ARTICLE IN PRESS

Biomaterials 27 (2006) 785795 www.elsevier.com/locate/biomaterials

Mechanical properties of human stratum corneum: Effects of temperature, hydration, and chemical treatment
Kenneth S. Wua, William W. van Osdolb, Reinhold H. Dauskardtc,
Department of Mechanical Engineering, Stanford University, Stanford, CA 94305, USA b ALZA Corporation, Mountain View, CA 94039, USA c Department of Materials Science and Engineering, Stanford University, Stanford, 416 Escondido Mall, Bldg 550, Rm 550G, CA 94305-2205, USA Received 4 March 2005; accepted 24 June 2005 Available online 10 August 2005
a

Abstract An in vitro mechanics approach to quantify the intercellular delamination energy and mechanical behavior of isolated human stratum corneum (SC) in a direction perpendicular to the skin surface is presented. The effects of temperature, hydration, and a chloroformmethanol treatment to remove intercellular lipids were explored. The delamination energy for debonding of cells within the SC layer was found to be sensitive to the moisture content of the tissue and to the test temperature. Delamination energies for untreated stratum corneum were measured in the range of 18 J/m2 depending on test temperature. Fully hydrated specimen energies decreased with increasing temperature, while room-humidity-hydrated specimens exhibited more constant values of 24 J/ m2. Lipid-extracted specimens exhibited higher delamination energies of 12 J/m2, with values decreasing to 4 J/m2 with increasing test temperature. The peak separation stress decreased with increasing temperature and hydration, but lipid-extracted specimens exhibited higher peak stresses than untreated controls. The delaminated surfaces revealed an intercellular failure path with no evidence of tearing or fracture of cells. The highly anisotropic mechanical behavior of the SC is discussed in relation to the underlying SC structure. r 2005 Elsevier Ltd. All rights reserved.
Keywords: Mechanical properties; Fracture toughness; Epithelial cell; Stratum corneum; Tissue treatment

1. Introduction The layered construction of skin has components that possess mechanical properties needed to accommodate intrinsic and imposed mechanical stresses, abrasions and penetration of foreign objects under variable ambient moisture and temperature conditions. It must also have other physical properties that resist the presence of toxic environmental chemicals, pathogens, and radiation [1].
Abbreviations: SC, stratum corneum; RH, relative humidity; FH, fully hydrated; RHH, room-humidity hydrated; SS, stress separation; DCB, double-cantilever beam; CMT, chloroformmethanol treated; SEM, standard error of the mean Corresponding author. Tel.: +1 650 725 0679; fax: +1 650 725 4034. E-mail address: dauskardt@stanford.edu (R.H. Dauskardt). 0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.06.019

More severe environmental conditions may result from occlusion of the skin via application of adhesive dressings or transdermal drug delivery patches which can locally elevate moisture content and affect mechanical behavior. The detailed cellular and intercellular structures of the outermost layer of the epidermis, the stratum corneum (SC), have been widely studied [25]. However, relatively few studies have examined the mechanical and fracture properties of SC to determine their dependence on tissue treatment and environmental conditions [611]. In particular, surprisingly few studies have reported on the mechanical and delamination properties of the SC in the direction normal to the skin surface [1215]. The SC consists of layered anucleated cells that mature and subsequently detach in the natural renewing

ARTICLE IN PRESS
786 K.S. Wu et al. / Biomaterials 27 (2006) 785795

process of desquamation. The disk-shaped SC cells, or corneocytes, composed largely of aligned keratin laments, have been likened to bricks bound together by a lipid-rich mortar [16]. Intercellular lipids have been identied as the primary pathway for chemical diffusion and as the barrier to water permeability through the SC layer [1720]. In addition to the intercellular lipids, degraded desmosomal protein junctions, or corneosomes, at the cell boundaries are central to SC cohesion and renewal, and their degradation is necessary for desquamation [2126]. Ultrastructural studies of the SC have suggested that increased moisture content is associated with elevated desmosome degradation and disruption of intercellular lipid structures, which affect both mechanical properties and permeability [25,27]. Changes in delamination properties of SC with environmental conditioning have been simply demonstrated by a cellophane tape-stripping technique. SC hydrated for 24 h prior to tape stripping permits cells to be removed more easily, requiring fewer applications of tape to fully strip the SC from the epidermis [12]. Another approach involved use of a cohesograph to measure SC cohesive strength [13]. Similar devices have been used to examine the bond strength of materials applied to the skin, such as adhesive dressings [14]. While these techniques provide relative measures of SC delamination strength, several inherent problems exist that make the results qualitative and unreliable. Most signicantly, the delamination of SC adhered to underlying tissue leads to a combined measurement of both SC and substrate properties. Additionally, the SC loading is nonuniform and can result in highly variable results. In the case of the cohesograph, the force necessary to pull the SC apart may be affected substantially by the deformation behavior of the underlying tissue substrate. The tape-stripping technique effectively peels the cell layers apart but, similar to the peel adhesion test for thin lms on elastic substrates, provides qualitative results that are difcult to quantify [28]. There is clearly a need for quantitative test techniques to measure cohesion and strength properties of SC accurately and reproducibly. In the present study, a quantitative in vitro experimental mechanics approach to examine the SC intercellular delamination energy and out-of-plane mechanical behavior is presented. The delamination energy of human SC was examined as a function of selected testing temperatures and moisture preconditioning treatments and related to the underlying cellular structure. For comparison, human SC delipidized with a chloroformmethanol treatment was examined in the same manner. While some of the treatments and temperatures are nonphysiological, they provide insight into the microstructural mechanisms of SC cohesion and help to isolate the role of individual components of the SC such as intercellular lipids. SC delamination energy

was quantied in terms of the energy required to propagate a debond through the SC layer and dened in terms of the strain energy release rate G, measured in units of J/m2. Stress-separation (SS) tests were performed to measure the out-of-plane mechanical behavior of the SC. Resulting failure surface morphologies were examined with scanning electron microscopy to provide an indication of the delamination mechanisms. The delamination energy and strength property dependence on temperature and moisture are rationalized in terms of the underlying SC cell structure and intercellular lipids.

2. Materials and methods 2.1. Tissue preparation Human cadaver SC tissue was isolated for these experiments from three female Caucasian donors, 7688 years of age, from the thigh, abdomen or lower back. For each study as detailed in the following sections, comparative tests were performed on the same donor tissue to avoid variability between donors. The SC tissue was separated from the underlying epidermis via a trypsin enzymatic digest, then stored at 4 1C in a fully hydrated (FH) state on water-moistened lter paper (Grade 595 General-Purpose Filter Paper, Schleicher & Schuell MicroScience GmbH, Dassel, Germany). For the untreated SC, two sets of tissue were prepared. One set consisted of SC that was FH at 100% relative humidity (RH) by storing on watermoistened lter paper, and the other was room-humidity hydrated (RHH) at 45% RH. Both sets were allowed to equilibrate for at least 24 h. These initial hydration and testing conditions represent signicantly different equilibrium SC water contents corresponding to 300400% wt/wt and 5% wt/wt SC water content, respectively [29,30]. The thickness of untreated RHH SC was measured with a micrometer to be between 1535 mm. The FH SC thickness was observed to increase by 30% similar to values reported by others [31]. Additional tissue, 60 60 mm2, was delipidized with a 120min 30 mL chloroform:methanol (2:1 by volume) soak with two subsequent 30 min 30 mL water rinses. The SC thickness was not observed to change signicantly with treatment. Chemically treated SC for SS tests was treated similarly, but with a shortened 60-min chloroformmethanol soak. The different treatment time is not believed to inuence the lipid extraction signicantly. Studies with a similar chloroform methanol treatment have shown that the majority of lipids are removed within the rst 15 min [32]. The treated SC was either fabricated into FH test specimens immediately after treatment or allowed to dehydrate in an ambient 45% RH environment similar to the untreated SC. 2.2. Delamination energy measurements The delamination energy of the SC tissue was examined using fracture mechanics techniques developed to measure the adhesive properties of highly viscoelastic pressure-sensitive adhesives [33]. Similar techniques have been employed to

ARTICLE IN PRESS
K.S. Wu et al. / Biomaterials 27 (2006) 785795 787

measure the adhesive properties of polymer bone cements and polymer layers in elastic substrates [3436]. The technique involves sandwiching the SC between two elastic polycarbonate substrates (Hyzods GP, Shefeld Plastics Inc., Shefeld, MA) with a cyanoacrylate adhesive (Instant Krazyglue Gel, Elmers Products Inc., Columbus, OH) to form fracturemechanics-based double-cantilever beam (DCB) specimens (Fig. 1(a)). Cyanoacrylate adhesive polymerization is readily initiated by the presence of small amounts of water on the bonding surfaces limiting the adhesive to the SC exterior. The transparent polycarbonate beams facilitate optical inspection of the inner sandwich structure during specimen preparation and testing. To enable the use of linear elastic fracture mechanics to determine the strain energy release rates, substrate dimensions were chosen to ensure purely elastic deformation of the substrates during testing [33,34]. To fabricate the specimens, a thin layer of cyanoacrylate adhesive was applied to one face of a nominally 40 10 2.88 mm3 polycarbonate substrate, leaving a 710mm region of the beam end uncoated. The substrate was pressed against the SC on the lter paper backing and a scalpel was used to cut around the substrate to detach the adhered SC from surrounding tissue. In the case of RHH SC, the freestanding lm was prepared similarly with a sheet of paper as a backing. To form the nal sandwich structure, another substrate coated with adhesive in the same manner was pressed against the SC face of the complimentary beam with adhesivefree ends aligned. Excess adhesive along the sandwich edges was removed with a scalpel to ensure that the two halves of the sandwich structure were bound together by SC only. DCB specimens containing either untreated or chloroform methanol-treated (CMT) SC were placed in an environmental chamber (Model LH-6, Associated Environmental Systems, Ayer, MA) at selected temperature (10, 25, 75 1C) and RH

(45, 85% RH) conditions and allowed to equilibrate for 10 min. Tests involving FH specimens were conducted in an 85% RH environment with RHH specimens examined at 45% RH unless specied otherwise. The specimens were loaded via attached loading tabs to propagate a debond within the SC layers. The specimens were tested in a custom-built mechanical test system with a computer-controlled DC servoelectric actuator operated in displacement control. Tests were performed at a constant displacement rate of 2 mm/s. Corresponding loads were measured using a 222 N load cell. The delamination length, a, was measured from recorded load displacement, PD, and their elastic compliance relationship: C D 2 a 0:64h3 , P 3 E0I (1)

where I bh3 =12, and C is the specimen compliance, P is the load, D=2 is the corresponding displacement of each beam from its original position at the loading point, E 0 E =1 n2 is the plane strain Youngs modulus for the polycarbonate, n is Poissons ratio, I is the area moment of inertia, b is the polycarbonate substrate width, and h is the height of each polycarbonate beam. By measuring the critical load, P, and the delamination length, a, at incipient crack extension, the delamination resistance, Gc, was determined from critical values of the strain energy release rate, G [33,34,37]: p   ! 5h 1 h 2 12P2 a2 . (2) G 2 3 0 1 2 a 2 a b h E Multiple delamination energies, Gc, were measured for each DCB specimen by recording the critical loads, Pc, and associated delamination lengths, a, during delamination extension (Fig. 2). For the present specimens, the values of E and n for the polycarbonate were 2.379 GPa and 0.38, respectively. Given the thin lm nature of the SC compared to the massive polycarbonate substrate, the contribution of the

Delamination Stratum h Corneum Polycarbonate h

3.0 Delamination Energy, Gc (J/m2)

Polycarbonate b

(a)

P P Stratum h Corneum Polycarbonate b b P h

2.0

1.0

(b)

0.0 0 10 Crack Extension, a (mm) 20

Fig. 1. Fracture mechanics specimen geometries. (a) DCB geometry illustrating relevant loading parameters (P, D), delamination extension as measured from loading axis points (a) and relevant specimen dimensions (b 10 mm, h 2:88 mm, length 40 mm). (b) SS specimen conguration showing loading axis and relevant dimensions (b 10 mm, h 2:88 mm).

Fig. 2. Typical delamination energy as a function of delamination extension. Illustration of the variation in delamination energy (Gc) as a function of crack length (Da) for a RHH (45% RH) SC specimen tested in a DCB conguration at 10 1C, 45% RH.

ARTICLE IN PRESS
788 K.S. Wu et al. / Biomaterials 27 (2006) 785795

elastic strain energy in the SC layer can be ignored in the analysis [28].
Delamination Energy, Gc (J/m2)

6.0 * 5.0 + 4.0 3.0 2.0 1.0 0.0

(a)

45% R.H. 100% R.H.

2.3. Stress-separation measurements Structures containing SC sandwiched between polycarbonate substrates were fabricated in a similar manner to the DCB specimens (Fig. 1(b)). The nominal substrate dimensions were 10 10 2.88 mm3. The specimens were loaded normal to the SC face via attached loading tabs in an environmental chamber with controlled temperature and humidity. SS measurements were performed under the same conditions and SC hydrations as those for the DCB experiments. Testing was conducted at a constant displacement rate of 1 mm/s, yielding a machine compliance-corrected initial strain rate of 0.007 s1 in the SC for an SC thickness of 10 mm. The specimens were allowed to equilibrate for 20 min prior to testing. 2.4. Scanning electron microscopy Both DCB and SS specimens were examined after mechanical testing using a scanning electron microscope (Hitachi S-2500, Hitachi, Tokyo, Japan) to characterize the fracture surface morphologies. Selected specimens were allowed to dry in ambient conditions (25 1C, 45% RH), gold or goldpalladium coated, then examined in the electron microscope operated at 15 kV. Multiple specimens from each testing condition were inspected to ensure representative characterization. 2.5. Statistical analysis Delamination energies are presented as mean 71.96 standard error of the mean (SEM) in which the mean values reported are expected to fall within these bounds with 95% condence. On average, n 38 for each test condition. Delamination energies were compared using the Wilcoxon test for independent samples. Statistical signicance was set at 1% or 5% as specied in the gures. SS measurements are presented as mean7standard deviation (SD). Further statistical analysis was not performed due to the small sample sizes for each test condition (n 324).

*+ +

10

20

30 40 50 60 Temperature (C)

70

80

1.0 45% R.H. 0.8 Peak Stress, (MPa) 100% R.H.

0.6

0.4

0.2

0.0
(b)

10

20

30 40 50 Temperature (C)

60

70

80

Fig. 3. The variation in delamination energy and peak stress with temperature and hydration for untreated SC. (a) Delamination energy values from DCB experiments performed at selected temperatures and hydrations (Po0:01 except *Po0:05, +P40:05). (b) Peak stresses from SS experiments similarly performed at selected temperatures and hydrations. Error bars: (A) mean71.96 SEM and (B) mean7SD.

3. Results 3.1. Quantifying delamination energies The delamination energy, Gc, measured as a function of the delamination length, Da, for a specimen containing RHH SC is shown in Fig. 2. The data represent a socalled delamination resistance curve, and were obtained from testing of a single DCB specimen. The variability in the measured Gc values as seen in the curve was predominantly related to inhomogeneities in the SC along the DCB length. Typically, four cantilever specimens were tested for each condition, each yielding approximately 712 data points. All data points

(navg 38) for a given test condition were subsequently averaged to produce the delamination resistance values, Gc, represented in Figs. 3 and 4. Note that the delamination tip strain rates do not vary signicantly as a function of Da during such tests, as determined by modeling and experimental examination of similar tests on pressure-sensitive adhesives [38]. This reduces the effects of varying strain rate on the measured delamination energies. 3.2. Delamination energy variation with temperature and hydration Delamination energy values, Gc, for specimens containing RHH and FH SC measured at selected temperatures are presented in Figs. 3(a) and 4(a). Each

ARTICLE IN PRESS
K.S. Wu et al. / Biomaterials 27 (2006) 785795
16.0 45% R.H. + Delamination Energy, Gc (J/m2) 12.0 *+ * 8.0 100% R.H. CMT 45% R.H. CMT 100% R.H.

789

remain unclear. One possible explanation is that SC specimens from different donors possess variable properties that elicit different delamination energy measurements. 3.3. Comparison to stress-separation results The results of SS tests performed on RHH and FH specimens from the same SC as that represented in Fig. 3(a) and under the same environmental testing conditions are presented in Fig. 3(b). Typically, 34 measurements were taken at each testing condition, with each measurement requiring one specimen. SS tests resulted in specimen loading up to a peak stress, after which the stress decayed rapidly with continued displacement. Peak stress values obtained from the SS curves are shown in Fig. 3(b). The peak stresses were observed to decrease with increasing hydration and testing temperature, exhibiting similar behavior to the delamination energy trends seen for the DCB specimens. 3.4. Comparison to delipidized stratum corneum Delamination in the CMT SC specimens exhibited markedly different behavior compared to the untreated specimens. The corresponding data are shown in Fig. 4(a). Most notably, at each test temperature the measured delamination energies were signicantly higher than those of their untreated counterparts and proportionally smaller differences were observed between the delamination energies of the RHH and FH specimens. In contrast to the untreated specimens, statistically signicant decreases in Gc values were observed for the highest test temperature compared to the lower test temperatures regardless of initial hydration. A comparison between untreated and CMT SS specimens reveals similar trends to those observed during delamination testing. Delipidized SS specimens exhibited generally higher peak stresses compared to their untreated counterparts as shown in Fig. 4(b). The signicant scatter in the untreated control specimens in Fig. 4(b) was not characteristic of most of the SS tests performed. These specimens possessed the same loading characteristics with loading to a peak stress, followed by signicant load decreases with increasing displacement. Unlike the control group, the treated specimen peak stresses exhibited the same trend with testing temperature regardless of hydration. 3.5. Examination of delamination surface morphologies Representative scanning electron micrographs of the fracture surfaces of DCB specimens containing FH and RHH hydrated SC tested at 25 1C are shown in Fig. 5. Similar micrographs were obtained for such specimens

4.0 ++ 0.0 ++

** **

10

20

(a)
2.0

30 40 50 60 Temperature (C)

70

80

45% R.H. 100% R.H. 1.5 Peak Stress, (MPa) CMT 45% R.H. CMT 100% R.H.

1.0

0.5

0.0

10

20

(b)

30 40 50 60 Temperature (C)

70

80

Fig. 4. Comparison of untreated and chloroformmethanol-extracted SC. (a) Delamination energy measured in DCB tests performed on untreated and CMT SC at selected temperature (10, 25, 75 1C) and hydration (45, 100% RH) (Po0:01 except *Po0:05, +,++,z,** P40:05). (b) Peak stress data from SS measurements conducted at the same temperatures and hydrations for both untreated and CMT SC. Data offset for clarity. Error bars: (a) mean71.96SEM and (b) mean7SD.

plot represents data from a single SC donor to mitigate tissue specimen variations. SS experiments were performed on the same SC as that in Fig. 3(a). The results are shown in Fig. 3(b). Further DCB tests were performed on the same SC as that in Fig. 4(a) after chloroformmethanol treatment. These results are presented in Fig. 4(a). With increasing test temperature, the untreated FH specimens exhibited markedly lower delamination energies. In contrast, the RHH specimen energies were not statistically different with increasing test temperature, except between 10 and 25 1C for results shown in Fig. 4(a). Note that the Gc values for the untreated 45% RH specimens in Fig. 3(a) were signicantly higher than the untreated 45% RH specimens of Fig. 4(a). The reasons for these differences

ARTICLE IN PRESS
790 K.S. Wu et al. / Biomaterials 27 (2006) 785795

Fig. 5. DCB specimen delamination surface morphology. Scanning electron micrographs of DCB specimens containing (a) FH (100% RH), and (b) RHH (45% RH) SC. Specimens tested at 25 1C corresponding to delamination energies reported in Fig. 4(a). Images representative of DCB specimen delamination surfaces regardless of treatment.

at each temperature. For the FH specimens, individual SC cells are difcult to discern and surface features were characterized by tortuous surface undulations (Fig. 5(a)). The RHH specimen micrograph reveals a different surface morphology compared to FH fracture surfaces, showing partial pull-up of individual SC cells. This produced fracture surface roughness characterized by individual cellular features (Fig. 5(b)). The micrographs of delipidized specimens revealed the same difference in fracture surface appearance between RHH and FH specimens. Fracture surfaces of SS specimens yielded morphologies similar to those seen in the DCB specimens in which the RHH specimens exhibited SC cell pull-up, while the FH specimens lacked similar cellular features.

4. Discussion 4.1. Origin of delamination energies As far as fracture properties of materials are concerned, the SC delamination energy values are comparatively low. Polymer layers bonded weakly to elastic substrates may exhibit similar values, but with more strongly covalently bonded interfaces, fracture energies increase signicantly above 10 J/m2 [35,36,39]. In most of these cases, plastic deformation at the crack tip contributes signicantly to the delamination fracture energy. During delamination of the SC layer, energy is dissipated by separation of intercellular boundaries, plastic deformation of the SC layer, and by the work done during cell pull-up. To determine the extent of possible plastic deformation near the crack tip, plastic zone radius estimates, rp, were obtained using the well-known plane strain plastic zone expression: rp  1 Gc E 0 , 6p s2 ys (3)

where the plane strain modulus, E0 , and the yield strength, sys, are for the SC layer. The values of E0 were obtained from the SS initial loading slopes and found to be in the range 125 MPa. We note that there is considerable experimental uncertainty in the values of E0 estimated. These are due to inherent difculties in determining the through thickness strain in the thin SC layer, and values should be treated accordingly. The sys values were taken as the peak stresses in the SS tests as no obvious yielding was apparent prior to the peak. The resulting plastic zone size estimates were 212 mm, corresponding to an average of 25% of the original SC thickness, suggesting that only a fraction of the SC thickness may have undergone plastic deformation. For all of the tissue conditions examined, the failure path occurred between corneocytes. Given the average corneocyte thickness of 0.51 mm, multiple layers of cells on either side of the delamination may have undergone plastic deformation. However, energy dissipation by plastic deformation in a plastic zone represents only one possible energy dissipation mechanism. Alternatively, the cohesive properties of the cell boundaries themselves may dominate the energy dissipation with minimal general plastic deformation. This has been substantiated by studies of graded delamination properties in which delamination energies increased towards the lower SC as upper layers were removed [40]. Plasticity-dominated dissipation would suggest the opposite trend of decreasing delamination energy with decreasing SC thickness. Examination of the delamination surfaces of the DCB and SS specimens supports the idea that changes in the intercellular cohesive properties play a central role in determining delamination energy. The scanning electron micrographs reveal signicant differences between the RHH and FH specimens (Fig. 5), with both untreated and delipidized SC showing similar topographic characteristics. Notably, cell pull-up was present in the RHH specimens (Fig. 5(b)), and similar results were observed for the CMT specimens. The FH specimen delamination

ARTICLE IN PRESS
K.S. Wu et al. / Biomaterials 27 (2006) 785795 791

surfaces presented no obvious cell pull-up and individual corneocytes were not easily discernible (Fig. 5(a)). Similar features were observed for CMT SC. Consistent with the small variations in Gc values with temperature for RHH-untreated specimens (Figs. 3(a) and 4(a)), corresponding fracture surface morphologies exhibited no discernible differences in cell pull-up density with temperature. However, no obvious morphological changes were observed for the FH-untreated specimens despite large variations in Gc values with temperature (Figs. 3(a) and 4(a)). These observations suggest that the delamination energy must be more intimately tied to the cohesive properties of the intercellular boundaries of the SC and not to the failure morphology. This notion is supported by the similar trends exhibited by the peak stress data. Similar to the untreated specimens, the delipidized SC fracture surface morphologies remained unchanged as a function of temperature for each hydration condition despite signicant Gc variations with temperature. 4.2. Environmental effects on SC structure The marginal changes in delamination energy and peak cohesive strength with increasing test temperature exhibited by the untreated RHH SC compared to that of the FH specimens indicate that SC cohesive strength depends strongly on initial hydration. The commonly observed failure of SC between corneocytes highlights the relevance of the intercellular space to SC fracture properties [22]. In a simplistic model, SC cells are viewed as highly keratinized bricks surrounded by a lipid mortar (Fig. 6) [16]. This notion led to the proposal that SC cohesive behavior is largely inuenced by the intercellular lipid characteristics [16,41,42,]. Additional studies probing the effect of hydration and chemical treatment provide further insights. The microstructure of FH human SC has been examined using

Fig. 6. Illustration of SC delamination front and microstructural components. Schematic of the SC during delamination, illustrating intercellular failure and showing cellular structure, including aligned keratin laments as well as lipid intercellular space with corneosomes.

freeze-fracture electron microscopy to reveal the swelling of corneocytes, as well as the presence of water pools in the intercellular spaces [30,43]. Other studies have indicated that water disrupts the lipid lamellae to varying degrees and that intercellular corneosomes become degraded with time in the presence of water [25,27,30]. Hydration also has been shown to inuence lamellar lipid spacing in hairless mouse SC [44]. Examination of the affect of hydration on the lipid orthorhombic to hexagonal phase transition temperature near 35 1C has revealed small decreases in transition temperature with increasing hydration, suggesting that hydration helps to uidize the SC lipids [30]. However, the precise effects of water on lipid ordering within the lamellae remain undetermined with current evidence suggesting that the lipids remain relatively unperturbed by changes in SC hydration [30,45,46,]. The separation of corneocyte interfaces by the presence of intercellular water, particularly in conjunction with increases in temperature, may be the cause of decreases in delamination energy and peak stress values for FH untreated SC as seen in Figs. 3 and 4. While the precise effects of hydration on lipid ordering remain unclear, temperature has been observed to affect the ordering of intercellular lipids by inducing lipid disorder and phase changes. Specically, heatinginduced alkyl-chain disordering of porcine SC lipids has been correlated with increased permeability of water through the SC membrane [47]. In addition, transmission electron diffraction has revealed gradual changes in human SC lipid organization from an orthorhombic to hexagonal to uid phase when heated from 20 to 90 1C, representing changes toward a more disordered state [48,49]. Differential scanning calorimetry of human SC reveals major transitions near 70 and 80 1C, which have been deduced to reect intercellular lamellar lipid melting then subsequent dissociation of lipidprotein complexes formed at corneocyte envelope interfaces, respectively [5052]. Despite such transitions, abrupt changes in delamination energy with temperature are not necessarily expected due to a competition between decreased lipid cohesion and increased plastic dissipation from greater lipid uidity. The lack of substantial changes in delamination energy for RHH specimens as seen in Figs. 3(a) and 4(a) as well as unpublished results examining a larger temperature range from 10 to 90 1C further corroborate this notion. The lack of drastic changes in delamination energies and peak stress values for the RHH SC compared to FH SC indicate that lipid disordering is not enough to weaken the intercellular structure signicantly (Figs. 3 and 4). Increased hydration in concert with increasing temperature seems to play a key role in reducing delamination energy and peak stress values. In pathologic skin states including atopic dermatitis and lamellar ichthyosis, impaired barrier properties and a prevalence

ARTICLE IN PRESS
792 K.S. Wu et al. / Biomaterials 27 (2006) 785795

of hexagonal and uid lipid phases suggest that these lipid regimes may be associated with higher diffusive mobility of substances through the SC [49]. At intermediate temperatures where multiple phases coexist, certain substances exhibit increased SC penetration attributed to diffusion-enhancing phase boundaries [45,49,53]. As temperature is increased, the lipid phase transitions likely enable additional diffusion paths that allow water molecules to penetrate and separate interfaces otherwise inaccessible at lower temperatures consistent with our observed decreases in delamination energy for FH SC and the lack of large changes for the RHH SC. Given the implied signicance of the lipid structures on the cohesive behavior of the SC, removal of the intercellular lipids remains important to understanding their role in SC delamination properties. The intercellular domain has been noted as a pathway for chemical diffusion, as observed in transepidermal water loss experiments and other chemical species tests [17,20,54,55]. Studies have shown that isolated and fully delipidized non-plantar SC cells require lipids for aggregation and proper cohesion [42]. Furthermore, mechanical tests on individual SC cells have yielded substantially higher modulus values compared to coherent SC tissue, highlighting the effect of intercellular constituents on SC mechanical behavior [56]. Still, understanding of the contribution of the intercellular lipids to the cohesive strength of the SC remains incomplete [54,57,58]. The chloroformmethanol lipid extractions used in the present study have been noted to remove key lipids believed to be responsible for regulation of desquamation, in particular, cholesterol sulfate, whose excess is partly responsible for ichthyosis [5860]. However, such chemical treatment does not lead to cell dissociation as observed here and by others [60]. Indeed, the measured delamination energy for the treated specimens was substantially higher compared to the untreated controls (Fig. 4(a)). SS tests on the treated specimens also resulted in higher peak stresses (Fig. 4(b)). Similar observations have been noted by others using less quantitative cohesometry techniques [22]. The effects of chloroformmethanol lipid extraction on the SC reveal marked modications to the intercellular space. Treatment with similar chloroform methanol solutions has shown via thin-layer chromatography that the majority of the lipid lamellae are leached from the specimen [32]. The treatment results in intimate contact between unextracted lipids covalently bound to the cornied envelopes of adjacent corneocytes and seemingly leaves corneosomes unaltered [22,6164]. The increased cohesion of delipidized SC has been attributed to the interaction between remaining lipids in opposing corneocyte envelopes and postulated to result from interdigitation of the opposing envelope lipids [64,65]. Supporting this notion, extraction of the lamellae and

envelope lipids leads to dissociation of SC tissue into individual cells [42]. While lipid extraction signicantly increased delamination energies compared to those of untreated SC, the resulting fracture surfaces exhibited similar features dependent on initial tissue hydration (Fig. 5). Even with these differences in failure surface morphologies, the chemically treated RHH and FH specimens possessed delamination energy values and peak stresses proportionately more similar than their untreated counterparts, further highlighting the importance of intercellular lipids on SC cohesive properties (Fig. 4). In addition to intercellular lipid contributions to SC integrity, the corneosome protein linkages between cells are known to play a critical role in SC cohesion. Corneosome degradation is accelerated with increasing SC hydration [25]. Chapman et al. have shown that adhesion testing on porcine SC reveals a gradient in SC intercellular cohesion with progressively weaker bonding from the interior toward the more supercial SC layers [22]. These results were correlated with a gradient in corneosome areal density, in which the number of corneosomes between adjacent SC layers, or nonperipheral corneosomes, was observed to decrease progressively from the inner to outer cell layers [21,22]. Similar observations have been made in human SC [26]. However, the DCB test results reveal more complex trends than simply that increased hydration leads to decreased cohesion (Figs. 3 and 4). It is likely that the delamination energies are not correlated highly with expected corneosome cohesive contributions due to the location of the delamination in the few outer layers of the SC, as determined by examination of the graded properties through the thickness of the SC [40]. 4.3. Comparison of out-of-plane to in-plane mechanical studies of stratum corneum The reported in-plane properties of SC are signicantly different from the out-of-plane behavior presented in this study. These differences reveal the highly anisotropic nature of the SC composite. With respect to elastic behavior, in-plane tensile tests on newborn rat SC led to measurements of decreasing tensile moduli ranging from 8800 to 12 MPa with increasing humidity, while similar tests on human SC revealed moduli decreasing from 80 to 20 MPa, which we calculated assuming a SC thickness of 10 mm since only force, not stress, data were reported [6,9]. Our own initial in-plane tensile studies have yielded moduli decreasing from 1000 to 5 MPa with increasing hydration. Similar decreases in modulus for rat SC have been reported with increasing temperature [9]. These variations have been associated with the in-plane orientation of the intermediate laments in the keratinized corneocytes and the substitution of existing proteinprotein hydrogen bonds

ARTICLE IN PRESS
K.S. Wu et al. / Biomaterials 27 (2006) 785795 793

with water-mediated bonding to facilitate greater ber mobility [7,9,66]. Associated changes in the intercellular lipids with temperature and hydration may also affect in-plane tensile moduli. In FH SC tensile tests, corneocytes were observed to slide past each other, highlighting the importance of intercellular components. Compared to our human SC in-plane modulus measurements from 1000 to 5 MPa with increasing hydration, initial out-of-plane modulus measurements yielded values in the range of 125 MPa, substantially lower than those of the in-plane values. Interestingly, while SC lipid transitions have been observed to be largely hydration independent, in the present study mechanical properties are sensitive to the presence of water to facilitate decreases in structural integrity and cohesive strength, albeit for different reasons out-ofplane and in-plane [30,45,46,66]. The out-of-plane SC peak cohesive strengths were also found to be substantially reduced compared to inplane values. For human SC, in-plane strength values have been reported to decrease from 22 to 5 MPa when testing in 0100% RH environments with ether-treated specimens yielded elevated strengths ranging from 25 to 19 MPa in the same environments [6]. Our preliminary in-plane tensile measurements gave similar results with strengths ranging from 18 to 2 MPa in testing environments of 10100% RH. In contrast, the out-ofplane strengths are signicantly lower in the range 0.10.8 MPa (Figs. 3(b) and 4(b)). The CMT SC specimens exhibited somewhat higher out-of-plane strength values of 0.41.4 MPa, but still signicantly lower than in-plane values. The lower out-of-plane strength values are not unexpected, given the continual renewal of the SC which sheds corneocytes perpendicular to the skin surface. Only limited work to measure in-plane fracture properties of human SC as a function of varying environmental conditions has been reported. Mean fracture energies of 3600 J/m2 for 76% RH-conditioned SC specimens have been measured in a tearing conguration [10]. Subsequent work suggested that delamination energy increases with increasing hydration [11]. This trend was not observed in the delamination results of the present study. The in-plane fracture energy is expected to be higher than that of the out-of-plane direction, where lower cohesive values of 18 J/m2 are needed to facilitate natural desquamation. The magnitude of the difference is, however, surprising. Peripheral corneosomes that help to connect the cells in-plane may contribute to the high toughness values. Note, however, that the in-plane tearing energies reported above were calculated assuming linear elastic behavior. Viscoplastic behavior of the SC, particularly at higher hydrations, and the unconstrained nature of the tearing conguration suggest that the reported in-plane fracture energies should be treated with caution.

5. Conclusions In conclusion, a fracture-mechanics-based approach using DCB specimens has been developed to examine quantitatively the delamination properties of human SC. In conjunction with out-of-plane SS measurements, a decrease in delamination energy with increasing test temperature was observed in FH (100% RH) SC tissue. Little effect of testing temperature was observed for RHH (45% RH) SC tissue. The observed decrease in delamination energy was associated with a reduction in SC cohesive strength. In comparison, delamination energy values of CMT specimens exhibited little hydration dependence and were higher than those of untreated controls, as explained by corneocyte envelope interactions. Further exploration is necessary to understand the microstructural changes that occur with delipidization, particularly in conjunction with temperature variation. Initial measurements of modulus reveal that SC stiffness in this mode is much lower than that reported in-plane. From comparisons to in-plane experiments, the fracture energy and modulus data reveal that a simple bricks and mortal model requires renement to explain the highly anisotropic mechanical behavior exhibited by SC.

Acknowledgements The authors would like to thank Dr. Jay Audett and the ALZA Corporation for funding in support of the research, and Ms. Nieves Crisologo and Ms. Mrudula Patel from ALZA Corporation for their assistance in providing the stratum corneum. K.W. was supported by a Stanford Graduate Fellowship. References
[1] Thody AJ, Friedmann PS. Functions of the skin. In: Thody AJ, Friedmann PS, editors. Scientic basis of dermatology: a physiological approach. Edinburgh, London: Churchill Livingstone Inc.; 1986. p. 15. [2] Montagna W, Lobitz Jr. WC, editors. The epidermis. New York: Academic Press, 1964. [3] Wilkes GL, Brown IA, Wildnauer RH. The biomechanical properties of skin. CRC Crit Rev Biomed Eng 1973;1:45395. [4] Marks R, Plewig G, editors. Stratum corneum. Berlin: Springer; 1983. [5] Harding CR. The stratum corneum: structure and function in health and disease. Dermatol Ther 2004;17:615. [6] Wildnauer RH, Bothwell JW, Douglass AB. Stratum corneum biomechanical properties. I. Inuence of relative humidity on normal and extracted human stratum corneum. J Invest Dermatol 1971;56:728. [7] Wilkes GL, Wildnauer RH. Structureproperty relationships of the stratum corneum of human and neonatal rat. II. Dynamic mechanical studies. Biochim Biophys Acta 1973;304:27689. [8] Wolfram MA, Wolejsza NF, Laden K. Biomechanical properties of delipidized stratum corneum. J Invest Dermatol 1973;59:4216.

ARTICLE IN PRESS
794 K.S. Wu et al. / Biomaterials 27 (2006) 785795 lamellae: damage is similar to surfactants. J Invest Dermatol 1999;113:9606. Dauskardt RH, Lane M, Ma Q, Krishna N. Adhesion and debonding of multi-layer thin lm structures. Eng Fract Mech 1998;61:14162. Takahashi M, Machida Y. The inuence of hydroxy acids on the rheological properties of stratum corneum. J Soc Cosmet Chem 1985;36:17787. Bouwstra JA, de Graff A, Gooris GS, Nijsse J, Wieches JW, van Aelst AC. Water distribution and related morphology in human stratum corneum at different hydration levels. J Invest Dermatol 2003;120:7508. n L, Emilson A, Forslind B. Stratum corneum swelling. Norle Biophysical and computer assisted quantitative assessments. Arch Dermatol Res 1997;289:50613. Sznitowska M, Janicki S, Williams A, Lau S, Sto"yhwo A. pHinduced modications to stratum corneum lipids investigated using thermal, spectroscopic, and chromatographic techniques. J Pharm Sci 2003;92:1739. Taub MB, Dauskardt RH. Adhesion of pressure sensitive adhesives with applications in transdermal drug delivery. In: Mallapragada S, Korsmeyer R, Mathiowitz E, Narasimhan B, Tracy M, editors. Biomaterials for drug delivery and tissue engineering. Materials Research Society Proceedings, 2000 November 27December 1, Boston, MA, USA. Warrendale, PA: Materials Research Society; 2001. v. 662. p. NN4.9.1NN4.9.6. Ohashi KL, Romero AC, McGowan PD, Maloney WJ, Dauskardt RH. Adhesion and reliability of interfaces in cemented total joint arthroplasties. J Orthopaed Res 1998;16:70514. Hohlfelder RJ, Maidenberg DA, Dauskardt RH, Wei Y, Hutchinson JW. Adhesion of benzocyclobutene-passivated silicon in epoxy layered structures. J Mater Res 2001;16:24355. Snodgrass JM, Pantelidis D, Jenkins ML, Bravman JC, Dauskardt RH. Subcritical debonding of polymer/silica interfaces under monotonic and cyclic loading. Acta Mater 2002;50: 2395411. Kanninen MF. An augmented double cantilever beam model for studying crack propagation and arrest. Int J Fract 1973;9:8392. Taub MB, Dauskardt RH. Unpublished results. Stanford University; 2004. Jenkins ML, Dauskardt RH, Bravman JC. Important factors for silane adhesion promoter efcacy: surface coverage, functionality, and chain length. J Adhesion Sci Tech 2004;18:1483588. Wu KS, Dauskardt RH. Unpublished results. Stanford University; 2005. Williams ML, Elias PM. Stratum corneum lipids in disorders of cornication. Increased cholesterol sulfate content of stratum corneum in recessive xlinked ichthyosis. J Clin Invest 1981; 68:140410. Smith WP, Christensen MS, Nacht S, Gans EH. Effect of lipids on the aggregation and permeability of human stratum corneum. J Invest Dermatol 1982;78:711. van Hal DA, Jeremiasse E, Junginger HE, Spies F, Bouwstra JA. Structure of fully hydrated human stratum corneum: a freeze fracture electron microscopy study. J Invest Dermatol 1996; 106:8995. Ohta N, Ban S, Tanaka H, Nakata S, Hatta I. Swelling of intercellular lipid lamellar structure with short repeat distance in hairless mouse stratum corneum as studied by X-ray diffraction. Chem Phys Lipids 2003;123:18. Bouwstra JA, Gooris GS, Bras W, Downing DT. Lipid organization in pig stratum corneum. J Lipid Res 1995;36:68595. Mak VH, Potts RO, Guy RH. Does hydration affect intercellular lipid organization in the stratum corneum? Pharm Res 1991;8:10645. [9] Papir YS, Hsu K-H, Wildnauer RH. The mechanical properties of the stratum corneum. I. The effect of water and ambient temperature on the tensile properties of newborn rat stratum corneum. Biochim Biophys Acta 1975;399:17080. [10] Koutroupi KS, Barbenel JC. Mechanical and failure behavior of the stratum corneum. J Biomech 1990;23:2817. [11] Nicolopoulos CS, Giannoudis PV, Glaros KD, Barbenel JC. In vitro study of the failure of skin surface after inuence of hydration and preconditioning. Arch Dermatol Res 1998;290: 63840. [12] Weigand D, Gaylor JR. Removal of stratum corneum in vivo: an improvement on the cellophane tape stripping technique. J Invest Dermatol 1973;60:847. [13] Marks R, Nicholls S, Ritzbeorge D. Measurement of intracorneal cohesion in man using in vivo techniques. J Invest Dermatol 1977;69:299302. [14] Dykes PJ, Heggie R, Hill SA. Effects of adhesive dressings on the stratum corneum of the skin. J Wound Care 2001;10:710. [15] Wu KS, Van Osdol WW, Dauskardt RH. Mechanical and microstructural properties of stratum corneum. In: McKittrick J, Aizenberg J, Orme C, Vekilov P, editors. Biological and biomimetic materialsProperties to function. Materials Research Society Proceedings, 2002 April 15, San Francisco, CA, USA. Warrendale, PA: Materials Research Society; 2002. v. 724. p. 2733. [16] Elias PM, Grayson S, Lampe MA, Williams ML, Brown BE. The intercorneocyte space. In: Marks R, Plewig G, editors. Stratum corneum. Berlin: Springer-Verlag; 1983. p. 5367. [17] Friedmann PS. The skin as a permeability barrier. In: Thody AJ, Friedmann PS, editors. Scientic basis of dermatology: a physiological approach. Edinburgh, London: Churchill Livingstone, Inc.; 1986. p. 2635. [18] Simonetti O, Hoogstraate AJ, Bialik W, Kempenaar JA, HE, Ponec M. Visualization of diffusion Schrijvers AH, Bodde pathways across the stratum corneum of native and in-vitroreconstructed epidermis by confocal laser scanning microscopy. Arch Dermatol Res 1995;287:46573. [19] Johnson ME, Mitragotri S, Patel A, Blankschtein D, Langer R. Synergistic effects of chemical enhancers and therapeutic ultrasound on transdermal drug delivery. J Pharm Sci 1996;85:6709. [20] Fartasch M. Epidermal barrier in disorders of the skin. Microsc Res Tech 1997;38:36172. [21] Chapman SJ, Walsh A. Desmosomes, corneosomes and desquamation An ultrastructural study of adult pig epidermis. Arch Dermatol Res 1990;282:30410. [22] Chapman SJ, Walsh A, Jackson SM, Friedmann PS. Lipids, proteins and corneocyte adhesion. Arch Dermatol Res 1991;283:16773. [23] Walsh A, Chapman SJ. Sugars protect desmosome and corneosome glycoproteins from proteolysis. Arch Dermatol Res 1991; 283:1749. [24] Lundstro m A, Serre G, Haftek M, Egelrud T. Evidence for a role of corneodesmosin, a protein which may serve to modify desmosomes during cornication in stratum corneum cell cohesion and desquamation. Arch Dermatol Res 1994;286:36975. [25] Rawlings A, Harding C, Watkinson A, Banks J, Ackerman C, Sabin R. The effect of glycerol and humidity on desmosome degradation in stratum corneum. Arch Dermatol Res 1995;287: 45764. [26] Simon M, Bernard D, Minondo A-M, Camus C, Fiat F, Corcuff P, Schmidt R, Serrre G. Persistence of both peripheral and nonperipheral corneodesmosomes in the upper stratum corneum of winter xerosis skin versus only peripheral in normal skin. J Invest Dermatol 2001;116:2330. [27] Warner RR, Boissy YL, Lilly NA, Spears MJ, McKillop K, Marshall JL, Stone KJ. Water disrupts stratum corneum lipid

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37] [38] [39]

[40] [41]

[42]

[43]

[44]

[45] [46]

ARTICLE IN PRESS
K.S. Wu et al. / Biomaterials 27 (2006) 785795 [47] Potts RO, Francoeur ML. Lipid biophysics of water loss through the skin. Proc Natl Acad Sci USA 1990;87:38713. [48] Pilgram GS, Engelsma-van Pelt AM, Bouwstra JA, Koerten HK. Electron diffraction provides new information on human stratum corneum lipid organization studied in relation to depth and temperature. J Invest Dermatol 1999;113:4039. [49] Pilgram GS, Vissers DC, van der Meulen H, Pavel S, Lavrijsen SP, Bouwstra JA, Koerten HK. Aberrant lipid organization in stratum corneum of patients with atopic dermatitis and lamellar ichthyosis. J Invest Dermatol 2001;117:7107. [50] Van Duzee BF. Thermal analysis of human stratum corneum. J Invest Dermatol 1975;65:4048. [51] Golden GM, Guzek DB, Harris RR, McKie JE, Potts RO. Lipid thermotropic transitions in human stratum corneum. J Invest Dermatol 1986;86:2559. [52] Gay CL, Guy RH, Golden GM, Mak VHW, Francoeur ML. Characterization of low-temperature (i.e., o651C) lipid transitions in human stratum corneum. J Invest Dermatol 1994; 103:2339. [53] Ogiso T, Hirota T, Iwaki M, Hina T, Tanino T. Effect of temperature on percutaneous absorption of terodiline, and relationship between penetration and uidity of the stratum corneum lipids. Int J Pharm 1998;176:6372. [54] Madison K. Barrier function of the skin: la raison de tre of the epidermis. J Invest Dermatol 2003;121:23141. [55] Yu B, Kim KH, So PT, Blankschtein D, Langer R. Visualization of oleic acid-induced transdermal diffusion pathways using twophoton uorescence microscopy. J Invest Dermatol 2003;120: 44855. ve [56] Le que JL, Poelman MC, de Rigal J, Kligman AM. Are corneocytes elastic? Dermatologica 1988;176:659. 795 [57] Elias PM. Epidermal lipids, barrier function, and desquamation. J Invest Dermatol 1983;80:44s9s. [58] Egelrud T, Lundstro m A. The dependence of detergent-induced cell dissociation in non-palmo-plantar stratum corneum on endogenous proteolysis. J Invest Dermatol 1990;95:4569. [59] Long SA, Wertz PW, Strauss JS, Downing DT. Human stratum corneum polar lipids and desquamation. Arch Dermatol Res 1985;277:2847. [60] Takahashi M, Aizawa M, Miyazawa K, Machida Y. Effects of surface active agents on stratum corneum cell cohesion. J Soc Cosmet Chem 1987;38:218. [61] Swartzendruber DC, Wertz PW, Madison KC, Downing DT. Evidence that the corneocyte has a chemically bound lipid envelope. J Invest Dermatol 1987;88:70913. [62] Wertz PW, Swartzendruber DC, Madison KC, Downing DT. Composition and morphology of epidermal cyst lipids. J Invest Dermatol 1987;89:41925. [63] Wertz PW, Madison KC, Downing DT. Covalently bound lipid of human stratum corneum. J Invest Dermatol 1989;92:10911. [64] Wertz PW, Swartzendruber DC, Kitko DJ, Madison KC, Downing DT. The role of the corneocyte lipid envelopes in cohesion of the stratum corneum. J Invest Dermatol 1989; 93:16972. [65] Stewart ME, Downing DT. The o-hydrocyceramides of pig epidermis are attached to corneocytes solely through o-hydroxyl groups. J Lipid Res 2001;42:110510. [66] Wilkes GL, Nguyen A-L, Wildnauer R. Structureproperty relations of human and neonatal rat stratum corneum. I. Thermal stability of the crystalline lipid structure as studied by X-ray diffraction and differential thermal analysis. Biochim Biophys Acta 1973;304:26775.