Causative risk factors

Tobacco use and alcohol consumption are well-established risk factors for HNSCC. The larynx has been suggested to be the organ most susceptible to carcinogenic insult from tobacco smoke in the head and neck region . The enzymes that metabolize the carcinogenic compounds in cigarettes and alcohol, such as glutathione S-transferase, N-acetyltransferase, cytochrome p450, alcohol dehydrogenase, and aldehyde dehydrogenase, have been studied to assess an individual’s risk for cancer development, and although considerable polymorphisms of these enzymes have been found among the population, the results are controversial . Although tobacco exposure and alcohol consumption are the etiologic causes for approximately 75% to 80% of HNSCC, infection with the human papillomavirus (HPV) has been etiologically linked with the remaining 20% to 25% of all head and neck cancers . Upon infection of the cell, the HPV E6 and E7 viral proteins bind and disrupt two important cellular gatekeeper proteins, p53 and pRb , respectively . In laryngeal cancer, the prevalence of HPV DNA has been reported to range between 3% and 47% . The wide variation in prevalence is most likely attributable to the various detection techniques used. In addition to detection of the “high-risk” HPV types, such as HPV-16 and HPV-18, which are associated with oropharyngeal cancers, there are several studies that have reported the integration of the HPV-11 genome in patients who have recurrent respiratory papillomatosis and who progressed to develop carcinoma . However, further studies are necessary before a clear association between HPV and laryngeal cancer can be established.

Genetic progression model

In 1971, Knudson proposed that loss of both parental alleles of a tumor suppressor gene (TSG) can lead to a malignant phenotype . In 1990, Fearon and Vogelstein expanded this concept and proposed that either loss of TSGs and/or activation of oncogenes can lead to cellular transformation, and it is the progressive accumulation of such genetic alterations which can lead to a selective growth advantage of a clonal population of transformed cells which can then develop into cancer. Statistical analysis based on the age-specific incidence of head and neck cancer suggests that HNSCCs arise after the accumulation of 6-10 independent genetic events. Taking these concepts, a progression model can be established by tracking the genetic changes acquired through time within cloneal populations and then correlating these genetic changes with histologic progression from hyperplasia, to dysplasia, to carcinoma in situ, and, finally, to invasive carcinoma. The techniques that can be used to track the molecular alterations include cytogenetics, loss of heterozygosity (LOH), mutation analysis, and, more recently, gene expression profiles.

Cytogenetics is the analysis of chromosomes for structural changes, such as translocations, deletions, or amplifications. The classic karyotype and fluorescence in situ hybridization (FISH) assays are commonly used cytogenetic techniques. The former uses trypsin-Giemsa staining to compare the banding of the chromosomes, and the latter localizes particular regions of the chromosome with DNA probes. Since the size of the alterations needs to be large enough to be visualized on a microscopic scale, these tests are considered to have low sensitivity.

A more sensitive test is the detection of specific allelic loss or LOH. The assay consists of a polymerase chain reaction (PCR) that amplifies microsatellite markers, which are short repeat polymorphic regions distributed throughout the noncoding genome. The loss of microsatellite markers in the assay identifies areas of neighboring alleles harboring potential TSGs, and the implication of the functional deficit of the protein is the development of carcinogenesis. This technique evaluates gene expression or deletion within a localized region of the chromosome. With the advent of microarray technology, whole genome gene expression profiling is now possible. Comparison of gene expression patterns between normal and cancerous cells enables the characterization of gene expression patterns that can be correlated with functionally important genes or states. For example, using microarray technology, tumor classifications, treatment responses, and clinical outcome predictions may become a reality in many cancers. This area of research is actively evolving with the generation of novel assays that provide reliable, sensitive, high-throughput assays with accurate and efficient read-out systems of thousands of genes.

In 1996, Califano and colleagues correlated histologic progression from hyperplasia, to dysplasia, to carcinoma in situ, and to invasive carcinoma by tracking genetic changes acquired through time using LOH and FISH analysis and proposed a genetic progression model for head and neck cancer. The group found that 30% of head and neck precursor cancer lesions, such as hyperplasia, showed loss of chromosome 9p21, indicating that loss in this region was an early event. This locus was found to encode p16 (CDKN2A/MTS1), an important TSG that inhibits cell cycle progression from the G1 checkpoint to the S phase by inhibiting the phosphorylation of p Rb. Dyplasia demonstrated genetic alterations with loss of 3p21 and 17p13. The latter locus contains p53 , which is a major TSG that plays an important regulatory role in DNA repair, cell cycle progression, and apoptosis. Carcinoma in situ had loss of 13q21 and 14q32 and amplification of 11q3. p Rb is contained in the 13q21 region and controls cell cycle progression Cyclin D (PRAD1) is localized to the 11q3 locus and overexpression of this gene results in phosphorylation, and thus activation of p Rb. Invasive carcinomas were found to have LOH on 6p, 8p, 4p27, and 10q12 . Subsequent multiple studies have confirmed chromosomal losses on 3p, 5q, 8p, 9p, 13p, 18q, and 21q in head and neck cancers .

The establishment of a genetic progression model has several advantages. The use of allelic loss may facilitate the prediction of the malignant potential of low-grade premalignant lesions. It has been found that patients with 3p and 9p loss have a 3.8-fold risk for progression to cancer. Additional losses, such as 4q, 8p, 11q, or 17p, increase the risk to 33-fold, however . Therefore, patients with LOH at 3p or 9p are at risk for progression, and their relative risk increases with loss of additional arms. This information can be used to tailor treatment to the individual and his or her lesion. For example, patients with LOH at 3p or 9p with additional loss on other chromosomal arms may need more aggressive treatment of their lesions, whereas, patients with only 3p or 9p LOH may benefit from close monitoring for further alterations.

A genetic progression model can also guide the development of novel treatment strategies through gene therapy and the replacement of altered genes with their wild-type phenotype. Laryngeal cancer has a particular pattern of p53 mutations as compared with other regions of the head and neck. The laryngeal pattern is more consistent with that of lung cancer, and it shows lower mutation indexes (35.4% versus 60%) with different concentrated regions of p53 mutations. In most head and neck cancers, exons 5 through 8 are altered, whereas in laryngeal cancer, the mutations tend to cluster in exon 5, especially in the S2 protein domain and in codon 248 . A therapeutic strategy that is currently in development is the delivery of wild-type p53 into dysplastic or cancerous cells to restore wild-type p53 function within the cell. This approach has been explored using a recombinant adenovirus that delivers wild-type p53 (Ad5CMV-p53) in a laryngeal cancer model. In vivo studies demonstrated that intratumoral injections of Ad5CMV-p53 significantly inhibited the further growth of established laryngeal tumor xenografts . Furthermore, studies have demonstrated that the cotransfer of wild-type p53 with immunomodulatory genes, such as granulocyte-macrophage colony-stimulating factor (GM-CSF) and B7-1 genes using a recombinant adenovirus, induced cellular apoptosis and enhanced the immunogenicity of cancer cells, suggesting a potential combinatorial gene therapy strategy for laryngeal cancer .

Genetic progression model

In 1971, Knudson proposed that loss of both parental alleles of a tumor suppressor gene (TSG) can lead to a malignant phenotype . In 1990, Fearon and Vogelstein expanded this concept and proposed that either loss of TSGs and/or activation of oncogenes can lead to cellular transformation, and it is the progressive accumulation of such genetic alterations which can lead to a selective growth advantage of a clonal population of transformed cells which can then develop into cancer. Statistical analysis based on the age-specific incidence of head and neck cancer suggests that HNSCCs arise after the accumulation of 6-10 independent genetic events. Taking these concepts, a progression model can be established by tracking the genetic changes acquired through time within cloneal populations and then correlating these genetic changes with histologic progression from hyperplasia, to dysplasia, to carcinoma in situ, and, finally, to invasive carcinoma. The techniques that can be used to track the molecular alterations include cytogenetics, loss of heterozygosity (LOH), mutation analysis, and, more recently, gene expression profiles.

Cytogenetics is the analysis of chromosomes for structural changes, such as translocations, deletions, or amplifications. The classic karyotype and fluorescence in situ hybridization (FISH) assays are commonly used cytogenetic techniques. The former uses trypsin-Giemsa staining to compare the banding of the chromosomes, and the latter localizes particular regions of the chromosome with DNA probes. Since the size of the alterations needs to be large enough to be visualized on a microscopic scale, these tests are considered to have low sensitivity.

A more sensitive test is the detection of specific allelic loss or LOH. The assay consists of a polymerase chain reaction (PCR) that amplifies microsatellite markers, which are short repeat polymorphic regions distributed throughout the noncoding genome. The loss of microsatellite markers in the assay identifies areas of neighboring alleles harboring potential TSGs, and the implication of the functional deficit of the protein is the development of carcinogenesis. This technique evaluates gene expression or deletion within a localized region of the chromosome. With the advent of microarray technology, whole genome gene expression profiling is now possible. Comparison of gene expression patterns between normal and cancerous cells enables the characterization of gene expression patterns that can be correlated with functionally important genes or states. For example, using microarray technology, tumor classifications, treatment responses, and clinical outcome predictions may become a reality in many cancers. This area of research is actively evolving with the generation of novel assays that provide reliable, sensitive, high-throughput assays with accurate and efficient read-out systems of thousands of genes.

In 1996, Califano and colleagues correlated histologic progression from hyperplasia, to dysplasia, to carcinoma in situ, and to invasive carcinoma by tracking genetic changes acquired through time using LOH and FISH analysis and proposed a genetic progression model for head and neck cancer. The group found that 30% of head and neck precursor cancer lesions, such as hyperplasia, showed loss of chromosome 9p21, indicating that loss in this region was an early event. This locus was found to encode p16 (CDKN2A/MTS1), an important TSG that inhibits cell cycle progression from the G1 checkpoint to the S phase by inhibiting the phosphorylation of p Rb. Dyplasia demonstrated genetic alterations with loss of 3p21 and 17p13. The latter locus contains p53 , which is a major TSG that plays an important regulatory role in DNA repair, cell cycle progression, and apoptosis. Carcinoma in situ had loss of 13q21 and 14q32 and amplification of 11q3. p Rb is contained in the 13q21 region and controls cell cycle progression Cyclin D (PRAD1) is localized to the 11q3 locus and overexpression of this gene results in phosphorylation, and thus activation of p Rb. Invasive carcinomas were found to have LOH on 6p, 8p, 4p27, and 10q12 . Subsequent multiple studies have confirmed chromosomal losses on 3p, 5q, 8p, 9p, 13p, 18q, and 21q in head and neck cancers .

The establishment of a genetic progression model has several advantages. The use of allelic loss may facilitate the prediction of the malignant potential of low-grade premalignant lesions. It has been found that patients with 3p and 9p loss have a 3.8-fold risk for progression to cancer. Additional losses, such as 4q, 8p, 11q, or 17p, increase the risk to 33-fold, however . Therefore, patients with LOH at 3p or 9p are at risk for progression, and their relative risk increases with loss of additional arms. This information can be used to tailor treatment to the individual and his or her lesion. For example, patients with LOH at 3p or 9p with additional loss on other chromosomal arms may need more aggressive treatment of their lesions, whereas, patients with only 3p or 9p LOH may benefit from close monitoring for further alterations.

A genetic progression model can also guide the development of novel treatment strategies through gene therapy and the replacement of altered genes with their wild-type phenotype. Laryngeal cancer has a particular pattern of p53 mutations as compared with other regions of the head and neck. The laryngeal pattern is more consistent with that of lung cancer, and it shows lower mutation indexes (35.4% versus 60%) with different concentrated regions of p53 mutations. In most head and neck cancers, exons 5 through 8 are altered, whereas in laryngeal cancer, the mutations tend to cluster in exon 5, especially in the S2 protein domain and in codon 248 . A therapeutic strategy that is currently in development is the delivery of wild-type p53 into dysplastic or cancerous cells to restore wild-type p53 function within the cell. This approach has been explored using a recombinant adenovirus that delivers wild-type p53 (Ad5CMV-p53) in a laryngeal cancer model. In vivo studies demonstrated that intratumoral injections of Ad5CMV-p53 significantly inhibited the further growth of established laryngeal tumor xenografts . Furthermore, studies have demonstrated that the cotransfer of wild-type p53 with immunomodulatory genes, such as granulocyte-macrophage colony-stimulating factor (GM-CSF) and B7-1 genes using a recombinant adenovirus, induced cellular apoptosis and enhanced the immunogenicity of cancer cells, suggesting a potential combinatorial gene therapy strategy for laryngeal cancer .

Epigenetic alterations

Genetic loss of function mutations, such as point mutations, deletions, and translocations, was one of the first identified means of silencing TSGs; however, it has been found that gene silencing can also be achieved by epigenetic alterations of the genome, such as DNA hypermethylation of CpG islands in the promoter regions of TSGs. DNA hypermethylation of a promoter region results in transcriptional suppression of the gene through the recruitment of chromatin remodeling complexes. Different epigenetic profiles have been found when comparing tumor with normal tissue, which has provided insight into genetic oncogenesis.

The main mechanisms for epigenetic silencing are DNA hypermethylation and chromatin modification. DNA hypermethylation refers to the addition of a methyl group to the cytosine ring from a methyl donor by a DNA methyltransferase (DNMT). This occurs in regions of the DNA in which there are high concentrations of cytosines followed by guanosines or CpG dinucleotides. These regions are called CpG islands and are present in the noncoding promoter regions. Detection assays for methylation can be targeted to specific genes or global genomes. Among the targeted techniques, a methylation-specific restriction enzyme assay with Southern blot analysis was the first to be used. Currently, PCR techniques are used after sodium bisulfite treatment, which converts nonmethylated cytosines into uracil and makes methylation detectable by the presence or absence of a PCR product on an agarose gel.

The p16 gene which is located on chromosome 9p21 is a region that has been found to be silenced through promoter hypermethylation. Genomic analysis of the p16 locus demonstrated that 67% of patients had a homozygous deletion of the p16 locus and 21% had hypermethylation of the p16 promoter . In another study of 14 HNSCC cell lines, 21% of the lines had homozygous deletions of the locus, 29% had exonic mutations, 14% had intronic mutations, and 29% showed hypermethylation of the p16 promoter . Investigations of other genes known to exhibit promoter hypermethylation in other cancers have revealed similar promoter hypermethylation patterns in HNSCC as well. For example, E-cadherin (CDH1) is involved in cell-to-cell adhesion and is methylated in 0% to 85% of HNSCC, O6-methylguanine-DNA methyltransferase (MGMT) plays a role in detoxifying DNA adducts and is methylated in 25% to 52% of HNSCC, and death-associated protein kinase (DAPK) is involved in apoptosis and is methylated in 7% to 68% of HNSCC . The wide range of variation might be attributed to the different detection techniques used. The current challenge in epigenetics is identifying genes that are relevant and uniquely involved in HNSCC.

The methylation pattern of tumors has been evaluated within the tumor bed and in body fluids, such as serum or exfoliated cells shed by the tumor. A study involving 33 patients diagnosed with early stage laryngeal cancer detected p16 methylation in 42% of the pharyngoesophageal collections; combining it with UT5085 tetranucleotide microsatellite instability, this combination of genetic and epigenetic alterations were found in 76% of the tumors . This study highlights the importance of combining genetic and epigenetic alterations when evaluating molecular changes in any cancer, because the combination of alterations may provide a better “fingerprint” of the cancer biology than any single approach.

Despite regional promoter hypermethylation, global hypomethylation is also present in HNSCC and other solid tumors. In a study looking at 134 samples, the global methylation level for HNSCC specimens was 46.8% and the methylation level for normal controls was 54% ( P <.001). Hypomethylation levels were found to increase with tumor stage and in patients exposed to alcohol and tobacco smoke . The implication of global hypomethylation is the potential reactivation of proto-oncogenes in contrast to hypermethylation, which is the silencing of TSGs. Epigenetic alterations are an interesting phenomenon, and as more insight is gained into this area of research, it should be interesting to see how epigenetic changes can be integrated into a head and neck tumor progression model.

Altered protein expression

In addition to genetic and epigenetic modifications, protein-based alterations may contribute to carcinogenesis. The epidermal growth factor receptor (EGFR; also known as HER1 and ErbB1) is a transmembrane tyrosine kinase (TK) receptor that plays a critical role in cell survival and proliferation. Activation of EGFR through ligand binding with the epidermal growth factor (EGF) or transforming growth factor-α (TGFα) leads to receptor dimerization, kinase activation, and autophosphorylation, which activate various cellular pathways involved in cellular proliferation, angiogenesis, metastases, and inhibition of apoptosis. More than 90% of HNSCC overexpresses EGFR . In HPV-associated head and neck cancers, the viral protein E5 upregulates the expression of EGFR ; however, in non–HPV-associated HNSCC, the mechanisms for upregulation are not fully understood with gene amplification of the EGFR gene occuring in a low percentage of HNSCC (0%–25%) . Therefore, alternative proposed mechanisms for overexpression include mutational activation of EGFR , overexpression of EGFR ligands, establishment of autocrine or paracrine loops , and transactivation by other receptor and nonreceptor TKs .

EGFR overexpression has been associated with an unfavorable prognosis in early glottic carcinomas and has been linked to early disease progression, poor survival, and resistance to chemotherapy . Therefore, blocking the EGFR signal transduction pathway has been evaluated as a potential target for anticancer therapy. The EGFR pathway can be targeted by monoclonal antibodies (mAbs) and by tyrosine kinase inhibitors (TKIs). Anti-EGFR antibodies, such as cetuximab, which is a chimerized mAb, act as competitive antagonists to the receptor ligands, such as EGF and TGFα. Suggested mechanisms of action of anti-EGFR mAb-based therapy include inhibition of ligand-induced activation of this receptor and induction of receptor degradation. Other potential mechanisms of therapy are antibody-dependent cellular cytotoxicity, complement-dependent cytotoxicity, and complement-dependent cell-mediated cytotoxicity . In early clinical studies, single-agent activity of cetuximab was shown to be effective and safe in HNSCC . The most common side effect was a skin rash, and less common side effects included fatigue, nausea, vomiting, diarrhea, mucositis, and hypersensitivity reactions. In a phase III randomized clinical trial that compared radiation therapy (RT) and cetuximab with RT alone, patients who had locally advanced HNSCC demonstrated better survival and locoregional control by 10% to 15%. Laryngeal preservation rates were also improved . Consequently, in February 2006, the US Food and Drug Administration (FDA) approved cetuximab in combination with RT as a frontline treatment for patients who have locally advanced HNSCC . Additional trials have been undertaken to assess the feasibility of combining cetuximab with chemo-RT. A phase II trial demonstrated that the combination of cetuximab with RT and cisplatin yielded 3-year overall survival, progression-free survival, and locoregional control rates of 76%, 56%, and 71%, respectively . In addition, the combination of cetuximab with RT and gemcitabine in HNSCC yielded a complete response rate of 77%, with 89% patient compliance to chemotherapy .

In light of these promising results, several groups have addressed the potential benefit of inhibiting the EGFR pathway downstream of the extracellular EGFR domain by using other molecules, such as TKIs, that target the EGFR intracellular machinery. TKIs are small molecules that cross the plasma membrane and interact with the cytoplasmic domain of cell-surface receptors, which modulate intracellular signaling. They are less specific than mAbs and may be associated with increased toxicities because of the associated inhibition of several signaling pathways. Similar to the results obtained with single-agent mAbs, modest responses ranging between 1% and 11% were obtained with single-agent TKIs in phase II trials using gefitinib or erlotinib in patients who had recurrent or metastatic HNSCC. The combination of TKIs with chemotherapy or RT resulted in more favorable results, however . A phase I/II trial of erlotinib and cisplatin in patients who had recurrent or metastatic HNSCC demonstrated a favorable toxicity profile and antitumor activity comparable to standard combination chemotherapy regimens .

Several recent reports suggest that dual-agent targeting of the EGFR pathway may overcome the limitations of a single agent to suppress EGFR-mediated signaling sufficiently. In a human tumor xenograft model, the combined treatment of gefitinib and cetuximab resulted in a synergistic effect on inhibiting cell proliferation . A phase I study demonstrated that cetuximab combined with gefitinib enhanced tumor cell apoptosis and reduced proliferation rates of tumors as compared with a single agent alone in patients who had HNSCC .

As clinical trials that target the EGFR pathway continue to grow, the mechanisms underlying tumor resistance to or lack of sensitization by EGFR inhibitors will need to be addressed. One hypothesis for resistance to EGFR inhibitors is the presence of EGFR mutations in the extracellular domain, which results in a lack of recognition by mAbs, and in the intracellular domain, which results in a diminished response to EGFR antagonists. There are only a few somatic mutations of the EGFR gene reported in patients who had HNSCC, however. In Asian patients, the incidence rate was reported to be 7.3% (3 of 65 patients) ; in European patients, it was 1% (1 of 100 patients) ; and in American patients, no mutation has yet been detectable (0 of 65 patients) . Other possible explanations for the limited efficacy of EGFR-directed therapy are the constitutive activation of signaling pathways downstream of EGFR or activation of EGFR-independent pathways, such as G protein–coupled receptors (GPCRs) , which may promote survival and resistance to EGFR inhibitors. Because of the potential of developing resistance to directed therapies, multimodality treatments that target multiple aberrant cellular pathways may be the appropriate strategy to improve overall clinical outcomes.