comparison of an in vitro skin model to normal human skin for dermatological research

Comparison of an In Vitro Skin Model to Normal Human Skinfor Dermatological ResearchNANCYA. MONTEIRO-RIVIERE,1* ALFRED O. INMAN,1 THOMAS H. SNIDER,2JIMA. BLANK,2 AND DAVID W. HOBSON3

1Cutaneous Pharmacology and Toxicology Center, North Carolina State University, Raleigh, North Carolina 276062Battelle Memorial Research Institute, Columbus, Ohio 43201-26933Dermatological Products of Texas, San Antonio, Texas 78215

KEY WORDS in vitro skin model; human skin; viability; cell culture; electron microscopy;morphology; skin

ABSTRACT EpiDermy, an in vitro human skin equivalent (HSE), was compared to normalhuman breast skin (NHS) tomorphologically and biochemically assess its feasibility for dermatologi-cal research. Intralot and interlot variability was studied in day 0, 1, 2, and 3 in vitro cultures and inday 0, 3, 5, and 7 NHS. For NHS, light microscopy (LM) at day 0 showed stratified epidermis whichexhibited an increase in vacuoles and dark basal cells as storage increased to 3, 5, and 7 days.Transmission electron microscopy (TEM) revealed typical organelles in the epidermis and aconvoluted basement membrane at day 0. With increased storage, vacuoles and paranuclear cleftsbecame numerous, necrosis increased, tonofilaments became less organized, and overall cellularintegrity decreased. Biochemical data showed consistent MTT and glucose utilization (GU) throughday 5, while lactate production decreased to 75% by day 3. By LM, day 0 HSE consisted of a thick,compact, stratum corneum that sent projections between the stratum granulosum cells. By TEM,the configuration, organization, differentiation, distribution, and frequency of the organellesdiffered slightly from NHS. In addition, the basement membrane of the HSE was not completelydifferentiated, and the dermis was thin and acellular. Although day 1 and 2 cultures showed littlechange, day 3 exhibited an overall degeneration. Biochemical analysis showed GU and lactateproduction decreased through day 3. In conclusion, the EpiDermy HSE, although exhibiting slightdifferences, was morphologically and biochemically similar to normal human epidermis and may bea valuable model in assessing the toxicology, metabolism, or pharmacology of nonvesicatingcompounds.Microsc. Res. Tech., 37:172–179, 1997. r 1997 Wiley-Liss, Inc.

INTRODUCTIONThe increased awareness of animal welfare within

the general public and the scientific community hasstimulated recent advances toward the development ofcommercial and noncommercial in vitro human skinequivalent (HSE) models. The need for alternativemodels to assess dermal toxicity is great in biomedicalresearch to test or evaluate the toxicity of numerouscompounds. Although these in vitro models may serveas alternatives, they cannot completely satisfy thephysiological or anatomical requirements found in invivo models. In order to reduce or replace animals inscientific research, ‘‘animal alternatives’’ are being pur-sued. These HSE cell culture models have been used tostudy percutaneous absorption, metabolism (Slivka etal., 1993), and toxicology (Klausner et al., 1995). How-ever, the most important application of such models indermatological researchmay be to screen potential cutane-ous irritants for use in safety assessments. A number ofpharmaceutical and cosmetic companies are using in vitromodels to assure the safety of their products and to meetregulatory requirements to classify irritancy potential. Inaddition, these results are combined with other in vitrobiochemical tests to determine the safety of the product.Commercialmodels includeTestskinyLivingSkinEquiva-

lent (Organogenesis, Inc., Canton, MA) (no longeravailable in the US), Skin Models ZK 1300 (AdvancedTissue Sciences, La Jolla, CA), and EpiDermy (MatTekCorporation, Ashland, MA). Although Advanced TissueSciences (ATS) and MatTek (MT) are currently market-ing these models to assess cutaneous irritants andtoxicity, Organogenesis (OR) has limited this model totherapeutic applications. Similarly, ATS has expandedthe scope of its model to include clinical applications.The purpose of this study was to directly compare

EpiDermy to normal human skin (NHS), to determinethemagnitude of intralot and interlot variability, and toassess cell viability at different storage times.

MATERIALS AND METHODSThe MatTek EpiDermy model consists of normal

human-derived epidermal keratinocytes cultured toform a multilayered differentiated model of humanepidermis. Each culture has an 8.0 mm diameter which

Received 15April 1995; Accepted in revised form 15 July 1995.Contract grant sponsor: U.S. Army Medical and Development Command;

Contract grant number: DAMD 17-89-C-9050, Battelle subcontract 53155-G155524.*Correspondence to: NancyA. Monteiro-Riviere, Cutaneous Pharmacology and

Toxicology Center, North Carolina State University, 4700 Hillsborough Street,Raleigh, NC 27606.

MICROSCOPY RESEARCH AND TECHNIQUE 37:172–179 (1997)

r 1997 WILEY-LISS, INC.

is supported on a microporous membrane measuringapproximately 28 µm in thickness.The EpiDermy HSE, purchased fromMatTek Corpo-

ration, was shipped next day air in a 24-well cultureplate. Upon arrival, the wells were removed from theshipping agar, briefly rinsed, and then placed in themaintenance media supplied by MatTek. Based on therecommendation of themanufacturer, the culture shouldbe used immediately or stored at 4°C and used the nextday. In this study, cultures were kept in an incubator at37°C in a 5% CO2, moisture-saturated atmosphere(90% relative humidity) and extended for up to 3 dayswith daily medium changes to assess viability. Thenormal human breast skin (NHS) was obtained throughthe Cooperative Human Tissue Network (Ohio StateUniversity), dermatomed to approximately 1 mm,wrapped in sterile gauze, and placed into 50 ml centri-fuge tubes filled with maintenance media (AdvancedTissue Sciences). Since only limited amounts could beobtained at one time, it was necessary to store the NHSat 4°C. For the HSE and the NHS, three samples fromeach of four lots were used to assess the day 0 intralotand interlot variability. In addition, samples from oneof the day 0 lots were stored in culture or at 4°C (1, 2,and 3 days for EpiDermy and 3, 5, and 7 days for NHS),and viability was assessed on each day. Since NHSdisplays few signs of morphological degeneration underminimal culture conditions, the samples were held inmedia for up to 7 days. Twenty-one sample areas wereharvested from each tissue type for a total of 42samples.

Morphological StudiesAll tissue samples were trimmed and placed in vials

containing cold (4°C) half-strength Karnovsky’s fixative(Monteiro-Riviere and Manning, 1987). Following abrief fixation, the samples were fine trimmed to theappropriate processing size (approximately 1 mm2 forNHS and 2 mm 3 4 mm for the HSE), postfixed in 1%osmium tetroxide, dehydrated in a graded ethanolseries, cleared in acetone, infiltrated, and embedded inSpurr resin. Thick plastic sections (1 µm) were cutusing glass knives, stained with 1% toluidine blue, andscreened with an Olympus BH-2 photomicroscope(Olympus Optical, Ltd., Tokyo, Japan) to assess theoverall morphology of the sample and to determine therepresentative area for transmission electron micros-copy (TEM) evaluation. Three random fields per slidewere measured using an eyepiece reticle to determinethe thickness variations of NHS and the HSE. Thinsections (800 Å) were cut on a Reichert Ultracut Eultramicrotome (Leica, Nussloch, Germany) with adiamond knife, picked up on 75 3 200 mesh coppergrids, stained with uranyl acetate and Reynolds’ leadcitrate, and then viewed with a Philips EM 410LStransmission electron microscope (Philips ElectronicInstruments, Inc., Mahwah, NJ) operating at an accel-erating voltage of 80 KV. Each section was thoroughlyevaluated, and representative areaswere photographed.

Biochemical StudiesAnalysis was performed on the HSE and NHS.

D-glucose and lactate concentrations were determinedsimultaneously with a model 2700 SELECT Biochemis-try Analyzer (Yellow Springs Instrument Co., Yellow

Springs, OH). These assays were performed on bothwhole HSE samples and on 5 mm biopsies of NHS. Aninitial volume of 0.9 ml of MT maintenance media wasplaced into the EpiDermy specimen wells, while theNHS was incubated in 0.9 ml of ATS maintenancemedia. The media were replaced after 1 h to remove theaccumulation of lactate from overnight incubations.Thereafter, 55 µl aliquots were collected from eachspecimen well at hourly intervals for 4 h. The sampledmedia was not replaced, but the assay values werecorrected for this repeated loss of volume. Concentra-tion values were regressed against time from 1–4 hafter the media change to determine slope estimatesrepresenting metabolic rates.Representative whole samples of four HSE lots were

assayed upon receipt using the MTT (3-[4,5-dimethyl-thiazol-2-yl]-2,5-diphenyltetrazolium bromide) reduc-tion method to determine the overall viability of thecultures. The values were equivalent or greater thanthe preshipped MTT values supplied by the manufac-turer. Viability as a function of storage was evaluatedon one HSE lot. MTT reduction determinations werethen performed on whole HSE samples and 5 mmbiopsies of split thickness NHS. Specimens were incu-bated with MTT, rinsed, and solubilized in eitherisopropanol (for the HSE) or Unisolt (Isolab, Inc.,Akron, OH) (for the NHS). The amount of reduced MTTformed by viable tissue was determined spectrophoto-metrically.

RESULTSNormal Human Skin

Light Microscopy. The epidermis of the NHS wasmade up of the keratinized cells of the stratum cor-neum, the flattened cells of the stratum granulosum,the polyhedral cells of the stratum spinosum, and thecuboidal to columnar cells of the stratum basale. Theentire thickness of the viable epidermis (stratum granu-losum, stratum spinosum, and stratum basale) variedfrom about 45–65 µm, with a thickness up to 83 µm inthe rete pegs. The compact stratum corneum measuredfrom 6–11 µm in thickness. Langerhans cells and theirprocesses were frequently found within the epidermis.The day 0 samples contained numerous melanocytes intwo sample lots (Fig. 1). The severity of intercellularepidermal edema varied from slight to moderate orsevere, depending on the sample lot.Morphological differences were apparent in the skin

stored in the culture medium prior to fixation. In allsamples, the culture medium was retained above andbetween the cell layers of the stratum corneum. A fewvacuoles were present within the stratum spinosumcells of day 3 samples, with the number of vacuolesincreasing in the day 5 samples. At day 7, additionalvacuoles were present in the stratum spinosum andstratum basale cells. A few dark basal pyknotic cellswere present within the epidermis. Samples within thesame lot were similar. The underlying dermis in allsamples contained the normal cell types, includingmacrophages, fibroblasts, and mast cells.Electron Microscopy. The stratum corneum of day

0 NHS consisted of keratinized stratified squamouscells connected by desmosomes. Membrane-coatinggranules and desmosomes were fairly numerous withinthe upper stratum granulosum layer. The membrane-

173COMPARISON OF A SKIN MODEL TO HUMAN SKIN

coating granules in the upper stratum granulosumwere occasionally seen fusing to the lower stratumcorneum cell layer (Fig. 2). The distribution of tonofila-ments, keratohyalin granules, and lipids was normal.Occasional degenerating cells indicative of normal epi-dermal keratinization were seen in the stratum granu-losum layer.The polyhedral stratum spinosum cells and cuboidal

stratum basale cells were connected to adjoining cellsby desmosomes. Intercellular epidermal edema in thestratum spinosum and stratum basale layers rangedfrom slight to moderate or severe, depending on lot.Melanocytes were present within the stratum basalelayer, and stage IV melanosomes were scatteredthroughout the stratum spinosum and stratum basale

layers of two lots. Stellate tonofilament bundles werenumerous, and Langerhans cells and their processescontaining Langerhans cell granules were scatteredthroughout the stratum spinosum layers.The basement membrane was highly convoluted and

appeared normal. Hemidesmosomes were aligned alongthe plasma membrane of the basal cells. The electron-lucent lamina lucida separating the basal cell plasmamembrane from the underlying lamina densa con-tained thin anchoring filaments (Fig. 3). Anchoringfibrils as well as microfibril bundles extended from thelamina densa into the dermis (Fig. 3).Slight cellular degeneration was apparent at day 3.

The residual culture medium detected by light micros-copy appeared as an amorphous substance between thestratum corneum layers of all day 3, day 5, and day 7samples. Occasional necrotic cells and organelles werescattered throughout the stratum spinosum and stra-tum basale cell layers of day 3 samples. Slight nuclearenvelope separation and degenerative mitochondriawith ruptured cristae were sometimes present in theselayers. Langerhans cells usually contained the typicalrod-shaped Langerhans cell granules, an indentednucleus, swollen and ruptured mitochondria, and anelectron-lucent cytoplasm but lacked desmosomes andtonofilaments. Paranuclear clefts present in some stra-tum basale cells (Fig. 4) often contained membranouswhorls.At day 5, the stratum corneum appeared more frag-

mented and electron-lucent. Membrane-coating gran-ules had degenerated to small vacuoles in the stratumgranulosum. Cytoplasmic vacuoles were scatteredthroughout the stratum spinosum and stratum basalelayers. Paranuclear clefts were also present and morenumerous than at day 3. Also, necrotic cells withchromatin clumping were obvious. The stratum basalelayer became distorted, with a loss of cellular integrity.Focal areas of the convoluted basement membrane hadthickened hemidesmosomes that appeared to extendinto the lamina lucida.

Fig. 1. Light micrograph of normal human skin showing melano-cytes (arrows) in stratum basale layer. Note epidermis (E), dermis (D),and residual maintenance medium (m). NHS, day 0. 3300.

Fig. 2. Transmission electron micrograph showing membrane-coating granules fusing (arrows) to the lower stratum corneum (SC)layer. Note normal stratum granulosum (SG), membrane-coatinggranules (arrowheads), desmosomes (d), tonofilaments (T), and kerato-hyalin granules (k). NHS, day 0. 323,600.

Fig. 3. Transmission electron micrograph showing a normal base-ment membrane. Note hemidesmosomes (h), lamina lucida (L), laminadensa (D), anchoring filaments (arrowheads), anchoring fibrils (smallarrow), and microfibril bundles (large arrow). NHS, day 0. 339,400.

174 N.A. MONTEIRO-RIVIERE ET AL.

Overall cellular degeneration evidenced by a de-crease in cellular integrity was greater at day 7.Mitochondria were damaged and cristae completelyruptured, leaving only cytoplasmic vacuoles through-out the cells. An increase in vacuoles, paranuclearclefts, degenerative organelles, and necrotic cellsthroughout the epidermis was noted. Tonofilaments didnot appear as recognizable bundles but rather ashomogeneous hyalin masses within these cells (Fig. 5).In addition, vacuoles and ruptured mitochondria werefound within the dermal fibroblasts.Biochemical Studies. For NHS, viability assessed

by MTT dye reduction and glucose utilization (GU)remained consistent through 5 days of storage. Afterday 7, MTT reduction fell to 65% of day 0 assay, and GUdecreased to 77%. By day 3, lactate production de-creased to 75% and remained constant through day 7.

EpiDermy Human Skin EquivalentLight Microscopy. The EpiDermy HSE consisted

of a thick and compact stratum corneum, a prominentstratum granulosum, a stratum spinosum layer, astratum basale layer, and a thin acellular dermis (Fig.6). The stratum corneum thickness was 19 µm in day 0,25 µm in day 1, 27–56 µm in day 2, and 47–64 µm in day3 samples. The junction between the stratum corneumand the stratum granulosum was very irregular, withnumerous keratinized fingerlike projections of the stra-tum corneum layer extending down into the stratumgranulosum (Fig. 7). The lighter staining cells of thestratum granulosum varied in shape from circular toflat and contained prominent keratohyalin granules.Although the stratum granulosum layer was typicallyone to three cells thick, focal areas up to five cell layerswere sometimes noted. Cells of the stratum spinosumwere polyhedral and contained round nuclei. The stra-tum basale cell layer consisted of cuboidal to slightlyelliptical cells containing cytoplasmic tonofilamentsand round to oval nuclei with compact nucleoli.Although the thickness of the viable epidermis varied

between 80 and 130 µm in all samples, no rete pegswere present. Since the thickness of the stratum cor-neum increased from day 0 to day 3, the remaininglayers of the epidermis subsequently decreased inthickness. The overall stratification and morphology ofthe epidermis was comparable to human skin, with theexception of focal areas of swollen stratum granulosumcells and moderate intercellular epidermal edema andvacuoles within the stratum basale cells. The underly-ing dermis in day 0 to day 3 samples varied in thicknessfrom 3–8 µm.The epidermal morphology at day 1, day 2, and day 3

was similar to that of day 0. The notable exceptionswere that projections of stratum corneum into thestratum granulosumwere more pronounced and, in oneinstance, had abnormal epidermal stratification (Fig. 8).Electron Microscopy. The majority of the stratum

corneum of day 0 samples consisted of fairly compact

Fig. 4. Transmission electron micrograph depicting a paranuclearcleft (pc). NHS, day 3. 38,500.

Fig. 5. Transmission electron micrograph showing a necrotic cell(nc) and hyalinized tonofilaments (T) within the stratum spinosumlayer. NHS, day 7. 37,000.

Fig. 6. Light micrograph of a HSE showing a thick compactstratum corneum (SC), a light prominent stratum granulosum (SG),stratum spinosum (SS), stratum basale (SB), and acellular dermis (D).MT, day 0. 3420.


cell layers connected with occasional desmosomes. Iso-lated areas of the stratum corneumwere highly disorga-nized. Remnants of numerous degenerated cellularorganelles such as membrane-coating granules (appar-ently devoid of lipid) and filaments were present intra-cellularly and intercellularly throughout the organized(Fig. 9) and disorganized layers. In some cases, theupper layer of the stratum corneum contained electron-dense profiles and fibrous strands embedded within anamorphous ground substance. This may represent anarea of undifferentiated keratin containing the rem-nants of membrane-coating granules and mitochon-dria. The bottom layer of the stratum corneum varied agreat deal in thickness and often sent fingerlike projec-tions extending down between the cells of the stratumgranulosum (Fig. 10).

The large, flattened cells of the stratum granulosumcontained round keratohyalin granules which wereattached to adjacent cells by desmosomes (Fig. 10).Membrane-coating granules within these cells wereoften seen fusing with the lower layer of the stratumcorneum to release their lipid contents (Fig. 11). Thestratumgranulosum cells contained tonofilaments, dam-aged mitochondria, and vacuoles.The cytoplasm of cells within the stratum spinosum

was more compact than that of cells within the stratumgranulosum and contained numerous tonofilamentbundles, glycogen, and mitochondria. The occasionalelectron-dense compact masses seen within the cellswere probably the cross-section of tonofilament bundles.The stratum basale cells were cuboidal to slightly

elliptical. Although the basal cells appeared fairlynormal (Fig. 12), a few cells did contain small vacuoles,

Fig. 7. Light micrograph showing numerous fingerlike projections(arrows) of the stratum corneum (SC) extending between the cells ofthe stratum granulosum. MT, day 0. 3400.

Fig. 8. Light micrograph exhibiting abnormal stratification of theepidermis. Note stratum corneum (SC), stratum granulosum (SG),stratum spinosum (SS), stratum basale (SB), and dermis (arrows). MT,day 0. 3250.

Fig. 9. Transmission electron micrograph showing intracellular(arrows) organelle remnants. MT, day 0. 37,000.

Fig. 10. Transmission electronmicrograph of the stratum corneum(SC) projecting into the stratum granulosum (SG) layer. Note kerato-hyalin granules (k). MT, day 0. 34,500.


nuclear envelope separation, random rather than sub-basal mitochondria with ruptured cristae, large cyto-plasmic vacuoles, and thickened tonofilaments. Moder-ate intercellular epidermal edema between basal cellswas common, and desmosomes between adjacent basalcells appeared to have large dense plaques but notonofilaments.Hemidesmosomes were very numerous along the

plasma membrane of these stratum basale cells. Al-though a broken, slightly amorphous zone similar tothe lamina lucida and lamina densa was present be-neath the hemidesomosomes, no true basement mem-brane was actually present. This ‘‘pseudo basementmembrane’’ was relatively flat, not highly convolutedlike the basement membrane of normal human skin.Focal areas of the basement membrane beneath thehemidesmosomes did appear to have fine anchoring

filaments. However, anchoring fibrils and microfibrilbundles were not identified between the basementmembrane and the dermis (Fig. 13). The underlyingcollagen-rich dermis was acellular (Fig. 13). In onesample, the upper 20% of the dermis appeared devoid ofa collagen matrix. A layer of amorphous ground sub-stance containing vacuoles very similar to that nor-mally present within basal intercellular spaces wassituated between the pseudo basement membrane andthe dermal collagen matrix.The day 1 samples were similar to the day 0 samples

with the following exceptions. The bottom layer of thestratum corneum had more extensive stratum corneumprojections into the stratum granulosum. The promi-nent stratum granulosum contained round keratohya-lin granules and thick tonofilament bundles, while thestratum basale nuclei exhibited chromatin clumpingand differentiation. Fine fibrils in the dermis alignedparallel to the pseudo basement membrane.The stratum corneum of day 2 samples remained

compact and consisted of up to 19 cell layers. Within theviable epidermis, tonofilaments were thick, and desmo-some connections between adjacent cells had thick,dense attachment plaques. Microvesicles and an in-crease in collagen fibrils (Fig. 14) were seen in someareas of the basement membrane.Cellular integrity was less in the day 3 samples,

indicating an overall cellular degeneration of the cul-ture. The stratum corneum became less organized dueto fewer desmosomes and possessed fewer compact celllayers. The stratum granulosum contained remnants ofmitochondria and fewmembrane-coating granules.Also,nuclear margination and slight nuclear envelope sepa-ration were more obvious. The stratum basale cellswere occasionally vacuolated. The large paranuclearvacuoles may be caused by the coalescing of rupturedmitochondria. Tonofilaments were not as thick or dense,and fewer hemidesmosomes were located along thebasal cell membrane. In several areas, there was anincrease in the fine fibrils that were oriented parallel tothe basement membrane.

Fig. 11. Transmission electron micrograph showing membranecoating granules (arrowhead) and subsequent fusing (arrows) to astratum corneum (SC) projection. MT, day 0. 310,600.

Fig. 12. Transmission electron micrograph showing normal stra-tum basale cells containingmitochondia (M), tonofilaments (T), desmo-somes (d), vacuoles (V), and numerous hemidesmosomes (arrows)along the basement membrane. MT, day 0. 37,000.

Fig. 13. Transmission electron micrograph of the pseudo basementmembrane. Note hemidesmosomes (h), poorly developed lamina lucida(arrowheads) and lamina densa (arrows), tonofilaments (T), anddermal collagen (c). MT, day 0. 322,900.


Biochemical Studies. For the one HSE lot exam-ined, GU decreased gradually to 60% of day 0 values byday 3. Lactate production remained consistent throughday 1 and then decreased to 85% by day 2 and 65% byday 3.

DISCUSSIONThe objective of this study was to morphologically

and biochemically evaluate EpiDermy, a commerciallyavailable in vitro HSE, and NHS as potential modelsfor cutaneous toxicity and percutaneous absorptionstudies. In addition, this study assessed the interlotand intralot variability of NHS and the HSE. To bevalid, an in vitro skin model must correlate morphologi-cally and biochemically with human skin.Although the general morphology of the NHS and the

HSE at the light microscopic level appeared similar,specific anatomical differences were present (Table 1).In general, both consisted of a stratified squamousepithelium made up of a typical stratum corneum,stratum granulosum, stratum spinosum, and stratumbasale cell layers supported by a dermal matrix. Themorphological composition of NHS and EpiDermy issimilar to other commercial (Bilbo et al., 1993; Contardet al., 1993; Stoppie et al., 1993) and noncommercial(Cook et al., 1993; Mak et al., 1991; Michel and Auger,1993; Roguet et al., 1992) cultured skin models.The stratum corneum of the HSE was more compact

and smooth than the NHS. The stratum corneumthickness of all NHS samples was fairly consistent,while that of the HSE was thicker, more variable, andincreased with culture incubation time. Stoppie et al.(1993) noted a twofold increase in the thickness of thestratum corneum of a commercial HSE fromATS after 7days of culture. Also, the stratum corneum of OGTestskiny continued to increase in thickness through-out the culture period, eventually leading to an overallreduction by 31 days in thickness of the viable epider-mis (Bilbo et al., 1993). The stratum corneum layers ofthe HSE retained numerous cellular remnants andwere not always completely keratinized. The stratum

corneum of the HSE cultures sent numerous finger-likeprojections between the cells of the stratum granulo-sum.The cells within the viable epidermis of the HSE

differed somewhat in shape from those of NHS. TheHSE possessed cuboidal- to elliptical-shaped cells inthe stratum basale layer rather than the typical cuboi-dal to columnar. While Fleischmajer et al. (1993) foundthe stratum basale of theATS model to consist of two tothree layers of polygonal cells, other investigatorsfound cuboidal to ‘‘long-shaped’’ cells in ATS models ZK1300/2000 (Stoppie et al., 1993). The epidermis of allNHS and HSE samples during all culture storage timesdid fall within specific thickness ranges. The overallreduction of epidermal thickness coupled with an in-crease of stratum corneum thickness with time indi-cates that the cells of the HSE are in a different growthphase than the cells of NHS. The epidermal cells of thein vitro cultures contained such typical organelles asmembrane-coating granules, nuclei, nucleoli, mitochon-dria, endoplasmic reticulum, Golgi apparatus, lipiddroplets, and tonofilaments. Arrangement of mitochon-dria in the stratum basale of HSE samples appearedrandom rather than concentrated beneath the nuclei asin NHS. Desmosomes in the HSE samples were sparsecompared to NHS. While numerous keratohyalin gran-ules within the stratum granulosum of NHS werestellate, they were round in the HSE. Melanocytes andLangerhans cells were present in NHS and absent inthe HSE.The degree of intercellular edema in the stratum

spinosum and stratum basale of NHS was consistent

Fig. 14. Transmission electron micrograph of the pseudo basementmembrane. Note hemidesmosomes (h), poorly developed lamina lucida(arrowheads) and lamina densa (arrows), and dermal collagen (c). MT,day 2. 356,500.

TABLE 1. Comparison of normal human skin and MatTek HSE1

Morphology NHS MT

Stratum corneum Convoluted SmoothThickness D-0 6–11 µm 19 µm

D-1 N/A 25 µmD-2 N/A 27–56 µmD-3 6–11 µm2 47–64 µm

EpidermisRete pegs2 Present AbsentThickness D-0 45–65 µm3 80–130 µm

D-1 N/A 80–130 µmD-2 N/A 80–130 µmD-3 45–65 µm2,3 80–130 µm

Cell shape SG2 Flat Circular–flatSS2 Polyhedral PolyhedralSB2 Cuboidal-columnar Cuboidal-elliptical

Melanocytes2 Present AbsentLangerhans cells2 Present AbsentIntercellular edema SS2 None–severe None–moderate

SB2 None–severe None–moderateBasement membrane2 Convoluted SmoothHemidesmosomes Numerous More numerousLamina lucida Continuous PseudoLamina densa Continuous PseudoAnchoring filaments Present PresentAnchoring fibrils Present AbsentMicrofibrils Present Absent

Dermis2Thickness .2,000 µm4 3–8 µmFibroblasts Numerous5 AbsentCollagen Numerous fibers Numerous fibrilsAdnexial structures Present Absent

1D-0, day 0; D-1, day 1; D-2, day 2; D-3, day 3; SB, stratum basale; SG, stratumgranulosum; SS, stratum spinosum.2Identical with increased culture storage times.3Measurements excluded rete peg areas.4Biopsies included only partial dermis.5Additional cell types were observed.


within but not between lots. The HSE had occasionalintracellular edema within the stratum granulosumand moderate intercellular edema in the stratum ba-sale. Overall, edema in the HSE was fairly consistentfrom sample to sample.The basement membrane of the HSE, unlike that of

the NHS, was smooth due to the physical cultureconstraints and did not possess the epidermal rete pegspresent in NHS. An underdeveloped basement mem-brane zone was present in the HSE. It consisted of apoorly developed pseudo basement membrane thatlacked a true lamina lucida and lamina densa as well asthe accompanying anchoring filaments, anchoringfibrils, and microfibril bundles. Cook et al. (1993), usinga collagen gel substrate supplied byMT, found a similarstructure resembling the basement membrane in cul-ture preparations whose basal layers were planar. OGcultures following 11 days at the air-liquid interfaceshowed morphological evidence of the beginnings ofhemidesmosome formation while lacking the cytoplas-mic plaque and associated lamina densa present in thebasement membrane of NHS (Bilbo et al., 1993). Incontrast, 5 weekATS cultures showed a distinct laminadensa in most areas with an associated lamina lucida,anchoring filaments, and hemidesmosomes (Contard etal., 1993). In day 2 samples, occasional microvesicles werepresent along theHSEpseudo basementmembrane.The dermis of the cultures consisted of a collagen-rich

matrix, denoted by numerous collagen fibers and/orfibrils. The dermis of NHS contained numerous celltypes (fibroblasts, macrophages, and mast cells) andsite-specific adnexial appendages (sebaceous glands,sweat glands, and hair follicles). The dermis of the HSEcultures was completely acellular and lacked adnexialappendages and nonkeratinocytes (melanocytes,Merkelcells, and Langerhans cells).Cellular degeneration increased in the NHS from day

3 to day 7 and in the HSE from day 1 through day 3.This was generally characterized by an increase in thenumber of vacuoles, nuclear envelope separation, swol-len and ruptured mitochondria, dark basal cells, andthe later appearance of additional vacuoles, dyskera-totic cells, and necrotic cells.The MTT assay and GU showed the biochemical viabil-

ity of NHS remained consistent from day 0 through day 5,while HSE showed a decrease in GU by day 3. Lactateproduction decreased in both NHS and HSE by day 3.The HSE is not a monolayer culture but rather a

multilayered, highly differentiated model that re-sembles NHS at the light level. Subtle differencesbetween the in vitro model and NHS are more apparentby TEM. The validity of such in vitro models has beenquestioned in percutaneous absorption and metabolismstudies. Since the stratum corneum and its constituentlipids are the primary barrier to absorption, the properstructure and function must be expressed in an in vitromodel. Therefore, considerable debate exists on how tocompare these in vitro cultures to normal human skin.Although the model has limitations, it has advantagesover NHS such as commercial availability, little in-tralot and interlot variability, and a lack of variabilitydue to site specificity. Extensive validation of this modelis necessary since no single in vitro model can predictwhat actually occurs in vivo.In conclusion, the morphology and biochemistry of

the day 0 cultures was good and correlates well between

EpiDermy andNHS. Since theHSE possesses a pseudobasement membrane that consists of an underdevel-oped lamina lucida and lamina densa, it may notadequately respond to cutaneous vesicants. Althoughthis may eliminate the use of the HSE for morphologi-cal studies with vesicants, it is still a powerful biochemi-cal tool. In addition, EpiDermy may be appropriate forthe toxicology, pharmacology, and metabolic studiesusing nonvesicating compounds if the rate of penetra-tion is determined to be comparable to NHS. Mappingof basement membrane epitopes is needed to determinethe antigenic similarities between the HSE and NHS.Normal human skin and HSEs stored beyond day 0exhibited an increase in epidermal necrosis, vacuoliza-tion, ruptured mitochondria, and dark basal cells,which correlates to the reduction in biochemical viabil-ity. As indicated by the morphological and biochemicaldata, interpreting dermatotoxicity from cellular necro-sis due to extended culture times may be difficult. Tominimize any possible interlot and/or intralot variabil-ity found within the NHS and the HSE, it is essentialthat all samples exposed to a cutaneous toxicant beaccompanied by a corresponding control.

ACKNOWLEDGMENTSThis work was supported by the U.S. Army Medical

and Development Command under contract DAMD17-89-C-9050, Battelle subcontract 53155-G155524. Theviews, opinions, and/or findings contained in thismanu-script are those of the authors and should not beconstrued as an official Department of the Army posi-tion, policy, or decision unless so designated by otherdocumentation.

REFERENCESBilbo, P.R., Nolte, C.J.M., Oleson, M.A., Mason, V.S., and Parenteau,N.L. (1993) Skin in complex culture: The transition from ‘‘culture’’phenotype to organotypic phenotype. J. Toxicol. Cut. Ocular Toxicol.,12:183–196.

Contard, P., Bartel, R.L., Jacobs L., II, Perlish, J.S., MacDonald, E.D.,II, Handler, L., Cone, D., and Fleischmajer, R. (1993) Culturingkeratinocytes and fibroblasts in a three-dimensional mesh results inepidermal differentiation and formation of a basal lamina-anchor-ing zone. J. Invest. Dermatol., 100:35–39.

Cook, J.R., Patrone, L.M., Rhoads, L.S., and Van Buskirk, R.G. (1993)Ahuman epidermal model grown on an acellular gel. J. Toxicol. Cut.Ocular Toxicol., 12:109–128.

Fleischmajer, R., MacDonald, E.D., II, Contard, P., and Persish, J.S.(1993) Immunochemistry of a keratinocyte-fibroblast co-culturemodel for reconstruction of human skin. J. Histochem. Cytochem.,41:1359–1366.

Klausner, M., Neal, P.J., and Kubilus, J. (1995) UVB irradiation of anorganotypic skin model, Epidermy, results in significant release ofcytokines. Toxicologist, 15:325.

Mak, V.H.W., Cumpstone, M.B., Kennedy, A.H., Harmon, C.S., Guy,R.H., and Potts, R.O. (1991) Barrier function of human keratinocytecultures grown at the air-liquid interface. J. Invest. Dermatol.,96:323–327.

Michel, M., and Auger, F.A. (1993) Anchored skin equivalent culturedin vitro: A new tool for percutaneous absorption studies. In VitroCell. Dev. Biol., 29A:834–837.

Monteiro-Riviere, N.A., and Manning, T.O. (1987) The effects ofdifferent fixatives on the porcine integument. In: 45th AnnualProceedings, Electron Microscopy Society of America. G.W. Bailey,ed. San Francisco Press, San Francisco, pp. 948–949.

Roguet, R., Dossou, K.G., and Rougier, A. (1992) Use of in vitro skinrecombinants to evaluate cutaneous toxicity: A preliminary study. J.Toxicol. Cut. Ocular Toxicol., 11:305–315.

Slivka, S.R., Landeen, L.K., Zeigler, F., Zimber, M.P., and Bartel, R.L.(1993) Characterization, barrier function, and drug metabolism ofan in vitro skin model. J. Invest. Dermatol., 100:40–46.

Stoppie, P., Borghgraef, P., De Wever, B., Geysen, J., and Borgers, M.(1993) The epidermal architecture of an in vitro reconstructedhuman skin equivalent. Eur. J. Morphol., 31:26–29.


comparison of an in vitro skin model to normal human skin for dermatological research

Documents