Cyclin D1 and p16INK4A are molecules with pivotal roles in cell cycle control and the development of diverse human cancers, and overexpression of cyclin D1 and loss of p16INK4A expression are common genetic events in head and neck squamous cell carcinoma. The prognostic significance of these molecular events at different sites within the head and neck, however, remains controversial. Thus, we sought to determine the relationship between cyclin D1 and/or p16INK4A expression and disease outcome in squamous cell carcinoma of the anterior tongue. Immunohistochemical detection of nuclear proteins cyclin D1, p53, and p16INK4A, and the Ki-67 labeling index was undertaken in tissue sections from 148 tongue cancers treated by surgical resection. Nuclear antigen status was analyzed in relation to pathological variables, tumor recurrence, and patient survival. Statistical significance was assessed using χ2 analysis for pathological variables and the Kaplan-Meier method, log rank test, and the Cox proportional hazards model for survival parameters. Overexpression of cyclin D1 occurred in 68% of tumors (100 of 147) and was associated with increased lymph node stage (P = 0.014), increased tumor grade (P = 0.003), and reduced disease-free (P = 0.006) and overall (P = 0.01) survival. Loss of p16INK4A expression was demonstrated in 55% of tumors (78 of 143) and was associated with reduced disease-free (P = 0.007) and overall (P = 0.014) survival. Multivariate analysis confirmed that in addition to pathological stage and regional lymph node status, cyclin D1 overexpression and loss of p16INK4A expression are independent predictors of death from tongue cancer. Loss of p16INK4A in the presence of cyclin D1 overexpression conferred a significantly worse disease-free (P = 0.011) and overall (P = 0.002) survival at 5 years. p53 nuclear accumulation and the Ki-67 labeling index were not prognostic. These data indicate that cyclin D1 overexpression and loss of p16INK4A expression predict early relapse and reduced survival in squamous cell carcinoma of the anterior tongue. Simultaneous assessment of cyclin D1 and p16INK4A protein levels define subgroups of patients at increased risk of relapse and may be of clinical utility in optimizing therapy.

HNSCC3 accounts for ∼5% of all cancers in industrialized countries and has a worldwide incidence of ∼500,000 new cases/annum (1). Up to one-half of all HNSCCs occur in the oral cavity. Squamous cell carcinoma of the tongue is the most common intraoral malignancy, with tobacco and alcohol use being the most important risk factors (2, 3, 4). Approximately 50% of these cancers are cured by primary surgical treatment, but identifying those at increased risk of relapse is a major issue in the clinical management of the disease. Presently, lymph node stage is the most accurate predictor of disease recurrence (5, 6). However, with an increased understanding of the biology of the disease, molecular markers predictive of locoregional recurrence and reduced survival are likely to be of benefit in identifying patients suited to more intensive treatment regimes. It is now well established that a number of oncogenes and tumor suppressor genes with roles in cell cycle control are aberrantly expressed or inactivated during the evolution of diverse human cancers (7, 8). Recent studies in progression models for HNSCCs have identified the likely importance of genes controlling the G1 to S phase progression in the cell cycle including: CCND1, INK4A, Rb, and p53(9).

pRb acts as a gatekeeper, controlling the passage of cells through the G1 to S phase restriction point of the cell cycle (10). The phosphorylation of pRb results in its functional inactivation, leading to the release of transcription factors necessary for S phase entry and progression (10). Cyclin D1 and p16INK4A are proteins that modulate the phosphorylation of pRb and in the normal physiological state provide tight control over G1 to S phase progression and cellular proliferation rates. Cyclin D1 is the regulatory subunit of Cdk-4 and Cdk-6, and overexpression of cyclin D1, with increased formation of Cdk-4 and Cdk-6 complexes, results in hyperphosphorylation and hence functional inactivation of pRb. The CCND1 gene that encodes the cyclin D1 protein is located on chromosome 11q13, and this locus is amplified in ∼45% of HNSCC (11).

The role of cyclin D1 as a prognostic marker in HNSCC remains controversial. Overexpression has been associated with a more aggressive tumor phenotype and reduced survival in patients with operable HNSCC from different anatomical sites (12, 13, 14). In one study of 115 HNSCCs, overexpression of cyclin D1 was an independent prognostic variable; however, this study comprised predominantly hypopharyngeal and laryngeal carcinomas (14). In contrast, a study of 45 patients with a significantly higher proportion of oral carcinoma found no significant correlation between overexpression of cyclin D1 and any of the clinicopathological parameters studied (15). It is well documented that 11q13 amplification, and consequent overexpression of cyclin D1, are most common in hypopharyngeal carcinoma, which have the highest mortality of tumors of the head and neck (14, 16, 17). Hence, it is possible that the poor prognosis associated with both 11q13 amplification and cyclin D1 overexpression in HNSCC may reflect the prevalence of hypopharyngeal carcinoma in the published study populations.

The tumor suppressor gene, CDKN2/MTS1/INK4A, maps to chromosome 9p21 and encodes the CDK inhibitor, p16INK4A, which binds to and inhibits Cdk-4 and Cdk-6, resulting in inhibition of pRb phosphorylation and decreased cell cycle progression (18, 19). Loss of heterozygosity at 9p21 occurs at high frequency in HNSCC (20) and has been proposed as an early genetic event in the evolution of oral HNSCC (21). Inactivation of INK4A occurs by at least three distinct mechanisms: homozygous gene deletion, transcriptional silencing through promoter methylation, and by point mutations (18, 19). Immunohistochemical analysis of p16INK4A gene expression is an accurate method for evaluating INK4A gene inactivation, with loss of expression occurring in ∼75–85% of HNSCCs (20). To date, there are no published data on the relationship between p16INK4A expression and outcome in HNSCC.

There have been few large immunohistochemical studies dedicated to defining relationships between the expression of cell cycle proteins, clinicopathological features, and patient outcome from cancers at a single anatomical site in the head and neck. However, a recent study of 149 laryngeal squamous cell carcinomas revealed overexpression of cyclin D1 in 32% of cases, and this was associated with tumor extension, advanced clinical stage, lymph node metastases, and a shorter disease-free and overall survival (6). Focusing on specific anatomical subsites is likely to provide more accurate and clinically useful information on the prognostic significance of specific gene alterations in HNSCC. In this study, we sought to determine whether the expression of cyclin D1 and p16INK4A, alone or in combination, were important predictors of outcome in a large series of patients with squamous cell carcinoma of the anterior tongue. Given the pivotal role of these genes in cell proliferation control, the Ki-67 labeling index was assessed in an attempt to correlate changes in gene expression with a marker of cycling cells. p53, which is commonly mutated and overexpressed in HNSCC and may be associated with local recurrence but apparently not survival, was also assessed (22, 23, 24, 25, 26).

Patient and Tissue Samples.

Following Ethics Committee approval, 148 patients with primary operable squamous cell carcinoma of the anterior tongue and treated with curative intent were identified from the case records of the Departments of Head and Neck Surgery at St. Vincent’s Hospital and Westmead Hospital, Sydney, New South Wales. The clinicopathological characteristics of this series are summarized in Table 1.

Surgery was the sole treatment modality in 96 patients, whereas surgery with adjuvant radiotherapy was the treatment in 52 patients. In all cases, the margins of resection were microscopically negative for tumor. Clinical follow-up to the time of treatment failure or a minimum of 2 years disease-free survival was obtained from clinical records, New South Wales State Register of Births, Deaths and Marriages, and by consultation with the patient’s treating physician. Mean follow-up was 57 months (range, 1–186 months).

Immunohistochemistry.

Four-μm tissue sections were cut from archival formalin-fixed, paraffin-embedded tissue and deparaffinized in xylene for 10 min, followed by rehydration in a graded series of alcohol. A microwave antigen unmasking technique was used with citrate buffer (pH 6.0) for p16INK4A, p53, and Ki- 67 and EDTA buffer (pH 8) for cyclin D1. Sections were heated in buffer solution for 10 min at 750 W in a microwave oven and allowed to cool to room temperature. Endogenous peroxidase activity was quenched using 0.3% hydrogen peroxide in methanol for 15 min. Sections were then blocked with 10% normal horse serum in phosphate-buffered saline for 20 min. Slides were incubated overnight at 4°C with monoclonal antibodies to human cyclin D1 (1:50; Novocastra NCL-cyclin D1-GM; Novocastra, Rockdale, Australia), p16INK4A protein (1:60; clone ZJ11; Neomarkers, Fremont, CA), Ki-67 nuclear antigen (1:50; clone 7B11; Zymed Laboratories, San Francisco, CA), and p53 protein (1:50; clone DO-7; Dako, Carpinteria, CA). After washing with PBS for 20 min, the slides were incubated for 30 min at room temperature with biotinylated horse antimouse antibody (1:100; Vector Laboratories, Burlingame, CA). The Vectastain Elite ABC kit (Vector Labs) provided the secondary detection system, and color development was obtained using 3,3-diaminobenzidine (DAB kit; Vector Laboratories) in accordance with the manufacturer’s instructions. Hematoxylin was used as a counterstain.

Formalin-fixed, paraffin-embedded sections of breast cancer cell lines with known differential expression of cyclin D1 (27, 28) and p16INK4A protein (29) were used as positive and negative controls in each experiment. Positive and negative controls for cyclin D1 included cell lines MDA-MB-134 and HBL-100, respectively. Positive controls for p16INK4A included the cell lines MDA-MB-157 and MDA-MB-468. Admixed stromal tissue and adjacent normal mucosa also served as internal positive controls for this antigen. Negative controls for p16INK4A included the cell lines MCF-7 and MDA-MB-231, which do not express p16INK4A(29). The prostatic carcinoma cell line DU145 (30) was used as a positive control, whereas substitution of the primary antibody with horse serum was used as a negative control for p53 staining.

Scoring.

Immunohistochemical staining was scored by at least two independent observers (R. J. B., D. I. Q., and J. S. N.) without knowledge of the patients’ outcome. The immunoreactivity of the cyclin D1, p53, p16INK4A, and Ki-67 protein staining was scored according to the percentage of positively stained tumor cell nuclei. At least 10 high-powered tumor fields were counted. For cyclin D1 and p53 staining, scores were ranked as: negative, defined by <10% of the tumor cells positive; and positive, if >10% of the tumor cells were positive.

The immunohistochemical evaluation of p16INK4A staining was similar to a protocol published previously (31). Only positive nuclear staining was scored positive for p16INK4A, regardless of the degree of cytoplasmic staining. Stromal and adjacent mucosal cells served as positive internal controls. If the stromal cells or adjacent mucosal cells failed to display nuclear positivity in the presence of negative tumor staining after three staining attempts, then the immunohistochemical staining was considered a technical failure. Four samples were deleted from the analysis by this criterion. A section was considered positive if >1% of the tumor cells demonstrated positive nuclear staining for p16INK4A. This level of expression was chosen as the cutoff after a stepwise analysis of the percentage of positive cells versus outcome. Using the more commonly used cutoff of >5% positive cells rendered the vast majority of tumors negative and resulted in no relationship with other clinicopathological parameters, including outcome. For each tumor, the Ki-67 labeling index was determined by estimating the percentage of tumor cell nuclei staining positive.

Statistical Analysis.

The relationship between cyclin D1, p53, and p16INK4A protein expression and various clinical and pathological parameters was determined using the χ2 test after the expression level for each variable was dichotomized as described above. The Kruskal-Wallis test was used to compare the median Ki-67 values across the levels of the categorical variables including tumor stage, nodal stage, tumor grade, cyclin D1, p53, and p16INK4A expression.

Overall survival was calculated from the time of surgery to the date of death, as the event of interest, or the date of last follow-up. Disease-free survival was calculated from the time of surgery to the time of local or regional disease recurrence, as the event of interest, or last follow-up. Survival analysis was undertaken using the Kaplan-Meier method (32), and curves were compared using log-rank analysis. Univariate analysis for known prognostic factors and the expression of each of the four nuclear antigens was performed to define variables predicative of relapse and/or death. Predictive variables in univariate analysis were then incorporated into models of multivariate analysis using the Cox proportional hazard method (33) to identify which factors were independent predictors of disease-free and overall survival. All statistical analyses were performed using Statview 4.5 software (Abacus Concepts, Berkeley, CA).

One hundred and forty-eight patients with diagnoses of a primary squamous cell carcinoma of the anterior tongue entered the study. The clinicopathological, treatment, and outcome characteristics of the series are summarized in Table 1. Seventy % of patients were male, 63% were >65 years, 53% had T1 tumors, and 37% were T2; 114 patients (76%) had no lymph node involvement. The majority of tumors were moderately differentiated (53%), with the remainder equally split between well differentiated (25%) and poorly differentiated (22%). All patients underwent surgical resection with curative intent as their primary treatment; 52 patients (35%) received adjuvant radiotherapy. During a mean follow-up period of 57 months (range 1–186 months), 38 (26%) of the patients had a local or regional recurrence, and 34 (23%) had died from their disease.

Association of Cyclin D1, p53, and p16INK4A Expression and Ki-67 Labeling Index with Clinicopathological Parameters.

Data for cyclin D1 were available on 147 patients, p16INK4A and p53 on 143 patients, and Ki-67 labeling index was available on all 148 patients. The relationships between the expression of these antigens and the clinicopathological features of the tumors are presented in Table 2. Cyclin D1 expression in tumor cells was detected in 100 (68%) specimens, of which 44 (30%) showed high expression, i.e., >50% of positive tumor cells (Fig. 1, A and B). In those sections that included adjacent normal and dysplastic mucosa, uniform staining for cyclin D1 was observed in the majority of cases, predominantly in the suprabasal epithelial cells. However, the staining in adjacent mucosa was generally of reduced intensity compared with invasive carcinoma.

Cyclin D1 expression did not correlate with tumor size (P = 0.18) or pathological stage (P = 0.16). There was, however, a significant correlation between cyclin D1 expression and tumor grade, with only 46% of well-differentiated tumors overexpressing cyclin D1 compared with 73% of moderately differentiated and 81% of poorly differentiated tumors (P = 0.003). Cyclin D1 was overexpressed in a significantly greater percentage of N1,2,3 tumors (85%) than in N0 tumors (62%; P = 0.014). There was no association between cyclin D1 and p16INK4A status or the Ki-67 labeling index, but cyclin D1 expression was significantly correlated with nuclear accumulation of p53. Of the tumors that overexpressed cyclin D1, 70% had concurrent increase in p53 accumulation in contrast to the cyclin D1 negative tumors, where only 44% were p53 positive (P = 0.003). A significantly greater percentage of patients treated with adjuvant radiation therapy were cyclin D1 positive (P = 0.028), presumably as a result of the relationship between cyclin D1, nodal stage, and tumor grade (Table 2).

Forty-six % of tumors showed positive immunoreactivity for p16INK4A protein. The majority of these sections demonstrated both nuclear and cytoplasmic staining in the tumor cells, whereas the p16INK4A-negative tumors showed complete absence of nuclear staining in the presence of adjacent admixed nonneoplastic elements scoring positive (Fig. 1, C and D). Four patients were omitted from the p16INK4A analysis because complete absence of staining was noted in both tumor and stromal elements, despite repeated staining attempts. There was no association between the p16INK4A expression status and any of the clinicopathological parameters (Table 2), nor was there an association with p53 protein accumulation (P = 0.32) or overexpression of cyclin D1 (P = 0.21) or Ki-67 labeling index (P = 0.61).

Sixty-two % of tumors demonstrated excess accumulation of nuclear p53 protein, with 89% of positive tumors showing >50% of nuclear staining positive for p53 protein. Staining was generally intense and uniform in the invasive tumor front and in the basal layers of adjacent dysplastic epithelium. The adjacent superficial epithelial cells and the most differentiated cells at the center of tumor whorls demonstrated the least amount of staining. There was no correlation between the p53 nuclear accumulation and any of the clinical or pathological parameters studied, nor was there any relationship with any of the other molecular markers, except cyclin D1, as noted above (Table 2).

The Ki-67 labeling index was evaluated in all patients. The staining was confined to the nucleus and was homogeneous in most cases. Staining in adjacent mucosa was confined to the suprabasal layers of epithelial cells, with the occasional basal cell displaying positivity. There was no significant relationship between the Ki-67 labeling index and any of the clinicopathological characteristics or molecular markers assessed in this study (data not shown).

Association between Cyclin D1 and p16INK4A Status, Disease Relapse, and Overall Survival.

Table 3 presents univariate Cox regression analysis of disease-free and overall survival for each of the clinicopathological and molecular parameters assessed in the study. In agreement with previous studies, lymph node stage was a significant predictor of disease-free (P = 0.006) and overall (P < 0.001) survival, whereas overall pathological stage predicted for overall survival (P = 0.005) but did not reach significance for disease-free survival (P = 0.09). Overexpression of cyclin D1 was associated with both reduced disease-free (P = 0.006) and overall (P = 0.01) survival. There was no significant survival difference between low cyclin D1 overexpressors, as defined by 10–50% positive tumor cell nuclei, and high cyclin D1 overexpression, i.e., >51% positive nuclei (data not shown). Kaplan-Meier survival curves for disease-free and overall survival are presented in Figs. 2,A and 3 A, respectively, and clearly demonstrate the adverse impact of cyclin D1 overexpression on both disease recurrence (log-rank P = 0.024) and overall survival (log-rank P = 0.005).

Complete loss of p16INK4A expression was also associated with reduced disease-free survival (P = 0.007) and overall survival (P = 0.02; Table 3), and this is demonstrated by the Kaplan-Meier analysis (Figs. 2,B and 3,B; log-rank P = 0.005 and P = 0.014, respectively). p53 nuclear accumulation and the Ki-67 labeling index had no significant association with disease outcome. Patients receiving adjuvant radiation therapy had a worse disease-free and overall survival than those treated with surgery alone (Table 3) because of selection bias in favor of patients with higher grade (P = 0.007) and stage (P < 0.0001) receiving adjuvant radiation therapy.

In a multivariate analysis where cyclin D1, p16INK4A, and treatment were included in two models with either lymph node stage or overall pathological stage, both cyclin D1 and p16INK4A were independent predictors of tongue cancer recurrence and death (Table 4). In both models, loss of p16INK4A expression was the most significant independent prognostic marker for disease-free survival. For overall survival, pathological stage (P = 0.01) and lymph node stage (P = 0.005) were the most significant predictors in the two models (Table 4). Nonetheless, both cyclin D1 and p16INK4A were independent predictors of overall survival.

Association of Combined Cyclin D1 and p16INK4A Status with Disease Outcome.

Given that cyclin D1 and p16INK4A were independent predictors of disease outcome in this study, we determined whether more information could be obtained by analyzing subgroups based on combined cyclin D1 and p16 status. Four subgroups based on positivity or negativity for each antigen were assigned. Kaplan-Meier curves for disease-free and overall survival are presented in Figs. 2,C and 3,C, respectively, and 5-year survival data are presented in Table 5. All of 20 2patients with tumors that were negative for cyclin D1 expression and positive for p16INK4A protein expression were alive with no locoregional recurrence at the time of follow-up. In contrast, of 53 patients with tumors demonstrating simultaneous overexpression of cyclin D1 protein and loss of p16INK4A protein, 21 had relapsed and 19 had died of their disease (Table 5; Figs. 2,C and 3 C). The other two subgroups, i.e., cyclin D1 negative, p16INK4A negative, and cyclin D1 positive, p16INK4A positive had an intermediate outcome, with the latter group having a worse overall survival.

The multistep theory of carcinogenesis proposes that cancer of the head and neck results from the accumulation of genetic changes in mucosal cells resulting from sustained exposure to carcinogens such as tobacco and ethanol. The consequent genetic instability resulting from deregulation of certain cell cycle proteins results in the temporal transition of a mucosal cell from a normal state to dysplasia and eventually carcinoma. Although the temporal order of genetic alterations preceding overt malignancy is unknown, a genetic progression model of HNSCC presented by Califano et al.(9) proposes that p16INK4A is the earliest known tumor suppressor gene to be inactivated in HNSCC, whereas deregulation of p53 and cyclin D1 occur later. Irrespective of the temporal sequence of those events, it is clear that loss or deregulated expression of several genes with pivotal roles in cell cycle control are critical events in the evolution of HNSCC (6, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 34, 35, 36, 37). Whether changes in the expression of these molecules are associated with clinical tumor behavior and thus might provide molecular markers of disease outcome was investigated in a series of 148 HNSCC from a single anatomical site, the anterior tongue. Our data confirm earlier studies that failed to establish an association between p53 nuclear accumulation (14, 25) or markers of cell proliferation (38) with survival and extend the evidence for cyclin D1 being an independent prognostic marker in HNSCC (6, 13, 14, 15). More importantly, we define for the first time a significant relationship between loss of p16INK4A expression and an adverse disease outcome in HNSCC and demonstrate that simultaneous assessment of cyclin D1 and p16INK4A status allow segregation of subgroups of patients with significantly different probabilities of relapse.

There are several published studies reporting cyclin D1 overexpression in HNSCC, but only recently have data from large series at a single anatomical site, the larynx, become available (6). The earliest studies, using a number of chromosome 11 markers, demonstrated amplification at 11q13, including the CCND1 gene, in ∼45% of HNSCC. Several studies reported a positive correlation between 11q13 amplification and advanced stage, lymph node involvement, and reduced survival, but often this did not reach statistical significance, perhaps because of sample size (11). More recent studies support a relationship between CCND1 amplification and adverse prognosis (13, 39, 40), although data confined to 50 tumors of the oral cavity and oropharynx found no relationship with lymph node invasion or disease recurrence (41).

Immunohistochemical detection of cyclin D1 is likely to be a more accurate estimate of overexpression than CCND1 amplification. Cyclin D1 overexpression occurred in 68% of specimens, which is slightly higher than other reported series (range, 35–64%; Refs. 12, 13, 14, 34, and 35). Although 11q13 amplification correlates with increased tumor grade (15), this is the first demonstration of a significant correlation between cyclin D1 expression and tumor grade. In our series, moderately and poorly differentiated tumors demonstrated a significantly higher incidence of cyclin D1 overexpression than well-differentiated tumors. The reason for this relationship remains obscure. However, both in vitro and in vivo studies have found that cyclin D1 overexpression can inhibit the differentiation of myoblasts (15, 42) and intestinal epithelial cells (43), thus raising the possibility that cyclin D1 overexpression may play a role in the inhibition of tumor cell differentiation in some cell types.

Overexpression of cyclin D1 was also associated with reduced disease-free and overall survival with a higher incidence of local and regional recurrence. This is in agreement with several recent publications investigating the prognostic significance of cyclin D1 in HNSCC (6, 12, 14, 15, 40), including studies restricted to tumors of the hypopharynx (12) and the larynx (6). Thus, our data provide further evidence that overexpression of cyclin D1 imparts a poor prognosis in HNSCC independent of tumor site and is an independent prognostic tumor marker in patients with anterior tongue cancer.

The mechanisms responsible for cyclin D1 overexpression and by which the overexpression confers a more aggressive malignant phenotype remain unclear. Although the contribution of CCND1 gene amplification to overexpression has been well documented (11), other likely mechanisms include up-regulation of receptor and signaling pathways that converge on cyclin D1 gene expression (10). Recent evidence that levels of TGF-α and EGFR are independent predictors of outcome in patients with HNSCC (44) raises the interesting possibility that part of this effect may be mediated via increased expression of cyclin D1. Overexpression of cyclin D1 shortens the G1 phase and reduces dependence on growth factors (45), which in turn may result in loss of cell cycle control and increased cell proliferation. However, when Ki-67 expression was used as a proliferative marker, no correlation with cyclin D1 expression was observed. This is consistent with other published data where an in vivo infusion of iododeoxyuridine was used as a measure of cell proliferation in a series of HNSCCs, and no correlation between cyclin D1 expression and cell proliferation was observed (34).

A positive correlation between p53 accumulation and cyclin D1 overexpression was noted in our study (P = 0.001). Seventy % of cyclin D1-positive tumors accumulated nuclear p53, whereas only 40% of cyclin D1 negative tumors showed evidence of p53 accumulation. This is in agreement with a study of 39 patients with HNSCCs that demonstrated a statistically significant correlation between P53 mutation and amplification of the CCND1 gene (36). Because it has been postulated that deregulation of P53 is an early event in HNSCCs (22, 23), loss of p53 function may render cells susceptible to further genetic alterations, such as amplification and/or overexpression of CCND1. In contrast, a more recent study found no correlation between loss of heterozygosity at the P53 locus and amplification of the CCND1 gene in a group of 56 HNSCCs. That study, however, found that tumors with both loss of heterozygosity at the P53 locus and CCND1 amplification were associated with advanced stage and increased frequency of recurrence (37). Our study did not demonstrate a significant survival difference when p53 protein accumulation was included in the analysis. The 5-year survival for cyclin D1-negative and cyclin D1-positive tumors was not significantly altered by the p53 status of the tumors (data not shown). The high rate of p53 nuclear accumulation in this study is similar to rates observed in other immunohistochemical analyses of patients with intraoral squamous cell carcinomas (38, 46). The lack of association of p53 protein accumulation with any of the clinicopathological parameters, including survival, in this study is consistent with other reported series (47, 48).

Loss of expression of p16INK4A protein, as assessed by immunohistochemistry, has been reported in up to 47% of premalignant lesions (21) and up to 80% of invasive HNSCCs (20). The 54% loss of p16INK4A expression in the present study is lower than reported by others, and this may reflect the different scoring system used in this study. In contrast to other studies that have used 5% of positive tumor cells as the discriminatory threshold, we reduced the threshold to 1%, resulting in a greater proportion of positive tumors. Although there is no real consensus in the literature on the criteria for distinguishing p16-positive tumors from p16-negative using immunohistochemistry, it is noteworthy that limits as low as one positive cell per field have been used (49). Although nuclear staining was required to confirm positivity, the additional presence of cytoplasmic staining was uniform in most positive tumor sections. This cytoplasmic staining was usually weaker than its nuclear counterpart. It remains unclear whether cytoplasmic staining in tissue sections represents nonspecific background or extravasation of intranuclear p16INK4A protein (31). Thus, further interpretation of an appropriate discriminating threshold of p16INK4A expression and the relevance of any cytoplasmic staining must await further studies on relationships with outcome. The marked disparity between the high staining intensity of some positive tumor cells compared with weakly staining adjacent stromal cells in this study raises the possibility that some tumor cells were overexpressing p16INK4A. It is well documented that inactivating mutations in Rb result in elevated expression of p16INK4A because of a loss of the transcriptional inhibition of INK4A gene expression characteristic of wild-type Rb (50, 51). The relationship between Rb mutations, p16INK4A expression, and outcome in HNSCCs may further assist in identifying subgroups of patients at increased risk of relapse.

No significant association was apparent between the absence of p16INK4A protein expression and any of the pathological or molecular variables studied. Although loss of heterozygosity at 9p21 correlates with recurrence in HNSCCs (52), this is the first study to report an association between p16INK4A expression and patient outcome in this disease. Loss of p16INK4A protein expression was associated with reduced disease-free and overall survival and on multivariate analysis was an independent prognostic variable. The observed association between a favorable clinical outcome and tumors displaying a combination of low expression of cyclin D1 and positive p16INK4A protein expression is an interesting and potentially important finding in this study. It raises the question as to whether the combination of cyclin D1 and p16INK4A protein expression is potentially a more accurate prognostic tumor marker than either parameter alone in patients with tongue carcinoma.

In conclusion, this study confirms that deregulation of cell cycle proteins such as p53, p16INK4A, and cyclin D1 are common events in squamous cell carcinoma of the tongue and provides strong evidence that cyclin D1 overexpression and loss of p16INK4A are both independent prognostic tumor markers in this disease. Together with recently published data on transforming growth factor α and epidermal growth factor receptor (44), cyclin D1 and p16INK4A are among the most promising molecular markers of disease outcome yet identified in HNSCCs and may identify subgroups of patients for which more aggressive therapy is warranted.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

        
1

This work was supported by research grants from the Garnett Passe and Rodney Williams Memorial Foundation, the St. Vincent’s Clinic Foundation, the Royal Australasian College of Surgeons, the R. T. Hall Trust, the National Health and Medical Research Council of Australia, and the New South Wales Cancer Council. R. J. B. was the recipient of a Research Fellowship from the Royal Australasian College of Surgeons and D. I. Q. was supported by a Medical Postgraduate Scholarship from the National Health and Medical Research Council of Australia.

                
3

The abbreviations used are: HNSCC, head and neck squamous cell carcinoma; pRb, retinoblastoma gene product; Cdk, cyclin-dependent kinase.

Fig. 1.

Cyclin D1 and p16 immunostaining in squamous cell carcinoma of the anterior tongue. A, cyclin D1-positive tumor. ×200. B, cyclin D1-negative tumor. ×200. C, p16-positive tumor. ×400. D, p16-negative tumor. ×200.

Fig. 1.

Cyclin D1 and p16 immunostaining in squamous cell carcinoma of the anterior tongue. A, cyclin D1-positive tumor. ×200. B, cyclin D1-negative tumor. ×200. C, p16-positive tumor. ×400. D, p16-negative tumor. ×200.

Close modal
Fig. 2.

Disease-free survival of 147 patients with tongue carcinoma according to cyclin D1 and p16INK4A status. A, Kaplan-Meier curve for disease-free survival according to cyclin D1 status, i.e., <10% cyclin D1-positive cells versus >10% cyclin D1 positive cells (P = 0.024). B, Kaplan-Meier curve comparing patients with <1% p16INK4A expression to those with >1% p16INK4A expression (P = 0.005). C, Kaplan-Meier curve for disease-free survival comparing patients according to their cyclin D1 and p16INK4A status. Patients were assigned to four subgroups according to the percentage of positive tumor cell nuclei for each of the two antigens: ▴, <10% cyclin D1, >1% p16INK4A; ▵, <10% cyclin D1, <1% p16INK4A; □, >10% cyclin D1, >1% p16INK4A; +, >10% cyclin D1, <1% p16INK4A.

Fig. 2.

Disease-free survival of 147 patients with tongue carcinoma according to cyclin D1 and p16INK4A status. A, Kaplan-Meier curve for disease-free survival according to cyclin D1 status, i.e., <10% cyclin D1-positive cells versus >10% cyclin D1 positive cells (P = 0.024). B, Kaplan-Meier curve comparing patients with <1% p16INK4A expression to those with >1% p16INK4A expression (P = 0.005). C, Kaplan-Meier curve for disease-free survival comparing patients according to their cyclin D1 and p16INK4A status. Patients were assigned to four subgroups according to the percentage of positive tumor cell nuclei for each of the two antigens: ▴, <10% cyclin D1, >1% p16INK4A; ▵, <10% cyclin D1, <1% p16INK4A; □, >10% cyclin D1, >1% p16INK4A; +, >10% cyclin D1, <1% p16INK4A.

Close modal
Fig. 3.

Overall survival of 147 patients with tongue carcinoma according to cyclin D1 and p16INK4A status. The survival curves presented in A, B, and C are for the subgroups of patients as outlined in the legend to Fig. 2.

Fig. 3.

Overall survival of 147 patients with tongue carcinoma according to cyclin D1 and p16INK4A status. The survival curves presented in A, B, and C are for the subgroups of patients as outlined in the legend to Fig. 2.

Close modal
Table 1

Clinicopathological, treatment, and outcome characteristics of 148 patients treated for squamous cell carcinoma of the anterior tongue

No. of patients% of patients
Sex   
 Male 104 70 
 Female 44 30 
Age   
 <65 93 63 
 >65 55 37 
Tumor stagea   
 T1 79 53 
 T2 55 37 
 T3 11 
 T4 
Lymph node stagea   
 N0 114 76 
 N1, N2, N3 34 24 
Tumor grade   
 Well differentiated 29 25 
 Moderately differentiated 84 53 
 Poorly differentiated 35 22 
Radiation therapy   
 Yes 52 35 
 No 96 65 
Recurrence   
 Yes 38 26 
 No 110 74 
Death   
 Yes 34 23 
 No 114 77 
No. of patients% of patients
Sex   
 Male 104 70 
 Female 44 30 
Age   
 <65 93 63 
 >65 55 37 
Tumor stagea   
 T1 79 53 
 T2 55 37 
 T3 11 
 T4 
Lymph node stagea   
 N0 114 76 
 N1, N2, N3 34 24 
Tumor grade   
 Well differentiated 29 25 
 Moderately differentiated 84 53 
 Poorly differentiated 35 22 
Radiation therapy   
 Yes 52 35 
 No 96 65 
Recurrence   
 Yes 38 26 
 No 110 74 
Death   
 Yes 34 23 
 No 114 77 
a

Stage was determined by pathological analysis.

Table 2

Correlations between nuclear cyclin D1, p16INK4A, and p53 protein and clinicopathological parameters in 148 squamous cell cancers of the tongue

Number (n = 148) (%)Cyclin D1 (n = 147)p16INK4A (n = 143)p53 (n = 143)
NegativePositive (%)PNegativePositive (%)PNegativePositive (%)P
Tumor size           
 T1 79 (53) 31 47 (59)  38 39 (51)  31 44 (59)  
 T2 55 (37) 12 43 (78)  33 20 (38)  21 33 (61)  
 T3 11 (7) 8 (73) 0.18 4 (40) 0.43 9 (82) 0.26 
 T4 3 (2) 2 (68)  2 (67)  3 (100)  
Nodal stage           
 N0 114 (77) 42 71 (62)  60 49 (45)  43 67 (61)  
 N1, N2, N3 34 (23) 29 (85) 0.014 18 16 (47) 0.83 11 22 (67) 0.55 
Pathologic stage           
 I 70 (47) 28 40 (59)  32 35 (52)  30 36 (54)  
 II 37 (25) 28 (76)  24 11 (31)  14 23 (62)  
 III 26 (18) 19 (73) 0.16 13 12 (48) 0.2 19 (76) 0.22 
 IV 16 (11) 13 (81)  7 (44)  11 (73)  
Tumor grade           
 Well 37 (25) 20 17 (46)  19 18 (49)  13 22 (63)  
 Moderate 78 (53) 21 57 (73) 0.003 42 32 (0) 0.85 33 44 (57) 0.25 
 Poor 33 (22) 26 (81)  17 15 (47)  23 (74)  
Treatment           
 Surgery 92 (65) 36 56 (61)  53 36 (40)  40 49 (55)  
 Surgery + radiotherapy 55 (35) 11 44 (80) 0.028 25 29 (54) 0.35 14 40 (74) 0.12 
Number (n = 148) (%)Cyclin D1 (n = 147)p16INK4A (n = 143)p53 (n = 143)
NegativePositive (%)PNegativePositive (%)PNegativePositive (%)P
Tumor size           
 T1 79 (53) 31 47 (59)  38 39 (51)  31 44 (59)  
 T2 55 (37) 12 43 (78)  33 20 (38)  21 33 (61)  
 T3 11 (7) 8 (73) 0.18 4 (40) 0.43 9 (82) 0.26 
 T4 3 (2) 2 (68)  2 (67)  3 (100)  
Nodal stage           
 N0 114 (77) 42 71 (62)  60 49 (45)  43 67 (61)  
 N1, N2, N3 34 (23) 29 (85) 0.014 18 16 (47) 0.83 11 22 (67) 0.55 
Pathologic stage           
 I 70 (47) 28 40 (59)  32 35 (52)  30 36 (54)  
 II 37 (25) 28 (76)  24 11 (31)  14 23 (62)  
 III 26 (18) 19 (73) 0.16 13 12 (48) 0.2 19 (76) 0.22 
 IV 16 (11) 13 (81)  7 (44)  11 (73)  
Tumor grade           
 Well 37 (25) 20 17 (46)  19 18 (49)  13 22 (63)  
 Moderate 78 (53) 21 57 (73) 0.003 42 32 (0) 0.85 33 44 (57) 0.25 
 Poor 33 (22) 26 (81)  17 15 (47)  23 (74)  
Treatment           
 Surgery 92 (65) 36 56 (61)  53 36 (40)  40 49 (55)  
 Surgery + radiotherapy 55 (35) 11 44 (80) 0.028 25 29 (54) 0.35 14 40 (74) 0.12 
Table 3

Univariate Cox regression analysis of disease-free and overall survival in squamous cell carcinoma of the anterior tongue

Disease-free survivalOverall survival
Relative risk95% confidence intervalPRelative risk95% confidence intervalP
Sex       
 Male vs. female 1.32 0.67 –2.57 0.43 1.17 0.57 –2.40 0.67 
Age       
 <65 vs. ≥65 0.71 0.37 –1.35 0.29 0.98 0.48 –1.98 0.95 
Tumor stagea       
 T1–T2vs. T3–T4 0.55 0.13 –2.31 0.42 1.78 0.68 –4.60 0.24 
Nodal stagea       
 N1, N2, N3vs.N0 2.60 1.32 –5.14 0.006 4.15 2.10 –8.19 <0.001 
Overall pathological stage       
 Stage 3–4 vs.Stage 1–2 1.77 0.91 –3.42 0.09 3.34 1.70 –6.57 0.005 
Tumor grade       
 Moderate, Poor vs.Well 2.48 0.97 –6.35 0.06 2.70 0.95 –7.67 0.06 
Cyclin D1 expression       
 Positive vs. Negativeb 2.50 1.32 –5.15 0.006 3.89 1.37 –11.07 0.01 
p53 nuclear accumulation       
 Positive vs. Negativec 1.18 0.60 –2.32 0.64 1.46 0.69 –3.06 0.32 
p16INK4A expression       
 Negative vs. Positived 2.73 1.32 –5.67 0.007 2.50 1.16 –5.41 0.02 
Ki-67 index       
 ≥50% vs.<50% 1.17 0.61 –2.27 0.63 1.24 0.62 –2.47 0.54 
Treatment       
 Surgery + Radiotherapy vs. Surgery 2.76 1.40 –2.48 0.004 2.02 1.07 –3.82 0.003 
Disease-free survivalOverall survival
Relative risk95% confidence intervalPRelative risk95% confidence intervalP
Sex       
 Male vs. female 1.32 0.67 –2.57 0.43 1.17 0.57 –2.40 0.67 
Age       
 <65 vs. ≥65 0.71 0.37 –1.35 0.29 0.98 0.48 –1.98 0.95 
Tumor stagea       
 T1–T2vs. T3–T4 0.55 0.13 –2.31 0.42 1.78 0.68 –4.60 0.24 
Nodal stagea       
 N1, N2, N3vs.N0 2.60 1.32 –5.14 0.006 4.15 2.10 –8.19 <0.001 
Overall pathological stage       
 Stage 3–4 vs.Stage 1–2 1.77 0.91 –3.42 0.09 3.34 1.70 –6.57 0.005 
Tumor grade       
 Moderate, Poor vs.Well 2.48 0.97 –6.35 0.06 2.70 0.95 –7.67 0.06 
Cyclin D1 expression       
 Positive vs. Negativeb 2.50 1.32 –5.15 0.006 3.89 1.37 –11.07 0.01 
p53 nuclear accumulation       
 Positive vs. Negativec 1.18 0.60 –2.32 0.64 1.46 0.69 –3.06 0.32 
p16INK4A expression       
 Negative vs. Positived 2.73 1.32 –5.67 0.007 2.50 1.16 –5.41 0.02 
Ki-67 index       
 ≥50% vs.<50% 1.17 0.61 –2.27 0.63 1.24 0.62 –2.47 0.54 
Treatment       
 Surgery + Radiotherapy vs. Surgery 2.76 1.40 –2.48 0.004 2.02 1.07 –3.82 0.003 
a

Stage was determined by pathological analysis.

b

Cyclin D1 expression: Positive, >10%; Negative, <10%.

c

p53 nuclear accumulation: Positive, >10%; Negative, <10%.

d

p16INK4A protein expression: Positive, >1%; Negative, <1%.

Table 4

Multivariate Cox regression analysis for disease-free and overall survival in squamous cell carcinoma of the anterior tongue

Disease-free survivalOverall survival
Relative risk95% confidence intervalPRelative risk95% confidence intervalP
Model 1       
 Overall pathological stagea       
  Stage 3–4 vs. Stage 1–2 0.94 0.39 –2.26 0.94 3.09 1.29 –7.47 0.01 
 Cyclin D1 expression       
  Positive vs. Negativeb 2.48 1.0 –6.15 0.05 4.2 1.23 –14.09 0.02 
 p16INK4A expression       
  Negative vs. Positivec 3.15 1.65 –7.50 0.001 2.77 1.27 –6.06 0.03 
 Treatment       
  Surgery + Radiotherapy vs. Surgery 2.51 1.14 –5.53 0.03 1.69 0.67 –4.27 0.26 
Model 2       
 Nodal stagea       
  N1, N2, N3vs. N0 1.42 0.56 –3.63 0.46 3.76 1.49 –9.45 0.005 
 Cyclin D1 expression       
  Positive vs. Negativeb 2.44 0.98 –6.06 0.05 3.64 1.07 –12.25 0.03 
 p16INK4A expression       
  Negative vs. Positivec 3.23 1.49 –6.99 0.003 2.32 1.04 –5.16 0.04 
 Treatment       
  Surgery + Radiotherapy vs. Surgery 2.09 0.92 –4.75 0.08 1.58 0.62 –4.00 0.33 
Disease-free survivalOverall survival
Relative risk95% confidence intervalPRelative risk95% confidence intervalP
Model 1       
 Overall pathological stagea       
  Stage 3–4 vs. Stage 1–2 0.94 0.39 –2.26 0.94 3.09 1.29 –7.47 0.01 
 Cyclin D1 expression       
  Positive vs. Negativeb 2.48 1.0 –6.15 0.05 4.2 1.23 –14.09 0.02 
 p16INK4A expression       
  Negative vs. Positivec 3.15 1.65 –7.50 0.001 2.77 1.27 –6.06 0.03 
 Treatment       
  Surgery + Radiotherapy vs. Surgery 2.51 1.14 –5.53 0.03 1.69 0.67 –4.27 0.26 
Model 2       
 Nodal stagea       
  N1, N2, N3vs. N0 1.42 0.56 –3.63 0.46 3.76 1.49 –9.45 0.005 
 Cyclin D1 expression       
  Positive vs. Negativeb 2.44 0.98 –6.06 0.05 3.64 1.07 –12.25 0.03 
 p16INK4A expression       
  Negative vs. Positivec 3.23 1.49 –6.99 0.003 2.32 1.04 –5.16 0.04 
 Treatment       
  Surgery + Radiotherapy vs. Surgery 2.09 0.92 –4.75 0.08 1.58 0.62 –4.00 0.33 
a

Stage was determined by pathologic analysis.

b

Cyclin D1 expression: Positive, >10%; Negative, <10%

c

p16INK4A protein expression: Positive, >1%; Negative, <1%.

Table 5

Comparison of 5-year disease-free and overall survival according to cyclin D1, p16INK4A, and p53 status

Tumor no.5-year disease-free survival (%)P5-year overall survival (%)P
Cyclin D1a      
 Positive 100 67 0.013 69 0.008 
 Negative 47 83  88  
p16INK4Ab      
 Positive 65 85 0.003 81 0.022 
 Negative 78 64  72  
p53c      
 Positive 89 73  71  
 Negative 54 74 0.810 80 0.230 
Cyclin D1 positivea 53 60  64  
p16INK4A negativeb      
Cyclin D1 positivea 45 73  75  
p16INK4A positiveb   0.011  0.002 
Cyclin D1 negativea 25 80  87  
p16INK4A negativeb      
Cyclin D1 negativea 20 100  100  
p16INK4A positiveb      
Tumor no.5-year disease-free survival (%)P5-year overall survival (%)P
Cyclin D1a      
 Positive 100 67 0.013 69 0.008 
 Negative 47 83  88  
p16INK4Ab      
 Positive 65 85 0.003 81 0.022 
 Negative 78 64  72  
p53c      
 Positive 89 73  71  
 Negative 54 74 0.810 80 0.230 
Cyclin D1 positivea 53 60  64  
p16INK4A negativeb      
Cyclin D1 positivea 45 73  75  
p16INK4A positiveb   0.011  0.002 
Cyclin D1 negativea 25 80  87  
p16INK4A negativeb      
Cyclin D1 negativea 20 100  100  
p16INK4A positiveb      
a

Cyclin D1 expression: Positive, >10%; Negative, <10%.

b

p16INK4A protein expression: Positive, >1%; Negative, <1%.

c

p53 nuclear accumulation: Positive, >10%; Negative, <10%.

We thank members of the Departments of Anatomical Pathology at St. Vincent’s and Westmead Hospitals for providing the specimens and Dr. Susan Henshall and Professors John Grygiel and Ken Ho for critical review of the manuscript

1
Parkin D. M., Pisani P., Ferlay J. Estimates of the worldwide incidence of 25 major cancers in 1990.
Int. J. Cancer
,
80
:
827
-841,  
1999
.
2
Spitz M. R., Fueger J. J., Goepfert H., Hong W. K., Newell G. R. Squamous cell carcinoma of the upper aerodigestive tract. A case comparison analysis.
Cancer (Phila.)
,
61
:
203
-208,  
1988
.
3
Schottenfeld D. Alcohol as a co-factor in the etiology of cancer.
Cancer (Phila.)
,
43
:
1962
-1966,  
1979
.
4
Merletti F., Boffetta P., Ciccone G., Mashberg A., Terracini B. Role of tobacco and alcoholic beverages in the etiology of cancer of the oral cavity/oropharynx in Torino, Italy.
Cancer Res.
,
49
:
4919
-4924,  
1989
.
5
Spiro R. H., Strong E. W. Surgical treatment of cancer of the tongue.
Surg. Clin. N. Am.
,
54
:
759
-765,  
1974
.
6
Pignataro L., Pruneri G., Carboni N., Capaccio P., Cesana B. M., Neri A., Buffa R. Clinical relevance of cyclin D1 protein overexpression in laryngeal squamous cell carcinoma.
J. Clin. Oncol.
,
16
:
3069
-3077,  
1998
.
7
Sherr C. J. Cancer cell cycles.
Science (Washington DC)
,
274
:
1672
-1677,  
1996
.
8
Hunter T., Pines J. Cyclins and cancer. II. Cyclin D and CDK inhibitors come of age.
Cell
,
79
:
573
-582,  
1994
.
9
Califano J., Vanderriet P., Westra W., Nawroz H., Clayman G., Piantadosi S., Corio R., Lee D., Greenberg B., Koch W., Sidransky D. Genetic progression model for head and neck cancer: implications for field cancerization.
Cancer Res.
,
56
:
2488
-2492,  
1996
.
10
Weinberg R. A. The retinoblastoma protein and cell cycle control.
Cell
,
81
:
323
-330,  
1995
.
11
Hall M., Peters G. Genetic alterations of cyclins, cyclin-dependent kinases, and Cdk inhibitors in human cancer.
Adv. Cancer Res.
,
68
:
65
-108,  
1996
.
12
Masuda M., Hirikawa N., Nackashima T., Kuratomi Y., Komiyama S. Cyclin D1 overexpression in primary hypopharyngeal carcinomas.
Cancer (Phila.)
,
78
:
390
-395,  
1996
.
13
Akervall J. A., Michalaides R. J., Mineta H., Balm A., Borg A., Dictor M. Amplification of cyclin D1 in squamous cell carcinoma of the head and neck and the prognostic value of chromosomal abnormalities and cyclin D1 overexpression.
Cancer (Phila.)
,
79
:
380
-389,  
1997
.
14
Michalaides R. J., van Heelen N. M. J., Kristel P. M. P., Hart A. A. M., Loftus B. M., Hilgers F. J. M. Overexpression of cyclin D1 indicates a poor prognosis in squamous cell carcinoma of the head and neck.
Arch. Otolaryngol. Head Neck Surg.
,
123
:
497
-502,  
1997
.
15
Kyomoto R., Kumazawa H., Toda Y., Sakaida N., Okamura A., Iwanga M., Shintaku M., Yamashita T., Hiai H., Fukumoto M. Cyclin D1 gene amplification is a more potent prognostic factor than its protein overexpression in human head and neck squamous cell carcinoma.
Int. J. Cancer
,
74
:
576
-581,  
1997
.
16
Williams M. E., Gaffey M. J., Lawrence M. W., Wilczynski S. P., Schuuring E., Levine P. A. Chromosome 11q13 amplification in head and neck squamous cell carcinoma.
Arch. Otolarynyngol. Head Neck Surg.
,
119
:
1238
-1243,  
1993
.
17
Muller D., Millon M., Velten M., Bronner G., Jung G., Engelmen H. Amplification of 11q13 DNA markers in head and neck squamous cell carcinomas: correlation with clinical outcome.
Eur. J. Cancer
,
33
:
2203
-2210,  
1997
.
18
Ruas M., Peters G. The p16INK4a/CDKN2A tumor suppressor and its relatives.
Biochim. Biophys. Acta
,
1378
:
F115
-F177,  
1998
.
19
Liggett W. H., Sidransky D. Role of the p16 tumor suppressor gene in cancer.
J. Clin. Oncol.
,
16
:
1197
-1206,  
1998
.
20
Reed A. L., Califano J., Cairns P., Westra W. H., Jones R. M., Koch W. High frequency of p16 (CDKN2/MTS-1/INK4A) inactivation in head and neck squamous cell carcinoma.
Cancer Res.
,
351
:
3630
-3633,  
1996
.
21
Papadimitrakopoulou V., Izzo J., Lippman S. M., Lee J. S., Fan Y. H., Clayman G., Ro J. Y., Hittelman W. N., Lotan R., Hong W. K., Mao L. Frequent inactivation of p16INK4a in oral premalignant lesions.
Oncogene
,
14
:
1799
-1803,  
1997
.
22
Boyle J. O., Hakim J., Koch W., van der Riet P., Hruban R. H., Roa R. A., Correo R., Eby Y. J., Ruppert J. M., Sidransky D. The incidence of p53 mutations increases with progression of head and neck cancer.
Cancer Res.
,
53
:
4477
-4480,  
1993
.
23
Shim D. M., Kim J., Ro J. Y. Activation of p53 gene expression in premalignant lesions during head and neck tumorigenesis.
Cancer Res.
,
54
:
321
-326,  
1994
.
24
Brennan J. A., Mao L., Hruban R. H., Boyle J. O., Eby Y. J., Koch W. M., Goodman S. N., Sidransky D. Molecular assessment of histopathological staging in squamous-cell carcinoma of the head and neck.
N. Engl. J. Med.
,
332
:
429
-435,  
1995
.
25
Ahomadegbe J. C., Barrois M., Fogel S. High incidence of p53 alterations (mutation, deletion, overexpression) in head and neck primary tumors and metastases. Absence of correlation with clinical outcome: frequent protein overexpression in normal epithelium and in early non invasive lesions.
Oncogene
,
10
:
1217
-1227,  
1995
.
26
Koch W. M., Brennan J. A., Zahvrak M., Goodman S. N., Westra W. H., Schwab D., Yoo G. H., Jenlee D., Sidransky D. p53 mutation and locoregional treatment failure in head and neck squamous cell carcinoma.
J. Natl. Cancer Inst.
,
88
:
1580
-1586,  
1996
.
27
Alle K. M., Henshall S. M., Field A. S., Sutherland R. L. Cyclin D1 protein is overexpressed in hyperplasia and intraductal carcinoma of the breast.
Clin. Cancer Res.
,
4
:
847
-854,  
1998
.
28
Buckley M. F., Sweeney K. J., Hamilton J. A., Sini R. L., Manning D. L., Nicholson R. I., deFazio A., Watts C. K., Musgrove E. A., Sutherland R. L. Expression and amplification of cyclin genes in human breast cancer.
Oncogene
,
8
:
2127
-2133,  
1993
.
29
Musgrove E. A., Lilischkis R., Cornish A. L., Lee C. S., Setlur V., Seshadri R., Sutherland R. L. Expression of the cyclin-dependent kinase inhibitors p16INK4, p15INK4b and p21WAF1/CIP1 in human breast cancer.
Int. J. Cancer
,
63
:
584
-591,  
1995
.
30
Stone K. R., Mickey D. D., Wunderli H., Mickey G. H., Paulson D. F. Isolation of a human prostatic carcinoma cell line (DU 145).
Int. J. Cancer
,
21
:
274
-281,  
1978
.
31
Geradts J., Kratzke R. A., Niehans G. A., Lincoln C. E. Immunohistochemical detection of the cyclin-dependent kinase inhibitor 2/multiple tumor suppressor gene 1 (CDKN2/MTS1) product p16INK4a in archival human solid tumors: correlation with retinoblastoma protein expression.
Cancer Res,
,
55
:
6006
-6011,  
1995
.
32
Kaplan E. L., Meier P. Nonparametric estimation from incomplete observations.
J. Am. Stat. Assoc.
,
53
:
457
-481,  
1958
.
33
Cox D. R. Regression models and life tables (with discussion).
J. R. Stat. Soc.
,
34
:
187
-189,  
1972
.
34
Kotelnikov V. M., Coon J. S., Mundle S., Kelanic S., Lafollette S., Taylor S., Hutchinson J., Panje W., Caldarelli D. D., Preisler H. D. Cyclin D1 expression in squamous cell carcinomas of the head and neck and in oral mucosa in relation to proliferation and apoptosis.
Clin. Cancer Res.
,
3
:
95
-101,  
1997
.
35
Bartkova J., Lukas J., Muller H., Strauss M., Gusterson B., Bartek J. Abnormal patterns of D-type cyclin expression and G1 regulation in human head and neck cancer.
Cancer Res.
,
55
:
949
-956,  
1995
.
36
Mineta H., Borg A., Dictor M., Walhberg P., Wennerberg J. Correlation between p53 and cyclin D1 amplification in head and neck carcinoma.
Oral Oncol.
,
33
:
42
-46,  
1997
.
37
Nogueira C. P., Dolan R. W., Gooey J., Byahatti S., Vaughan C. W., Fuleihan N. S., Grillone G., Baker E., Domanowski G. Inactivation of p53 and amplification of cyclin D1 correlate with clinical outcome in head and neck cancer.
Laryngoscope
,
108
:
345
-350,  
1998
.
38
Nylander K., Schildt E. B., Eriksson M., Roos G. PCNA, Ki-67, p53, bcl-2 and prognosis in intraoral squamous cell carcinoma of the head and neck.
Ann. Cell Pathol.
,
14
:
101
-110,  
1997
.
39
James P., Fernandez P. L., Campo E., Nadel A., Bosch F., Aiza G. PRAD 1/CYCLIN D1 gene amplification correlates with messenger RNA overexpression and tumor progression in human laryngeal carcinomas.
Cancer Res.
,
54
:
4813
-4817,  
1994
.
40
Wang X., Pavelic Z. P., Li Y. Q., Gleich L., Radack K., Gluckman J. L., Sambrook P. J. Amplification and overexpression of the cyclin D1 gene in head and neck squamous cell carcinoma.
J. Clin. Pathol. Mol. Pathol.
,
48
:
M256
-M259,  
1995
.
41
Fortin A., Guerry M., Guerry R., Talbot M., Parise O., Schwaab G. Chromosome 11q13 gene amplifications in oral and oropharyngeal carcinomas: no correlation with subclinical lymph node invasion and disease recurrence.
Clin. Cancer Res.
,
3
:
1609
-1614,  
1997
.
42
Skapek S. X., Rhee J., Spicer D. B., Lassar A. B. Inhibition of myogenic differentiation in proliferating myoblasts by cyclin D1-dependent kinase.
Science (Washington DC)
,
267
:
1022
-1024,  
1995
.
43
Chandrasekaran C., Coopersmith C. M., Gordon J. I. Use of normal and transgenic mice to examine the relationship between terminal differentiation of intestinal epithelial cells and accumulation of their cell cycle regulators.
J. Biol. Chem.
,
271
:
28414
-28421,  
1996
.
44
Grandis J. R., Melhem M. F., Gooding W. E., Day R., Holst V. A., Wagener M. M., Drenning S. D., Tweardy D. J. Levels of TGF-α and EGFR protein in head and neck squamous cell carcinoma and patient survival.
J. Natl. Cancer Inst.
,
90
:
824
-832,  
1998
.
45
Musgrove E. A., Lee C. S. K., Buckley M. F., Sutherland R. L. Cyclin D1 induction in breast cancer cells shortens G1 and is sufficient for cells arrested in G1 to complete the cell cycle.
Proc. Natl. Acad. Sci. USA
,
91
:
8022
-8026,  
1994
.
46
Portugal L. G., Goldenberg J. D., Wenig B. L., Ferrer K. T., Nodzenski E., Sabnani J. B., Javier C., Eichselbaum R. R., Vokes E. E. Human papilloma virus expression and p53 gene mutations in squamous cell carcinoma.
Arch. Otalarynogol. Head Neck Surg.
,
123
:
1230
-1234,  
1997
.
47
Veneroni S., Silverstrini R., Costa A., Salvatori P., Faranda A., Molinari R. Biological indicators of survival in patients treated by surgery for squamous cell carcinoma of the oral cavity and oropharynx.
Oral Oncol.
,
33
:
408
-413,  
1997
.
48
van Heerden W. F., van Rensburg E. J., Hemmer J., Raubenheimer E. J., Engelbrecht S. Correlation between p53 gene mutation, p53 protein labeling and PCNA expression in oral squamous cell carcinoma.
Anticancer Res.
,
18
:
237
-240,  
1998
.
49
Wilentz R. E., Geradts J., Maynard R., Offerhaus G., Kang M., Goggins M., Yeo C. J., Kern S. E., Hruban R. H. Inactivation of the p16 (ink4a) tumor-suppressor gene in pancreatic duct lesions: loss of intranuclear expression.
Cancer Res.
,
58
:
4740
-4744,  
1998
.
50
Li H., Nichols M. A., Shay J. W., Xiong Y. Transcriptional repression of the D-type cyclin-dependent kinase inhibitor p16 by the retinoblastoma susceptibility gene product pRb.
Cancer Res.
,
54
:
6078
-6082,  
1994
.
51
Parry D., Bates S., Mann D. J., Peters G. Lack of cyclin D-Ckd complexes in Rb-negative cells correlates with high levels of p16INK4/MTS1 tumour suppressor gene product.
EMBO J.
,
14
:
503
-511,  
1995
.
52
Lydiatt W. M., Davidson B. J., Schantz S. P., Caruana S., Chaganti R. 9p21 deletion correlates with recurrence in head and neck cancer.
Head Neck
,
20
:
113
-118,  
1998
.