Interleukin (IL) 1α–mediated up-regulation of gene clusters in postauricular skin–derived fibroblasts (SFs) and middle ear cholesteatoma–derived fibroblasts (MECFs). Representation of messenger RNA expression levels in spontaneous SF, IL-1α–stimulated SF, spontaneous MECF, and IL-1α–stimulated MECF. The SFs and MECFs were stimulated with IL-1α (10 ng/mL) for 4 hours. The colored bars show the magnitude of the response of each gene, according to the scale of expression level shown. Cluster A contained 164 genes (1) in which expression in SFs increased at least 2-fold after stimulation with IL-1α and (2) in which increased gene expression was more than 2-fold that of IL-1α–stimulated MECFs. Cluster B contained 84 genes (1) in which expression in MECFs increased at least 2-fold after stimulation with IL-1α and (2) in which increased gene expression was more than 2-fold that of IL-1α–stimulated SFs.
Interleukin (IL) 1α–mediated down-regulation of gene clusters in postauricular skin-derived fibroblasts (SFs) and middle ear cholesteatoma–derived fibroblasts (MECFs). Representation of messenger RNA expression levels in spontaneous SF, IL-1α–stimulated SF, spontaneous MECF, and IL-1α–stimulated MECF. The SFs and MECFs were stimulated with IL-1α (10 ng/mL) for 4 hours. The colored bars show the magnitude of the response of each gene, according to the scale of the expression level shown. Cluster C contained 65 genes (1) in which expression in SFs decreased by at least half after stimulation with IL-1α and (2) in which expression in spontaneous SFs increased gene expression by more than 2-fold that of spontaneous MECFs. Cluster D contained 69 genes (1) in which expression in MECFs decreased by at least half after stimulation with IL-1α and (2) in which expression in spontaneous MECFs increased gene expression by more than 2-fold that of spontaneous SFs.
Validation of microarray expression levels by real-time polymerase chain reaction (PCR) in fibroblasts. To confirm the human genome U133A probe array (GeneChip; Affymetrix, Inc, Santa Clara, Calif) data, we additionally cultured postauricular skin–derived fibroblasts (SFs) and middle ear cholesteatoma–derived fibroblasts (MECFs) (some cultures stimulated with 10 ng/mL of interleukin [IL] 1α and/or IL-1β) from 5 other patients and determined the messenger RNA (mRNA) levels of liver and activation-regulated chemokine (LARC) (A), granulocyte-macrophage colony-stimulating factor (GM-CSF) (B), epiregulin (C), intercellular adhesion molecule 1 (ICAM-1) (D), and transforming growth factor α (TGF-α) (E) (extracted from Table 2) using real-time PCR analysis. The results are shown as mean ± SEM (*P<.05). The copy number is expressed as the number of transcripts. These data were divided by the expression level of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) for background correction.
Yoshikawa M, Kojima H, Wada K, Tsukidate T, Okada N, Saito H, Moriyama H. Identification of Specific Gene Expression Profiles in Fibroblasts Derived From Middle Ear Cholesteatoma. Arch Otolaryngol Head Neck Surg. 2006;132(7):734-742. doi:10.1001/archotol.132.7.734
To investigate the role of fibroblasts in the pathogenesis of cholesteatoma.
Tissue specimens were obtained from our patients. Middle ear cholesteatoma–derived fibroblasts (MECFs) and postauricular skin–derived fibroblasts (SFs) as controls were then cultured for a few weeks. These fibroblasts were stimulated with interleukin (IL) 1α and/or IL-1β before gene expression assays. We used the human genome U133A probe array (GeneChip) and real-time polymerase chain reaction to examine and compare the gene expression profiles of the MECFs and SFs.
Six patients who had undergone tympanoplasty.
The IL-1α–regulated genes were classified into 4 distinct clusters on the basis of profiles differentially regulated by SF and MECF using a hierarchical clustering analysis. The messenger RNA expressions of LARC (liver and activation-regulated chemokine), GMCSF (granulocyte-macrophage colony-stimulating factor), epiregulin, ICAM1 (intercellular adhesion molecule 1), and TGFA (transforming growth factor α) were more strongly up-regulated by IL-1α and/or IL-1β in MECF than in SF, suggesting that these fibroblasts derived from different tissues retained their typical gene expression profiles.
Fibroblasts may play a role in hyperkeratosis of middle ear cholesteatoma by releasing molecules involved in inflammation and epidermal growth. These fibroblasts may retain tissue-specific characteristics presumably controlled by epigenetic mechanisms.
Recently, the interaction between mesenchymal cells such as fibroblasts and epithelial cells or keratinocytes has been proposed to be involved in inflammation, homeostasis, and tissue regeneration.1- 5 For example, interleukin (IL) 1 produced by keratinocytes induces the release of keratinocyte growth factor (KGF), granulocyte-macrophage colony-stimulating factor (GM-CSF), and transforming growth factor α (TGF-α) from fibroblasts.6- 8 These fibroblast-derived cytokines support the proliferation and differentiation of keratinocytes. This paracrine loop is thought to be important in repair processes of tissue inflammation. In the study presented herein, we used the human genome U133A probe array (GeneChip; Affymetrix, Inc, Santa Clara, Calif) and real-time polymerase chain reaction (PCR) to examine the gene expression profiles of fibroblasts derived from middle ear cholesteatoma and postauricular skin obtained from patients to determine whether these fibroblasts express some molecules that may interact with keratinocytes and be involved in the pathogenesis of middle ear cholesteatoma.
Middle ear cholesteatoma and normal postauricular skin samples were obtained from patients undergoing tympanoplasty, and single fibroblasts from each sample were cultured in Dulbecco Modified Eagle Medium/F12 (DMEM/F12; Invitrogen Corp, Carlsbad, Calif) with 10% fetal calf serum (JRH Biosciences, Lenexa, Kan) and a combination of 60-μg/mL penicillin and 100-μg/mL streptomycin (Invitrogen Corp) for a few weeks. The cells were incubated at 37°C in a humidified incubator containing 5% carbon dioxide in air and were analyzed after 4 passages. Patient consent and the approval of our university's ethics review board were obtained before the start of the study.
Middle ear cholesteatoma fibroblasts (MECFs) and skin fibroblasts (SFs) were stimulated with 10 ng/mL of IL-1α or 10 ng/mL of IL-1β (both from R&D Systems Inc, Minneapolis, Minn) for 4 hours before messenger RNA (mRNA) extraction.
Human genome-wide gene expression was examined using the human genome U133A probe array (GeneChip), which contains the oligonucleotide probe set for approximately 22 000 full-length genes, according to the manufacturer's protocol and previously reported strategies.9- 12
Total RNA (5 μg) was extracted from the fibroblasts, and double-stranded complementary DNA (cDNA) was synthesized using a SuperScript Choice system (Invitrogen Corp) and a T7-(dT)24 primer (Amersham Pharmacia Biotech, Buckinghamshire, England). The cDNA was subjected to in vitro transcription in the presence of biotinylated nucleoside triphosphates using a high-yield RNA transcript labeling kit (BioArray; Enzo Diagnostics, Farmingdale, NY). The biotinylated complementary RNA was then hybridized with the probe array for 16 hours at 45°C. After washing, the hybridized biotinylated complementary RNA was stained with streptavidin-phycoerythrin (Molecular Probes, Eugene, Ore) and scanned with a gene array scanner (Hewlett-Packard Company, Palo Alto, Calif). The fluorescence intensity of each probe was quantified using the GeneChip Analysis Suite program, version 5.0 (Affymetrix, Inc). The expression level of single mRNA was determined as the mean fluorescence intensity among the intensities obtained with 11 paired (perfect-matched and single nucleotide–mismatched) probes. If the intensities of the mismatched probes were very high, gene expression was judged to be absent, even if a high mean fluorescence was obtained with the GeneChip Analysis Suite 5.0 program.
The resulting data were analyzed using GeneSpring software, version 7.2 (Silicon Genetics, San Carlos, Calif). To normalize the staining intensity variations among the chips, the values for all genes on a given chip were divided by the median of all measurements on that chip. To eliminate changes within the range of background noise and to select the most differentially expressed genes, data were used only if the raw data values were less than 100 and gene expression was judged to be present by a gene expression data analysis. Hierarchical clustering analysis with standard correlation was used to identify gene clusters.
To confirm the GeneChip microarray expression analysis data, we quantified the gene expression in fibroblasts derived from 5 other patients using quantitative real-time PCR. Total RNA was isolated using an RNA purification kit that included DNase digestion (RNeasy Mini Kit; Qiagen GmbH, Hilden, Germany). The RNA was then transcribed into cDNA using reverse transcriptase (Superscript II: Invitrogen Corp). Quantitative PCR was performed using a sequence detector system (ABI/PRISM 7700; Applied Biosystems, Foster City, Calif) and TaqMan Universal PCR Master Mix (Applied Biosystems), according to the manufacturers’ instructions. The primers and TaqMan probes used for the genes LARC (liver and activation-regulated chemokine), GMCSF (granulocyte-macrophage colony-stimulating factor), ICAM1 (intercellular adhesion molecule 1), and TGFA (transforming growth factor-α) were as follows: LARC: forward, 5′-TGTCAGTGCTGCTACTCCACCT-3′; reverse, 5′-CTGTGTATCCAAGACAGCAGTCAA-3′; and TaqMan probe, 5′-TGCGGCGAATCAGAAGCAGCAA-3′; GMCSF: forward, 5′-GCCTCACCAAGCTCAAGGG-3′; reverse, 5′-GGTTGGAGGGCAGTGCTG-3′; and TaqMan probe, 5′-CCCTTGACCATGATGGCCAGCC-3′; ICAM1: forward, 5′-CTGTGTCCCCCTCAAAAGTCA-3′; reverse, 5′-ATACACCTTCCGGTTGTTCCC-3′; and TaqMan probe, 5′-TGCGGCGAATCAGAAGCAGCAA-3′; TGFA: forward, 5′-AGGAGACCCCTGCCCTCTAGT-3′; reverse, 5′-TCTGCAATGTGTTCTTGGTTTTG-3′; and TaqMan probe, 5′-TTCCAACCTGCCCAGTCACAGAAGG-3′. For epiregulin, quantitative PCR was performed using SYBR green PCR master mix (Applied Biosystems). The primers for epiregulin were as follows: forward, 5′-ATCCTGGCATGTGCTAGGGT-3′ and reverse, 5′-GTGCTCCAGAGGTCAGCCAT-3′. The expression levels of mRNA were normalized by the mean expression of a housekeeping gene (glyceraldehyde-3-phosphate dehydrogenase [GAPDH]), which was measured using Pre-Developed Assay Reagents (Applied Biosystems).
Data are presented as mean ± SEM. Statistical significance was determined using the paired t test, and differences were considered significant at P<.05.
GeneChip was used to identify the gene expression pattern induced by IL-1α in the MECFs (n = 1) and SFs (n = 1). Following the identification of candidate disease–related genes with GeneChip, quantitative real-time PCR was used to confirm the expression of the selected gene (n = 5 for each).
To assign a “fold-change cutoff” threshold, we used a GeneSpring analysis for selected genes in which the mean expression level had increased or decreased by more than 2-fold after 4 hours. As shown in Figure 1 and Figure 2, IL-1α had an effect on numerous genes in the SF and MECF cultures.
Using hierarchical clustering analysis of the gene expression profiles of 22 283 genes, we identified a cluster containing genes that were up-regulated in SF or MECF after IL-1α stimulation (Figure 1). Visual inspection identified 2 major groups among the up-regulated genes. The first group of 164 genes displayed an increase in gene expression in SF that was more than 2-fold of that in MECF (Figure 1, Table 1), while the second group of 84 genes displayed an increase in gene expression in MECF that was more than 2-fold of that in SF (Figure 1, Table 2).
Among the genes in Table 1, several were identified whose expressions are generally known to be induced by IL-1α in fibroblasts, including monocyte chemoattractant protein 2 (MCP2); MCP3; granulocyte colony-stimulating factor (GCSF); regulated on activation normal T cell expressed and secreted (RANTES); matrix metalloproteinase 1 (MMP1); interferon-inducible protein 10 (IP10); IL-15 (IL15); vascular cell adhesion molecule 1 (VCAM1); KGF; eotaxin; and interferon-inducible T-cell α chemoattractant (ITAC). Genes for multiple profibrotic cytokines and chemokines that exhibited elevated expressions are shown in Table 2. These cytokines included TGFA, GMCSF, ICAM1, epiregulin, and LARC.
Figure 2 shows the genes that were down-regulated in SF or MECF after 4 hours of stimulation with IL-1α. Two major groups were identified among these genes: 65 genes showed a decrease in gene expression in SF that was less than half of that in MECF (Figure 2, Table 3), while the other 69 genes displayed a decrease in gene expression in MECF that was less than half of that in SF (Figure 2, Table 4).
We determined the mRNA levels using real-time PCR to confirm the GeneChip data of cluster B because it was thought that cluster B might contain some of the genes associated with the pathogenesis of middle ear cholesteatoma. Our results showed that the mRNA expression of LARC, GMCSF, epiregulin, ICAM1, and TGFA was significantly more strongly up-regulated by IL-1α and/or IL-1β in MECF than in SF (Figure 3).
This study is the first report, to our knowledge, of a global gene expression analysis used to identify specific genes in fibroblasts derived from middle ear cholesteatoma and normal postauricular skin in vitro. We identified many differences in the gene expression patterns of SFs and MECFs, although the fibroblasts were cultured and stimulated under the same conditions. Thus, SFs and MECFs appear to have different phenotypes because they were derived from a single genotype. Several reports13- 19 have stated that fibroblasts have different phenotypes in lesions and normal tissues. A recent report20 found that the cell-type–specific absence of CCCAT/enhancer-binding protein α (C/EBPα) was responsible for the enhanced proliferation of bronchial smooth muscle cells derived from subjects with asthma, explaining the failure of glucocorticoids to inhibit proliferation in vitro. Our findings suggest that tissue-derived cells may retain their characteristic features even when cultured for long periods. Consequently, these data suggest that fibroblasts derived from different tissues retain their typical gene expression profiles. Furthermore, the differing characteristics might be controlled by epigenetic mechanisms.
Real-time PCR analysis showed that the subepidermal fibroblasts obtained from the middle ear cholesteatoma produced much more GMCSF, epiregulin, and TGFA than did fibroblasts obtained from postauricular skin. Thus, activated MECFs may induce the exuberant growth of keratinocytes, resulting in the production of IL-1α and/or IL-1β in the injured and/or infected tissues. Transforming growth factor α is thought to be the main growth factor influencing keratinocytes via the previously mentioned paracrine loop. Many reports on TGF-α expression in cholesteatoma and the autocrine mechanism of TGF-α have been published.
21- 25 However, this is the first report, to our knowledge, that describes the expression of epiregulin in middle ear cholesteatoma. Epiregulin has been purified from the conditioned media of a mouse fibroblast-derived tumor cell line, NIH3T3/clone T7, and is a member of the epidermal growth factor (EGF) family.26 In recent studies,27,28 epiregulin was shown to be an autocrine growth factor in normal human keratinocytes, organizing the epidermal structure by regulating keratinocyte proliferation and differentiation, as well as the expression of TGF-α, heparin-binding–EGF, and amphiregulin. Consequently, a tendency for these growth factors to be expressed may be the origin of middle ear cholesteatoma. However, many other growth factors actually participate in the growth of keratinocytes. For example, the mRNA expression of KGF was enhanced in SFs after stimulation with IL-1α and/or IL-1β (data not shown). Our results contradict those for GM-CSF, epiregulin, and TGF-α, and future study is needed to determine the cause of this discrepancy.
In addition, the mRNA expressions of LARC and ICAM-1 were more strongly up-regulated in MECF than in SF. The chemokine LARC is thought to contribute to the initiation of the immune response of T lymphocytes during the early phase of inflammation because it promotes the migration of immature dendritic cells and memory T lymphocytes to the area of local inflammation.29- 32 In other words, compared with SFs, MECFs allow more CCR6-positive cells, such as immature dendritic cells and memory T lymphocytes, to accumulate during the early phase of inflammation triggered by a foreign antigen. Thus, the phenotype of activated fibroblasts residing in a given tissue may directly influence the nature and magnitude of leukocyte recruitment. For example, LARC has been previously shown to be related to the onset of rheumatoid arthritis.33,34 However, further examination of middle ear cholesteatoma is needed to determine whether fibroblasts are the main producers of LARC. The adhesion molecule ICAM-1 belongs to an immunoglobulin superfamily and is expressed in endothelial cells. ICAM-1 is mainly responsible for the migration of white blood cells to areas of inflammation. In middle ear cholesteatoma, ICAM-1 was shown by immunohistochemical analysis to be present in areas of inflammatory change.35- 37 The significance of ICAM-1 being expressed on fibroblasts is that the fibroblasts then adhere to inflammatory cells, which migrate from blood vessels and remain in tissue. Consequently, local inflammation depending on the activation of inflammatory cells may be initiated by the binding of inflammatory cells to integrin, such as lymphocyte function–associated antigen 1 (LFA-1) and Mac-1 (CD11b/CD18), on cell surfaces. The present data for LARC and ICAM-1 suggest that MECFs may be able to evoke inflammation and contribute to its persistence more easily than SFs.
Because these genes were differentially expressed in IL-1–stimulated MECFs but not in SFs, they seem to be related specifically to local immunity, such as the prominent hyperkeratosis seen in middle ear cholesteatoma. However, because these results were obtained with the use of cultured cells, further investigation is needed to elucidate how they actually contribute to the pathogenesis of middle ear cholesteatoma.
We identified many differences in the gene expression patterns of SFs and MECFs, although the fibroblasts were cultured and stimulated under the same conditions. Thus, SFs and MECFs appear to have different phenotypes, since they were derived from a single genotype. These differential gene expressions suggest that subepidermal fibroblasts may play a role in the hyperkeratosis that occurs during middle ear cholesteatoma by releasing molecules involved in inflammation and epidermal growth; they also suggest that these fibroblasts retain tissue-specific characteristics that are presumably controlled by epigenetic mechanisms. These results may contribute to our understanding of the pathogenesis of middle ear cholesteatoma.
Correspondence: Mamoru Yoshikawa, MD, PhD, Department of Otorhinolaryngology, Jikei University School of Medicine, 3-25-8 Nishi-Shimbashi, Minato-ku, Tokyo 105-8461, Japan (firstname.lastname@example.org).
Submitted for Publication: December 3, 2005; final revision received February 22, 2006; accepted March 15, 2006.
Financial Disclosure: None reported.
Funding/Support: This study was supported by Grants-in-Aid from the National Institute of Biomedical Innovation, Osaka, Japan.
Acknowledgment: We thank Noriko Hashimoto, of the National Research Institute for Child Health and Development, for her skillful technical assistance.