| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Review |
1 Department of Surgery, Postgraduate Medical Institute in Association with Hull York Medical School, University of Hull, Wolfson Building, Cottingham Road, Kingston Upon Hull, United Kingdom, HU6 7RX
2 Department of Oncology, Postgraduate Medical Institute in Association with Hull York Medical School, University of Hull, Wolfson Building, Cottingham Road, Kingston Upon Hull, United Kingdom, HU6 7RX
Correspondence: Address correspondence and reprint requests to: John Greenman, PhD; E-mail: j.greenman{at}hull.ac.uk.
ABSTRACT
Information from the genetic analysis of head and neck cancer has grown enormously in the last 20 years. The advent of high-resolution genetic analysis techniques such as microarray technology will further expand this field in the future. Here we review the data on chromosomal aberrations of head and neck squamous cell carcinoma, focusing on the data generated by comparative genomic hybridization analysis, and suggest how such findings will be taken forward over the next decade. With the search engine PUBMED, the key words "comparative genomic hybridisation," "head and neck," "oral," "hypopharyngeal," "laryngeal," and "squamous cell carcinoma" were used. Publications unavailable in English were excluded.
Key Words: Head and neck cancer • Genetic analysis • Chromosomal aberrations • Comparative genomic hybridization
Cancer genetics has been revolutionized in the last 30 years with the development of genome-wide genetic analysis techniques—i.e., comparative genomic hybridization (CGH) and multifluorescence in situ hybridization—that complement classic cytogenetic methods. These techniques have been used to characterize the genetic profiles of multiple solid cancers and have contributed significantly to our understanding of the genetic control of tumorigenesis and metastatic progression in many different cancer types. This review discusses the genetic alterations found in head and neck squamous cell carcinoma (HNSCC), excluding squamous cell carcinoma (SCC) of the skin, and concentrates on alterations identified by CGH.
GENETIC ALTERATIONS IN HEAD AND NECK CANCER
HNSCC is the sixth most common cancer worldwide,1 90% of which are known to be associated with prolonged tobacco and alcohol abuse and are said to be "sporadic."2 It is understood that HNSCC develops via a multistep process and progresses through a series of well-defined histopathologic stages. The multistep progression model theory was first proposed by Vogelstein et al.3 for colorectal adenocarcinoma. Vogelstein et al. also suggested that it is the accumulation, rather than a particular order, of genetic alterations that leads to the invasive phenotype. Eight years later, a revised version of this model was suggested for HNSCC, including some of the genetic aberrations associated with the steps between hyperplasia, dysplasia, carcinoma-in-situ, and invasive cancer (Fig. 1
).4 The model was constructed by using 10 specific tumor-suppressor regions that had previously been shown to be commonly altered in >40% of invasive HNSCCs by allelotype analysis with microsatellite markers.
|
The conclusions about HNSCC drawn from CGH analysis alone should be viewed with some caution, however, because of its relatively low resolution, the generally small cohort sizes, and the heterogeneity of the samples studied. For example, the first CGH data were published by Speicher et al.5: 13 tumors were analyzed, of which 6 were pharyngeal, 3 tongue, 2 larynx, 1 lip, and 1 "neck." Many other CGH studies have since been reported, most also analyzing heterogeneous groups, including cell lines derived from HNSCC. Fortunately, some groups have focused on single subsites—in most cases, oral SCC. All the published reports are summarized in Table 1. Over time, many centers have expanded their cohorts and republished their results; these are clearly marked in Table 1 (with a superscript letter identifying related cohorts). In this review, an attempt has been made to support any conclusions drawn from CGH data with evidence from classic cytogenetic and immunohistochemical staining studies that have investigated the presence or absence of specific genes or gene products. Some genetic loci previously identified as harboring putative oncogenes or TSG have not been shown to be significant with CGH, probably because of the low resolution of the technique. For example, chromosome 17 gain was found in only 20% of cases in one study,6 even though it is well established that the p53 gene is located at 17p13 and even though its loss of function is thought to be a key event in the transformation of preinvasive to invasive carcinoma.7 Therefore, this review does not attempt to discuss specific genes in depth, even though some are highly likely to be involved in HNSCC, but, where appropriate, brief comments have been made with reference to recent seminal reviews.8,9
|
In an effort to summarize the CGH data in the most useful manner, only the most frequent or potentially important (supported by other cytogenetic techniques), but less common, aberrations are reviewed.
Chromosome 3
Copy number changes on chromosome 3 could be considered to be the most important in the genetic development of HNSCC. Speicher et al.5 found that 10 (75%) of 13 samples had a DNA copy number increase on the long arm of chromosome 3 (3q) and that 5 (38%) of 13 had a loss of DNA on the short arm (3p). Later, isochromosome formation (one arm translocated to the other) was suggested,14 but CGH analysis alone cannot prove this, because the technique can detect only a change in the ratio of normal to tumor DNA content and not specific chromosomal rearrangements. Chromosome 3 alterations have been consistently reported in nearly all CGH studies since these early ones.
3p Deletions
The deletion of the short arm of chromosome 3 is considered one of the most pivotal mutations in a range of cancers, e.g., lung, cervix, and renal.16 Weber et al.11 reported this deletion in premalignant oral samples (n = 12) and invasive oral SCC (n = 14); in most cases, the entire arm was deleted. This is in agreement with Califano and colleagues’ model that 3p deletion is an early event in the development of HNSCC. Bockmühl et al.14 reported three distinct nonoverlapping regions on the short arm of chromosome 3 (3p11-14, 3p21-22, and 3p25) to be commonly involved. Stafford et al.17 concurred, finding loss of the entire short arm of chromosome 3 in 10 (52%) of 19 cases and partial loss in another 3 tumors.
It is interesting to note that 3p14.2 is the most common breakage site in all epithelial cancers.18 The fragile histidine triad (FHIT) gene, a member of the histidine triad superfamily of nucleotide-binding proteins, among others, spans this location. It seems to function as a proapoptotic TSG that, when deleted, is associated with the development of multiple epithelial-derived cancers.19 The FHIT gene has been shown to be mutated in up to 85% of all HNSCCs.20 Other studies have looked at FHIT tumor-suppressor protein expression in HNSCC. Correlation has been found between low protein expression, low rates of apoptosis, and high tumor cell proliferation,21–24 and this correlation strongly supports the indication that FHIT is a key mutation involved in HNSCC development.
Lin et al.25 postulated that deletion of 3p is also associated with the development of metastasis, thus implying the involvement of other genes. Several candidate TSGs have been located to the regions reported by Bockmühl et al.14 These include the DNA mismatch-repair gene MLH1, which resides at 3p21.3-p23; XPC, located at 3p25; and the von Hippel--Lindau (VHL) TSG, located at 3p25-p26. The VHL gene has been extensively investigated, and it is known that this gene product is required to degrade hypoxia-inducible factor 1. Hypoxia-inducible factor 1 is a transcription complex that plays a key role in oxygen homeostasis by controlling the expression of many genes, including angiogenic growth factors, glucose transporters, and glycolytic enzymes.26 Loss of VHL gene function leads to the overexpression of hypoxia-inducible factor 1 and the subsequent upregulation of proangiogenic mediators such as vascular endothelial factor. The VHL gene is also thought to regulate extracellular matrix formation.27 If this were so, loss of the VHL gene would offer a possible mechanism causing disruption of the extracellular matrix and new vessel formation, thus beginning the development of an invasive phenotype. Unfortunately, very little evidence supports the specific role of VHL in HNSCC. By using microsatellite analysis and hypermethylation studies, the VHL gene status was investigated in 26 upper aerodigestive tract SCCs that had previously been shown to exhibit allelic loss of 3p. In no case was LOH or gene inactivation found in the region of the VHL gene.28 No other evidence for or against the role of the VHL gene in HNSCC has been reported to date, but it remains an attractive gene for further investigation.
3q Gain
Bockmühl et al.14 reported a gain of DNA material on chromosome 3q in 44 (88%) of 50 tumors, of which 9 were high-level amplifications (copy number increase 
1.5) locating to 3q26-q27. Previously Speicher et al.5 reported similar percentages of gains, as well as high-level amplifications also locating to 3q25-27 in 3 of 13 tumors. Four studies have attempted to further define this 3q amplification in HNSCC cell lines with CGH and other forms of chromosome analysis; all have concurred that 3q26 (ranging from 3q24 to 3q28) is an important chromosomal aberration in HNSCC.29–32 Okafuji et al.33 postulated that 3q26-28 gains were associated with stage T2 or above in oral SCC. However, Bockmühl et al.34 constructed a putative progression model that suggested that the gain on 3q was an earlier aberration leading to the development of an invasive phenotype in HNSCC. These discrepancies are most likely due to the differences in cohort number and subsite composition of the tumors analyzed.
Many putative oncogenes have been located to 3q26-28, such as LAZ3, BCL-6, PIK3CA, telomerase RNA, and the AIS gene. AIS gene products have been shown to be amplified independently in HNSCC,35 along with PIK3CA.36 The function of AIS is not yet known, but it is thought to be important in stem cell renewal; thus, when overexpressed, it gives the cell the capacity for neoplastic growth. Hibi et al.35 observed that AIS and p53 mutations commonly occur together. This suggests a specific tumorigenic pathway, but no genotypic studies have confirmed their findings.
An alternative gene, PIK3CA, encodes for the catalytic subunit p110 
of phosphatidylinositol 3-kinase, one component of a lipid-signaling pathway involved in multiple cancer-related functions: cell survival, proliferation, and cell migration. Although strong data support the role of PIK3CA in various cancers (e.g., ovarian37), there is less evidence of its involvement in HNSCC. Redon et al.36 reported that overexpression of PIK3CA in low-grade HNSCC (node-negative tumors) was associated with an increased gene copy number. This implies that PIK3CA may participate in early HNSCC development. They also found that overexpression of 3q26 predicts clinical outcome in these early-stage tumors. A year later, p110 messenger RNA (mRNA) and protein levels were assessed in dysplastic and invasive HNSCC, along with 3q26 gain.38 Woenckhaus et al.38 suggested that 3q gain was associated with the development of high-grade dysplasia in 7 (78%) of 9 cases and of invasive HNSCC in 11 (100%) of 11 cases. They also supported the role of PIK3CA in HNSCC, because tumors that had shown a 3q gain in >50% of the cells or a 3q amplification in >10% showed increased p110 mRNA and protein expression. The minor drawback of this study is, once again, the size (n = 20); verification with a larger, independent series of tumors is needed to confirm these results.
Chromosome 5
Speicher et al.5 proposed the involvement of chromosome 5 in the head and neck multistep tumorigenesis cascade, an aberration previously not found by classic karyotyping techniques. Deletions of the long arm of chromosome 5 are often associated with gains of DNA on the short arm, thus suggesting possible formation of 5p isochromosomes, as has been suggested with chromosome 3. Isochromosome formation has also been considered for other chromosomes—namely, chromosomes 7, 8, and 9—but, as has been already stated, CGH is limited to detection of DNA copy number alterations and not balanced translocations.
5p Gain
The presence of a gain on the short arm of chromosome 5 varies in frequency between studies (range, 0%–86%); the highest incidence is found in oral cell lines (86%). In tumor samples, Wolff et al.39 reported a similar incidence of 80% (16 of 20) in oral SCC. The frequencies of 5p gains in heterogeneous series are much lower—approximately 30% to 45%. This suggests subsite differences. Huang et al.,40 however, reported a gain of 5p to be a common aberration in all HNSCC (40% of their cohort; n = 75) and suggested that it was important in the progression of HNSCC, independently of site, and was not specific to tumors of oral origin.
In Bockmühl and colleagues’ earliest series,41 they also found that 16 (89%) of 18 node-negative HNSCC tumors had amplifications on the short arm of chromosome 5, thus implying an association with metastasis; unfortunately, the loci of these amplifications were not documented. Welkoborsky et al.42 observed similar frequencies of 5p14-ter gains between both metastasizing and nonmetastasizing cohorts (n = 20) but did not comment on the frequencies of amplifications rather than gains. The significance of the difference between gains and amplifications remains to be clarified in gene-expression or phenotype studies, as does the location of these changes, because two discrete areas may be involved. As yet, no putative oncogenes that could explain these differences have been identified on the short arm of the chromosome.
5q Deletion
Deletions of the long arm of chromosome 5 and gains on the short arm are not as common an occurrence together as would be expected if isochromosome formation were always the cause. Bockmühl et al.14 reported 70% (21 of 30) of samples to have a 5q deletion, whereas only 37% (11 of 30) had a gain on 5p. Later, Bockmühl et al.,43 investigating the differences between grades of tumor, reported a deletion on the long arm of chromosome 5 to be associated with well-differentiated tumors (along with a gain of 3q and deletions on 3p and 9q).
On the long arm, the TSG APC is located at 5q21, and although its importance in familial adenomatous polyposis and subsequent colon cancer is well established,44 a role in HNSCC remains to be clarified. Most of the evidence for the APC gene’s involvement in HNSCC has been produced from investigation of oral SCC. A study of 43 oral SCCs reported LOH at 5q21 in 42% of samples and found an association with poor prognosis.45 Other studies support the involvement of the APC gene in Chinese and Caucasian oral SCC, but it is interesting to note that there is less evidence for its involvement in Indian patients.23,46,47 This may simply represent study variance, or different etiological factors may initiate tumorigenesis by separate pathways.
Another gene on this chromosome arm is one of the RAS family, RA83c. In HNSCC in the Western world, Ras mutations are rare48 (<5%), whereas in India, 35% of HNSCC patients harbor mutations of the H-ras and K-ras genes.49 The role of Ras genes in tumorigenesis has been relatively extensively investigated; they are responsible for signaling pathways, which regulate proliferation and differentiation in normal and cancer tissue.50
Chromosome 7
7p Gain
Hermsen et al.51 and Bockmühl et al.14 were the first to report a gain on the short arm of chromosome 7 to be a common finding in HNSCC. Bockmühl et al. reported a statistically significant association between this aberration (7p15) and node-negative tumors. Bergamo et al.52 reported a gain on 7p in 32% of a node-negative cohort tumors (n = 12) and localized the mutation to 7p13-p22. However, more recently, Gebhart et al.53 reported 54% (19 of 35) of their cohort to have an aberration at 7p12. They reported that specific aberrations were associated consistently with 7p12 gain and suggested that these tumors have a distinct genetic pathway compared with tumors that do not express this aberration. In addition, they reported this aberration to be associated with a poor prognosis; however, this finding was not substantiated by Ashman et al.54 in a similar-sized study.
The epidermal growth factor receptor (EGFR) and the insulin-like growth factors IGFB1 and IGFB2 are three potentially interesting genes located in the 7p13-22 region. The expression of EGFR has been extensively investigated, particularly with respect to therapeutic targeting of HNSCC. As a member of the c-erb family of transmembrane proteins, it is involved in the transcriptional regulation of proteases and cytokines implicated in tumor invasion and angiogenesis.55 The quantity of evidence to support the role of EGFR protein involvement in HNSCC is compelling, and there are also data to show gene-expression changes,56–58 but few studies have investigated genotype and phenotype in the same tumor samples. Weichselbaum et al.59 examined 11 HNSCC samples, 3 of which showed increased EGFR expression and 10 of which showed increased EGFR mRNA. This work is not definitive but suggests that at least two pathways regulate EGFR protein expression.
Huang et al.40 reported aberrations on chromosome 7 to be associated with specific subsites. They observed an increase in copy number along the short arm to be associated with laryngeal SCC. Kujawski et al.,13 studying laryngeal samples only, did not support this conclusion, because 7p gain was not reported as a common finding (only 11% [4 of 38] samples showed a copy number increase at 7p15-ter).
7q Gain
Huang et al.40 also reported a gain on the long arm of chromosome 7 to be associated with pharyngeal SCC. It is interesting to note that both Tremmel et al.60 and Bockmühl et al.34 reported a higher frequency of 7q (7q11.2) gains in metastases compared with the corresponding primary tumor. Both Tremmel’s and Bockmühl’s series were mainly made up of pharyngeal samples (Tremmel, 22 of 32; Bockmühl, 24 of 52); it is thus possible that gain of the long arm is more commonly found in pharyngeal SCC and that it is indicative of metastatic potential. If this is the case, it may explain one of the reasons for the aggressive behavior and the poor prognosis of pharyngeal carcinoma. Dahlgren et al.,7 studying 25 primary tonsillar SCCs, also found an association between gain at 7q11.2-q22 in human papillomavirus–negative tumors (P = .017), together with a decreased disease-specific survival (P = .002). The decreased survival may once again be a reflection of the increased potential for metastatic progression.
Oncogenes on chromosome 7q are yet to be discovered. Cromer et al.61 have identified a few potential genes associated with tumorigenesis and metastatic potential of hypopharyngeal cancer by using microarray analysis on 34 tumor samples. They reported overexpression of the breast cancer metastasis suppressor 1 gene (BRMS1), which is located at 7q11.2-22; however, its role in HNSCC tumorigenesis remains unproven.
Chromosome 8
8p Deletion
There are conflicting reports regarding the role of deletions on the short arm of chromosome 8 in HNSCC. Bockmühl et al.41 originally reported that 53% of mixed HNSCC tumors had a deletion of chromosome 8p that was associated with poor differentiation and high-grade tumors. The only other group to report this deletion at a similar frequency was our own.17 In the latter series, 14 of 19 tumors were graded as moderately or poorly differentiated. Neither Speicher and colleagues5 (n = 13) nor Brzoska and associates62 (n = 10) found this association, even though their cohorts contained mostly high-grade tumors. Thus, the association between a deletion of 8p and poor differentiation should be viewed with caution.
8q Gain
Copy number gains are frequently found on the long arm of chromosome 8 in HNSCC. Fifty-seven percent (n = 30) of Bockmühl and associates’41 series showed a gain on this arm that specifically involved 8q23-q24. Stafford et al.17 reported a lower frequency (31%) and also suggested a second location, 8qcent-q13, to be of interest (26%). Bergamo et al.,52 publishing a year later, confirmed both locations but found a gain at 8q12 in only two of their series (n = 19). Most studies have shown both these areas to yield copy number increases (Table 1
).
Several genes of interest have been reported to be located on the long arm of chromosome 8: C-myc at 8q24, PTK2 at 8q23-q24, and lyn at 8q11-q12. LOH at 8q24, the location of the C-myc gene, has been previously reported in HNSCC, as has overexpression of the C-myc gene.63 A chromosomal gain in the area of 8q24 has also been shown to be associated with deregulation of protein expression.32 The C-myc gene encodes for a 62-kDa transcription factor that represses cell-cycle inhibitors, regulates apoptotic pathways, and, therefore, gives the cell a growth advantage.64 Overexpression is thus thought to lead to malignant transformation and has been found to be associated with poor survival.65
The PTK2 gene (focal adhesion kinase protein/ tyrosine kinase 2) has been found to be amplified in several primary tumor sites, e.g., glioblastomas and hepatocellular carcinomas.66,67 It has been shown to play an important role in controlling adhesion and growth-regulatory signal pathways; thus, overexpression may confer a growth advantage.68 As yet, there is no evidence for involvement in HNSCC tumorigenesis, but because of its close proximity to the C-myc gene, coamplification may be important.
Chromosome 9
Classic karyotyping originally suggested that a loss of DNA on the short arm of chromosome 9 was an early occurrence in the carcinogenic process that was associated with the development of dysplasia in HNSCC.69 Miracca et al.70 reported that 48% of HNSCCs demonstrated LOH at 9p21. Bockmühl et al.41 reported a 9p gain in 68% of their primary cases. Other studies have reported considerably lower values (0% and 36%). Although there seems to be inconsistency in the results from the CGH studies, previous work using other techniques has shown 9p21-22 to be one of the most common aberrations found in HNSCC; therefore, this region is extremely likely to harbor a putative TSG.69 The p16 (CDNKN2/MTS1) and p14ARF genes have been identified at this locus. Loss or inactivation of CDNKN2/MTS1 is thought to initiate G1 arrest and prolong cell life by blocking the cyclin D–dependent kinases Cdk4 and Cdk6, thereby inactivating apoptosis and other cell-regulatory pathways usually controlled by the retino-blastoma (Rb) gene. The INK4a-ARF gene, also found at 9p21, responds to hyperproliferative states by activating both the Rb and p53 genes and by arresting cell proliferation. Therefore, loss of this region prevents the initiation of a key apoptotic pathway, thereby prolonging cell viability. Sherr71 has recently published an authoritative review of this topic.
Redon et al.36 suggested that the loss of 9p, along with a loss on the long arm of chromosome 18, was indicative of stage T3 or T4 tumors, and this suggestion was supported by Brieger et al.72 It contradicts previous findings that this mutation was a relatively early event (Fig. 1). The disparity between results of classic karyotyping techniques and most CGH findings can probably be explained by the relatively low resolution of CGH. CDKN2A is approximately 27 kilobases long. CGH can accurately detect deletions of only approximately 10 megabase pairs and above and, thus, could very easily miss specific, small aberrations.
Chromosome 11
A further consistent finding in most studies is the involvement of chromosome 11. It has independent gains and losses of chromosomal material along the q arm. DNA is gained at 11q13, whereas the losses are located at 11q23. Redon et al.,36 comparing stage I and II with stage III and IV SCC tumors, observed a specific pattern of aberrations in the early tumors that affected 3q, 3p, 8q, and 11q. They suggested that 11q13 gain, along with these other aberrations, represented the key early events in tumorigenesis. Tremmel et al.60 confirmed the importance of the specific gain at 11q13, reporting a high frequency (78% of tumors) in their primary cohort (n = 35) of stage III and IV tumors, most of which (n = 26) originated in the pharynx. Huang et al.,40 however, reported that a gain at 11q13 was indicative of laryngeal and oral cancer, but not pharyngeal SCC, when comparing the genetic aberrations found in different subsites. Dahlgren et al.7 found an 11q deletion in human papilloma virus–negative oropharyngeal SCC, again contradicting Huang et al.40 These apparent differences could be explained by different etiological factors but are most likely due to chance events in low cohort sizes that may be unrepresentative of the genetic changes found in different subsites.
A gene possibly involved in carcinogenesis, located at 11q13, is the PRAD1 gene, which codes for cyclin D1. Cyclin D1 is a cell-cycle regulatory protein involved in control of the transition between the G1 and S phases of the cell cycle (see above). Other potentially interesting genes are fibroblast growth factors 3 and 4, but overexpression has not been reported in HNSCC.73
A deletion spanning 11q14-qter was first reported by Bockmühl et al.14 Stafford et al.17 also observed this finding but isolated the location of the aberration to 11q23-qter; this was independently confirmed with LOH studies.74 Subsequently, Bockmühl et al.34 suggested that the deletion was associated with metastatic progression in HNSCC (along with 11p14). Other studies have also reported an association between LOH at 11q23 and recurrence,75 but no putative genes have been identified at this location.
ABERRATIONS ASSOCIATED WITH HEAD AND NECK SUBSITES AND ETIOLOGICAL FACTORS
Some aberrations have been found to be present in HNSCC, but not as prevalently (generally <50%) as the chromosomal aberrations discussed previously. These aberrations include deletions on chromosomes 4 (4q) and 18 (18q). As has been already implied by Huang et al.,40 a potential reason for some aberrations to be at a lower frequency is that most studies have looked at head and neck SCC and not at specific subsites or etiological factors, thus diluting possible findings. Obviously, separation of the tumors into defined anatomical divisions can be difficult; e.g., patients often present with piriform fossae carcinoma that extends into the supraglottis or vice versa. There are also many patients who smoke and drink alcohol but few who do one without the other. The same can be said of betel nut consumers and tobacco smokers in Eastern Asia. In a unique study, Lin et al.25 investigated the genetic profiles of oral SCC associated with betel nut use (n = 33) compared with cigarette smoking (n = 14). As expected, most aberrations were the same in the two groups (both groups smoked tobacco), but the frequencies of a gain on 8q and/or deletions on 4q were significantly lower in the betel nut group. This suggests that the etiological factor does influence the genetic signature in carcinogenesis. Singh et al.76 attempted to establish whether the severity of tobacco and alcohol abuse affected the genetic aberrations found. Overall there was no correlation, but they did report that 1p gain and 3q amplification were significantly more common in patients with a history of alcohol and smoking than in those who did not. Obviously, these are two relatively small studies; however, the results are interesting, and further work needs to confirm or refute their findings.
SCC of the oral cavity has been studied most extensively as an individual subsite of the head and neck region (Table 1). Oga et al.77 reported an average of 5.6 aberrations per tumor in oral SCC, compared with 16 in HNSCC; however, it must be noted that approximately half the oral tumors in this study were stage I and II, whereas most head and neck cohorts contain predominantly advanced-stage tumors. Lin et al.25 reported similar frequencies in their cohort of oral tumors, whereas Singh et al.76 reported a slightly higher average of nine aberrations per tumor, again in oral tumors. Wolff et al.,39 in stark contrast to these findings, reported an average of 23 aberrations per oral tumor. The common aberrations (gain of 1p, 3q, 5p 8q, 9q, and 11q and deletion of 3p and 4p) do not differ markedly in location or frequency between these studies, and only 9q gain and 4p deletion seem more frequent in oral SCC than in other HNSCC, thus suggesting a particular role for these locations.
GENETIC ABERRATIONS ASSOCIATED WITH METASTASIS
No clear genetic model has emerged that predicts metastatic potential. Six groups, including our own, have looked at the differences among the genetic profiles of primary HNSCC, their matched lymph nodes, and node-negative tumors.34,42,78–81 It has been postulated that comparison of the primary tumor with matched metastatic deposits would elucidate genetic aberrations that could not be detected in the primary tumor because of its heterogeneity. All six groups observed that, although there were more aberrations in the metastatic nodes, the genetic profiles were clonally related to the dominant population in the primary SCC.
As has been already stated, Bockmühl et al.41 found that gains on 5p and 10q were associated with node-negative tumors, but these results were not supported by Welkoborsky et al.,42 who found that 22q gain and deletion of 18q were associated with metastasis. Bockmühl and colleagues41 observed more specific aberrations than most others—namely, gains on 7q11.2 and 1q21-22 and losses on 5q, 8p, 11p14, 11q14-qter, 10p12, 10q, and 14q. Some of these aberrations, i.e., 10p11-12 and 11p gain and 4q22, 9p13-24, and 14q deletion, were also found to be associated with nodal metastases by Wreesmann et al.78 in a recent study. Kujawski et al.,13 investigating laryngeal SCC, also reported losses of 8p, 9q, and 13q to be associated with metastasis. Together these findings are intriguing, because they suggest that some common aberrations do exist that are responsible for lymph node spread. However, the changes may simply be due to chance because of the relatively low numbers of tumors in each cohort. It is of paramount importance that large, homogeneous cohorts of HNSCC be studied to answer this question, because identifying metastatically competent tumors would potentially significantly affect the clinical management of patients.
SUMMARY
Genetic analysis of HNSCC has found chromosomal aberrations in all subsites and developmental stages. CGH has contributed a significant amount of information to such analysis and has facilitated our understanding of tumorigenesis. Although CGH has limitations, it is a useful tool for identifying areas across the entire genome, thus allowing specific chromosomal analysis to be performed more efficiently.
The complexity of the genetic profile of HNSCC is being slowly unraveled as more tumors are analyzed and techniques are standardized and perfected. High-resolution techniques such as microarrays are more specific and will add significant amounts of information to the genetic foundations discussed in this review.
This review has shown that several common aberrations, in combination, seem to control the biological behavior of HNSCC. Furthermore, there is a distinct possibility that some of these aberrations are associated with particular etiological factors, stage of disease, and specific anatomical subsites. The significance of spontaneous aberrations, which occur less frequently than the common aberrations discussed previously, remains to be elucidated. As more studies are performed on large cohorts of homogeneous samples, together with the advent of high-throughput proteomics and genomics techniques, the answers to these questions will be forthcoming. Ultimately, the relative importance of specific aberrations, alone and in combination, as well as other associated factors, will become apparent and hence yield a clear genetic profile that can be used to aid in treatment selection for patients with HNSCC.
Received for publication September 17, 2004. Accepted for publication April 25, 2005.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
