- Email: [email protected]
Inhibition of dermal fibrosis in self-assembled skin equivalents by undifferentiated keratinocytes
Journal of Dermatological Science 53 (2009) 103–111
Contents lists available at ScienceDirect
Journal of Dermatological Science journal homepage: www.intl.elsevierhealth.com/journals/jods
Inhibition of dermal fibrosis in self-assembled skin equivalents by undifferentiated keratinocytes Xinwen Wang a,b,1, Yuan Liu a,1, Zhihong Deng c, Rui Dong a, Yanli Liu a, Shijie Hu a, Yuan Li a, Yan Jin a,* a b c
Research and Development Center for Tissue Engineering, School of Stomatology, FMMU, China Department of Periodontology and Oral Medicine, School of Stomatology, FMMU, China Department of Otolaryngology, Xijing Hospital, FMMU, China
A R T I C L E I N F O
S U M M A R Y
Article history: Received 24 November 2007 Received in revised form 5 May 2008 Accepted 16 August 2008
Background: Previous studies showed that keratinocyte plays a major role in dermal cell behavior and hypertrophic scar formation. Further investigations showed that keratinocytes derived from normal skin and hypertrophic scar have different effects on dermal fibroblasts. Objective: To investigate the role of undifferentiated keratinocytes in epidermal–dermal interaction and dermal fibrosis. Methods: A tissue-engineered model of self-assembled reconstructed skin was used in this study to mimic interactions between dermal and epidermal cells. Transmission electron microscope, RT and Western blot analysis were performed to show extracellular matrix morphology, collagen synthesis and associated factors expression changes. Results: The dermal extracellular matrix co-cultured with undifferentiated keratinocytes was well distributed, collagen bundles were not seen, and the levels of collagen mRNA and protein expression declined to 46%, 20% of that in the presence of differentiated keratinocytes. Undifferentiated keratinocytes inhibited dermal fibrosis through down-regulation of TGFb1, promoting bFGF expression and desmosome formation. Conclusions: Undifferentiated keratinocytes have the ability to preserve normal epidermal–dermal interaction and inhibit dermal fibrosis. Absence or diminution of undifferentiated keratinocytes may take part in initiating events leading to pathological fibrosis. ß 2008 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved.
Keywords: Keratinocyte Beta 1 integrin Keratin 5 Keratin 10 Hypertrophic scar Fibrosis
1. Introduction The healing process for normal wounds induces stop signals that halt the repair process when the dermal defect is closed and epithelialization is complete. When these signals are absent or ineffective, the repair process may continue, resulting in the formation of excessive scar tissue as the result of imbalance in collagen production and maturation [1]. The extracellular matrix (ECM) is an area where epidermal cells have a potential influence on collagen produced by fibroblasts, and hypertrophic scar tissue occurs frequently when re-epithelialization has been delayed [2], and abnormalities in epidermal–dermal cross-talk have been
* Corresponding author. Tel.: +86 29 84776147; fax: +86 29 83218039. E-mail address: [email protected] (Y. Jin). 1 These authors contribute equally to this work.
hypothesized [3,4]. An increasing body of evidence shows that epidermal–mesenchymal interaction may have a major role in dermal cell behavior and hypertrophic scar formation [3,5], but the underlying mechanisms are not known. Bellemare et al. [3] studied interactions between dermis and epidermis using a tissue-engineered skin model and demonstrated that keratinocytes have a role in pathological fibrosis development in the dermis, by influencing proliferation of dermal cells and matrix accumulation. The analysis showed that keratinocytes in normal skin and hypertrophic scar tissue have different effects on fibroblasts; hypertrophic keratinocytes induce significantly thicker dermis or more extracellular matrix formation than normal keratinocytes. The epidermis is a stratified squamous epithelium composed primarily of keratinocytes, which can be divided into basale, spinosum, granulosum and corneum cells on the basis of morphology. They can also be divided into undifferentiated
0923-1811/$30.00 ß 2008 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.jdermsci.2008.08.010
104
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
keratinocytes, which consist of epidermal stem cells (ESCs) and transit amplifying cells (TAs), and terminally differentiated keratinocytes. Undifferentiated keratinocytes are the foundation of epidermal regeneration [6,7]. When skin is damaged, ESCs and TAs divide, differentiate, migrate over the wound matrix, re-
establish the epithelial layer and restore epidermal continuity [8,9], while producing numerous protein factors and have critical influences on the overall skin environment [10,11]. Undifferentiated keratinocytes are most active in the epidermis [12,13], and likely have a critical role in epidermal–mesenchymal interaction.
Fig. 1. Expressions of b1 integrin, K5 and K10 in hypertrophic scar and normal skin. In normal skin, expressions of b1 integrin (A) and K5 (B) were restricted to cells over connective tissue, expression of K10 could be observed in all keratinocytes except basal cells (C). The expression pattern in hypertrophic scar was extremely different from this. In hypertrophic scar, b1 integrin (D) and K5 (E) positive cells were barely detectable, but K10 expressed all over the epidermis (F). The scale bar represents 50 mm.
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
The absence or diminution of undifferentiated keratinocytes may lead to altered epidermal–mesenchymal interaction and initiate events leading to pathological fibrosis. To test this hypothesis, undifferentiated keratinocytes in normal skin and in hypertrophic scar tissue were compared, and their inhibition of extracellular matrix accumulation was proved using a tissue-engineered model of self-assembled reconstructed skin.
105
2. Materials and methods 2.1. Immunohistochemistry of cutaneous biopsies 30–40 days hypertrophic scar tissue with site-matched normal skin samples were collected from healthy individuals who gave oral and written informed consent. Five normal skin and five
Fig. 2. Construction of skin equivalents and histological analysis. The b1 integrin-positive undifferentiated keratinocytes (UKs) dispersed among keratinocytes before sorting (A). Magnetic cell sorting allowed the separation of UKs (B), which were absent from the remaining keratinocytes (C). H&E staining showed that the UK epidermis had the expected layers, with round basal cells arranged in good order (D). The skin equivalents constructed with DKs had a thin epidermis without typical layers; moreover, the epidermis was prone to separate from the dermis (arrows) (E). The USK epidermis did not have very distinct layers, although it contained obvious stratum corneum (F). The FB formed typical dermis which was used as a negative control (G). The scale bar represents 50 mm. Immunofluorescent staining showed that b1 integrin was expressed in the basal layer of UK epidermis. The positive cells arranged in good order in basal area (H); in DK epidermis, b1 integrin expression was still absent (I); nearly extinct b1 integrin expression could be observed occasionally in USK epidermis (white arrow) (J). The scale bar represents 25 mm.
106
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
Fig. 2. (Continued ).
hypertrophic scar specimens paraformaldehyde-fixed and routinely embedded were used. Paraffin sections (5 mm) were immunohistochemically stained using EnVision+1 system (DAKO, USA). An indirect method using peroxidase-labeled secondary antibodies was carried out, with reagent dilutions and incubation conditions selected according to manufacturer’s instructions. Briefly, the sections were deparaffinized and rehydrated in an ethanol series, recovered protein antigenicity in glycin/HCl buffer (50 mM, pH 3.5). Sections were incubated with monoclonal anti-b1 integrin (Oncogene, UK), monoclonal anti-K5 (DAKO, USA) and monoclonal anti-K10 (DAKO, USA) separately at 4 8C overnight. For controls, the primary antibody was omitted. The sections were then immunostained with horseradish peroxidase-conjugated goat anti-mouse IgGs (at a 1:50 dilution) for 1 h at 37 8C. The colorimetric reagent was diaminobenzidine/hydrogen peroxide and all sections were counterstained with hematoxylin and examined by light microscopy. 2.2. Cell culture Normal keratinocytes and dermal fibroblasts were prepared from foreskin obtained from patients aged 6–18 years undergoing
circumcision. All samples were processed as described by Barrandon and Green [14] and Germain et al. [15]. Fibroblasts were cultured in DMEM supplemented with 10% fetal calf serum (FCS). Cells under passage 5 were used for experiments. Keratinocytes were cultured in Keratinocyte Serum-Free Medium (GIBCO, USA) supplemented with epidermal growth factor (0.15 ng/ml), and bovine pituitary extract (25 mg/ml). 1–2 passage keratinocytes were used for cell sorting. 2.3. Separation of undifferentiated keratinocytes Samples of cell preparations were labeled using an indirect technique, in which the cells were incubated with monoclonal mouse anti-b1 integrin IgG (Oncogene, UK; dilution 1:50) for 30 min at 4 8C. After two washes with PBS supplemented with 1% FCS, the cells at a final concentration of 107 cells/100 ml were labeled with goat anti-mouse IgG, separated magnetically for 10 min and the magnetic pellet saved for a positive selection. The remainder was saved as differentiated keratinocytes. Indirect immunocytochemical staining method was used for sorting evaluation. The sorted cells were grown on coverslips for 1 day and then fixed. After being treated with 0.3% Triton X-100, the coverslips were blocked with 5% normal goat serum, then were incubated with mouse anti-b1 integrin (1:50; Oncogene, UK)
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
107
Fig. 3. Electron micrograph of epidermal basal layer in skin equivalents. TEM analysis showed that desmosomes (arrows) were seen in the basal layer of UK epidermis (A). Further enlargement revealed structural details in the marked area, showing that desmosomes were characterized by a much wider intermembrane space and denser distribution (arrows) (B). TEM micrographs showed an absence of cell–cell junctions in the DK epidermis (C), and enlarged spaces between cells (arrows) (D). Desmosomes could be seen in the USK epidermis, but were sparse (arrows) (E and F). The scale bar represents 2 mm. Quantitative analysis showed that the number of desmosomes in the UK epidermis were increased >5-fold compared to that in the USK epidermis (G). The results are given as the mean S.E.M. (n = 3).
108
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
Fig. 4. Ultrastructure of dermal ECM and expressions of collagen, bFGF and TGFb1 in different skin equivalents. TEM showed that in the presence of UKs, microfibrils were scattered without directionality, and collagenous bundles were not seen (A). In contrast, when DKs were added onto dermal sheets, the newly synthesized ECM formed collagen fiber bundles with longitudinal orientation (arrows) (B). In the presence of USKs, microfibril bundles were sparse, short and orientated randomly (arrows) (C). In the
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
antibodies at 4 8C overnight. At last, immunostained with horseradish peroxidase-conjugated goat anti-mouse IgGs (at a 1:50 dilution). All coverslips were counterstained with hematoxylin and examined by light microscopy. 2.4. Production of skin equivalents Dermal fibroblasts were grown for 35 days in the presence of 50 mg/ml ascorbic acid (GIBCO, USA) to form a dermal sheet as described [16]. Four layers were superimposed and cultured for 1 week to form reconstructed dermis. Undifferentiated (UK), differentiated (DK) and unsorted (USK) keratinocytes were seeded onto the dermal sheets at a density of 2 105 cells/cm2. Reconstructed dermis without the addition of keratinocytes (FB) served as a negative control. Keratinocytes reached confluence after 8 days of submerged culture and were then raised at the air– liquid interface. Samples were used after 2 weeks of air–liquid culture for analysis. Sections 5 mm thick were stained with hematoxylin and eosin (H&E) for histological observation. 2.5. Immunofluorescence staining
b1 integrin expression in skin equivalents was detected by immunofluorescence staining. Sections were deparaffinized in xylene, hydrated sections through graded alcohols. Incubated slides for 20 min in 10% goat serum, and then incubated with mouse anti-b1 integrin overnight at 4 8C, FITC-conjugated secondary antibody for 30 min, washed three times with PBS and mounted aqueous coverslips. Images of staining were captured on a fluorescence microscope. 2.6. Transmission electron microscopy Samples were fixed with 0.1 M sodium cacodylate buffer containing 2.5% glutaraldehyde and post-fixed with osmium tetroxide (SIGMA, Canada). Tissues were stained with 1% uranyl acetate, and examined with a JEM-100SX transmission electron microscope. Estimation of desmosome frequency was done as described [17]. Micrographs were taken at magnifications of 4000 and 15,000. The percentage of cell membrane length occupied by desmosomes was calculated for different micrographs. The significance of differences in desmosome frequency between different skin equivalents was assessed using Student’s t-test, with P-values of <0.05 considered significant. 2.7. RNA extraction, reverse transcription (RT), and PCR analysis Total RNA was isolated from skin equivalents using TRIzol1 reagent (Invitrogen, USA) according to the manufacturer’s protocol. PCR was performed in a volume of 25 ml and the primers were as follows: the human a1(I) chain gene for collagen (forward) TGTTCAGCTTTGTGGACCTC, (reverse) CTTGGTCTCGTCACAGATCA; the human bFGF gene (forward) CCTCACATCAAGCTACAACT, (reverse) TCAGCTCTTAGCAGACATTG; the human TGFb1 gene (forward) ACCTGCAAGACTATCGACAT, (reverse) GTACTCTGCTTGAACTTGTC; and the human b-actin gene (forward) ATCATGTTTGAGACCTTCAA, (reverse) CATCTCTTGCTCGAAGTCCA. After PCR, the products were subjected to electrophoresis in a 1.5% agarose gel and semi-quantified by comparing the intensity of
109
the band of interest to that of an internal standard (b-actin) using LabWorks Image Acquisition and Analysis Software (UVP, USA). The differences in mRNA expression between skin equivalents were determined and compared. The statistical significance was evaluated using Student’s t-test, with P-values of <0.05 considered significant. 2.8. Western blot analysis Western blot analysis was done with antibodies specific for procollagen type I (1:2000, polyclone, Maixin, China), bFGF (1:1000, monoclone, Santa Cruz, USA), and TGFb1 (1:2000, polyclone, Maixin, China). Extracted proteins were separated by SDS-PAGE (10% polyacrylamide gel), and the separated proteins were transferred to nitrocellulose membranes (Millipore, USA). Membranes were incubated with primary antibodies followed by incubation with horseradish peroxidase-conjugated secondary antibodies. Bound antibodies were detected by enhanced chemiluminescence (ECL plus kit, UK). A human antibody specific for b-actin (1:1000 42 kDa, Maixin, China) was used as a loading control. 3. Results 3.1. b1 integrin, K5, K10 expression in hypertrophic scar tissue Immunohistochemical staining showed that b1 integrin and Keratin 5 (K5) antibodies stained only the basal cell layer of normal skin (Fig. 1A and B). A consistent staining could be found in suprabasal keratinocytes with K10 antibody but not in the basal cells (Fig. 1C). By contrast, in hypertrophic scar there was a strong expression of K10 all over the epidermis (Fig. 1F), b1 integrin and K5 expressing keratinocytes were nearly undetectable (Fig. 1D and E). 3.2. Cell sorting evaluation and construction of skin equivalents Immunocytochemical staining showed that UKs expressing b1 integrin scattered in keratinocytes before sorting (Fig. 2A). Using magnetic beads, UKs were separated efficiently (Fig. 2B). In remainder cells b1 integrin expression was hardly detected, they were saved as differentiated keratinocytes (Fig. 2C). Skin equivalents were constructed with different keratinocyte population, H&E staining showed that UKs formed maturated epidermis with typical morphological character (Fig. 2D), whereas DKs formed very thin epidermis which was easy to separate from connective tissue (Fig. 2E). In USK epidermis, although there was obvious stratum corneum, the layers were not as typical as those in the UK epidermis (Fig. 2F). In negative group, FB formed typical dermis (Fig. 2G). Immunofluorescent staining revealed that in UK epidermis only the basal keratinocytes situated over the connective tissue still keep the ability to express b1 integrin (Fig. 2H). Few of b1 integrin strongly expressing keratinocytes were detected in USK epidermis (Fig. 2J). 3.3. The formation of desmosome and dermal matrix morphology in skin equivalents Transmission electron microscopy demonstrated that basal cells in the UK epidermis were connected tightly through desmosomes
FB dermis, microfibrils were arranged much more densely with many visible collagen clusters (arrows) around active fibroblasts (D). The scale bar represents 2 mm. Expression of collagen, bFGF and TGFb1 mRNA in different skin equivalents was detected by RT-PCR (E) and semi-quantified by image analysis (F). The results are given as mean S.E.M. (n = 5). Protein lysates from different skin equivalents were used for Western blot analysis to detect protein levels of collagen, bFGF and TGFb1 (G). Semiquantification was done with image analysis (H). The results are expressed as mean S.E.M. (n = 5). Significant results vs. FB dermis are shown as *, significant results vs. DK skin equivalent are shown as # and significant results vs. USK skin equivalent are shown as ~.
110
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
(Fig. 3A and B). The frequency (Fig. 3G) and morphological appearance of desmosomes were similar in normal skin. However, very few desmosomes were found in the DK epidermis (Fig. 3C); the keratinocytes did not show tight cell-to-cell connection and were prone to separate from each other (Fig. 3D). Desmosomes were found in the USK epidermis (Fig. 3E and F), but fewer than in the UK epidermis. Statistical analysis of desmosome frequency indicated a significant difference between different skin equivalents (Fig. 3G). When UKs were added onto dermal sheets, the dermal ECM was well distributed, microfibrils were deposited without directionality, and bundles were not seen (Fig. 4A). When USKs were added onto dermal sheets, the ECM showed collagenous fibers that were short and orientated randomly (Fig. 4C). The presence of DKs or the absence of keratinocytes led to a greater production of collagen in the dermis. Tight bundles of microfibrils were formed with a longitudinal orientation (Fig. 4B and D). Especially in the FB dermis, microfibrillar material was more abundant and dense around active fibroblasts, with a disordered appearance (Fig. 4D). 3.4. Undifferentiated keratinocytes reduce collagen expression in skin equivalents Expression of collagen is essential for dermal fibrosis, in which this protein is deposited in excess amounts [18]. This study showed that when DKs were added onto dermal sheets, the levels of collagen mRNA and protein were very high (0.35 0.01 and 1.50 0.23, respectively), but were downregulated markedly when the dermal sheets were seeded with UKs (Fig. 4E–H). The level of mRNA expression declined to 46% of that in the presence of DKs (0.16 0.03 vs. 0.35 0.01), and the level of protein expression declined to 20% (0.32 0.09 vs. 1.50 0.23). When USKs were added, the levels of mRNA and protein for collagen were between those of the UK and DK groups (0.20 0.01 and 1.28 0.19, respectively). Consistent with previous studies [3], the FB dermis (no addition of keratinocytes) expressed the highest level of mRNA and protein of collagen (0.45 0.02 and 1.89 0.28, respectively). 3.5. bFGF and TGFb1 mRNA, protein levels were altered by undifferentiated keratinocytes The profile of TGFb1 and bFGF mRNA expression in different skin equivalents is shown in Fig. 4E and F. The addition of UKs led to the highest level of bFGF mRNA (0.72 0.01). However, replacing UKs with DKs resulted in a significant reduction (0.50 0.05). The level of bFGF mRNA in the USK group (0.53 0.07) was between those of the UK and DK groups, and the lowest level of bFGF mRNA was found in the FB dermis. TGFb1 mRNA was barely detectable in the presence of UKs. However, the addition of DKs resulted in an almost >5-fold increase (0.10 0.03 vs. 0.02 0.01) in the production of TGFb1 mRNA. The level of TGFb1 mRNA in the USK group was between those of the UK and DK groups, and the FB dermis expressed the highest level of TGFb1 mRNA. Western blot analysis revealed that the pattern of bFGF and TGFb1 protein expression (Fig. 4G and H) was consistent with that of mRNA expression. When UKs were added onto dermal sheets, the highest level of bFGF was observed and no TGFb1 was detectable. In the presence of DKs, the level of bFGF was relatively low (0.45 0.15). Moreover, the expression of bFGF and TGFb1 in the FB dermis and in the USK group was consistent with that of mRNAs. 4. Discussion
b1 integrin is a cell surface molecule, via which basal keratinocytes attach to the underlying basement [19], and its
expression is downregulated in keratinocytes that have initiated the differentiation program. So b1 integrin is considered a marker of ESCs and TAs. Keratins are serial proteins, which make internal skeleton of keratinocytes. Keratinocytes express different keratins during differentiation. K5 is a protein expressed by basal keratinocytes in normal skin. Once keratinocytes stop dividing and detach from the basement membrane, switch from the expression of K5 to K10, marker of keratinocyte differentiation [20]. In this study, we observed that b1 integrin and keratin 5 expression was restricted to cells over connective tissue in normal skin, and were barely detectable in hypertrophic scar tissue. However, K10 not expressed in basal layer in normal skin were strongly expressed throughout hypertrophic scar epidermis, which suggested that differentiation stage of keratinocytes in hypertrophic scar differs from normal skin, there was no typical undifferentiated keratinocytes in basal layer of hypertrophic scar. Earlier studies showed that keratinocytes at the wound margin are activated when skin is injured [21,22]. Surrounding keratinocytes migrate and divide more frequently to restore the injured ‘‘barrier’’. It has been shown that keratinocytes in hypertrophic scar epidermis enter an alternative differentiation pathway [23] and this, coupled with unbalanced renewal, leads to the loss of undifferentiated keratinocytes. We therefore, hypothesized that absence or diminution of undifferentiated keratinocytes may further lead to altered epidermal–mesenchymal interaction and take part in initiating events leading to pathological fibrosis. Reconstructed human skin is a three-dimensional model made from skin epithelial cells and dermal fibroblasts, which can be used to investigate epidermal–dermal interaction. A reconstructed human skin model similar to in vivo skin was used to evaluate the role of undifferentiated keratinocytes in epidermal–dermal interaction and the effects they produce on extracellular matrix accumulation [16]. Undifferentiated keratinocytes isolated by an immunomagnetic method and added onto the self-assembled dermis directly formed the basis for this study. H&E staining showed that it was easier to construct well-organized skin equivalents with undifferentiated keratinocytes compared with differentiated or unsorted keratinocytes. This is in line with the facts that ESC and TA cells have the ability to form fully differentiated epidermis both in vivo and in vitro [24]. Immunofluorescence results showed that the cells located in basal region of UK epidermis still could express b1 integrin after long-term organotypic culture, but not in USK epidermis, two possibilities are: (1) in the condition of air–liquid culture undifferentiated keratinocytes differentiate into mature keratinocytes at certain ratio, which results in b1 integrin positive keratinocytes remained in USK epidermis much less than UK epidermis and (2) keratinocyte differentiation is controlled by particular microenvironments [25,26]. USKs, a mix of keratinocytes in viable degrees of differentiation provide an environment; in which undifferentiated keratinocytes mature easier. Results showed that UKs inhibited collagen expression at either the mRNA level or the protein level more efficiently than other populations, suggesting that undifferentiated keratinocytes may have a significant role in pathologic collagen matrix remodeling, which is compatible with the results reported by Hohelfeld, who applied expanded fetal skin constructs with a high proportion of undifferentiated keratinocytes to eight patients and showed that the wound healed with little hypertrophy of new skin and no retraction [27]. TGFb1 is an important fibrosis-related cytokine that stimulates the deposition of extracellular matrix components and degradation of the blocking matrix [28,29]. Earlier studies showed decreased expression of collagen in bFGF-treated cells in culture [30,31]. Here, we showed that undifferentiated keratinocytes
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
promoted bFGF expression but suppressed the expression of TGFb1 in contrast to other keratinocytes, which may contribute to a decreased and well-distributed collagen secretion. Our experiments highlight the roles of undifferentiated keratinocytes in dermal fibrosis inhibition. They may release signals that regulate expression of factors associated with matrix remodeling, such as bFGF and TGFb1. Therefore, the lack of undifferentiated keratinocytes leads to failure of the regulation, resulting in pathological accumulation of collagen and dermal fibrosis. Interestingly, we found that cells in the UK epidermis were prone to form cell–cell desmosomes that connected keratinocytes in the nearby basal membrane very tightly. In the DK epidermis, however, it was difficult to see the cell–cell contact. Earlier studies provided strong support for the importance of contact between epithelial cells, and showed that intracellular contacts are active regulators of fibrosis [27,32]. Accordingly, cell contact integrity regulates the extracellular matrix, and inhibits organ fibrosis, which is in line with previous results and further supports the hypothesis. In this study, we examined the role of undifferentiated keratinocytes in epidermal–dermal interaction and fibroblast matrix accumulation, and identified the absence of undifferentiated keratinocytes as one of the key factors leading to pathological fibrosis. These data identify a novel regulatory mechanism for scar formation. Undifferentiated keratinocytes may be a promising option for hypertrophic scar prevention and an alternative target for anti-therapies. Acknowledgments The study was supported by National Natural Science Foundation of China (grant number: 30700870, 30700770) and the NHTRD Program of China (863, grant number: 2005AA205241). References [1] Ghahary A, Shen YJ, Scott PG, Gong Y. Enhanced expression of mRNA for transforming growth factor-beta, type I and type III procollagen in human post-burn hypertrophic scar tissues. J Lab Clin Med 1993;122:465–73. [2] Deitch EA, Wheelahan TM, Rose MP, Clothier J, Cotter J. Hypertrophic burn scars: analysis of variables. J Trauma 1983;23:895–8. [3] Bellemare J, Roberge CJ, Bergeron D, Lopez-Valle CA, Roy M, Moulin VJ. Epidermis promotes dermal fibrosis: role in the pathogenesis of hypertrophic scars. J Pathol 2005;206:1–8. [4] Hakvoort TE, Altun V, Ramrattan RS, Kwast TH, Benner R, Zuijlen PP, et al. Epidermal participation in post-burn hypertrophic scar development. Virchows Arch 1999;434:221–6. [5] Garner WL. Epidermal regulation of dermal fibroblast activity. Plast Reconstr Surg 1998;102:135–9. [6] Barrandon Y, Green H. Cell migration is essential for sustained growth of keratinocyte colonies: the roles of transforming growth factor-alpha and epidermal growth factor. Cell 1987;50:1131–7. [7] Santoro MM, Gaudino G. Cellular and molecular facets of keratinocyte reepithelization during wound healing. Exp Cell Res 2005;304:274–86.
111
[8] Cotsarelis G, Cheng SZ, Dong G, Sun TT, Lavker RM. Existence of slow-cycling limbal epithelial basal cells that can be preferentially stimulated to proliferate: implications on epithelial stem cells. Cell 1989;57:201–9. [9] Hall PA, Watt FM. Stem cells: the generation and maintenance of cellular diversity. Development 1989;106:619–33. [10] Werner S, Smola H, Liao X, Longaker MT, Krieq T, Hofschneider PH, et al. The function of KGF in morphogenesis of epithelium and reepithelialization of wounds. Science 1994;266:819–22. [11] Abraham JA, Klagsburn M. Modulation of wound repair by members of the fibroblast growth factor family. In: Clark RAF, editor. The molecular and cellular biology of wound repair. 2nd ed., New York: Plenum Press; 1996. p. 195–248. [12] Maria IM, Marjana TC. Epidermal stem cells: the cradle of epidermal determination differentiation and wound healing. Biol Cell 2005;97:173–83. [13] Tomic-Canic M, Komine M, Freedberg IM, Blumenberg M. Epidermal signal transduction and transcription factor activation in activated keratinocytes. J Dermatol Sci 1998;17:167–81. [14] Barrandon Y, Green H. Cell size as a determinant of the clone-forming ability of human keratinocytes. Proc Natl Acad Sci USA 1985;82:5390–4. [15] Germain L, Guignard R, Rouabhia M, Auger FA. Early basement membrane formation following the grafting of cultured epidermal sheets detached with thermolysin or Dispase. Burns 1995;21:175–80. [16] Michel M, L’Heureux N, Pouliot R, Xu W, Auger FA, Germain L. Characterization of a new tissue-engineered human skin equivalent with hair. In Vitro Cell Dev Biol Anim 1999;35:318–26. [17] Skerrow CJ, Clelland DG, Skerrow D. Changes to desmosomal antigens and lectin-binding sites during differentiation in normal human epidermis: a quantitative ultrastructural study. J Cell Sci 1989;92(Pt 4):667–77. [18] Clark RA. Biology of dermal wound repair. Am J Med Sci 1993;306:42–8. [19] Jones PH, Watt FM. Separation of human epidermal stem cells from transit amplifying cells on the basis of differences in integrin function and expression. Cell 1993;73:713–24. [20] Rao KS, Babu KK, Gupta PD. Keratins and skin disorders. Cell Biol Int 1996;20:261–74. [21] Grinnell F. The activated keratinocytes: up regulation of cell adhesion and migration during wound healing. J Trauma 1990;30:144–9. [22] Galvin S, loomis C, Manabe M, Dhouailly D, Sun TT. The major pathways of keratinocyte differentiation as defined by keratin expression: an overview. Adv Dermatol 1989;4:277–99. [23] Machesney M, Tidman N, Waseem A, Kirby L, Leigh I. Activated keratinocytes in the epidermis of hypertrophic scars. Am J Pathol 1998;152:1133–41. [24] Li A, Pouliot N, Redvers R, Kaur P. Extensive tissue-regenerative capacity of neonatal human keratinocytes stem cells and their progeny. J Clin Invest 2004;113:390–400. [25] Moore KA, Lemischka IR. Stem cells and their niches. Science 2006;311:1880– 5. [26] David TS. The stem cell niche as an entity of action. Nature 2006;441:1075– 9. [27] Hohlfeld J, de Buys Roessingh A, Hirt-Burri N, Chaubert P, Gerber A, Scaletta C, et al. Tissue engineered fetal skin constructs for paediatric burns. Lancet 2005;366:740–842. [28] Border WA, Ruoslahti E. Transforming growth factor-beta in disease: the dark side of tissue repair. J Clin Invest 1992;90:1–7. [29] Kenyon NJ, Ward RW, McGrew G, Last JA. TGF-beta1 causes airway fibrosis and increased collagen I and III mRNA in mice. Thorax 2003;58:772–7. [30] Kennedy SH, Qin H, Lin L, Tan EM. Basic fibroblast growth factor regulates type I collagen and collagenase gene expression in human smooth muscle cells. Am J Pathol 1995;146:764–71. [31] Kypreos KE, Nugent MA, Sonenshein GE. Basic fibroblast growth factorinduced decrease in type I collagen gene transcription is mediated by Bmyb. Cell Growth Differ 1998;9:723–30. [32] Fan L, Sebe A, Peterfi Z, Masszi A, Thirone AC, Rotstein OD, et al. Cell contactdependent regulation of epithelial-myofibroblast transition via the rho–rho kinase-phospho-myosin pathway. Mol Biol Cell 2007;18:1083–97.
Contents lists available at ScienceDirect
Journal of Dermatological Science journal homepage: www.intl.elsevierhealth.com/journals/jods
Inhibition of dermal fibrosis in self-assembled skin equivalents by undifferentiated keratinocytes Xinwen Wang a,b,1, Yuan Liu a,1, Zhihong Deng c, Rui Dong a, Yanli Liu a, Shijie Hu a, Yuan Li a, Yan Jin a,* a b c
Research and Development Center for Tissue Engineering, School of Stomatology, FMMU, China Department of Periodontology and Oral Medicine, School of Stomatology, FMMU, China Department of Otolaryngology, Xijing Hospital, FMMU, China
A R T I C L E I N F O
S U M M A R Y
Article history: Received 24 November 2007 Received in revised form 5 May 2008 Accepted 16 August 2008
Background: Previous studies showed that keratinocyte plays a major role in dermal cell behavior and hypertrophic scar formation. Further investigations showed that keratinocytes derived from normal skin and hypertrophic scar have different effects on dermal fibroblasts. Objective: To investigate the role of undifferentiated keratinocytes in epidermal–dermal interaction and dermal fibrosis. Methods: A tissue-engineered model of self-assembled reconstructed skin was used in this study to mimic interactions between dermal and epidermal cells. Transmission electron microscope, RT and Western blot analysis were performed to show extracellular matrix morphology, collagen synthesis and associated factors expression changes. Results: The dermal extracellular matrix co-cultured with undifferentiated keratinocytes was well distributed, collagen bundles were not seen, and the levels of collagen mRNA and protein expression declined to 46%, 20% of that in the presence of differentiated keratinocytes. Undifferentiated keratinocytes inhibited dermal fibrosis through down-regulation of TGFb1, promoting bFGF expression and desmosome formation. Conclusions: Undifferentiated keratinocytes have the ability to preserve normal epidermal–dermal interaction and inhibit dermal fibrosis. Absence or diminution of undifferentiated keratinocytes may take part in initiating events leading to pathological fibrosis. ß 2008 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved.
Keywords: Keratinocyte Beta 1 integrin Keratin 5 Keratin 10 Hypertrophic scar Fibrosis
1. Introduction The healing process for normal wounds induces stop signals that halt the repair process when the dermal defect is closed and epithelialization is complete. When these signals are absent or ineffective, the repair process may continue, resulting in the formation of excessive scar tissue as the result of imbalance in collagen production and maturation [1]. The extracellular matrix (ECM) is an area where epidermal cells have a potential influence on collagen produced by fibroblasts, and hypertrophic scar tissue occurs frequently when re-epithelialization has been delayed [2], and abnormalities in epidermal–dermal cross-talk have been
* Corresponding author. Tel.: +86 29 84776147; fax: +86 29 83218039. E-mail address: [email protected] (Y. Jin). 1 These authors contribute equally to this work.
hypothesized [3,4]. An increasing body of evidence shows that epidermal–mesenchymal interaction may have a major role in dermal cell behavior and hypertrophic scar formation [3,5], but the underlying mechanisms are not known. Bellemare et al. [3] studied interactions between dermis and epidermis using a tissue-engineered skin model and demonstrated that keratinocytes have a role in pathological fibrosis development in the dermis, by influencing proliferation of dermal cells and matrix accumulation. The analysis showed that keratinocytes in normal skin and hypertrophic scar tissue have different effects on fibroblasts; hypertrophic keratinocytes induce significantly thicker dermis or more extracellular matrix formation than normal keratinocytes. The epidermis is a stratified squamous epithelium composed primarily of keratinocytes, which can be divided into basale, spinosum, granulosum and corneum cells on the basis of morphology. They can also be divided into undifferentiated
0923-1811/$30.00 ß 2008 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.jdermsci.2008.08.010
104
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
keratinocytes, which consist of epidermal stem cells (ESCs) and transit amplifying cells (TAs), and terminally differentiated keratinocytes. Undifferentiated keratinocytes are the foundation of epidermal regeneration [6,7]. When skin is damaged, ESCs and TAs divide, differentiate, migrate over the wound matrix, re-
establish the epithelial layer and restore epidermal continuity [8,9], while producing numerous protein factors and have critical influences on the overall skin environment [10,11]. Undifferentiated keratinocytes are most active in the epidermis [12,13], and likely have a critical role in epidermal–mesenchymal interaction.
Fig. 1. Expressions of b1 integrin, K5 and K10 in hypertrophic scar and normal skin. In normal skin, expressions of b1 integrin (A) and K5 (B) were restricted to cells over connective tissue, expression of K10 could be observed in all keratinocytes except basal cells (C). The expression pattern in hypertrophic scar was extremely different from this. In hypertrophic scar, b1 integrin (D) and K5 (E) positive cells were barely detectable, but K10 expressed all over the epidermis (F). The scale bar represents 50 mm.
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
The absence or diminution of undifferentiated keratinocytes may lead to altered epidermal–mesenchymal interaction and initiate events leading to pathological fibrosis. To test this hypothesis, undifferentiated keratinocytes in normal skin and in hypertrophic scar tissue were compared, and their inhibition of extracellular matrix accumulation was proved using a tissue-engineered model of self-assembled reconstructed skin.
105
2. Materials and methods 2.1. Immunohistochemistry of cutaneous biopsies 30–40 days hypertrophic scar tissue with site-matched normal skin samples were collected from healthy individuals who gave oral and written informed consent. Five normal skin and five
Fig. 2. Construction of skin equivalents and histological analysis. The b1 integrin-positive undifferentiated keratinocytes (UKs) dispersed among keratinocytes before sorting (A). Magnetic cell sorting allowed the separation of UKs (B), which were absent from the remaining keratinocytes (C). H&E staining showed that the UK epidermis had the expected layers, with round basal cells arranged in good order (D). The skin equivalents constructed with DKs had a thin epidermis without typical layers; moreover, the epidermis was prone to separate from the dermis (arrows) (E). The USK epidermis did not have very distinct layers, although it contained obvious stratum corneum (F). The FB formed typical dermis which was used as a negative control (G). The scale bar represents 50 mm. Immunofluorescent staining showed that b1 integrin was expressed in the basal layer of UK epidermis. The positive cells arranged in good order in basal area (H); in DK epidermis, b1 integrin expression was still absent (I); nearly extinct b1 integrin expression could be observed occasionally in USK epidermis (white arrow) (J). The scale bar represents 25 mm.
106
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
Fig. 2. (Continued ).
hypertrophic scar specimens paraformaldehyde-fixed and routinely embedded were used. Paraffin sections (5 mm) were immunohistochemically stained using EnVision+1 system (DAKO, USA). An indirect method using peroxidase-labeled secondary antibodies was carried out, with reagent dilutions and incubation conditions selected according to manufacturer’s instructions. Briefly, the sections were deparaffinized and rehydrated in an ethanol series, recovered protein antigenicity in glycin/HCl buffer (50 mM, pH 3.5). Sections were incubated with monoclonal anti-b1 integrin (Oncogene, UK), monoclonal anti-K5 (DAKO, USA) and monoclonal anti-K10 (DAKO, USA) separately at 4 8C overnight. For controls, the primary antibody was omitted. The sections were then immunostained with horseradish peroxidase-conjugated goat anti-mouse IgGs (at a 1:50 dilution) for 1 h at 37 8C. The colorimetric reagent was diaminobenzidine/hydrogen peroxide and all sections were counterstained with hematoxylin and examined by light microscopy. 2.2. Cell culture Normal keratinocytes and dermal fibroblasts were prepared from foreskin obtained from patients aged 6–18 years undergoing
circumcision. All samples were processed as described by Barrandon and Green [14] and Germain et al. [15]. Fibroblasts were cultured in DMEM supplemented with 10% fetal calf serum (FCS). Cells under passage 5 were used for experiments. Keratinocytes were cultured in Keratinocyte Serum-Free Medium (GIBCO, USA) supplemented with epidermal growth factor (0.15 ng/ml), and bovine pituitary extract (25 mg/ml). 1–2 passage keratinocytes were used for cell sorting. 2.3. Separation of undifferentiated keratinocytes Samples of cell preparations were labeled using an indirect technique, in which the cells were incubated with monoclonal mouse anti-b1 integrin IgG (Oncogene, UK; dilution 1:50) for 30 min at 4 8C. After two washes with PBS supplemented with 1% FCS, the cells at a final concentration of 107 cells/100 ml were labeled with goat anti-mouse IgG, separated magnetically for 10 min and the magnetic pellet saved for a positive selection. The remainder was saved as differentiated keratinocytes. Indirect immunocytochemical staining method was used for sorting evaluation. The sorted cells were grown on coverslips for 1 day and then fixed. After being treated with 0.3% Triton X-100, the coverslips were blocked with 5% normal goat serum, then were incubated with mouse anti-b1 integrin (1:50; Oncogene, UK)
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
107
Fig. 3. Electron micrograph of epidermal basal layer in skin equivalents. TEM analysis showed that desmosomes (arrows) were seen in the basal layer of UK epidermis (A). Further enlargement revealed structural details in the marked area, showing that desmosomes were characterized by a much wider intermembrane space and denser distribution (arrows) (B). TEM micrographs showed an absence of cell–cell junctions in the DK epidermis (C), and enlarged spaces between cells (arrows) (D). Desmosomes could be seen in the USK epidermis, but were sparse (arrows) (E and F). The scale bar represents 2 mm. Quantitative analysis showed that the number of desmosomes in the UK epidermis were increased >5-fold compared to that in the USK epidermis (G). The results are given as the mean S.E.M. (n = 3).
108
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
Fig. 4. Ultrastructure of dermal ECM and expressions of collagen, bFGF and TGFb1 in different skin equivalents. TEM showed that in the presence of UKs, microfibrils were scattered without directionality, and collagenous bundles were not seen (A). In contrast, when DKs were added onto dermal sheets, the newly synthesized ECM formed collagen fiber bundles with longitudinal orientation (arrows) (B). In the presence of USKs, microfibril bundles were sparse, short and orientated randomly (arrows) (C). In the
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
antibodies at 4 8C overnight. At last, immunostained with horseradish peroxidase-conjugated goat anti-mouse IgGs (at a 1:50 dilution). All coverslips were counterstained with hematoxylin and examined by light microscopy. 2.4. Production of skin equivalents Dermal fibroblasts were grown for 35 days in the presence of 50 mg/ml ascorbic acid (GIBCO, USA) to form a dermal sheet as described [16]. Four layers were superimposed and cultured for 1 week to form reconstructed dermis. Undifferentiated (UK), differentiated (DK) and unsorted (USK) keratinocytes were seeded onto the dermal sheets at a density of 2 105 cells/cm2. Reconstructed dermis without the addition of keratinocytes (FB) served as a negative control. Keratinocytes reached confluence after 8 days of submerged culture and were then raised at the air– liquid interface. Samples were used after 2 weeks of air–liquid culture for analysis. Sections 5 mm thick were stained with hematoxylin and eosin (H&E) for histological observation. 2.5. Immunofluorescence staining
b1 integrin expression in skin equivalents was detected by immunofluorescence staining. Sections were deparaffinized in xylene, hydrated sections through graded alcohols. Incubated slides for 20 min in 10% goat serum, and then incubated with mouse anti-b1 integrin overnight at 4 8C, FITC-conjugated secondary antibody for 30 min, washed three times with PBS and mounted aqueous coverslips. Images of staining were captured on a fluorescence microscope. 2.6. Transmission electron microscopy Samples were fixed with 0.1 M sodium cacodylate buffer containing 2.5% glutaraldehyde and post-fixed with osmium tetroxide (SIGMA, Canada). Tissues were stained with 1% uranyl acetate, and examined with a JEM-100SX transmission electron microscope. Estimation of desmosome frequency was done as described [17]. Micrographs were taken at magnifications of 4000 and 15,000. The percentage of cell membrane length occupied by desmosomes was calculated for different micrographs. The significance of differences in desmosome frequency between different skin equivalents was assessed using Student’s t-test, with P-values of <0.05 considered significant. 2.7. RNA extraction, reverse transcription (RT), and PCR analysis Total RNA was isolated from skin equivalents using TRIzol1 reagent (Invitrogen, USA) according to the manufacturer’s protocol. PCR was performed in a volume of 25 ml and the primers were as follows: the human a1(I) chain gene for collagen (forward) TGTTCAGCTTTGTGGACCTC, (reverse) CTTGGTCTCGTCACAGATCA; the human bFGF gene (forward) CCTCACATCAAGCTACAACT, (reverse) TCAGCTCTTAGCAGACATTG; the human TGFb1 gene (forward) ACCTGCAAGACTATCGACAT, (reverse) GTACTCTGCTTGAACTTGTC; and the human b-actin gene (forward) ATCATGTTTGAGACCTTCAA, (reverse) CATCTCTTGCTCGAAGTCCA. After PCR, the products were subjected to electrophoresis in a 1.5% agarose gel and semi-quantified by comparing the intensity of
109
the band of interest to that of an internal standard (b-actin) using LabWorks Image Acquisition and Analysis Software (UVP, USA). The differences in mRNA expression between skin equivalents were determined and compared. The statistical significance was evaluated using Student’s t-test, with P-values of <0.05 considered significant. 2.8. Western blot analysis Western blot analysis was done with antibodies specific for procollagen type I (1:2000, polyclone, Maixin, China), bFGF (1:1000, monoclone, Santa Cruz, USA), and TGFb1 (1:2000, polyclone, Maixin, China). Extracted proteins were separated by SDS-PAGE (10% polyacrylamide gel), and the separated proteins were transferred to nitrocellulose membranes (Millipore, USA). Membranes were incubated with primary antibodies followed by incubation with horseradish peroxidase-conjugated secondary antibodies. Bound antibodies were detected by enhanced chemiluminescence (ECL plus kit, UK). A human antibody specific for b-actin (1:1000 42 kDa, Maixin, China) was used as a loading control. 3. Results 3.1. b1 integrin, K5, K10 expression in hypertrophic scar tissue Immunohistochemical staining showed that b1 integrin and Keratin 5 (K5) antibodies stained only the basal cell layer of normal skin (Fig. 1A and B). A consistent staining could be found in suprabasal keratinocytes with K10 antibody but not in the basal cells (Fig. 1C). By contrast, in hypertrophic scar there was a strong expression of K10 all over the epidermis (Fig. 1F), b1 integrin and K5 expressing keratinocytes were nearly undetectable (Fig. 1D and E). 3.2. Cell sorting evaluation and construction of skin equivalents Immunocytochemical staining showed that UKs expressing b1 integrin scattered in keratinocytes before sorting (Fig. 2A). Using magnetic beads, UKs were separated efficiently (Fig. 2B). In remainder cells b1 integrin expression was hardly detected, they were saved as differentiated keratinocytes (Fig. 2C). Skin equivalents were constructed with different keratinocyte population, H&E staining showed that UKs formed maturated epidermis with typical morphological character (Fig. 2D), whereas DKs formed very thin epidermis which was easy to separate from connective tissue (Fig. 2E). In USK epidermis, although there was obvious stratum corneum, the layers were not as typical as those in the UK epidermis (Fig. 2F). In negative group, FB formed typical dermis (Fig. 2G). Immunofluorescent staining revealed that in UK epidermis only the basal keratinocytes situated over the connective tissue still keep the ability to express b1 integrin (Fig. 2H). Few of b1 integrin strongly expressing keratinocytes were detected in USK epidermis (Fig. 2J). 3.3. The formation of desmosome and dermal matrix morphology in skin equivalents Transmission electron microscopy demonstrated that basal cells in the UK epidermis were connected tightly through desmosomes
FB dermis, microfibrils were arranged much more densely with many visible collagen clusters (arrows) around active fibroblasts (D). The scale bar represents 2 mm. Expression of collagen, bFGF and TGFb1 mRNA in different skin equivalents was detected by RT-PCR (E) and semi-quantified by image analysis (F). The results are given as mean S.E.M. (n = 5). Protein lysates from different skin equivalents were used for Western blot analysis to detect protein levels of collagen, bFGF and TGFb1 (G). Semiquantification was done with image analysis (H). The results are expressed as mean S.E.M. (n = 5). Significant results vs. FB dermis are shown as *, significant results vs. DK skin equivalent are shown as # and significant results vs. USK skin equivalent are shown as ~.
110
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
(Fig. 3A and B). The frequency (Fig. 3G) and morphological appearance of desmosomes were similar in normal skin. However, very few desmosomes were found in the DK epidermis (Fig. 3C); the keratinocytes did not show tight cell-to-cell connection and were prone to separate from each other (Fig. 3D). Desmosomes were found in the USK epidermis (Fig. 3E and F), but fewer than in the UK epidermis. Statistical analysis of desmosome frequency indicated a significant difference between different skin equivalents (Fig. 3G). When UKs were added onto dermal sheets, the dermal ECM was well distributed, microfibrils were deposited without directionality, and bundles were not seen (Fig. 4A). When USKs were added onto dermal sheets, the ECM showed collagenous fibers that were short and orientated randomly (Fig. 4C). The presence of DKs or the absence of keratinocytes led to a greater production of collagen in the dermis. Tight bundles of microfibrils were formed with a longitudinal orientation (Fig. 4B and D). Especially in the FB dermis, microfibrillar material was more abundant and dense around active fibroblasts, with a disordered appearance (Fig. 4D). 3.4. Undifferentiated keratinocytes reduce collagen expression in skin equivalents Expression of collagen is essential for dermal fibrosis, in which this protein is deposited in excess amounts [18]. This study showed that when DKs were added onto dermal sheets, the levels of collagen mRNA and protein were very high (0.35 0.01 and 1.50 0.23, respectively), but were downregulated markedly when the dermal sheets were seeded with UKs (Fig. 4E–H). The level of mRNA expression declined to 46% of that in the presence of DKs (0.16 0.03 vs. 0.35 0.01), and the level of protein expression declined to 20% (0.32 0.09 vs. 1.50 0.23). When USKs were added, the levels of mRNA and protein for collagen were between those of the UK and DK groups (0.20 0.01 and 1.28 0.19, respectively). Consistent with previous studies [3], the FB dermis (no addition of keratinocytes) expressed the highest level of mRNA and protein of collagen (0.45 0.02 and 1.89 0.28, respectively). 3.5. bFGF and TGFb1 mRNA, protein levels were altered by undifferentiated keratinocytes The profile of TGFb1 and bFGF mRNA expression in different skin equivalents is shown in Fig. 4E and F. The addition of UKs led to the highest level of bFGF mRNA (0.72 0.01). However, replacing UKs with DKs resulted in a significant reduction (0.50 0.05). The level of bFGF mRNA in the USK group (0.53 0.07) was between those of the UK and DK groups, and the lowest level of bFGF mRNA was found in the FB dermis. TGFb1 mRNA was barely detectable in the presence of UKs. However, the addition of DKs resulted in an almost >5-fold increase (0.10 0.03 vs. 0.02 0.01) in the production of TGFb1 mRNA. The level of TGFb1 mRNA in the USK group was between those of the UK and DK groups, and the FB dermis expressed the highest level of TGFb1 mRNA. Western blot analysis revealed that the pattern of bFGF and TGFb1 protein expression (Fig. 4G and H) was consistent with that of mRNA expression. When UKs were added onto dermal sheets, the highest level of bFGF was observed and no TGFb1 was detectable. In the presence of DKs, the level of bFGF was relatively low (0.45 0.15). Moreover, the expression of bFGF and TGFb1 in the FB dermis and in the USK group was consistent with that of mRNAs. 4. Discussion
b1 integrin is a cell surface molecule, via which basal keratinocytes attach to the underlying basement [19], and its
expression is downregulated in keratinocytes that have initiated the differentiation program. So b1 integrin is considered a marker of ESCs and TAs. Keratins are serial proteins, which make internal skeleton of keratinocytes. Keratinocytes express different keratins during differentiation. K5 is a protein expressed by basal keratinocytes in normal skin. Once keratinocytes stop dividing and detach from the basement membrane, switch from the expression of K5 to K10, marker of keratinocyte differentiation [20]. In this study, we observed that b1 integrin and keratin 5 expression was restricted to cells over connective tissue in normal skin, and were barely detectable in hypertrophic scar tissue. However, K10 not expressed in basal layer in normal skin were strongly expressed throughout hypertrophic scar epidermis, which suggested that differentiation stage of keratinocytes in hypertrophic scar differs from normal skin, there was no typical undifferentiated keratinocytes in basal layer of hypertrophic scar. Earlier studies showed that keratinocytes at the wound margin are activated when skin is injured [21,22]. Surrounding keratinocytes migrate and divide more frequently to restore the injured ‘‘barrier’’. It has been shown that keratinocytes in hypertrophic scar epidermis enter an alternative differentiation pathway [23] and this, coupled with unbalanced renewal, leads to the loss of undifferentiated keratinocytes. We therefore, hypothesized that absence or diminution of undifferentiated keratinocytes may further lead to altered epidermal–mesenchymal interaction and take part in initiating events leading to pathological fibrosis. Reconstructed human skin is a three-dimensional model made from skin epithelial cells and dermal fibroblasts, which can be used to investigate epidermal–dermal interaction. A reconstructed human skin model similar to in vivo skin was used to evaluate the role of undifferentiated keratinocytes in epidermal–dermal interaction and the effects they produce on extracellular matrix accumulation [16]. Undifferentiated keratinocytes isolated by an immunomagnetic method and added onto the self-assembled dermis directly formed the basis for this study. H&E staining showed that it was easier to construct well-organized skin equivalents with undifferentiated keratinocytes compared with differentiated or unsorted keratinocytes. This is in line with the facts that ESC and TA cells have the ability to form fully differentiated epidermis both in vivo and in vitro [24]. Immunofluorescence results showed that the cells located in basal region of UK epidermis still could express b1 integrin after long-term organotypic culture, but not in USK epidermis, two possibilities are: (1) in the condition of air–liquid culture undifferentiated keratinocytes differentiate into mature keratinocytes at certain ratio, which results in b1 integrin positive keratinocytes remained in USK epidermis much less than UK epidermis and (2) keratinocyte differentiation is controlled by particular microenvironments [25,26]. USKs, a mix of keratinocytes in viable degrees of differentiation provide an environment; in which undifferentiated keratinocytes mature easier. Results showed that UKs inhibited collagen expression at either the mRNA level or the protein level more efficiently than other populations, suggesting that undifferentiated keratinocytes may have a significant role in pathologic collagen matrix remodeling, which is compatible with the results reported by Hohelfeld, who applied expanded fetal skin constructs with a high proportion of undifferentiated keratinocytes to eight patients and showed that the wound healed with little hypertrophy of new skin and no retraction [27]. TGFb1 is an important fibrosis-related cytokine that stimulates the deposition of extracellular matrix components and degradation of the blocking matrix [28,29]. Earlier studies showed decreased expression of collagen in bFGF-treated cells in culture [30,31]. Here, we showed that undifferentiated keratinocytes
X. Wang et al. / Journal of Dermatological Science 53 (2009) 103–111
promoted bFGF expression but suppressed the expression of TGFb1 in contrast to other keratinocytes, which may contribute to a decreased and well-distributed collagen secretion. Our experiments highlight the roles of undifferentiated keratinocytes in dermal fibrosis inhibition. They may release signals that regulate expression of factors associated with matrix remodeling, such as bFGF and TGFb1. Therefore, the lack of undifferentiated keratinocytes leads to failure of the regulation, resulting in pathological accumulation of collagen and dermal fibrosis. Interestingly, we found that cells in the UK epidermis were prone to form cell–cell desmosomes that connected keratinocytes in the nearby basal membrane very tightly. In the DK epidermis, however, it was difficult to see the cell–cell contact. Earlier studies provided strong support for the importance of contact between epithelial cells, and showed that intracellular contacts are active regulators of fibrosis [27,32]. Accordingly, cell contact integrity regulates the extracellular matrix, and inhibits organ fibrosis, which is in line with previous results and further supports the hypothesis. In this study, we examined the role of undifferentiated keratinocytes in epidermal–dermal interaction and fibroblast matrix accumulation, and identified the absence of undifferentiated keratinocytes as one of the key factors leading to pathological fibrosis. These data identify a novel regulatory mechanism for scar formation. Undifferentiated keratinocytes may be a promising option for hypertrophic scar prevention and an alternative target for anti-therapies. Acknowledgments The study was supported by National Natural Science Foundation of China (grant number: 30700870, 30700770) and the NHTRD Program of China (863, grant number: 2005AA205241). References [1] Ghahary A, Shen YJ, Scott PG, Gong Y. Enhanced expression of mRNA for transforming growth factor-beta, type I and type III procollagen in human post-burn hypertrophic scar tissues. J Lab Clin Med 1993;122:465–73. [2] Deitch EA, Wheelahan TM, Rose MP, Clothier J, Cotter J. Hypertrophic burn scars: analysis of variables. J Trauma 1983;23:895–8. [3] Bellemare J, Roberge CJ, Bergeron D, Lopez-Valle CA, Roy M, Moulin VJ. Epidermis promotes dermal fibrosis: role in the pathogenesis of hypertrophic scars. J Pathol 2005;206:1–8. [4] Hakvoort TE, Altun V, Ramrattan RS, Kwast TH, Benner R, Zuijlen PP, et al. Epidermal participation in post-burn hypertrophic scar development. Virchows Arch 1999;434:221–6. [5] Garner WL. Epidermal regulation of dermal fibroblast activity. Plast Reconstr Surg 1998;102:135–9. [6] Barrandon Y, Green H. Cell migration is essential for sustained growth of keratinocyte colonies: the roles of transforming growth factor-alpha and epidermal growth factor. Cell 1987;50:1131–7. [7] Santoro MM, Gaudino G. Cellular and molecular facets of keratinocyte reepithelization during wound healing. Exp Cell Res 2005;304:274–86.
111
[8] Cotsarelis G, Cheng SZ, Dong G, Sun TT, Lavker RM. Existence of slow-cycling limbal epithelial basal cells that can be preferentially stimulated to proliferate: implications on epithelial stem cells. Cell 1989;57:201–9. [9] Hall PA, Watt FM. Stem cells: the generation and maintenance of cellular diversity. Development 1989;106:619–33. [10] Werner S, Smola H, Liao X, Longaker MT, Krieq T, Hofschneider PH, et al. The function of KGF in morphogenesis of epithelium and reepithelialization of wounds. Science 1994;266:819–22. [11] Abraham JA, Klagsburn M. Modulation of wound repair by members of the fibroblast growth factor family. In: Clark RAF, editor. The molecular and cellular biology of wound repair. 2nd ed., New York: Plenum Press; 1996. p. 195–248. [12] Maria IM, Marjana TC. Epidermal stem cells: the cradle of epidermal determination differentiation and wound healing. Biol Cell 2005;97:173–83. [13] Tomic-Canic M, Komine M, Freedberg IM, Blumenberg M. Epidermal signal transduction and transcription factor activation in activated keratinocytes. J Dermatol Sci 1998;17:167–81. [14] Barrandon Y, Green H. Cell size as a determinant of the clone-forming ability of human keratinocytes. Proc Natl Acad Sci USA 1985;82:5390–4. [15] Germain L, Guignard R, Rouabhia M, Auger FA. Early basement membrane formation following the grafting of cultured epidermal sheets detached with thermolysin or Dispase. Burns 1995;21:175–80. [16] Michel M, L’Heureux N, Pouliot R, Xu W, Auger FA, Germain L. Characterization of a new tissue-engineered human skin equivalent with hair. In Vitro Cell Dev Biol Anim 1999;35:318–26. [17] Skerrow CJ, Clelland DG, Skerrow D. Changes to desmosomal antigens and lectin-binding sites during differentiation in normal human epidermis: a quantitative ultrastructural study. J Cell Sci 1989;92(Pt 4):667–77. [18] Clark RA. Biology of dermal wound repair. Am J Med Sci 1993;306:42–8. [19] Jones PH, Watt FM. Separation of human epidermal stem cells from transit amplifying cells on the basis of differences in integrin function and expression. Cell 1993;73:713–24. [20] Rao KS, Babu KK, Gupta PD. Keratins and skin disorders. Cell Biol Int 1996;20:261–74. [21] Grinnell F. The activated keratinocytes: up regulation of cell adhesion and migration during wound healing. J Trauma 1990;30:144–9. [22] Galvin S, loomis C, Manabe M, Dhouailly D, Sun TT. The major pathways of keratinocyte differentiation as defined by keratin expression: an overview. Adv Dermatol 1989;4:277–99. [23] Machesney M, Tidman N, Waseem A, Kirby L, Leigh I. Activated keratinocytes in the epidermis of hypertrophic scars. Am J Pathol 1998;152:1133–41. [24] Li A, Pouliot N, Redvers R, Kaur P. Extensive tissue-regenerative capacity of neonatal human keratinocytes stem cells and their progeny. J Clin Invest 2004;113:390–400. [25] Moore KA, Lemischka IR. Stem cells and their niches. Science 2006;311:1880– 5. [26] David TS. The stem cell niche as an entity of action. Nature 2006;441:1075– 9. [27] Hohlfeld J, de Buys Roessingh A, Hirt-Burri N, Chaubert P, Gerber A, Scaletta C, et al. Tissue engineered fetal skin constructs for paediatric burns. Lancet 2005;366:740–842. [28] Border WA, Ruoslahti E. Transforming growth factor-beta in disease: the dark side of tissue repair. J Clin Invest 1992;90:1–7. [29] Kenyon NJ, Ward RW, McGrew G, Last JA. TGF-beta1 causes airway fibrosis and increased collagen I and III mRNA in mice. Thorax 2003;58:772–7. [30] Kennedy SH, Qin H, Lin L, Tan EM. Basic fibroblast growth factor regulates type I collagen and collagenase gene expression in human smooth muscle cells. Am J Pathol 1995;146:764–71. [31] Kypreos KE, Nugent MA, Sonenshein GE. Basic fibroblast growth factorinduced decrease in type I collagen gene transcription is mediated by Bmyb. Cell Growth Differ 1998;9:723–30. [32] Fan L, Sebe A, Peterfi Z, Masszi A, Thirone AC, Rotstein OD, et al. Cell contactdependent regulation of epithelial-myofibroblast transition via the rho–rho kinase-phospho-myosin pathway. Mol Biol Cell 2007;18:1083–97.