticalsignificances
weredeterminedusingaStudent'stwo-tailedt test.
Gasliquidchromatography of epidermal lipids. Full-thickness skin
from senescent hairless male micewas collected, the epidermis was
splitfrom theunderlyingskinbyexposuretodryheat(60°Cfor 60 s),
and theepidermallipidsextracted withBligh/Dyersolvents,asabove.
Epidermallipids weresubfractionatedbyHPTLC inpetroleumether/
diethylether/acetic acid(80:20:1vol),asdescribed above. Theneutral
lipids were pooled and subsequentlymethylated in boron trichloride
(12% wt/wt)overnightat800C.Thepooled methyletherswereresolved andquantitated on acapillarycolumn 30m X0.25mminnerdiameter,
0.2
MiM
film (SP2380; Supelco Inc.,Bellefonte,PA) in a gas chromato-graph(5890; Hewlett-PackardCo.,PaloAlto, CA), equippedwith anautosampler, under thefollowingconditions: column temperature 8 min at 150'C then increased by 3TC/min to 1900C and held for 30 min; linear velocity 20 cm/s; vector type: FID (2 X 1011 attenuated full spectrum);split ratio: 100:1; 9psihead pressure. Peakswereidentified
against authentic methylester standards from Supelco or fatty acids from Sigma Chemical Co. (St. Louis, MO), methylated asdescribed
above.
Results
Structure,function, and lipid compositionof aged
epidermis under basal conditions
Basal barrier function.We first determined whetheraged epi-dermis displays abnormalities in barrier function under basal conditions. Under basalconditions, agedhuman skinappeared paler, with arougher, drierappearing surface than in the young cohort. In comparison toyoung mice, aged mice displayed a
thickened, yellow appearing integument, withacoarse,dry
sur-face, as well. Inboth humans and inmice, basaltransepidermal water loss rates were slightly decreased in aged vs younger individuals. MeanTEWLrates inaged humanswere4.43±0.16 g/m2 perh vs a mean of6.41±0.93 g/m2 per h in the young (P < 0.05). Likewise, basal TEWL rates were decreased in aged vs young mice (aged: 0.24±0.07 g/m2 per h vsyoung: 0.74±0.13 g/m2per h;P < 0.01). These results demonstrate that barrier function, assessed by TEWL, is not deficient in intrinsically aged skin under basal conditions.
Stratum corneum membrane structure and the lamellar
body secretory system. We next evaluated the quantities and
internal structure of the lamellarbody and its secreted contents in aged epidermis. Electron microscopy of the epidermis of young humans and mice appeared similar, with comparable numbers of lamellar bodies in the granular cell cytosol, and abundantsecreted lamellarbodycontents atthestratum
granulo-sum-stratum corneuminterface (Fig. 1 F, human). Moreover,
as previously described (30), extensive multilamellar arrays, displayingacharacteristicbasic unit structure, were present at all levelsof thestratum corneuminterstices(Fig. 1 G, human).
Inaged human and mouse epidermis both the internal
struc-tureoflamellar bodiesaswellastheir numbers inthe granular cellcytosolweresimilartoyoungsubjects (Fig. 1 E, mouse).
Stereologicalmeasurements showed that lamellar bodies
occu-pied 5.0±0.4vs 5.1±0.6 (mean±SEM; n = 8 samples each) percent of the granular cell cytosol in young vs old human epidermis, respectively. However,apaucityof secreted lamellar bodycontents waspresentatthestratumgranulosum-stratum
corneuminterfaceinaged epidermis (Fig. 1, C [mouse] and D
[human]).Moreover, thestratumcorneumintersticescontained focal domains thatwereeither depleted or virtually devoid of lamellar structures (Fig. 1,A andB, human). Yet, where the lamellar bilayers werepresent in eithernormal or diminished numbers, the lamellar arrays themselves displayed both normal substructure and dimensions (interlamellar dimension [hu-mans]: young = 13.2 nm; aged = 13.3 nm). These results demonstrate that aged epidermis displays adiminution in
se-creted lamellarbody-derived contents, and a failure of these
contents to form acontinuous series of multilamellar bilayers
within the intercellularspacesof thestratumcorneum.
Lipid composition of aged murine stratum corneur. To delineate thelipidbiochemical basis for the structural abnormal-ities, we next compared both the lipid content and the lipid distribution ofsenescent vs young murine stratum corneum.
Lipidextractsdemonstratedasignificantdecrease(~-30%)in
thetotal lipidcontentin aged (0.120+0.13 jg/cm2)vsyoung
(0.177+0.02 jg/cm2;P<0.05)murinestratumcorneum(Fig.
2). In contrast, the lipid distribution of senescent vs young murinestratumcorneum showedasimilarprofileand
distribu-tion of themajorneutrallipids (free sterols, sterolesters,free
fatty acids, andtriglycerides)andsphingolipid species,
includ-ingacylceramides (ceramide 1) (Table I). These results
indi-catethatepidermal aging isaccompanied byaglobaldecrease
in stratum corneum lipid content, resulting in a comparable reduction of each of themajor lipid species. However,no
selec-tive reductionof any specific lipid speciesoccurs under basal conditions.
Fatty acid composition ofsenescentepidermal lipids. We
nextexamined thefattyacid distribution of the combined neutral
and sphingolipid species to determine whether a decrease in
either essentialfatty acidsorin verylongchain speciesexists inagedepidermis.As seeninFig.3 the spectrum offattyacids in senescent murine epidermal lipids does differ from young controls, i.e.,itdisplays adisproportionateincreasein medium
andlong chain(C14-C18) species.Moreover, there isno
corre-spondingdecrease in any of the verylongchainspecies
(C20-C24), orin linoleic acid (quantitativedatanotshown). Thus, analysis of the constituent fatty acids of the epidermal lipids does notreveal anysignificant abnormalityin aging epidermis under basal conditions.
Dynamic studies
of aged epidermal
barrierfunction
Barrierintegrity. We nextassessed theintegrityof thestratum
corneum in agedvs youngsubjects. Whereas only 18+2 tape strippingswererequiredtoperturbthebarrier of the volar
fore-armsofagedhumansubjects,asignificantlygreater number of tapestrippings (31±5;P< 0.05)wererequiredtoperturbthe barrierover comparable sites in the younger human subjects. Althoughdifficulttocompareaccuratelybecauseof differences in eachinvestigator's technique,and becausedifferentamounts
ofacetone may actuallycontact the test area, only 3-10 min ofacetonesoaking wasrequiredfor barrierperturbationin the agedvs30-60 mintoperturbthe barrier in the young.Likewise, the barrier in aged mice was more readily disrupted by tape strippingthan in younger mice(- 3vs - 5strippings,
respec-tively,foragedvsyounganimals). Thesestudies demonstrate
that, while barrier function appearsto be normal under basal
conditions, theintegrity of the stratumcorneum is diminished inchronologically aged epidermis.
Barrierrecovery. Todetermine whether themeasurements
ofTEWLunder basal conditions maskedanunderlying abnor-malityinbarrierfunction,we nextcomparedtherateof barrier recovery in aged vs young humans and hairless mice. After
comparableinsults (TEWL 20-30 g/m2 perh)the epidermis
of agedand young each display abarelydetectable shiny ap-pearance, withnoobviousdamagetotheepidermis.Thischange isno longervisible in either youngoroldat subsequenttime
points. DespitesubnormalTEWL ratesunderbasalconditions,
barrier recovery after comparable insults was much slower in
aged than in younghumans, regardless of the mode of barrier insult, i.e., acetone treatment or tape stripping (Fig. 4,A and
LIPIDCONTENT LIPIDDISTRIBUTION
(igg
/ cm2) (%)Young
NL 0.177 ± 0.024 69.23±8.2
SL
.38±3.27
14.39
nonrecovered P < 0.05 -32%Total
Lipid
NL
Aged
66.25+8.46
0.120 ±0.13
SL
16.57±3.
nonrecovered
Figure 2. Lipidcontentofaged murinestratumcorneum. Totallipid
content wasdecreasedby 32% in agedmurine stratum corneumvs
youngstratumcorneum. However, thelipid speciesdistribution
re-mained unchanged. Data shown are mean±SEM. n= 6.
onlyrecoveredby 15% at 24 hafteracetone treatment, barrier function inthe youngersubjects typicallyrecoveredby50%(P
< 0.01). Moreover, whereas the youngerbarrier recovered by
- 70-80%by 48 h, 5-6d wererequiredforagedskin to attain this levelofbarrier recoveryregardlessof the typeofinsult. In
the agedthe barrier had recovered by >90% by 7d, whereas in the young, >90%recovery was seenby4d.Barrier recovery in aged vs young hairless mice (Fig. 5) displayed a similar difference, although both age groups recovered over shorter
timeperiods thandid human subjects. Forexample, while the
aged barrier only recovered by 15% by 6 h, itrecovered by 50% in younger animals (Fig. 5, P < 0.001). These results
demonstrateaprofound abnormality of barrier functioninaged epidermis that only becomes apparent when the organ is
stressed, i.e., followingbarrierdisruption.
Stratum corneum neutral lipid distribution (Fig. 6). To determine the structural basis for the functional abnormalities
in aged epidermis,we nextassessed the distribution ofstratum
corneumlipids beforeandafter barrierdisruption. Immediately before barrierdisruption, asimilar pattern of neutrallipid
stain-ing was seen in the stratum corneum ofboth aged andyoung
hairless mice (Fig. 6, A and B). Moreover, immediately after
acetonewiping, histochemical staining forneutral lipids using
sudan red 7B was absent in both young and oldmurinestratum
TableL Lipid DistributioninAgedvs YoungMurine
Stratum Corneum Lipid Content
Young Aged %total lipid+SEM %total lipid±SEM
Totalsphingolipid 16.38+3.27 17.78+3.56 Acylceramide 7.52±+1.38 8.18±+1.56
CeramideIII 2.30±0.42 2.30+0.34
Ceramide IV 2.43+0.66 2.48±0.41 Acylglucosylceramide 0.82+0.35 0.80±0.42 Totalneutral lipid 69.23+8.2 66.25+8.46 Total identifiedlipid 85.61+5.68 84.0±2.53
corneum. Whereas at 3 h no visible return of lipid could be seen yet in eitheryoung or old stratum corneum, by 6 and 16 h significant amounts of stainable neutral lipid material had
reappearedin young murine stratum corneum(Fig. 6, Cand E, respectively).In contrast,staining didnotreturn toagedmurine stratumcorneum until after 16h(Fig. 6, D-G). By24 h
equiva-lent amounts of lipid appeared to have reaccumulated in the stratumcorneum of bothyoung andagedstratumcorneum(not shown), and both were indistinguishable from untreated
con-trols. Asimilardelay inreappearanceof neutrallipidswas seen infrozen sections ofsenescentmurinestratumcorneumstained
with the hydrophobic fluorophore, nile red, at 6 h (data not
shown). Theseresultsshow that the abnormalityin barrier
re-covery in aged murine epidermis is accompanied by a
corre-sponding delay in the reappearance of stainable neutral lipids
in the stratum corneum.
Lamellar body secretory response to barrier disruption. Electron microscopy with osmium tetroxide after fixation was used todetermine the lamellarbodyresponse tobarrier
disrup-tion. 3 hafter barrierdisruption largenumbersof newly formed
lamellar bodies, filled with abundantlipid, werevisible in the stratum granulosum of young murine epidermis (data not
shown; seereference39).Intheaged there also wasabundant
lamellar body formation, but these structures were not replete
with lipid and many contained only a few bilayer structures
(Fig. 7). Furthermore, after 3 h,whilethestratum
granulosum-stratum corneum interface in the young contained extensive
lipid material (see reference 39), the stratum
granulosum-stratumcorneuminterface in theagedis still devoidoflamellar
material (Fig.7).
Discussion
Becausepermeability barrierfunction in aging epidermisdoes not appear to be impaired under basal conditions, it has been
Figure 1. (A-E)Aged human and murine epidermis.A andBdemonstrate abnormalitiesinintercellular lamellae in the stratum corneum (SC) of
twodifferentagedhumansubjects. Note decreased number of lamellae along extended membrane regions (single arrows), althoughinsome areas
normalnumbers of lamellaearepresent(double arrows). Extensive intercellular clefts (asterisks) are also more common in aged than in young
SC. C-E demonstrate normal numbers of lamellar bodies (C, arrowheads) in aged stratumgranulosum (SG), and lamellar bodieshavenormal internalstructure(E, arrowheads) (C and E, human; D, mouse). However, the quantity of secreted lamellar body contents at theSG-SCinterface appearstobe reduced in bothagedhuman (C, single arrows) and murine epidermis (D, singlearrowsdepict apical plasmamembraneof SG cell withnoevidence of secreted intercellularlamellae).FandG depict young human epidermis for comparison. Note abundant intercellular lamellae inSC(G, double arrows) and secreted lamellar bodycontents atSG-SC interface (F, single arrows). A, x55,000;B, x60,000, C, x21,000;D,
A
II
1 1
111U
1
1
11
L1
generally assumed that barrier function is not altered signifi-cantlyin aging (40).Thecauseofthe reducedratesof TEWL
ofagedepidermis under basal conditionsareunknown.Possible
causesofadecreased TEWLcould includedecreasedsweating,
decreased microcirculationresultingin decreasedintegumental moisturecontent,decreasedskintemperature,orincreased
cor-neocytecell surfacearea.Although neither decreased sweating
(41),decreased microvasculature(42),nordecreased skin
sur-face temperature (14) alone can account for the decreased
TEWL inaged epidermis.Acombination ofthese factorscould be operative. In contrast, it has been long known that aged epidermis is abnormally permeabletoavariety ofdrugsin in
vitro systems (14-16). Thus, it seemed likely to us that the
seemingly normal and even subnormal levels of TEWL
dis-played by intrinsically aged epidermis mightmaskan
underly-ing lack of functional reserve. The functional, structural, and
lipidbiochemical studies shedlight onthis apparentparadox.
When theepidermal permeabilitybarrier issubjectedtostress,
i.e., barrier abrogation, marked differences in the function of
intrinsically aged vsyoungstratum corneumappear.
Inseekingthebasis for the functionalabnormality,wefound that bothagedhuman andmouseepidermis displayedanormal number of lamellar bodies in the granular cell cytosol with normal internalstructure.However,apaucityof secreted lamel-larbodycontentswas present atthestratum
granulosum-stra-tumcomeuminterface. Moreover, the stratum corneum
inter-stices contained focal domains thatwereeitherdepletedor
virtu-allydevoid of lamellarstructures. Yet, althoughdiminished in
numbers, the lamellar arrays themselves possessed a normal substructure. Thesefindingsmatch thelipidbiochemical obser-vations which showed thatstratumcorneumlipidsinagingare
globally reduced in quantity, without exhibiting any specific abnormalityin speciesdistributionorfattyacidcomposition.It is difficult to compare our results with prior reports because
mostof these studiesonly sampledthesuperficialstratum
cor-Figure 3.Fatty acid contentofsenescentepidermal
lipids. There isadisproportionate increasein
me-diumandlongchain(C14-C18)speciesin aged .
Z,
(A)vs young(Y)murineepidermis. However,nochangesareapparentin verylongchainspecies
(C20-C24),nor inlinoleicacid(C18).
neum (surface lipids) with either solvents or tape stripping (e.g., 18, 19). One of these workers reported a decrease in sterolestersandtriglycerideswithaging(18). In another recent study, onlyceramidecontentwasassessed (19),butthis study, usingasinglecyanoacrylate stripping ofstratumcorneum, dem-onstratedadecrease in the ceramidecontentinnormal human-aged skin. These workers then correlated the decrease in cera-midecontentwithadecrease insphingomyelinaseactivity (43). We, too, found that ceramides were decreased. However, the decreasewas notspecific forsphingolipids,but insteadextended
toeach of the majorlipid speciesanalyzed.
Whereasthese studies have shown that baselinefunctional parameters are only minimally altered in aged skin, dynamic
assessments revealedprofound abnormalities in thefunctional capacityof the agedepidermalpermeabilitybarrier. For
exam-ple, barrierintegrity, as indicated by resistance to sequential tapestrippingor acetonewiping,ismarkedlyreduced inaging
epidermis. Agedstratum corneumhas been notedpreviouslyto
exhibit altered cohesiveness (44), which could result in in-creasedamountsof material obtained with each tapestripping. The decreasedintegrityof the barrier inintrinsically aged stra-tum corneummay relatetothis alteredcohesiveness, which in turn, could be due to the decreased intercellularlipidcontent
thatwehave described here.Whether the decreasedlipid con-tentandfocalabnormalities in the intercellular spaces ofaged
stratumcorneum eitheraccount fororcontributetothe abnor-mality inintegrityremains speculative.
In addition to alteredintegrity, barrier recovery isdelayed in aged human and murine skin after barrier abrogation by eitheracetone or tape stripping.
Delipidization
or removal of theepidermalpermeabilitybarrier isfollowedbyarapid phase of barrier recoveryover6 h in young hairlessmice,
followed byaslower recoveryphase requiring 30-36htoreach comple-tion(45).Whereas theacuterecoveryphasestimulatesepider-mal cholesterol andfattyacidsynthesis(46),the late recovery
2286 Ghadially, Brown, Sequeira-Martin, Feingold, and Elias
Aa..0i
!A
kA-4-..
A. Acetone
1 T
X
Ih-
AgedN s 0
Nw X' t
- Young
o N d t0 0 0 N t X 0 0
T im' 0 0) 0 Nm C) IV co Time (hours)
-3- Aged * Young
Time (hours)
Figure 4. Human barrierrecovery. (A) Human barrierrecoveryafter
acetonewiping:recoveryisdelayedintheaged (P< 0.01 atall time
points). Note thatat24 h theagedbarrieris recoveredonly 15% while theyoungbarrier isalmost50% recovered. (B) Human barrierrecovery
aftertapestrippingrevealsasimilardivergentpattern of barrierrecovery
inagedvsyounghumanepidermis. Data shownaremean±SEMat
each time point. n= 5.
phase is accompanied by increases in sphingolipid (47) and DNA(24) synthesis.This burst insyntheticactivityis
accompa-nied byareturnoflipidto thestratumcorneuminterstices, as
shown by histochemical techniques (45), and an early burst in lamellar body synthesis and secretion, as seen by electron
microscopy (39). Thus,theaged epidermal permeabilitybarrier isboth easiertoperturb and slowertorepair.This isaconsistent
findingin both humans andmice, usingtwodifferentmethods
ofacutebarrierperturbation.Theuseofbothacetoneandtape
stripping parallel studies allowedustocontrol for thepossible
unwanted effects of each model. Hence, neither cytotoxicity
from acetone, nor the added variable of cellularreplacement
with tape stripping are likely to account for our results. The
basis for thedelayinbarrierrecoveryinaging epidermisappears
to be a delay in the return of stainable neutral lipids to the
stratumcorneuminterstices,asdemonstratedby histochemistry
and electronmicroscopy.Thedelayinlipid reaccumulation,in
~80-0
40-020
0 1 0 20 30 40 50
Time
(hours)
Figure5. Murinebarrierrecovery. Murine barrier recovery after tape
strippingisdelayedinagedvsyoungepidermis (P<0.01).Recovery
takesplace withasimilarpatterntohumanepidermis (comparewith
Fig.4), but over ashorter time periodthan in humans. Data shownare
mean±SEM at each timepoint. n= 10.
turn, may be attributed to diminished lamellarbody secretion, andthereforeadecrease inlipiddeliverytothestratumcorneum interstices. The retardation in lipid delivery could lead to a
delay in the normalrateof lamellarbilayer reformationin the
intercellular spaces. Thus, when the epidermal permeability
bar-rier is subjected to stress, i.e., barrier abrogation, there are
markeddifferences in the function ofintrinsically agedvsyoung
stratum corneum. Our results show further that measurements
of basal TEWL donotreflect the actual functionalstatusof the tissue,and can even bemisleading. In this respect, skin, as is the case with other organs (21-23), may display profound abnormalities only when the organ is stressed or undergoes repair processes. Mostimportantly,measurementsof basal func-tionprovide noinsights about thecapacity of theepidermisto
maintain barrier homeostasis whensubjectedtoinjury,chemical exposure,orenvironmentalstress.
The metabolic basis for the biochemical and structural ab-normalities inagingstratumcorneum is not known. Our
previ-ous studies have shown that an acute burst in both lipidand
DNA synthesis follows barrier abrogation (reviewed in 46). Moreover,epidermispossesses thecapacitytosynthesizeafull spectrum oflipids,including ceramides (48) cholesterol (49), and freefatty acids (46).Thesynthesisof all threekeylipids is notonly modulated by barrierrequirements (46), but also required for barrier homeostasis (27, 50). The burst in lipid biosynthesis after experimental perturbation leads to lipid
re-plenishment in the stratumcorneum, which is followed by a
return of barrier function (45). The increases in cholesterol, ceramide,andfattyacidsynthesis that accompany barrierrepair resultfrom a corresponding increase in the total activity of the rate-limiting enzymes for these lipids, HMG CoA reductase (25), serine palmitoyl transferase(26),andacetyl CoA
carbox-ylase(unpublished data). Recentstudies also have shown that
barrier disruption induces a rapid increase in the mRNA for HMG CoA reductase (51 ),aswell aschangesinthe phosphory-lationstateof this enzyme (25).Futurestudies should beable
to determine whether the decreased ability of aged epidermis
torepairthe barriercanbe attributedto specificorgloballipid metabolicabnormalities, andto determine the molecular bases for these alterations.
100 90
80
'70
60
50- 40- 30-20
10
0
b.
0
a
0)p
0
e
0a
Y-Pre
0-Pre
> i'. 9' 4ni k +1*s *J ~ t
Y-6
0-6
C4
jit.iMn
Y-16
A1-16
-V C~~~~~~~~~~~~~~~~~~~~~~t
F
A~~~~~~~~~~~~~1
"v-ILE
G
Figure 6. Histochemical localization of neutral lipids in young (Y) vs old(0)munineepidermis. A and B are before and C-G after acetone treatment. Immediately before acetone treatment the stratum corneum (A andB, brackets) of both young and old reveals abundant neutral lipid. By 6 hafter acetone treatment, substantial neutral lipid has returned to the stratum corneum of young animals (C, arrows), while old stratum corneum displays no return of neutral lipids (D, brackets). By 16 h, more neutral lipid staining has appeared in young stratum corneum (E, arrows), but aged stratum corneum at 16 h still displays virtually noreturnof neutral lipid(F, brackets). Only at 19 h, does stainable lipid begin to appear in stratum corneum of aged mice (G, brackets and arrows). Note intense staining of the sebaceous glands, which serve as a positive control. Open
Figure 7. Lamellar body response to murine barrier disruption. 3 h after barrier disruption in the aged the stratum granulosum-stratum corneum junction is devoid of lamellar structures (open arrows). Although newly formed lamellar bodies are evident, their contents are diminished or virtually absent (closed arrows). x75,000.
Finally, we have described here a new murine model of cutaneoussenescence.The Charles River hairless mouse strain displays a normal, murine life expectancy without the premature development of neoplasms, a common occurrence with other commercially available, inbred, hairless mouse strains. By all parameters examined to date, the epidermal abnormalities in this model parallel those that occur in the intrinsically aged epidermis of humans of advanced age (> 80 yr). In both spe-cies, we did not find consistent abnormalities in individuals of less advanced age (humans< 80 yr; mice< 14 mo;Ghadially,
R. and P. M. Elias, unpublished observations.). Thus, the changes described in this report are detected only with relatively advancedsenescence. Whether the program ofchronologic
se-nescence is accelerated when chronic photodamage is superim-posed upon intrinsic aging remains to be determined.
In conclusion, in aged skin, functional assessments under basalconditionsdonotprovide an accurateassessmentof epi-dermal function.Incontrast,stress tothe system reveals under-lyingalterations inepidermal integrity and barrier recovery after experimental perturbation. A global decrease in lipid content
leadstoalterations in lamellarbilayer morphology which appear
tounderlie the impaired functional abnormalities. In addition, wehavedescribed a new murine model of skin aging which may beuseful for future animal studiestoelucidate the biochemical mechanisms responsible for epidermal aging. Finally, these findingshavepotentialclinical significance. First, the decrease
inlipidcontentmay contribute to thecommonly xerotic skin
of theaged. Second,the decreased barrierintegrity implies an
increased susceptibilityto theinsults of the environment (e.g.,
decreasedhumidity, solvents, and detergents), as well as a di-minished capacity to recover from environmental insult.
Acknowledgments
Wegratefully acknowledge the excellent technical assistance of Debra A.Crumrine and Deborah Lawson, M.D.
This work was supported by the Dermatology Foundation, the Veter-ansAffairs Medical ResearchService, the Lester I. Conrad Research Foundation, and National Institutes of Health grants AR19098, AR
39448,and AR39639.
References
1.Christopher, E., and A. M. Kligman. 1965. Percutaneous absorption in aged skin. In Advances inBiologyof the Skin. W. Montagna editor.Pergamon Press,
Inc.New York. 163-175.
2.Grove, G. L., S. Duncan, and A. M. Kligman. 1982. Effect of aging on the
blisteringof human skin with ammoniumhydroxide.Br. J.Dermatol. 107:393-400.
3. Lavker, R. 1979. Structural alterations in exposed and unexposed skin. J. Invest.Dermatol. 73:259-266.
4.Lavker, R., P. Zheng, and G. Doug. 1987. Aged skin: a study oflight,
transmission electron, and scanning electron microscopy. J. Invest. Dermatol.
88:44s-5 1s.
5.Black,M. 1969. A modifiedradiographicmethodformeasuringskin thick-ness.Br. J.Dermatol.81:661-666.
6.Shuster, S.,and M. M. Black. 1975. The influence of age and sex on skin
thickness,skincollagenanddensity.Br.J. Dermatol.93:639-643.
7.Marks,R. 1981. Measurement ofbiological aginginhumanepidenrnis.Br. J.Dermatol. 104:627-633.
8.Thune,P.1989. Evaluation of thehydrationand thewater-holding capacity
9.Elsner,P., D.Wilhelm,and H. I.Maibach.1990.Frictional propertiesof humanforearmand vulvarskin:influenceofage andcorrelationwith
transepider-mal waterlossandcapacitance. Dermatologica(Basel). 181:88-91.
10.Wilhelm,K.-P., A.B.Cua, and H.I.Maibach. 1991.Skin aging.Effect
ontransepidermalwaterloss,stratumcorneumhydration, skin surfacepH,and casualsebumcontent.Arch.Dermatol.127:1806-1809.
11.Cua, A. B., K.-P. Wilhelm, and M. I.Maibach.1990. Frictional properties of humanskin: relation to age, sex and anatomical region, stratum corneum hydrationandtransepidermalwater loss. Br. J.Dernatol. 123:473-479.
12. Potts, R. O., E. M.Buras,and D.A.Christman. 1984.Changeswithage inthemoisture content of human skin. J. Invest. Dermatol.82:97-100.
13. Menon,G. K., K. R.Feingold,A. H. Moser, B.E.Brown, andP.M.
Elias. 1985. De novo sterologenesis in the skin. HI. Regulationbycutaneous
barrierrequirements. J.Lipid Res.26:418-427.
14.Leveque,J. L., P.Corcuff, J. deRigal,and P. Agache. 1984. In vivo
studies oftheevolution of physicalproperties of thehumanskinwith age.Int.J.
Dermatol. 23:322-328.
15.Tagami,H.1972.Functional characteristicsof aged skin. ActaDermatol. 66:19-21.
16.Malkinson,F. D. 1958. Studiesonthepercutaneous absorptionof14C
labelledsteroid byuseof thegasflowcell.J.Invest.Dermatol.31:19-28. 17.Grubauer, G.,K. R.Feingold,R. M.Harris, andP.M.Elias. 1989. Lipid
contentandlipidtype asdeterminants of theepidermal permeability barrier.J.
LipidRes.30:89-96.
18.LUger,D. S.,A.M.Frangois,J. L.Leveque,T.J.Stoudemayer, G.L.
Grove,and A. M. Kligman. 1988. Age-associated changesin stratum corneum
lipidsandtheir relationtodryness. Dermatologica (Basel). 177:159-164.
19.Imokawa, G., A.Akihito,J. Kumi,Y. Higaki,M. Kawashima,and A. Hidano. 1991.Decreasedlevelofceramidesinstratumcorneum of atopic
derma-titis: anetiologicfactor inatopic dryskin? J. Invest. Dermatol.96:523-526. 20. Stewart,M.E.,and D. T. Downing. 1991.Chemistryand function of
mammalian sebaceouslipids.In AdvancesinLipidResearch.Skin Lipids.P.M.
Elias,editor. AcademicPress,NewYork.282-283.
21.Katzung,B.G.1992.Specialaspects ofgeriatricpharmacology.In Basic and ClinicalPharmacology.B. G.Katzung,editor.Appleton&Lange,East
Nor-walk, CT.862-863.
22.Davies,I. 1992.Theoriesandgeneralprinciples ofaging.InTextbookof
Geriatric MedicineandGerontology.4thed. J. C.Brocklehurst,R. C.Tallis, and
H. M.Fillit,editors.Churchill-Livingstone, Inc.,New York.26-60.
23.Miller,R. A.1994.Thebiologyofagingandlongevity.InPrinciplesof
Geriatric Medicine andGerontology. 3rd ed. W. R. Hazzard, E. L.Bievman,
J. P.Blass,W. H.Ettinger,and J. B.Halter,editors. McGraw-Hill, Inc.,New York.3-18.
24.Proksch, E.,K. R.Feingold,M.Mao-Qiang, andP. M.Elias.1991. Barrier functionregulates epidermalDNAsynthesis.J.Clin. Invest. 87:1668-1673.
25.Proksch, E., P. M.Elias, and K. R. Feingold. 1990. Regulationof
3-hydroxy-3-methyl-glutarylcoenzyme Areductase activityinmurineepidermis:
modulation of enzyme content and activation stateby barrier requirements. J.
Clin.Invest. 85:874-882.
26.Holleran, W. M.,K. R. Feingold, M. Mao-Qiang, W. N. Gao, J. M.
Lee,and P. M.Elias. 1991. Regulationofepidermalsphingolipidsynthesisby permeability barrierfunction. J. LipidRes.32:1151-1158.
27.Ottey, K.,L.C.Wood,and P. M. Elias. 1995.Effectofcutaneous
perme-ability barrier distruptiononfattyacidsynthetic enzymesin the epidermisof
hairlessmice.J.LipidRes. In press.
28.Blichmann,C.W.,and J.Serup.1987.Reproducibilityandvariabilityof
transepidermalwaterloss measurement. ActaDernato-Venereol. 67:206-210. 29.Ghadially, R.G.,L.Halkier-Sorensen,and P. M.Elias. 1992.Effects of petrolatumonstratumcorneumstructureand function. J. Am. Acad.Dermatol. 26:387-396.
30.Hou,S. Y.E.,A.K.Mitra,S. H.White,G. K.Menon,R.Ghadially,and P. M. Elias. 1991. Membranestructuresin normal and essentialfattyacid deficient
stratumcorneum: characterization by ruthenium tetroxide staining and x-ray dif-fraction.J. Invest. Dermatol.96:215-223.
31. McNutt, N. S., and W. L.Crain. 1981.Quantitative electronmicroscope
comparisonoflymphaticnuclear contoursin mycosis fungoides and inbenign infiltrates oftheskin.Cancer(Phila). 47:163-166.
32. Elias, H.,A.Henning, andD.Schwartz. 1971. Stereology: applications
inbiomedical research. Physiol. Rev. 51:158-200.
33.Menon, G.K., S. K. Aggarwal, andA. M. Lucas. 1981.Evidenceforthe holocrinenature oflipidsecretionin avian epidermalcells. Ahistochemicaland
fine structural study of the rictus andtheuropyogialgland. J.Morphol.
167:185-199.
34. Fowler, S. D.,andP. H.Greenspan. 1985. Application of nilered, a
fluorescent hydrophilicprobe forthedetection ofneutral lipiddeposits intissue sections. J. Histochem. Cytochenm 33:833-836.
35.Lampe, M.A.,M. L.Williams, andP. M.Elias.1983.Human epidermal lipids: characterization and modulations duringdifferentiation. J. Lipid Res. 24:131-140.
36. Elias,P.M., B. E.Brown,P.Fritsch,J.Goerke,G. M.Gray, andR. J.
White. 1979.Localization andcompositionoflipids inneonatalmousestratum
granulosumandstratum granulosumand stratumcorneum.J.Invest.Dermatol. 73:339-348.
37. Bligh,E. G., and W.J. Dyer.1959.Arapidmethod of total lipidextraction andpurification.Can. J.Biochem. Physiol. 37:911-917.
38. Ponec, M., A. Weerheim, J. Kempenaar, A. M.Mommaas, andD.H.
Nugteren. 1988.Lipidcompositionofculturedhumankeratinocytesin relation totheirdifferentiation.J.LipidRes.29:949-961.
39. Menon, G.K.,K. R.Feingold,and P. M. Elias. 1992. Lamellarbody secretoryresponse tobarrier disruption. J. Invest.Dermatol.98:279-289.
40. Wilhelm, K. P., and H.L. Maibach. 1993. Influence ofagingonthe barrier function of human skinevaluated byinvivo transepidermalwaterloss
measurements.InAging Skin:PropertiesandFunctionalChanges. J.L. Leveque
andP.G.Agache,editors. Marcel Dekker,Inc. New York 239-250.
41.Pinnagoda, J.,R. A.Tupker,T.Agner, and J.Serup. 1990.Guidelines for transepidermalwaterloss(TEWL)measurement. A reportfrom the
standardiza-tion groupof the European Society of ContactDermatitis.ContactDermatitis.
22:164-178.
42.Roskos, K.,and H. I. Maibach.1992.Percutaneousabsorptionand age. Implicationsfor therapy.Drugs &Aging.2:432-449.
43.Yamamura, T., and T. Tezuka.1989.Change insphingomyelinaseactivity in humanepidermisduring aging. J.Dermatol. Sci. 10:617-641.
44.Long, C.C.,and R. Marks. 1992. Stratum corneumchangesinpatients
with senile pruritus. J.Am. Acad. Dennatol. 27:560-564.
45. Grubauer, G.,P. M. Elias, and K. R. Feingold. 1989. Transepidermal waterloss; thesignalforrecoveryof barrierstructureand function. J.LipidRes.
30:323-333.
46.Elias,P.M., andK. R.Feingold.1992.Lipidsand theepidermalwater barrier:metabolism,regulationandpathophysiology.Semin. Dermatol.
11:176-182.
47.Holleran,W. M., M.Mao-Qiang,W.N.Gao,G. K.Menon,P. M.Elias, andK. R.Feingold. 1991.Sphingolipidsarerequired for mammalian epidermal barrier function. Inhibition ofsphingolipid synthesis delays barrier recovery after
acuteperturbation. J. Clin. Invest.88:1338-1345.
48.Berardesca,E.,N.Farinelli,G.Rabbiosi, and H.L. Maibach.1991. Skin
bioengineeringin the noninvasive assessment of cutaneousaging.Dernmtologica (Basel). 182:1-6.
49. Feingold, K.R., B. E. Brown, S. R. Lear, A. H. Moser, and P. M.
Elias.1983. Localization of denovosterologenesis in mammalian skin. J.Invest.
Dermatol.81:365-369.
50. Feingold, K.R.,M.Mao-Qiang, G.K.Menon, S. S. Cho,B. E.Brown,
and P. M. Elias. 1990. Cholesterolsynthesis isrequired forcutaneousbarrier
function in mice. J.Clin.Invest.86:1738-1745.
51.Jackson, S.M.,L. C.Wood, S.Lauer, J.M.Taylor,A. D.Cooper,P. M.
Elias,and K. R.Feingold.1992.Effect ofcutaneouspermeabilitybarrier
