Although there is a genetic basis for pathogenesis of the sporadic form of ESCC, genetic predisposition appears to be rare (3). Tylosis is a rare autosomal dominant disease characterized by diffuse hyperkeratosis of the squamous epithelium of the esophagus, palms of the hands, and soles of the feet, and is associated with a high incidence of SCC (4). Also known as keratosis palmaris et plantaris, this syndrome may also present with oral leukokeratosis and follicular hyperkeratosis (5). The early dermatologic manifestations usually begin between 7 and 8 years of age, and Howel-Evans et al. are credited with the initial observation that these patients develop EC at an early age (average age 45 years) (6). The disease locus, termed the tylosis oesophageal cancer (TOC) locus, has been mapped to chromosome band 17q25 by linkage analyses (7). Thus, a target gene locus of potential importance in sporadic ESCC has been unearthed, in keeping with the principle that genes mutated in familial syndromes are also implicated in sporadic forms of the same cancer (8). Iwaya et al. reported frequent loss of heterozygosity (LOH) on chromosome arm 17q using 20 microsatellite markers from the TOC locus in ESCC. They observed LOH within the region of the TOC locus in 33 of 52 informative cases (63%) (8). However, despite these findings, research until recently was unable to identify the culpable gene (9). At the time of this writing, however, a group from the University of Liverpool, led by John K. Field and Janet M. Risk, appears to have identified a promising candidate gene at 17q25, cytoglobin, which is downregulated and transallelically repressed in TOC (10).

The Shanxi Province in north-central China has among the highest rates of ESCC in the world (11). Studies have shown a strong tendency toward familial aggregation in this high-risk area, suggesting a potential genetic mechanism of carcinogenesis (12). Researchers have identified a high degree of genetic instability in the early precursor lesions of ESCC in patients from this region (13). In particular, germline mutations of BRCA2 in patients from this high-risk area of China are more frequent in ESCC patients who have a positive family history of ESCC (9 of 78, 12%) than in those without a family history of ESCC (0 of 48, 0%; p = 0.013) (14). In addition, GSTM1 (glutathione S-transferase) deletions were found to be a predisposing risk factor (OR 1.91, 95% CI 1.03–3.81) for ESCC development in Chinese patients who smoked tobacco, drank alcohol, and had a positive family history of EC (15). It should be noted that these GSTM1 deletions represent constitutive (i.e., germline) polymorphisms, rather than acquired (i.e., somatic) events that occur during tumorigenesis. The GSTM1 gene product is a metabolic enzyme that can detoxify a number of reactive electrophilic compounds, including carcinogenic polyaromatic hydrocarbons, and polymorphisms of this gene have been implicated in the pathogenesis of multiple human cancers (16). GSTM1 deletions have also been described in Indian, French, and Brazilian patients who smoked tobacco and had ESCC (17,18,19).

A familial form of EAC has been described in the setting of families with increased gastroesophageal reflux. There
are case reports of families in which multiple members are affected by gastroesophageal reflux disease (GERD), Barrett esophagus (BE), and EAC (20,21,22). Pedigree analysis suggests an autosomal dominant inheritance, with variable degrees of penetrance (23). Thus far, the only report of a susceptibility locus (at chromosome 13q14) was identified in a kindred of patients with pediatric esophageal reflux, and the incidence of BE or EAC in these patients was not known (24).

ESCC is one of the leading causes of cancer mortality in men, particularly among African Americans (3). Tobacco smoking and excessive alcohol intake constitute the largest risk factors for the development of ESCC in North America and Western Europe (25,26). In certain high-risk regions, such as the Henan Province of China, northern Iran, southern Turkey, and northern Africa, dietary factors may represent additional major determinants of carcinoma risk (3).

The pathogenesis of sporadic ESCC is driven largely by these environmental factors; this topic has been reviewed in depth elsewhere (27). Underscoring the importance of environmental factors is the fact that seven of eight ESCCs in an American pedigree with tylosis occurred in smokers, which suggests that even in the presence of familial genetic predisposition, environmental triggers are necessary for tumorigenesis (28).

Additional risk factors for ESCC are achalasia, Plummer-Vinson syndrome (dysphagia, iron deficiency, and esophageal webs), caustic injury to the esophagus, a history of head and neck cancer, a history of breast cancer treated with radiotherapy, and poverty (27).

BE is the most important precursor condition to sporadic EAC (29,30,31). BE results from long-standing GERD. Patients with weekly reflux symptoms are at greatest risk for developing both BE and EAC (27,32). Risk factors for GERD (and therefore BE and EAC) are obesity, decreased lower esophageal sphincter pressure, and the presence of a hiatal hernia (27,33). In addition, scleroderma is a condition characterized by diminished lower esophageal pressure and complicated by BE and an increased likelihood of EAC (34). Patients with BE have a 2% to 5% lifetime risk of developing EAC (35).

In addition to reflux of gastric acid per se, there is some evidence suggesting that reflux of alkaline small bowel contents may contribute to metaplastic transformation of the esophagus (36). The controversy surrounding the relative clinical importance of alkaline reflux has not been resolved. Some authors suggest that alkaline reflux is a minor contributor, whereas others believe that the mixture of gastric and duodenal contents may cause synergistic damage to the esophageal epithelium (37). In one study of patients who received gastric surgery for benign peptic ulcers, no increased incidence of BE was found, suggesting that bile reflux alone is not sufficient to produce a metaplastic change in the distal esophagus (38).

In recent years, promoter hypermethylation–associated inactivation of gene expression has been demonstrated for a number of tumor suppressor genes in neoplasia, and has created a paradigm shift in the understanding of the molecular pathogenesis of EC and cancer in general. Genes frequently methylated in cancer include, but are not limited to, adenomatous polyposis coli (APC), tissue inhibitor of metalloproteinase 3 (TIMP3), p16 (CDKN2A), p15 (CDKNB), retinoblastoma (RB1), E-cadherin (CDH1), GSTP1, and O⁶-methylguanine-DNA methyltransferase (MGMT) (39,40,41,42,43). A host of genes found to be abnormally methylated in EC are shown in Table 15.1. The mutL homolog 1 (MLH1) mismatch repair gene promoter is often hypermethylated in cancers with high frequencies of mismatch repair deficiency, such as colorectal, endometrial, and gastric cancers (42,44,45,46). In fact, some of the best evidence for promoter hypermethylation as a direct cause of diminished gene expression derives from the study of MLH1 in repair-proficient and -deficient colorectal cancer cells (47). Colorectal cancer cells defective in DNA mismatch repair and known to be hypermethylated at MLH1 were demethylated with 5-azacytidine. Not only did this demethylation restore expression of MLH1, but the return of effective DNA repair was also documented (47). The precise mechanism whereby promoter hypermethylation inhibits gene expression is not well understood. One theory asserts that promoter methylation facilitates histone deacetylation, thereby altering the DNA ultrastructure so that transcription factor binding is inhibited (48).

This additional putative mechanism of gene inactivation in cancer has resulted in a dramatic shift in the understanding of molecular alterations. Now, in addition to mutation and deletion, a third epigenetic mechanism is understood to account for the silencing of genes in cancer. Loci harboring tumor suppressor genes in any cancer need not undergo accompanying mutations, but rather become hypermethylated at one or both alleles to explain diminished or silenced expression. Previous studies suggesting the role of a known gene contained within a particular locus frequently deleted in a given cancer can now be extended to investigate hypermethylation of its target gene promoter within the specific locus.

This approach has already been successfully applied to the study of CDKN2A in EC. As previously stated, the CDKN2A locus at 9p21 is frequently deleted but is infrequently mutated in EC. Reports of frequent CDKN2A hypermethylation in EC have been forthcoming (49,50). In addition, it appears that hypermethylation of multiple genes is an early, easily detectable abnormality that predicts which patients progress from Barrett metaplasia to develop high-grade dysplasia or adenocarcinoma (51,52). However, future research is needed to fully validate methylation biomarkers before they are clinically applicable.

The first genetic abnormalities to be described in EC were aneuploidy and abnormal chromosome complement. Aneuploidy refers to abnormal DNA content and encompasses both structural and numerical DNA abnormalities. Aneuploidy does not correlate with any single mutation but reflects widespread
DNA changes due to genomic instability (53). Abnormal DNA content has been detected in ESCC and its precursor lesion, squamous dysplasia (54,55). Aneuploidy occurs in 84% to 94% of ESCCs (54,55,56,57). The prevalence of aneuploidy in squamous dysplasia and carcinoma in situ has been reported at 22% to 28% (55). Aneuploidy also occurs in histologically normal mucosa adjacent to aneuploid tumors and is believed to represent increased potential for malignant transformation in response to environmental stimuli, such as cigarette smoke and alcohol (56). However, the reported prevalence of aneuploidy in normal mucosa varies widely, and brings into question the histologic classification of mucosa in some studies (55,56). One of these studies demonstrated that the percentage of S-phase fraction tends to increase with degree of dysplasia and correlates with progression from premalignant tissue to invasive carcinoma (55). A relationship between poorly differentiated tumors and aneuploidy has also been shown (54). Knowledge of specific genetic alterations implicated in esophageal tumorigenesis has expanded in recent years, allowing the study of associations between molecular findings and aneuploidy. One study established that ESCCs with aneuploidy are more likely to have genetic alterations, including amplification of the genes v-myc myelocytomatosis viral oncogene homolog (MYC), cyclin D1 (CCND1), and epidermal growth factor receptor (EGFR); LOH of multiple tumor suppressor loci including RB1; and APC (58). This latter study suggested that aneuploidy not only may constitute a marker for large-scale genomic alterations but also may be associated with specific genetic abnormalities important in esophageal tumorigenesis.

There is a preponderance of published data associating aneuploidy with advanced stages of disease and a poor prognosis (57,58,59). However, a role for the testing of DNA ploidy in the clinical management of patients with ESCC remains to be established.

Cellular subpopulations with aneuploidy or increased tetraploid (4N) DNA content occur within more than 90% to 95% of EAC, arise in premalignant epithelium, and appear to predict progression (60,61,62,63). One early study suggested that alterations in ploidy correlated with dysplasia in BE, although the histologic classification of dysplasia in this report was controversial (60). In another series of patients, the presence of aneuploid cells as detected by flow cytometric analyses of histologically equivocal biopsy specimens allowed for the identification of mildly dysplastic areas. Furthermore, aneuploidy was always associated with some morphologic abnormality, varying from mild dysplasia to frank carcinoma (64). Others have found that aneuploidy and dysplasia can be discordant (31,65). In a seminal study involving DNA ploidy analysis of mapped Barrett epithelium, clonal growth similar to that seen in fully developed cancer was present in metaplastic Barrett mucosa (66). Finally, in a prospective evaluation of patients with BE who were enrolled in a rigorous endoscopic biopsy protocol, patients with neither aneuploidy nor increased 4N fractions had a 0% 5-year cumulative cancer incidence. In comparison, patients with increased 4N fractions had a 56% 5-year cancer incidence, and patients with aneuploidy had a 43% 5-year cancer incidence (67).

In summary, although the results of certain studies indicate that flow cytometry appears to be useful in detecting a subset of patients who do not have high-grade dysplasia and yet show an increased risk of progression to EAC that cannot be identified by dysplasia grade (68), no precisely defined role for the determination of tetraploidy or aneuploidy in clinical practice has yet been proven for either EAC or ESCC.