Immunoexpression and prognostic significance of TIMP-1 and -2 in oral squamous cell carcinoma

Oral Oncology (2005) 41, 568–579

http://intl.elsevierhealth.com/journals/oron/

Immunoexpression and prognostic significance of TIMP-1 and -2 in oral squamous cell carcinoma Juan Carlos de Vicente a,*, Manuel Florentino Fresno b, Lucas Villalain c, ´n Lo ´pez Arranz a Jose Antonio Vega d, Juan Sebastia a

Department of Oral and Maxillofacial Surgery, Escuela de Estomalogia, University Hospital of Oviedo, c/Catedratico Jose Serrano, s/n 33006, Oviedo, Spain b Department of Pathology, University Hospital of Oviedo, Oviedo, Spain c ´n, Spain Department of Oral and Maxillofacial Surgery, Hospital of Virgen Blanca, Leo d Departamento de Anatomı´a y Embriologı´a, Facultad de Medicina, Universidad San Pablo-CEU, Madrid Received 16 December 2004; accepted 22 December 2004

KEYWORDS

Summary Matrix metalloproteinases (MMPs) are proteolytic enzymes that are capable of degrading different substrates within the extracellular matrix, and which are believed to be crucial for tumor invasion and metastasis. Tissue inhibitors of MMPs (TIMPs) can inhibit the action of MMPs but also can show a paradoxical poor prognostic effect. In order to evaluate the prognostic significance of TIMPs, we studied the expression of TIMP-1 and -2 in series of 68 oral squamous cell carcinomas (OSCC) by immunohistochemistry. Expression of TIMP-1 was detected in 45 cases (66.2%). In all of these TIMP-1 was expressed in tumoral tissue, and in 19 of them also in the surrounding stroma. In cancer tissue, TIMP-1 was observed in three patterns: homogeneous, central and irregular. Immunoreactivity for TIMP-2 was detected in 38 cases (56%) in tumoral tissue and 9 (13.2%) in the stroma. The expression pattern of TIMP-2 was the same three as TIMP-1 and one more: invasive front of tumoral nests. TIMP-1 expression was not correlated with clinical or pathological parameters. However, TIMP-2 was significantly correlated with T stage (p = 0.03), TNM stage (p = 0.01), local recurrence (p = 0.04), and poor survival (p = 0.03, odds ratio = 2.75). TIMP-1 and TIMP-2 were significantly correlated with cyclin D1 (p = 0.04; p = 0.015, respectively) and p53 expressions (p = 0.02; p = 0.04, respectively). Finally, TIMP-1 but no TIMP-2 was associated with the nuclear antigen Ki-67 (p = 0.001).

Immunohistochemistry; TIMP-1; TIMP-2; p53; Cyclin D1; Ki-67; Oral squamous cell carcinoma; Prognosis

* Corresponding author. Tel.: +34 985 103638; fax: +34 985 103673. E-mail address: [email protected] (J.C. de Vicente).




1368-8375/$ - see front matter c 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.oraloncology.2004.12.008

Immunoexpression and prognostic significance of TIMP-1 and -2

569

These results suggest that TIMP-1 and -2 are expressed in tumoral and stromal tissue in OSCC. TIMP-2 is related to advanced disease, recurrence and poor prognosis. c 2005 Elsevier Ltd. All rights reserved.




Introduction Oral squamous cell carcinoma (OSCC) is the sixth most common cause of cancer-related death, mainly by cervical lymph node metastasis, and occasionally by distant-organ metastasis. The process of metastasis in OSCC is based on the ability of epithelial cells to degrade the basement membrane and the adjacent extracellular matrix (ECM). The matrix metalloproteinases (MMPs) are a family of multidomain Zn2+-dependent proteolytic enzymes that are capable of degrading different substrates within the ECM,1,2 both in physiological and pathological conditions such as embryonic development, wound healing, angiogenesis, arthritis, inflammation, and tumor metastasis.3 They are classified into four groups according to their substrate specificity: collagenases (MMP-1, -8, and -13), gelatinases (MMP-2 and -9), stromelysins (MMP-3, -10 and -11), and membrane-type MMPs (MT-MMPs) (MMPs 14–17, 24, and 25). MMPs are produced by cancer cells or by induction in surrounding stromal cells. This suggests that tumor cells are capable of utilizing MMPs produced by stromal cells and indicates an active role for stroma in tumor invasion.4 The activity of MMPs can be regulated at the level of gene transcription, by proenzyme activation, and by tissue inhibitors of metalloproteinases (TIMPs). TIMPs consist of four homologous members identified to date5 and among them, TIMP-1 and -2 are the best known.6,7 TIMPs are multifunctional proteins that inhibit cell invasion in vitro and can suppress angiogenesis, tumorigenesis, tumor invasion and metastasis in vivo.8,9 Although each TIMP appears capable of inhibiting several MMPs, these proteins exhibit preferential inhibitory capacity. TIMP-1 and -2 have molecular weights of 28.5 and 21.0 kd, and inhibit gelatinases B and A, respectively, as well as other MMPs by forming 1:1 stoichiometric complexes with the active MMPs.5,10 But TIMPs have not only MMPs inhibitory activities. In the context of tumor invasion, originally it was considered that TIMPs had a role in inhibition of MMPs, thus serving as anti-invasive factors. However, more recently, a multifunctional complex role was considered for TIMPs which could show prognostic relevance. Furthermore, most of the MMPs, except MT-MMPs and

MMP-11 (stromelysin 3), are secreted as inactive zymogens and are activated extracellularly by serine proteinases11 that cleave their amino-terminal domains. Similarly to other MMPs, MMP-2 (gelatinase A) is secreted in an inactive zymogen form (pro-MMP2), but its conversion to an active form (active-MMP2) is due to membrane type 1-metalloproteinase (MT1-MMP or MMP-14).12 In fact, proMMP-2 binds to TIMP-2 in combination with MT1-MMP on the cell surface, forming a ternary complex. Then, pro-MMP-2 in the complex is activated by adjacent MT1-MMP that is free from TIMP-2.13 Thus, although previous studies have shown that TIMPs have antiangiogenic and antimetastatic potential, more recent reports indicate a dual function with positive correlation between increased TIMP levels and poor outcome in colorectal and breast carcinomas.14,15 Although treatment guidelines have been developed according to clinical (TNM stage, location) and histological (grade of differentiation, depth of invasion) parameters, these parameters do not always predict clinical outcome accurately. For example, patients of the same TNM groups have different survival rates. Based on the fact that death is often the result of local recurrence and distant metastasis, molecular factors related to the potential of invasion and metastasis such as MMPs and TIMPs could predict prognosis of OSCC. The purpose of this study was to evaluate the immunoexpression of TIMP-1 and TIMP-2 in a series of 68 OSCC, and to determine whether the expression of these proteins correlates with clinicopathologic and prognostic variables. In addition, the cell proliferation and the expression of genes involved in tumor progression16 such as cyclin D1 and p53 were analyzed and correlated with those of TIMP-1 and -2.

Materials and methods Patients This study is based on a retrospective cohort of 68 patients suffering from a primary OSCC who were diagnosed at the Department of Oral and Maxillofacial Surgery, Asturias University Hospital, Oviedo, Spain, between January of 1990 and December of

570 Table 1

J.C. de Vicente et al. Patient characteristics

Variable

Number (68)

%

Age (years) Mean (range)

60 (27–87)

Gender Male Female

53 15

77.9 22.1

Tumor site Lip Tongue Floor of the mouth Gingiva Buccal mucosa Palate

4 29 17 8 6 4

5.9 42.6 25.0 11.8 8.8 5.9

T-category T1 T2 T3 T4

20 29 3 16

29.4 42.6 4.4 23.5

N-category N0 N1 N2 N3

45 13 10 0

66.2 19.1 14.7 0.0

Stage I II III IV

14 19 12 23

20.6 27.9 17.6 33.8

Histologic grade Poorly Moderately Well

2 15 51

2.9 22.1 75.0

1992. Inclusion criteria were surgical treatment performed according to standard procedures and consisting of the resection of the primary tumor and a radical or selective ipsi- or bilateral neck dissection, complete clinicopathologic data and availability of sufficient paraffin-embedded tumor material. Clinicopathologic information on each case, including age, gender, smoking and alcohol intake history, tumor size, nodal status, location, histologic grade, treatment, and presence or absence of tumor recurrence was obtained from patient files. There were 15 (22%) women and 53 (78%) men, ranging in age from 27 to 87 years (mean, 59.8 years). The main clinical characteristics of the 68 patients selected for this study are detailed in Table 1. All patients were staged according to the 1997 UICC TNM Classification of Malignant Tumors.17 The follow-up period ranged from 4 to 128 months (average: 56.4 months). All

patients had been treated surgically with curative intention, and 29 (42.6%) underwent postoperative radiotherapy, receiving 40–70 Gy, according to the following indications: T4, poorly differentiated carcinomas, tumor-positive resection margins, and metastases in neck nodes. Surgical margins of resection were free of tumor infiltration in 63 of 68 cases (92.6%). Extracapsular extension was observed in 6 of 23 cases (26%) of neck node metastases. The recurrence-free survival and the survival of the patients were quantified. Clinical outcome was measured by two end points: death caused by disease recurrence and non-treatable disease presence at the end of the follow-up time. At the end of this period 27 patients (39.7%) had died of tumoral recurrence, and 41 cases (60.3%) were alive and free of recurrence.

Immunohistochemistry Immunohistochemistry was performed on paraffin sections of the 68 OSCC mounted on glass slides. Surgical specimens were fixed in 10% neutralbuffered formalin (pH 7.4) at 4
C for 72 h and then embedded in paraffin. Four-lm-thick tissue sections were mounted on poly-L-lysine-coated slides. Briefly, the sections were dewaxed with xylene, and rehydrated in graded ethanol. Endogenous peroxidase activity was blocked by immersion of slides in methanol with 0.03% hydrogen peroxide for 30 min. The sections were then heated in 10 mM citrate buffer (pH 6.0) three times for 10 min each in a microwave oven at 500 W to retrieve antigenicity. The sections were rinsed in distilled water, and then in phosphate-buffered saline (PBS). Non-specific conjugation was blocked with a solution of 20% rabbit serum (DAKO, Glostrup, Denmark) applied to the sections for 10 min. The sections were incubated with primary antibodies against TIMP-1 (mouse monoclonal anti-TIMP-1, clone 63515, RD systems, diluted 1:40) and TIMP-2 (mouse monoclonal anti-TIMP-2, clone 89025.11, RD systems, diluted 1:40), for 30 min at room temperature. A randomized group of 26 patients of the whole sample was also incubated with mouse monoclonal antihuman cyclin D1 (clones P2D11F11 and DCS-6, Novocastra, diluted 1:25), MIB-1 monoclonal antibody against Ki-67 nuclear antigen (Immunotech, diluted 1:50), and mouse monoclonal anti-p53 (clone DO-7, Dako, diluted 1:50). They were then rinsed in PBS, and bound antibody was detected using EnVision polymer technology (K4001, DAKO, Carpinteria, CA) for 30 min according to the manufacturer’s instructions. After washing in PBS, staining was developed with 3,30 -diaminobenzi-

Immunoexpression and prognostic significance of TIMP-1 and -2 dine-tetrahydrochloride in 50 mM Tris–HCL (pH 7.5) containing 0.001% hydrogen peroxide for 5 min, and then lightly counterstained with Mayer’s hematoxylin. One positive and one negative control were included in each batch of immunostained sections. Formalin-fixed, paraffin-embedded samples of breast carcinoma were used as positive controls for TIMP-1 and TIMP-2. For negative control, the primary antibody was replaced with non-immune mouse serum. All slides were scored by two investigators without knowledge of the clinical outcome. Occasional disagreements were discussed to reach a consensus. In cases of persistent differences between

571

them, the sections were studied by a third independent observer and the majority decision was thus considered. Staining for TIMP-1 and TIMP-2 was measured as the percentage of positively stained nuclei, and assigned to four categories, whereby the intense immunoreaction observed in breast carcinoma served as a positive control. The degree of staining was scored as follows: 0, negative; 1, less than 10%; 2, more than 10% and less than 50%; 3, more than 50% positive staining. Group 0 was defined as negative TIMP-1 or TIMP-2 expression, and groups 1, 2, and 3 as positive TIMP-1 or TIMP-2 immunoexpression. Areas with pronounced inflammation or necrosis were avoided. Staining

Figure 1 Immunoexpression of TIMP-1 and TIMP-2 in oral squamous cell carcinoma. (A) TIMP-1 is immunolocalized in both cancer and stromal cells (original magnification · 100); (B) TIMP-1 diffusely localized in the carcinoma cells (original magnification · 500); (C) representative example of TIMP-2 immunostaining in oral squamous cell carcinoma (original magnification · 100); (D) the immunoreactivity to TIMP-2 was expressed in cancer cells (original magnification · 500) as well as in the stromal cells. Expression of these proteins was scored as described in Materials and Methods.

572

Statistical analysis An SPSS for Windows computer program (Release 11.5, SPSS Inc., Chicago, IL) was used for statistical analyses. The association between clinical parameters (tumor size, regional lymph node status, tumor location, and histopathologic grade) and immunohistochemical results was analyzed with the chi-square or Fisher’s exact test (if N < 5). To analyze the correlation between the immunoexpressions of TIMP-1 and TIMP-2, and TIMPs and cyclin D1, the Spearman correlation coefficient was used. Differences between means of p53 immunostained cells in TIMP (-1, and -2) positive and negative cases, were tested using the Student t-test preceded by Levene’s test. Survival analysis was performed with the Kaplan– Meier product limit method and the log rank test, used to compare survival among groups of patients. A significance level of 0.05 was used for all statistical tests.

Results Immunohistologic expression of TIMP-1 and TIMP-2 in primary OSCC Immunoreactivity for TIMP-1 was detected in 45 of 68 cases (66.2%). Of these 45 cases, 11 cases (16.2%) strongly expressed TIMP-1; 16 (23.5%) showed immunostaining between 10% and 50% of cancer cells, and finally, 18 (26.5%) expressed less than 10% of positive cell staining. The remaining 23 cases (33.8%) were negative for TIMP-1. The 45 cases that expressed TIMP-1 showed immunoreactivity in tumoral tissue, and 19 (28%) expressed TIMP-1 also in the surrounding stroma. In cancer nests and cords, TIMP-1 was observed in three patterns, i.e., 26 cases (38.2%) expressed TIMP-1 in a homogenous picture, 8 (11.8%) showed immunoreactivity only in the central part of the cancer nests, and finally, 11 (16.2%) showed TIMP-1 reactivity in a scattered or irregular pattern (Fig. 1). Immunoreactivity

1.1 1.0 .9

Cumulated Survival

for p53 was measured as the percentage of positively stained nuclei counting of at least 700 tumor cells in consecutively chosen fields. Cyclin D1 and MIB-1 were scored into four groups as follows: 0, negative; 1, low (less than 25%); 2, intermediate (more than 25% and less than 50%); and 3, high (more than 50% of positively stained nuclei).

J.C. de Vicente et al.

.8 .7

TIMP-1

.6

Positive .5 Censored .4

Negative Censored

.3 0

20

40

60

80

100

120

140

Time

Figure 2 Survival curve for patients with OSCC according to the immunostaining of TIMP-1 (p = 0.22).

for TIMP-2 was detected in 38 of 68 cases (56%). Of these 38 cases, 8 cases (11.8%) strongly expressed TIMP-2; 12 (17.6%) showed immunostaining in 10–50% of cancer cells, and finally, 18 (26.5%) expressed less than 10% of positive cell staining. The remaining 30 cases (44.1%) were negative for TIMP-2. The 38 immunopositive cases expressed TIMP-2 in cancer nests, and 9 (13.2%) also in adjacent stroma. In tumoral tissue, TIMP-2 showed four patterns of immunoreactivity. Nine cases (13.2%) expressed this protein in a homogenous pattern, 16 (23.5%) only in the central part of the cancer nests, and 12 (17.6%) in an irregular pattern. Finally, one case (1.5%) expressed TIMP-2 only in the invasive front of tumoral areas (Fig. 2). The expression of TIMP-1 correlated significantly with that of TIMP-2 (r = 0.5; p = 0.001).

Immunohistologic expression of TIMP-1 in OSCC with regard to clinicopathological data The TIMP-1 immunoexpression according to the localization of the tumor showed no difference of statistical significance. Moreover, we could not find a correlation between the TIMP-1 expression level and the age, gender, T-stage, N-stage, TNM stage, histological grading, tumor recurrence, smoking habit, and alcohol consumption (Table 2). Of the 23 cases with lymph node metastasis, 16 cases (69.6%) expressed TIMP-1,

Immunoexpression and prognostic significance of TIMP-1 and -2 Table 2

573

Relationship between TIMP-1 and TIMP-2 expressions with clinicopathological factors

TIMP-1 (+)

TIMP-2 (+)

Total number

Number

%

p

Number

%

p

Age \60 P60

33 35

21 24

63.6 68.6

0.43

17 21

51.5 60.0

0.49

Sex Female Male

15 53

11 34

73.3 64.2

0.55

9 29

60.0 54.7

0.72

Tumor location Lip Tongue Floor Gum Palate Cheek

4 29 17 8 4 6

2 19 12 4 3 5

50.0 65.5 70.6 50.0 75.0 83.3

0.8

2 14 12 4 3 3

50.0 48.3 70.6 50.0 75.0 50.0

0.70

Tumor status T1 T2 T3 T4

20 29 3 16

11 20 3 11

55.0 69.0 100 68.8

0.54

10 13 2 13

50.0 44.8 66.7 81.3

0.03

Nodal status pN0 pN1 pN2 pN3

45 13 10 0

29 8 8

64.4 61.5 80.0

0.65

25 5 8 0

55.6 38.5 80.0

0.14

Nodal status N0 N+

45 23

29 16

64.4 69.6

0.67

25 13

55.6 56.5

0.93

TNM stage I II III IV

14 19 12 23

8 11 9 17

57.1 57.9 75.0 73.9

0.59

6 8 6 18

42.9 42.1 50.0 78.3

0.01

Tumor recurrence No 40 Yes 28

24 21

60.0 75.0

0.29

19 19

47.5 67.9

0.04

Histological grade Poor 2 Moderate 15 Well 51

1 9 35

50.0 60.0 68.6

0.59

1 6 31

50.0 40.0 60.8

0.38

Tobacco No Yes

22 46

17 28

77.3 60.9

0.27

13 25

59.1 54.3

0.79

Alcohol No Yes

24 44

18 27

75.0 61.4

0.29

14 24

58.3 54.5

0.80

with 5 (21.7%) belonging to group 3, 5 (21.7%) to group 2, and 6 (26.1%) to group 1. Of the 45 cases without lymph node metastasis 29 (64.4%) expressed TIMP-1. Performing Kaplan–Meier anal-

ysis for TIMP-1 expression, statistical significance was not reached (p = 0.22). Therefore, TIMP-1 immunoexpression was not related to survival of the patients (Fig. 2).

574

J.C. de Vicente et al. 95% confidence interval from 55 to 90 months) had a longer survival (p = 0.03 by log-rank test; Odds Ratio = 2.75).

1.1 1.0

Cumulated Survival

.9

Correlation between the immunoexpressions of TIMP-1 and -2 and cyclin D1

.8 .7

The correlation of TIMP-1 and TIMP-2 with cyclin D1 was examined. The expression of cyclin D1 correlated significantly with that of TIMP-1 (r = 0.4; p = 0.04) and with that of TIMP-2 (r = 0.47; p = 0.015) (Table 3).

TIMP- 2

.6

Positive Censored

.5

Negative .4

Censored

Relation between the immunoexpressions of TIMP-1 and -2 and p53

.3 0

20

40

60

80

100

120

140

Time

Figure 3 Survival curve for patients with OSCC according to the immunostaining of TIMP-2 (p = 0.03).

Immunohistologic expression of TIMP-2 in OSCC with regard to clinicopathological data Immunoexpression of TIMP-2 did not correlate with the patient’s age or sex, location of the tumor, N-stage, histological grading, smoking habit, or alcohol consumption (Table 2). Of the 23 cases with lymph node metastasis, 13 cases (56.5%) expressed TIMP-2, with three belonging to group 3 (13%), three (13%) to group 2, and seven (30.4%) to group 1. Of the 45 cases without lymph node metastasis 25 (55.6%) expressed TIMP-2. However, TIMP-2 immunoreactivity was significantly correlated with T (p = 0.03) and TNM stages (p = 0.01), and also with local recurrence (p = 0.04). Performing Kaplan–Meier analysis, positive TIMP-2 immunostaining was related with poor survival rates (Fig. 3). Thus, patients with TIMP-2 negative tumors (mean survival: 99 months—95% confidence interval from 83 to 116 months) versus TIMP-2 positive ones (mean survival: 72 months; Table 3

The average of cells positively stained for p53 was of 48.95 ± 40.8% (mean ± standard deviation; median, 50) in TIMP-1 negative cases and of 26.1% ± 32.8% (mean ± standard deviation; median, 10) in the TIMP-1 positive ones. The difference between both groups was statistically significant (t = 2.3; p = 0.02; 95% confidence interval of the difference from 2.8% to 42.8%). The mean of cells that showed positive immunostaining for p53 in TIMP-2 negative cases was of 44.57% ± 40.14% (mean ± standard deviation; median, 45) and in TIMP-2 positive ones was of 26.37% ± 33.24% (mean ± standard deviation; median, 10). The difference between both means reached, as was noted with TIMP-1, statistical significance (t = 1.95; p = 0.04; 95% confidence interval of the difference from 0.13% to 36.27%).

Relation between the immunoexpressions of TIMP-1 and -2 and Ki-67 The cell proliferation measured by Ki-67 expression was associated with TIMP-1 (Fisher exact test = 13.8, p = 0.001) (Table 4), but not with TIMP-2 immunoexpression (Fiser exact test = 6.3, p = 0.07).

Association between TIMP-1 and -2 and Cyclin D1

Cyclin D1

Negative Low Intermediate High Total

TIMP-1

TIMP-2

Negative

Positive

Negative

Positive

0 3 3 2 8

4 8 6 0 18

0 4 5 2 11

4 7 4 0 15

Immunoexpression and prognostic significance of TIMP-1 and -2 Table 4

575

Association between TIMP-1 and Ki-67 immunoexpression

Ki-67

TIMP-1 negative

TIMP-1 positive

Total

Negative Low Intermediate High Total

1 5 0 2 8

1 1 12 4 18

2 6 12 6 26

Discussion In the present study, we have shown, using immunohistochemistry, the expression and distribution of TIMP-1 and TIMP-2 in primary OSCC. Immunohistochemistry was chosen as the method of study because it allows a direct and visual localization of TIMP-1 and -2 and a correlation with morphology, especially with the cellular events on the tumor–stromal interface, and with clinical parameters and outcome. Furthermore, immunohistochemistry can be performed on paraffin embedded specimens of patients followed by a long lifetime span. Tumor invasion is a complex biological process, during which tumor cells detach from the primary tumor and infiltrate the surrounding tissue. This process requires loss of cells contacts between tumor cells, active cell migration, adhesion to ECM and proteolytic degradation of the ECM. On the molecular level, several different molecules, including cadherins, integrins, proteases and their inhibitors, and growth factors, have been implicated in the regulation of cancer invasiveness.18 The matrix metalloproteinases (MMPs) are commonly expressed in invasive tumors and adjacent stroma, and it is thought that they play an important role in tumor invasion and metastasis. In situ hybridization techniques showed that most MMPs are not produced by tumor cells but by neighbouring stromal cells. It is suggested that tumor cells produce a stimulatory factor (extracellular matrix metalloproteinase inducer, EMMPRIN) that induces stromal fibroblasts to produce MMPs.19 The activities of the MMPs are modulated and counterbalanced, in part, by endogenous tissue inhibitors of metalloproteinases (TIMPs). TIMPs bind to the MMPs and inhibit their action. When an imbalance exists between the activation and inhibition of MMPs, in favor of the MMP activity, degradation of ECM might ensue playing an important role in the pathophysiology of cancer facilitating the invasion of tumor cells through the ECM20 as well as the process of angiogenesis.21 Degradation of interstitial collagen in ECM is accomplished by many pro-

teinases, such as MMP-1, -8, -13, and MMP-2 and MT1-MMP.22 On the basis of these observations it could be suggested that the degradation of ECM is not dependent on the expression of a single MMP, and that the combined action of several MMPs and TIMPs is essential for the efficient degradation of the basement membrane and the interstitial stroma. Thus, TIMPs inhibit the proteolytic activity of MMPs, but on the other hand, TIMP-1 and -2 are the homologues of erythroid potentiating activity (EPA) factors which stimulate the growth of erythroid precursor cells,23 and also stimulate the growth of different types of cells, such as fibroblasts, chondrocytes, several cancer cell lines and melanoma cells.24,25 Previous studies on MMPs and TIMPs expression in tumors have been carried out, but no consensus has been reached as to their prognostic importance in OSCC (Table 5).10,13,26–36 And importantly, the documented paradoxical functions of TIMPs have not been characterized in OSCC. The correlation between increased TIMP-1 and TIMP-2 levels with less aggressive tumors was found in some studies,26,28,37 but the opposite pattern was also reported.10,13,27,29,32–34 To our knowledge, no previous studies considering TIMP1 or -2 as isolated prognostic factors in OSCC, have been performed. In OSCC, a low TIMP-1 gene expression has been reported in fibrous connective tissue38 and also in well-differentiated cancer cells and in endothelial cells, whereas TIMP-2 mRNA was localized in stromal cells near the invasive front of cancer nests and in endothelial cells as well.37 Sutinen et al.,27 using immunohistochemistry in 10 OSCC, observed that a weak positive staining reaction for TIMP-1 was located in tumoral stroma, whereas the staining for TIMP-2 was negative in all carcinoma and lymph node metastasis studied. In the present study, we observed a positive staining for TIMP-1 in 45 of 68 cases (66.2%) and not only in cancer cells but also in stromal cells. In all these 45 cases the staining was located in tumoral cells, and in 19 (28% of the whole sample) of them an additional staining was observed in surrounding stroma. Immunopositivity

TIMP-1 and/or TIMP-2 expression and clinical significance in oral squamous cell carcinoma and other head and neck carcinomas, reported in the literature Sample size/ country

Method

TIMP-1 positivity (%)/location

TIMP-2 positivity (%)/location

Association with clinicopathological parameters

Charous et al., 199726

29\ USA

In situ hibridation

95/stroma and endothelial cells

Not studied

Not found

Sutinen et al., 199827

10 Finland

Immunohistochemistry In situ hibridation

+/stroma +/stroma

Not detected Negative

Not studied

Ikebe et al., 199928

57 Japan

ELISA

+

Not studied

Higher levels of TIMP-1 in non-metastatic cases

Kuraharaet al., 199929

96 Japan

Immunohistochemistry

75/stroma

87.5/stroma

TIMP-1 was significantly increased in metastatic cases. No significant relationship was observed between TIMP-2 expression and nodal involvement

Birkedal-Hansen et al., 200030

20\\ USA

RT-PCR

Similar levels in tumor and normal tissues

Similar levels in tumor and normal tissues

Not studied

Shimada et al., 200031

24 Japan

Immunohistochemistry EIA

Not studied Higher levels in tumoral tissue

+/tumor Similar levels in tumor and normal tissues

Not found

Liu et al., 200132

30 China

In situ hibridation

+/tumor and stroma

+/tumor and stroma

Neck lymph node metastasis

O-charoenrat et al., 200110

54\\\ Thailand, England

RT-PCR

mRNA expression was significantly greater in malignant tissues

No significant difference was found between TIMP-2 levels in tumors and normal tissues

Higher T status (T3-T4) (only TIMP-1)

Yoshizaki et al., 200133

51 Japan

Immunohistochemistry

Not studied

43.1/tumor

Nodal status Clinical stage (not with tumor status) Poor survival MMP-2 immunoexpression

Culhaci et al., 200434

78\\\\ Turkey

Immunohistochemistry

53.8/stroma

Not studied

Invasive histology pattern

23 Israel

Immunohistochemistry

65.2

Not studied

Not found

Katayama et al., 200413

53 Japan

Immunohistochemistry

Not studied

88.7/tumor

Regional lymph node and distant metastasis Poor survival MMP-2 immunoexpression

Schmalbach et al., 200436

102\\\\\ USA

Immunohistochemistry

Higher levels in tumoral tissue

Not studied

Not found

All reports concern to OSCC, except * Supraglottis, oropharynx, hypopharynx, larynx, oral cavity (4), and maxillary sinus; ** pharynx, larynx, esophagus, skin; ***oral cavity (14), oropharynx, hypopharynx, and larynx, **** oral cavity, oropharynx, larynx, skin. ISH, in situ hibridation; RT-PCR, semiquantitative reverse transcription-polymerase chain reaction; ELISA, enzyme linked immunoadsorbent assay; EIA, sandwich enzyme immunoassay.

J.C. de Vicente et al.

Guttman et al., 2004

35

576

Table 5

Author/year (Reference)

Immunoexpression and prognostic significance of TIMP-1 and -2 for TIMP-2 was detected in 38 cases (56%) in cancer cells, and 9 (13.2%) in tumoral stroma. These findings suggest that OSCC is heterogeneous for the potential to produce TIMP-1 and TIMP-2. Sawatsubashi et al.20 examined 83 laryngeal squamous cell carcinomas using immunohistichemistry for TIMP-1, and 20 of them by in situ hybridization. They observed that the localization of this protein is similar to that of their transcript. These authors20 found TIMP-1 expression in 49 (59%) cases in cancer cells as well as stromal cells such as fibroblasts, macrophages, and mononuclear and endothelial cells. They did not find correlation between TIMP-1 staining and age, T category, nodal status, distant metastasis, or cancer stage. However, they found that TIMP-1 staining was correlated significantly with histologic type, and there was a trend toward the well-differentiated type in carcinomas with positive TIMP-1 expression compared with those with negative TIMP-1 expression. In the present study, we found also a trend toward the well-differentiated carcinomas in TIMP-1 immunoexpression, but we did not reach statistical significance (Table 2). TIMP-2 has been believed to suppress tumor invasion by inhibiting MMPs. However, TIMP-2 has also an important role in the activation mechanism of MMP-2 forming a pro-MMP-2/TIMP-2/MT1-MMP ternary complex13 on the cell surface, promoting hydrolysis or pro-MMP2 to its active form (MMP2) and resulting in degradation of ECM. It has also been reported that some TIMPs can directly affect cell growth and or cell survival independent of their actions of MMPs.5 The expression of TIMP-2 has been correlated with the metastatic ability and poor prognosis of the tongue squamous cell carcinoma,3 and some studies reported the positive role of TIMP-2 in tumor metastasis and in decreased survival of patients.39,40 A relationship between overexpression of TIMP-1 or -2 and a higher stage of disease progression in different types of lung cancer has been reported.41,42 However, other43 did not report influence of TIMP-2 expression on the survival time of patients with laryngeal squamous cell carcinoma. The findings of the current study strongly support that TIMP-2 plays a positive role in the tumor recurrence and patient survival in OSCC. Our results suggest that not only cancer cells but also stromal cells have a significant role in the prognosis of patients with OSCC. Thus, the treatment of OSCC should be targeted not only against the tumor cells but also against the stromal cells that may be related with the invasive behavior of the cancer and with its local recurrence. Katayama et al.13 showed signifi-

577

cantly shorter overall survival in patients afflicted of OSCC with marked expression of TIMP-2 than patients with negative or moderate expression of TIMP-2. Kanayama et al.39 stated that a high mRNA level of TIMP-2 was associated with poor prognosis in bladder carcinoma. Ross et al.44 immunohistochemically studied 138 prostatic adenocarcinomas using monoclonal antibodies against MMP2 and TIMP-2. They found that TIMP-2 expression individually correlated with advanced tumor stage and reached near significance with disease recurrence. We observed these same results in OSCC. Kallakury et al.5 observed that TIMP-1 ad -2 immunostaining correlate with poor prognosis and shortened survival in renal cell carcinoma. Thus, taking together all these findings, we could hypothesized that, contrary to the previously documented anti-tumor effects of TIMPs, TIMP-2 immunoexpression could have a tumor-promoting role in tumor recurrence and poor prognosis. Although in the present study no significantly differences were observed in the expression of TIMP-1 and TIMP-2 between the cases with neck node metastasis and those without them, the expression of TIMPs were higher in the metastatic cases than in non-metastatic ones (Table 2). These results disagree with those of Ikebe et al.28 which observed that the levels of TIMP-1 were higher in the non-metastatic cases than in the metastatic cases among 57 cases of OSCC. Despite our sample size is higher than many other studies (Table 5) it is possible that the lack of statistical significance between TIMP-1 or -2 expression and lymph node metastasis could be due to a type II error, and another study with a larger sample size might show a different result. To explore the relationship between TIMP-1 and -2 and other molecular markers of metastasis and poor prognosis, we studied cyclin D1, p53 and proliferative cell activity through the antigen ki67, using immunohistochemistry. The cyclin D1 gene encodes a protein forming a complex with cyclin-dependent kinases (CDKs), CDK4 and CDK6. Cyclin D-CDK4/CDK6 complexes phosphorylate Rb protein during the G1-S transition. Thus, cyclin D1 overexpression leads to an acceleration of G1 progression entering the cell in the phase S.45 The proliferation capacity of cancer cells have prognostic significance and can be studied by the Ki-67, a nuclear basic protein associated with cell proliferation and expressed in all phases of the cell cycle, except in G0.46 The p53 gene is the best investigated tumor suppressor gene. Mutations of p53 gene lead to a non-functional protein lacking tumor-suppressive activity.

578 Abnormalities of this gene are common in human cancers, suggesting that this event plays a crucial role in the onset or progression of many types of cancer, and it has been demostrated a prognostic role of this gene.47 In the present study we observed that TIMP-1 is significantly associated with the cell proliferation index, which represents a support of the paradoxical effect of TIMPs in a hypothetical role of stimuli of growth cell (the association between TIMP-2 and Ki-67 is close to statistical significance level). On the other hand, these two TIMPs are associated with cyclin D1 and p53. Cyclin D1 tends to be overexpressed when TIMP-1 and -2 are negatives, and the average of tumoral cells positively stained for p53 was higher when both TIMPs were also negatives. Taken together these results strongly support the hypothesis that TIMPs are not only inhibitors of MMPs but they are probably involved on the growth of cells of OSCC acting a several points in tumorigenicity and cancer progression. TIMP-2 shows a poor prognostic significance in univariate analysis, but the reduced expression of TIMP-1 or -2 is related with the overexpression of cyclin D1 and p53 proteins, both related with the progression of the cell cycle. Further studies to elucidate exactly how the function of TIMPs could be related with the cell proliferation in OSCC might be useful in evaluating the potential progression of an individual tumor. In summary, based on our results, we conclude that both cancerous and stromal cells have the capacity to express TIMPs, and it seems that TIMPs-1 and -2 expressions, measured by immunohistochemistry, are involved in extracellular matrix remodelling during cancer invasion and this event could have prognostic significance. Specifically, TIMP-2 expression is related with biological aggression in OSCC, and specifically to local recurrence and poor prognosis. This paradoxical poor prognostic significance warrants further studies to fully understand its biological role in OSCC.

Acknowledgement The authors thank Ms. Aurora Ferna ´ndez Garcı´a for technical assistance. We are grateful to Dr. Jonas Hannestad (Department of Psychiatry, Duke University) for help in reviewing the manuscript for English syntax. This work was supported by a grant for scientific research from the Ministry of Health, Spain (Instituto de Salud Carlosiii, PI020137).

J.C. de Vicente et al.

References 1. Mignatti P, Rifkin DB. Biology and biochemistry of proteinases in tumor invasion. Physiol Rev 1993;73:161–95. 2. Stettler-Stevenson WG, Aznavoorian S, Liotta LA. Tumor cell interactions with the extracellular matrix during invasion and metastasis. Annu Rev Cell Biol 1993;9:541–73. 3. Matrisian L. The matrix degrading metalloproteinases. Bioessays 1992;14:455–9. 4. Thomas GH, Lewis MP, Speight PM. Matrix metalloproteinases and oral cancer. Oral Oncol 1999;35:227–33. 5. Kallakury BVS, Karikehalli S, Haholu A, Sheehan CE, Azumi JS, Ross JS. Increased expression of matrix metalloproteinases 2 and 9 and tissue inhibitors of metalloproteinases 1 and 2 correlate with poor prognostic variables in renal cell carcinoma. Clin Cancer Res 2001;7:3113–9. 6. Welgus HG, Stricklin GP. Human skin fibroblast collagenase inhibitor. Comparative studies in human connective tissues, serum, and amniotic fluid. J Biol Chem 1983;258:12259–64. 7. Stetler-Stevenson WG, Krutzcsh HC, Liotta LA. Tissueinhibitor of metalloproteinase (TIMP-2). A new member of the metalloproteinase inhibitor family. J Biol Chem 1989;264:17374–8. 8. Liotta LA, Steeg PS, Stetler-Stevenson WG. Cancer metastasis and angiogenesis: an imbalance of positive and negative regulation. Cell 1991;64:327–36. 9. DeClerck YA, Perez N, Shimada H, Boone TC, Langley KE, Taylor SM. Inhibition of invasion and metastasis in cells transfected with an inhibitor of metalloproteinases. Cancer Res 1992;52:701–8. 10. 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;127:813–20. 11. Nagase H. Activation mechanisms of matrix metalloproteinases. Biol Chem 1997;378:151–60. 12. Sato H, Takino T, Okada Y, et al. A matrix metalloproteinase expressed on the surface of invasive tumor cells. Nature 1994;370:61–5. 13. Katayama A, Bandoh N, Kishibe K, et al. Expressions of matrix metalloproteinases in early-stage oral squamous cell carcinoma as predictive indicators for tumor metastasis and prognosis. Clin Cancer Res 2004;10:634–40. 14. Murashige M, Miyahara M, Shiraishi N, Saito T, Kohno K, Kobayashi M. Enhanced expression of tissue inhibitors of metalloproteinases in human colorectal tumors. Jpn J Clin Oncol 1996;26:303–9. 15. Ree AH, Florenes VA, Berg JP, Malandsmo GM, Nesland JM, Fodstad O. High levels of messenger RNAs for tissue inhibitors of metalloproteinase (TIMP-1 and TIMP-2) in primary breast carcinomas are associated with development of distant metastasis. Clin Cancer Res 1997;3:1623–8. 16. Takes RP, Baatenburg de Jong RJ, Wijffels K, et al. Expression of genetic markers in lymph node metastases compared with their primary tumors in head and neck cancer. J Pathol 2001;194:298–302. 17. Sobin LH, Witterkind Ch, editors. TNM classification of malignant tumors. New York: Wiley-Liss, Inc.; 1997, p. 17–50. 18. Engers R, Springer E, Michiels F, Collard JG, Gabbert HE. Rac affects invasion of human renal cell carcinomas by upregulating tissue inhibitor of metalloproteinases (TIMP)-1 and TIMP-2 expression. J Biol Chem 2001;276:41889–97. 19. Van den Oord JJ, Paemen L, Opdenakker G, de Wolf-Peeters C. Expression of gelatinase B and the extracellular matrix metalloproteinase inducer EMMPRIN in benign and malig-

Immunoexpression and prognostic significance of TIMP-1 and -2

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

nant pigment cell lesions of the skin. Am J Pathol 1997;151:665–70. Sawatsubashi M, Mizokami H, Tokunaga O, Shin T. Expression of MMP-1, TIMP-1, and type I collagen in laryngeal carcinoma. Mod Pathol 1998;11(9):878–85. Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J Clin Invent 1999;103:1237–41. Ziober BL, Turner MA, Palefsky JM, Banda MJ, Kramer RH. Type I collagen degradation by invasive oral squamous cell carcinoma. Oral Oncol 2000;36:365–72. Stetler-Stevenson WG, Bersch N, Golde DW. Tissue inhibitor of metalloproteinase-2 (TIMP-2) has erythroid-potentiating activity. FEBS Lett 1992;296:231–4. Hayakawa T, Yamashita K, Ohuchi E, Shinagawa A. Cell growth-promoting activity of tissue inhibitor of metalloproteinase-2 (TIMP-2). J Cell Sci 1994;107:2373–9. Corcoran ML, Stetler-Stevenson WG. Tissue inhibitor of metalloproteinase-2 stimulates fibroblasts proliferation via a cAMP-dependent mechanism. J Biol Chem 1999;270:13453–9. Charous SJ, Strickling GP, Nanney LB, Netterville JL, Burkey BB. Expression of matrix metalloproteinases and tissue inhibitor of metalloproteinases in head and neck squamous cell carcinoma. Ann Otol Rhinol Laryngol 1997;106:271–8. Sutinen M, Kainulainen T, Hurskainen T, et al. Expression of matrix metalloproteinases (MMP-1 and -2) and their inhibitors (TIMP-1, -2 and -3) in oral lichen planus, dysplasia, squamous cell carcinoma and lymph node metastasis. Br J Cancer 1998;77:2239–45. Ikebe T, Shinohara M, Takeuchi H, et al. Gelatinolytic activity of matrix metalloproteinase in tumor tissues correlates with the invasiveness of oral cancer. Clin Exp Metastasis 1999;17:315–23. Kurahara S, Shinohara M, Ikebe T, et al. Expression of MMPs, MT-MMP, and TIMPs in squamous cell carcinoma of the oral cavity: correlations with tumor invasion and metastasis. Head Neck 1999;21:627–38. Birkedal-Hansen B, Pavelic ZP, Gluckman JL, Stambrook P, Li YQ, Stetler-Stevenson WG. MMP and TIMP gene expression in head and neck squamous cell carcinomas and adjacent tissues. Oral Dis 2000;6:376–82. Shimada T, Nakamura H, Yamashita K, et al. Enhanced production and activation of progelatinase A mediated by membrane-type 1 matrix metalloproteinase in human oral squamous cell carcinoma: implications for lymph node metastasis. Clin Exp Metastasis 2000;18:179–88. Liu G, Li J, Li Z, Chen X. The effects of MMPs and TIMPs on the metastasis of oral squamous cell carcinoma to neck lymph nodes. Hua Xi Kou Qiang Yi Xue Za Zhi 2001;19:216–8. Yoshizaki T, Maruyama Y, Sato H, Furukawa M. Expression of tissue inhibitor of matrix metalloproteinase-2 correlates with activation of matrix metalloproteinase-2 and predicts poor prognosis in tongue squamous cell carcinoma. Iny J Cancer (Pred Oncol) 2001;95:44–50. Culhaci N, Metin K, Copcu E, Dikicioglu E. Elevated expresio ´n of MMP-13 and TIMP-1 in head and neck squamous

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

579

cell carcinomas may reflect increased tumor invasiveness. BMJ Cancer 2004;4:42. Available from URL: http:// www.biomedcentral.com/1471-2407/4/42. Guttman D, Stern Y, Shpitzer T, Ulanovski D, Druzd T, Feinmesser R. Expression of MMP-9, TIMP-1, CD-34 and factor-8 as prognostic markers for squamous cell carcinoma of the tongue. Oral Oncol 2004;40:798–803. Schmalbach CE, Chepeha DB, Giordano TJ, et al. Molecular profiling and the identification of genes associated with metasatatic oral cavity/pharynx squamous cell carcinoma. Arch Otolaryngol Head Neck Surg 2004;130(Mar):295– 302. Polette M, Clavel C, Birembaut P, DeClerck YA. Localization by in situ hybridization of mRNAs encoding stromelysin 3 and tissue inhibitors of metallo-proteinases TIMP-1 and TIMP-2 in human head and neck carcinomas. Path Res Pract 1993;189:1052–7. Gray ST, Wilkins RJ, Yun K. Interstitial collagenase gene expression in oral squamous cell carcinoma. Am J Pathol 1992;141:301–6. Kanayama H, Yokota K, Kurokawa Y, Murakami Y, Nishitani S, Kagawa S. Prognostic value of matrix metalloproteinase-2 and tissue inhibitor of metalloproteinase-2 expression in bladder cancer. Cancer 1998;82:1359–66. Visscher DW, Hoyhtya M, Ottosen SK, et al. Enhanced expression of tissue inhibitor of metalloproteinase-2 (TIMP-2) in the stroma of the breast carcinomas correlates with tumor recurrence. Int J Cancer 1994;59:339– 44. Thomas P, Khokha R, Shepherd FA, Feld R, Tsao MS. Differential expression of matrix metalloproteinases and their inhibitors in non-small cell lung cancer. J Pathol 2000;190:150–6. Suzuki M, Iizasa T, Fujisawa T, et al. Expression of matrix metalloproteinases and tissue inhibitor of matrix metalloproteinases in non-small-cell lung cancer. Invasionmetastasis 1998;18:134–41. Danilewicz M, Sikorska B, Wagrowska-Danilewicz M. Prognostic significance of the immunoexpression of matrix metalloproetinase MMP2 and its inhibitor TIMP2 in laryngeal cancer. Med Sci Monit 2003;9:42–7. Ross JS, Kaur P, Sheehan CE, Fisher HA, Kaufman Jr RA, Kallakury BV. Prognostic significance of matrix metalloproteinase 2 and tissue inhibitor of metalloproteinase 2 expression in prostate cancer. Mod Pathol 2003;16:198–205. de Vicente JC, Herrero-Zapatero A, Fresno MF, Lo ´pezArranz JS. Expression of cyclin D1 and Ki-67 in squamous cell carcinoma of the oral cavity: clinicopathological and prognostic significance. Oral Oncol 2002;38:301–8. van Diest PJ, Brugal G, Baak JPA. Proliferation markers in tumours: interpretation and clinical value. J Clin Pathol 1998;51:716–24. de Vicente JC, Junquera LM, HerreroZapatero A, Fresno MF, Hernandez-Vallejo G, Lo ´pez-Arranz JS. Prognostic significance of p53 expression in oral squamous cell carcinoma without neck node metastases. Head Neck 2004;26: 22–30.