Head-and-neck squamous-cell cancer (HNSCC) is an important public-health problem, accounting for approximately 40,300 new cancer cases and 12,000 cancer deaths annually in the United States (Greenlee et al., 2000). Patients with early-stage disease are often cured with surgery or radiotherapy but are at high risk for second primary tumor (SPT) development (Lippman and Hong, 1989), and the majority of patients present with advanced disease, for which the outcomes have not markedly improved despite advances in combined-modality therapy (Vokes et al., 1993). HNSCC arises from transformation of the genetic material of normal cells, followed by successive genetic alterations in a multistep fashion, leading to clonal evolution of progeny cells with a proliferative advantage (Vogelstein and Kinzler, 1993), induced by tobacco carcinogens (Slaughter et al., 1953). Chemoprevention aims at reversal of this process through re-regulation of growth and differentiation and possibly elimination of genetically and phenotypically aberrant clones. Chemoprevention studies in upper aerodigestive tract (UADT) cancers are based on these fundamental premises and the identification of molecular genetic and biologic cellular changes. These alterations represent biomarkers of the carcinogenesis process and ultimately, if validated, could serve as intermediate end points for these studies. Although smoking is the major risk factor for UADT cancers, inherent host susceptibility factors also play an important role. The presence of mutagen sensitivity, an in vitro chromosomal marker of sensitivity to genotoxicity, can be used as an independent risk factor for HNSCC (Spitz et al., 1989) and is a substantial predictor of risk of multiple primary cancers (Spitz et al., 1994). Consistent chromosomal gains and losses have been identified through cytogenetics at 3q21, 5p, 7p, 8q, 11q13-23, 3p13-q24, 5q12-q23, 8p22-p23, 9p21-p24 and 18q22-23 (Jin et al., 1995). Allelic losses at marker loci in specific genomic regions have also been identified at 3p, 5q, 9p, 11q, 13q,18q and 17p (Nawroz et al., 1994). Among the several candidate tumor-suppressor genes at 3p, FHIT is altered in 80% of HNSCC cell lines (Mao et al., 1996a). Loss of heterozygosity (LOH) at 3p14 and 9p21 has been observed in oral pre-malignant lesions (OPLs) with high frequency and is significantly associated with subsequent development of HNSCC (Mao et al., 1996b). Histologic progression to invasive cancer correlates with allelic loss at increasing numbers of critical loci (Califano et al., 1996). The degree of chromosomal polysomy (i.e., 3 or more copies), detected by in situ hybridization, is associated with histologic progression from normal epithelium to dysplastic mucosa (Voravud et al., 1993) and with risk of development of oral cancer, indicating that polysomy could be a useful marker of generalized genomic instability in tissues at risk (Lee et al., 1993). The p16MTS1/INK4a gene, an inhibitor of the cyclin D–cyclin-dependent kinase complex, is the most likely tumor-suppressor gene at the 9p21-p24 region (Serrano et al., 1993). The gene is frequently inactivated in HNSCC through point mutation, homozygous deletion or promoter methylation (Cairns et al., 1995; Merlo et al., 1995; Reed et al., 1996), and loss of p16 protein expression occurs early in OPL (Papadimitrakoulou et al., 1997). Cyclin D1 plays a central role in G1/S transition and gene amplification, and over-expression has been frequently described in HNSCC and pre-malignancy (Motokura et al., 1991; Izzo et al., 1998). Retinoid-dependent cyclin D1 proteolysis, described in normal and neoplastic human bronchial epithelial cells, may have potential as an intermediate marker for effective retinoid chemoprevention (Boyle et al., 1999). A combination of persistent dysregulation of cyclin D1 and p16 loss in lesions of patients undergoing chemoprevention appears to predict for adverse outcome in terms of histologic progression and malignant transformation (Papadimitrakopoulou et al., 2000). Alterations of the tumor-suppressor gene p53 disrupt its role in cell-cycle checkpoint control and apoptosis, leading to genomic instability, and occur frequently in HNSCC. Alterations of p53 in pre-malignant lesions contiguous with head-and-neck tumors occur with increasing frequency along with histologic progression (Boyle et al., 1993; Shin et al., 1994a). p53 over-expression in initial head-and-neck tumors predicts for increased incidence of SPT and recurrence of the primary tumor (Shin et al., 1996). The epidermal growth factor receptor (EGFR) binds and mediates signal transduction for both EGF and transforming growth factor-α (TGF-α). Increased EGFR levels have been documented frequently in HNSCC and in histologically normal epithelium of HNSCC-bearing patients (Grandis and Tweardy, 1993), while in pre-malignant lesions, a stepwise increase of EGFR expression has been observed following histologic progression (Shin et al., 1994b) and appears to be a sensitive predictor of neoplastic potential in dysplastic tissues (Uhlman et al., 1996). Modulation of EGFR gene transcription by retinoic acid in head-and-neck cancer cells strengthens its potential use as an intermediate end point in chemoprevention trials (Grandis et al., 1996) Vitamin A and its analogs, the retinoids, are established modulators of epithelial differentiation and have been widely used in pre-clinical and clinical chemoprevention studies. They regulate gene expression through nuclear retinoid receptors. Expression of RAR-β is suppressed in early stages of head-and-neck carcinogenesis in OPLs in patients without cancer and in dysplastic tissue adjacent to HNSCC (Xu et al., 1994), and this loss is thought to be associated with HNSCC development. Its expression is up-regulated in vivo by 13-cis retinoic acid (13-cRA), which correlates well with the clinical response (Lotan et al., 1995). There is compelling evidence that aberrant DNA methylation at CpG islands in the promoter region can silence certain critical genes contributing to tumorigenesis. In one study, promoter hypermethylation of 4 genes (including p16) was seen in 50% of HNSCC cases and was predictive for presence of lymph node involvement and development of distant metastasis (Sanchez-Cespedes et al., 2000). Aberrant p16 promoter methylation was observed in precursor lesions of squamous-cell lung carcinomas with a frequency increasing with disease progression from hyperplasia to cancer (Belinsky et al., 1998), a finding reciprocated in head-and-neck pre-malignancy (Papadimitrakopoulou et al., 2000). The role of retinoids as chemopreventive agents has been evaluated and their efficacy demonstrated in a few landmark trials summarized below. 13-cRA has been shown to produce a significant delay in the development of SPT when used after therapy of a primary HNSCC (Hong et al., 1990). The oral pre-malignancy chemoprevention trials conducted at the M.D. Anderson Cancer Center have demonstrated that treatment with high-dose 13-cRA results in major regression of leukoplakia in 67% of patients (Hong et al., 1986) and that low-dose 13-cRA is significantly more effective than β-carotene at maintaining clinical/histologic remission (Lippman et al., 1993). Retinoid resistance was associated with abnormal p53 expression, increased degree of genomic instability and lack of RAR-β up-regulation following treatment with 13-cRA (Lippman et al., 1995). The biochemoprevention study was developed to address the issue of retinoid resistance in advanced pre-malignant lesions of the UADT using a combination of 13-cRA with α-tocopherol and IFN-α. A striking difference in response was observed in favor of laryngeal lesions (Papadimitrakopoulou et al., 1999), while biomarker analysis confirmed the negative prognostic value of p53 over-expression and revealed persistent 9p21 loss in the majority of lesions despite response (Mao et al., 1998). However, even such aggressive chemopreventive intervention is not effective at reversing advanced pre-malignant lesions of the oral cavity and oropharynx and is associated with substantial toxicity, suggesting an urgent need for innovative approaches and targeted chemoprevention. Among these newer approaches is the development of cyclo-oxygenase-2 (COX-2) inhibitors and the ONYX-015 virus, targeting selectively p53-deficient cells. None of the biomarkers discussed has been validated for use as an intermediate end point in chemoprevention trials. Furthermore, it is unclear whether the presence of certain biomarkers marks distinct stages in carcinogenesis, though their occurrence has been “mapped” to histologic stages of pre-neoplastic progression (Fig. 1). Therefore, the activity of chemopreventive agents that target specific abnormalities (biomarkers) can accurately be assessed only on their specific effect on the biomarker in question. Nevertheless, it is possible that such agents affect pathways downstream or parallel to that derived from the specific biomarker, as for agents affecting cell-cycle distribution, signal transduction or the retinoids. With the current level of knowledge, we cannot use any specific biomarker as an intermediate end point because we would be running the risk of evaluating the activity of an agent based on the wrong biomarker occurring either at an earlier or later stage than the one the agent is affecting. Another issue that has not been clarified is how accurate a predictor any specific biomarker is for subsequent cancer development. Although multiple studies have been performed to address this issue with individual biomarkers, it appears that a model incorporating multiple biomarkers should be developed to facilitate this task as progression to cancer is multifactorial. The answers to these questions can come only from detailed and prospective studies that incorporate systematic analyses of biomarkers in correlation with response to intervention and development of invasive cancer. A histologic progression model for UADT pre-malignancy with molecular abnormalities occurring at each step of the process (potential biomarkers) and a proposed intervention schema according to increasing risk of cancer development. Finally, continued study of the molecular pathogenesis of head-and-neck cancer is necessary for the rational development of sound chemopreventive strategies that will impact on the incidence of head-and-neck cancer.
No takes yet. Share an insight, caveat, or question.
Papadimitrakopoulou et al. (2000) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: