Semiquantitative reverse transcription–polymerase chain reaction analysis of matrix metalloproteinases (MMPs) and tissue inhibitors of MMPs (TIMPs). Each value is the ratio of MMP or TIMP to β-actin messenger RNA level and is the mean of 3 independent RNA samples. The MMPs and TIMPs for which no messenger RNA was detected by 50 or more polymerase chain reaction cycles are indicated by zeros in the graphs. Horizontal lines indicate mean values; LNM, lymph node metastasis; and MT1, membrane-type 1.
Quantitation of matrix metalloproteinase (MMP) 1 (pro form) (A), tissue inhibitor of MMP (TIMP) 1 (B), and TIMP-2 (C) proteins in tissue homogenates. Error bars represent SEM. LNM indicates lymph node metastasis.
A, Gelatin zymography; B, casein zymography; and C, Western blot analysis of matrix metalloproteinase 14 protein. Tissue homogenates from 5 representative matched pairs of primary tumors (T), metastatic lymph nodes (LN), and histologically normal tissues (N) were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Conditioned medium from 12-O-tetradecanoyl-phorbol-13-acetate–treated HT-1080 fibrosarcoma cell line and MDA-MB 231 mammary carcinoma cells served as a positive control and as a standard for intergel variations for gelatin zymography and casein zymography, respectively. Molecular mass markers, in kilodaltons, are indicated on the left (Mr × 103).
Gelatinolytic and caseinolytic activities in tissue homogenates. Values are obtained from computerized image analysis of transparent bands in substrate zymography. Error bars represent SEM.
Correlation between lymph node metastasis (A) or advanced pathological stages (B) and the messenger RNA expression of matrix metalloproteinase (MMP) 2 and MMP-9 in primary head and neck squamous cell carcinoma tissues. Node – and Node + are primary tumors without and with neck node metastasis, respectively. The cutoff values obtained from the mean values of MMP messenger RNAs determined by reverse transcription–polymerase chain reaction in primary tumors were used to separate primary head and neck squamous cell carcinoma into groups with low and high expression.
O-charoenrat P, Rhys-Evans PH, Eccles SA. Expression of Matrix Metalloproteinases and Their Inhibitors Correlates With Invasion and Metastasis in Squamous Cell Carcinoma of the Head and Neck. Arch Otolaryngol Head Neck Surg. 2001;127(7):813-820. doi:10-1001/pubs.Arch Otolaryngol. Head Neck Surg.-ISSN-0886-4470-127-7-ooa00225
Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001
Matrix metalloproteinases (MMPs) have been implicated in the invasion and metastasis of head and neck squamous cell carcinoma (HNSCC). However, a detailed analysis of MMPs and tissue inhibitors of MMPs (TIMPs) in relation to the biological behavior of HNSCC has yet to be performed in clinical material.
To study a comprehensive profile of MMPs and their 2 main inhibitors in HNSCC tissue samples and to correlate the patterns of expression with clinicopathological characteristics, invasion, and metastasis.
This study included 54 consecutive patients with primary HNSCC, 27 of which showed lymph node metastasis. Expression of MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-10, MMP-11, MMP-13, MMP-14, TIMP-1, and TIMP-2 was simultaneously analyzed in tissue homogenates using semiquantitative reverse transcription–polymerase chain reaction assay. Where feasible, levels of protein and enzyme activity were confirmed by Western blot, enzyme-linked immunosorbent assay, and substrate zymography. Conventional clinicopathological features, including mode of tumor invasion, were also examined.
Significantly higher MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-10, MMP-11, MMP-13, and TIMP-1 levels were found in tumors vs specimens of matched normal mucosa. No difference in the distribution of MMPs and TIMPs in relation to age, sex, tumor site, or histological grade was observed. A significant correlation was demonstrated between levels of MMP-1, MMP-9, and TIMP-1 and advanced T stage and between MMP-9 expression and an infiltrative pattern of growth. Enhanced expression of MMP-9 was strongly correlated (P<.001) and levels of MMP-2, MMP-7, and MMP-11 were weakly correlated (P = .03-.05) with lymph node involvement.
Overexpression of multiple MMPs and TIMPs is characteristic of HNSCC, and analysis of specific MMPs, MMP-9 in particular, might be useful for evaluating the malignant potential in individual HNSCC.
SQUAMOUS CELL carcinoma of the head and neck (HNSCC) is a major problem worldwide.1 One of the most characteristic clinical features of HNSCC is its capacity to invade adjacent tissues and metastasize locoregionally. Cancer cell invasion, metastasis, and angiogenesis is a complex, multistep process involving the cooperation of multiple proteolytic enzymes secreted by tumor or host cells and whose substrates include extracellular matrix components.2 Evidence3- 22 suggests that matrix metalloproteinases (MMPs) and their physiological tissue inhibitors (TIMPs) might play a causal role in HNSCC progression.
The MMPs are a family of highly homologous extracellular zinc- and calcium-dependent endopeptidases with enzymatic activity against almost all protein components of the extracellular matrix. Based on the protein domain structure and substrate specificity, the MMPs can be divided into 4 subclasses.2 The first group, which degrades types I, II, and III fibrillar collagens, is composed of MMP-1 (interstitial collagenase), MMP-8 (neutrophil collagenase), and MMP-13 (collagenase-3). The second group, stromelysins, includes 4 members: MMP-7 (matrilysin) contains the minimal number of domains, ie, a predomain, a prodomain, and a catalytic domain, and MMP-3 (stromelysin-1), MMP-10 (stromelysin-2), and MMP-11 (stromelysin-3) contain an additional carboxy-terminal hemopexinlike domain. The stromelysins have a broad substrate specificity and are capable of degrading many extracellular components, eg, laminin, fibronectin, and proteoglycans. Matrix metalloproteinase 2 (gelatinase-A) and MMP-9 (gelatinase-B) account for a separate class based on the presence of a fibronectinlike domain. Gelatinases are able to cleave both the denatured forms of collagen and type IV collagen found in basement membrane. Matrix metalloproteinase 2 and MMP-9 also contain a gelatin-binding domain that endows them with high affinity for gelatin. The last group of MMPs contains the membrane-type MMPs (MT-MMPs), which are composed of MMP-14 (MT1-MMP), MMP-15 (MT2-MMP), MMP-16 (MT3-MMP), MMP-17 (MT4-MMP), and MMP-24 (MT5-MMP). Membrane-type MMPs have the unique property of possessing a hydrophobic sequence at the C-terminus, which allows insertion of the protein into the cell membrane. Some MMPs cannot be grouped into any of these classes, including MMP-12 (metalloelastase), MMP-18, MMP-19, MMP-20 (enamelysin), and MMP-23. These enzymes differ in substrate specificity, regulation, tissue-specific expression, and potential interactions with additional MMP and TIMP family members. Expression of MMP activity can be controlled at the level of gene transcription, by proenzyme activation and by broad-spectrum and specific inhibitors. Tumor cells might induce the host cells within the surrounding stroma to secrete these enzymes or vice versa. Most MMPs are secreted as latent proenzymes that undergo proteolytic cleavage of an amino-terminal domain during activation. The net activity of MMPs is determined by the amount of proenzyme expressed, the extent to which the proenzyme is activated, and the local concentration of specific tissue inhibitors of MMPs, ie, TIMPs.
A multigene family of proteins named TIMPs has been demonstrated to inhibit fully activated MMPs. Tissue inhibitors of MMPs comprise at least 4 members, and, together, they provide a tightly regulated mechanism for control of MMP activation and function. Tissue inhibitor of MMP-1 and TIMP-2 have molecular weights of 28.5 and 21.0 kd, respectively, and seem to act by forming 1:1 stoichiometric complexes with the active MMP.23 Tissue inhibitor of MMP-1 can inhibit the collagenases, MMP-3, and the gelatinases.24 Tissue inhibitor of MMP-2 binds preferentially to MMP-2 but also inhibits the activities of MMP-1, MMP-3, MMP-7, and MMP-9.25 The local balance of these enzymes and inhibitors seems to be a crucial factor in tumor invasion and metastasis.
We20 recently demonstrated the expression of several members of the MMP and TIMP family in a large panel of HNSCC cell lines and a strong correlation between some MMPs (MMP-9 in particular) and their in vitro invasiveness. The aim of this investigation was to perform a comprehensive analysis of MMPs previously identified in HNSCC, including MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-10, MMP-11, MMP-13, and MMP-14 (MT1-MMP), and 2 inhibitors, TIMP-1 and TIMP-2, in fresh tissue samples and to correlate with the clinicopathological characteristics of HNSCC.
The protocol for the following studies was approved by the Ethical Committee of the Royal Marsden Hospital Trust, London, England. Fresh tissue samples were obtained from 54 patients undergoing major surgical resection for HNSCC at the Department of Head and Neck Surgery, the Royal Marsden Hospital, between July 1, 1997, and October 31, 1999. The clinical and pathological characteristics of patients are summarized in Table 1. There were 44 men and 10 women (median age, 58.5 years; range, 27-83 years). Patients had no detectable metastases in distant organs at the time of surgery. None of the patients had previously received preoperative chemotherapy or radiotherapy. Adjuvant treatment was given after radical surgery in appropriate cases following the hospital's protocol. In each case, the portion of tumor was resected near the advancing edge of the tumor, avoiding its necrotic center. After excision, tissue samples were immediately snap-frozen and stored in liquid nitrogen until use. Samples of the adjacent tissues were submitted for histopathological study, which revealed that most cells were malignant. Tumors were staged according to the TNM classification (5th edition)26 and were graded as well (G1), moderately (G2), and poorly (G3) differentiated. T classification was evaluated according to tumor size for tumors from the oral cavity or the oropharynx and tumor size and extensiveness for tumors from the hypopharynx and the larynx. The mode of cancer invasion was histologically classified as described previously27; grade 1 has a well-defined borderline, grade 2 has a less well-defined borderline, grade 3 has groups of cells and no distinct borderline, and grade 4 has diffuse invasion. In 27 patients, tissue samples of metastatic lymph nodes (LNM) were also available for analysis. Matched histologically normal mucosa of the upper aerodigestive tract, resected at least 5 cm distant from the tumor area,3,8 was obtained from 32 patients. As of June 30, 2000, median follow-up for living patients was 20 months (range, 9-35 months); 35 patients (65%) were alive, 15 (28%) died of tumors, and 4 (7%) died of unrelated causes. Node-positive cases in this study are patients in whom positive cervical nodes were identified based on histological diagnosis after a neck dissection, and patients who experienced no metastasis for at least 12 months after surgery are node-negative cases.
The semiquantitative reverse transcription–polymerase chain reaction (RT-PCR) assay and the primer sequences have been described previously.20 The following genes were assayed using specific primer pairs: MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-10, MMP-11, MMP-13, MMP-14 (MT1-MMP), TIMP-1, and TIMP-2. β-Actin was used to check RNA integrity and as an internal control. Oligonucleotide primers were purchased from Genosys (Cambridge, England). To control gel-to-gel variability, each PCR product from HT-1080 fibrosarcoma cells or MDA-MB 231 mammary carcinoma cells was also electrophoresed as a control on every gel. The level of messenger RNA (mRNA) was calculated as the ratio of tissue sample to control cell line on the same scan and was then corrected as a ratio to the corresponding β-actin level.
Tissues samples (wet weight, 300-400 mg) were homogenized in ice-cold 75-mM Tris-hydrochloride buffer, pH 7.4, containing 0.5% Triton X-100. After removal of debris and nuclear pellets by centrifugation (10 000g for 10 minutes at 4°C), the protein concentration was determined using a protein assay reagent kit (BCA; Pierce, Rockford, Ill). The supernatants containing protein were stored at −70°C until required.
Forty micrograms of the extract proteins was resolved under nonreducing conditions in 10% sodium dodecyl sulfate–polyacrylamide gels copolymerized with 0.1% (wt/vol) gelatin or 0.05% β-casein. Gelatinolytic or caseinolytic enzymes were detected as transparent bands on the blue background, and the intensity of the bands was measured using image analysis software (Quantiscan, Cambridge, England), as described previously.20 Results were expressed in arbitrary units per 40 µg of total protein. Conditioned medium from 12-O-tetradecanoyl-phorbol-13-acetate (TPA)–treated HT-1080 fibrosarcoma cell line and TPA-treated MDA-MB 231 mammary carcinoma cells served as a positive control and a standard for intergel variations for gelatin zymography and casein zymography, respectively.
Concentrations of MMP-1, TIMP-1, and TIMP-2 in the same tissue homogenates (500 µg per sample) as used in zymography were measured using commercially available enzyme-linked immunosorbent assay (ELISA) kits (Chemicon International Inc, Temecula, Calif). The values measured represent pro–MMP-1, TIMP-1, and TIMP-2 concentrations, with a range of detection at 0.16 to 10.0, 1.2 to 49.0, and 20.0 to 320.0 ng/mL, respectively. Results were calculated as nanograms per 1 mg of total protein tissue extracts. Two independent experiments were performed. In each experiment, tissue lysates were prepared from 2 separate pieces of the same tissue specimen, and ELISA values were measured in duplicate for each sample.
Equal amounts of protein (100 µg) from the same tissue homogenates as used in zymography were resolved under reducing condition in 10% and 15% sodium dodecyl sulfate–polyacrylamide gel electrophoresis for MMP and TIMP detection, respectively, and transferred onto a nitrocellulose membrane (Highbond-C extra; Amersham International, Buckinghamshire, England), then probed with the appropriate primary antibody. Antibodies to MMP-2, MMP-3, MMP-7, MMP-9, and MT1-MMP were provided by British Biotech (Oxford, England). Antibodies to TIMP-1 and TIMP-2 were purchased from Chemicon (Harrow, England). Blots were washed, incubated with a secondary antibody coupled to horseradish peroxidase (Serotec, Oxford), and developed using the luminol reagent (Santa Cruz Inc, Santa Cruz, Calif) and Kodak X-OMAT AR film (Eastman Kodak, New York, NY) with an intensifying screen. Levels of proteins were determined by image analysis using Quantiscan software. As a negative control, the primary antibody, which was preabsorbed with corresponding proteins overnight at a ratio of 1:10, or normal serum was reacted with the membrane filter. The specific bands were absent when the preabsorbed antibody or normal serum was used. Purified human MMP-2/MMP-9, MMP-3/MMP-7, TIMP-1 (28 kd), and TIMP-2 (24 kd) were obtained from CalBiochem (Nottingham, England).
All statistical analyses were performed using statistical software (GraphPad Prism version 2.01; GraphPad Software Inc, San Diego, Calif). To compare levels of mRNA expression between tumor tissues and control (histologically normal) tissues and to determine the significance of increased mRNA expression of MMPs and TIMPs with various clinicopathological variables, the Mann-Whitney test and the Kruskal-Wallis test with the Dunn multiple comparison test were used when comparing 2 groups and 3 or more groups, respectively. The 2-tailed Fisher exact test was used to analyze the contingency table. Levels of mRNA and protein were measured from 2 parts of the same specimens in duplicate where feasible. Correlations between the mRNA and protein levels were computed using the 2-tailed Spearman nonparametric correlation. Results are expressed as mean ± SEM. P<.05 was considered statistically significant. Unless otherwise stated, each experiment was performed twice with virtually identical results.
Lesions from primary HNSCC (n = 54) and LNM (n = 27) and histologically normal adjacent mucosa (n = 32) were examined for mRNA levels of multiple MMPs and TIMPs. Using semiquantitative RT-PCR assay, mRNA expression of MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-10, MMP-13, and TIMP-1 was significantly greater in malignant tissues (primary tumors and/or LNM) compared with mRNA levels in histologically normal mucosa (P = .02 to <.001) (Figure 1). No significant differences were found between mRNA levels of MMP-14 (MT1-MMP) and TIMP-2 in tumors and control tissues. In addition, no differences were found between expression of most MMPs tested in primary HNSCC vs LNM, except for the MMP-3 gene, where the levels in primary tumors (14.55 ± 3.91) were significantly higher than those in LNM (1.09 ± 0.32) (P = .03).
Levels of MMP-1, TIMP-1, and TIMP-2 proteins were also measured from tissue homogenates using the ELISA kit (Figure 2). Compared with control tissues, (1) MMP-1 protein levels were 4.1-fold greater in primary tumors and 3.8-fold greater in LNM (P<.001 for both) and (2) TIMP-1 protein levels were 1.6-fold higher in primary tumors (P = .009) and 1.9-fold in LNM (P = .007); TIMP-2 protein levels were detected in tumors and in control tissues, but no significant differences were observed, corresponding to the RT-PCR results.
We used substrate zymography to determine the activity of MMP-2/MMP-9, MMP-3/MMP-7, and TIMP-1/TIMP-2. Gelatin zymography revealed a varied profile of MMPs: lysis zones corresponding to molecular weights of 92, 84, 72, and 62 kd and high-molecular-weight (approximately 120-kd) gelatinases were seen (Figure 3A). Gelatinases of 92 and 84 kd correspond to pro–MMP-9 and an active form of MMP-9, respectively. Those of 72 and 62 kd are considered to be pro–MMP-2 and its active form, respectively. These results are supported by the profile of TPA-treated HT-1080 cells, which express all of these enzyme species and whose gelatinolytic zones matched those of the HNSCC samples. The high-molecular-weight gelatinase detected in some samples might be due to a complex of MMP and TIMP. Indeed, 2 inhibitors tested, TIMP-1 and TIMP-2, were detected in most HNSCC tissues by ELISA (Figure 2) and gelatin reverse zymography (data not shown). Casein zymography demonstrated lytic zones corresponding to molecular weights of of approximately 57, 45, 29, and 21 kd and several small bands between 57 and 45 kd and 29 and 21 kd (Figure 3B). The 57- and 45-kd bands correspond to latent and active forms of MMP-3, respectively. The 29- and 21-kd bands are considered to be pro–MMP-7 and its active form, respectively. The nature of gelatinolytic (MMP-2, MMP-9, and MMP-14) and caseinolytic (MMP-3 and MMP-7) enzymes or inhibitors (TIMP-1 and TIMP-2) was confirmed by Western blotting of the same tissue homogenates with specific antibodies (only data of MMP-14 are shown in Figure 3C).
Using computerized image analysis of transparent bands, we quantitated the gelatinolytic and caseinolytic activities in tissue homogenates. The results (Figure 4) showed that the MMP-9 (92 and 84 kd, respectively) and MMP-2 (72 and 62 kd, respectively) activities were significantly greater in primary tumors and LNM compared with the levels in control tissues. Compared with control tissues, the levels of high-molecular-weight gelatinase (approximately 120 kd), MMP-3 (45-51 kd), and MMP-7 (21-29 kd) were also greater in the primary HNSCC, although the levels in LNM did not reach statistical significance (Figure 4).
Our results imply that several members of the MMP and TIMP family are present in primary and secondary HNSCC. Apart from MMP-14 and TIMP-2, all the molecules tested (including TIMP-1) were up-regulated in primary HNSCC and/or LNM compared with adjacent histologically normal mucosa.
Relationships between the mRNA expression of MMPs and TIMPs in 54 primary tumors and their clinicopathological variables were analyzed. As shown in Table 2, MMP-1, MMP-9, and TIMP-1 mRNA expression in primary HNSCC showed a statistically significant relationship with a higher T classification (T3-T4) (P = .004, .001, and .003, respectively). A significant correlation was found between MMP-9 expression and an infiltrating pattern of growth (P = .002). In addition, enhanced mRNA expression of MMP-9 was strongly correlated with the presence of lymph node involvement (P<.001), whereas MMP-2, MMP-7, and MMP-11 levels were weakly correlated (P = .04, .03, and .049, respectively). Comparing primary tumors with early lesions (pathological stages I and II) and advanced diseases (stages III and IV), higher expression levels of MMP-2 and MMP-9 were observed in the latter group (P = .01 and <.001, respectively). On the other hand, when patients were separated into 4 groups according to the cutoff values for MMP-2 and MMP-9 expression obtained from the mean mRNA levels determined by RT-PCR in the primary tumors, groups with high MMP-2 and high MMP-9 expression showed the highest incidence of LNM (100%; P = .002; odds ratio, 28.24) and advanced pathological stages (100%; P = .02; odds ratio, 12.78), whereas groups with low MMP-2 and low MMP-9 expression showed the lowest incidence of nodal metastasis (23%; P<.001) and advanced stage (38%; P<.001) compared with other groups (Figure 5). In contrast, there was no association between MMP-1, MMP-2, MMP-9, MMP-11, or TIMP-1 levels and age, sex, site of primary tumors, or histological grade (Table 2). Furthermore, no association between expression of MMP-3, MMP-10, MMP-13, MMP-14, TIMP-2, ratio of MMP-9 to TIMP-1, or ratio of MMP-2 to TIMP-2 and any clinicopathological variables was observed (data not shown).
A number of studies have attempted to delineate which, if any, of the MMPs and TIMPs are required for HNSCC to grow and spread.28 Until now, the predictive value of the MMPs and TIMPs in invasion and metastasis of HNSCC has been controversial, partly because of the varying methods used to detect MMP expression. Because the components of the extracellular matrix are complex, the combined action of various MMPs is essential for the efficient degradation of the structure. Thus, a comprehensive study of the expression of multiple MMPs and their inhibitors (TIMPs) is important for understanding the complex processes by which tumors acquire their invasive and metastatic potential. In the present study, we quantified the expression of a comprehensive set of MMPs and TIMPs previously identified in HNSCC and studied their relationship with clinicopathological variables in an attempt to determine whether overexpression of certain specific proteases could be particularly relevant to progression in this disease.
Using the highly sensitive RT-PCR assay, we studied the expression of all genes of interest within the same tissue samples. Where available, we also confirmed the presence of proteins and enzyme activities using immunoblot, enzyme immunoassay, and substrate zymography. We used tumor margins in the present studies based on the hypothesis that the cellular events in the tumor-stromal interface might be more closely related to the metastatic potential of the tumor than the (often necrotic) center. Coexpression of several members of the MMP family seems to be a general characteristic of human HNSCC. The simultaneous expression of several MMPs is consistent with the concept that extracellular matrix remodeling during tumor progression requires the synergistic action of several proteolytic enzymes produced by tumor cells or by stromal cells.2 Evidence suggests that each of these MMP genes might have a distinct role in tumor progression. Some members of the MMP family, such as MMP-2 and MMP-11, are expressed mainly in stromal fibroblasts and might be regarded as paracrine, stroma-derived factors necessary for the progression of HNSCC.8,11 Both MMP-7 and MMP-9 were expressed exclusively in epithelial cells3,10,11; MMP-1 and MMP-10 were found principally within fibroblasts surrounding tumor, in endothelial cells, and also in neoplastic cells.3- 5,8 Muller et al8 reported that increased MMP-11 gene expression might be a useful marker for defining subpopulations of aggressive HNSCC. Increased expression of MMP-14 was detected at the tumor cell surface, especially at the invasive edge of tumor cell nests, in most HNSCC tissues assayed.13
Although inhibition of in vitro and in vivo tumor invasion by TIMPs has been demonstrated,20,29 increased, rather than decreased, TIMP levels have been shown to be related to poor outcome in several malignant tumors, such as bladder cancer.30 The present finding of increased TIMP-1 expression in HNSCC might be explained by the growth-promoting activity of TIMPs on a variety of cell types31 or the induction of TIMPs by secreted MMPs (or vice versa) from tumor-host interaction in the extracellular milieu. The correlation between increased TIMP-1 and TIMP-2 levels with less aggressive tumors was found in some studies,7,11,17 although the opposite pattern was also reported16,18 (and confirmed in the present study).
Several studies have examined relationships between the expression of gelatinases and malignant potential in HNSCC, but the results are still inconclusive. In oral cancers, MMP-2, but not MMP-9, was found to correlate with LNM and poor clinical outcome.6,10 One study17 found that high levels of MMP-2 and MMP-9 were related to the invasiveness of oral SCC, whereas another11 showed no difference in MMP-2 and MMP-9 levels between primary HNSCC and LNM. Most recently, several studies19- 22 have shown that MMP-9 might play a more important role than MMP-2 in the invasive and metastatic potential of HNSCC. Among 9 MMPs and 2 TIMPs tested in the present study, MMP-9 overexpression showed the strongest correlation with the presence of neck nodal metastases and advanced pathological stages. The underlying roles that some MMP family members play in the process of lymphatic metastasis remain to be elucidated. In addition, it is possible that some MMPs and TIMPs did not show statistical significance because of a type II error (relatively small sample size with high variability), and another study with a larger sample size might be required.
The up-regulation of several MMPs in lymph node–positive patients suggests that the evaluation of MMPs, MMP-9 in particular, in HNSCC tissues at the time of presentation might allow identification of a subset of patients with HNSCC who are more susceptible to metastatic spread via lymphatic pathways and permit therapy to be offered accordingly. With the application of highly sensitive RT-PCR analysis, preoperative assessment from small tissue biopsies or even needle aspirates will become more useful in assessing the malignant potential of HNSCC. Aggressive neck management in tumors showing multiple MMPs positive, MMP-9 in particular, might be considered to avoid later lymphatic spread. Because of the relatively short follow-up, we are unable to demonstrate yet whether MMP/TIMP expression is related to survival. This awaits confirmation in a longer follow-up period.
In conclusion, the results of these studies suggest that expression of multiple MMPs and TIMPs is characteristic of HNSCC and that no specific member of the MMP family is solely responsible for HNSCC progression. The correlative studies of MMP/TIMP expression in human head and neck tumor tissues suggest the potential role of MMP-2, MMP-7, MMP-9, and MMP-11 in progression and metastasis of human HNSCC. Combined analysis of these MMPs, MMP-9 in particular, might be useful in evaluating the malignant potential in individual HNSCC.
Accepted for publication March 27, 2001.
This project was supported by a Fellowship from the Faculty of Medicine, Siriraj Hospital Medical School (Dr O-charoenrat), and by the Siriraj Hospital Fund, Bangkok, Thailand.
Presented at the annual meeting of the American Head and Neck Society, Fifth International Conference on Head and Neck Cancer, San Francisco, Calif, August 1, 2000.
We are grateful to all staff in the Department of Head and Neck Surgery and operating theaters at the Royal Marsden Hospital for help with collecting the clinical specimens. We also thank Gary M. Box, BSc, and William J. Court, MSc, for excellent technical assistance.
Corresponding author and reprints: Pornchai O-charoenrat, MD, Division of Head and Neck Surgery, Department of Surgery, Siriraj Hospital Medical School, Bangkok 10700, Thailand (e-mail: firstname.lastname@example.org).