This study has identified molecular changes characteristic of early oral cancer progression. We reported previously that acquisition of the immortal phenotype is an early event in oral cancer development (F. McGregor et al., Cancer Res., 57: 3886–3889, 1997); our current data indicate that about half of oral dysplasia cultures are immortal, and this is associated with loss of expression of retinoic acid receptor (RAR)-β and the cell cycle inhibitor p16ink4a (p16), p53 mutations, and increased levels of telomerase/human telomerase reverse transcriptase mRNA. In contrast, increased expression of the epidermal growth factor receptor, known to be a characteristic of oral cancer, does not occur until after the dysplasia stage in squamous cell carcinomas. Acquisition of invasive properties as judged by an in vitro Matrigel invasion assay also does not occur until the carcinoma stage and is further increased in metastases. Interestingly, one atypical mortal dysplasia with a considerably extended life span has lost expression of RAR-β and p16, but it still expresses only wild-type p53 (albeit at a higher level than normal) and has not activated telomerase. RAR-β and/or p16 re-expression can be induced by treatment with 5-aza-2-deoxycytidine (Aza-C) in some immortal dysplasias, and this has been shown to be due to silencing of gene expression by promoter methylation. Aza-C treatment also down-regulated telomerase activity and human telomerase reverse transcriptase mRNA. Interestingly, with one dysplasia, Aza-C was able to reverse its immortal phenotype, as judged by morphological criteria and expression of the senescence-associated acid β-galactosidase activity during terminal growth arrest; this immortal dysplasia was the only one in which Aza-C treatment not only down-regulated telomerase activity but also induced re-expression of both RAR-β and p16. The possibility of reversing the immortal phenotype of some dysplasias by Aza-C may be of clinical usefulness.

Whereas early detection of oral cancer has improved, survival rates are poor and have remained at the same level for over two decades, with a 5-year survival rate of approximately 50% from diagnosis of SCC.4 Only a small proportion (5–10%) of the early dysplasia lesions progress to SCC, and early identification of this subset could potentially alter treatment and improve the poor survival rate for such cancers. Understanding the molecular events involved in oral cancer progression is vital if progress is to be made.

Several molecular changes in mitogenic pathways or cell cycle regulators have been identified in oral and HNSCC, for example, the EGFR (1, 2), cyclin D1 (3, 4), PLK1 (5), p21WAF1/CIP1(6, 7), p27KIP1(8), p53 (reviewed in Ref. 9), p16ink4a (p16) (9, 10, 11, 12, 13), and the CDC25A and CDC23B phosphatases (14). Inactivation of other putative suppressor genes on several chromosomes has also been demonstrated in HNSCC, as judged by loss of heterozygosity or other approaches, and such changes have recently been used to predict malignant risk for low-grade oral dysplasias (15).

Another feature of cancer development is the loss of replicative senescence and acquisition of an immortal phenotype and an unlimited life span potential (see Ref. 16 for a recent general review). Control of senescence involves multiple growth-regulatory pathways, and increased expression of a number of factors involved in cell cycle regulation, such as the cell cycle inhibitors p16ink4A and p21WAF1/CIP1, has been shown to occur as cells approach senescence (reviewed in Ref. 17). An important trigger for entry into senescence in human cells is the progressive erosion of the chromosome telomeres, with senescence commencing once a critical length is reached (reviewed in Ref. 18. Activation of telomerase, which allows cells to maintain and increase the length of their telomeres, has been observed in a variety of cancers, although alternative mechanisms for maintenance of telomere length have also been described (18). However, although expression of telomerase in normal fibroblasts confers an extended life span, it does not result in transformation of the cells (19, 20), and telomerase-positive keratinocytes require additional alterations such as p16ink4 loss or pRb inactivation before immortalization occurs (21, 22).

Retinoids and their receptors are central to the normal growth and differentiation of a variety of epithelial cells, and, in particular, loss of RAR-β has been associated with malignant progression in a variety of cancers, including HNSCC, where loss of expression is found in premalignant lesions (23, 24). Our previous study showed that loss of expression of both RAR-β and p16ink4 was associated with a change in proliferative life span potential from mortality to immortality (25). In the present study, we have therefore determined the stage of oral cancer progression at which characteristic molecular changes occur and how they are related to the acquisition of the immortal phenotype. We have also investigated whether the immortal phenotype can be reversed by treatment with Aza-C in view of evidence that inactivation of genes such as p16 by methylation can be reversed by Aza-C and its potential relevance for use of demethylating agents in cancer therapy.

Cell Culture.

The derivation, characterization, and maintenance of the primary human keratinocyte cultures have been described previously (25). Normal cultures (FNB, FNB5, and FNB6) were obtained from three healthy volunteers; these cultures senesced at 10, 20, and 11.5 PDs, respectively. Several previously described (25) cultures from apparently normal tissue adjacent to tumors (NB1, NB5, NB9, and NT11) senesced after 20.5, 13, 7.5, and 8 PDs, respectively. The characteristics of the dysplasias are summarized in Table 1. Most were leukoplakias, except for D19 and D35, which were erythroleukoplakias. Most were from high-risk sites (lateral tongue, 11 of 16; floor of mouth, 3 of 16; or encompassing both sites, 2 of 16), and only 1 was from buccal mucosa. Cultures from these dysplasia biopsies were divided according to their proliferative fate: D6, D8, D25, D30, D41, D47, D48, and D17 senesced after 3.5–45 PDs; whereas D4, D9, D19, D20, D34, D35, and D38 were considered immortal after having completed >100 PDs. Cultures were maintained on a feeder layer of irradiated 3T3 fibroblasts that was routinely removed by EDTA treatment before analysis of the keratinocyte culture.

RNA Analysis.

RNA extraction and Northern blotting were carried out as described previously (25). The RAR-β probe was made from a plasmid containing the full-length RAR-β2 cDNA sequence provided by Prof. P. Chambon (Strasbourg, France).

Western Blot Analysis.

Protein extraction and Western blotting were carried out as described previously (25). Primary antibody incubations were carried out using the following antibodies: p16ink4A (C-20; Santa Cruz Biotechnology, Santa Cruz, CA); p21WAF1/CIP1 (Affiniti; Transduction Laboratories, Lexington, KY); p27KIP1 (C-19; Santa Cruz Biotechnology); EGFR (Santa Cruz Biotechnology); PLK1 (Affiniti; Transduction Laboratories); cyclin D1 (Ab3; Oncogene Research Products, Boston, MA); and p53 (D01; Santa Cruz Biotechnology). Membranes were reprobed with a polyclonal antibody against total ERK1 and ERK2 [ERK1(C-16)-G; Santa Cruz Biotechnology] to ensure even loading and transfer because they were found to be uniformly expressed in all keratinocyte cultures regardless of proliferative fate. Unless otherwise stated, protein was extracted from cultures during the proliferative phase of growth.

p53 Sequencing.

This was performed using proofreading DNA polymerase (Qiagen, Crawley, United Kingdom) and the following primers: exon 2, 2F/2R for amplification and sequencing; exon 3, 3F/4R for amplification and sequencing; exon 4, 3F/4R for amplification and 3F/4F/4R for sequencing; exons 5 and 6, 5F/5R for amplification and sequencing; exon 7, 7F/8R for amplification and 7F/7R for sequencing; exons 8 and 9, 7F/8R for amplification and 8F/8R for sequencing; exon 10, 10F/10R for amplification and sequencing; and exon 11, 11F/11R for amplification and 11F/11BR for sequencing. Primer sequences were as follows: 2F, 5′-CTCAAAGAAGTGCATGGCTGG-3′; 2R, 5′-CAGGAAGTCTGAAAGACAAGAGC-3′; 3F, 5′-GAAGCGAAAATTCATGGGACTGAC-3′; 4R, 5′-CATTGAAGTCTCATGGAAGCCA-3′; 4F, 5′-ACCTGGTCCTCTGACTGCTCT-3′; 5F, 5′-CGTGTTCCAGTTGCTTTATCTG-3′; 5R, 5′-GGAGGGCCACTGACAACCA-3′; 7F, 5′-AGGCGCACTGGCCTCATCTT-3′; 7R, 5′-AGGGGTCAGCGGCAAGCAGA-3′; 8F, 5′-TTGGGAGTAGATGGAGCCT-3′; 8R, 5′-AACTTTCCACTTGATAAGAGGTC-3′; 10F, 5′-GTACTGTGAATATACTTACTTCTCC-3′; 10R, 5′-CAAAGAAGTATCCACACTCGTC-3′; 11F, 5′-CACTCATGTGATGTCATCTCTCC-3′; 11R, 5′-CAAGACTTGACAACTCCCTC-3′; and 11BR, 5′-ATCTAAGCTGGTATGTCCTACTC-3′. Sequencing of fragments in both directions was performed according to the manufacturer’s protocols (ABI prism big dye terminator reaction kit; Applied Biosystems). Where a mutation was found, the result was confirmed using an independent PCR amplification product.

SABG Assay.

Cells were maintained in culture for a minimum of 3 weeks to confirm that proliferative growth had ceased. The 3T3 feeder layer was removed, and the cells were fixed and stained essentially as described by Dimri et al.(26). Cultures were incubated with stain for approximately 16 h at 37°C in a humid box.

Telomerase Assay.

Telomerase activity was measured in cell lysates containing 20–200 ng of protein, using the TRAPeze telomerase detection kit (Intergen, Oxford, United Kingdom) according to the manufacturer’s recommendations. The feeder layer of irradiated 3T3 fibroblasts was carefully removed before analysis to eliminate any residual contaminating activity from these cells. WI38 and HeLa cells were used as negative and positive controls, respectively. The PCR products were analyzed by electrophoresis on a 12.5% polyacrylamide gel and visualized using ethidium bromide staining.

Semiquantitative and Quantitative RT-PCR Assays for hTERT mRNA.

This was performed using the Taqman reverse transcription reagents and Taqman PCR core reagent kit (Perkin-Elmer, Branchberg, NJ), according to the manufacturer’s protocols. The hTERT primers used were ACAGCACTTGAAGAGGGTGCA and TTCCCGATGCTGCCTGAC, and the fluorescent probe was 6-carboxy-fluorescein-TGCGGGAGCTGTCGGAAGCA-6-carboxy-tetramethyl-rhodomine. This method measures the number of PCR cycles required to obtain exponential amplification of specific PCR product and comparing this with a standard curve. Similar analysis was also performed of the levels of 7S RNase as an internal control in the same samples, and the relative levels of hTERT to RNase were calculated.

In Vitro Invasion Assay.

This was performed as described previously by our colleagues (27). A total of 105 FNB, D6, or D4 cells or 5 × 104 B56, B31, and B22 cells were seeded as an almost confluent monolayer on the bottom of a porous filter of the transwell, and the cells were washed three times with PBS before the filter was inverted and placed in growth medium, and the inside was filled with a plug of Matrigel covered with a layer of medium with or without 100 ng/ml epidermal growth factor. After 5 days, the wells were stained with propidium iodide and then scanned on the confocal microscope to determine the number of cells on the underside of the filter and at 10-μm intervals above it, taking four fields at random for each measurement. The extent of invasion was calculated as the percentage of cells that had migrated 20 μm or more from the bottom of the filter.

Methylation-specific PCR.

This was performed using the CpGenome DNA modification kit (catalogue number S7820; Intergene) according to the manufacturer’s instructions. The primers used were those used previously by Virmani et al.(28) for RAR-β [methylated form, 5′-TCGAGAACGCGAGCGATTCG-3′ (sense) and 5′-GACCAATCCAACCGAAACGA-3′ (antisense); unmethylated form, 5′-TTGAGAATGTGAGTGATTTGA-3′ (sense) and 5′-AACCAATCCAACCAAAACAA-3′ (antisense)] and Baur et al.(29) for p16 exon 1 [methylated form, 5′-TTATTAGAGGGTGGGGCGGATCGC-3′ (sense) and 5′-GACCCCGAACCGCGACCGTAA-3′ (antisense); unmethylated form, 5′-TTATTAGAGGGTGGGGTGGATTGT-3′ (sense) and 5′-CAACCCCAAACCACAACCATAA-3′ (antisense)].

Acquisition of the Immortal and Invasive Phenotypes

We have investigated the molecular changes associated with oral cancer progression using primary cultures of biopsies from very early lesions as well as SCC, using the 3T3 feeder layer culture system of Rheinwald and Beckett (30) that has been shown to facilitate growth of keratinocytes at all stages of cancer progression and to maintain the characteristics of the original tumors; it has also been reported to avoid the premature growth arrest associated with culture of epithelial cells in chemically defined keratinocyte growth medium directly on plastic culture dishes (31). The series of 14 dysplasia cultures derived mainly from biopsies from similar “high-risk” sites (lateral tongue and floor of mouth) is an extremely valuable resource for studies of the early stages of oral cancer progression. We showed previously using a smaller series of cultures that acquisition of the immortal phenotype usually occurs at the dysplasia stage of oral cancer progression (25), and this has been confirmed by more recent work with all 14 cultures. There is no obvious correlation between pathology (i.e., mild/moderate/severe dysplasia status) and mortality phenotype (Table 1). Both of our erythroleukoplakia cultures were immortal (Table 1), although this is not generally the case because four erythroplakia cultures isolated previously were all mortal (32).

We have also investigated the stage at which cells acquire invasive properties, using an in vitro Matrigel invasion assay (27). This showed clearly that significant invasive properties could be detected with SCCs, particularly those derived from a nodal metastasis, but not with normal oral mucosa (as expected) or dysplasias, even if they were immortal (Fig. 1). Thus, in general, acquisition of immortality is an earlier event during oral cancer progression than acquisition of invasive properties.

Molecular and Phenotypic Changes During Oral Cancer Progression

EGFR, PLK1, and p53.

The expression of EGFR (1, 2), PLK1 (5), and cyclin D1 (3, 4) has been reported to be elevated in oral cancers, but it is not known at what stage this occurs. To address this point, we investigated the expression of these proteins in our panel of primary cultures of normal oral mucosa, MDs and IDs, and carcinomas. No reproducible differences in expression of EGFR were found between normal oral mucosa and dysplasias, but there was a significant increase in SCCs (Fig. 2). No consistent stage-specific changes in expression of PLK1 or cyclin D1 were observed, although there was some variability between individual cultures (Fig. 2). In contrast, the expression of p53 was at a similar low level in normal oral mucosa and all but one of the MDs but was significantly increased in IDs and SSCs (Fig. 2). Interestingly, the MD with atypically high p53 expression (D17) had an extended life span (but nevertheless eventually became senescent; Fig. 2). Complete sequencing of the p53 gene coding regions has confirmed the presence of mutations in all of the IDs, whereas all of the MDs (D38, D8, D47, D48, D41, D30, D25, and D6), including D17, had wild-type p53 (data not shown). The mutations were base changes or insertions mainly affecting the protein sequence in the DNA-binding domain or, in one case, the tetramerization domain; about half the mutations were heterozygous. p53 mutations in our oral SCC lines have been reported previously (32, 33, 34).

RAR-β and the Cell Cycle Inhibitors p16, p21, and p27.

Our previous analysis of dysplasias revealed a change in the regulation of RAR-β expression during the normal to dysplasia transition because its expression in MDs was constitutive in tissue culture, whereas with cultures of normal oral mucosa, it was retinoic acid dependent (25). The constitutive high-level expression of RAR-β in MDs has been confirmed in the current larger series of dysplasias (Fig. 3,A), with one interesting exception, D17 (see below), which has an extended life span but eventually becomes senescent. Because various isoforms of RAR-β have been identified in cells, some of which may act as dominant negative forms, we determined by isoform-specific RT-PCR methods (28, 35) precisely which RAR-β isoform(s) are expressed in normal oral mucosa and whether there are any changes in MDs; however, in all cases, the RAR-β2 isoform was the predominant form expressed, with minor amounts of RAR-β4 (data not shown). p16 was also expressed in six of seven MD cultures (although perhaps generally at somewhat lower levels than in normal cultures); however, like RAR-β, p16 was undetectable in D17 (Fig. 3,B). In contrast to MDs, RAR-β and p16 expression was undetectable in six of seven IDs, with the exception of D38, which expresses p16 but not RAR-β (Fig. 3); by sequencing, the p16 gene appears normal, but it is possible that an unknown mutation has occurred elsewhere in the p16-regulated pathway. Rare SCCs also express normal levels of wild-type p16 (11). In contrast to p16, there were no consistent changes in the expression of the other cell cycle inhibitors, p21 or p27, in dysplasias (Fig. 3 B) or SCC (data not shown; note that the apparently lower amount of p21 in D19 is not a consistent finding in other experiments).

Further evidence that RAR-β expression is associated with the program of senescence is that in several experiments RAR-β expression was found to increase during culture as normal oral mucosa cells entered the terminal PDs immediately before senescence, although in some cases the level then decreased in cells that had been irreversibly stationary for about 2 weeks due to senescence (Fig. 4). This increase in RAR-β was not due simply to prolonged passage in culture or to the culture having completed an increased number of PDs because, for example, an ID culture showed no expression of RAR-β after completing 88 PDs during culture for 9 months (data not shown). In contrast, p16 expression remained constant from the earliest passage available (Fig. 4), suggesting that p16 may be involved earlier, possibly as an initiator of the senescence program. The high level of p16 expression even in the earliest passages of our normal oral mucosa cultures, despite being isolated and maintained on 3T3 feeders (31), is probably due to the fact that our biopsies were obtained from adult donors (as were our biopsies of dysplasias and SCCs).

In view of the role of telomerase in generating the immortal phenotype, telomerase activity was also determined in oral cultures from biopsies at all stages of cancer progression using the sensitive (but only semiquantitative) TRAPeze assay. WI38 and HeLa cells were also included as negative and positive controls. Telomerase assays were performed using different amounts of extract to ensure that the assay was not saturated by the amounts of telomerase in the samples. Using 200 ng of extract in the TRAP assay, all of the ID cultures had readily detectable telomerase activities, whereas the telomerase activities of MDs and normal oral mucosa were barely detectable or not detectable (Fig. 5,A). However, using lower amounts of extract in the TRAP assays revealed that the telomerase activities of IDs and SCCs varied considerably, with some showing significant activity even using 10 ng of extract [particularly SCC B22 (from a metastasis) and IDs D19 and D20 (Fig. 5 B)].

To determine the mechanism of activation of telomerase in more detail, the levels of the telomerase reverse transcriptase subunit (hTERT) mRNA and template RNA (hTERC) were also measured. By Northern blotting, the level of hTERC RNA did not change dramatically in cultures of biopsies from all stages of cancer progression, although IDs and SCC may have somewhat higher levels of hTERC RNA than normal oral mucosa and MDs (Fig. 6,A). In contrast, by semiquantitative RT-PCR, there was a dramatic increase in hTERT mRNA in IDs and carcinomas compared with MDs or normal oral mucosa (Fig. 6,A), and this was confirmed by quantitative real-time RT-PCR (Fig. 6,B). These data also confirmed that the much higher level of telomerase activity in the subset of IDs (such as D19 and D20) was due to a large increase in the level of hTERT (compare Figs. 5,B and 6 B). It is presently too early to assess whether high expression of hTERT in these dysplasias has any clinical significance, although it is interesting that the D19 ID did subsequently progress rapidly to aggressive carcinoma within 3 years.

Because hTERT gene expression has been reported to be up-regulated by c-Myc (36) and down-regulated by Mad (37), we measured the levels of Myc and Mad proteins in cultures of normal oral mucosa, dysplasias, and SCCs, but no changes were found (data not shown). Wild-type p53 is also known to down-regulate hTERT gene expression (38), so the mutations and changes in the levels of p53 found in IDs could be responsible, at least in part, for the associated increases in hTERT mRNA.

Inactivation of p16 (10, 12) and RAR-β (28) in HNSCC or lung cancer has been shown to be sometimes due to promoter methylation, and this can be reversed by treatment with Aza-C. We tested whether similar methylation mechanisms may be responsible for the inactivation of RAR-β in IDs in our oral cancer model. These experiments revealed that treatment with Aza-C induced expression of RAR-β and/or p16 in some but not all IDs, but not in any of the carcinomas tested (B31, B56, B68, or B22; Fig. 7). RAR-β was strongly induced in IDs D34 and D19 and weakly induced in D35, D38, and D9 as well as in the SK D36 (Fig. 7,A), whereas p16 was strongly induced in IDs D9 and D34 (Fig. 7,B). Moreover, p16 was not induced in D17, the MD with an extended life span, and RAR-β was very weakly induced, if at all (Fig. 7, A and B). Thus, D34 was the only ID in which both RAR-β and p16 were strongly induced. We then determined whether these effects of Aza-C were due to changes in the methylation status of CpG islands in the RAR-β P2 and p16 exon 1 promoters known to be involved in their regulation (28, 29), using the bisulfite modification method to detect unmethylated cytosine residues followed by methylation-specific PCR (Fig. 8,A). This showed that the p16 promoter was completely methylated in D9 and D34, whereas it was unmethylated in D20 and D35. No PCR product was found with D19, suggesting that the p16 promoter contains a deletion in the region of the potential methylation site or is mutated in one of the primer sites (Fig. 8,A). The p16 promoter site was unmethylated in the D25 dysplasia, in which p16 is expressed. The RAR-β promoter site was mainly methylated in D9, D19, D20, and D35, whereas in D34, about half of the sites were methylated (Fig. 8,A). We then determined whether the methylation of the RAR-β and p16 promoters was reversed by Aza-C treatment in D19 and D34; the results show an increase in unmethylation at the RAR-β promoter site in both D19 and D34 and at the p16 promoter site in D34 (Fig. 8 B). Thus, lack of expression of p16 and RAR-β in some IDs is associated with promoter methylation, and this can be reversed by Aza-C; however, in other cases, the promoters are unmethylated, implying that gene expression is prevented for other reasons, e.g., mutation or deletion.

We also tested whether such Aza-C treatment affected the expression of telomerase or hTERT mRNA in immortal oral dysplasias or carcinomas; this showed that the levels of telomerase activity and hTERT mRNA in the IDs tested were reduced but that the change in the carcinoma was less marked (Fig. 9, A and B).

Finally, we investigated whether Aza-C treatment could induce senescence in IDs or carcinomas, as judged by morphological changes and expression of the SABG (26). For this experiment, we compared an ID in which both p16 and RAR-β were reactivated by Aza-C (D34), an ID in which only RAR-β was reactivated (D19), and an ID (D20) and a SCC (B56) in which neither RAR-β nor p16 was reactivated. Aza-C induced a slower growth rate and a flatter morphology in all cultures (Fig. 10). However, SABG expression was strongly induced by Aza-C only in the D34 ID (in which both RAR-β and p16 were inducible by Aza-C), with only occasional and weak expression in the other cultures (e.g., D20; Fig. 10). We also tested whether the various cultures responded to retinoic acid treatment for 7 days; although this retarded the growth of the colonies, there was no evidence of induction of SABG (data not shown).

Molecular Changes at the Dysplasia Stage of Oral Cancer Progression.

We showed previously that whereas RAR-β expression is retinoic acid dependent in normal oral mucosa in culture, it becomes constitutively expressed in MDs but is completely suppressed in IDs and SCCs (25). Several isoforms of RAR-β have been reported, including some with dominant negative functions (35), but we have found no difference in RAR-β isoform expression (RAR-β2 is primarily expressed, with a small amount of RAR-β4) between MDs and normal oral mucosa. This strongly suggests that RAR-β expression is associated with maintaining the mortal phenotype and therefore has to be suppressed for a cell to become immortal. In this report, we present further evidence supporting this hypothesis because RAR-β expression increases in normal oral mucosa as the cells approach the final stages of senescence, as is the case also with normal human mammary epithelial cells (39). In contrast, p16 is highly expressed even in the earliest passages of normal oral mucosa cultures available, consistent with its putative role in initiating the senescence program (17). We hypothesize therefore that MDs are equivalent to late-stage senescent normal cells in terms of RAR-β and p16ink4A expression but have acquired other genetic changes that give them an extended but finite life span. What these genetic changes might be is unclear at present, but p53 mutations are excluded because these have only been found in IDs.

Molecular Changes Associated with the Development of the Immortal Phenotype.

Increasing evidence points to senescence control as an important tumor suppressor mechanism. Loss of RAR-β (23, 24, 25) and p16 (13) expression has been well documented at the dysplasia stage of head and neck cancer. The functional importance of RAR-β loss is supported by transgenic experiments showing that transgenic mice overexpressing an antisense RAR-β2 transgene (40) or a truncated (dominant negative) RAR-β transgene (41) develop hyperplasias and tumors, whereas overexpression of the β2 isoform of RAR-β inhibits the growth of lung cancer cells both in vitro and in vivo(42, 43). Our previous work was the first to provide evidence suggesting that the function of RAR-β and p16 may be to regulate the mortal-to-immortal transition (10, 25), and we now show that activation of telomerase/hTERT and p53 mutations are also involved. The fact that about half of our IDs are heterozygous for wild-type/mutant p53, whereas 11 of 13 SCCs were found previously to be homozygous for a mutant p53 allele (33), is of interest because the combination of heterozygous p53 mutations and short telomeres in mice has been shown to promote epithelial cancers involving a type of genetic instability postulated to be important in the process of human cancer development (44). The presence of p53 mutations in a high proportion of oral carcinomas has been demonstrated previously (33, 34, 46), but the proportion of dysplasias with p53 mutations is lower (45). No p53 mutations were found in erythroplakias that were mortal in culture (32), consistent with our own conclusions with dysplasias in general. Of particular interest is our intermediate-phenotype D17 dysplasia culture that has an extended life span but remains mortal; like IDs, it has extremely low levels of RAR-β and p16, but p53 is wild-type (albeit expressed at a higher than normal level), and telomerase activity is extremely low. Wild-type p53 can repress hTERT mRNA transcription in human tumor cells, but this repressor activity is lost in p53 mutants (38). Repression of hTERT expression by wild-type p53 may therefore explain the association of p53 mutations and hTERT activation at the mortal-to-immortal transition in our oral dysplasias, including the atypical D17 dysplasia. However, other work expressing papilloma virus E6 in primary human foreskin keratinocytes shows that telomerase activation and p53 degradation are insufficient per se for their immortalization (46).

It is unclear at present whether RAR-β is directly involved in regulating the senescent phenotype, and, if so, what its relevant target genes are. It is unlikely that p16 is a RAR-β target gene because p16 is expressed in normal oral mucosa cells in culture in the absence of RAR-β. It is unlikely that loss of RAR-β expression in conjunction with p53 mutations is required to activate hTERT because p53 regulates hTERT in HeLa cells (38) that do not express RAR-β.

Later Changes.

Other molecular and phenotypic characteristics of HNSCCs occur after the dysplasia stage, for example, up-regulation of EGFR expression and acquisition of invasive properties. Increased expression of EGFR has been reported in severe but not mild dysplasias adjacent to HNSCC (47), but not in erythroplakias (48). Interestingly, EGFR overexpression may be corrected by restoration of wild-type p53 function in a HNSCC line (49) and also by retinoic acid treatment (50). However, no clear stage-specific changes in expression of cyclin D1 or PLK1 were found in our cultures. It is possible that the differences in expression of these markers between normal oral mucosa and SCCs found in vivo(3, 4, 5) are masked when the cells are triggered to proliferate rapidly in tissue culture.

Prognostic Significance of Molecular Changes During Oral Cancer Progression?

The tongue and the floor of the mouth are well established as high-risk sites within the oral cavity. The differences between our dysplasias are clearly not due to site differences because virtually all are from these two high-risk sites. There is a well established difference in progression rate between erythroplakia and leukoplakia types of dysplasia (about 5% for leukoplakias and about 25–50% for erythroplakias within 5–10 years), and this is reflected in the fact that, to date, both our erythroleukoplakias (D19 and D35) have progressed, in contrast to only 1 of 12 leukoplakias. Histological grading is not thought to be a reliable prognostic indicator, but in fact, all three of our dysplasias that have progressed were graded as moderate/severe to severe. Overall, there is no absolute association of histological grading and mortal/immortal phenotype, although dysplasias that prove to be mortal in culture tend to have been classified as mild or mild/moderate at biopsy, and conversely, IDs tend to have been classified as severe (Table 1). Clearly, not all IDs progress rapidly in vivo because only a small minority of dysplasias progress to SCC. To date, two of seven of our IDs and one of eight of our MDs have progressed. Moreover, in our previous study, three of four erythroplakias progressed to malignant tumors, despite being mortal in culture (32); conversely, a subset of quite advanced SCCs are mortal (34). Thus, it is not at all certain that the rapidly progressing dysplasias are actually a subset of the IDs. Our ongoing prospective study should clarify this and determine whether any of the molecular changes reported in this study or other changes in gene expression profile identified by microarray analysis have useful prognostic significance for oral dysplasia progression.

In other studies, elevated p53 expression or functional loss of p53 has been shown to be of prognostic significance in HNSCC (51, 52), as has loss of p16 expression in premalignant laryngeal lesions, especially in combination with p53 overexpression (53). However, no relationship between high RAR-β expression level and prognosis was found in oral leukoplakia (54), whereas in stage I non-small cell lung cancer, it correlated with worse rather than better survival (55). Increased EGFR levels have been found to be predictive of reduced survival in HNSCC (56), as have elevated PLK1 levels in non-small cell lung cancer (5, 57). Increased expression of cyclins D1 and E is also frequent in bronchial preneoplasia in vivo before carcinoma development (58), and cyclin D1 overexpression in conjunction with p16 loss can predict reduced survival in carcinoma of the anterior tongue (59). Furthermore, loss of p27 expression has been found to be associated with poorer survival in non-small cell lung cancer (60).

Reversal of the Immortal Phenotype by Aza-C.

The final important point of interest from our present study concerns the possibility that the immortal phenotype of certain dysplasias can be reversed by treatment with Aza-C. Our experiments show that expression of RAR-β or p16 (and sometimes both) can be reactivated in some IDs but not in any of the carcinomas tested, and this correlates with the methylation status of the RAR-β P2 and p16 promoters in the various cultures. This indicates that expression of RAR-β and/or p16 is frequently initially suppressed in IDs by methylation, but later in cancer progression, other genetic changes silence expression irreversibly. This is consistent with previous data from our colleagues and others that p16 can be suppressed by methylation in some HNSCCs (10, 12), but most HNSCCs have deletions in p16 (11). Inactivation of RAR-β due to methylation has been reported previously for lung cancer (28). Our experiments also show that telomerase activity and hTERT mRNA are also suppressed by Aza-C in all of the IDs tested, but not in carcinomas. Suppression of hTERT by Aza-C has also very recently been demonstrated in prostate cancer cells (61). Finally, with one dysplasia, we found that Aza-C treatment reversed its immortal phenotype, as judged by morphological criteria and expression of acid β-galactosidase activity during the terminal growth arrest; significantly, this was the only ID to show Aza-C-induced reexpression of both RAR-β and p16 expression as well as suppression of telomerase activity. This conclusion could have implications for identifying the dysplasias for which chemotherapy by demethylating agents might be useful.

Fig. 1.

In vitro invasion assay using primary cultures of three normal oral mucosa biopsies from buccal mucosa (FNB and NB9) or tongue (NT11), a MD (D6), an ID (D4), two stage 4 SCCs (B56 and B31), and a lymph node metastasis (B22). The percentage of cells migrating into Matrigel was measured with or without the presence of EGF as a chemoattractant in the medium above the Matrigel (see “Materials and Methods” for details). The results are the averages (with SEs) of several experiments (FNB, three experiments; D4, four experiments; D6, four experiments; B56, nine experiments; B31, three experiments; B22, five experiments), except for NT11 and NB9, for which data from a single experiment are shown.

Fig. 1.

In vitro invasion assay using primary cultures of three normal oral mucosa biopsies from buccal mucosa (FNB and NB9) or tongue (NT11), a MD (D6), an ID (D4), two stage 4 SCCs (B56 and B31), and a lymph node metastasis (B22). The percentage of cells migrating into Matrigel was measured with or without the presence of EGF as a chemoattractant in the medium above the Matrigel (see “Materials and Methods” for details). The results are the averages (with SEs) of several experiments (FNB, three experiments; D4, four experiments; D6, four experiments; B56, nine experiments; B31, three experiments; B22, five experiments), except for NT11 and NB9, for which data from a single experiment are shown.

Close modal
Fig. 2.

Levels of EGFR, PLK1, cyclin D1, and p53 proteins in normal tissue (N), MDs, IDs, and SCCs. The level of the proteins was measured by Western blotting, using the total levels of ERK1 and ERK2 or tubulin as loading controls.

Fig. 2.

Levels of EGFR, PLK1, cyclin D1, and p53 proteins in normal tissue (N), MDs, IDs, and SCCs. The level of the proteins was measured by Western blotting, using the total levels of ERK1 and ERK2 or tubulin as loading controls.

Close modal
Fig. 3.

Expression of RAR-β mRNA and the cell cycle inhibitors p16INK4A, p21, and p27 in normal oral mucosa (N), MDs, IDs, and SCCs. A, RAR-β mRNA expression was analyzed by Northern blotting using 7S RNase as a loading control. The left panel shows alternative blots for selected MDs or IDs, in some cases in duplicate to show the consistency of the results. RNA was extracted from the epithelial cell cultures after removal of irradiated 3T3 feeders. Control experiments showed that irradiated 3T3 feeders contain very little RNA and that neither live or irradiated 3T3 cells contain significant amounts of RAR-β mRNA (A, right panel). Thus the RAR-β measurements are not affected by any residual irradiated feeders still contaminating the epithelial cells. B, p16INK4A, p21, and p27 protein levels were analyzed by Western blotting, using the total level of ERK1 and ERK2 as a loading control.

Fig. 3.

Expression of RAR-β mRNA and the cell cycle inhibitors p16INK4A, p21, and p27 in normal oral mucosa (N), MDs, IDs, and SCCs. A, RAR-β mRNA expression was analyzed by Northern blotting using 7S RNase as a loading control. The left panel shows alternative blots for selected MDs or IDs, in some cases in duplicate to show the consistency of the results. RNA was extracted from the epithelial cell cultures after removal of irradiated 3T3 feeders. Control experiments showed that irradiated 3T3 feeders contain very little RNA and that neither live or irradiated 3T3 cells contain significant amounts of RAR-β mRNA (A, right panel). Thus the RAR-β measurements are not affected by any residual irradiated feeders still contaminating the epithelial cells. B, p16INK4A, p21, and p27 protein levels were analyzed by Western blotting, using the total level of ERK1 and ERK2 as a loading control.

Close modal
Fig. 4.

Expression of RAR-β and p16ink4A in a senescing culture of normal oral mucosa, FNB5. Paired protein and RNA samples were obtained, and the corresponding PDs (dbs) attained were determined at each passage of the culture. Top panel, Northern blotting of RAR-β expression, with a loading control of 7S RNA. The increase in RAR-β RNA as cells approach senescence followed by a decline in cells that are actually senescent has been found in two separate experiments. Bottom panel, Western blotting of p16ink4A expression, using an anti-ERK1/ERK2 antibody as loading control.

Fig. 4.

Expression of RAR-β and p16ink4A in a senescing culture of normal oral mucosa, FNB5. Paired protein and RNA samples were obtained, and the corresponding PDs (dbs) attained were determined at each passage of the culture. Top panel, Northern blotting of RAR-β expression, with a loading control of 7S RNA. The increase in RAR-β RNA as cells approach senescence followed by a decline in cells that are actually senescent has been found in two separate experiments. Bottom panel, Western blotting of p16ink4A expression, using an anti-ERK1/ERK2 antibody as loading control.

Close modal
Fig. 5.

Measurement of telomerase activity in oral cultures. Lysates from the same number of cells were prepared, and telomerase activity was measured by the TRAPeze assay. The gel shows the PCR products generated visualized using ethidium bromide staining. N, normal oral mucosa. WI38 and HeLa cells were also included as telomerase-negative and -positive controls, respectively. An important feature of the TRAPeze telomerase detection kit is the inclusion of an internal standard, a 36-bp band (S-IC), in every lane. The internal control band appears weaker in those samples with high telomerase activity because amplification of the TRAP products and that of the S-IC band are semicompetitive. In A, 200 ng of extract were used in all samples; in B, 200 ng (H) or 10 ng (L) were used to control for saturation of the assay in the samples with the highest telomerase activity.

Fig. 5.

Measurement of telomerase activity in oral cultures. Lysates from the same number of cells were prepared, and telomerase activity was measured by the TRAPeze assay. The gel shows the PCR products generated visualized using ethidium bromide staining. N, normal oral mucosa. WI38 and HeLa cells were also included as telomerase-negative and -positive controls, respectively. An important feature of the TRAPeze telomerase detection kit is the inclusion of an internal standard, a 36-bp band (S-IC), in every lane. The internal control band appears weaker in those samples with high telomerase activity because amplification of the TRAP products and that of the S-IC band are semicompetitive. In A, 200 ng of extract were used in all samples; in B, 200 ng (H) or 10 ng (L) were used to control for saturation of the assay in the samples with the highest telomerase activity.

Close modal
Fig. 6.

Quantitation of hTERC RNA and hTERT mRNA levels in oral cultures. A, measurement of hTERT and glyceraldehyde-3-phosphate dehydrogenase mRNAs by semiquantitative RT-PCR, and measurement of hTERC RNA and 7S RNase by Northern blotting. The numbers below the Northern blot are the relative amounts of hTERC/7S RNase by densitometry. B, the levels of hTERT mRNA and 7S RNase were determined by real-time RT-PCR (see “Materials and Methods” for details). The results are expressed as the relative levels of hTERT mRNA to RNase. The figures shown are the mean of three experiments where all of the samples were analyzed in triplicate together with SEs as bars. Culture types are as described in the Fig. 5 legend.

Fig. 6.

Quantitation of hTERC RNA and hTERT mRNA levels in oral cultures. A, measurement of hTERT and glyceraldehyde-3-phosphate dehydrogenase mRNAs by semiquantitative RT-PCR, and measurement of hTERC RNA and 7S RNase by Northern blotting. The numbers below the Northern blot are the relative amounts of hTERC/7S RNase by densitometry. B, the levels of hTERT mRNA and 7S RNase were determined by real-time RT-PCR (see “Materials and Methods” for details). The results are expressed as the relative levels of hTERT mRNA to RNase. The figures shown are the mean of three experiments where all of the samples were analyzed in triplicate together with SEs as bars. Culture types are as described in the Fig. 5 legend.

Close modal
Fig. 7.

Reactivation of RAR-β (A) and p16 (B) by Aza-C. Overnight cultures of ID cultures (D4, D9, D19, D20, D34, and D35) and SCC cultures (B22, B31, and B56) were fed daily with the indicated concentrations of Aza-C for 7 consecutive days, and the levels of RAR-β and p16 were measured as described in the Fig. 4 legend. As positive controls, normal oral mucosa cells, MDs (D25, D30, D6, and D17), or a SK culture were included on some blots.

Fig. 7.

Reactivation of RAR-β (A) and p16 (B) by Aza-C. Overnight cultures of ID cultures (D4, D9, D19, D20, D34, and D35) and SCC cultures (B22, B31, and B56) were fed daily with the indicated concentrations of Aza-C for 7 consecutive days, and the levels of RAR-β and p16 were measured as described in the Fig. 4 legend. As positive controls, normal oral mucosa cells, MDs (D25, D30, D6, and D17), or a SK culture were included on some blots.

Close modal
Fig. 8.

Methylation status of RAR-β and p16 promoters in IDs. A, DNAs from the various ID cultures (D9, D19, D20, D34, and D35) were treated with bisulfite and then analyzed by methylation-specific PCR to determine the presence of methylated (m) and unmethylated (u) sites in the RAR-β and p16 promoters (see “Materials and Methods” for details). As a control, a MD (D25), in which both RAR-β and p16 are expressed, was also included. B, cells were untreated (C) or treated with 2 μm Aza-C for 7 daya (+Aza), and the DNA was extracted and analyzed as described in A.

Fig. 8.

Methylation status of RAR-β and p16 promoters in IDs. A, DNAs from the various ID cultures (D9, D19, D20, D34, and D35) were treated with bisulfite and then analyzed by methylation-specific PCR to determine the presence of methylated (m) and unmethylated (u) sites in the RAR-β and p16 promoters (see “Materials and Methods” for details). As a control, a MD (D25), in which both RAR-β and p16 are expressed, was also included. B, cells were untreated (C) or treated with 2 μm Aza-C for 7 daya (+Aza), and the DNA was extracted and analyzed as described in A.

Close modal
Fig. 9.

Inhibition of telomerase and hTERT mRNA expression in oral cultures by Aza-C treatment. Cells were either left untreated or treated with 2 μm Aza-C for 48 h and then analyzed for (A) telomerase activity by TRAP assays as described in the Fig. 5 legend or (B) hTERT mRNA levels by real-time RT-PCR as described in the Fig. 6 legend. All of the samples in B were analyzed in parallel.

Fig. 9.

Inhibition of telomerase and hTERT mRNA expression in oral cultures by Aza-C treatment. Cells were either left untreated or treated with 2 μm Aza-C for 48 h and then analyzed for (A) telomerase activity by TRAP assays as described in the Fig. 5 legend or (B) hTERT mRNA levels by real-time RT-PCR as described in the Fig. 6 legend. All of the samples in B were analyzed in parallel.

Close modal
Fig. 10.

Induction of senescence marker expression by Aza-C treatment of ID cultures (D34, D20, and D19) or a carcinoma culture (B56). Cells were grown in culture for 14 days with daily refeeding with medium containing 2 μm Aza-C and then stained for SABG expression as described in “Materials and Methods.” SABG staining is shown of three typical colonies after Aza-C treatment, photographed under phase-contrast to identify the location of the colonies. The results are from five experiments using various combinations of dysplasias (D34, five experiments; D20, four experiments; B56, two experiments; D19, two experiments).

Fig. 10.

Induction of senescence marker expression by Aza-C treatment of ID cultures (D34, D20, and D19) or a carcinoma culture (B56). Cells were grown in culture for 14 days with daily refeeding with medium containing 2 μm Aza-C and then stained for SABG expression as described in “Materials and Methods.” SABG staining is shown of three typical colonies after Aza-C treatment, photographed under phase-contrast to identify the location of the colonies. The results are from five experiments using various combinations of dysplasias (D34, five experiments; D20, four experiments; B56, two experiments; D19, two experiments).

Close modal

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

The work of F. M. and J. B. on this project was supported by the Association for International Cancer Research. A. M. is supported by the Glasgow Cancer School and White Lily Trust. P. R. H., J. F., and E. K. P. are supported by the Cancer Research Campaign, and D. H. F. and D. G. M. are supported by the Dental School, Glasgow.

4

The abbreviations used are: SCC, squamous cell carcinoma; Aza-C, 5-aza-2-deoxycytidine; HNSCC, head and neck SCC; hTERT, human telomerase reverse transcriptase; hTERC, human telomerase RNA component; ID, immortal dysplasia; MD, mortal dysplasia; PD, population doubling; PLK1, polo-like kinase 1; RAR-β, retinoic acid receptor-β; SK, smoker’s keratosis; EGFR, epidermal growth factor receptor; ERK, extracellular signal-regulated kinase; SABG, senescence-associated acid β-galactosidase; RT-PCR, reverse transcription-PCR; TRAP, telomeric repeat amplification protocol.

Table 1

Characteristics of dysplasias

TypeDysplasiaSiteDysplasia pathologyMortality phenotype (maximum no. of PDs)
Leukoplakia D41 Posterolateral tongue Mild 3.5 
 D6 Posterior tongue Moderate/severe 25 
 D8 Ventral tongue Mild/moderate 
 D25 Floor of mouth Severe 28 
 D30 Floor of mouth Mild 30 
 D47 Floor of mouth Moderate 20 
 D17 Buccal mucosa Mild/moderate 45 
 D48 Ventral tongue Moderate/severe 25.5 
 D20 Lateral tongue Moderate Immortal 
 D34 Posterolateral tongue Moderate Immortal 
 D9 Ventral tongue Mild/moderate Immortal 
 D4 Ventral tongue/floor of mouth Early SCC Immortal 
 D38 Lateral tongue Mild Immortal 
Erythroleukoplakia D19 Lateral tongue Severe/CISa Immortal 
 D35 Floor of mouth/ventral tongue Severe/CIS Immortal 
SK SK36 Retromolar area Nondysplastic 25.5 
TypeDysplasiaSiteDysplasia pathologyMortality phenotype (maximum no. of PDs)
Leukoplakia D41 Posterolateral tongue Mild 3.5 
 D6 Posterior tongue Moderate/severe 25 
 D8 Ventral tongue Mild/moderate 
 D25 Floor of mouth Severe 28 
 D30 Floor of mouth Mild 30 
 D47 Floor of mouth Moderate 20 
 D17 Buccal mucosa Mild/moderate 45 
 D48 Ventral tongue Moderate/severe 25.5 
 D20 Lateral tongue Moderate Immortal 
 D34 Posterolateral tongue Moderate Immortal 
 D9 Ventral tongue Mild/moderate Immortal 
 D4 Ventral tongue/floor of mouth Early SCC Immortal 
 D38 Lateral tongue Mild Immortal 
Erythroleukoplakia D19 Lateral tongue Severe/CISa Immortal 
 D35 Floor of mouth/ventral tongue Severe/CIS Immortal 
SK SK36 Retromolar area Nondysplastic 25.5 
a

CIS, carcinoma in situ.

We are grateful to Prof. P. Chambon (Strasbourg, France) for the RAR-β2 cDNA and to Robert McFarlane (Beatson Institute) for assistance with the real-time RT-PCR and p53 sequencing experiments. We also thank Prof. J. Wyke and Prof. B. Ozanne for critical reading of the manuscript.

1
Cowley G., Smith J. A., Gusterson B. A., Hendler F. J., Ozanne B. The amount of EGF receptor is elevated on squamous cell carcinomas.
Cancer Cells
,
1
:
5
-10,  
1984
.
2
Grandis J. R., Tweardy D. J. Elevated level of transforming growth factor α and epidermal growth factor receptor messenger RNA are early markers of carcinogenesis in the head and neck.
Cancer Res.
,
53
:
3579
-3584,  
1993
.
3
Callender T., El-Naggar A. K., Lee M. S., Frankenthaler R., Luna M. A., Batsakis J. G. PRAD-1 (CCND1)/cyclin D1 oncogene amplification in primary head and neck squamous cell carcinoma.
Cancer (Phila.)
,
74
:
152
-158,  
1994
.
4
Liu S-C., Sauter E. R., Clapper M. L., Feldman R. S., Levin L., Chen S. Y., Yen T. J., Ross E., Engstom P. F., Klein-Szanto A. J. P. Markers of cell proliferation in normal epithelia and dysplastic leukoplakias of the oral cavity.
Cancer Epidemiol. Biomark. Prev.
,
7
:
597
-603,  
1998
.
5
Knecht R., Elez R., Oechler M., Solbach C., von Ilberg C., Strebhardt K. Prognostic significance of polo-like kinase (PLK) in squamous cell carcinomas of the head and neck.
Cancer Res.
,
59
:
2794
-2797,  
1999
.
6
van Oijen M. G., Tilanus M. G., Medema R. H., Slootweg P. J. Expression of p21Waf1/Cip1 in head and neck cancer in relation to proliferation, differentiation, p53 status and cyclin D1 expression.
J. Oral Pathol. Med.
,
27
:
367
-375,  
1998
.
7
Erber R., Klein W., Andl T., Enders C., Born A. I., Conradt C., Bartek J., Bosch F. X. Aberrant p21CIP1/WAF1 protein accumulation in head and neck cancer.
Int. J. Cancer
,
74
:
383
-389,  
1997
.
8
Mineta H., Miura K., Suzuki I., Takebayashi S., Amano H., Araki K., Harada H., Ichimura K., Wennerberg J. P., Dictor M. R. Low p27 expression correlates with poor prognosis for patients with oral tongue squamous cell carcinoma.
Cancer (Phila.)
,
85
:
1011
-1017,  
1999
.
9
Nylander K., Dabelsteen E., Hall P. A. The p53 molecule and its prognostic role in squamous cell carcinomas of the head and neck.
J. Oral Pathol. Med.
,
29
:
413
-425,  
2000
.
10
Loughran O., Malliri A., Owens D., Gallimore P. H., Stanley M. A., Ozanne B., Frame M. C., Parkinson E. K. Association of CDKN2A/p16INK4A with human head and neck keratinocyte replicative senescence: relationship of dysfunction to immortality and neoplasia.
Oncogene
,
13
:
561
-568,  
1996
.
11
Munro J., Scott F. J., Vousden K. H., Peters G., Parkinson E. K. Role of the alternative INK4A proteins in human keratinocyte senescence: evidence for the specific inactivation of p16INK4A upon immortalization.
Cancer Res.
,
59
:
2516
-2521,  
1999
.
12
Reed A. L., Califano J., Cairns P., Westra W. H., Jones R. M., Koch W., Ahrendt S., Eby Y., Sewell D., Nawroz H., Bartek J., Sidransky D. High frequency of p16 (CDKN2/MTS-1/INK4A) inactivation in head and neck squamous cell carcinoma.
Cancer Res.
,
56
:
3630
-3633,  
1996
.
13
Papadimitrakopoulou V., Izzo J., Lippman S. M., Lee J. S., Fan Y. H., Clayman G., Ro J. Y., Hittleman W. N., Lotan R., Hong W. K., Mao L. Frequent inactivation of p16INK4a in oral premalignant lesions.
Oncogene
,
14
:
1799
-1803,  
1997
.
14
Gasparotto D., Maestro R., Piccinin S., Vukosaljevic T., Barzan L., Sulfaro S., Boiocchi M. Overexpression of CDC25A and CDC25B in head and neck cancers.
Cancer Res.
,
57
:
2366
-2368,  
1997
.
15
Rosin M. P., Cheng X., Poh C., Lam W. L., Huang Y., Lovas J., Berean K., Epstein J. B., Priddy R., Le N. D., Zhang L. Use of allelic loss to predict malignant risk for low-grade oral epithelial dysplasia.
Clin. Cancer Res.
,
6
:
357
-362,  
2000
.
16
Campisi J. Cancer, aging and cellular senescence.
In Vivo
,
14
:
183
-188,  
2000
.
17
Huschtscha L. I., Reddel R. R. p16INK4a and the control of cellular proliferative life span.
Carcinogenesis (Lond.)
,
20
:
921
-926,  
1999
.
18
Colgin L. M., Reddel R. R. Telomere maintenance mechanisms and cellular immortalization.
Curr. Opin. Genet. Dev.
,
9
:
97
-103,  
1999
.
19
Jiang X. R., Jimenez G., Chang E., Frolkis M., Kusler B., Sage M., Beeche M., Bodnar A. G., Wahl G. M., Tlsty T. D., Chiu C. P. Telomerase expression in human somatic cells does not induce changes associated with a transformed phenotype.
Nat. Genet.
,
21
:
111
-114,  
1999
.
20
Dickson M. A., Hahn W. C., Ino Y., Ronfard V., Wu J. Y., Weinberg R. A., Louis D. N., Li F. P., Rheinwald J. G. Human keratinocytes that express hTERT and also bypass a p16INK4a-enforced mechanism that limits life span become immortal yet retain normal growth and differentiation characteristics.
Mol. Cell. Biol.
,
20
:
1436
-1447,  
2000
.
21
Kiyono T., Foster S. A., Koop J. I., McDougall J. K., Galloway D. A., Klingelhutz A. J. Both Rb/p16INK4a inactivation and telomerase activity are required to immortalise human epithelial cells.
Nature (Lond.)
,
396
:
84
-88,  
1998
.
22
Farwell D. G., Shera K. A., Koop J. I., Bonnet G. A., Matthews C. P., Reuther G. W., Coltrera M. D., McDougall J. K., Klingelhutz A. J. Genetic and epigenetic changes in human epithelial cells immortalized by telomerase.
Am. J. Pathol.
,
156
:
1537
-1547,  
2000
.
23
Hu L., Crowe D. L., Rheinwald J. G., Chambon P., Gudas L. J. Abnormal expression of retinoic acid receptors and keratin 19 by human oral and epidermal squamous cell carcinoma cell lines.
Cancer Res.
,
51
:
3972
-3981,  
1991
.
24
Xu X-C., Ro J. Y., Lee J. S., Shin D. M., Hong W. K., Lotan R. Differential expression of nuclear retinoid receptors in normal, premalignant, and malignant head and neck tissues.
Cancer Res.
,
54
:
3580
-3587,  
1994
.
25
McGregor F., Wagner E., Felix D., Soutar D., Parkinson K., Harrison P. R. Inappropriate retinoic acid receptor-β expression in oral dysplasias: correlation with acquisition of the immortal phenotype.
Cancer Res.
,
57
:
3886
-3889,  
1997
.
26
Dimri G. P., Lee X., Basile G., Acosta M., Scott G., Roskelley C., Medrano E. E., Linskens M., Rubelj I., Pereira-Smith O., Peacocke M., Campisi J. A biomarker that identifies senescent human cells in culture and in aging skin in vivo.
Proc. Natl. Acad. Sci. USA
,
92
:
9363
-9367,  
1995
.
27
Malliri A., Symons M., Hennigan R. F., Hurlstone A. F. L., Lamb R. F., Wheeler T., Ozanne B. W. The transcription factor AP-1 is required for EGF-induced activation of Rho-like GTPases, cytoskeletal rearrangements, motility and in vitro invasion of A431 cells.
J. Cell Biol.
,
143
:
1087
-1099,  
1998
.
28
Virmani A. K., Rathi A., Zochbauer-Muller S., Sacchi N., Fukuyama Y., Bryant D., Maitra A., Heda S., Fong K. M., Thunnissen F., Minna J. D., Gazdar A. F. Promoter methylation and silencing of the retinoic acid receptor β gene in lung carcinomas.
J. Natl. Cancer Inst. (Bethesda)
,
92
:
1303
-1307,  
2000
.
29
Baur A. S., Shaw P., Burri N., Delacretaz F., Bosman F. T., Chaubert P. Frequent methylation silencing of p15INK4b (MST2) and p16INK4a (MST1) in B-cell and T-cell lymphomas.
Blood
,
94
:
1773
-1781,  
1999
.
30
Rheinwald J. G., Beckett M. A. Tumorigenic keratinocyte lines requiring anchorage and fibroblast support cultures from squamous cell carcinomas.
Cancer Res.
,
41
:
1657
-1663,  
1981
.
31
Ramirez R. D., Morales C. P., Herbert B-S., Rohde J. M., Passons C., Shay J. W., Wright W. E. Putative telomere-independent mechanisms of replicative aging reflect inadequate growth conditions.
Genes Dev.
,
15
:
398
-403,  
2001
.
32
Burns J. E., Clark L. J., Yeudall W. A., Mitchell R., MacKenzie K., Chang S. E., Parkinson E. K. The p53 status of cultured human premalignant oral keratinocytes.
Br. J. Cancer
,
70
:
591
-595,  
1994
.
33
Burns J. E., Baird M. C., Clark L. J., Burns P. A., Edington K., Chapman C., Mitchell R., Robertson G., Soutar D., Parkinson E. K. Gene mutations and increased levels of p53 protein in human squamous cell carcinomas and their cell lines.
Br. J. Cancer
,
67
:
1274
-1284,  
1993
.
34
Edington K., Loughran O. P., Berry I. J., Parkinson E. K. Cellular immortality: a late event in the progression of human squamous cell carcinoma of the head and neck associated with p53 alteration and a high frequency of allele loss.
Mol. Carcinog.
,
13
:
254
-265,  
1995
.
35
Zelent A., Mendelsohn C., Kastner P., Krust A., Garnier J. M., Ruffenach F., Leroy P., Chambon P. Differentially expressed isoforms of the mouse retinoic acid receptor β generated by usage of two promoters and alternative splicing.
EMBO J.
,
10
:
71
-81,  
1991
.
36
Greenberg R. A., O’Hagan R. C., Deng H., Xiao Q., Hann S. R., Adams R. R., Lichtsteiner S., Chin L., Morin G. B., DePinho R. A. Telomerase reverse transcriptase gene is a direct target of c-Myc but is not functionally equivalent in cellular transformation.
Oncogene
,
18
:
1219
-1226,  
1999
.
37
Oh S., Song Y-H., Yim J., Kim T. K. Identification of Mad as a repressor of human telomerase (hTERT) gene.
Oncogene
,
19
:
1485
-1490,  
2000
.
38
Xu D., Wang Q., Gruber A., Bjorkholm M., Chen Z., Zaid A., Selivanova G., Peterson C., Wiman K. G., Pisa P. Downregulation of telomerase reverse transcriptase mRNA expression by wild type p53 in human tumor cells.
Oncogene
,
19
:
5123
-5133,  
2000
.
39
Swisshelm K., Ryan K., Lee X., Tsou H. C., Peacocke M., Sager R. Down-regulation of retinoic acid receptor β in mammary carcinoma cell lines and its up-regulation in senescing normal mammary epithelial cells.
Cell Growth Differ.
,
5
:
133
-141,  
1994
.
40
Berard J., Laboune F., Mukuna M., Masse S., Kothary R., Bradley W. E. C. Lung tumors in mice expressing an antisense RARβ2 transgene.
FASEB J.
,
10
:
1091
-1097,  
1996
.
41
Berard J., Gaboury L., Landers M., De Repentingy Y., Houle B., Kothary R., Bradley W. E. Hyperplasia and tumours in lung, breast and other tissues in mice carrying a RARβ4-like transgene.
EMBO J.
,
13
:
5570
-5580,  
1994
.
42
Toulouse A., Morin J., Dion P. A., Houle B., Bradley W. E. RARβ2 specificity in mediating RA inhibition of growth of lung cancer-derived cells.
Lung Cancer
,
28
:
127
-137,  
2000
.
43
Houle B., Rochette-Egly C., Bradley W. E. Tumor-suppressive effect of the retinoic acid receptor β in human epidermoid lung cancer cells.
Proc. Natl. Acad. Sci. USA
,
90
:
985
-989,  
1993
.
44
Artandi S. E., Chang S., Lee S. L., Alson S., Gottlieb G. J., Chin L., DePinho R. A. Telomere dysfunction promotes non-reciprocal translocations and epithelial cancers in mice.
Nature (Lond.)
,
406
:
573
-574,  
2000
.
45
Kashiwazaki I., Tonoki H., Tada M., Chiba I., Shindoh M., Totsuka Y., Iggo R., Moriuchi T. High frequency of p53 mutations in human oral epithelial dysplasia and primary squamous cell carcinoma detected by yeast functional assay.
Oncogene
,
15
:
2667
-2674,  
1997
.
46
Klingelhutz A. J., Foster S. A., McGougall J. K. Telomerase activation by the E6 gene product of human papillomavirus type 16.
Nature (Lond.)
,
380
:
79
-82,  
1996
.
47
Grandis J. R., Tweardy D. J., Melham M. F. Asynchronous modulation of transforming growth factor α and epidermal growth factor protein expression in progression of premalignant lesions to head and neck squamous cell carcinoma.
Clin. Cancer Res.
,
4
:
13
-20,  
1998
.
48
Stanton P., Richard S., Reeves J., Nikolic M., Edington K., Clark L., Robertson G., Souter D., Mitchell R., Hendler F. J., Cooke T., Parkinson E. K., Ozanne B. W. Epidermal growth factor receptor expression by human squamous cell carcinomas of the head and neck, cell lines and xenografts.
Br. J. Cancer
,
70
:
427
-433,  
1994
.
49
Grandis J. R., Zeng Q., Drenning S. D., Tweardy D. J. Normalisation of EGFR mRNA levels following restoration of wild-type p53 in a head and neck squamous cell carcinoma cell line.
Int. J. Oncol.
,
13
:
375
-378,  
1998
.
50
Grandis J. R., Zeng Q., Tweardy D. J. Retinoic acid normalises the increased gene transcription rate of TGF-α and EGFR in head and neck cancer cell lines.
Nat. Med.
,
2
:
237
-240,  
1996
.
51
Shin D. M., Lee J. S., Lippman S. M., Lee J. J., Tu Z. N., Choi T. G., Heyne K., Shin H. J. C., Ro J. Y., Goepfert H., Hong W. K., Hittleman W. N. p53 expression: predicting recurrence and second primary tumors in head and neck squamous cell carcinoma.
J. Natl. Cancer Inst. (Bethesda)
,
88
:
519
-529,  
1996
.
52
Obata A., Eura M., Sasaki J., Saya H., Chikamatsu K., Tada M., Iggo R. D., Yumoto E. Clinical significance of p53 functional loss in squamous cell carcinoma of the oropharynx.
Int. J. Cancer Predict. Oncol.
,
89
:
187
-193,  
2000
.
53
Gallo O., Santucci M., Franchi A. Cumulative prognostic value of p16/CDKN2 and p53 oncoprotein expression in premalignant laryngeal lesions.
J. Natl. Cancer Inst. (Bethesda)
,
89
:
1161
-1163,  
1997
.
54
Lee J. J., Hong W. K., Hittleman W. N., Mao L., Lotan R., Shin D. M., Benner S. E., Xu X-C., Lee J. S., Papadimtrakopoulou V. M., Geyer C., Perez C., Martin J. W., El-Naggar A. K., Lippmann S. M. Predicting cancer development in oral leukoplakia: ten years of translational research.
Clin. Cancer Res.
,
6
:
1702
-1710,  
2000
.
55
Khuri F. R., Lotan R., Kemp B. L., Lippman S. M., Wu H., Feng L., Lee J. J., Cooksley C. S., Parr B., Chang E., Walsh G. L., Lee J. S., Hong W. K., Xu X-C. Retinoic acid receptor-β as a prognostic indicator in stage I non-small-cell lung cancer.
J. Clin. Oncol.
,
18
:
2798
-2804,  
2000
.
56
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. (Bethesda)
,
90
:
824
-832,  
1998
.
57
Wolf G., Elez R., Doermer A., Holtrich U., Ackermann H., Stutte H. J., Altsmannsberger H. M., Rubsamen-Waigmann H., Strebhardt K. Prognostic significance of polo-like (PLK) expression in non-small-cell lung cancer.
Oncogene
,
14
:
543
-549,  
1997
.
58
Lonardo F., Rusch V., Langenfeld E., Klimstra D. S. Overexpression of cyclins D1 and E is frequent in bronchial preneoplasia and preceded squamous cell carcinoma.
Cancer Res.
,
59
:
2470
-2476,  
1999
.
59
Bova R. J., Quinn D. I., Nankervis J. S., Cole I. E., Sheridan B. F., Jensen M. J., Morgan G. J., Hughes C. J., Sutherland R. L. Cyclin D1 and p16INK4A expression predict reduced survival in carcinoma of the anterior tongue.
Clin. Cancer Res.
,
5
:
2810
-2819,  
1999
.
60
Catzavelos C., Tsao M. S., DeBoer G., Bhattacharya N., Shepherd F. A., Slingerland J. M. Reduced expression of the cell cycle inhibitor p27Kip1 in non-small-cell lung carcinoma: a prognostic factor independent of Ras.
Cancer Res.
,
59
:
684
-688,  
1999
.
61
Kitagawa Y., Kyo S., Takakura M., Namiki M., Inoue M. Demethylating reagent 5-azacytidine inhibits telomerase activity in human prostate cancer cells through transcriptional repression of hTERT.
Clin. Cancer Res.
,
6
:
2868
-2875,  
2000
.