Experimental and Molecular Pathology 80 (2006) 147 – 159 www.elsevier.com/locate/yexmp
Increased EMMPRIN (CD 147) expression during oral carcinogenesis Nadarajah Vigneswaran a,⁎, Simone Beckers b , Sabine Waigel b , John Mensah b , Jean Wu a , Juan Mo c , Kenneth E. Fleisher d , Jerry Bouquot a , Peter G. Sacks c , Wolfgang Zacharias b a
b
Department of Diagnostic Sciences, The University of Texas Health Science Center at Houston, Dental Branch, 6516 M.D. Anderson Blvd., Room 3.094G, Houston, TX 77030, USA Departments of Medicine, Pharmacology and Toxicology, James Graham Brown Cancer Center, University of Louisville, Louisville, KY 40202, USA c Department of Basic Science and Craniofacial Biology, New York University College of Dentistry, New York, NY 10010, USA d Department of Oral and Maxillofacial Surgery, New York University College of Dentistry, New York, NY 10010, USA Received 12 September 2005 Available online 28 November 2005
Abstract Gene expression profiling of oral premalignant (OPM) cells and normal oral epithelial (NOR) cells showed that EMMPRIN expression was markedly upregulated in OPM cells compared to NOR cells. We used an oral squamous cell carcinoma (OSCC) progression model composed of cell lines, organotypic cultures and tissue specimens to characterize EMMPRIN expression patterns by microarray analysis, qRT-PCR, Western blotting and immunohistochemistry. EMMPRIN levels are elevated in OPM and primary and metastatic OSCC cells as compared to NOR. EMMPRIN was detected as high and low glycosylated forms in the OPM and OSCC cellular extracts and was released in the media by OSCC cells but not by OPM cells. EMMPRIN expression in an organotypic culture model of normal and OPM mucosae mirrored the expression patterns in the respective tissues in vivo. EMMPRIN expression was limited to basal cells of normal, benign hyperkeratotic and inflammatory (lichen planus) oral mucosa. EMMPRIN expression is increased in dysplastic leukoplakias spreading to more superficial layers, and its expression levels correlated significantly with the degree of dysplasia. Primary and metastatic OSCC showed strong cell surface expression of EMMPRIN. These results suggest that EMMPRIN overexpression occurs at a very early stage of oral carcinogenesis and plays a contributing role in OSCC tumorigenesis. © 2005 Elsevier Inc. All rights reserved. Keywords: Microarray analysis; EMMPRIN; Oral cancer; Organotypic culture; Leukoplakia
Introduction Oral squamous cell carcinoma (OSCC) is the most common cancer detected in the head and neck region and ranks among the top ten most frequently diagnosed cancers worldwide with an estimated 500,000 patients affected annually worldwide (Parkin et al., 1999; Pisani et al., 1999). Moreover, OSCC incidence is increasing in most developing countries, especially among younger persons (Macfarlane et al., 1994). Despite Abbreviations: EMMPRIN, extracellular matrix metalloproteinase inducer; MMP, matrix metalloproteinases; NHEK, normal epidermal keratinocytes; NOR, normal oral epithelial cells; OED, oral epithelial dysplasia; OPM, oral premalignancy; OSCC, oral squamous cell carcinoma. ⁎ Corresponding author. Fax: +1 713 500 4416. E-mail address:
[email protected] (N. Vigneswaran). 0014-4800/$ - see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.yexmp.2005.09.011
remarkable recent advances in cancer treatment, the long-term survival of patients with OSCC has only marginally improved over the past three decades, with only 50% of OSCC patients surviving 5 years after diagnosis (Forastiere et al., 2001). Early diagnosis and treatment appear to be the most powerful means by which to improve prognosis, especially if the diagnosis can be made in the preinvasive, intraepithelial stage (Boyle et al., 1993; O’Shaughnessy et al., 2002; Lippman et al., 2005). Development of OSCC is a multistep process of genetic and epigenetic changes that typically manifest clinically in two stages: (1) an oral premalignant (OPM) lesion and (2) progression or transformation to OSCC (Braakhuis et al., 2004; Kim and Califano, 2004; Lippman et al., 2005). OPM lesions are readily identifiable clinically and usually present as white patches (leukoplakia) or as red patch (erythroplakia) (Neville et al., 2002). Leukoplakia is the most common,
148
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
representing 85% of such lesions and showing a prevalence rate of 2–8% in people over the age of 70 (Neville et al., 2002; Neville and Day, 2002). Since only up to 20% of oral leukoplakias progress to malignancy (Neville et al., 2002), a major challenge is to accurately measure risk of OSCC development in individual leukoplakias so that they may be treated accordingly. If the disease can be halted at this stage, other benefits accrue. Patients with OSCC have a significantly increased risk of developing subsequent second primary or recurrent OSCC (Kopelovich et al., 1999). Recent genetic studies indicate that the second primary OSCC shows similar genetic alterations as the original tumor and thus appears to be derived from the spread of the original OPM clone (Braakhuis et al., 2003), perhaps because a ‘field’ of premalignant clones remained after resection of the primary tumor (Tabor et al., 2001). Along a similar fashion, OPM lesions which have the potential to develop into new tumors persist in one fourth of the OSCC patients with cancer-free surgical margins (Weijers et al., 2002). It would be ideal to evaluate these for markers of high risk since so many of these lesions have extensive involvement. Malignancy risk of oral leukoplakia is currently assessed by the presence and degree of oral epithelial dysplasia (OED). Histologic grading classifies oral leukoplakias into stages with increasing risk of malignant transformation, namely: epithelial hyperkeratosis and hyperplasia; mild, moderate and severe OED; and carcinoma-in-situ. Among the leukoplakias with OED, a substantial portion (15–20%) progress to OSCC with a constant rate of 2–3% per year (Lumerman et al., 1995; Schepman et al., 1999). Although the severity of OED is known to be proportional to the risk of subsequent OSCC development, histologic grading of OED is an imprecise art, and considerable improvement is needed for predicting OSCC risk for the individual lesions (Sudbo et al., 2001a,b). The ability to change our current paradigms of diagnosis and treatment of oral leukoplakias depends in large part on the development of refined and clinically significant molecular risk markers that can assist the microscopic grading of OED (O’Shaughnessy et al., 2002; Lippman et al., 2005). In order to identify gene products that can be used as molecular risk markers for OPM lesions, we used oligonucleotide microarrays for global gene expression profiling of normal oral epithelial cells and OPM cell lines. It should be noted that these OPM cell lines were derived from leukoplakias adjacent to existing OSCC and therefore have proven malignancy risk (Bouquot, 1999). Our data showed that expression of EMMPRIN (extracellular matrix metalloproteinase inducer) was markedly upregulated in OPM-derived cells compared to NOR. EMMPRIN (also known as CD147 and M6 in humans; basigin and gp42 in mouse) was originally isolated from the plasma membranes of malignant tumor cells (Biswas, 1982; Biswas, 1984). EMMPRIN has recently been recognized as an important modulator of tumor–stroma cross talk and mediates a wide range of tumor-promoting molecular events that include acquisition of anchorage-independent growth and invasive phenotype by tumor cells, as well as tumor-cell-induced angiogenesis (Yan et al., 2005). This critical role during early tumorigenesis
prompted us to focus on EMMPRIN as potential molecular diagnostic marker and therapeutic target for early diagnosis and treatment of OPM lesions. We used microarray RNA expression profiling, qRT-PCR, Western blotting and immunohistochemistry to characterize EMMPRIN expression patterns in an OSCC tumor progression model composed of differently staged cell lines, in vitro 3dimensional organotypic culture models of normal oral mucosa and OPM lesion, and archival tissue specimens. Our data show that EMMPRIN transcript levels are markedly upregulated in cell lines derived from OED, primary OSCC and metastatic OSCC compared to normal human oral epithelial and epidermal keratinocytes. Our data show that upregulation of EMMPRIN expression occurs during the early stage of OPM lesions and appears to be a critical phenotypic alteration that precedes the development of OSCC. Materials and methods Cell lines, organotypic culture and tissue specimens All protocols for the use of human cell lines and tissue specimens in this work were approved by the Institutional Review Boards of The University of Louisville and the University of Texas at Houston. Primary cultures of NOR were prepared from discarded human gingival tissue. Normal human epidermal keratinocytes (NHEK) were obtained from Cambrex BioScience (Walkersville, MD). The Leuk1 cell line was derived from a dysplastic leukoplakia adjacent an early invasive OSCC (T1N0M0) involving the tongue of a 47-year-old female. The Leuk2 cell line was derived from dysplastic leukoplakia in a 72-year-old female with a history of recurrent new disease. Both of these OPM cell lines (Leuk1 and Leuk2) show loss of response to calcium and defective terminal differentiation which are characteristics of premalignant keratinocytes (Sacks, 1996). The source of patient-matched primary (686Tu and 1386Tu) and lymph node metastatic oral cancer cell lines (686Ln and 1386Ln) has been described elsewhere (Vigneswaran et al., 2005). The NOR, NHEK and Leuk cells were grown in serum-free keratinocyte growth media (KGM-2) supplemented with growth factor bullit kit (Cambrex). The 686 and 1386 cell line pairs were maintained in DMEM/F12 50/50 mix (Cambrex) containing 10% fetal bovine serum, 0.4 μg/ml hydrocortisone and penicillin–streptomycin–amphotericin antibiotic mix (Cambrex) at 37°C with 5% CO2. For each cell line, approximately the same (±2) passages were used for microarray, qRT-PCR, Western blots and immunohistochemical analyses. Primary cultures of NOR and Leuk1 cells were used to establish a threedimensional organotypic co-culture model of homotypic normal oral mucosa, OPM mucosa and heterotopic combination of normal and OMP mucosa. Oral mucosal tissue-like organotypic culture was established by growing NOR and Leuk1 cells on a human oral fibroblast embedded collagen gel at the liquid–air interface on Gelfoam. Each organotypic culture was fixed in formalin and processed for embedding in paraffin. Three-micrometer sections were cut and processed for routine histologic and immunohistochemical staining for EMMPRIN. We used archival tissue specimens of leukoplakia exhibiting hyperkeratosis (n = 27), mild OED (n = 17), moderate OED (n = 18) and severe OED/ carcinoma-in-situ (n = 17) to examine the EMMPRIN expression pattern by immunohistochemistry. In addition, we examined the EMMPRIN expression in proliferative verrucous leukoplakia (n = 14), lichen planus (n = 10), as well as in primary OSCC (n = 20) and metastatic OSCC (n = 17). Five-micrometer sections were cut from each block and processed for H&E staining and immunohistochemical detection of EMMPRIN.
Microarray analysis Each cell line was grown in 100 mm dishes as two independent biological replicates. Total cellular RNA was extracted from 75 to 80% confluent cultures,
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159 and the quantitation and integrity tests were performed using the Bioanalyzer Nano-Chip (Agilent, Foster City, CA) and by UV absorbance (GeneQuant II, Pharmacia Biotech). All samples showed the desired 18S and 28S ribosomal peak; no degradation was detected. The standard protocols and reagents recommended by Affymetrix were used for cDNA synthesis, in vitro transcription and biotin labeling and cRNA fragmentation. For cDNA synthesis, the Superscript cDNA Synthesis kit (Invitrogen, Carlsbad, CA) and an oligo-dTT7 promoter primer were used. Biotin-labeled cRNA was transcribed from the T7-primed cDNA pool through in vitro transcription using the BioArray High Yield RNA Transcript Labeling kit (Enzo Diagnostics, Farmingdale, NY). RNA expression profiles were collected after hybridization to Affymetrix HG-U133A arrays, which interrogate 22,500 known and annotated human genes. After applying the antibody amplification protocol for signal enhancement, stained arrays were scanned with an Agilent GeneArray Scanner 2500 (Agilent, Palo Alto, CA). All absolute analysis data were uniformly scaled to a target intensity of 150 for comparisons across samples using Affymetrix MAS v5.0 software. All array data report files showed acceptable 3′/5′ signal ratios for GAPDH and β-actin as internal housekeeping controls and showed comparable values for “Noise”, “Scale factor” and “Number of genes present”. In the absolute analysis of CHP files, all genes with absent (A) calls in all samples were eliminated. Only genes with detection P values of b0.065 in at least 1 of the samples and change P values of b0.006 in all pairwise comparisons were included. Data files were further imported into GeneSpring v7.0 (Silicon Genetics, Redwood City, CA) and Partek Pro v6.0 (Partek, St. Charles, MO) data analysis software to study similarities and behavior of the different samples. Two-way ANOVA, false discovery rate (FDR) and principle component analysis were performed with Partek Pro v6.0. Spearman correlation clustering (heatmaps) and hierarchical clustering analyses were performed with GeneSpring v7.0.
Quantitative real-time RT-PCR Aliquots of the same cellular RNAs used for microarray studies were reverse transcribed in duplicates using Multiscribe reverse transcriptase (Applied Biosystems, Foster City, CA). Real-time quantitative PCR was performed with SybrGreen PCR Master Mix (Applied Biosystems) on an ABI Prism 7000 Sequence Detection System. Gene-specific primer sequences for EMMPRIN (GenBank accession number NM_001728) were obtained from the PrimerBank website (http://pga.mgh.harvard.edu/primerbank): forward, AGTGGTGGTTTGAAGGGCAG; reverse, CACGAGCGTGTCGATGGAG; amplicon size 134 bp. Primers for human β-actin internal control (GenBank accession number BC002409) were: forward, GGATGCAGAAGGAGATCACTG; reverse, CGATCCACACGGAGTACTTG; amplicon size 150 bp. Primers were commercially synthesized (Invitrogen, Carlsbad, CA). Each 25-μl PCR reaction contained 2 μl of cDNA template and 3 μl of forward and reverse primer (0.3 μM each) in SYBR® Green PCR Master Mix. The following amplification conditions were used: one step denaturation 95°C/ 10 min, then 40 cycles of denaturation 95°C/15 s, annealing and extension 60°C/ 1 min. All sample data were normalized against β-actin, and the mean fold changes for EMMPRIN RNA levels were calculated for EMMPRIN relative to β-actin using the 2−ΔΔCT method (Livak and Schmittgen, 2001). Standard deviation calculations were performed using the Comparative Method (Applied Biosystems User Bulletin #2).
Western blotting Cells grown in 100-mm plates to 75–80% confluence and secreted proteins were obtained by replacing the regular media with 2 ml of serum-free media and incubation overnight. The next day, the conditioned media were aspirated and concentrated using the iCON protein concentrators (PIERCE, Rockford, IL, USA). Cellular proteins were extracted from the same cells using the M-PER™ Protein Extraction Reagent (PIERCE, Rockford, IL, USA). Protein concentrations were determined using the Protein dotMETRIC™ assay kit (Geno Technology, Inc, St. Louis, MO, USA). Equal amounts of media and cellular proteins (25 μg) were separated by 10% SDS-PAGE and transferred to nitrocellulose membrane (Novex, San Diego, CA). The membranes were incubated in a blocking solution consisting of 5% milk powder in 10 mM Tris– HCl (pH 8.0), 150 mM NaCl and 0.1% Tween 20 at room temperature for 1
149
h then immunoblotted with polyclonal rabbit anti-human EMMPRIN (1:700, Santa Cruz Biotechnology, Santa Cruz, CA) or monoclonal mouse anti-β-actin (Sigma-Aldrich Co, St. Louis, MO) antibody followed by horseradishperoxidase-conjugated secondary antibody and chemiluminescence detection.
Immunohistochemistry NHEK, Leuk1, Leuk2 and 686Tu cells were grown in microscope slide chambers to 50–60% confluence, washed in PBS and fixed in 4% paraformaldehyde for immunohistochemical and immunofluorescence detection of EMMPRIN. For immunofluorescence staining, Leuk1 and 686Tu cells were probed with rabbit anti-human EMMPRIN antibody (dilution = 1:250) followed by biotinylated goat–anti-rabbit IgG and streptavidin–FITC. Cells were counterstained with propidium iodide. Three-dimensional images of EMMPRIN staining patterns of Leuk1 and 686Tu cells were obtained using laser scanning confocal microscopic imaging system with appropriate filters. Formalin-fixed paraffin-embedded tissue sections of organotypic cultures, leukoplakias, lichen planus and OSCC were deparaffinized in xylene, rehydrated in graded ethanol and used for immunohistochemical staining. Tissue sections were subjected to antigen retrieval by heating in target antigen retrieval solution (DAKO Corporation Carpentaria, CA, USA). Immunoreactivity for EMMPRIN was detected using the rabbit anti-human EMMPRIN antibody (1:250) and visualized using the streptavidin–biotin–HRP detection method. Immunostained sections were evaluated and graded without prior knowledge of the clinicopathologic details. A reproducible semiquantitative analysis of IHC staining was used to evaluate the expression level of EMMPRIN in tissue samples (Vigneswaran et al., 2000). The sections were initially scanned at low power, and areas with high and low staining intensity were selected. At each selected field, cells were classified with respect to staining intensity ranging from 0 to 4, and the percentage of cells for each staining intensity was estimated. A numerical value for EMMPRIN staining in each tissue section was then calculated by multiplying the fraction of cells (decimal equivalent of percentage of cells) at each staining intensity by the numerical value of that intensity, resulting in scores ranging from 0 to 4. Immunohistochemical data are presented as mean ± SEM and were analyzed by ANOVA and Student’s t test.
Results Global gene expression analysis revealed elevated expression of EMMPRIN in OED and OSCC cell lines We performed a global analysis of gene expression of 22,500 genes in a panel of cell lines to identify genes that are differentially expressed during premalignant transformation of oral epithelial cells. These cell lines represented different phenotypes ranging from normal (NOR, NHEK) to premalignant (Leuk1, Leuk2) to primary OSCC (686Tu, 1386Tu) and to metastatic OSCC (686Ln, 1386Ln). While examining the gene expression profiles of these cells, we detected markedly upregulated expression of EMMPRIN in OPM cells compared to normal oral and epidermal keratinocytes (Fig. 1A). Interestingly, the increase was most pronounced in the OPM cell lines (Leuk1 and Leuk2) followed by the metastatic OSCC lines (686Ln and 1386Ln) and then in the primary tumor lines (686Tu and 1386Tu) (Fig. 1B). Interrogating the array signals for phenotype-specific associations by principal component analysis placed the different phenotypes into distinct components based on their EMMPRIN signal intensities and clearly separated the normal from dysplastic from malignant cells (Fig. 1C). For validation of the microarray data, we obtained good correlations between qRT-PCR and array data for the majority of samples (Fig. 2). Although for a few of the samples there
150
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
Fig. 1. Microarray-based EMMPRIN RNA expression levels in oral cell lines. (A) Signal intensities for EMMPRIN were averaged across all replicates of all cell lines, and individual differences presented as fold change relative to this average (displayed as a color-coded intensity). (B) Microarray data for EMMPRIN, averaged over all samples within each phenotype/malignancy (normal 6×, dysplasia 4×, primary tumor 4×, metastasis 4× samples). (C) Principle component analysis based on EMMPRIN signal intensities. The four different phenotypes displayed (magenta = normal; red = dysplasia/OPM; blue = primary (Tu); green = metastasis (Ln)) cluster as distinct and not overlapping components, indicating that EMMPRIN signals change characteristically between phenotypes. Dysplasia and Tu cells have the highest consistency within their groups, whereas normal and Ln cells have a broader range of data points. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
were deviations in the quantitative fold changes obtained by the two methods, all qRT-PCR data confirmed the increased EMMPRIN transcript levels in OPM and OSCC cells compared to normal cells. Such occasional deviations between microarray and PCR data are most likely due to the different quantitation principles of the two techniques since array quantitation uses the average of multiple gene-specific probe hybridizations, whereas PCR uses only one specific primer pair for amplification. Therefore, we performed qRT-PCR analyses in two independent experiments for each sample in triplicate wells. The qRT-PCR data, obtained after normalization to the internal β-actin control
signals and averaging over six data sets (triplicate wells ×2), correlated well with the respective microarray data for most of the cell lines with the exception of 1386Tu and -Ln pairs (Fig. 2). This confirmed the conclusion that there is a malignancydependent enhancement of EMMRIN RNA expression in this panel of 8 different cell lines. EMMPRIN protein expression in OPM and OSCC cell lines EMMPRIN protein expression levels in OPM and OSCC cell lines were analyzed by Western blotting and
Fig. 2. Comparison of EMMPRIN expression levels determined by microarray analysis and quantitative real-time RT-PCR. Since several different levels of normalizations and error calculations were necessary for each of the displayed values, error bars are not included, but the mean and range of data for each are listed below the corresponding bars. Both microarray and qRT-PCR data show a gradual increase in EMMPRIN transcript levels in OPM cell lines (Leuk1 and Leuk2) and primary (686Tu and 1386Tu) and metastatic (686Ln and 1386Ln) OSCC cell lines compared to normal human oral (NOR) and epidermal (NHEK) keratinocytes.
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
Fig. 3. Western blotting analysis of EMMPRIN glycoprotein in cell lysate (top; Extract) and concentrated media protein (bottom; Media). Cell lysates and media proteins were loaded and probed with polyclonal rabbit anti-human EMMPRIN antibody. For cell lysate proteins, reprobing for β-actin was used as loading control. Lanes #1: Leuk1; #2: Leuk2; #3: 686Tu; #4: 1386Tu; #5: 686Ln; #6: 1386Ln. HG: highly glycosylated form (∼45–65 kDa); LG: less glycosylated form (∼44–32 kDa); CP: core protein (27 kDa).
immunohistochemistry. EMMPRIN is a variably glycosylated protein with a molecular weight of 32–60 kDa, depending on the glycosylation of the core protein (27 kDa). Total cell lysates of all cell lines demonstrated the expression of EMMPRIN protein at various levels (Fig. 3). The highest level of EMMPRIN protein was noted in OPM cells followed by metastatic OSCC and then primary OSCC cells in accordance with the microarray and qRT-PCR data. EMMPRIN exists in tumor and transformed cell lines as a highly glycosylated (HG) form migrating at ∼45–65 kDa and as a less glycosylated (LG) form migrating at ∼44–32 kDa (Tang et al., 2004a,b; Tang and Hemler, 2004). Total cell lysates of OPM and primary and metastatic OSCC cell lines contained both HG and LG forms of EMPPRIN. Both OPM and metastatic OSCC cells produced predominantly the HG form of EMMPRIN, and only a minor portion represented the LG form. However, substantial amounts of EMMPRIN in LG form relative to HG form were found in the primary OSCC cells (686Tu and 1386Tu) compared to their respective metastatic cells. Furthermore, EMMPRIN in the HG form
151
was detected in the conditioned media of primary and metastatic OSCC cell lines, but interestingly not in the media of either OPM cell line. A recent report concluded that full-length EMMPRIN is released from tumor cells by microvesicle shedding (Sidhu et al., 2004). Therefore, we examined whether or not there are any distinguishing characteristics in the expression pattern of EMMPRIN between OPM cells (Leuk1) and primary OSCC cells (686Tu) using three-dimensional immunofluorescence microscopic imagining. Our data showed that EMMPRIN was expressed in the cytoplasm and on the cell surface of both OPM and OSCC cells (Fig. 4). Moreover, focal plaque-like augmented staining for EMMPRIN was noted on the invasive pseudopodia of 686Tu cells but not in Leuk1 cells, suggesting early EMMPRIN-specific microvesicle formations only in OSCC cells (Fig. 4) Immunohistochemical examination revealed strong cell surface and peri-cytoplasmic immunoreactivity for EMMPRIN in both OPM cell lines Leuk1 and Leuk2, whereas NHEK cells showed only a weak cytoplasmic staining (Fig. 5). EMMPRIN expression in organotypic culture models of normal oral mucosa and oral epithelial dysplasia Organotypic cultures are three-dimensional tissue cultures used to reconstruct a tissue/organ in vitro that closely resemble the original tissue/organ in vivo (Chinnathambi et al., 2003). Organotypic cultures are commonly used to investigate cell behavior, differentiation, drug effects and cell–matrix interactions in vitro. We used NOR and OPM (Leuk1) cells to establish organotypic cultures of normal and premalignant oral mucosae and examined the differential expression patterns of EMMPRIN in these models (Fig. 6). In homotypic normal oral mucosal organotypic culture, we found that EMMPRIN expression is restricted to actively differentiating basal cells, mimicking normal oral mucosa in vivo (Fig. 6B). This model also exhibited a band-like EMMPRIN staining in an area that corresponds to the basement membrane zone in vivo (Fig. 6B). We detected strong EMMPRIN expression in all layers of the homotypic
Fig. 4. Immunofluorescence staining for EMMPRIN in oral premalignant and primary OSCC cell lines. Three-dimensional immunofluorescence images show intense cell surface and cytoplasmic staining for EMMPRIN in both Leuk1 (A) and 686Tu (B) cells. However, only 686Tu cells demonstrate microvesicle-like EMMPRINpositive structures (arrows) on their invasive pseudopodia surface. Cells were counterstained with propidium iodide which binds the nuclear DNA and produces red fluorescence (×400). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
152
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
Fig. 5. Immunohistochemical localization of EMMPRIN in normal human epidermal keratinocytes (NHEK) and oral premalignant cell lines. Cell surface and pericytoplasmic localization of EMMPRIN was noted in both Leuk1 and Leuk2 cells. Isolated NHEK cells showed weak cytoplasmic positivity for EMMPRIN.
OPM organotypic culture but no band-like EMMPRIN staining beneath the basal cells (Fig. 6D). The EMMPRIN expression pattern at the junction of the normal oral mucosa and OPM in a heterotypic organotypic confrontational model established with normal and OPM mucosa on each side was also examined (Fig. 6E, F). Interestingly, EMMPRIN expression remained strong in OPM mucosa which included a region of OPM cells that expanded laterally beneath the adjacent normal oral mucosa Fig. 6F). Significantly, EMMPRIN expression was not upregulated in NOR in direct contact with OPM cells (Fig. 6F).
EMMPRIN expression in benign, premalignant and malignant lesions of oral mucosa Immunohistochemical examination of tissue sections showed that EMMPRIN expression was limited to the actively differentiating basal cells of normal keratinized and nonkeratinized oral mucosa (score range = 0.3–0.4). In addition, endothelial cells, salivary gland ductal cells and striated muscle were positive for EMMPRIN. Leukoplakias showing hyperkeratosis ± epithelial hyperplasia (benign; n = 27)
Fig. 6. EMMPRIN expression pattern in organotypic cultures. Histologic and immunohistochemical staining was done on three-dimensional organotypic culture models of normal (A and B) and premalignant (C and D) oral mucosa and their co-culture (E and F). EMMPRIN expression is mostly restricted to the basal layer of the normal oral mucosal organotypic culture, whereas it is expressed in all layers of oral premalignant organotypic culture. A prominent band like EMMPRIN reactivity (arrow) was noted at the epithelial–connective tissue interface of normal oral mucosal models but not in the premalignant mucosal model. In the co-culture model, premalignant epithelium confronts and grows underneath the normal epithelium (arrows). However, EMMPRIN expression level remains unchanged in normal epithelial cell that are in direct contact with premalignant cells. Left panel: H&E; right panel: EMMPRIN staining (×200).
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
153
Fig. 7. Semiquantitative analyses of EMMPRIN expression levels in various benign, premalignant and malignant oral lesions. Tissue sections of these specimens were immunohistochemically stained for EMMPRIN, and expression levels were graded using a semiquantitative analysis of immunohistochemical staining as described in the Materials and methods section. Scores are shown for Keratosis: oral leukoplakias without OED; Mild: mild OED; Moderate: moderate OED; Severe: severe OED; PVL: proliferative verrucous leukoplakia; SCC-Iry: primary OSCC; SCC-Met: metastatic OSCC specimens; LP: lichen planus. n = number of cases examined. * = differences were statistically significant.
revealed moderate to strong EMMPRIN staining mostly restricted to the basal and parabasal layers (score range = 0. 4–1.0) (Figs. 7 and 8). Lichen planus is a chronic inflammatory and hyperkeratotic condition of unknown etiology which commonly presents as intraoral white keratosis with a very low malignant transformation risk (b2%) (van der Meij et al., 2003) compared to oral leukoplakias (b20%). In lichen planus (n = 10) EMMPRIN expression was restricted to the basal layer (score range = 0.45–1.2) (Figs. 7 and 8). More importantly,
EMMPRIN expression remained the same, limited to the basal cells, in areas of reactive atypia found in lichen planus. Relative to OED, EMMPRIN expression gradually increased to higher epithelial layers with increasing grades of OED: mild OED (n = 17; score range = 0.5–1.7); moderate OED (n = 18; score range = 0.95–2.5); severe OED/carcinoma-in-situ (n = 17; score range = 1.1–3.6) (Figs. 7 and 9). Since it is difficult to distinguish between severe dysplasia and carcinoma-in-situ (Lippman et al., 2005), we combined these two pathologic entities together in our data analysis. It is noteworthy that although EMPPRIN expression overall became more pronounced with increasing OED grades in oral leukoplakias there are remarkable variations in its expression levels within each of these OED grades (Fig. 7). Increased expression of EMMPRIN was also noted in primary OSCC (n = 20; score range = 1.8–3.5) and metastatic OSCC (n = 17; score range = 1.5–3.5) specimens (Figs. 7 and 10). Strong EMMPRIN expression was noted in N90% of the tumor cells of carcinomas-in-situ and early invasive (micro-invasive) squamous cell carcinomas (Fig. 10). Moreover, epithelium adjacent to OSCC showed an increasingly strong expression of EMMPRIN with decreasing distance from the invasive tumor, even in the absence of histologically evident dysplasia (Fig. 10). Submucosal and peritumoral fibroblasts and inflammatory cells adjacent to OED and OSCC also expressed EMMPRIN. In more advanced OSCC, however, EMPPRIN expression was restricted mostly to the invasive fronts of the tumor, while the more differentiated areas in the center of the tumor remained negative or only weakly positive (Fig. 10). Discussion
Fig. 8. Immunohistochemical staining for EMMPRIN in frictional keratosis without OED (Keratosis) and lichen planus (LP). EMMPRIN expression is mostly localized to the basal epithelial layer of both keratosis and lichen planus. Chronic inflammatory cells adjacent to the basal layer of the lichen planus (arrow) are also positive for EMMPRIN. Left panel: H&E; right panel: EMMPRIN staining (×200).
EMMPRIN: a potential molecular risk marker for OSCC development in OPM lesions OPM lesions, especially leukoplakia, represent an excellent model system to study molecular markers of cancer risk assessment and response to therapy for the following reasons:
154
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
Fig. 9. Immunohistochemical staining for EMMPRIN in oral leukoplakias with mild, moderate and severe degrees of OED. EMMPRIN expression gradually increases with increasing grade of OED. Left panel: H&E; right panel: EMMPRIN IHC (×200).
Fig. 10. EMMPRIN expression in oral squamous cell carcinoma (OSCC) specimens. EMMPRIN is strongly positive in most tumor cells of carcinoma-in-situ and early invasive OSCC. Mucosa adjacent to the microinvasive OSCC (arrow) demonstrates elevated EMMPRIN expression without any significant epithelial dysplasia as shown in the adjacent H&E-stained section. In advanced OSCC, EMMPRIN expression is noted predominantly in tumor cells found at the invasive front of the tumors, whereas tumor cells found within the center of the tumor are mostly negative. Left panel: H&E; right panel: EMMPRIN IHC (×200; ×100).
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
(1) these lesions are easy to monitor and biopsy compared to the precancers of most other anatomic sites; (2) several follow-up studies have documented the risk of subsequent OSCC development in these lesions; (3) most of these lesions are clearly related to tobacco exposure, which is the primary risk factor for OSCC development. Because of this, recent studies have focused on the identification of molecular markers in oral leukoplakia, which might aid in the recognition of those lesions with the highest risk of malignant transformation (Lee et al., 2000; Banerjee et al., 2003; Hasina et al., 2003). Among the numerous molecular markers evaluated, the DNA content (aneuploidy) and loss of heterozygosity (LOH) containing known or presumptive tumor suppressor genes (Rosin et al., 2000; Sudbo et al., 2001a,b; Tabor et al., 2003) have shown the most consistent reliability in predicting cancer risk as concluded by a recent WHO consensus conference on oral precancers (Odell and Johnson, 2005). Identifying high-grade preinvasive OPM lesion based upon DNA content and/or LOH is, however, a technically complex procedure and is highly dependent upon the skills and experience of the technician. Hence, there is a recognized need to identify easily performed supplemental molecular diagnostic techniques with improved sensitivity, specificity, reproducibility and utility. EMMPRIN analysis may, in fact, meet this need. In the present study, using an in vitro OSCC progression model as well as in vivo tissue specimens, we report that EMMPRIN expression gradually increases during early oral carcinogenesis. Our microarray, qRT-PCR, Western blotting and immunohistochemical data convincingly demonstrated that EMMPRIN mRNA and protein are expressed at very low levels in normal oral and epidermal keratinocytes and at very high levels in OPM cell lines Leuk1 and Leuk2. EMMPRIN expression levels remain high in both primary and metastatic OSCC cell lines, although somewhat lower than in the OPM cells. This indicates that upregulation of EMMPRIN expression occurs during the preinvasive stage of OSCC development and its expression remains stable during invasive and metastatic progression. It appears that there is no further increase in EMMPRIN expression levels with progression of OPM to OSCC, but clearly these levels are consistently higher for all premalignant and malignant cells compared to the non-malignant cells. As shown in our OSCC tumor specimen, EMPPRIN expression becomes repressed among differentiated OSCC cells which may partially explain the differences in the EMMPRIN expression levels between OPM and OSCC cells. It should be also noted that the original tumors of OSCC cell lines 686Tu/Ln and 1386Tu/Ln are not derived from the same patients that were the sources for the OPM lesions of Leuk1 or Leuk2 cells. The distinct tumor microenvironment within those different tissue sources presumably contributes to quantitative variations of such expression levels. EMMPRIN: multiple roles in promoting carcinogenesis The transition from premalignancy to invasive cancer is preceded by the activation of local host stroma that is favorable
155
to the microinvasion of cancer cells (Liotta and Kohn, 2001). Thus, the microenvironment of the host stroma is, therefore, an active participant in this process, functioning as a “carcinogen” during tumorigenesis (Bissell and Radisky, 2001). EMMPRIN is a multifunctional transmembrane protein that mediates critical molecular events that precede the transition of premalignant lesions to invasive carcinoma. Remodeling of the extracellular matrix (ECM) and basement membrane is confined to the immediate pericellular environment of the premalignant cells and may be the first step necessary for local invasion. The principal enzymes that degrade extracellular matrix and basement membrane are the matrix metalloproteinases, a family of secreted and membrane-anchored (MTMMPs) proteinases. One of the well-known functions of EMMPRIN is its ability to stimulate host stromal fibroblasts to produce multiple MMPs, MT-MMPs and their endogenous activators, in part via a mitogen-activated protein kinase (MAPK) p38 kinase signaling pathway (Yan et al., 2005). EMMPRIN is primarily expressed by tumor cells, whereas the MMP expression is localized to the peritumoral fibroblasts (Yan et al., 2005). Transition from OPM to invasive OSCC also depends on the capability of these cells to induce tumor angiogenesis. Recent in vitro and animal model investigations have demonstrated that EMMPRIN stimulates the production of VEGF by both tumor and stromal cells via MMP-dependent and -independent pathways (Tang et al., 2005). Disengagement of integrinmediated adhesion to the matrix, an early event in carcinogenesis, can elicit apoptosis or anoikis if not followed by attachment and re-adhesion. Consequently, for premalignant cells to initiate invasion and migration, pro-invasive and antianoikis signals must occur in concert to enable anchorageindependent growth. EMMPRIN stimulates the production of hyaluronan by tumor cells which in turn promotes cell survival and anchorage-independent growth via activation of Akt, Erk and FAK (Marieb et al., 2004). Multidrug resistance in tumor cells is also linked to the overexpression of EMMPRIN (Yang et al., 2003). Taken together, EMMPRIN seems to play a very crucial role in mediating anchorage-independent growth, angiogenesis and invasion, all of which are essential molecular events of carcinogenesis (Fig. 11). EMMPRIN and malignant transformation of OPM lesions Since EMMPRIN plays such an important role in epithelial–connective tissue interactions, we evaluated the EMMPRIN expression patterns in tissue-engineered models that recapitulate normal oral mucosa, OPM mucosa and a combination of both. Membranous reactivity for EMMPRIN was noted only in the basal layer of normal oral mucosal organotypic culture, whereas all epithelial layers of OPM organotypic culture showed intense staining for EMMPRIN. Recent studies have suggested that EMMPRIN serves as its own counter-receptor on the surface of cancer cells and peritumoral fibroblasts and induces the production of MMP or VEGF by these cells via a homophilic interaction mediated by the extracellular Ig domain (Sun and Hemler, 2001). It has
156
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
Fig. 11. Hypothetical model for the role of EMMPRIN in mediating tumor–host interaction during transition from oral premalignant lesion to invasive oral cancer. BM: basement membrane; ECM: extracellular matrix; EGFR: epidermal growth factor receptor; VEGF: vascular endothelial growth factor.
been shown, for example, that EMMPRIN expression is upregulated in fibroblasts and presented on their cell surfaces as its own counter-receptor when these cells are pretreated with EMMPRIN (Tang et al., 2004a,b). Therefore, we evaluated the effect of EMMPRIN produced by OPM cells (Leuk1) on adjacent normal oral epithelial cells using organotypic oral mucosal models composed of normal and OPM cells. Upregulated EMMPRIN found on the OPM cells did not enhance the expression of EMMPRIN on normal cells in direct contact with the OPM cells. This suggests that the novel positive feedback regulatory mechanism reported for the increased expression of EMMPRIN by peritumoral fibroblasts cannot be applied to normal oral epithelial cells adjacent to EMMPRIN-overexpressing OPM cells. Therefore, upregulated EMMPRIN expression in oral epithelial cells seems to only occur when there is dysregulated signaling of critical molecules implicated in oral carcinogenesis. Abnormally high levels of EMMPRIN expression by cancer cells have been attributed to dysregulation of epidermal growth factor receptor (EGFR) signaling (Menashi et al., 2003). Furthermore, the EGFR ligand amphiregulin promotes tumor progression via EMMPRIN-mediated increased production of MMPs by both fibroblasts and endothelial cells (Menashi et al., 2003). It should be noted that EGFR overexpression occurs in approximately 80–100% of OPM (Grandis and Tweardy, 1993; Shin et al., 1994). Consequently, EMMPRIN overexpression in OPM should be closely tied to the deregulated EGFR signaling in these lesions. Therefore, molecular targeted therapy towards EGFR in OPM may
also have a beneficial effect by downregulating EMMPRIN expression in these lesions. Taken together, EMMPRIN expression patterns of both normal and OPM organotypic mucosal models closely mimicked those of the tissues in vivo. Therefore, these in vitro models would appear to be valuable for the evaluation of molecular-targeted strategies to block the expression and function of EMMPRIN. The unique discovery of a band-like EMMPRIN deposit in an area that recapitulates the basement membrane zone in vivo suggests that it may play a role in establishing the basal cell polarity during oral mucosal morphogenesis and development. A similar role for EMMPRIN has been reported in the development and morphogenesis of the retinal pigment epithelium (Marmorstein et al., 1998). It is significant that a similar EMMPRIN expression pattern was not observed with our OPM model. This supports the dysregulated morphogenesis and differentiation of this model. EMMPRIN expression was also evaluated in an oral cancer progression model in vivo, represented by archival specimens of leukoplakias without OED, with varying degrees of OED, carcinoma-in-situ and of primary and metastatic OSCC. EMMPRIN expression is generally low in normal oral mucosa and exclusively limited to the actively differentiating basal cells. EMMPRIN expression was restricted to the basal cells of benign leukoplakias showing no significant OED. However, expression of EMMPRIN rose, spreading to the upper epithelial layers, with increasing severity of OED in most of the leukoplakias and was noted in N90% of the epithelial cells in leukoplakias with severe OED and carcinoma-in-situ. Thus, the
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
polarized distribution of EMMPRIN to the basal cells of normal oral epithelia and benign oral leukoplakias is lost in leukoplakias with OED. Similarly, EMMPRIN expression was noted in N90% of the tumor cells of early or superficially invasive OSCC. EMMPRIN remained overexpressed in advanced primary and metastatic OSCC tumors, but its overexpression was mostly localized to the invasive fronts of these tumors. Such expression is repressed and becomes negligible in most of the differentiated areas of these tumors. This disassociated expression pattern for EMMPRIN during the early and advanced stages of OSCC progression suggests different mechanisms controlling the expression of EMMPRIN during each stage. Moreover, these in vivo expression patterns of EMMPRIN in OPM lesions and OSCC tumors mirror the expression levels observed in vitro with their respective cell lines. The EMMPRIN expression patterns were also assessed in two distinct clinical entities, oral lichen planus and proliferative verrucous leukoplakia, one with low/negligible and one with high malignancy transformation risks, respectively. Malignant transformation of oral lichen planus per year ranges between 0.04% and 1.74% (van der Meij et al., 2003). Proliferative verrucous leukoplakia is a specific clinical subtype of leukoplakia with a risk of OSCC development in the range of 60–100% (Neville et al., 2002; Bagan et al., 2004). EMMPRIN was restricted to the basal cells of most of the lichen planus cases examined, irrespective of the presence of reactive inflammatory atypia found in some of these biopsy specimens. On the other hand, EMMPRIN expression was markedly upregulated in all of the proliferative leukoplakia lesions examined, presumably reflecting their high malignancy risk. Thus, EMMPRIN as a molecular risk marker for malignancy holds true in two clinical entities with distinct and well-defined malignant transformation potential. Accumulating evidence suggests a prominent role for EMMPRIN in promoting tumorigenesis and metastasis in various types of malignant tumors. Increased expression of EMMPRIN both at the mRNA and protein levels has been reported in various types of malignant tumors (Bordador et al., 2000; Kanekura et al., 2002; Rosenthal et al., 2003; Ishibashi et al., 2004). Significantly, EMMPRIN expression in breast and esophageal cancer correlates with tumor size and stage and is associated with poor prognosis (Ishibashi et al., 2004; Reimers et al., 2004). EMMPRIN is also frequently expressed in micrometastases of breast cancers, suggesting a potential role for EMMPRIN in promoting metastatic spread of tumor cells (Reimers et al., 2004). Breast cancer cells engineered to overexpress EMPPRIN, moreover, have demonstrated an accelerated growth rate and metastatic progression, confirming a causal role for EMMPRIN in tumor growth and metastasis (Zucker et al., 2001). EMMPRIN is a membrane-associated glycoprotein and is composed of two extracellular immunoglobulin (Ig) domains, a transmembrane domain and a 39-amino acid cytoplasmic domain (Miyauchi et al., 1990; Miyauchi et al., 1991). In tumor cells, EMMPRIN is found in various forms, which include highly (HG ∼45–65 kDa) and low (LG ∼32–44 kDa)
157
glycosylated fractions as well as the core protein (∼27 kDa) (Tang et al., 2004a,b; Tang and Hemler, 2004). The HG form of EMMPRIN is the most effective inducer of MMPs (Sun and Hemler, 2001). The present data show that the ratio of HG/ LG-forms of EMMPRIN is markedly higher in OPM cells and metastatic OSCC cells compared to primary OSCC cells. This suggests a functional relevance for the HG form of EMMPRIN in oral carcinogenesis and metastatic progression. Caveolin-1, by associating with the LG-form, inhibits the formation of HG form and cell surface clustering of EMMPRIN which leads to impaired induction of MMP (Tang et al., 2004a,b; Tang and Hemler, 2004). Caveolin-1 is a principal component of plasma membrane caveolae and an important regulator of caveolaedependent signaling and endocytosis. Caveolin-1 is putative tumor suppressor gene, mapped to the locus 7q31.1 which is frequently deleted or mutated in various epithelial malignancies including OSCC (Engelman et al., 1998; Han et al., 2004). The tumor suppressor activity of caveolin-1 has been attributed to its inhibitory effect on EMMPRIN glycosylation (Tang et al., 2004a,b; Tang and Hemler, 2004). Our investigation showed that the HG form of EMMPRIN is secreted in the media by primary and metastatic OSCC cells but not by OPM cells. In fact, our three-dimensional immunofluorescence images demonstrated very strong EMMPRIN-positive microvesicle-like structures on the surface of OSCC cells, but not on OPM cells. These findings are supported by a recent report concluding that full-length EMMPRIN is frequently released from the tumor cell surface via microvesicle shedding (Sidhu et al., 2004). It can be speculated that, in invasive OSCC, EMMPRIN is released from the tumor cells and modulates stromal and endothelial cells that are not in direct contact with the tumor cells. Such a distant modulator activity appears not to be present in OPM lesions. Conclusion In vivo expression patterns of EMMPRIN in oral premalignant and malignant lesions mirror the expression levels observed in vitro with cell lines and organotypic cultures established from cells derived from normal oral mucosa, premalignant leukoplakia and primary and metastatic OSCC. Upregulation of EMMPRIN expression occurs at a very early stage as OPM lesions progress through hyperplasia/ hyperkeratosis to dysplasia and to invasive OSCC. Thus, upregulated expression of EMMPRIN seems to be an important epigenetic alteration that accompanies oral carcinogenic progression and most likely represents a cause rather than a consequence of malignant transformation of OPM lesions (Fig. 11). Therefore, we propose that EMMPRIN should be evaluated further as (1) an objective molecular diagnostic marker for the detection of high-risk preinvasive oral leukoplakias by comparison with proven molecular risk markers such as aneuploidy and LOH; (2) a biomarker for therapies targeting EGFR expression in oral leukoplakias; (3) a therapeutic target to prevent the progression high-risk oral leukoplakias to invasive OSCC.
158
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159
Acknowledgments This work was supported by NIH/NIDCR grants DE13150 (W.Z.) and DE14395 (P.G. S.), a Philip Morris USA Inc. and Philip Morris International External Research Program grant (W.Z) and by a Fleming and Davenport Award, Texas Medical Center (N.V.).
References Bagan, J.V., Murillo, J., Poveda, R., et al., 2004. Proliferative verrucous leukoplakia: unusual locations of oral squamous cell carcinomas, and field cancerization as shown by the appearance of multiple OSCCs. Oral. Oncol. 40, 440–443. Banerjee, A.G., Bhattacharyya, I., Lydiatt, W.M., et al., 2003. Aberrant expression and localization of decorin in human oral dysplasia and squamous cell carcinoma. Cancer Res. 63, 7769–7776. Bissell, M.J., Radisky, D., 2001. Putting tumours in context. Nat. Rev., Cancer 1, 46–54. Biswas, C., 1982. Tumor cell stimulation of collagenase production by fibroblasts. Biochem. Biophys. Res. Commun. 109, 1026–1034. Biswas, C., 1984. Collagenase stimulation in cocultures of human fibroblasts and human tumor cells. Cancer Lett. 24, 201–207. Bordador, L.C., Li, X., Toole, B., et al., 2000. Expression of emmprin by oral squamous cell carcinoma. Int. J. Cancer 85, 347–352. Bouquot, J., 1999. Oral cancers with leukoplakia. Oral Dis. 5, 183–184. Boyle, P., Macfarlane, G.J., Scully, C., 1993. Oral cancer: necessity for prevention strategies. Lancet 342, 1129. Braakhuis, B.J., Tabor, M.P., Kummer, J.A., et al., 2003. A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications. Cancer Res. 63, 1727–1730. Braakhuis, B.J., Leemans, C.R., Brakenhoff, R.H., 2004. A genetic progression model of oral cancer: current evidence and clinical implications. J. Oral Pathol. & Med. 33, 317–322. Chinnathambi, S., Tomanek-Chalkley, A., Ludwig, N., et al., 2003. Recapitulation of oral mucosal tissues in long-term organotypic culture. Anat. Rec. A Discov. Mol. Cell Evol. Biol. 270, 162–174. Engelman, J.A., Zhang, X.L., Lisanti, M.P., 1998. Genes encoding human caveolin-1 and -2 are co-localized to the D7S522 locus (7q31.1), a known fragile site (FRA7G) that is frequently deleted in human cancers. FEBS Lett. 436, 403–410. Forastiere, A., Koch, W., Trotti, A., et al., 2001. Head and neck cancer. N. Engl. J. Med. 345, 1890–1900. Grandis, J.R., Tweardy, D.J., 1993. Elevated levels of transforming growth factor alpha and epidermal growth factor receptor messenger RNA are early markers of carcinogenesis in head and neck cancer. Cancer Res. 53, 3579–3584. Han, S.E., Park, K.H., Lee, G., et al., 2004. Mutation and aberrant expression of Caveolin-1 in human oral squamous cell carcinomas and oral cancer cell lines. Int. J. Oncol. 24, 435–440. Hasina, R., Hulett, K., Bicciato, S., et al., 2003. Plasminogen activator inhibitor2: a molecular biomarker for head and neck cancer progression. Cancer Res. 63, 555–559. Ishibashi, Y., Matsumoto, T., Niwa, M., et al., 2004. CD147 and matrix metalloproteinase-2 protein expression as significant prognostic factors in esophageal squamous cell carcinoma. Cancer 101, 1994–2000. Kanekura, T., Chen, X., Kanzaki, T., 2002. Basigin (CD147) is expressed on melanoma cells and induces tumor cell invasion by stimulating production of matrix metalloproteinases by fibroblasts. Int. J. Cancer 99, 520–528. Kim, M.M., Califano, J.A., 2004. Molecular pathology of head-and-neck cancer. Int. J. Cancer 112, 545–553. Kopelovich, L., Henson, D.E., Gazdar, A.F., et al., 1999. Surrogate anatomic/ functional sites for evaluating cancer risk: an extension of the field effect. Clin Cancer Res. 5, 3899–3905. Lee, J.J., Hong, W.K., Hittelman, W.N., et al., 2000. Predicting cancer
development in oral leukoplakia: ten years of translational research. Clin. Cancer Res. 6, 1702–1710. Liotta, L.A., Kohn, E.C., 2001. The microenvironment of the tumour–host interface. Nature 411, 375–379. Lippman, S.M., Sudbo, J., Hong, W.K., 2005. Oral cancer prevention and the evolution of molecular-targeted drug development. J. Clin. Oncol. 23, 346–356. Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) method. Methods 25, 402–408. Lumerman, H., Freedman, P., Kerpel, S., 1995. Oral epithelial dysplasia and the development of invasive squamous cell carcinoma. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endo. 79, 321–329. Macfarlane, G.J., Boyle, P., Evstifeeva, T.V., et al., 1994. Rising trends of oral cancer mortality among males worldwide: the return of an old public health problem. Cancer Causes Control 5, 259–265. Marieb, E.A., Zoltan-Jones, A., Li, R., et al., 2004. Emmprin promotes anchorage-independent growth in human mammary carcinoma cells by stimulating hyaluronan production. Cancer Res. 64, 1229–1232. Marmorstein, A.D., Gan, Y.C., Bonilha, V.L., et al., 1998. Apical polarity of NCAM and EMMPRIN in retinal pigment epithelium resulting from suppression of basolateral signal recognition. J. Cell Biol. 142, 697–710. Menashi, S., Serova, M., Ma, L., et al., 2003. Regulation of extracellular matrix metalloproteinase inducer and matrix metalloproteinase expression by amphiregulin in transformed human breast epithelial cells. Cancer Res. 63, 7575–7580. Miyauchi, T., Kanekura, T., Yamaoka, A., et al., 1990. Basigin, a new, broadly distributed member of the immunoglobulin superfamily, has strong homology with both the immunoglobulin V domain and the beta-chain of major histocompatibility complex class II antigen. J. Biochem. (Tokyo) 107, 316–323. Miyauchi, T., Masuzawa, Y., Muramatsu, T., 1991. The basigin group of the immunoglobulin superfamily: complete conservation of a segment in and around transmembrane domains of human and mouse basigin and chicken HT7 antigen. J. Biochem. (Tokyo) 110, 770–774. Neville, B.W., Day, T.A., 2002. Oral cancer and precancerous lesions. CA Cancer J. Clin. 52, 195–215. Neville, B., Damm, D., Allen, C., Bouquot, J., 2002. Oral and Maxillofacial Pathology: Epithelial Pathology. W.B. Saunders, Philadelphia, pp. 315–387. Odell, E., Johnson, N.W., 2005. The evidence base and practicalities of genetic screening for risk of subsequent malignancy: cell and molecular markers. WHO Collaborating Centre Workshop on Classification and Diagnosis of Oral Potentially Malignant Lesions, London, U.K. O’Shaughnessy, J.A., Kelloff, G.J., Gordon, G.B., et al., 2002. Treatment and prevention of intraepithelial neoplasia: an important target for accelerated new agent development. Clin. Cancer Res. 8, 314–346. Parkin, D.M., Pisani, P., Ferlay, J., 1999. Global cancer statistics. CA Cancer J. Clin. 49, 33–64 (31). Pisani, P., Parkin, D.M., Bray, F., et al., 1999. Estimates of the worldwide mortality from 25 cancers in 1990. Int. J. Cancer 83, 18–29. Reimers, N., Zafrakas, K., Assmann, V., et al., 2004. Expression of extracellular matrix metalloproteases inducer on micrometastatic and primary mammary carcinoma cells. Clin. Cancer Res. 10, 3422–3428. Rosenthal, E.L., Shreenivas, S., Peters, G.E., et al., 2003. Expression of extracellular matrix metalloprotease inducer in laryngeal squamous cell carcinoma. Laryngoscope 113, 1406–1410. Rosin, M.P., Cheng, X., Poh, C., et al., 2000. Use of allelic loss to predict malignant risk for low-grade oral epithelial dysplasia. Clin. Cancer Res. 6, 357–362. Sacks, P.G., 1996. Cell, tissue and organ culture as in vitro models to study the biology of squamous cell carcinomas of the head and neck. Cancer Metastasis Rev. 15, 27–51. Schepman, K., der Meij, E., Smeele, L., et al., 1999. Concomitant leukoplakia in patients with oral squamous cell carcinoma. Oral Dis. 5, 206–209. Shin, D.M., Ro, J.Y., Hong, W.K., et al., 1994. Dysregulation of epidermal growth factor receptor expression in premalignant lesions during head and neck tumorigenesis. Cancer Res. 54, 3153–3159. Sidhu, S.S., Mengistab, A.T., Tauscher, A.N., et al., 2004. The microvesicle as a
N. Vigneswaran et al. / Experimental and Molecular Pathology 80 (2006) 147–159 vehicle for EMMPRIN in tumor–stromal interactions. Oncogene 23, 956–963. Sudbo, J., Bryne, M., Johannessen, A.C., et al., 2001a. Comparison of histological grading and large-scale genomic status (DNA ploidy) as prognostic tools in oral dysplasia. J. Pathol. 194, 303–310. Sudbo, J., Kildal, W., Risberg, B., et al., 2001b. DNA content as a prognostic marker in patients with oral leukoplakia. N. Engl. J. Med. 344, 1270–1278. Sun, J., Hemler, M.E., 2001. Regulation of MMP-1 and MMP-2 production through CD147/extracellular matrix metalloproteinase inducer interactions. Cancer Res. 61, 2276–2281. Tabor, M.P., Brakenhoff, R.H., van Houten, V.M., et al., 2001. Persistence of genetically altered fields in head and neck cancer patients: biological and clinical implications. Clin. Cancer Res. 7, 1523–1532. Tabor, M.P., Braakhuis, B.J., van der Wal, J.E., et al., 2003. Comparative molecular and histological grading of epithelial dysplasia of the oral cavity and the oropharynx. J. Pathol. 199, 354–360. Tang, W., Hemler, M.E., 2004. Caveolin-1 regulates matrix metalloproteinases1 induction and CD147/EMMPRIN cell surface clustering. J. Biol. Chem. 279, 11112–11118. Tang, W., Chang, S.B., Hemler, M.E., 2004a. Links between CD147 function, glycosylation, and caveolin-1. Mol. Biol. Cell 15, 4043–4050. Tang, Y., Kesavan, P., Nakada, M.T., et al., 2004b. Tumor–stroma interaction: positive feedback regulation of extracellular matrix metalloproteinase inducer (EMMPRIN) expression and matrix metalloproteinase-dependent generation of soluble EMMPRIN. Mol. Cancer Res. 2, 73–80. Tang, Y., Nakada, M.T., Kesavan, P., et al., 2005. Extracellular matrix metalloproteinase inducer stimulates tumor angiogenesis by elevating
159
vascular endothelial cell growth factor and matrix metalloproteinases. Cancer Res. 65, 3193–3199. van der Meij, E.H., Schepman, K.P., van der Waal, I., 2003. The possible premalignant character of oral lichen planus and oral lichenoid lesions: a prospective study. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endo. 96, 164–171. Vigneswaran, N., Zhao, W., Dassanayake, A., et al., 2000. Variable expression of cathepsin B and D correlates with highly invasive and metastatic phenotype of oral cancer. Hum. Pathol. 31, 931–937. Vigneswaran, N., Wu, J., Sacks, P., et al., 2005. Microarray gene expression profiling of cell lines from primary and metastatic tongue squamous cell carcinoma: possible insights from emerging technology. J. Oral. Pathol. Med. 34, 77–86. Weijers, M., Snow, G.B., Bezemer, P.D., et al., 2002. The clinical relevance of epithelial dysplasia in the surgical margins of tongue and floor of mouth squamous cell carcinoma: an analysis of 37 patients. J. Oral. Pathol. Med. 31, 11–15. Yan, L., Zucker, S., Toole, B.P., 2005. Roles of the multifunctional glycoprotein, emmprin (basigin; CD147), in tumour progression. Thromb. Haemost. 93, 199–204. Yang, J.M., Xu, Z., Wu, H., et al., 2003. Overexpression of extracellular matrix metalloproteinase inducer in multidrug resistant cancer cells. Mol. Cancer Res. 1, 420–427. Zucker, S., Hymowitz, M., Rollo, E.E., et al., 2001. Tumorigenic potential of extracellular matrix metalloproteinase inducer. Am. J. Pathol. 158, 1921–1928.
