High Viral Loads of Epstein-Barr Virus DNA in Peripheral Blood of Patients with Chronic Lymphocytic Leukemia Associated with Unfavorable Prognosis

Epstein-Barr virus (EBV) is a ubiquitous γ-herpesvirus that infects more than 90% of the world population. The potential involvement of EBV in the clinical course of chronic lymphocytic leukemia (CLL) remains unexplained. The aim of this study was to determine whether EBV-DNA load in the peripheral blood mononuclear cells (PBMCs) of CLL patients may influence heterogeneity in the course of the disease. The study included peripheral blood samples from 115 previously untreated patients with CLL (54 women and 61 men) and 40 healthy controls (16 women and 24 men). We analyzed the association between the EBV-DNA load in PBMCs and the stage of the disease, adverse prognostic factors, and clinical outcome. Detectable numbers of EBV-DNA copies in PBMCs were found in 62 out of 115 CLL patients (53.91%). The EBV-DNA copy number/μg DNA was significantly higher in patients who required early implementation of treatment, presented with lymphocyte count doubling time <12 months, displayed CD38-positive or ZAP-70-positive phenotype, and with the del(11q22.3) cytogenetic abnormality. Furthermore, the EBV-DNA copy number/μg DNA showed significant positive correlation with the concentrations of lactate dehydrogenase (LDH) and beta-2-microglobulin. We have shown that in CLL patients, higher EBV-DNA copy number predicted shorter survival and shorter time to disease progression, and it was associated with other established unfavorable prognostic factors. This suggests that EBV may negatively affect the outcome of CLL.


Introduction
Epstein-Barr virus (EBV), also referred to as human herpesvirus 4 (HHV-4), is the first identified human virus with documented involvement in carcinogenesis [1]. Many cells of the non-Hodgkin lymphomas (NHL) show the presence of the monoclonal form of EBV genome (i.e. the EBV-positive phenotype). This finding points to the origin of the malignancy from a single infected cell and involvement of EBV in its pathogenesis [2,3]. A number of hypotheses have been proposed to explain the etiology of NHL in individuals from the general population with inadequately controlled EBV infection. One postulated mechanism is the EBV-induced proliferation of B lymphocytes and a resultant increase in the number of B-cells at risk of oncogenic mutations and clonal expansion [4].
Despite its chronic character by definition, chronic lymphocytic leukemia (CLL) is characterized by marked heterogeneity [5,6]. Only 30% of patients survive up to 10-20 years after diagnosis [7]. The remaining CLL patients develop terminal phase within 5-10 years, despite mild onset of the disease. The individuals with the aggressive form of CLL survive no more than 2-3 years after diagnosis [8]. The reasons for such heterogeneous natural history of the condition remain unclear.
Potential involvement of EBV in the clinical course of CLL is still unexplained. Latent EBV infection is controlled by a cell-mediated immune response in healthy carriers. This immune response is impaired in CLL patients and might result in poor control of reactivation and replication of the virus. Since EBV may activate B cells, stimulate their proliferation, and inhibit their apoptosis, we hypothesized that it could contribute to unfavorable clinical course of CLL and may be one of the reasons for the observed disease heterogeneity.
The aim of this study was to define a role of EBV in the etiopathogenesis of CLL. The detailed objectives included the determination of the EBV-DNA copy number in mononuclear cells and isolated B lymphocytes from peripheral blood of CLL patients and healthy individuals and the analysis of association between this parameter and the established prognostic factors, stage of the disease, and its clinical manifestation.

Characteristics of CLL patients and healthy volunteers
The study included peripheral blood samples from 115 previously untreated patients with CLL (54 women and 61 men). The control group comprised 40 healthy subjects (16 women and 24 men). Neither the CLL patients nor the controls used immunomodulating agents or hormonal preparations, showed signs of infection within at least 3 months prior to the study, underwent blood transfusion, or presented with autoimmune condition or allergy. Moreover, none of the controls had a history of oncological therapy or prior treatment for tuberculosis or other chronic conditions that could be associated with impaired cellular or humoral immunity.
The diagnosis of CLL was established on the basis of diagnostic criteria included in the IWCLL guidelines of the American National Cancer Institute (NCI) [9,10]. Detailed characteristics of patients and controls are presented in Table 1 and Table 2.
This study was approved by the Ethics Committee of the Medical University of Lublin (decision no. KE-0254/227/2010). Written informed consent was obtained from all patients with respect to the use of their blood for scientific purposes.

Examined material
Peripheral blood (20 mL) from the basilic vein of CLL patients and healthy controls was collected into EDTA-treated tubes (15 mL) and into tubes containing clot activator (5 mL) (aspiration and vacuum systems Sarstedt, Germany). Immediately after collection, the samples were used for immunophenotyping of lymphocytes, isolation of mononuclear cells for the EBV-DNA copy number determination, serum collection for the determination of specific anti-EBV antibodies concentration, and cytogenetic studies.

Isolation of mononuclear cells and serum
Peripheral blood was diluted with 0.9% buffered saline (PBS) without calcium (Ca 2+ ) and magnesium (Mg 2+ ) (Biochrome AG, Germany) in 1: 1 ratio. The diluted material was built up with 3 mL of Gradisol L (specific gravity 1.077 g/ml; Aqua Medica, Poland), and centrifuged in a density gradient at 700 × g for 20 min. The obtained fraction of peripheral blood mononuclear Isolation of DNA and determination of the EBV copy number DNA from 5 × 10 6 PBMCs was isolated manually with the QIAamp DNA Blood Mini Kit (QIAGEN, Germany). The procedure for isolation followed the manufacturer's protocol, with a modified volume of DNA elution. Concentration and purity of the isolated DNA were verified with the BioSpec-nano spectrophotometer (Shimadzu, Japan). The EBV-DNA copy number in PBMCs was determined with the ISEX variant of the EBV PCR kit (GeneProof, Czech Republic). Qualitative and quantitative diagnostics of EBV was performed using the Real Time Polymerase Chain Reaction (RT-PCR). Specific conservative DNA sequence of a single-copy gene for the EBV nuclear antigen 1 (EBNA-1) was amplified in the course of the PCR process. The number of viral DNA copies/μL of the eluent was adjusted for the efficiency of the DNA isolation process, and then it was expressed as the viral DNA copy number/μg DNA. All the samples were examined in duplicates. A negative control, i.e. the pure buffer used for DNA elution, was amplified in every case. As the sensitivity of the system amounts to 10 copies/μL, all the samples with the EBV-DNA copy number below this detection threshold were considered EBV-negative [EBV(-)]. The PCR was performed with the 7300 Real Time PCR System (Applied Biosystems). The reaction was conducted on MicroAmp 1 Optical 96-Well Reaction Plates (Life Technologies) with MicroAmp 1 Optical Adhesive Film (Life Technologies).

Assessment of activated T and B cells
A standard, whole-blood assay with erythrocyte cell lysis was used for preparing the peripheral blood specimens. The cells were phenotypically characterized by incubation (20 min in the dark at room temperature) with a combination of relevant fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, and CyChrome-labelled monoclonal antibodies (MoAbs). Immunofluorescence studies were performed using a combination of the following MoAbs: CD3 PE, CD19 PE, CD5 FITC/CD19 PE, CD4 PE, CD8 PE, CD8 FITC/CD4 PE, CD25 CyChrome, and CD69 CyChrome, purchased from BD Biosciences (USA). Finally, cells were washed and analyzed by flow cytometry, performed on a BD FACSCalibur System. Five data parameters were acquired and stored: linear forward and side scatter (FSC, SSC), log FL-1 (FITC), log FL-2 (PE), and log FL-3 (PE-Cy5). For each analysis, 20,000 events were acquired and analyzed using CellQuest Pro software. Isotype-matched antibodies were used to verify the staining specificity and as a guide for setting the markers to delineate positive and negative populations. Mean fluorescence intensity (MFI) and the percentage of cells expressing surface markers were analyzed.

Analysis of CD38 and ZAP-70 expression in CLL cells
CLL cells were stained for CD38 antigen and ZAP-70 protein expression (as described previously by Hus et al. [11]) and analyzed using flow cytometry. The cut-off point for CD38 and ZAP-70 positivity in leukemic cells was 30% and 20%, respectively.

I-FISH analysis
PBMCs were cultivated for 24 hours in RPMI 1640 medium without mitogen stimulation. After hypotonic treatment and methanol-acetic acid 3: 1 fixation, cell suspensions were dropped onto microscopic slides and used directly for I-FISH. The commercially available Vysis probes (Abbott Molecular Europe, Wiesbaden, Germany) LSI ATM SpectrumOrange/ CEP 11 SpectrumGreen Probe, and LSI TP53 SpectrumOrange/CEP 17 SpectrumGreen Probe were used. At least 200 nuclei were analyzed for each probe. The cut-off levels for positive values for normal controls were 2.5% (mean ± SD).

Anti-EBV immunoassay
Commercial enzyme-linked immunosorbent assay (ELISA) kits, purchased from IBL International (Germany) for a quantitative determination of specific anti-EBV antibodies in human serum were used. Protocols followed were in accordance with the manufacturer's recommendations. The ELISA Reader Victor TM3 (PerkinElmer, USA) was used.

Statistical analysis
Normal distribution of continuous variables was tested using the Shapiro-Wilk test. Statistical characteristics of the continuous variables were presented as medians, minimum and maximum values, as well as arithmetic means and their standard deviations (SD). The Student t-test was used for independent variables, and the Mann-Whitney U-test was used for intergroup comparisons. The power and direction of relationships between pairs of continuous variables were determined on the basis of the values of Spearman's coefficient of rank correlation (R). The distributions of discrete variables in the studied groups were compared with the Pearson's Chi-square test or the Fisher's exact test. The survival curves were constructed with the Kaplan-Meier method, and the proportions of survivors within the studied groups were compared with the log-rank test. Univariable and multivariable Cox proportional hazard regression models were used to determined association between patients characteristics and time to first treatment. For the multivariable model, stepwise variable selection was used. Significant variables (p<0.1) were tested for inclusion in the regression models, and nonsignificant variables were removed sequentially until only those significant at p<0.05 remained.
Receiver operating characteristic (ROC) curves were generated for significant predictor variables of EBV(+) CLL patients. Areas under the ROC curves (AUCs) were calculated for each parameter and compared. All the calculations were carried out with Statistica 10 (StatSoft 1 , USA) package, with the level of significance set at P < 0.05.

Results
The PBMCs of 115 CLL patients and 40 controls were tested for the presence of EBV-DNA with an aid of real-time PCR. A total of 62 (53.91%) CLL patients presented with a large number of EBV-DNA copies (more than 10 copies/μL). The following three groups were identified on the basis of this criterion: (a) CLL patients whose PBMCs showed the presence of the EBV-DNA, i.e. the EBV(+) group, (b) CLL patients whose PBMCs lacked the EBV-DNA, i.e. the EBV(-) group and (c) healthy controls.
In univariable analysis, among clinical and biochemical parameters of CLL patients both traditional and new prognostic factors associated with shorter time to first treatment turn out to be statistically significant (Table 1). We found that categorical variables such as the splenomegaly, trisomy 12, CD38, and positive anti-EBV EA IgA were associated with time to first treatment (Table 1). Among continuous variables lower hemoglobin, higher beta-2 microglobulin and LDH, higher CD19+CD25+ cells, CD19+CD69+ cells, CD3+CD69+ cells, CD19+CD5+CD38+ cells and CD19+CD5+CD23+ cells had significant association with time to treatment. Higher level of anti-EBV: EA IgG, EBNA IgG and VCA IgG were associated with shorter time to first treatment. Interestingly, EBV(+) patients had more than 7 times shorter time to first treatment comparing to EBV(-) patients. This phenomenon was confirmed during multivariable analysis. The multivariable Cox analysis identified the higher beta-2-microglobulin (HR = 1.43; p = 0.0033), enlarged spleen (HR = 4.70; p<0.0001), EBV presence (HR = 23.39; p<0.0001) and lower anti-EBV EBNA-1 IgM level (HR = 0.98; p = 0.0089) as independent predictors for shorter time to first treatment (Table 3a). Since the Cox proportional hazard regression model showed highly statistically significant association of EBV(+) with progression of illness, a ROC curve was drawn to test the variables predictive validity in EBV(+) CLL patients (Table 3b). As the area under the curve shows (Fig 1), CD19+CD25 + cells [%] in the peripheral blood parameter was the most sensitive and specific to determine EBV(+) (AUC = 0.854) . Fig 2 presents a Kaplan-Meier curve illustrating the time to first treatment depending on the EBV-DNA copy number/μg DNA isolated from PBMCs (Fig 2a) and a Kaplan-Meier curve illustrating the probability of lymphocyte doubling-free survival depending on the EBV-DNA copy number/μg DNA isolated from PBMCs (Fig 2b). Table 4 presents the comparison between clinical and laboratory parameters in CLL EBV(+) patients, CLL EBV (-) patients and the study group. Table 5 presents an assessment of anti-EBV antibody concentrations in relation to the presence or absence of EBV-DNA copies in PBMCs of CLL patients     [12], according to whom the EBV-DNA copy number per μg DNA extracted from PBMCs of patients with EBV-related lymphoproliferative disorders and infectious mononucleosis ranges between 10 and 10000. In contrast, the EBV-DNA copy number in most healthy controls and immunocompromised patients after organ transplantation did not exceed 10 copies/μg DNA. Both Kimura et al. [12] and Stevens et al. [13] postulated that more than 100-1000 EBV-DNA copies/μg DNA isolated from examined material are associated with clinical signs of EBV infection, manifesting as the chronic active EBV disease (CAEBV) or post-transplant lymphoproliferative disease. These findings and the results of our study point to a potential presence of a CLL subtype being associated with EBV infection. Moreover, an increase in the EBV-DNA copy number was documented in most of our patients during approximately 2-year follow-up. We revealed the presence of EBV-DNA in PBMCs and isolated B lymphocytes in more than a half of our CLL patients. To the best of our knowledge, no previous studies distinguished between the CLL forms being associated with EBV infection or unrelated to this virus. The EBV-associated form of CLL seems to be characterized by more aggressive phenotype. We showed that the EBV-DNA copy number in PBMCs of patients with hepatomegaly or thrombocytopenia and individuals who required earlier implementation of treatment was significantly higher than that in the remaining individuals. A number of previous studies documented the role of EBV in induction of thrombocytopenia. While the presence of EBV in patients with infectious mononucleosis is usually associated with a slight decrease in platelet count, in the case of CAEBV, it can be associated with severe thrombocytopenia, anemia (usually of autoimmune origin), and splenomegaly (resulting from lymphocyte infiltration) or even liver failure [14,[15][16]. Moreover, we showed that the EBV-DNA copy number correlated significantly with serum concentrations of beta-2-microglobulin and LDH. As early as 1981, Ibsen et al. [17] revealed that the level of beta-2-microglobulin is at its highest during initial stages of infectious mononucleosis, and subsequently, within 3 weeks to 3 months after recovery, it normalizes to its baseline level. The fact that concentration of beta-2-microglobulin constitutes an established predictive factor in CLL patients may suggest that the elevated level of this protein is associated with EBV infection in at least some of the cases [18]. Furthermore, we revealed significant associations between other negative prognostic factors such as high cytoplasmic expression of ZAP-70 [19], surface expression of CD38 in leukemic cells [20], surface expression of CD23, CD25, and CD69 [21,22], as well as unfavorable genetic mutations [23], and EBV-DNA copy number.
Tsimberidou et al. reported that 38% of CLL patients had evidence of EBV infection by in situ hybridization for EBV EBER1, a small noncoding RNA species [24]. Tarrand et al. [25] reported that LMP1 mRNA levels were higher in CLL patients than in healthy subjects (14% vs. 1% of healthy controls), suggesting that EBV late gene expression occurs at least in a subset of CLL cells. We demonstrate significant associations between viral load of EBV-DNA and  various clinical and pathologic variables among CLL patients, including associations with progression and time to treatment. These findings are in line with conclusions made by Visco et al. [26] who postulated that EBV-DNA load at presentation is an independent predictor of overall survival in patients with CLL.

Conclusions
In conclusion, more than a half of CLL patients presented with CLL EBV-DNA in their PBMCs, whereas no detectable amounts of genetic material for this pathogen were found in healthy controls. Greater EBV-DNA copy number was associated with shorter overall survival and time to progression in CLL patients. Positive correlation between EBV-DNA copy number and established unfavorable prognostic factors of CLL implies that increased EBV load in peripheral blood may predict poor clinical outcomes of CLL.