点击显示 收起
National Microbiology Center, Instituto de Salud Carlos III, Ctra. Majadahonda-Pozuelo s/n, 28220 Majadahonda, Madrid, Spain
ABSTRACT
Clinical research suggests a role for viral load measurement in predicting and monitoring Epstein-Barr virus (EBV)-associated diseases. The aim of this study was to assess the performance of the recently commercially available quantitative assays for EBV based on real-time PCR: the RealArt EBV LC PCR kit and the LightCycler EBV quantification kit. A total of 87 samples were analyzed: 67 samples were obtained from transplant recipients and patients with EBV-associated diseases, 8 samples were obtained from the Quality Control for Molecular Diagnostics 2002 EBV Proficiency Program, and 12 negative qualitative nested PCR samples were used as negative controls. Inter- and intra-assay variabilities were determined by running replicates of two samples. All samples were run in a LightCycler instrument. The differences between positive and negative results were not considered statistically significant (P = 0.5355). There were no false-positive results using either method for nested PCR negative-control samples. The difference in viral load values using the two different methods was considered statistically significant (P < 0.01). The logarithmic linear correlation for both assays was low (r = 0.449) but significant (P < 0.01). The LightCycler EBV quantification kit showed a wider dispersal in results but produced substantially more-accurate melting temperature profile curves. The bias towards lower measurements was considerable in comparison with higher viral load. The differences in PCR efficiency and the presence of mutations could explain the disparity between the two methods. It was concluded that confidence intervals would be required to report the results rather than plain absolute values of viral load for patient monitoring.
INTRODUCTION
Epstein-Barr virus (EBV) is a human herpesvirus included in the Gammaherpesvirinae subfamily and is the only human species belonging to the genus Lymphocryptovirus (11). EBV infects more than 90% of the world's population, leaving the majority of individuals with a lifelong silent infection (14). Although most primary EBV infections are asymptomatic, the virus is the causative agent of infectious mononucleosis, a mild and self-limited lymphoproliferative disease (13). EBV is associated with the development of two epithelial tumors, nasopharyngeal carcinoma (5) and oral hairy leukoplakia, seen mainly in human immunodeficiency virus infection (3), as well as with various lymphoid carcinomas, lymphoproliferative disease in immunosuppressed patients, Burkitt's lymphoma, Hodgkin's disease, and T-cell non-Hodgkin's lymphoma (25).
EBV DNA is present in a small fraction of lymphoid cells, and healthy virus carriers harbor 1 to 50 EBV genomes per 106 mononuclear cells, with B lymphocytes representing the major cellular reservoir (9). Qualitative PCR assays are unable to distinguish between active and latent infection. Consequently, clinical interpretation of positive results is difficult. However, clinical research suggests a role for viral load measurement in predicting and monitoring EBV-associated tumors, including nasopharyngeal carcinoma, posttransplant lymphoproliferative disorders (PTLD), Hodgkin's disease, and AIDS-related lymphoma (7). Furthermore, real-time quantitative assays have been performed for diagnosis of primary EBV infection on the basis of the relatively poor sensitivity of some serology markers in younger children (19).
Real-time amplification technology reduces labor costs, is less time-consuming, and also reduces the risk of amplicon contamination. Since its development in the 1990s, many different assay formats and applications have been developed and the number of real-time PCR machines of different designs is increasing. Several reports describe the development of "in-house" real-time PCR protocols for the detection of EBV genomes using TaqMan probes (6, 12) and fluorescence resonance energy transfer (FRET) probes (4, 18, 21). Two commercial kits for the measurement of EBV DNA loads in different clinical samples (serum, whole blood, cerebrospinal fluid, peripheral blood mononuclear cells, and peripheral blood polymorphonuclear cells) have recently become available: the RealArt EBV LC PCR kit (Artus, Hamburg, Germany) and the LightCycler EBV quantification kit (Roche Diagnostics, Mannheim, Germany).
The aim of this study was to evaluate the performance of the recently available commercial quantitative assays for EBV based on real-time PCR.
MATERIALS AND METHODS
Samples. A total of 87 samples were analyzed. All the samples were tested with a qualitative nested PCR (n-PCR) for lymphotropic herpesviruses as previously described (20). Characteristics of samples and patients were as follows.
Ten serum samples were obtained from patients with symptomatic primary EBV infection diagnosed by detection of anti-viral capsid antigen (VCA)-specific immunoglobulin M (IgM), and four of them were also positive by n-PCR. Eleven cerebrospinal fluid (CSF) specimens were taken from patients with central nervous system (CNS) EBV-associated diseases who were positive for EBV DNA by n-PCR. Nineteen cross-sectional whole-blood samples were taken from 6 kidney and 13 liver transplant recipients positive for EBV by n-PCR. Eight follow-up samples were taken from one kidney transplant recipient with recurrent EBV infection with persistent positivity for IgM VCA: 7 serum samples and 1 whole-blood sample positive by n-PCR. Nineteen samples were taken from 11 patients diagnosed with EBV-related tumors, PTLD (n = 4), lymphoma (n = 4), chronic lymphatic leukemia (n = 2), and nasopharyngeal carcinoma (n = 1). A total of 10 serum and 3 whole-blood samples from these patients were obtained, and a set of follow-up specimens (2 serum, 2 peripheral blood mononuclear cell, and 2 peripheral blood polymorphonuclear cell samples) were taken from a patient with PTLD. Eight samples belonged to the Quality Control for Molecular Diagnostics (Epstein-Barr virus proficiency program 2002) of known EBV DNA load. Finally, 4 whole-blood, 4 serum, and 4 CSF samples from patients with no EBV-related diseases and who were negative for EBV DNA by qualitative n-PCR were used as negative controls.
Viral DNA isolation. Viral DNA was extracted from 100 μl of samples by the automated MagNA Pure LC system (Roche Diagnostics, Mannheim, Germany) following the total nucleic acid external lysis protocol, according to the manufacturer's instructions, to obtain 50 μl of DNA solution. Two aliquots of the extracted DNA were made to avoid further freezing and thawing.
Quantitative LightCycler PCR. All samples were run in a LightCycler instrument (Roche Diagnostics, Mannheim, Germany). The RealArt EBV LC PCR kit amplifies a 97-bp region of the EBV genome and is based on the detection of the amplified product with FRET hybridization probes. The capillaries were loaded with 15 μl of reaction mixture containing deoxynucleoside triphosphates, primers, probes, MgCl2, Taq DNA polymerase, the internal control for checking PCR inhibition, and 5 μl of sample DNA. The four enclosed quantification standards were included in each run at the same volume as purified samples. PCR was performed on the LightCycler instrument using the cycling program described in the user manual. The data were analyzed by activating the color compensation file to separate signals from the EBV PCR (channel F2) and from the internal control (channel F3). Finally, a melting curve analysis was carried out to check the specificity of the assay. The detection limit, according to the manufacturer's instructions, is 5.78 copies/μl (P = 0.05) for a cloned PCR product.
The LightCycler EBV quantification kit uses a specific pair of FRET hybridization probes to detect a fragment of the latent membrane protein. The master mix contained 2 μl of LightCycler EBV reaction mix with FastStart Taq DNA polymerase, reaction buffer, and deoxynucleoside triphosphates plus 2 μl of LightCycler EBV detection mix with the hybridization probe mixture and PCR-grade water to achieve a final volume of 15 μl per capillary. Five μl of sample DNA was added, and EBV DNA standards were included in each run. The values of the cycling were programmed according to the user manual. Data were analyzed using a previously generated color compensation file. After amplification was completed, a melting curve analysis program was run. The lower detection limit was 10 copies per reaction, and the linear range of the assay was 102 to 106 copies per reaction, as specified in the user manual.
None of the purified DNA samples inhibited the amplification of the internal control for the two kits. To avoid user-borne influences, the cycle threshold values were calculated using the second derivate maximum method with an arithmetic baseline adjustment. The concentration units were converted into copies per milliliter and a log of copies/ml.
Inter- and intra-assay variability. To assess the level of precision, inter- and intra-assay variabilities were determined by running repetitions of two different whole-blood samples from one liver (sample A) and one kidney (sample B) transplant recipient. The samples were extracted in four different runs, using the MagNA Pure LC system to assess variability in viral DNA recovery. Real-time PCR of the DNA template from the same and different extraction assays and their replicates were run on three consecutive days.
Statistics. The qualitative results were compared with the use of Fisher's exact test. The differences in viral load values were calculated with Mann-Whitney U and Wilcoxon W tests. Pearson's coefficients served to compare the correlation between the two kits. Bland and Altman's method (1) was used to assess the degree of agreement of viral load single measurements as well as repeatability for samples A and B. The intra- and interassay variations were evaluated with descriptive statistics.
Statistics were worked out using the SPSS 11.5.1 software package (SPSS, Inc.) and SigmaPlot 8.02 (Systat Software, Inc., California) scientific graphic software.
RESULTS
Qualitative results. Excluding the Quality Control for Molecular Diagnosis (QCMD) samples, there were 50 positive samples with the RealArt EBV LC PCR kit and 45 positive samples with the LightCycler EBV quantification kit (Table 1). Five samples (7.5%) which were positive with the LightCycler EBV quantification kit were negative with the RealArt EBV LC PCR kit, and 10 samples (14.9%) which were positive with the RealArt EBV LC PCR kit were negative with the LightCycler EBV quantification kit. There were 12 (17.9%) samples which were negative in both tests, most of these being serum samples from patients with primary EBV infection, diagnosed by serological reactivity for IgM anti-VCA and negative by n-PCR. There were no false-positive results using either method for n-PCR-negative samples (n = 12) included as negative controls. The observed differences between the two methods were not considered statistically significant (P = 0.5355). The QCMD samples with 500 copies/ml were not detected by either method (Table 2). Considering the dilution factor of the extraction and amplification, 500 copies/ml means 5 DNA copies per reaction tube.
Quantitative results. The mean copy number for positive samples, including the QCMD samples, was 18,561 copies/ml (n = 55) using the RealArt EBV LC PCR kit and 99,961 copies/ml (n = 50) using the LightCycler EBV quantification kit. This difference was statistically significant (P < 0.01). However, the LightCycler EBV quantification kit also showed more scattered results, with a range of concentrations from 4,821 copies/ml to 1,494,000 copies/ml. The highest viral loads were found in the groups of CNS EBV-associated diseases and EBV-related tumors (Table 1). The quantitative results obtained with the QCMD 2002 EBV Proficiency Program are summarized in Table 2. The logarithmic linear correlation (Fig. 1) between the two assays was low (r = 0.449) but significant (P < 0.01). The linear correlation with absolute values, rather than logarithmic values, was higher: r = 0.711 (P < 0.01). The best correlation in absolute values was given by CSF samples: r = 0.958 (P < 0.01). The mean of the values obtained with each method was plotted against the difference in these same values for each individual sample (Fig. 2) to improve the study of the statistical significance of observed differences (1). Most samples grouped into a single cluster within the 2x standard deviation lines, showing that most of the differences have normal distribution. However, samples positive by only one technique made up two completely separate groups: one group on the upper left made up of samples which were positive by the RealArt EBV LC PCR kit but negative by the LightCycler EBV quantification kit and another group on the lower left made up of samples positive by the LightCycler EBV quantification kit and negative by the RealArt EBV LC PCR kit. As quantitative values in the respective positive technique were not low, most samples in these two groups are close to or even outside the admitted 2x standard deviation value.
Intra- and interassay variations. The results for intra- and interassay variability are summarized in Table 3. Sample B showed the lower viral DNA load but had the highest variation coefficients for the two methods. The largest intra- and interassay variations were found with the LightCycler EBV quantification kit.
Melting temperature values. The mean melting temperature (Tm) for the LightCycler EBV quantification kit was 62.61 ± 0.42°C. Similar Tms were obtained for samples A and B run to measure intra- and interassay variations: 62.12 ± 0.26°C and 62.53 ± 0.22°C, respectively. Nevertheless, a wider range of Tms was found in the case of the RealArt EBV LC PCR kit, 67.56 ± 1.36°C, with a slight difference in mean Tm between samples A and B, 64.11 ± 0.63°C and 64.49 ± 0.63°C, respectively. Profiles of the Tm curves for the LightCycler EBV quantification kit were substantially more accurate. Only one sample showed polymorphism within the hybridization probe binding region, with a melting temperature of 55.5°C for the LightCycler EBV quantification kit (Fig. 3).
PCR efficiency. PCR efficiency is a technical term to describe the quality of the standard curve plot (8, 16). The ideal PCR efficiency should be 100% with a slope value of –3.32 (8). The mean percentage of PCR efficiency was 97.3% for the LightCycler EBV quantification kit and 82.5% for the RealArt EBV LC PCR kit. Similar results were obtained from the runs to measure inter- and intra-assay variations: 95.5% and 88.1%, respectively.
DISCUSSION
Qualitative detection of EBV generally does not necessarily correlate with the presence of disease. However, quantification of EBV DNA could potentially link viral load fluctuations and clinical symptoms to monitoring EBV-associated diseases. Given the lack of ability of most clinical laboratories to develop "in-house" methods, available commercial assays would be the way to provide them with this new diagnostic tool.
Both the LightCycler EBV quantification kit and the RealArt EBV LC PCR kit showed good sensitivity, as the qualitative results illustrate, and ranged between 5 to 10 copies/tube depending on the QCMD panel. The results are in agreement with the QCMD 2002 EBV Proficiency Program report (17), which shows that none of the laboratories (n = 4) gave correct results with the RealArt EBV LC PCR kit containing 100 and 500 copies/ml, quantified by electron microscopy, whereas 3 of 12 and 4 of 12 laboratories reported correct results for both samples, respectively, with the LightCycler EBV quantification kit.
We found that the majority of our negative results using the two techniques corresponded to serum from patients with primary EBV infection, diagnosed serologically and negative by n-PCR, suggesting that EBV does not remain long in the serum of patients with acute infection, in contrast to IgM (19).
The high EBV viral load in CSF (5.22 log copies/ml for the LightCycler EBV quantification kit and 4.67 log copies/ml for the RealArt EBV LC PCR kit) are in agreement with values previously reported (2, 24). The LightCycler EBV quantification kit showed higher but wider distribution of viral load values than the RealArt EBV LC PCR kit. This did not correlate with better analytical or clinical sensitivity. PCR efficiency, which was more than 14% higher for the LightCycler EBV quantification kit, could account for the lower values obtained with the RealArt EBV LC PCR kit. The Tm values obtained with the LightCycler EBV quantification kit were well defined and within the range of the manufacturer's instructions (62 ± 2°C). Nevertheless, a wider range of temperatures (62.89°C to 69.32°C) was obtained with the RealArt EBV LC PCR kit, showing less sharp Tm profiles. This could be the cause of lower slope values and, consequently, the lower PCR efficiency. Other additional variables likely to influence PCR efficiency are MgCl2 and primer and probe concentrations (8).
We found 15 samples that showed negative results by one technique but high viral load values by the other. They did not show any continuity with the cluster containing the majority of the samples for which results are shown in Fig. 2. This suggests the presence of mutations involving the sites of primer annealing and/or probes in one or another method. This is another issue that should be taken into consideration by manufacturers to avoid false-negative results.
The results of intra- and interassay variability support the belief that, for patient monitoring, confidence intervals are required to report results instead of plain absolute values for viral load. Otherwise, "technical" fluctuations in viral load values could be misinterpreted as having clinical significance, leading to improper handling of patient treatment. As an example, sample B ranged between 16,820 and 105,700 in different assays (Table 3). An additional limiting step that could worsen this effect is the reproducibility and sensitivity of the nucleic acid isolation method. Follow-up samples should be analyzed together with the previous sample in the same assay to evaluate viral load fluctuations during patient monitoring, as this is typically undertaken in serological assays to consider differences in antibody titers. Published cutoff values defined for patient management by "in-house" assays (22, 23) should not be considered for these commercial techniques until further comparative studies are performed. Commercial quantitative assays should be normalized to a universal standard, based either on electron microscopy-counted viral particles or, at least, on full EBV genomes contained in EBV-infected cell lines, such as the case of Namalwa cell line which contains two copies of EBV integrated within chromosome 1 (10). Moreover, some reports (7, 15, 21), as a future objective, comment on the use of patient-individualized EBV DNA load kinetics rather than absolute clinical cutoff values.
To sum up, the techniques evaluated in this study could be used for the diagnosis of EBV infections. However, confidence intervals should be added to the absolute result values to achieve effective patient management based on viral loads.
ACKNOWLEDGMENTS
We thank Roche Diagnostics Spain and IZASA for providing the LightCycler EBV quantification kit and the RealArt EBV LC PCR kit, respectively. We thank Tobias Ruckes (Artus GmbH) for reviewing the results. Finally, we are grateful for the collaboration of technicians Irene González, Ana Castellanos, Paloma Lucas, Francisco Salvador, Sara Hierro, Meyi González, and María Eulalia Guisasola.
Present address: Department of Clinical Microbiology and Parasitology, University Hospital La Paz, Paseo de la Castellana, 261, 28046 Madrid, Spain.
REFERENCES
Bland, J. M., and D. G. Altman. 1986. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i:307-310.
Bossolasco, S., P. Cinque, M. Ponzoni, M. G. Vigano, A. Lazzarin, A. Linde, and K. I. Falk. 2002. Epstein-Barr virus DNA load in cerebrospinal fluid and plasma of patients with AIDS-related lymphoma. J. Neurovirol. 8:432-438.
Boulter, A. W., N. Soltanpoor, A. V. Swan, W. Birnbaum, N. W. Johnson, and C. G. Teo. 1996. Risk factors associated with Epstein-Barr virus replication in oral epithelial cells of HIV-infected individuals. AIDS 10:935-940.
Brengel-Pesce, K., P. Morand, A. Schmuck, M. J. Bourgeat, M. Buisson, G. Bargues, M. Bouzid, and J. M. Seigneurin. 2002. Routine use of real-time quantitative PCR for laboratory diagnosis of Epstein-Barr virus infections. J. Med. Virol. 66:360-369.
Chan, K. C., and Y. M. Lo. 2002. Circulating EBV DNA as a tumor marker for nasopharyngeal carcinoma. Semin. Cancer Biol. 12:489-496.
Dehee, A., C. Asselot, T. Piolot, C. Jacomet, W. Rozenbaum, M. Vidaud, A. Garbarg-Chenon, and J. C. Nicolas. 2001. Quantification of Epstein-Barr virus load in peripheral blood of human immunodeficiency virus-infected patients using real-time PCR. J. Med. Virol. 65:543-552.
Fan, H., and M. L. Gulley. 2001. Epstein-Barr viral load measurement as a marker of EBV-related disease. Mol. Diagn. 6:279-289.
Ginzinger, D. G. 2002. Gene quantification using real-time quantitative PCR: an emerging technology hits the mainstream. Exp. Hematol. 30:503-512.
Gulley, M. L. 2001. Molecular diagnosis of Epstein-Barr virus-related diseases. J. Mol. Diag. 3:1-10.
Henderson, A., S. Ripley, M. Heller, and E. Kieff. 1983. Chromosome site for Epstein-Barr virus DNA in a Burkitt tumor cell line and in lymphocytes growth-transformed in vitro. Proc. Natl. Acad. Sci. USA 80:1987-1991.
Kieff, E., and A. B. Rickinson. 2001. Epstein-Barr virus and its replication, p. 2511-2573. In D. M. Knipe, P. M. Howley, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. E. Straus (ed.), Fields virology, 4th ed., vol. 1. Lippincott Williams & Wilkins, Philadelphia, Pa.
Kimura, H., K. Nishikawa, Y. Hoshino, A. Sofue, Y. Nishiyama, and T. Morishima. 1999. Quantitative analysis of Epstein-Barr virus load by using a real-time PCR assay. J. Clin. Microbiol. 37:132-136.
Kobbervig, C., D. Norback, and B. Kahl. 2003. Infectious mononucleosis progressing to fatal malignant lymphoma: a case report and review of the literature. Leuk. Lymphoma 44:1215-1221.
Macsween, K. F., and D. H. Crawford. 2003. Epstein-Barr virus-recent advances. Lancet Infect. Dis. 3:131-140.
Matsukura, T., A. Yokoi, H. Egawa, T. Kudo, M. Kawashima, Y. Hirata, H. Tanaka, K. Kagajo, H. Wada, and K. Tanaka. 2002. Significance of serial real-time PCR monitoring of EBV genome load in living donor liver transplantation. Clin. Transplant. 16:107-112.
Meijerink, J., C. Mandigers, L. van de Locht, E. Tonnissen, F. Goodsaid, and J. Raemaekers. 2001. A novel method to compensate for different amplification efficiencies between patient DNA samples in quantitative real-time PCR. J. Mol. Diagn. 3:55-61.
Niesters, H. G. M. 2003. Summary of results QCMD 2002 Epstein-Barr virus proficiency programme. Quality Control for Molecular Diagnostics, Glasgow, United Kingdom.
Patel, S., M. Zuckerman, and M. Smith. 2003. Real-time quantitative PCR of Epstein-Barr virus BZLF1 DNA using the LightCycler. J. Virol. Methods 109:227-233.
Pitetti, R. D., S. Laus, and R. M. Wadowsky. 2003. Clinical evaluation of a quantitative real time polymerase chain reaction assay for diagnosis of primary Epstein-Barr virus infection in children. Pediatr. Infect. Dis. J. 22:736-739.
Pozo, F., and A. Tenorio. 1999. Detection and typing of lymphotropic herpesviruses by multiplex polymerase chain reaction. J. Virol. Methods 79:9-19.
Stevens, S. J., S. A. Verkuijlen, A. J. Brule, and J. M. Middeldorp. 2002. Comparison of quantitative competitive PCR with LightCycler-based PCR for measuring Epstein-Barr virus DNA load in clinical specimens. J. Clin. Microbiol. 40:3986-3992.
Stevens, S. J., E. A. Verschuuren, I. Pronk, W. van Der Bij, M. C. Harmsen, T. H. The, C. J. Meijer, A. J. van Den Brule, and J. M. Middeldorp. 2001. Frequent monitoring of Epstein-Barr virus DNA load in unfractionated whole blood is essential for early detection of posttransplant lymphoproliferative disease in high-risk patients. Blood 97:1165-1171.
Wagner, H. J., M. Wessel, W. Jabs, F. Smets, L. Fischer, G. Offner, and P. Bucsky. 2001. Patients at risk for development of posttransplant lymphoproliferative disorder: plasma versus peripheral blood mononuclear cells as material for quantification of Epstein-Barr viral load by using real-time quantitative polymerase chain reaction. Transplantation 72:1012-1019.
Weinberg, A., S. Li, M. Palmer, and K. L. Tyler. 2002. Quantitative CSF PCR in Epstein-Barr virus infections of the central nervous system. Ann. Neurol. 52:543-548.
Young, L. S., and P. G. Murray. 2003. Epstein-Barr virus and oncogenesis: from latent genes to tumours. Oncogene 22:5108-5121