• polyomavirus;
  • oncogenesis;
  • catenins;
  • T antigen;
  • agnoprotein


  1. Top of page
  2. Abstract
  6. Acknowledgements


The human polyomavirus JC virus (JCV) causes progressive multifocal leukoencephalopathy. Subclinical infection with JCV occurs in 85–90% of the population worldwide. The virus usually remains latent but can reactivate under immunosuppressive conditions, resulting in progressive multifocal leukoencephalopathy. JCV is oncogenic in experimental animals and is associated with human brain tumors. JCV is found in normal mucosa of the gastrointestinal tract, and some colon carcinomas express the oncogenic JCV T-antigen protein. The objective of this study was to examine the presence of JCV DNA sequences and JCV protein expression in normal and malignant human esophageal tissues.


The authors examined the presence of JCV DNA sequences and protein expression in normal and malignant human esophageal tissues. Seventy well characterized biopsy specimens from patients with a spectrum of esophageal disorders were studied by immunohistochemistry, and 18 specimens were analyzed further by polymerase chain reaction amplification.


JC viral DNA was isolated from 11 of 13 normal esophageal biopsy specimens (85%) and from 5 of 5 esophageal carcinomas (100%). Using immunohistochemistry, JCV T antigen was detected in 10 of 19 carcinomas (53%), agnoprotein was detected in 8 carcinomas (42%), p53 tumor suppressor was detected in 11 carcinomas (58%), and β-catenin was detected in 4 carcinomas (21%). Zero of 51 normal, benign, and premalignant esophageal samples expressed viral proteins. Laser-capture microdissection verified the presence and specificity of JCV DNA sequences. β-Catenin and p53 were colocalized with JCV T-antigen in the nuclei of neoplastic cells.


The results provide evidence for infection of gastrointestinal tract cells by JCV and suggest a potential role of JCV in the development of upper digestive tract carcinomas. Cancer 2005. © 2005 American Cancer Society.

In industrialized countries, the incidence of adenocarcinoma of the esophagus has risen dramatically in the last decades at a rate of approximately 5–10% per year. The epidemiologic features among patients with esophageal adenocarcinomas are similar to those among patients with intestinal metaplasia of the distal esophagus, also known as Barrett esophagus, which has been identified as the single most important risk factor for the development of esophageal adenocarcinomas.1 Although chronic gastroesophageal reflux is the usual underlying cause of the repetitive mucosal injury and also provides an abnormal environment during the healing process that predisposes to intestinal metaplasia, the etiology of adenocarcinomas of the esophagus remains unclear. Factors predisposing to the development of Barrett esophagus and subsequent adenocarcinoma include a markedly increased exposure to gastric contents due to defective barrier function of the lower esophageal sphincter.1 Molecular changes associated with the development of esophageal tumors include mutations in TP53,2, 3 APC,2, 3 and β-catenin4 and changes in the level of Ki-67 expression.5

JC virus (JCV) is a human polyomavirus that was isolated first from the brain of a patient suffering from progressive multifocal leukoencephalopathy (PML).6 Polyomaviruses are a subfamily of nonenveloped DNA viruses with icosahedral capsids that contain small, circular, double-stranded DNA genomes.7 It is known that JCV is the etiologic infectious agent of PML, a fatal demyelinating disease of the central nervous system. The virus is widespread throughout the population, with > 80% of adults exhibiting JCV-specific antibodies. It is believed that JCV infection takes place during early childhood and usually is subclinical. However, under conditions of immunosuppression, e.g., in patients with acquired immunodeficiency syndrome, JCV can emerge from latency to cause PML.8 JCV can transform cells in culture and is oncogenic in laboratory animals.9 JCV DNA sequences have been detected in several kinds of human malignancies, including brain tumors of glial origin,10, 11 medulloblastomas,12 and colon carcinoma.13 The role of JCV in the development of human malignancies has been reviewed recently.9

The double-stranded circular DNA of JCV contains three functional regions: the early and late coding genes and the noncoding regulatory sequence.7 The early region encodes the large T- and small t-antigens, whereas the late region encodes the viral capsid proteins (VP1, VP2, and VP3) and a small regulatory protein known as agnoprotein, which is encoded near the 5′ end of the primary late transcript. A major function of T-antigen is to modulate cellular signaling pathways, which induces cells to enter the S-phase of the cell cycle. This deregulation of cell-cycle control accounts for the ability of T-antigen to transform cells and is accomplished by its interaction with cell-regulatory proteins, including tumor suppressors. JCV T- antigen is able to bind and inactivate the retinoblastoma gene product pRb14, 15 and p53.15, 16 It has also been found that the Wnt signaling pathways is disrupted by the binding of JCV T-antigen to β-catenin.13, 17

The mechanism of human-to-human transmission of JCV has not been established firmly, but the presence of infectious JCV in raw sewage suggests that ingestion of contaminated water or food may represent a possible portal of entrance of JCV into the human population.18–20 Earlier studies revealed the presence of JCV DNA sequences in a high percentage of normal tissue samples taken from the upper and lower human gastrointestinal tract.21–23 JCV DNA sequences also have been associated with colon carcinoma.13, 21, 23 We recently reported the presence of JCV DNA sequences in 22 of 27 well characterized epithelial malignant tumors of the large intestine.13 In 17 of those 27 samples, JCV T-antigen expression was detected by immunohistochemistry, and 12 of those 17 tumors also were positive for agnoprotein expression. In the current study, we examine the presence of JCV DNA sequences and JCV protein expression in human esophageal tissues using a collection of 70 well characterized biopsy samples from patients who had a spectrum of esophageal disorders. Evidence is presented for a potential role of JCV in the development of adenocarcinomas of the upper digestive tract.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Clinical Samples

For this study, in total, 70 biopsy samples from patients with specific esophageal disorders were collected from the pathology archives of Temple University Hospital (Philadelphia, PA). Clinically, the patients were suffering from a variety of esophageal disorders, including achalasia, reflux esophagitis, Barrett esophagus, Barrett esophagus with dysplasia, adenocarcinoma of the esophagus, and squamous cell carcinoma of the esophagus. Table 1 summarizes the diagnoses and clinical data from the patients.

Table 1. Clinical, Polymerase Chain Reaction, and Immunohistochemical Data in Esophageal Disorders
Diagnosis/patient no.Age (yrs)GenderPCR/Southern blotImmunohistochemistry
PepsVPsAgnoCRJCV T-antigenp53agnoprotein
  1. EGD: esophageal disorder; Peps: T-antigen region; VPs: VP1 region; Agno: agnogene region; CR: control region; JCV: JC virus; ND: not done (due to insufficient material); GERD: gastroesophageal reflux disease; GI: gastrointestinal.

  2. a For immunohistochemistry, − indicates negative, + indicates 1–30% positive cells, ++ indicates 31–60 positive cells, and +++ indicates > 61% positive cells.

Normal tissues         
 1 Dyspepsia50MaleNDNDNDND
 2 Heartburn66FemaleNDNDNDND
 3 GERD syndrome48FemaleNDNDNDND
 4 Dysphagia55FemaleNDNDNDND
 5 GERD syndrome69MaleNDNDNDND
 6 GERD syndrome49FemaleNDNDNDND
 7 GERD syndrome56Female+++
 8 GERD syndrome75FemaleNDNDNDND
 9 Dysphagia, GERD syndrome34FemaleNDNDNDND
 10 Heartburn55MaleNDNDNDND
 11 Dysphagia51FemaleNDNDNDND
 12 Dysphagia85MaleNDNDNDND
 13 Heartburn54MaleNDNDNDND
 14 Dysphagia64FemaleNDNDNDND
 15 Chest pain61Male+
 16 Dysphagia38MaleNDNDNDND
 17 Dysphagia37FemaleNDNDNDND
 18 Chest pain54MaleNDNDNDND
 19 Dysphagia82FemaleNDNDNDND
 20 Chest pain52Female
 21 Dysphagia40FemaleNDNDNDND
 22 Heartburn56MaleNDNDNDND
Reflux esophagitis         
 23 Heartburn59FemaleNDNDNDND
 24 GERD syndrome57FemaleNDNDNDND
 25 Dysphagia72FemaleNDNDNDND
 26 Heartburn59FemaleNDNDNDND
 27 Heartburn71FemaleNDNDNDND
 28 GERD syndrome40FemaleNDNDNDND
 29 GERD syndrome37FemaleNDNDNDND
 30 Dysphagia, GERD syndrome70Female
 31 Heartburn16MaleNDNDNDND
Barrett esophagus, no dysplasia         
 32 Barrett surveillance75MaleNDNDNDND
 33 Barrett surveillance36MaleNDNDNDND
 34 Barrett surveillance78Male+
 35 Barrett surveillance70FemaleNDNDNDND
 36 Barrett surveillance65Male++
 37 Barrett surveillance61MaleNDNDNDND
 38 Barrett surveillance61MaleNDNDNDND
 39 Barrett surveillance62Female+
 40 Barrett surveillance80MaleNDNDNDND
 41 Barrett surveillance58FemaleNDNDNDND
Barrett esophagus with dysplasia         
 42 Barrett surveillance52Female+
 43 Barrett surveillance84Male++
 44 Barrett surveillance78FemaleNDNDNDND
 45 Barrett surveillance52MaleNDNDNDND
 46 Barrett surveillance59Male+
 47 Barrett surveillance79Male++++
 48 Barrett surveillance57Male+
 49 Barrett surveillance77MaleNDNDNDND
 50 Barrett surveillance56MaleNDNDNDND++
 51 Barrett surveillance29Male++
 52 Esophageal mass53MaleNDNDNDND
 53 Esophageal mass44MaleNDNDNDND
 54 Esophageal mass59MaleNDNDNDND+++
 55 Esophageal mass28MaleNDNDNDND
 56 Upper GI bleeding85MaleNDNDNDND++
 57 Esophageal mass60MaleNDNDNDND++++++++
 58 Dysphagia75Female++
 59 Upper GI bleeding79MaleNDNDNDND+++++++
 60 Dysphagia47Male+
Squamous cell carcinoma         
 61 Dysphagia55Male+++++++
 62 Esophageal mass47MaleNDNDNDND++++++++
 63 Esophageal mass63FemaleNDNDNDND+++++++
 64 Dysphagia65MaleNDNDNDND
 65 Esophageal mass59Female++++
 66 Reflux symptoms73MaleNDNDNDND++++++
 67 Esophageal mass49Male+++
 68 Dysphagia38MaleNDNDNDND+++++++
 69 Dysphagia66MaleNDNDNDND++++++
 70 Dysphagia, esophageal mass54FemaleNDNDNDND+++++

DNA Extraction, Amplification, and Hybridization

DNA extraction, polymerase chain reaction (PCR) amplification, and Southern blot hybridization were performed essentially as described previously.10 DNA was extracted from several sections of 10 μm from each the tissue sample using the QIAamp tissue kit according to the manufacturer's instructions (Qiagen, Valencia, CA). PCR amplification was performed on DNA extracted from the tumors using 4 individual sets of primers: Pep1 and Pep2, which amplify sequences in the N-terminal region of the T-antigen of JCV (nucleotides 4255–4274 and 4408–4427, respectively); VP2 and VP3 to amplify a region of the JCV VP1 capsid protein (nucleotides 1828–1848 and 2019–2039, respectively); Agno1 and Agno2 to amplify a sequence within the coding region of the JCV agnogene (nucleotides 279–298 and 438–458, respectively); and CR1 and CR2 to amplify a portion of the noncoding control region of JCV (nucleotides 4986–5006 and 238–258, respectively). The primers CR1 and CR2 for the noncoding controls lie outside the region that is rearranged between JCV variants and would be expected to amplify the DNA of all JCV genotypes. Amplification was carried out on 500 ng of template DNA with Failsafe Taq polymerase in Failsafe buffer B (Epicenter) in a total volume of 50 μL in the presence of 500 nM of primers. After denaturation at 95 °C for 10 minutes, 45 cycles of denaturation at 95 °C for 15 seconds, annealing for 30 seconds, and extension at 72 °C for 30 seconds, a final extension step of 72 °C for 7 minutes was performed for termination. Annealing temperatures were 55 °C for Pep primers, 54 °C for VP primers, 57 °C for Agno primers, and 52 °C for CR primers. Ten microliters of the PCR products were separated by 2% agarose gel electrophoresis, depurinated, denatured, neutralized, and transferred to nylon membranes (Hybond-N; Amersham). The membranes were hybridized with 1 × 106 cpm/mL of γ-[32P] end-labeled oligonucleotide probe overnight, followed by washing and autoradiography, as described previously.24 Oligonucleotides homologous to the following JCV-specific sequences were utilized as probes: JC probe (Pep primers; nucleotides 4303–4327), VP probe (nucleotides 1872–1891), Agno probe (nucleotides 425–445), and CR probe (nucleotides 62–81).

Histologic and Immunohistochemical Analysis

The formalin fixed, paraffin embedded tissue specimens were sectioned at a thickness of 4 μm and were stained with hematoxylin and eosin for routine histologic characterization. Immunohistochemistry was performed using the avidin-biotin-peroxidase complex system according to the manufacturer's instructions (Vectastain Elite ABC peroxidase kit; Vector Laboratories, Burlingame, CA) Briefly, our protocol includes deparaffinization in xylenes; rehydration through alcohols; nonenzymatic antigen retrieval performed in 0.01 M sodium citrate buffer, pH 6.0, for 40 minutes at 95 °C; endogenous peroxidase quenching in MeOH/3% H2O2 for 20 minutes; and blocking in 5% normal horse serum for 2 hours at room temperature. Primary antibodies against viral and cellular proteins were incubated overnight at room temperature in a humidified chamber. The primary antibodies used in this study included a rabbit polyclonal anti-JCV agnoprotein (1:2000 dilution), which was described previously11; a rabbit polyclonal anti-JCV VP1 (1:1000 dilution; kindly provided by Dr. Walter Atwood, Brown University); a mouse monoclonal anti-simian virus T-antigen that cross reacts with JCV T-antigen (1:100 dilution; clone pAb416; Oncogene Science); a mouse monoclonal anti-p53 (1:100 dilution; clone DO-7; DAKO); a mouse monoclonal anti-β-catenin (1:100 dilution; clone E-5; Santa Cruz), a goat polyclonal anti-TCF-4 (1:300 dilution; N-20; Santa Cruz); and a goat polyclonal anti-LEF-1 (1:250 dilution; C-19; Santa Cruz). After an incubation with primary antibodies, biotinylated secondary anti-mouse or anti-rabbit antibodies were incubated for 1 hour at room temperature followed by incubation with avidin-biotin complex for 1 hour and were developed with diaminobenzidine. Slides were counterstained with hematoxylin, dehydrated, cleared in xylene, and mounted with Permount (Sigma Laboratories).

Laser-Capture Microdissection and DNA Extraction

Representative sections of tissue samples were labeled by immunohistochemistry, as described above. Slides were dehydrated, cleared with xylene, and air dried for 24 hours. Next, laser capture was performed under direct microscopic visualization of the immunolabeled areas by laser activation of thermoplastic film mounted on optically transparent CapSure HS laser-capture microdissection (LCM) caps (Arcturus Engineering). The PixCell II LCM System (Arcturus) was set to the following parameters: 15 μm spot size, 40 mW power, and 3.0 msec duration. A total of 50 neoplastic cells per category were captured by focal melting of the membrane through a carbon dioxide laser-pulse activation. DNA isolation was performed using the Arcturus PicoPure DNA extraction kit according to the manufacturer's instructions (Arcturus). PCR amplification and analyses were performed as described above.

Double-Labeling Immunofluorescence

Deparaffinization, antigen retrieval, endogenous peroxidase quenching, and blocking of paraffin-embedded sections were performed as described above. Sections were then incubated in either the mouse anti-T-antigen antibody (1:100 dilution) or in the mouse anti-β-catenin antibody (1:100 dilution) overnight at room temperature; this was followed by incubation in antimouse fluorescein antibody for 2 hours. Next, sections were washed, blocked again, and incubated with either the mouse anti-p53 antibody (1:100 dilution) or the mouse anti-T-antigen antibody (1:100 dilution) overnight followed by incubation in anti-mouse rhodamine antibody. Sections were washed and mounted in Vectashield aqueous mounting media and were visualized by fluorescence microscopy. Black-and-white images were acquired, pseudocolored, and overlaid using IPLab software.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Samples were categorized clinically and histopathologically into seven groups. The first group comprised biopsies from normal esophagus (10 samples). The second group comprised biopsies from patients suffering from achalasia (12 samples). The third group was from patients suffering from reflux esophagitis (9 samples). The fourth group was from patients with Barrett esophagus (10 samples). The fifth group was from patients with Barrett esophagus exhibiting mild or moderate dysplasia (10 samples). The sixth group was from patients with adenocarcinoma (9 samples), and the seventh group was from patients with squamous cell carcinoma (10 samples). Histologically, the samples from patients with achalasia were characterized by a stratified epithelium with no alterations. Mucosal breaks were the diagnostic feature of acute and/or chronic reflux esophagitis.25 The abnormal presence of gastric glandular epithelium in the lower one-third of the esophagus made the diagnosis of Barrett esophagus (Fig. 1A); intestinal metaplasia was characterized by the presence of abundant goblet cells within these epithelium (Fig. 1B). The dramatic reduction of goblet cells and nuclear atypia and pleomorphism in cells confined to the glandular structures was the criteria for Barrett esophagus with dysplasia (Fig. 1C). Finally, carcinomas were classified based on marked nuclear atypia and cellular pleomorphism with glandular cells invading the adjacent stroma (Fig. 1D). The characterization of the tumors was performed according to the World Health Organization Classification of Tumours of the Digestive System.26

thumbnail image

Figure 1. Histologic evaluation of esophageal clinical samples. (A) An area of normal stratified esophageal epithelium and the transitional area of glandular metaplasia that is characteristic of Barrett esophagus (hematoxylin and eosin [H&E] stain). (B) A glandular area of Barrett esophagus that exhibits gastric metaplasia (H&E stain). (C) In some samples, the Barrett epithelium shows loss of goblet cells and marked nuclear atypia in cells that still are confined to the gland, characteristic of dysplasia (H&E stain). (D) Samples of adenocarcinoma are characterized by numerous pleomorphic glandular cells with marked nuclear atypia and numerous mitotic figures (H&E stain). Original magnification ×100 (A); ×200 (C); ×400 (B,D).

Download figure to PowerPoint

The presence of JCV DNA sequence in esophageal biopsy samples was examined by performing PCR on DNA extracted from the sample followed by Southern blot hybridization, as described above (see Materials and Methods). Four regions of the JCV genome were analyzed, and these were an N-terminal region of the JCV T-antigen coding sequence, a region of the VP1 capsid protein coding sequence, a sequence within the coding region of the agnogene, and a portion of the noncoding control region of JCV near the origin of viral DNA replication. The relative positions of these regions on the JCV genome are shown in Figure 2A. In each sample, the oligonucleotides that were used as labeled probes in the Southern blot were nested internally with respect to the primers that were used in the PCR reaction to prevent hybridization to nonspecifically amplified DNA product.

thumbnail image

Figure 2. Polymerase chain reaction (PCR) amplification of JC virus (JCV) DNA sequences in esophageal disorders. (A) This schematic representation of the JCV genome shows the early and late coding regions separated by the noncoding regulatory region. The location and nucleotide positions are shown for the primers used for PCR amplification of the N-terminal of the T-antigen-encoding region, the VP-1 encoding region, the accessory protein agnoprotein encoding region, the noncoding regulatory region (outside numbers), and the probes used for Southern blot analysis (inner underlined numbers). Nucleotides are numbered relative to the Mad-1 reference strain (GenBank no. J02226). Mad-1 is the etiologic agent of progressive multifocal leukoencephalopathy (PML) and has a regulatory region that is characterized by 2 98-base pair (bp) repeats and by the absence of 23 bp and 66 bp that are present in the archetype. (B) The results are shown from the PCR/Southern blot amplification with the different JCV-specific primers. Numbers on top of the blots correspond to the sample numbers in Table 1. CR: control region; T-Ag: T-antigen region; VPs: VP1 region; Agno: agnogene region.

Download figure to PowerPoint

Southern blots for PCR products that were amplified from DNA extracted from biopsy samples for each of the 4 regions of the JCV genome are shown in Figure 2B. JCV DNA sequence could be detected from at least 1 region in 11 of the 13 samples (85%) from normal or benign esophagus that were examined (Table 1). In the samples of adenocarcinoma and squamous cell carcinoma that were examined, 5 of 5 samples (100%) were positive. The noncoding control region was detected in only 1 sample (Sample 47) and was the same size as the Mad-1-positive control, although its genotype was not determined by sequencing.

LCM was performed in 3 different regions of a sample of esophageal adenocarcinoma (Sample 57) and in a sample of squamous cell carcinoma (Sample 62) that were immunolabeled for T-antigen (Fig. 3). In the first region, approximately 50 cells from the normal stratified epithelium of the esophagus were dissected (Fig. 3A). In the second region, 25 glandular structures of an area of Barrett esophagus that was negative for T-antigen were microdissected (Fig. 3B). In the third area, 25 neoplastic glands on a region of invasive adenocarcinoma that exhibited nuclear immunoreactivity for T-antigen were selected (Fig. 3C). In the second sample, a squamous cell carcinoma, 50 neoplastic single cells were selected and microdissected (Fig. 3D). Southern blot analysis of the PCR-amplified regions demonstrated the presence of DNA fragments of the T-antigen and agnoprotein coding regions in the four areas selected (normal epithelium, Barrett esophagus, adenoma, and squamous cell carcinoma).

thumbnail image

Figure 3. Laser-capture microdissection and polymerase chain reaction PCR amplification in esophageal cancer. (A1–A3) Patient 57 had adenocarcinoma that originated in Barrett esophagus and was immunolabeled for T-antigen. Normal cells were microdissected. (B1–B3) Fifty glandular structures from an area of Barrett esophagus with intestinal metaplasia also were also collected. (C1–C3) Finally, in an area of invasive adenocarcinoma, 50 neoplastic glands with T-antigen-positive nuclei were selected and microdissected. (D1–D3) In addition, 50 separate neoplastic cells that also exhibited nuclear expression of T antigen were selected and microdissected from Patient 62 (squamous cell carcinoma). Southern blots from all the areas with primers specific for the T-antigen (T-Ag) and agnoprotein (Agno) regions demonstrate the presence of JC virus (JCV) DNA sequences in all microdissected areas. Original magnification ×10 (B1–B3); ×400 (A1–A3; C1–C3; D1–D3).

Download figure to PowerPoint

Figure 4 shows immunohistochemistry experiments that were performed to determine the presence and localization of the JCV T-antigen protein in the different clinical samples. There was a complete lack of immune reaction in samples of benign origin, i.e., normal esophagus, achalasia (Fig. 4A), reflux esophagitis, Barrett esophagus (Fig. 4B), and Barrett esophagus with dysplasia. However, we found robust immunolabeling in the nuclei of neoplastic cells in 10 of 19 carcinoma samples studied (53%) (Tables 1, 2). Figure 4C,D shows a T-antigen-positive adenocarcinoma at low and high magnification, respectively.

thumbnail image

Figure 4. Immunohistochemical detection of JC virus (JCV) T-antigen in esophageal carcinoma. (A) Normal stratified epithelium demonstrated no expression of T- antigen in samples of normal esophagus and achalasia (a representative section is shown in A). (B) Areas of Barrett esophagus were negative when tested with a T-antigen antibody. (C,D) Samples of squamous cell carcinoma (C) and adenocarcinoma (D) demonstrated robust immunoreactivity for T-antigen in the nuclei of most (but not all) neoplastic cells.

Download figure to PowerPoint

Table 2. Immunohistochemical Expression of JC Virus Proteins and Colocalization of T-Antigen and p53 in Patients with Esophageal Disorders
DiagnosisTotal no. of patientsNo. of patients (%)
T antigenp53T antigen/p53Agnoprotein
Normal epithelium100 (0.0)0 (0.0)0 (0.0)0 (0.0)
Achalasia120 (0.0)0 (0.0)0 (0.0)0 (0.0)
Reflux esophagitis90 (0.0)0 (0.0)0 (0.0)0 (0.0)
Barrett esophagus100 (0.0)0 (0.0)0 (0.0)0 (0.0)
Barrett esophagus with dysplasia100 (0.0)0 (0.0)0 (0.0)0 (0.0)
Adenocarcinoma92 (22.2)4 (44.4)2 (22.2)2 (22.2)
Squamous cell carcinoma108 (80.0)7 (70.0)7 (70.0)6 (60.0)

The tumor suppressor p53 is mutated in many malignancies27 and also can be inactivated by binding to JCV large T-antigen.15, 16 We examined the presence of p53 in the different clinical samples with an antibody that recognizes both wild-type and mutant p53. No p53 expression was observed in any of the samples of normal esophagus, reflux esophagitis, or Barrett esophagus with the exception of one sample of dysplastic Barrett esophagus (Table 1). In the carcinoma samples, we found robust immunolabeling in the nuclei of neoplastic cells in 11 of 19 samples (58%), including 9 of 10 samples that were positive for T- antigen (Tables 1, 2). Immunohistochemistry for p53 demonstrated robust reactivity in the nuclei of neoplastic glandular cells in samples of adenocarcinoma that also expressed T-antigen (Fig. 5A,B). Double-labeling immunohistochemistry of adenocarcinoma demonstrated nuclear expression of T-antigen (Fig. 5D) and robust nuclear presence of p53 (Fig. 5E). Superimposition of the images demonstrates some cells with exclusive T-antigen expression (green), some cells with exclusive p53 immunoreactivity (red), and a majority of cells with colocalization of both proteins in the nuclei of neoplastic cells (yellow) (Fig. 5F).

thumbnail image

Figure 5. Immunohistochemical detection of p53 and agnoprotein in esophageal adenocarcinomas. (A,B) Immunohistochemistry for p53 in neoplastic glandular cells in samples of adenocarcinoma that also expressed T-antigen shows nuclear staining. (C) Expression of the JCV accessory protein, agnoprotein, was detected in the cytoplasm of neoplastic cells in 3 samples of esophageal adenocarcinoma. (E,F) Double-labeling immunohistochemistry in samples of adenocarcinoma demonstrate nuclear expression of T-antigen (D) (fluorescein) and robust nuclear presence of p53 (E) (rhodamine). Superimposition of the images demonstrates cells with exclusive T-antigen expression (green), cells with exclusive p53 immunoreactivity (red), and a majority of cells with colocalization of both proteins in the nuclei of neoplastic cells (yellow) (F). Original magnification ×200 (A); ×600 (B,C).

Download figure to PowerPoint

JCV Agnoprotein is a small regulatory protein that can bind to T-antigen28 and p53.29 Expression of Agnoprotein was detected in the cytoplasm in six samples of squamous cell carcinoma and in two samples of adenocarcinoma (Tables 1, 2; Fig. 5C). Earlier studies also showed that agnoprotein localized to the cytoplasm.28 Of seven agnoprotein-positive samples, all seven also were positive for T-antigen and for p53 (Table 1).

JCV T-antigen binds to β-catenin,17 and β-catenin has been found in the nuclei of some T-antigen-positive colon carcinoma cells.13 We found that 6 of 19 carcinoma samples were immunopositive for β-catenin (21%) (Table 3). Figure 6A shows immunohistochemistry with an anti-β-catenin antibody in a sample of squamous cell carcinoma that demonstrates nuclear and cytoplasmic localization of the protein in neoplastic cells. Double-labeling immunofluorescence demonstrates the nuclear and cytoplasmic presence of β-catenin in a large number of neoplastic cells (Fig. 6D). Robust signal is detected in the nuclei of the neoplastic cells with an anti-T-antigen antibody (Fig. 6E). Superimposition of the images demonstrates expression of exclusively T-antigen in some cells (red), β-catenin in some cells (green), and the colocalization of both proteins in the majority of the neoplastic cells (yellow) (Fig. 6F).

Table 3. Immunohistochemical Detection of Wnt Signaling Pathway Proteins and Correlation with JC Virus Protein Expression in Patients with Esophageal Carcinoma
Diagnosis/patient no.Immunohistochemistrya
T antigenp53agnoproteinβ-cateninTCF-4LEF-1
  • a

    Immunohistochemistry: −; indicates negative immunoreactivity, + indicates 1–30% positive cells, ++ indicates 31–60% positive cells, and +++ indicates > 61% positive cells.

Squamous cell carcinoma      
thumbnail image

Figure 6. Expression of Wnt pathway proteins and colocalization of β-catenin and T-antigen in esophageal carcinoma. (A) Immunohistochemistry with an anti-β-catenin antibody in a patient with of squamous cell carcinoma demonstrates nuclear and cytoplasmic localization of the protein in neoplastic cells. (B) TCF-4 also was detected in the nuclei and cytoplasm of neoplastic cells. (C) Anti-LEF-1 antibody shows predominant cytoplasmic immunoreactivity. (D) Double-labeling immunofluorescence demonstrates the nuclear and cytoplasmic presence of β-catenin in a large number of neoplastic cells (fluorescein). (E) Robust signal is detected in the nuclei of neoplastic cells with an anti-T-antigen antibody (rhodamine). (F) Superimposition of the images demonstrates exclusive expression of T-antigen in some cells (red), exclusive expression of β-catenin in some cells (green), and colocalization of both proteins in the majority of the neoplastic cells (yellow). Original magnification ×200 (D–F); ×400 (A–C).

Download figure to PowerPoint

TCF-4 and LEF-1 are transcription factors that are located downstream of β-catenin in the Wnt signaling pathway.30 The 6 samples of carcinoma that expressed β-catenin also expressed TCF-4, and 4 of them expressed LEF-1 (Table 3). Staining for TCF-4 was seen in the nuclei and cytoplasm (Fig. 6B), whereas LEF-1 staining was cytoplasmic (Fig. 6C). Taken together, the data from Figures 5 and 6 suggest that the Wnt signaling pathway is dysregulated in these tumor cells, similar to what has been reported for colon carcinoma cells.13 We found that all nonneoplastic samples (patients 1–51) were negative for staining for β-catenin, TCF-4, and LEF-1.


  1. Top of page
  2. Abstract
  6. Acknowledgements

In light of the evidence implicating JCV in some patients with colon carcinoma,21–23 we were interested in extending our studies to carcinoma of the esophagus. Herein, we report that JCV DNA sequences are found in normal esophagus (1 of 1 sample), in benign (4 of 6 samples) and premalignant esophageal diseases (6 of 6 samples), and in esophageal carcinoma (5 of 5 samples), but it is only in the malignant tissue that viral genes are expressed. Most individuals have been exposed to JCV, as evidenced by the presence of antibodies specific for the virus in > 80% of adults. The virus itself is present in the environment and is found in sewage.18–20 Ingestion of contaminated water or food may represent a means of transmission for JCV. It is possible that the presence of JCV in the gastrointestinal epithelium is due to direct entry of ingested virus through the apical membrane of the epithelial cells. Alternatively, it is possible that JCV may enter the gut epithelium indirectly from the bloodstream. It has been reported that JCV can infect cells in the tonsils and can spread from there by replication in lymphoid cells.31 Presumably, this is the route whereby JCV reaches the kidney, which is the major organ of JCV persistence during latency.32

Although JCV DNA sequences were found in cells from normal esophagus, benign esophageal disease, and esophageal carcinoma, the virus was active only in carcinoma cells, in which the expression of viral proteins occurred. The viral capsid proteins were not detected in any of the samples that were examined (i.e., there was no evidence of productive infection of the esophageal cells by JCV; data not shown). However, T antigen and agnoprotein were expressed in many of the cancerous samples. The oncogenic potential of these proteins is well established;7–10, 14–17 thus, it is possible that their expression contributes to the malignant phenotype of the cells. It is known that carcinoma is a multistep process whereby the normal regulation of cellular proliferation is eroded progressively. This is illustrated clearly by the colorectal adenoma-carcinoma model for the sequence of genetic changes seen during oncogenesis.33 T-antigen is a highly multifunctional protein that is able to deliver several “hits” at once by knocking out key regulators, such as pRb,14, 15 p53,15, 16 and β-catenin.17 Some of the esophageal carcinoma samples showed changes in p53 and β-catenin, as expected. JCV T-antigen binds to β-catenin13, 17 and increases the level of β-catenin in the cell due to increased protein stability. This interaction causes β-catenin to translocate to the nucleus, where it enhances the expression of genes such as c-myc and cyclin D1. These observations indicate that stabilization of β-catenin can deregulate the Wnt pathway, which signals to the downstream transcription factors TCF-4 and LEF-1. It also has been shown that JCV T-antigen deregulates the insulin-like growth factor I receptor signaling system in some tumors.34 JCV agnoprotein also may contribute to the malignancy of esophageal carcinoma. We have shown that this protein can interfere with the cell cycle28 and disrupt DNA repair.35

The ability of JCV T antigen and agnoprotein to knock out several regulators at once is likely to be important in pushing normal epithelial cells along the multistep path to malignancy. It is not known whether JCV expression of T antigen and agnoprotein is an early or late step in tumorigenesis. First, it is possible that expression occurs at an early stage and that the genetic instability caused by T-antigen induces further pathologic changes.36 Alternatively, other early abnormal cellular changes may occur first that precede T-antigen and agnoprotein expression. For example, conceivably, there may be a mutation in the structure of one of the cellular transcription factors or a change in the expression level of one of the cellular transcription factors that bind to the control region of JCV.37 This may lead to a transcriptional activation of latent JCV during the course of tumorigenesis, leading to the onset of T-antigen production and, subsequently to a more malignant phenotype. The observation that T- antigen and agnoprotein were not expressed in any of the patients who had Barrett esophagus with dysplasia (Table 1) argues that JCV expression may occur relatively late in neoplastic progression. Similarly, the metaplastic area shown in Figure 3B that was microdissected was negative for T-antigen expression (but positive for T-antigen DNA), whereas the invasive adenocarcinoma cells in the same sample had become positive for T-antigen expression (Fig. 3C). Further studies are warranted to delineate the role of JCV oncoproteins in tumor progression. The possibility of the involvement of JCV T-antigen with gastrointestinal malignancy indicates that molecular strategies for the disruption of the interaction of T-antigen with cellular proteins may represent a fruitful avenue for the development of new types of therapeutic interventions.


  1. Top of page
  2. Abstract
  6. Acknowledgements

The authors thank past and present members of the Center for Neurovirology and Cancer Biology for their insightful discussion and sharing of ideas and reagents. They authors also thank J. Gordon for her comments on the article and C. Schriver for editorial assistance and preparation of the article.


  1. Top of page
  2. Abstract
  6. Acknowledgements