Prospective Study of Seroreactivity to JC Virus T-Antigen and Risk of Colorectal Cancers and Adenomas

John Cunningham virus (JCV) is a nonenveloped dsDNA virus with three viral capsid proteins (VP1, VP2, and VP3), small (t-Ag) and large transforming antigens (T-Ag; refs. 1, 2). JCV is highly prevalent worldwide, causing asymptomatic infection in 70% of adults (3, 4). JCV was first identified in the early 1970s in association with progressive multifocal leukoencephalopathy, a demyelinating disease of the brain with poor prognosis (5). JCV DNA has since been detected in a variety of human tumor tissues, including oligodendrogliomas (6), gastric (7), and esophageal (8) cancers. The International Agency for Research on Cancer recently classified JCV as a “group 2B” carcinogenic virus (9).

Several lines of evidence suggest that JCV may play a role in colorectal cancer. Although JCV is detected in 40% normal colon mucosa, a higher prevalence of JCV (90%) is observed in colorectal cancer (10). The expression of JCV DNA increases across the continuum of normal colon mucosa, adenoma, and colon cancer, and within colorectal cancer tumors, is significantly associated with high grade and poor prognosis of colorectal cancer (11). One cross-sectional study reported a significant correlation between circulating antibodies to JCV and colorectal cancer (12). In contrast, two prospective serologic studies, including our own (13, 14), and a case–control study measuring JCV DNA in urine (3), observed no associations between markers of JCV infection and colorectal cancer. However, we observed that seropositivity to JCV was associated with more than 2-fold increased risk of adenomas among men, with an inverse association observed among women (13).

Although previous studies measured antibodies to JCV capsid antigens, T-Ag oncoproteins are also capable of stimulating host IgG antibody response. JCV T-Ag is required for viral replication (3). Its expression promotes colorectal cancer metastasis (1) and is associated with p53 expression and chromosomal instability (15). Furthermore, JCV T-Ag DNA sequences have been detected in 82% of adenomas (16) and 77% colorectal cancers (17). Presence of JCV T-Ag DNA has been associated with methylation of tumor suppressor genes (17). JCV T-Ag sequences are more prevalent than JCV capsid sequences in tumors (7), suggesting that JCV T-Ag may be a more specific marker of oncogenic viral activity.

Collectively, these studies suggest that markers of JCV T-Ag could be important for elucidating the potential role of JCV infection in colorectal cancer. Therefore, we sought to extend our previous work by examining the association between seroreactivity to JCV T-Ag and the development of colorectal cancer and adenomas within the context of the same nested case–control study from which we previously reported our JCV capsid antibody findings (13).

A nested case–control study was conducted to investigate the association between baseline circulating antibodies to JCV T-Ag and the subsequent development of colorectal cancer and adenomas. Participant selection methods have been previously described (13). Briefly, a community-based cohort (CLUE II) was established in 1989 with 25,080 residents of Washington County, Maryland. At baseline, participants completed brief questionnaires providing information on demographic characteristics, medical history, medication use, and smoking status and provided blood samples. Additional follow-up questionnaires were mailed to the CLUE II participants every other year (1996, 1998, 2000, and 2003), capturing information on family history, medication use, screening exams, and diagnoses subsequent to baseline enrollment. Two hundred and fifty-seven incident cases of colorectal cancer were identified by linkage to the Washington County Cancer Registry and the Maryland Cancer Registry among CLUE II participants through July 2006. Controls included participants who did not develop colorectal cancer and were matched to colorectal cancer cases on age (±1 year), sex, race, and date of blood draw (±2 weeks). One hundred and twenty-three colorectal adenoma cases were identified through self-report on follow-up and verified through pathology report reviews. Controls for the adenoma cases included participants who were found not to have an adenoma after colorectal cancer screening and were matched to the adenoma cases on age, sex, race, date of blood draw, date of endoscopy (±1 year), and type of colorectal cancer screening (colonoscopy vs. sigmoidoscopy).

Plasma samples were stored at −70°C before the analysis. Recombinant protein-based ELISA was used to measure anti-JCV T-Ag antibodies. JCV large T-antigen (LTA) proteins were produced using recombinant baculovirus, expressing JCV LTA gene. For ELISA, 96-well polystyrene flat-bottom MaxiSorp plates (Nunc) were coated overnight at 4°C with 1 μg/mL of recombinant JCV LTA protein in PBS. The plates were blocked for 2 hours at room temperature with 300 μL of 0.5% (wt vol⁻¹) polyvinyl alcohol (PVA), MW 30,000 to 70,000 (Sigma) in Blocker Casein in PBS (Pierce). Serum samples, diluted 1:100 in blocking solution, were added to the antigen-coated plates and incubated at 37°C for 1 hour on a microplate shaker. After washing the plates four times with PBS, 0.05% Tween 20 in an automatic plate washer (Skanwasher 300; Skatron), goat anti-human IgG conjugated with horseradish peroxidase (Southern Biotech), diluted 1:4,000 in 0.5% PVA, 0.025% Tween 20, 0.8% (wt vol⁻¹) polyvinylpyrrolidone, MA 360,000 (Sigma-Aldrich) in PBS, were added. The plates were incubated at 37°C for 30 minutes, washed as described above, and then freshly prepared 2,2′-azino-di-(3-ethylbenzthiazoline-6-sulfonate) hydrogen peroxide solution (Kirkegaard & Perry.) prewarmed to 37°C was added. After incubation at room temperature in the dark for 20 minutes, the enzyme reaction was stopped by the addition of 1% sodium dodecyl sulfate. The plates were read at 405 nm in an automated microtiter plate reader (Molecular Devices) with a reference wavelength of 490 nm.

In the absence of a gold standard for the measurement of JCV T-Ag exposure, the median optical density (OD) value was calculated on the basis of the distribution among the controls and used as the binary cutoff for defining high versus low seroreactivity to JCV T-Ag.

McNemar test and Bowker test of symmetry were used, as appropriate, for comparing the baseline characteristics between cases and controls with respect to categorical variables. Potential confounders associated with adenomas and colorectal cancer at a significance level of P < 0.10 and other well-established risk factors for colorectal cancer were further assessed in subsequent multivariable analyses. Among the controls, the distribution of JCV T-Ag seroreactivity (mean, median, and% seroreactivity) was compared across baseline characteristics using the Wilcoxon and Fisher exact tests.

Conditional logistic regression models were used to estimate OR and 95% confidence intervals (CI) for the association between JCV T-Ag seroreactivity (low vs. high) and colorectal cancer/adenomas, with and without the inclusion of smoking status, body mass index, and family history of colorectal cancer as covariates. Additional adjustment for use of NSAIDs did not appreciably change the risk estimates, therefore, NSAID use was not included in the final models. Analyses were stratified by gender, anatomic site of the tumor/adenoma (colon, rectum, distal, proximal), and stage at colorectal cancer diagnosis. Additional stratification was conducted for colorectal cancer risk estimates by categories of time between blood draw and diagnosis (<10 years, 10–19 years) and for adenoma risk estimates by the number of adenomas (single, multiple), histology (tubular or tubulovillous/villous), and size [< or ≥0.55 cm (the median size of adenomas in the study population)]. Statistical significance of interactions between JCV T-antigen seroreactivity and stratified variables in relation to colorectal cancer/adenoma risk was tested by including interaction terms in the conditional logistic regression model, which were evaluated by the Wald test. All statistical analyses were conducted using SAS, version 9.3 (SAS Institute, Inc.).

Cases were more likely than controls to report a family history of colorectal cancer, a difference that was statistically significant for colorectal cancer cases, but not for adenoma cases (Table 1). Males had significantly (P < 0.0001) higher levels of JCV T-Ag seroreactivity compared with females (Table 2), as were study participants without family history of colorectal cancer compared with those with a family history (P = 0.01).

Overall, higher levels of JCV T-Ag seroreactivity were not statistically significantly associated with risk of colorectal cancer (Table 3; OR, 1.34; 95% CI, 0.89–2.01) or adenomas (OR, 1.30; 95% CI, 0.70–2.42), after adjusting for potential confounders. After stratification by gender, no association between baseline JCV T-Ag seroreactivity and colorectal cancer was observed among males (OR, 0.93; 95% CI, 0.51–1.67), whereas an increased risk of colorectal cancer was observed among females (OR, 1.82; 95% CI, 1.00–3.31, P_interaction = 0.09). After stratifying by length of time between baseline blood draw and diagnosis, no significant association was observed between JCV T-Ag seroreactivity and the risk of colorectal cancer (OR_{1–9 years} = 1.11; 95% CI, 0.66–1.87 and OR_{10–19 years} = 1.80; 95% CI, 0.92–3.52, P_interaction = 0.31), after adjusting for confounders.

We observed no overall significant associations between JCV T-Ag seroreactivity and colon adenomas or colorectal cancer. However, we observed a borderline significant, 82% increased risk of colorectal cancer among women in association with JCV T-Ag seroreactivity, whereas no statistically significant association was observed among men. Association of JCV T-Ag seroreactivity with adenomas did not differ by gender. In contrast, we previously reported a statistically significant positive association between seropositivity to JCV capsid and adenomas in men, and a significantly inverse association among women (13). This contrast in findings based on the same study population could be a reflection of the difference in the types of JCV antibodies measured. Because T-Ag is oncogenic and required for viral replication (3), it follows that antibody response to T-Ag is a more specific marker of active viral infection, as opposed to antibody formation against viral capsid antigen which results from asymptomatic infection (13). If this is true, our finding of no association between JCV T-Ag seroreactivity and either adenoma or colorectal cancer suggests that JCV does not play a role in colorectal cancer pathogenesis, consistent with previous studies (13, 14). Although, statistical power to detect an OR between 1.2 and 1.4 was limited given our small sample size, we had 80% power to detect an OR of 1.7, which is still a clinically meaningful risk estimate.

The borderline significant risk of colorectal cancer in association with JCV T-Ag seroreactivity observed among women is intriguing, given that women were less likely to have JCV seroreactivity to T-Ag than men (Table 2). On the contrary, if men are more likely than women to have seroreactivity to T-Ag, and if indeed JCV plays a role in colorectal cancer, then we would expect an association between seroreactivity and colorectal cancer among men. However, no statistically significant association was observed among men. The positive association between JCV T-Ag seroreactivity and colorectal cancer among women could be due to an unknown confounder that is associated with JCV T-Ag seroreactivity or due to chance.

A limitation of this study is the use of a single time point measurement of JCV T-Ag seroreactivity in banked serum samples from up to 17 years before diagnosis of colorectal cancer or adenoma. Some individuals might have acquired new JCV infection after their serum samples were collected, whereas in others, seroreactivity may have changed with increasing age. It is unclear how age affects seroreactivity to JCV T-Ag since this is a first report on JCV T-Ag seroreactivity. Between 33.3% to 54.2% controls showed high seroreactivity to JCV T-Ag, with no clear pattern seen across different age groups (Table 2). This is contrary to seroprevalence of JCV viral capsid, which increases with age (18, 19). Previously, our group observed that although seroreactivity to JCV viral capsids did not change in the majority of the study population over 15 years, the absolute level of antibody response to JCV decreased over time for more than two thirds of the population (20). Because absolute antibody level was used to define seroreactivity, our findings could be biased if change in antibodies over time varies by case–control status.

This is the first epidemiologic study of serologic response to JCV T-Ag in association with both adenomas and colorectal cancer in a well-defined cohort. In conclusion, we did not find a statistically significant association between circulating antibodies to JCV T-Ag and colon adenomas or colorectal cancer overall. Our past findings of no association between JCV capsid seropositivity and colorectal cancer (13), combined with our current findings of no association between higher JCV T-antigen seroreactivity and colorectal cancer, overall, and only a borderline significant association among women, do not support a notable etiologic role for JCV infection in colorectal cancer.

No potential conflicts of interest were disclosed.

The findings and conclusions of this report are those of the authors and do not necessarily represent the views of the Maryland Cancer Registry or the NIH.

Conception and design: R.P. Viscidi, K.J. Helzlsouer, A.R. Giuliano, D.E. Rollison

Development of methodology: R.P. Viscidi, J.-H. Lee

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R.P. Viscidi, K.J. Helzlsouer, E.A. Platz, D.E. Rollison

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R.P. Viscidi, J.-H. Lee, W.J. Fulp

Writing, review, and/or revision of the manuscript: S.S. Hampras, R.P. Viscidi, K.J. Helzlsouer, J.-H. Lee, W.J. Fulp, A.R. Giuliano, E.A. Platz, D.E. Rollison

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): W.J. Fulp

Study supervision: K.J. Helzlsouer, D.E. Rollison

The authors acknowledge the Maryland Cancer Registry, Prevention and Health Promotion Administration, and Department of Health and Mental Hygiene (http://phpa.dhmh.maryland.gov/cancer/SitePages/mcr_home.aspx) for providing cancer incidence data.

This work was supported by grant #1-R01-CA118348-01 (awarded to D.E. Rollison) and grants from the American Institute for Cancer Research (awarded to E.A. Platz). The authors acknowledge the State of Maryland, the Maryland Cigarette Restitution Fund, and the National Program of Cancer Registries of the Centers for Disease Control and Prevention for the funds that support the collection and availability of the cancer registry data.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.