Positional differences in the wound transcriptome of skin and oral mucosa
© Chen et al; licensee BioMed Central Ltd. 2010
Received: 2 February 2010
Accepted: 12 August 2010
Published: 12 August 2010
When compared to skin, oral mucosal wounds heal rapidly and with reduced scar formation. Recent studies suggest that intrinsic differences in inflammation, growth factor production, levels of stem cells, and cellular proliferation capacity may underlie the exceptional healing that occurs in oral mucosa. The current study was designed to compare the transcriptomes of oral mucosal and skin wounds in order to identify critical differences in the healing response at these two sites using an unbiased approach.
Using microarray analysis, we explored the differences in gene expression in skin and oral mucosal wound healing in a murine model of paired equivalent sized wounds. Samples were examined from days 0 to 10 and spanned all stages of the wound healing process. Using unwounded matched tissue as a control, filtering identified 1,479 probe sets in skin wounds yet only 502 probe sets in mucosal wounds that were significantly differentially expressed over time. Clusters of genes that showed similar patterns of expression were also identified in each wound type. Analysis of functionally related gene expression demonstrated dramatically different reactions to injury between skin and mucosal wounds. To explore whether site-specific differences might be derived from intrinsic differences in cellular responses at each site, we compared the response of isolated epithelial cells from skin and oral mucosa to a defined in vitro stimulus. When cytokine levels were measured, epithelial cells from skin produced significantly higher amounts of proinflammatory cytokines than cells from oral mucosa.
The results provide the first detailed molecular profile of the site-specific differences in the genetic response to injury in mucosa and skin, and suggest the divergent reactions to injury may derive from intrinsic differences in the cellular responses at each site.
Wound healing is a complicated pathophysiological process orchestrated by a variety of known and unknown factors. Although cutaneous and mucosal wound healing proceed through the same stages of hemostasis, inflammation, proliferation, and remodeling, mucosal wounds demonstrate accelerated healing compared to cutaneous wounds [1–4]. Mucosal wounds also generally heal with minimal scar formation, and hypertrophic scars are rare in the oral cavity .
Studies in at least three different models of oral mucosal wound healing now support the concept that rapid wound closure and reduced scar formation are near-universal features of the superior healing phenotype that is observed in the oral cavity [2, 5–7]. The one exception that has been seen is excisional wounds placed on the hard palate of the mouse. In this model, the underlying connective tissue is extremely thin, so the wound depth reaches the periosteal bony surface and healing is slow . Nearly all other oral mucosal wounds, including palatal wounds in humans and pigs, heal more quickly than skin [5, 6].
Re-epithelialization at 24 h
MPO (U/mg protein) at 24 h
2.82 ± 0.2
1.26 ± 0.2
Mϕ (HPF) at 72h
10.8 ± 0.3
6.7 ± 0.4
T cells (HPF) at day 7
15.1 ± 1.5
5.6 ± 0.3
Vessel density compared to normal tissue at day 5 (fold increase)
No marked change
Collagen fibril diameter in wound
No marked change
Several previous studies have examined the transcriptome in wounds in a more limited fashion than the current study. Microarray analysis has been used to determine the changes in the transcriptome at the edges of acute wounds in murine skin , in laser captured blood vessels from human chronic wounds , and in non-healing human venous ulcers . Other studies have compared gene expression in wounds produced at early and later gestational ages  and in prenatal and postnatal wounds . These studies have contributed to our understanding of the wound transcriptome and the complexity of the repair process. To date, a global and comprehensive profiling of the differentially expressed genes in normal cutaneous and mucosal wounds has not been reported. The purpose of this study was to utilize microarray analysis to discover differences in the repair processes of oral and cutaneous wounds. A systematic profiling of gene expression in matched, equivalent sized cutaneous and oral mucosal wounds was performed at seven time points from 6 hours to 10 days post wounding. Microarray analysis of gene expression in normal, unwounded tissues was also performed. Our results suggest that tissue repair in oral mucosal wounds involves a more rapid, yet more refined response than that of skin wounds. The global expression patterns show that a significantly fewer number of gene sets change over time in mucosal wounds than in skin wounds. These findings demonstrate that tissue repair has distinguishable and different genomic expression patterns in mucosal and cutaneous sites.
More gene sets are differentially expressed in the skin wounds than in mucosal wounds over time
Differentially expressed clusters of gene sets identified in skin and tongue wounds
Patterns of K-means cluster distribution
Skin only (#)
Early up-regulation (high)*
Early up-regulation (medium)*
64.0 (cluster 1, 2, & 3 combined)
Early up-regulation (low)*
Tongue only (#)
Early up-regulation (high)*
Early up-regulation (medium)*
66.5 (cluster 1, 2, & 3 combined)
Early up-regulation (low)*
K-means cluster analysis identified a few patterns of transcriptional changes in healing tissues (Fig. 3). In skin, the largest group of transcripts (k-means clusters 1, 2, and 3 comprised of 95, 264, and 588 gene sets respectively) undergoes rapid early up-regulation, plateaus at 12-24 hours and rapidly subsides by day 3 and more gradually thereafter. The difference between the three clusters lays mainly in the degree of up-regulation (from low to medium to high). The whole group represented by these three clusters is designated as early upregulated (Fig. 3A and Table 2). Another group of transcripts (k-means cluster 4 comprised of 331 gene sets) undergoes early down-regulation, reaches its lowest point at 12-24 hours and gradually returns to the initial expression level by day 10. This group was designated as skin early downregulated (Fig. 3A and Table 2). The expression of the last group of transcripts (k-means cluster 5 comprised of 201 gene sets) slightly decreases during the first 24 hours and then undergoes a low degree of up-regulation that peaks at day 3 and remains upregulated up to day 10. This group was designated as late upregulated (Fig. 3A and Table 2).
In tongue wounds, the largest group of transcripts (k-means clusters 1, 2, and 3 comprised of 27, 112, and 195 gene sets respectively) undergoes rapid early up-regulation, plateaus at 6-24 hours and subsides sharply by day 3 and gradually decreases thereafter. Similar to skin wounds, the difference between the three groups also lays mainly in the degree of up-regulation (from low to medium to high). The difference between skin and tongue wounds is that in tongue wounds up-regulation reaches its height at an earlier time point (6 hours in tongue vs. 12 hours in skin). This group, represented by three k-means clusters, was designated as early upregulated (Fig.3B and Table 2). Another group of transcripts (k-means cluster 4 comprised of 52 gene sets) is down regulated at 6 hours, and reaches its lowest point between 12 hours to day 5, then returns to initial expression level by day 7 after wounding. This group remains downregulated longer than the analogous group in skin wounds. This group was designated as early downregulated (Fig. 3B and Table 2). The last group of transcripts (k-means cluster 5 comprised of 116 gene sets) resembles the analogous group in skin wounds (k-means cluster 5) and is slightly down regulated in the first 24 hours after wounding, then upregulates at day 3 and remains elevated up to day 10. This group was designated as late upregulated (Fig. 3B and Table 2).
As described above, tongue wounds contained expression groups that were highly similar in pattern to those identified in skin. While the overall number of genes differed, the percentage of gene sets found in each expression group was quite similar in the two tissues (Table 2). In skin, 64% of all identified transcripts fell into the early up-regulated transcripts group, while 66.5% of the identified transcripts in tongue wounds were situated within the early up-regulated group. In addition, there were 22.4% and 13.6% of gene sets in the skin wounds, and 10.4% and 23.1% of gene sets in the tongue wounds that fell into the early down-regulated and late up-regulated transcripts group respectively (Table 2).
Functional motifs behind transcriptional changes
To uncover cellular processes likely affected by transcriptional changes, we applied gene enrichment and functional annotation analysis with the use of the web based application DAVID (Database for Annotation, Visualization and Integrated Discovery) at http://david.abcc.ncifcrf.gov.
Early up-regulated genes: enriched annotation terms
GO: Response to wounding
GO: Inflammatory response
KEGG: Toll-like receptor signaling pathway
GO: Receptor binding
KEGG: Cytokine-cytokine receptor interaction
KEGG: Jak-STAT signaling pathway
Early up-regulated genes: enriched groups of functionally related genes
Interferon α, β and δ/cytokine
16 (Group 1)
12 (Group 2)
10 (Group 1)
8 (Group 3)
Positive regulation of protein kinase activity
4 (Group 4)
Regulation of metabolic/biological process
4 (Group 5)
Pattern recognition receptors (TLR related)
7 (Group 6)
8 (Group 7)
Matrix metallopeptidase (MMP)
6 (Group 8)
Negative regulation of cell proliferation
4 (Group 2)
14 (Group 3)
4 (Group 4)
Functional Annotation of Early Upregulated Genes
We found that both types of wounds exhibited similar biological tendencies in early response (clusters 1, 2, and 3 from K means analysis) including those identified as 1) response to wounding, 2) inflammatory response, 3) cytokine, 4) cytokine-cytokine receptor interaction and 5) chemotaxis (Table 3). However, a number of group members is higher in skin, therefore the majority of adjusted p values for skin wounds were lower than that for tongue wounds (Table 3). In addition, analysis identified the cytokine-cytokine receptor interaction pathway as significantly expressed in both skin wounds and tongue wounds (Table 3 and additional file 1) with more numbers of cytokine/chemokines observed in skin wounds. Interestingly, the Toll-like receptor-signaling pathway (Table 3 and additional file 2) and Janus family tyrosine kinases-signal transducers and activators of transcription (JAK-STAT) signaling pathway (Table 3 and additional file 3) were only utilized in skin wounds. Furthermore, the DAVID gene functional classification analysis identified 27 groups in skin wounds but only 11 groups in tongue wounds. When 2.00 was used as the cutoff for enrichment scores, the number of groups found to be significantly enriched was reduced to 8 out of 27 in skin wounds (Table 4 and additional file 4) and just 4 out of 11 in tongue wounds (Table 4 and additional file 5).
The group of genes (cytokines) with the highest enrichment score (9.77) in skin wounds included interferon (IFN)-α, IFN-β, IL-23, IL-24, and CSF-3. None of these genes were found to be differentially expressed in tongue wounds (Additional file 4 and 5).
Chemokines were found to be prominently differentially expressed by wounds at both locations. Functional grouping gene demonstrated that Group 1 in tongue wounds and Group 2 in skin wounds had 10 and 12 chemokines respectively. Both wound types included CCL2, CCL4, CCL7, CXCL1, CXCL2, CXCL5, and CXCL11. However, CCL5, CCL12, and CXCL10 only appeared in tongue wounds, while CCL3, CCL20, CXCL3, CXCL7, and CXCL13 only appeared in skin wounds (Additional file 4 and 5).
One interesting set of genes that was found to be enriched in mucosa, but not in skin, was a group of several genes involved in the negative regulation of proliferation. More specifically, Group 2 in tongue wounds contained 4 such genes including IFN induced transmembrane protein 3, schlafen 1, schlafen 2, and schlafen 3 (Additional file 5). This group of genes was not identified as enriched as a group in skin wounds, although each was significantly expressed in skin wounds according to ANOVA.
Group 3 genes in skin wounds included myeloid cell leukemia sequence 1, pleckstrin homology-like domain, family a, member 1, lectin, galactose binding, soluble 7, egl nine homolog 3, B-cell leukemia/lymphoma 2 related protein a1a, B-cell leukemia/lymphoma 2 related protein a1 d, B-cell leukemia/lymphoma 2 related protein a1b, and axin1 up-regulated 1 in skin wounds are associated with cell differentiation and apoptosis (Additional file 4). Only one of these genes (lectin, galactose binding, soluble 7) was present in tongue wounds.
Group 3 in tongue wounds contains 14 gene members associated with GTP-binding activity (Additional file 5).
Group 4 in tongue wounds included small proline-rich proteins 2d, 2f, 2i, 2h and 2j which are associated with keratinization and epidermal development (Additional file 5). Small proline-rich protein 2a, 2d, and 2i were present in skin wounds, but not enriched as a group.
Group 4 in skin wounds contains a groups of 4 genes which are involved in positive regulation of protein kinase activity, including growth arrest and DNA-damage-inducible 45 β, sperm associated antigen 9, toll-interleukin 1 receptor (tir) domain-containing adaptor protein and arginine vasopressin-induced 1 (Additional file 4). None of these genes appeared as differentially expressed in tongue wounds.
Group 5 in skin wounds includes hematopoietic cell specific lyn substrate 1, cardiotrophin-like cytokine factor 1, CD80 antigen and yamaguchi sarcoma viral (v-yes-1) oncogene homolog. All of these genes play a positive role in regulation of metabolic/biological process (Additional file 4). None of these genes were present as differentially expressed in tongue wound.
Group 6 in skin wounds includes toll-like receptor (TLR) 2, TLR 4, TLR 6, TLR 13, c-type lectin domain family 7, member a, MyD88, and IL-1b. These molecules, with the exception of MyD88 and IL-1, are pattern recognition receptors for pathogens and/or endogenous ligands and play critical roles in innate immunity (Additional file 4). None of these genes appears as differentially expressed in tongue wound.
Group 7 in skin wounds includes ebna1 binding protein 2, bystin-like, tsr1/20 s rRNA accumulation/homolog (yeast), riken cDNA 4933411h20 gene, sda1 domain containing 1, nuclear import 7 homolog (S. cerevisiae), riken cDNA 5730405k23 gene, and riken cDNA 2610012o22 gene. These genes are all intracellular membrane-bound organelle and nucleus proteins that play roles in ribosome and organelle organization, biogenesis, and assembly (Additional file 4). None of these genes appear as differentially expressed in tongue wound.
Group 8 in skin wounds includes matrix metallopeptidase (MMP) 1a, MMP 1b, MMP 8t MMP 9, MMP 10, and MMP 13 (Additional file 4). None of these genes appears as differentially expressed in tongue wound.
Functional Annotation of Early Downregulated Genes
Early down-regulated genes: enriched annotation terms
GO: DNA binding
Early down-regulated genes: enriched groups of functionally related genes
Transcription regulation/DNA binding
52 (Group 1)
5 (Group 1)
Gene functional annotation clustering analysis of Cluster 4 genes suggests that the tongue wounds have similar gene clusters as the skin wounds but has much less numbers of members and higher p-values (Table 5). Only one functional group was identified in gene functional classification analysis; this group had an enrichment score of 1.25 and contained 5 transcripts annotated as regulators of transcription, and 4 as DNA binding (Table 6 and additional file 7). When the results of the functional classification analysis from skin and tongue were compared, four members of this group in tongue wounds except nuclear receptor subfamily 1, group d, member 1 are present in skin as well.
Functional Annotation of Late Upregulated Genes
Late up-regulated genes: enriched annotation terms
SP_PIR_KEYWORDS: Extracellular matrix
SP_PIR_KEYWORDS: Structural protein
GO: ECM-receptor interaction
KEGG: Cell communication
KEGG: Peptidase activity
Late up-regulated genes: enriched groups of functionally related genes
13 (Group 1)
9 (Group 1)
Serine-type endopeptidase activity
4 (Group 2)
4 (Group 3)
3 (Group 4)
8 (Group 5)
5 (Group 2)
Another functional group found in both types of wounds contains transcripts encoding elements of cytoskeleton (group 2 in tongue and group 5 in skin) (Table 8, additional file 8 and 9). These functional groups include actin, myosin, and troponin (Additional file 8 and 9). Skin wounds contain cardiac actin α, cardiac troponin t2, myosin/heavy polypeptide 3/skeletal muscle/embryonic, troponin i/skeletal/slow 1, troponin c/cardiac/slow skeletal, myosin/heavy polypeptide 13/skeletal muscle, kinesin family member 20a, and cDNA sequence bc056349. Similarly, tongue wounds were found to contain all of these genes except for the last three.
In skin wounds, functional Group 2 contains 4 peptidases including protease serine 19 (neuropsin), tryptase α/β1, transmembrane protease serine 11a, and protease serine 35 (Additional file 8). None of the members are present in tongue wounds. Group 4 in skin wounds also contained peptidases, including matrix metallopeptidase 23, a disintegrin/metallopeptidase domain 12, and a disintegrin-like and metallopeptidase with thrombospondin type 1 motif, 16 (Additional file 8). None of these genes were present in tongue wounds.
Functional Group 3 in the skin contains 4 calcium binding proteins including hypothetical protein 9430004m15, CD248, cadherin 11, and thrombospondin 4 (Additional file 8). CD248 and thrombospondin 4 are also involved in adhesion. Tongue wounds contained only the first two genes in this group.
Skin and mucosal keratinocytes respond differently upon IL-1β stimulation
Validation of microarray Data
Real time RT-PCR validation (Ratio of target/Glucuronidase β)
12 h or 24h
12 h or 24h
This is the first systemic, comprehensive and dynamic study of gene expression profiles in skin and mucosal wounds over all stages of wound healing. Using well established mouse models of skin and mucosal wound healing [1, 2, 7] and advanced microarray technology, similarly expressed as well as significantly differentially expressed genes in skin and mucosal wounds were successfully identified. Overall, the identification of five clusters of genes showing similar patterns of expression allowed for a general comparison of the global genomic response to injury in skin and tongue. This comparison suggests that the patterns of expression are similar in both types of wounds, but hardly identical. The results clearly show that the comprehensive genomic response to injury in the tongue is more rapid, shorter in duration, and of lesser intensity than the response of skin to a similar sized insult. This data implies that, as compared to skin, the tongue has an intrinsic genetic response that accelerates repair. One possible reason for the apparent increase in the intensity of the response to injury in skin as compared to tongue could be that baseline expression levels are simply higher in tongue for many genes. In this case, the mucosa, being "preactivated" would not require an increase in the expression of genes during the healing process. To examine this, we compared the baseline levels of several genes that are highly upregulated in the early skin wounds but not tongue wounds. Because inflammation is quite different at the two sites, we focused on genes from the cytokines/chemokines group, including IFN-α, IFN-β, IL-23, IL-24, CSF-3, CCL3, CCL20, CXCL3, CXCL7, and CXCL13. This analysis showed that the baseline levels of nine of these genes are very similar in normal mucosa and skin. The one exception was CXCL13, which is indeed higher at baseline in mucosal sites (Additional file 10). Therefore, any baseline differences can only partially explain the differential healing responses of skin and mucosa.
The comprehensive analysis of the transcriptome of skin and mucosal wounds allowed us to identify several specific groups of genes that are disparately expressed in the skin and mucosal wounds and that may be functionally important to healing outcomes. Skin and mucosal wounds demonstrated notable differences in the expression of proinflammatory and inflammation related elements. Mucosal wounds exhibited decreased levels of pro-inflammatory genes including cytokines and chemokines. Cytokine and chemokine induced inflammation in the early stage of wound healing is critical for tissue reconstruction and regeneration in the later stages, as several studies show that an appropriate inflammatory response in the early stage is critical for wound healing [15–18]. Cytokines and chemokines play important roles in the inflammatory response including recruiting and activating inflammatory cells such as neutrophils, macrophages, T cells, and mast cells, which are all involved in the wound healing process [15–18]. Among the early response genes, cytokines including IL-23, IL-24, CSF, IFN-α, and IFN-β and chemokines including CCL3, CCL20, CXCL3, CXCL7, and CXCL13 are significantly expressed in skin wounds but not in mucosal wounds. This data suggests that decreased inflammatory cytokine and/or chemokine levels may support rapid healing and reduced scar formation in the oral mucosal wounds. Another inflammatory pathway that showed tissue specific differences was the JAK-STAT pathway. The JAK-STAT signaling pathway is one of the important pathways to drive biological responses to cytokines . Since there are numbers of cytokines significantly upregulated in the skin wounds, it is not surprising that the JAK-STAT signaling pathway is involved in the inflammatory response to wounding seen in the skin, but not in mucosa (Additional file 3).
Interestingly, the TLRs were differentially expressed in skin and mucosal wounds. TLRs are a group of transmembrane receptors that traditionally are thought only to recognize structurally conserved molecules derived from various organisms including bacteria, viruses, fungi, and parasites [20–22]. When such threats breach physical barriers such as the skin or mucosa, they bind and activate TLRs on immune cells and induce the innate immune response [20, 22, 23]. However, recent studies strongly suggest that TLRs also recognize endogenous ligands such as heat shock proteins, hyaluronan, fibronectin, and fibrinogen when these molecules are released from damaged tissues or cells resulting from wounding, ischemia, or other injuries independent of the presence of microbes [24, 25]. DAVID gene analysis tools indicated that the TLR signaling pathway is involved in healing. TLR2, TLR4, TLR6, and TLR13 are highly expressed and enriched as a significant group in skin wounds, but not in tongue wounds. In a recent study, migration of human epidermal keratinocytes, in in vitro scratch wounds, was inhibited by LPS and the inhibition could be blocked partially by anti-TLR4 (75%) and anti-TLR2 (40%) . The emerging roles of TLRs in wound healing has received little experimental attention to date except in chemical or thermal-induced burn injuries [27–31]. For example, TLR4 plays a critical role in microvascular leakage and leukocyte adhesion under the inflammatory conditions associated with nonseptic thermal injury . Severe burn also primes the innate immune system for enhanced TLR2 and TLR4-mediated response which may contribute to the development of heightened systemic inflammation . TLRs have also been demonstrated changes in various skin diseases such as acne, leprosy, psoriasis, and Lyme disease . The importance of TLRs in the injury response is under active investigation by many laboratories.
Beyond the inflammation related elements, several additional pathways and groups of genes that are known to be important to injury responses were differentially expressed in skin and mucosal wounds. The MMPs, a family of zinc-dependent endopeptidases that play a critical role in wound healing process, also showed differential expression in skin and mucosal wounds. MMPs can regulate inflammation, and degrade components of ECM to facilitate the migration of cells. They are involved in tissue remodeling [33–35]. Upregulation of MMP 1a, MMP 1b, MMP 8, MMP 9, MMP 10 and MMP13 seen in early skin wounds, but not in oral mucosal wounds, suggests that cells in the skin reacts to wounding with a much more aggressive production of MMPs than do the cells in mucosa.
Another group of non-inflammatory related genes that were differentially expressed was the peptidases. A number of peptidase genes (Table 8 and additional file 8) that belong to the late upregulated gene cluster were found to only appear in skin wounds. This significantly more robust expression of peptidases in skin may suggest that tissue remodeling or reorganization processes are active in skin wounds at later time points than they are in tongue. This concept is consistent with clinical and pathological observations in these two types of wounds.
On the whole, the differences noted in the transcriptome of oral mucosal and skin wounds are striking, and raise questions as to whether some of the divergence might be due to anatomic variation itself. Oral mucosa and skin are both stratified epithelium, yet these tissues do exhibit multiple microscopic and anatomic differences, such as hair follicles and sweat glands (in skin), and taste buds (in mucosa). Certainly these various differences may account for some of the differential transcriptional events that are observed during the injury response at the two sites. However, studies in palate, buccal mucosa, and tongue models all demonstrate that oral mucosa, regardless of specific anatomical features, heals more rapidly than skin. Thus, the improved response to injury seen in oral mucosa seems to include elements that are not dependent upon adnexal structures.
In the context of our understanding of the wound healing response, it seems likely that many of the observed differences in the transcriptome of skin and mucosal wounds are derived from epithelium rather than connective tissue. An important role for the epithelial response in dictating wound healing outcomes has been suggested by numerous studies . At sites of injury, the epithelium is a rich source of both proinflammatory mediators, such as IL-1, and TNF-α, as well as growth factors, such as vascular endothelial growth factor (VEGF). Keratinocytes are also capable of modulating fibroblast behavior, including collagen synthesis, through the production and release of soluble factors [37, 38]. Given the importance of the epithelial response to healing outcomes, at least a portion of the wound healing phenotype probably reflects intrinsic differences in the epithelial response to injury at skin and mucosal sites. In support of this concept, we found that keratinocytes from skin and oral mucosa respond differently to equivalent IL-1β stimulation. Our experiments showed that, following identical stimulation in vitro, skin keratinocytes have an intrinsic capacity to express significantly more IL-6 and TNF-α than mucosal keratinocytes (p < 0.01, Fig. 5). This data strongly supports the concept that keratinocytes from skin and mucosa maintain differential regulatory pathways that lead to differential responsiveness at sites of injury. Such site-specific regulation may be responsible for the increased inflammation that is observed in skin versus mucosal wounds. Additional studies will be needed to isolate the contributions of epithelial and connective tissue responses to injury at the two sites.
Taken together, the current study provides a unique, comprehensive view of the different gene expression profiles in the process of skin and mucosal wound healing. The results strongly suggests that there is a fundamental difference in intrinsic genetic response to wounding between skin and mucosa, which makes mucosal wounds heal faster and with less inflammation and scar formation. The combination of the current findings with proteomic studies may permit the identification of genes or response elements that are responsible for superior mucosal wound healing and/or the more severe scar formation that is commonly seen in skin wounds.
Animals and wound models
Six- to 8-week-old female Balb/c mice (Harlan Inc., Indianapolis, IN) were anesthetized with ketamine (100 mg/kg) and xylazine (5 mg/kg). For dermal wounds the dorsal skin was shaved, wiped with isopropyl alcohol, and six full-thickness dermal wounds were placed on the opposite sides of the midline using a 1 mm punch biopsy instrument (Acu-Punch, Acuderm Inc., Ft. Lauderdale, FL) [1, 2, 7]. In a separate group of mice, oral mucosal wounds were placed on the lateral side of the tongue at an equal distance from the midline using the same 1 mm punch biopsy instrument [1, 2, 7]. At defined intervals (6 hours, 12 hours, 24 hours, day 3, day 5, day 7 and day 10) after injury, wounds and surrounding tissues were removed with a 2 mm biopsy punch and stored in RNAlater (Sigma, St. Louis, MO) until processing. Skin wounds include epidermis, dermis, subcutaneous, and muscle. Tongue wounds involve epithelia, connective tissue and muscle. The time points chosen (6, 12, 24 hours, 3, 5, 7, and 10 days after wounding) covered all these four stages. Six-12 hours, 24 hours-day 3, day 5-7, and day10 correspond to hemostasis, inflammation, proliferation, and remodeling respectively. All 6 skin wounds from each mouse were pooled and used as one sample in the following microarray and RT-PCR analyses. Each time point had 3 mice for either skin or tongue wound group. Additional skin or tongue wounds were made and samples were frozen in TBS tissue freezing medium (Triangle Biomedical Sciences, Durham, NC) for subsequent hematoxylin and eosin (H&E) staining to evaluate the reepithelialization as described in our previous publication  (n = 4). All animal procedures were approved by the University of Illinois Institutional Animal Care and Use Committee.
Total RNA preparation
Total RNA was extracted using TriZol (Invitrogen, Carlsbad, CA) and purified up by RNeasy kit (QIAGEN, Valencia, CA). The integrity (18S/28S) and concentrations of RNA was determined using an Experion (Bio-Rad, Hercules, CA) per the manufacturer's instruction.
Microarray data collection and data analysis
Total RNA from each sample was labeled and hybridized to Affymetrix GeneChip Mouse Genome 430 v 2.0 chip (Santa Clara, CA) containing 45,037 gene sets according to standard Affymetrix recommended protocols. Each hybridization image was inspected for the following quality metrics: total background, raw noise (Q), average signal present, signal intensity of species-specific house-keeping genes, 3'/5' signal ratio of house-keeping genes, relative signal intensities of labeling controls, absolute signal intensities of hybridization controls, and GCOS scale factors. All 48 hybridizations passed each quality criteria.
Hybridization data was analyzed in 'S-Plus' 6.2 statistical package (Insightful, Palo Alto, CA). Signal intensities were normalized and background was corrected according to Robust Multi-array Average (RMA) method . A one-way ANOVA test was applied to identify significant differentially expressed gene sets across all time points after wounding. Levels of gene expression at each time point in the wounded skin or wounded tongue were compared to normal skin or normal tongue tissue respectively. If any gene was identified as significant differentially expressed at one time point or more after injury, this gene was included for further analyses. Raw p-values were corrected for False Discovery Rate (FDR) according to the Benjamini-Hochberg (BH) procedure . Differentially expressed transcripts were annotated according to Affymetrix's 'NetAffx Analysis Center' (Annotation release R23). A subset of statistically significant transcripts was used for K means clustering in order to analyze patterns of expression changes across all time points. The following filtering criteria were applied: 1) absolute fold change > 2 in at least one time point in relation to normal tissue, and 2) FDR corrected p Value < 1E-5. The microarray data were deposited into the GEO repository and made public on July 20, 2010 with the accession number GSE23006.
Skin and mucosal keratinocyte isolation and cell culture
Human skin and oral mucosal (palate) tissues were obtained from healthy adult donors after consent under a protocol approved by Institutional Review Board at the University of Illinois at Chicago. The procedures for isolating keratinocyte from skin and palate tissues were the same as described previously . Keratinocytes from both skin and mucosa were plated in 12-well plates at a density of 4.5 × 104 cells/well in KBM-2 medium (Lonza, Basel, Switzerland) and incubated overnight. The following day, the cells were treated with human recombinant IL-1β (50 ng/ml) (PeproTech, Rocky Hills, NJ) or medium only for 24 hours. The cells were then harvested using Trizol followed by total RNA isolation and DNase I (Invitrogen) treatment. Three independent isolates of primary keratinocytes were obtained from normal human skin and oral mucosa. One isolate from skin and one isolate from mucosa were from the same individual and were prepared in our lab. Two additional cultures of primary mucosal keratinocytes were also isolated in our lab. The two additional isolates of skin keratinocyte were purchased from Lonza. Basel, Switzerland. Similar results were obtained from the skin keratinocytes isolated in our lab and purchased from Lonza. Representative pictures of skin and mucosal keratinocytes isolated in our lab are shown in (Additional file 11). From the morphology view, there were no contaminations from other cells.
Real-Time RT-PCR Verification of Affymetrix GeneChip Data
Primer sets for real time RT-PCR
Forward primers (5'--3')
Reverse primers (5'--3')
Real time PCR was also used to determine the mRNA expression of IL-6 and TNF-α by skin and mucosal keratinocytes after being treated by IL-1β. 2-∆∆Ct method was used to calculate fold changes of IL-6 and TNF-α mRNA expression after IL-1β treatment compared to medium treated cells. Student t test was used to determine the statistical difference between the fold changes in skin and mucosa. A p value less than 0.05 was considered statistically significant.
This publication was supported by NIH grant P20-GM078426 (LAD, PTM, and TAM). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIGMS or NIH.
- Szpaderska AM, Walsh CG, Steinberg MJ, DiPietro LA: Distinct patterns of angiogenesis in oral and skin wounds. J Dent Res. 2005, 84: 309-314. 10.1177/154405910508400403.PubMedView Article
- Szpaderska AM, Zuckerman JD, DiPietro LA: Differential injury responses in oral mucosal and cutaneous wounds. J Dent Res. 2003, 82: 621-626. 10.1177/154405910308200810.PubMedView Article
- Stephens P, Davies KJ, al-Khateeb T, Shepherd JP, Thomas DW: A comparison of the ability of intra-oral and extra-oral fibroblasts to stimulate extracellular matrix reorganization in a model of wound contraction. J Dent Res. 1996, 75: 1358-1364. 10.1177/00220345960750060601.PubMedView Article
- Warburton G, Nares S, Angelov N, Brahim JS, Dionne RA, Wahl SM: Transcriptional events in a clinical model of oral mucosal tissue injury and repair. Wound Repair Regen. 2005, 13: 19-26. 10.1111/j.1067-1927.2005.130104.x.PubMedView Article
- Wong JW, Gallant-Behm C, Wiebe C, Mak K, Hart DA, Larjava H, Hakkinen L: Wound healing in oral mucosa results in reduced scar formation as compared with skin: evidence from the red Duroc pig model and humans. Wound Repair Regen. 2009, 17: 717-729. 10.1111/j.1524-475X.2009.00531.x.PubMedView Article
- Mak K, Manji A, Gallant-Behm C, Wiebe C, Hart DA, Larjava H, Hakkinen L: Scarless healing of oral mucosa is characterized by faster resolution of inflammation and control of myofibroblast action compared to skin wounds in the red Duroc pig model. J Dermatol Sci. 2009, 56: 168-180. 10.1016/j.jdermsci.2009.09.005.PubMedView Article
- Schrementi ME, Ferreira AM, Zender C, DiPietro LA: Site-specific production of TGF-beta in oral mucosal and cutaneous wounds. Wound Repair Regen. 2008, 16: 80-86. 10.1111/j.1524-475X.2007.00320.x.PubMedView Article
- Graves DT, Nooh N, Gillen T, Davey M, Patel S, Cottrell D, Amar S: IL-1 plays a critical role in oral, but not dermal, wound healing. J Immunol. 2001, 167: 5316-5320.PubMedView Article
- Schrementi ME: Site-specific differences in wound healing. PhD thesis. 2006, 50-54.
- Roy S, Khanna S, Rink C, Biswas S, Sen CK: Characterization of the acute temporal changes in excisional murine cutaneous wound inflammation by screening of the wound-edge transcriptome. Physiol Genomics. 2008, 34: 162-184. 10.1152/physiolgenomics.00045.2008.PubMed CentralPubMedView Article
- Roy S, Patel D, Khanna S, Gordillo GM, Biswas S, Friedman A, Sen CK: Transcriptome-wide analysis of blood vessels laser captured from human skin and chronic wound-edge tissue. Proc Natl Acad Sci USA. 2007, 104: 14472-14477. 10.1073/pnas.0706793104.PubMed CentralPubMedView Article
- Charles CA, Tomic-Canic M, Vincek V, Nassiri M, Stojadinovic O, Eaglstein WH, Kirsner RS: A gene signature of nonhealing venous ulcers: Potential diagnostic markers. J Am Acad Dermatol. 2008, 59: 758-771. 10.1016/j.jaad.2008.07.018.PubMedView Article
- Chen W, Fu X, Ge S, Sun T, Zhou G, Han B, Li H, Sheng Z: Profiling of genes differentially expressed in a rat of early and later gestational ages with high-density oligonucleotide DNA array. Wound Repair Regen. 2007, 15: 147-155. 10.1111/j.1524-475X.2006.00195.x.PubMedView Article
- Colwell AS, Longaker MT, Peter Lorenz H: Identification of differentially regulated genes in fetal wounds during regenerative repair. Wound Repair Regen. 2008, 16: 450-459. 10.1111/j.1524-475X.2008.00383.x.PubMedView Article
- Eming SA, Krieg T, Davidson JM: Inflammation in wound repair: molecular and cellular mechanisms. J Invest Dermatol. 2007, 127: 514-525. 10.1038/sj.jid.5700701.PubMedView Article
- Szpaderska AM, DiPietro LA: Inflammation in surgical wound healing: friend or foe?. Surgery. 2005, 137: 571-573. 10.1016/j.surg.2005.01.006.PubMedView Article
- Martin P, Leibovich SJ: Inflammatory cells during wound repair: the good, the bad and the ugly. Trends Cell Biol. 2005, 15: 599-607. 10.1016/j.tcb.2005.09.002.PubMedView Article
- Singer AJ, Clark RA: Cutaneous wound healing. N Engl J Med. 1999, 341: 738-746. 10.1056/NEJM199909023411006.PubMedView Article
- Imada K, Leonard WJ: The Jak-STAT pathway. Mol Immunol. 2000, 37: 1-11. 10.1016/S0161-5890(00)00018-3.PubMedView Article
- Takeda K, Akira S: Toll-like receptors in innate immunity. Int Immunol. 2005, 17: 1-14. 10.1093/intimm/dxh186.PubMedView Article
- Uematsu S, Akira S: Toll-Like receptors (TLRs) and their ligands. Handb Exp Pharmacol. 2008, 183: 1-20. full_text.PubMedView Article
- Kawai T, Akira S: TLR signaling. Cell Death Differ. 2006, 13: 816-825. 10.1038/sj.cdd.4401850.PubMedView Article
- Uematsu S, Akira S: The role of Toll-like receptors in immune disorders. Expert Opin Biol Ther. 2006, 6: 203-214. 10.1517/14712522.214.171.124.PubMedView Article
- Mollen KP, Anand RJ, Tsung A, Prince JM, Levy RM, Billiar TR: Emerging paradigm: toll-like receptor 4-sentinel for the detection of tissue damage. Shock. 2006, 26: 430-437. 10.1097/01.shk.0000228797.41044.08.PubMedView Article
- Zhang Z, Schluesener HJ: Mammalian toll-like receptors: from endogenous ligands to tissue regeneration. Cell Mol Life Sci. 2006, 63: 2901-2907. 10.1007/s00018-006-6189-1.PubMedView Article
- Loryman C, Mansbridge J: Inhibition of keratinocyte migration by lipopolysaccharide. Wound Repair Regen. 2008, 16: 45-51. 10.1111/j.1524-475X.2007.00290.x.PubMedView Article
- Breslin JW, Wu MH, Guo M, Reynoso R, Yuan SY: Toll-like receptor 4 contributes to microvascular inflammation and barrier dysfunction in thermal injury. Shock. 2008, 29: 349-355.PubMed
- Paterson HM, Murphy TJ, Purcell EJ, Shelley O, Kriynovich SJ, Lien E, Mannick JA, Lederer JA: Injury primes the innate immune system for enhanced Toll-like receptor reactivity. J Immunol. 2003, 171: 1473-1483.PubMedView Article
- Price JA, Rogers JV, McDougal JN, Shaw MQ, Reid FM, Kiser RC, Graham JS: Gene expression analysis of bromine-induced burns in porcine skin. Toxicol Lett. 2008, 182: 69-78. 10.1016/j.toxlet.2008.08.007.PubMedView Article
- Rogers JV, McDougal JN, Price JA, Reid FM, Graham JS: Transcriptional responses associated with sulfur mustard and thermal burns in porcine skin. Cutan Ocul Toxicol. 2008, 27: 135-160. 10.1080/15569520802092054.PubMedView Article
- Gerecke DR, Chen M, Isukapalli SS, Gordon MK, Chang YC, Tong W, Androulakis IP, Georgopoulos PG: Differential gene expression profiling of mouse skin after sulfur mustard exposure: Extended time response and inhibitor effect. Toxicol Appl Pharmacol. 2009, 234: 156-165. 10.1016/j.taap.2008.09.020.PubMed CentralPubMedView Article
- McInturff JE, Modlin RL, Kim J: The role of toll-like receptors in the pathogenesis and treatment of dermatological disease. J Invest Dermatol. 2005, 125: 1-8. 10.1111/j.0022-202X.2004.23459.x.PubMedView Article
- Parks WC: Matrix metalloproteinases in repair. Wound Repair Regen. 1999, 7: 423-432. 10.1046/j.1524-475X.1999.00423.x.PubMedView Article
- Kahari VM, Saarialho-Kere U: Matrix metalloproteinases in skin. Exp Dermatol. 1997, 6: 199-213. 10.1111/j.1600-0625.1997.tb00164.x.PubMedView Article
- Toy LW: Matrix metalloproteinases: their function in tissue repair. J Wound Care. 2005, 14: 20-22.PubMedView Article
- O'Shaughnessy KD, De La Garza M, Roy NK, Mustoe TA: Homeostasis of the epidermal barrier layer: a theory of how occlusion reduces hypertrophic scarring. Wound Repair Regen. 2009, 17: 700-708. 10.1111/j.1524-475X.2009.00534.x.PubMedView Article
- Colwell AS, Yun R, Krummel TM, Longaker MT, Lorenz HP: Keratinocytes modulate fetal and postnatal fibroblast transforming growth factor-beta and Smad expression in co-culture. Plast Reconstr Surg. 2007, 119: 1440-1445. 10.1097/01.prs.0000256049.53562.39.PubMedView Article
- Tandara AA, Kloeters O, Mogford JE, Mustoe TA: Hydrated keratinocytes reduce collagen synthesis by fibroblasts via paracrine mechanisms. Wound Repair Regen. 2007, 15: 497-504. 10.1111/j.1524-475X.2007.00257.x.PubMedView Article
- Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP: Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003, 4: 249-264. 10.1093/biostatistics/4.2.249.PubMedView Article
- Benjamini Y, Yekutieli D: The control of the false discovery rate in multiple testing under dependency. Ann Statist. 2001, 29: 1165-1188. 10.1214/aos/1013699998.View Article
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.