Ectopic Expression of Guanylyl Cyclase C in Adenocarcinomas of the Esophagus and Stomach¹

GC-C³ is a surface receptor specifically expressed by epithelial cells of the intestine, from the duodenum to the rectum, in normal adult humans (1). Of significance, GC-C expression persists after intestinal epithelial cells undergo neoplastic transformation into colorectal adenocarcinoma cells (2). Indeed, all primary and metastatic colorectal tumors, but no extra-GI tissues or tumors, examined thus far express GC-C (2). GC-C expression, detected by RT-PCR, is more sensitive than histopathology for detecting micrometastatic disease in lymph nodes of patients undergoing staging for colorectal cancer (3). Detection of micrometastases in lymph nodes by GC-C RT-PCR is associated with an increased risk of mortality from recurrent disease (3). In addition, GC-C RT-PCR detects circulating colorectal cancer cells in blood and may represent a sensitive and specific marker for detecting disease recurrence earlier than current modalities (2, 4, 5).

Tissue-specific expression of GC-C is activated, in part, by CDX2, a member of the homeodomain family of transcription regulatory proteins specifically expressed in intestine (6). In addition to GC-C, CDX2 activates the expression of several proteins that are characteristic of the differentiated intestinal enterocyte phenotype including SI and lactase (7, 8). GC-C and CDX2 exhibit a parallel pattern of expression in normal adult humans limited to the duodenum, intestine, and rectum (1, 7, 9, 10). Of significance, development of adenocarcinomas of the upper GI tract is typically preceded by a process of metaplasia involving transformation to an intestinal phenotype associated with the expression of intestinal-specific proteins (11). Interestingly, CDX2 expression is absent from normal gastric mucosa but present in vitro in cell lines derived from gastric adenocarcinomas (12).

The expression of intestinal-specific genes in intestinal metaplasia and adenocarcinomas of the upper GI tract suggests that intestinal-specific markers such as GC-C may be ectopically expressed by adenocarcinomas of the esophagus and stomach. Previous studies demonstrated that GC-C is not expressed in normal epithelia of the esophagus and stomach (2, 10). The present study examined the ectopic expression of GC-C in adenocarcinomas of the esophagus and stomach. Detection of GC-C in primary and metastatic esophageal and gastric adenocarcinomas suggests the potential utility of this marker in the management of patients undergoing diagnosis, staging, and postoperative surveillance for these malignancies.

T84 and Caco2 human colon carcinoma cells (American Type Culture Collection, Manassas, VA) were used as a positive control for GC-C mRNA in RT-PCR analyses (13). Cells were grown in a 50:50 mixture of Ham’s F12 medium and DMEM (DMEM/F12; Mediatech, Inc., Herndon, VA) supplemented with 10% fetal bovine serum (BioWhittaker, Walkersville, MD), 10,000 units/ml penicillin, and 10,000 μg/ml streptomycin (BioWhittaker). Cells were maintained at 37°C in a humidified 5% CO₂/95% air atmosphere.

Eighteen patients with either esophageal, gastric, or GE junction adenocarcinomas were included in this study (Table 1). The patients included 13 males and 5 females ranging in age from 47 to 84 years. All tumors were identified as adenocarcinomas, and all tissues adjacent to tumors were free of disease by histopathology. Surgical specimens were obtained from the CHTN Eastern Division (Philadelphia, PA) and TJU-PATH (Philadelphia, PA). The National Cancer Institute funds CHTN, and, consequently, other investigators may have received specimens from the same subjects. All specimens were immediately frozen in liquid nitrogen and stored at −80°C within 2 h of resection. For extraction of total RNA, tissues were pulverized by mortar and pestle and then processed by RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. The concentration and purity of samples were determined by UV spectrophotometry.

The expression of specific RNA transcripts was determined by a two-step RT-PCR protocol using transcript-specific primer sets that flank exon-intron boundaries. The primer sequences employed were (5′ to 3′): GC-C sense, GTTTCCTATTTCTCCCACGAACTC; GC-C antisense, TTTCTTGGTGTCCACAGAGGTA; CDX2 sense, CCCGGCGGCCAGCGGCGGAACCTGT; CDX2 antisense, TATTTGTCTTTCGTCCTGGTTTTCA; β-actin sense, ATCCTCACCCTGAAGTACCC; and β-actin antisense, CAGCCTGGATAGCAACGTAC (3, 14). The predicted sizes of the amplicons are as follows: GC-C first round, 533 bp; CDX2, 95 bp; and β-actin, 226 bp.

Reverse transcription of 1.0 μg of total RNA was performed with 0.25 unit/μl avian myeloblastosis virus reverse transcriptase (Panvera; Madison, WI) and 10 mm Tris-HCl (pH 8.3); 50 mm KCl; 5 mm MgCl₂; 1 unit/μl Ribonuclease III Inhibitor (Panvera); 1 mm each of dATP, dCTP, dGTP, dTTP (Promega, Madison, WI); and 1 μm of the appropriate antisense primer in a total volume of 20 μl. This reaction was incubated at 50°C for 30 min, 95°C for 5 min, and 25°C for 5 min. For PCR, this entire cDNA product (20 μl) was combined with a reaction mix resulting in a final concentration of 0.05 unit/μl recombinant Taq polymerase (Panvera) and 10 mm Tris-HCl (pH 8.3), 50 mm KCl, 2.0 or 3.0 mm MgCl₂ (for β-actin or GC-C, respectively), and 0.2 μm of each of the appropriate sense and antisense primers in a final volume of 50 μl. Primers were designed to avoid possible interference by genomic DNA contamination by amplifying a product that spanned at least one exon-intron junction. Thermal cycling conditions for GC-C were 95°C for 2 min, 1 cycle; 94°C for 30 s, 60°C for 30 s, and 72°C for 1 min, 40 cycles; and 72°C for 5 min and 25°C for 5 min, 1 cycle. For β-actin, the thermal cycling conditions were 95°C for 2 min, 1 cycle; 94°C for 30 s, 58°C for 30 s, and 72°C for 1 min, 25 cycles; and 72°C for 5 min and 25°C for 5 min, 1 cycle.

A second round of GC-C PCR amplification utilizing primers internal to the initial amplification specifically confirmed the presence or absence of GC-C amplicons (3). This nested reaction contained 2% (1 μl) of the initial PCR reaction, 0.05 unit/μl recombinant Taq polymerase, 10 mm Tris-HCl (pH 8.3), 50 mm KCl, 2.0 mm MgCl₂, and 0.2 μm of each of the GC-C sense (GGACCACAACAGGAAAAGCAATG) and antisense (AGGCAAGACGAAAGTCTCGTTT) primers in a final volume of 50 μl. Thermal cycling conditions for nested PCR were 95°C for 2 min, 1 cycle; 94°C for 30 s, 60°C for 30 s, and 72°C for 1 min, 20 cycles; and 72°C for 5 min and 25°C for 5 min, 1 cycle. The predicted size of the amplicon from GC-C nested PCR is 262 bp.

Amplicons were visualized after electrophoretic separation in a 2% Nusieve 3:1 agarose gel (BioWhittaker Molecular Applications, Rockland, Maine) containing 0.5 μg/ml ethidium bromide (Fisher Scientific, Pittsburgh, PA).

Esophageal adenocarcinomas and adjacent surgical margins were obtained from four men and two women whose ages ranged from 56 to 81 years (Table 1). Their primary tumors exhibited varying degrees of differentiation by histopathology. Two lymph nodes that were free of disease by histopathology were obtained during a resection of an esophageal adenocarcinoma (patient 4). Gastric adenocarcinomas and surgical margins were obtained from six men and three women whose ages ranged from 47 to 73 years (Table 1). Their primary tumors exhibited varying degrees of differentiation by histopathology. One lymph node (patient 16) secondary to gastric adenocarcinoma contained cancer cells identifiable by histopathology. GE junction adenocarcinomas and associated surgical margins were obtained from two men whose ages were 58 and 84 years (Table 1). Patient 18 alone underwent chemoradiotherapy before surgery, and postsurgical examination revealed fibrosis at the site of the primary tumor with no evidence of residual adenocarcinoma. From this same patient, 11 lymph nodes that were free of disease by histopathology were obtained during resection of the tumor. It remains unknown whether pretreatment with chemoradiotherapy alters the expression of GC-C or CDX2 in upper GI malignancies.

Samples of primary esophageal and gastric adenocarcinomas and their adjacent margins were surveyed for ectopic expression of CDX2 and GC-C. Previous studies demonstrated that normal esophagus and stomach do not express CDX2 or GC-C, whereas CDX2 is expressed in cell lines derived from gastric adenocarcinomas (2, 9, 10, 12). In the present study, adenocarcinomas of the esophagus and stomach expressed CDX2 and GC-C mRNA detected by RT-PCR (Fig. 1). Histologically normal tissue adjacent to an esophageal tumor did not express GC-C or CDX2 mRNA, whereas histologically normal tissue adjacent to a gastric adenocarcinoma expressed both transcripts. The intestinal-specific transcript encoding CDX1 also was detected in esophageal and gastric adenocarcinomas (data not shown). Furthermore, ectopic expression of the GC-C transcript in gastric adenocarcinomas as well as in histologically normal tissues adjacent to tumors was detected in commercially available Matched cDNA Pairs of gastric tumors and adjacent tissues (Clontech, Palo Alto, CA; data not shown). Amplicons reflect amplification of mRNA rather than contaminating DNA because primer sets used for these analyses spanned intron-exon boundaries (2, 3).

On the basis of the initial findings (Fig. 1), a larger number of esophageal and gastric adenocarcinomas was examined for expression of GC-C mRNA by nested RT-PCR (Table 2). GC-C expression was detected in all (five of five) esophageal tumors, eight of nine gastric tumors, and the single GE junction tumor evaluated.

GC-C mRNA expression also was assessed in histologically normal tissues adjacent to tumors by nested RT-PCR. In tissues adjacent to esophageal tumors, amplification revealed expression of GC-C mRNA in three of five specimens (Table 2). Similarly, in tissues adjacent to gastric tumors, amplification demonstrated GC-C mRNA expression in six of seven specimens (Table 2). In contrast, no normal gastric specimen (from patients 5–7) expressed GC-C. In tissues adjacent to a GE junction tumor, amplification detected GC-C mRNA expression in all (four of four) specimens.

Lymph nodes obtained from a patient with esophageal adenocarcinoma that were free of disease by histopathology did not express GC-C by nested RT-PCR analysis (Fig. 2). One lymph node from patient 16 with gastric adenocarcinoma that contained tumor cells by histopathology also expressed GC-C mRNA by nested RT-PCR. Of significance, 3 of 11 lymph nodes from a patient with a GE junction tumor (patient 18) that were free of disease by histopathology expressed GC-C mRNA by nested RT-PCR analysis. These data suggest the presence of micrometastases in those lymph nodes that escaped detection by routine evaluation.

The incidence of esophageal adenocarcinoma in the United States has increased by 350% since the mid-1970s (15). In contrast, the overall incidence of gastric adenocarcinoma has decreased in the United States, although the incidence of adenocarcinomas of the proximal stomach has increased (15). Whereas the combined incidence of gastric and esophageal cancer in the United States is <50,000 cases/year, these tumors are associated with a 5-year survival of <25% (16). The most significant predictor of survival for patients with these tumors is the stage of disease at the time of diagnosis. Larger and more invasive tumors, particularly those that have spread to regional lymph nodes, carry significant prognostic risk associated with higher patient mortality (17). Methods that accurately determine the extent of tumor progression, including the presence of occult micrometastases in regional lymph nodes, will be useful to more precisely define patient prognosis. Indeed, patients with micrometastases of gastric or esophageal adenocarcinoma in regional lymph nodes detected by marker-specific RT-PCR or immunohistochemistry experienced a shorter disease-free survival compared with patients without evidence of occult lymph node metastases (18). Thus, staging of patients with esophageal or gastric adenocarcinoma could be significantly improved by identifying reliable markers specifically expressed by those tumors.

Detection of ectopic GC-C mRNA expression by RT-PCR in the upper GI tract may represent a particularly sensitive and specific method for identifying primary and metastatic gastric and esophageal tumors. GC-C mRNA and protein are specifically expressed only in intestinal epithelial cells, from the duodenum to the rectum, in normal adult humans (2, 10). Of significance, this protein continues to be expressed after neoplastic transformation of those epithelial cells into colorectal cancer cells (2). Indeed, ectopic expression of GC-C mRNA in blood and lymph nodes, detected by RT-PCR, serves as a sensitive and specific marker for metastatic tumor cells in patients with colorectal cancer (2, 3, 4, 5). In contrast, GC-C is not expressed in normal human esophagus and stomach. The GC-C-specific ligand Escherichia coli heat-stable enterotoxin does not activate guanylyl cyclase activity in membranes from normal stomach (10). Also, GC-C mRNA is not expressed in normal human esophagus or stomach, as assessed by Northern analysis or RT-PCR (2, 10). In addition, CDX2, a tissue-specific transcription factor required for GC-C expression in intestine, is not expressed in normal esophagus and stomach (2, 6, 9, 10).

The present study is the first characterization of ectopic CDX2 and GC-C expression in adenocarcinomas of the esophagus and stomach. Indeed, an expanded survey of tissues confirmed that GC-C is expressed in upper GI adenocarcinomas and in some apparently histologically normal tissues adjacent to tumors (Table 2). Expression of intestinal-specific genes such as GC-C in some esophageal and gastric tissues adjacent to tumors likely reflects molecular changes associated with the process of intestinal metaplasia that precedes overt neoplastic transformation (see below). Together, these data support the suggestion that GC-C is ectopically expressed in adenocarcinomas of the esophagus and stomach and may serve as a sensitive and specific marker for those tumors.

Similarly, examination of GC-C mRNA expression in lymph nodes from a patient with an adenocarcinoma of the GE junction revealed evidence of micrometastatic disease in 3 of 11 lymph nodes that were free of tumor by histopathology. Previous studies in colon cancer patients demonstrated that GC-C expression detected by RT-PCR identified micrometastatic disease that correlated closely with the risk of developing recurrence and mortality (3). The present study suggests that detection of GC-C transcripts by RT-PCR may be a useful marker for identifying metastatic adenocarcinoma originating from the upper GI tract. This technique may have specific utility in accurately staging patients with adenocarcinoma of the esophagus and stomach.

Ectopic expression of GC-C in adenocarcinomas of the esophagus and stomach reflects intestinal metaplasia of the involved epithelium, a process characteristic of neoplastic transformation in these tissues (11, 19). Thus, during neoplastic transformation, adenocarcinomas of the esophagus and stomach progress through an intermediate stage, during which normal epithelial cells acquire the molecular and histological characteristics of differentiated intestinal enterocytes; these cells continue to express intestinal markers after progression to invasive carcinoma (19). Normal epithelium of the upper GI tract does not express intestinal-specific genes such as CDX2 and SI (7, 9, 11, 12). However, these transcripts and their associated protein products are expressed in those cells after intestinal metaplasia and progression to invasive carcinoma (11, 12). In addition, ectopic expression of CDX2 in the gastric epithelium of transgenic mice induces expression of GC-C mRNA in that tissue.⁴ Indeed, ectopic expression of GC-C in intestinal metaplasia and adenocarcinoma of the upper GI tract likely reflects ectopic expression of CDX2 (6).⁴ These data support the suggestion that GC-C and other proteins characteristic of differentiated intestinal cells that are ectopically expressed during intestinal metaplasia may serve as markers for adenocarcinomas of the upper GI tract.

It is notable that intestinal-specific transcripts, such as GC-C and CDX2, were detected in some, but not all, tissue samples adjacent to adenocarcinomas of the esophagus and stomach that were apparently free of tumor by histopathology. These observations support the hypothesis that cancers of the upper GI tract arise from within generalized zones of transformation in which molecular changes precede histopathological alterations in tissue architecture (20, 21, 22). Previous studies demonstrated that preneoplastic epithelial cells of the esophagus exhibited loss of heterozygosity in the absence of overt neoplastic transformation assessed by histopathology (20). Similarly, fields of aberrant CpG island hypermethylation have been identified in preneoplastic and neoplastic esophagus (21). In addition, abnormal expression of FHIT precedes neoplastic transformation in the esophagus (22). Finally, genetic alterations in histologically normal tissues adjacent to tumors have been detected in gastric adenocarcinomas (23). These data suggest that expression of intestinal-specific transcripts, such as GC-C, detected by RT-PCR analysis in tissues adjacent to tumors that are free of disease by histopathology may reflect early neoplastic changes in patients at high risk for developing adenocarcinoma of the upper GI tract.

In summary, the present study demonstrates that GC-C transcripts are ectopically expressed in adenocarcinoma of the esophagus and stomach, likely reflecting the process of intestinal metaplasia characteristic of tumor progression at these sites. This study is the first to characterize the expression of GC-C in primary tumors, adjacent tissues, and lymph nodes from patients with esophageal and gastric adenocarcinomas. This suggests that GC-C mRNA detected by RT-PCR may have utility as a marker for identifying primary and metastatic tumors of the upper GI tract. In addition, this transcript may be useful for identifying the earliest stages of neoplastic transformation in these tissues. Furthermore, other intestinal-specific transcripts, including SI, are ectopically expressed at the protein level in those tumors (11). Thus, GC-C protein also may be expressed in adenocarcinomas of esophagus and stomach. This protein could serve as a useful target for novel diagnostic and therapeutic agents delivered by specific ligands for this receptor, including E. coli heat-stable enterotoxin. The expression of functional GC-C in adenocarcinomas of the esophagus and stomach, and its utility as a diagnostic marker and therapeutic target in those tumors are currently being explored.