1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Keloids are a dermal fibrotic disease whose etiology remains totally unknown and for which there is no successful treatment. Here, we employed cDNA microarray analysis to examine gene expression in keloid lesions and control skin. We found that 32 genes among the 9000 tested were strongly up-regulated in keloid lesions, of which 21 were confirmed by Northern blotting. These included at least seven chondrocyte/osteoblast marker genes, and RT-PCR analysis revealed that transcription factors specific for these genes, SOX9 and CBFA1, were induced. Immunostaining and in situ hybridization further supported that these markers are expressed in keloid lesions. Intriguingly, scleraxis, a transcription factor known as a marker of tendons and ligaments, was also induced in keloid fibroblasts. We propose that reprogramming of gene expression or disordered differentiation from a dermal pattern to that of a chondrocytic/osteogenic lineage, probably closer to that of tendon/ligament lineage, may be involved in the etiology of keloids.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Keloids are dermal tumors in which an overabundance of extracellular matrix is produced during the wound-healing process (Abergel et al. 1985). They appear as raised, red and inflexible scar tissue that is itchy and painful. The lesions expand over the boundaries of the initial injury site through the rapid proliferation of fibroblast-like cells (Heenan 1997). Thus, keloids represent a derailment of the protective wound-healing process. The genes encoding tumor growth factor β1 (TGFβ1) and type I and VI collagen (Peltonen et al. 1991; Naitoh et al. 2001) have been reported to be up-regulated in keloid lesions. Type I collagen is a principal component of excessive matrix in the keloid lesions, and TGFβ1 may play a role in stimulating its production and accumulation in the keloid. Type VI collagen is also abundant in the keloid matrix (Peltonen et al. 1991; Naitoh et al. 2001), suggesting an alteration of collagen fibrillogenesis in the pathological situation. Our previous study also indicated that collagen type III was strongly up-regulated in keloid lesions (Naitoh et al. 2001), although expression of this collagen was previously reported to be unaffected by the keloid (Peltonen et al. 1991). However, the characteristics of the keloid lesion and its matrix remain largely unknown.

Proliferation of fibroblastic cells is one of the typical characteristics of keloid lesions. Although Akasaka et al. (2000) reported that apoptosis was induced in keloid fibroblasts, Funayama et al. (2003) recently demonstrated that keratinocytes inhibit the apoptosis of keloid fibroblasts. Thus, there seems to be considerable confusion in understanding the pathogenesis of keloids. This ambiguity and the fact that keloids resist all clinical therapies and often recur, means that the quality of life of many patients is seriously compromised by keloid disease. Notably, keloids only occur in humans and are not observed in animals. The lack of relevant animal models hampers efforts to establish the etiology of keloid disease, and many attempts to explore treatment modalities have proved unsuccessful because of the lack of etiological information.

Here, we performed detailed DNA microarray analysis of keloid tissue specimens in comparison with normal skin, and unexpectedly found that chondrocytic and osteogenic lineage maker genes, including those expressed in tendons and ligaments, are expressed in keloid lesions. We therefore propose that reprogramming of gene expression or disordered differentiation of skin fibroblasts may play a role in keloid etiology.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Expression profile of genes strongly up-regulated in keloid lesions

We first characterized the lineage-specific gene expression of human keloid lesions in comparison to unaffected skin tissue specimens from the same patient. A Cy3-labeled keloid cDNA probe and a Cy5-labeled normal skin cDNA probe were simultaneously hybridized to a cDNA microarray containing 9182 cDNA fragments (HUMAN UniGEM version 2.0 array, Incyte Genomics). Thirty-two genes were found to be up-regulated more than 3.5-fold in the keloid lesions relative to normal skin (Table 1). Northern blotting of total RNA purified from keloids and normal skin showed that 19 genes were up-regulated in the lesions of keloid patients (Fig. 1). Of these, 17 genes were up-regulated in all three patients tested while two genes, neuronal leucine-rich repeat protein-3 (NLRR3) and secreted frizzled-related protein-4 (SFRP4), were up-regulated in two of three patients. In addition to these 19 genes, we previously confirmed the up-regulation of the genes encoding collagen α1(I) chain in keloids by Northern blotting (Naitoh et al. 2001), and another group reported the up-regulation of type VI collagen (Peltonen et al. 1991). The remaining 11 of 32 genes up-regulated on the array gave no signals in the Northern analysis, probably due to the detection limit and the conditions used (data not shown).

Table 1.  List of genes found to be up-regulated in keloid by cDNA microarray analysis
CloneTranscriptGenBank accession no.Differential expression*Northern blottingFunctional classification
  • *

    Genes up-regulated in the keloid lesion more than 3.5-fold relative to normal skin as determined by differential hybridization are shown.

  • The + symbols indicate the genes confirmed to be up-regulated in the keloid lesion by Northern blotting as shown in Fig. 1. NS, no signals.

  • The number indicates the functional classification as described in Fig. 1. NA, not applicable because of no signals on Northern blot.

  • §

    Up-regulation of these genes in keloid lesions were confirmed by Northern blotiing previously (Naitoh et al. 2001).

  • Up-regulation of collagen type VI in keloid lesions was described by Peltonen et al. (1991). Complete microarray data are deposited in the GEO databank ( under the accession number GSE2945.

KSG2fibronectin 1AW38569014.5+3
KSG3collagen α1(I)AU11776211.4+§3
KSG4collagen α2(V)AU11748411.1+3
KSG5O-mannoseβ-1, 2-N-acetylglucosanominyltransferase (POMGnT1)AI92208310.5+3
KSG6collagen α1(XI)J0417710.1+1
KSG7LPS binding protien (LBP)AF105067 9.4NSNA
KSG8mimecan (osteoglycin)AL110267 8.2+3
KSG9lumicanAU120442 6.2+3
KSG10serine protease 11, IGF binding (HtrA1)NM_002775 5.8+3
KSG11neuronal leucine-rich repeat protein-3 (NLRR3)AL442092 5.6+4
KSG12collagen α1(VI)X99135 5.5+3
KSG13cartilage oligomeric matrix protein (COMP)NM_000095 5.4NSNA
KSG14collagen α1(X)X72579 5+§1
KSG15p311AF119859 4.9+2
KSG16hypothetical protein P15-2AL03187 4.9NSNA
KSG17versicanX15998 4.8+3
KSG18thrombosponin 4NM_003248 4.7+1
KSG19opioid-binding protein/cell adhesion molecule-likeU79251 4.6NSNA
KSG20collagen α4(IV)NM_000092 4.5NSNA
KSG21collagen α1(XVI)S57132 4.5+3
KSG22EST(highly similar to IFT2 interferon induced protein with tetratricopeptide repeats 2)AI609324 4.3NSNA
KSG23collagen α1(XV)L01697 4.3+3
KSG24protein kinease, cAMP-dependant, catalytic, betaNM_002731 4.1NSNA
KSG25fibroblast activation protein, alpha (FAP-α)NM_004460 4.1+4
KSG26secreted frizzled-related protein 4 (SFRP4)AF026692 4.1+4
KSG27homo sapiens cDNA FLJ11362fis, cloneHEMBA1000244AW444842 3.9NSNA
KSG28solute carrier family 14 (urea transporter), member2X96969 3.8NSNA
KSG29procollagen-lysine, 2-oxoglutarate 5-dioxygenase (lysine hydroxylase) 2NM_000935 3.7NSNA
KSG30aggrecan 1NM_013227 3.7NSNA
KSG31cadherin 11 (OB-cadherin)L34056 3.6+1
KSG32transforming growth factor beta3 (TGFβ3)X14885 3.5+1

Figure 1. Northern blot analysis of genes up-regulated in keloid lesions. Total RNA (10 µg/lane) was electrophoresed, blotted and hybridized with 32P-labeled human cDNA probes. Normal skin of an unrelated patient (lane 1), unaffected regions of skin from a keloid patient (lane 2), a keloid lesion from the same patient (lane 3), and keloid lesions of two other patients (lane 4 and 5) were analyzed. The 19 genes up-regulated in the keloids were divided into four groups according to their functions, as follows. Group 1: osteogenesis and chondrogenesis-associated genes; Group 2: tumorigenesis-associated genes; Group 3: extracellular matrix and matrix-associated genes; and Group 4: genes of unknown functions. * Also reported in Naitoh et al. (2001) without raw data.

Download figure to PowerPoint

Twenty-one genes confirmed to be up-regulated in keloid lesions can be classified into four groups according to their reported functions (Fig. 1 and Table 1): Group 1, osteogenesis-associated and chondrogenesis-associated genes (seven genes in Fig. 1); Group 2, tumorigenesis-associated genes (two genes in Fig. 1); Group 3, extracellular matrix and related genes (eight genes in Fig. 1 and collagen types I and VI); Group 4, genes of unknown functions (two genes in Fig. 1). Although the up-regulation of extracellular matrix genes is consistent with previous reports, the strong ectopic expression of osteogenesis- and chondrogenesis-associated genes (Group 1) in keloid lesions, which was unexpectedly found, suggested that there is a marked alteration of gene expression pattern in the keloid lesions.

Ectopic expression of osteogenesis- and chondrogenesis-associated genes in keloid lesions

Of the genes in Group 1, periostin (also called osteoblast specific factor 2), is a secreted protein known to be specifically expressed in periosteum and periodontal ligaments (Saito et al. 2002). Collagen types XI and X are specifically produced in cartilage (Mendler et al. 1989) and hypertrophic cartilage (Nerlich et al. 1992), respectively. Thrombospondin 4, an extracellular matrix calcium-binding protein that stimulates cell proliferation, distributes in chondrogenic regions and prechondrogenic mesenchyme (Tucker et al. 1995). OB-cadherin (cadherin 11) is a cell-adhesion molecule that is selectively expressed in osteoblastic cells and plays an important role in bone cell differentiation (Okazaki et al. 1994). TGFβ3 is able to promote chondrogenic differentiation in cultures of human mesenchymal stem cells (Makay et al. 1998). Although HtrA1 (also called serine protease 11) was initially identified as a gene down-regulated by transformation with SV40, using a human fibroblast cell line (Zumbrunn & Trueb 1996), a recent report indicated that this gene is highly expressed in skeletal elements and that its protein product inhibits signaling from TGF-β family proteins including bone morphogenetic protein 2 (BMP2), BMP4 and TGFβ1 (Oka et al. 2004).

To confirm the expression of osteoblast and chondrocyte marker genes in keloid lesions, we performed in situ hybridization using periostin and OB-cadherin RNA probes on keloid tissue sections (Fig. 2A–D). The expression of these genes was clearly detected in the fibroblastic cells of the hypercellular and actively growing portion of keloid lesions, but not in those of normal skin. These osteogenesis- and chondrogenesis-associated genes were expressed in keloid lesions despite the fact that the keloids develop from dermal tissues, and thus these observations suggest that gene expression in keloid lesions is altered toward that of an osteo-chondrogenetic lineage.


Figure 2. In situ hybridization of osteoblast and chondrocyte marker genes. Sections from active lesions of keloid tissues (A, C, E, G) or normal skin tissues (B, D, F, H) were hybridized with probes specific to (A, B) periostin, (C, D) OB-cadherin, (E, F) lumican and (G, H) mimecan. Positive signals are visualized in red. Bar = 20 µm.

Download figure to PowerPoint

Up-regulation of genes encoding extracellular matrix proteins

In addition to the previously identified collagen types I and VI, many genes encoding general extracellular matrix proteins, including collagen types V, XV and XVI and fibronectin, were also found to be strongly up-regulated in the keloid lesions (Group 3 in Fig. 1). The data are consistent with the fact that keloid lesions are extremely abundant in extracellular matrix. While collagen type V is a fibrillar collagen that widely distributes in extracelluar matrix of many types of tissues, a recent report indicated that expression of this collagen is strongly up-regulated during mouse bone development (Kahai et al. 2004), which is consistent with the view that osteogenesis-associated gene expression occurs in keloids. Although type I collagen is constitutively expressed in dermal tissues at a level much lower than in bone, a dramatic induction of this gene (collagen α1(I) gene, 11.4-fold induction compared to normal skin, Table 1 and Naitoh et al. 2001) is also indicative that genes strongly expressed in the osteogenesis lineage are up-regulated in keloid lesions. Type XV collagen is known to be abundant in the basement membranes of capillary blood vessels, particularly in developing mouse embryos (Mouna et al. 2004), which is consistent with the fact that keloid lesions are rich in capillaries. Although type XVI collagen is a fibril-associated collagen with interrupted triple helices that binds to fibrillin-1 and fibronectin (Kassner et al. 2004), little is known about its biological role except that it may be involved in organization of fibrilar structures within the extracellular matrix. Fibronectin is a glycoprotein that connects cells to extracellular matrix components.

Proteoglycans including lumican, mimecan and versican, were also up-regulated in keloids as determined by Northern blotting. Lumican and mimecan (also called osteoglycin), which both belong to the small leucine rich proteoglycan (SLRP) subfamily, are the major components of cornea matrix in the eye and are also expressed in the skin. In situ hybridization further confirmed that fibroblastic cells in the active keloid lesions abundantly expressed lumican (Fig. 2E) and mimecan (Fig. 2G). As studies of knockout mice indicated that these proteins prevent abnormal acceleration of collagen fibril assembly (Chakravarti et al. 1998; Tasheva et al. 2002), it is possible that the high expression levels of these two proteoglycans may cause abnormal collagen fibrinogenesis in keloids. Versican is a large extracellular matrix proteoglycan that distributes in a variety of tissues (Wight 2002). Although immunohistocheimical studies of versican previously indicated that this protein localizes along with elastic fibers in the skin dermis, and is expressed in cultured fibroblasts (Zimmermann et al. 1994), cells producing versican mRNA have recently been determined to be dermal papilla cells (Kishimoto et al. 1999; Botchkarev & Kishimoto 2003). As the level of versican mRNA was markedly elevated in keloid lesions compared with normal skin (Fig. 1), it suggests that relative to normal skin, keloid lesions may be enriched in cells with papilla-like characteristics. Protein O-mannose-β-1,2-N-acetylglucosaminyltransferase 1 (POMGnT1) is a glycosyltransferase which is responsible for the formation of the GlcNAcβ1-2Man linkage of O-mannosyl glycans, such as α-dystroglycan. Mutations of the POMGnT1 gene cause muscle-eye-brain disease, an autosomal disorder characterized by congenital muscle dystrophy, brain malformation and ocular abnormalities (Yoshida et al. 2001). O-mannosyl glycans are rarely found in mammals, with α-dystroglycan the only mammalian example kown to date. However, α-dystroglycan expression level is very low in the skin, and its expression level in the keloid lesions was similar to that in normal skin, as measured by microarray analysis (1.1-fold). Thus, the identity of the O-mannosyl glycan recognized by POMGnT1 in keloids is currently unclear and remains to be determined. Overall, these results indicate abnormal up-regulation of a number of proteoglycans and related genes that are minor components of the normal skin. This suggests that in addition to major skin components like collagen, the regulatory mechanisms of matrix organization are also affected in keloid lesions.

Up-regulation of tumorigenesis-associated and other genes

In addition to the osteogenesis- and chondrogenesis-associated genes, two tumorigenesis-related genes, p311 and fibroblast activation protein alpha (FAP-α) (Group 2 in Fig. 1), were up-regulated in the keloid lesions. The p311 protein enhances glioma cell mobility during glioma invasion (Mariani et al. 2001). FAP-α is an integral membrane serine protease with both dipeptidyl-peptidase and gelatinase activities, and has been suggested to function in extracellular matrix degradation or activation of growth factors during tissue remodeling (Park et al. 1999). Up-regulation of the group 2 genes is consistent with the fact that fibroblastic cells in keloid lesions proliferate in a rapid and unregulated manner and invade into normal skin tissues, beyond the boundaries of the initial region of injury.

We identified two additional genes that are up-regulated in keloid lesions (Group 4): secreted frizzled-related protein 4 (sFRP4), and neuronal leucine rich repeat protein 3 (NLRR3), although little is known about their functions. The roles of these genes in keloid pathogenesis remain to be determined.

Accumulation of cartilage-like matrix and type XI collagen

Fibroblasts in keloid lesions stained strongly with a specific antibody against type XI collagen, which is one of the major components of cartilage and is never detected in the skin (Mendler et al. 1989) (Fig. 3A–C). To examine whether the keloid lesions show metachromasia, which indicates the accumulation of cartilage-like matrix, we performed toluidine blue staining at pH 2.5. The keloid lesion clearly showed metachromasia, while the normal skin region surrounding the keloid lesion did not (Fig. 2L). These results confirm the up-regulation of osteoblast and chondrocyte markers at the protein level and the accumulation of cartilage-like matrix in keloid lesions.


Figure 3. (A–F) Immunohistochemical staining of type XI collagen at the boundary region between keloid and normal skin tissues. Larger magnification of views in panel A are shown in panels B (normal region) and C (keloid lesion). Positive signals are visualized in red. (D) Toluidine blue staining for examining metachromasia, a marker of cartilage-like matrix accumulation. k, keloid lesion; n, unaffected skin. (E, F) Immunostaining of versican in keloid hypertrophic lesions (E), and normal skin (F). Versican accumulated in the interspace between large hyalinaized collagen bundles, and fibroblastic cells surrounding abnormal collagen bundles were also stained. In normal skin, weak staining was observed in an elastic fiber pattern. Bar = 100 µm (A, D), 20 µm (B, C), or 50 µm (E, F).

Download figure to PowerPoint

Metachromasia indicates excessive accumulation of chondroitin sulfate in the keloid lesions, while aggrecan and versican are major proteoglycans known to contain chondroitin sulfate. Versican mRNA was markedly increased in the lesions (Fig. 1, Group 2), whereas aggrecan mRNA expression was so low to be undetectable by Northen blot analysis (Table 1). Thus, this metachromasia was assumed to be due to excessive accumulation of versican. To confirm versican overabundance at the protein level, we performed immunohistochemical staining of the keloid lesions and normal skin samples with an anti-versican antibody. Fibroblast-like cells and the intercellular space surrounding typical large and hyalinized collagen bundlles stained strongly in the keloid lesions, while only faint staining was observed along elastic fibers (in the extracellular matrix) in the normal skin. These data indicated that the keloid matrix, like cartilage, contains abundant chondroitin sulfate as (Fig. 3E, F).

However, as the major source of chondroitin sulfate is versican in keloids but aggrecan in cartilage (Kiani et al. 2002), clearly the keloid matrix has similar, but not identical, characteristics to those of cartilage.

Expression of osteogenesis- and chondrogenesis-specific transcription factors, SOX9 and CBFA1

The transcription factors SOX9 and CBFA1 activate many genes involved in osteogenesis (Ducy et al. 1997) and chondrogenesis (Bi et al. 1999; Enomoto et al. 2000), including some of the genes that were up-regulated in keloid lesions (e.g. collagen type I (Kern et al. 2001) and XI (Bridgewater et al. 1998)). RT-PCR using specific primers revealed significant expression of SOX9 and CBFA1 gene in keloid lesions of independent patients, but not in normal skin (Fig. 4A). Furthermore, SOX9 and CBFA1 mRNAs were expressed in primary cultures of fibroblasts derived from keloids at levels similar to their expression in chondrosarcoma (OUMS-27) and osteosarcoma (HuO9) cells, respectively (Fig. 4B). In contrast, the RT-PCR products of SOX9 and CBFA1 were never detected in normal dermal fibroblasts. In situ hybridization using a specific probe confirmed that CBFA1 mRNA was highly expressed in the keloid lesions, but not in normal skin (Fig. 4C). These results indicate that both SOX9 and CBFA1 are ectopically expressed in keloid lesions, which supports the view that in keloids, gene expression at the transcriptional level is altered towards that of chondrocytes or osteoblasts.


Figure 4. Analysis of SOX9, CBFA1 and scleraxis mRNA expression in keloids and their fibroblasts. (A) SOX9, CBFA1 mRNAs were amplified by RT-PCR from total RNA extracted from keloid and normal skin tissues. Keloid tissues and normal skin tissues were obtained from five different keloid patients and one unaffected patient, respectively. (B) RT-PCR analysis of SOX9, CBFA1 and scleraxis mRNAs in keloid fibroblasts and normal fibroblasts. OUMS27 and HU09 are chondrosarcoma and osteosarcoma cell lines, respectively. (C) In situ hybridization for CBFA1 and keloid skin tissues. Bar = 20 µm.

Download figure to PowerPoint

Expression of scleraxis, a tendon/ligament lineage-associated transcription factor, in keloid-derived fibroblastic cells

Although the skeletogenesis-regulator transcription factors SOX9 and CBFA1 were ectopically expressed in keloids, some skeletogenesis marker genes, including collagen type II, were not induced in these lesions (data not shown). Among the tissues expressing chondrocyte- and osteoblast-lineage markers, fibroblastic cells derived from fibrous cartilage, including tendons and periodontal ligaments, are reported to express little or no type II collagen, while they strongly express type I collagen and CBFA1 genes (Saito et al. 2002; Salingcarnboriboon et al. 2003). We therefore assumed that the gene expression pattern in keloid lesions may more closely resemble that of tendons or ligaments than that in cartilage or bone. To test this possibility, we used RT-PCR to determine the expression of scleraxis, which is a transcription factor predominantly expressed in cells of tendon and ligament lineage, but not in differentiated cartilage or bone (Schweitzer et al. 2001), in the keloid-derived fibroblastic cells. Scleraxis was expressed in the cells derived from keloid legions, but not in those of normal skin (Fig. 4B). These results support the view that tendon/ligament-like gene expression occurs in keloids.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

In the present study, we found that many osteogenesis and chondrogenesis-associated genes are ectopically expressed in keloid lesions (Figs 1 and 2, Table 1). These genes included collagen types X and XI, periostin, thrombospondin 4, OB-cadherin, HtrA1, and TGFβ3. Collagen types X (Nerlich et al. 1992) and XI (Mendler et al. 1989), periostin (Saito et al. 2002) and thrombospondin 4 (Tucker et al. 1995) are extracellular matrix proteins in the cartilage and/or bone. OB-cadherin is an osteoblast cell adhesion molecule. TGFβ3 (Mackay et al. 1998) and HtrA1 (Oka et al. 2004) are a signaling molecule and inhibitor controlling chondro-osteogenesis, respectively. In addition, enhanced expression of collagen types I and V, which are known to be up-regulated during chondro-osteogenesis, was observed in keloids (Kahai et al. 2004). Furthermore, lumican, which is a major proteoglycan component of bone matrix (Raouf et al. 2002), and mimecan, which is reported to induce ectopic bone formation, were among the matrix components expressed in keloids (Bentz et al. 1989). Versican is expressed in prechondrogenic mesenchyme and chondrogenic regions during the early stages of differentiation (Shibata et al. 2003). Notably, the clinical appearance of keloid lesions also resembles the hard gel-like texture of cartilage.

Consistent with these observations, the genes encoding transcription factors SOX9 and CBFA1, which play essential roles in the regulation of gene expression in osteogenesis and chondrogenesis, are expressed at a high level in keloid legions, but not in normal skin (Fig. 4). SOX9 is critical for chondrocyte differentiation and cartilage formation, because no cartilage develops in teratomas derived from Sox9−/− embryonic stem cells and Sox9−/− cells are present as an undifferentiated mesenchyme in mouse chimeras (Bi et al. 1999). Furthermore, mutations in SOX9 are found in patients with campomelic dysplasia, which is characterized by severe malformation of essentially all cartilage-derived structures (Foster et al. 1994; Wagner et al. 1994). CBFA1 is essential for skeletal development, as CBFA1 (Runx2)-deficient mice completely lack bone formation (Komori et al. 1997; Otto et al. 1997) and mutations of this gene cause cleidocranial dysplasia (Lee et al. 1997; Mundlos et al. 1997). Thus, these transcription factors are most likely to play important roles in the ectopic expression of chondrogenesis/osteogenesis-marker genes in keloid lesions.

Although our present data indicate that chondrocyte-lineage marker genes, including collagen types X and XI (Fig. 1) and cartilage oligomeric matrix protein (COMP, Table 1), are ectopically expressed in keloid lesions, the expression of type II collagen, a typical chondrocyte marker, was not detected by Northern blot analysis (data not shown), suggesting that the gene expression pattern in keloids is not exactly identical to that of cartilage or bone. Among tissues expressing chondrocyte- and osteoblast-lineage markers, some fibrocartilages, including tendons and ligaments, contain only a small quantity of type II collagen, in the type I collagen-rich matrix (Webb et al. 1998; Benjamin & Ralphs 2004). Recently, tendon-derived and periodontal ligament-derived cell lines were established and these cells were shown to express both chondrocyte- and osteoblast-lineage markers including SOX9 and CBFA1 (Saito et al. 2002; Salingcarnboriboon et al. 2003). Although periodontal ligament-derived cells express periostin, an osteoblast-specific marker (Takeshita et al. 1993), no expression was observed for osteocalcin (OCN), a marker of osteoblasts in a late stage or bone sialoprotein (BSP), a marker of middle-stage osteoblasts (Saito et al. 2002). In the present study, periostin was strongly expressed in keloid tissues (Figs 1 and 2), but no significant expression of OCN and BSP was detected by our microarray analysis or RT-PCR (data not shown). Taken together, these observations suggested that the gene expression profile of keloid tissue is more similar to that of fibrous cartilage including periodontal ligaments. More importantly, scleraxis, a basic helix-loop-helix (bHLH) transcription factor that is expressed in tendons and ligaments during their developmental stages from early progenitors to mature tissues, but not in differentiated cartilage or bone (Schweitzer et al. 2001; Seo et al. 2004), was found to be expressed in fibroblasts derived from keloid lesions (Fig. 4). Thus, this indicates that the keloid gene expression pattern is likely to more closely resemble that of tendons or ligaments than that of cartilage or bone.

The strong HtrA1 expression in keloids is consistent with this view, because this protein is known to divert the fate of mesenchyme to periosteum, peichondrium, tendons and ligaments through the inhibition of BMP signaling (Oka et al. 2004). Intriguingly, HtrA1 is reported to be expressed in the hair follicles of skin (Oka et al. 2004). Similarly, versican is expressed by dermal papillae in the hair follicles (Kishimoto et al. 1999). These observations therefore present the possibility that the dermal papillae-like characteristics of keloid lesions may be involved in their pathogenesis, although this is an aspect that requires further study.

As keloid studies have to date been seriously hampered by a lack of information regarding their etiology, the finding that the gene expression profile of keloids is altered towards chondrocyte and/or osteoblast lineages, probably resembling those of tendons or ligaments, may enable the development of new treatment modalities for keloids. We are currently comparing the gene expression profiles of keloid lesions and dermal papillae.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Tissue specimens

Nine patients with keloids (aged 13–70 years) and three unrelated patients (aged 14–70 years) undergoing surgical treatment were enrolled in this study. Informed consent was obtained from all of the patients. Keloid diagnosis was performed on the basis of their clinical appearance, history and anatomical location. All lesions analyzed in this study satisfied the histopathological criteria for keloids. Keloids and surrounding unaffected skin tissue specimens were surgically obtained from the patients. One normal skin sample was obtained by full-thickness punch biopsy from an unrelated patient undergoing cosmetic surgery. Tissue samples were immediately frozen in liquid nitrogen and stored at −80 °C prior to analysis.

RNA preparation and cDNA microarray analysis

Total RNA was isolated from tissue specimens using the single-step method (Chomczynski & Sacchi 1987). For microarray analysis, mRNA was purified from total RNA using an Oligotex-dT30 mRNA Purification Kit (TAKARA, Kyoto, Japan). The cDNA microchip array experiment was performed using the public domain HUMAN Uni GEM version 2.0 array that contains 9182 cDNA fragments and was analyzed with GemTools software (Incyte Genomics, Palo Alto, CA, USA). cDNAs were synthesized from keloid and normal skin mRNA, and labeled with Cy3 and Cy5, respectively. Labeled cDNAs were hybridized simultaneously to the microarray. Fluorescent ratios were calculated for all elements and normalized using controls for sensitivity, differential expression, reverse transcription quality, and signal-to-background ratio.

Complete microarray data are deposited in the GEO databank ( under the accession number GSE2945.

Northern blot analysis

cDNA probes were generated by PCR from GEM Microarray Clones (Incyte Genomics) using specific primers. Total RNA was electrophoresed in 2.2 m formaldehyde/1% agarose gels, and transferred onto nylon filters. The filters were hybridized with 32P-labeled cDNA probes in ExpressHyb Hybridization Solution (Clontech, Palo Alto, CA, USA) at 65 °C for 1 h and washed in 0.1 × SSC, 0.1% SDS at 68 °C. Radioactivity on the filters was analyzed using a STORM 820 system (Molecular Dynamics, Sunnyvale, CA, USA).

Primary cell culture

Tissue samples were cut into 1–3 mm3 pieces, placed into plastic tissue culture dishes, and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum. Semiconfluent cultures of fibroblastic cells were passaged by trypsinization up to 3 times prior to analysis.

Cell lines

Human chondrosarcoma OUMS-27 and osteosarcoma Hu09 cell lines were obtained from the Health Science Research Resources Bank (Sennan, Japan) and cultured in 10% FCS/DMEM and 10% FCS/RPMI1640, respectively.

Histological and immunohistochemical studies

Tissue samples were fixed in 4% paraformaldehyde at 4 °C and paraffin sections (3 µm) were prepared. Deparaffinized sections were subjected to staining with toluidine blue at pH 2.5. Immunohistochemical analyses were performed after blocking endogenous peroxidase and nonspecific protein binding activities. Blocked sections were incubated with an antibody against type XI collagen (LSL, Tokyo, Japan) at a dilution of 1 : 400, followed by incubation with a peroxidase-conjugated anti-rabbit IgG.

For immunostaining of versican, sections were fixed in 95% ethanol containing 1% acetic acid. Sections were incubated with anti-versican antibody (Seikagaku Corp., Tokyo, Japan) at a dilution of 1 : 400, followed by incubation with a peroxidase-conjugated anti-mouse IgG. Sections were stained using a LSAB/HRP kit (DakoCytomation, Carpinteria, CA, USA) and counterstained with hematoxylin.


Total RNA was prepared from cultured cells using the RNeasy total RNA kit (Qiagen, Hilden, Germany). First-strand cDNA was synthesized from 1 µg total RNA by reverse transcription at 50 °C for 30 min, and specific DNA fragments were amplified with the One Step RNA PCR kit (Takara Bio, Otsu, Japan) using the following oligonucleotide primers: β-actin: 5′CAAGAGATGGCCACGGCTGCT and 5′TCCTTCTGCATCCTGTCGGCA; Cbfa1: 5′AGCTGCAATCACCAACCAC and 5′TGCTGTGGTTGGTGATTGCAGCT; Sox9: 5′CTCCTGGACTCAAAGGGCCTTTTCTC and 5′AGGAGAGAAAAGGCCCTTTGAGTCCAGGAG; Scleraxis: 5′GTGAACACGGCCTTCACGG and 5′CTGCGAATCGCTGTCTTTC; GAPDH: 5′TGGTATCGTGGAAGGACTCATGAC and 5′ATGCCAGTGAGCTTCCCGTTCAGC. PCR amplification was performed as 25 cycles for β-actin, CBFA1 and SOX9, and 40 cycles for scleraxis of denaturation at 94 °C for 30 s, annealing at 58 °C for 30 s, and extension at 72 °C for 1 min. PCR products were analyzed by agarose gel electrophoresis followed by ethidium bromide staining.

In situ hybridization

For periostin, lumican, mimecan and OB-cadherin, digoxigenin-UTP-labeled single-stranded RNA probes were generated from GEM microarray clones, using a DIG RNA Labeling Kit (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. For CBFA1, a 0.3 kb mouse exon I cDNA fragment (Horiuchi et al. 1999) was subcloned into pBluescript SK (Stratagene) and used for probe preparation. The tissue sections were deparaffinized and were treated with proteinase K (4 µg/mL) at 37 °C for 15 min, followed by rinsing in PBS. The specimens were incubated with chondroitin ABC lyase (0.02 units/mL) in blocking buffer (0.1 m Tris-HCl pH 8.0, 30 mm sodium acetate, 0.05% bovine serum albumin) at 37 °C for 20 min, according to the method of Tsukifuji et al. (1997).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Professor Y. Nishimura for helpful comments on experiments. This work was partly supported by a grant for Kasahara Memorial Foundation.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Abergel, R.P., Pizzurro, D., Meeker, C.A., et al. (1985) Biochemical composition of the connective tissue in keloids and analysis of collagen metabolism in keloid fibroblast cultures. J. Invest. Dermatol. 84, 384390.
  • Akasaka, Y., Ishikawa, Y., Ono, I., et al. (2000) Enhanced expression of caspase-3 in hypertrophic scars and keloid: induction of caspase-3 and apoptosis in keloid fibroblasts in vitro. Lab. Invest. 80, 345357.
  • Benjamin, M. & Ralphs, J.R. (2004) Biology of fibrocartilage cells. Int. Rev. Cytol. 233, 145.
  • Bentz, H., Nathan, R.M., Rosen, D.M., et al. (1989) Purification and characterization of a unique osteoinductive factor from bovine bone. J. Biol. Chem. 264, 2080520810.
  • Bi, W., Deng, J.M., Zhang, Z., Behringer, R.R. & de Crombrugghe, B. (1999) Sox9 is required for cartilage formation. Nature Genet. 22, 8589.
  • Botchkarev, V.A. & Kisimono, J. (2003) Molecular control of epithelial–mesenchymal interactions furing hair follicle cycling. J. Investig. Dermatol. Symp. Proc. 8, 4655.
  • Bridgewater, L.C., Lefebvre, V. & de Crombrugghe, B. (1998) Chondrocyte-specific enhancer elements in the Col11a2 gene resemble the Col2a1 tissue-specific enhancer. J. Biol. Chem. 273, 1499815006.
  • Chakravarti, S., Magnuson, T., Lass, J.H., et al. (1998) Lumican regulates collagen fibril assembly: skin fragility and corneal opacity in the absence of lumican. J. Cell Biol. 141, 12771286.
  • Chomczynski, P. & Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156159.
  • Ducy, P., Zhang, R., Geoffroy, V., Ridall, A.L. & Karsenty, G. (1997) Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89, 747754.
  • Enomoto, H., Emoto-Iwamoto, M., Iwamoto, M., et al. (2000) Cbfa1 is a positive regulatory factor in chondrocyte maturation. J. Biol. Chem. 275, 86958702.
  • Foster, J.W., Dominguez-Steglich, M.A., Guioli, S., et al. (1994) Campomelic dysplasia and autosomal sex reversal caused by mutations inan SRY-related gene. Nature 372, 525530.
  • Funayama, E., Chondon, T., Oyama, A. & Sugihara, T. (2003) Keratinocytes promote proliferation and inhibit apoptosis of the underlying fibroblasts: an important role in the pathogenesis of keloid. J. Invest. Dermatol. 121, 13261331.
  • Heenan, P.J. (1997) Tumors of the fibrous tissue involving the skin. In: Lever's Histopathology of the Skin (eds D.Elder, R.Elenitsas, C.Jaworsky & B.Johnson Jr), pp. 847887. Philadelphia, Pennsylvania: Lippincott-Raven Publishers.
  • Horiuchi, K., Amizuka, N., Takeshita, S., et al. (1999) Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. J. Bone Miner. Res. 14, 12391249.
  • Kahai, S., Vary, C.P., Gao, Y. & Seth, A. (2004) Collagen, type V, alpha1 (COL5A1) is regulated by TGF-beta in osteoblasts. Matrix Biol. 23, 445455.
  • Kassner, A., Tiedemann, K., Notbohm, H., et al. (2004) Molecular structure and interaction of recombinant human type XVI collagen. J. Mol. Biol. 339, 835853.
  • Kern, B., Shen, J., Starbuck, M. & Karsenty, G. (2001) Cbfa1 contributes to the osteoblast-specific expression of type I collagen genes. J. Biol. Chem. 276, 71017107.
  • Kiani, K., Chen, L., Wu, Y.J., Yee, A.J. & Yang, B.B. (2002) Structure and function of aggrecan. Cell Res. 12, 1932.
  • Kishimoto, J., Ehama, R., Wu, L., Jiang, S., Jiang, N. & Burgeson, R.E. (1999) Selective activation of versican promoter by epithelial–mesenchmal interactions during hair follicle development. Proc. Natl. Acad. Sci. USA 96, 736741.
  • Komori, T., Yagi, H., Nomura, S., et al. (1997) Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89, 755764.
  • Lee, B., Thirunavukkarasu, K., Zhou, L., et al. (1997) Missense mutations abolishing DNA binding of the osteoblast-specific transcription factor OSF2/CBFA1 in cleidocranial dysplasia. Nature Genet. 16, 307310.
  • Makay, A.M., Beck, S.C., Murphy, J.M., et al. (1998) Chondrogenic differentiation of cultured human mesenchymal stem cells from marrow. Tissue Eng. 4, 415428.
  • Mariani, L., McDonough, W.S., Hoelzinger, D.B., et al. (2001) Identification and validation of P311 as a glioblastoma invasion gene using laser capture microdissection. Cancer Res. 61, 41904196.
  • Mendler, M., Eich-Bender, S.G., Vaughan, L., Winterhalter, K.H. & Bruckner, P. (1989) Cartilage contains mixed fibrils of collagen types II, IX, and XI. J. Cell Biol. 108, 191197.
  • Mouna, A., Eklund, L., Vaisanen, T. & Pihlajaniemi, T. (2004) Developmentally regulated expression of type XV collagen correlates with abnormalities in Col15a1 (−/−) mice. Matrix Biol. 21, 89102.
  • Mundlos, S., Otto, F., Mundlos, C., et al. (1997) Mutations involving the transcription factor CBFA1 cause cleidocranial dysplasia. Cell 89, 773779.
  • Naitoh, M., Hosokawa, N., Kubota, H., et al. (2001) Upregulation of HSP47 and collagen type III in the dermal fibrotic disease, keloid. Biochem. Biophys. Res. Commun. 280, 13161322.
  • Nerlich, A.G., Kirsch, T., Wiest, I., Betz, P. & von Der Mark, K. (1992) Localization of collagen X in human fetal and juvenile articular cartilage and bone. Histochemistry 98, 275281.
  • Oka, C., Tsujimoto, R., Kajikawa, M., et al. (2004) HtrA1 serine prtoease inhibits signaling mediated by Tgfbeta family proteins. Development 131, 10411053.
  • Okazaki, M., Takeshita, S., Kawai, S., et al. (1994) Molecular cloning and characterization of OB-cadherin, a new member of cadherin family expressed in osteoblasts. J. Biol. Chem. 269, 1209212098.
  • Otto, F., Thornell, A.P., Crompton, T., et al. (1997) Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development. Cell 89, 765771.
  • Park, J.E., Lenter, M.C., Zimmermann, R.N., Garin-Chesa, P., Old, L.J. & Rettig, W.J. (1999) Fibroblast activation protein, a dual specificity serine protease expressed in reactive human tumor stromal fibroblasts. J. Biol. Chem. 274, 3650536512.
  • Peltonen, J., Hsiao, L.L., Jaakkola, S., et al. (1991) Activation of collagen gene expression in keloids: co-localization of type I and VI collagen and transforming growth factor-beta 1 mRNA. J. Invest. Dermatol. 97, 240248.
  • Raouf, A., Ganss, B., McMahon, C., Vary, C., Roughley, P.J. & Seth, A. (2002) Lumican is a major proteoglycan component of the bone matrix. Matrix Biol. 21, 361367.
  • Saito, Y., Yoshizawa, T., Takizawa, F., et al. (2002) A cell line with characteristics of the periodontal ligament fibroblasts is negatively regulated for mineralization and Runx2/Cbfa1/Osf2 activity, part of which can be overcome by bone morphogenetic protein-2. J. Cell Sci. 115, 41914200.
  • Salingcarnboriboon, R., Yoshitake, H., Tsuji, K., et al. (2003) Establishment of tendon-derived cell lines exhibiting pluripotent mesenchymal stem cell-like property. Exp. Cell Res. 287, 289300.
  • Schweitzer, R., Chyung, J.H., Murtaugh, L.C., et al. (2001) Analysis of tendon cell fate using Scleraxis, a specific marker for tendon and ligaments. Development 128, 38553866.
  • Seo, B.M., Miura, M., Gronthos, S., et al. (2004) Investigation of multipotent postnatal stem cells from human periodontal ligament. Lancet. 364, 149155.
  • Shibata, S., Fukuda, K., Imai, H., Abe, T. & Yamashita, Y. (2003) In situ hybridization and immunohistochemistry of versican, aggrecan and link protein, and histochemistry of hyaluronan in the developing mouse limb bud cartilage. J. Anat. 203, 425432.
  • Takeshita, S., Kikuno, R., Tezuka, K. & Amann, E. (1993) Osteoblast-specific factor 2: cloning of a putative cone adhesion protein with homology with the insect protein fasciclin I. Biochem. J. 294, 271278.
  • Tasheva, E.S., Koester, A., Palsen, A.Q., et al. (2002) Miemecan/ osoteoglycin-deficient mice have collagen fibril abnormalities. Mol. Vis. 31, 407415.
  • Tsukifuji, R., Sakai, Y., Hatamochi, A. & Shinkai, H. (1997) Gene expression of matrix metalloproteinase-1 (interstitial collagenase) and matrix metalloproteinase-3 (stromelysin-1) in basal cell carcinoma by in situ hybridization using chondroitin ABC lyase. Histochem. J. 29, 401407.
  • Tucker, R.P., Adams, J.C. & Lawler, J. (1995) Thrombospondin-4 is expressed by early osteogenic tissues in the chick embryo. Dev. Dyn. 203, 477490.
  • Wagner, T., Wirth, J., Meyer, J., et al. (1994) Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell 79, 11111120.
  • Webb, P.P., Moxham, B.J., Ralphs, J.R. & Benjamin, M. (1998) Immunolocalisation of collagens in the developing rat molar tooth. Eur. J. Oral. Sci. 106 (Suppl. 1), 147155.
  • Wight, T.N. (2002) Versican: a versatile extracellular matrix proteoglycan in cell biology. Curr. Opin. Cell Biol. 14, 617623.
  • Yoshida, A., Kobayashi, K., Manya, H., et al. (2001) Muscular dystrophy and neuronal migration disorder caused by mutations in a glycosyltransferase, POMGnT1. Dev. Cell 1, 717724.
  • Zimmermann, D.R., Dours-Zimmermann, M.T., Schubert, M. & Bruckner-Tudrman. L. (1994) Versican is expressed in the proliferating zone in the epidermis and in association with the elastic network of the derimis. J. Cell Biol. 124, 817825.
  • Zumbrunn, J. & Trueb, B. (1996) Primary structure of a putative serine protease specific for IGF-binding proteins. FEBS Lett. 398, 187192.