Measurement of Epstein-Barr Virus DNA Loads in Whole Blood and Plasma by TaqMan PCR and in Peripheral Blood Lymphocytes by Competitive PCR

Size: px

Start display at page:

Download "Measurement of Epstein-Barr Virus DNA Loads in Whole Blood and Plasma by TaqMan PCR and in Peripheral Blood Lymphocytes by Competitive PCR"

Gavin Underwood
5 years ago
Views:

1 JOURNAL OF CLNCAL MCROBOLOGY, Nov. 2003, p Vol. 41, No /03/$ DO: /JCM Copyright 2003, American Society for Microbiology. All Rights Reserved. Measurement of Epstein-Barr Virus DNA Loads in Whole Blood and Plasma by TaqMan PCR and in Peripheral Blood Lymphocytes by Competitive PCR Robert M. Wadowsky, 1,2,3 * Stella Laus, 3 Michael Green, 3,4,5 Steven A. Webber, 3,4 and David Rowe 2 Departments of Pathology, 1 Pediatrics, 4 and Surgery, 5 School of Medicine, and Department of nfectious Diseases and Microbiology, 2 Graduate School of Public Health, University of Pittsburgh, and Children s Hospital of Pittsburgh, 3 Pittsburgh, Pennsylvania Received 2 April 2003/Returned for modification 19 July 2003/Accepted 5 August 2003 Epstein-Barr virus (EBV) DNA load values were measured in samples of whole blood (n 60) and plasma (n 59) by TaqMan PCR and in samples of peripheral blood lymphocytes (PBLs) (n 60) by competitive PCR (cpcr). The samples were obtained from 44 transplant recipients. The whole-blood and PBL loads correlated highly (r 2 > 0.900), whereas the plasma and PBL loads correlated poorly (r ). Testing of whole blood by TaqMan PCR is an acceptable alternative to testing of PBLs by cpcr for quantifying EBV DNA load. Monitoring of Epstein-Barr virus (EBV) DNA load in peripheral blood lymphocytes (PBLs) from transplant patients by competitive PCR (cpcr) has been shown to be useful in the diagnosis and management of EBV infection and EBV-associated posttransplant lymphoproliferative disease (PTLD) (2, 3, 9, 12). Although the viral load has also been measured in plasma and whole blood by real-time TaqMan PCR (4, 5, 8), the comparability of viral load results in different sample types, determined by different PCR methods, is unclear. The purpose of this study was to evaluate the comparability of EBV DNA loads measured in whole blood and plasma by a TaqMan PCR assay (R. D. Pitetti, S. Laus, and R. M. Wadowsky, abstract from the American Academy of Pediatrics National Conference 2002; Pediatr. Emerg. Care 18: , 2002) with the load in PBLs measured by a cpcr assay (9). Samples of whole blood and plasma for testing by TaqMan PCR were derived from clinical samples that had been collected between October 2001 and April 2002 for determination of EBV viral load in PBLs by cpcr at the EBV Laboratory at the Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pa. (9). The whole-blood and plasma samples were stored at 80 C until testing. Paired samples of whole blood and plasma were chosen from the frozen bank in a manner that provided a relatively even distribution of results from the cpcr assay. To facilitate selection, results from the cpcr assay were categorized into one of four load groups: group, 8 copies/10 5 PBLs; group, 8 to 100 copies/10 5 PBLs; group, 200 to 1,000 copies/10 5 PBLs; and group, 1,000 copies/10 5 PBLs. Viral load cutoffs for these groups were based on how we have typically interpreted results from the testing of PBLs by the cpcr assay for the past 7 years (2, 3, 9, 12). Group represents either a value not detected or a detectable (albeit very low) value, similar to values seen in * Corresponding author. Mailing address: Department of Pathology, Children s Hospital of Pittsburgh, 3705 Fifth Ave., Pittsburgh, PA Phone: (412) Fax: (412) Robert.Wadowsky@chp.edu. latent infection in immunocompetent individuals. Group values are 1 to 2 log 10 units greater than those expected in normal latent infection, but which have not typically been associated with clinical disease in transplant patients. Groups and values are considered highly and extremely elevated, respectively, and we have previously associated these values with an increased risk for either having or developing symptomatic EBV infection or EBV-associated PTLD (9). Testing by TaqMan PCR was done blindly in the Pediatric Molecular Microbiology Laboratory at Children s Hospital of Pittsburgh. DNA was extracted from 200- l portions of plasma and whole blood by a spin column method (QAamp DNA Blood Mini kit; Qiagen, Valencia, Calif.) and eluted in 50 lof buffer AE (Qiagen). A primer-probe set (4) for a 90-bp target within the EBV DNA polymerase gene (i.e., BALF-5) was used in this TaqMan PCR assay (Pitetti et al.); the nucleotide sequences (5 33 ) were as follows: upstream primer, CGGA AGCCCTCTGGACTTC; downstream primer, CCCTGTTTA TCCGATGGAATG; and probe, TGTACACGCACGAGAA ATGCGCC. The probe was synthesized with a FAM reporter molecule on the 5 end and a TAMRA quencher molecule on the 3 end (Applied Biosystems; Foster City, Calif.). PCR amplification was performed in a 50- l reaction mixture containing 25 l of2 TaqMan Universal PCR master mix (Applied Biosystems), 5 l of10 Exogenous nternal Positive Control Mix (Applied Biosystems), 1 l of50 Exogenous nternal Positive Control DNA (Applied Biosystems), 0.2 M (each) EBV forward and reverse primers, 0.1 M EBV probe, and 5 l of sample. The Exogenous nternal Positive Control reagents contain a fluorogenic DNA probe labeled with a VC reporter molecule for measuring the amplification of the internal positive control. The internal positive control identifies samples containing substances that inhibit PCR. Each run included the testing of a pgem-balf5 plasmid standard (4) in 10-fold increments ranging from 20 to 2,000,000 copies per reaction and a reagent control for every five test samples. The plasmid was kindly provided by Hiroshi Kimura, Nagoya University School of Medicine, Nagoya, Japan. DNA samples, 5245

2 5246 NOTES J. CLN. MCROBOL. Group TABLE 1. Comparison of cutoff values selected for the cpcr assay and the TaqMan PCR assay cpcr (copies/10 5 PBL) Cutoff values for viral loads a Whole blood TaqMan PCR Copies/ml Copies/ g of DNA Copies/10 5 PBLs Plasma (copies/ml) , , , ,000 16, ,999 4,000 29,999 2,000 16, ,999 1, ,000 30,000 17,000 21,000 a Determined by PCR assay type, sample type, and unit of measurement. controls, and standards were tested in duplicate. PCR was performed using the AB Prism 7700 sequence detector (Applied Biosystems) and a thermal program consisting of 50 C for 2 min, 95 C for 10 min, and 50 two-step cycles of 95 C for 15 s and 60 C for 1 min. The FAM and VC reporter signals were measured automatically in real time, and the concentration of EBV DNA was determined from a standard curve (7). The viral load in plasma was expressed as the number of copies per milliliter, and that in whole blood was expressed as the number of copies per milliliter, number of copies per microgram of DNA based on spectrophotometric measurement, and number of copies per 10 5 PBLs based on an absolute lymphocyte count (when available). Results from each of these assays were also categorized into one of four load groups (Table 1). These groups were determined by preparing a cross-tabulation of each set of results from the TaqMan PCR assay, ranked from lowest to highest value, versus the group-categorized result from the cpcr assay and visually inspecting the crosstabulations repetitively to identify cutoff values at major divisions on the quantitative scales for the TaqMan assay yielding the highest number of concordant results with the group-categorized results for the cpcr assay. The viral load groups (i.e., to ) determined in whole blood and plasma by the TaqMan PCR assay were then compared to those determined in PBLs by the cpcr assay. The percent correlation between the Taq- Man and cpcr assays was defined as the (number of sample pairs with identical group-categorized results/total number of sample pairs) 100%. Cross-tabulations of data, Spearman s r 2 values, and statistical significance were determined with SPSS version 11 software (SPSS, nc., Chicago, ll.). This study TABLE 2. Correlation of EBV DNA load in whole blood determined by TaqMan PCR and expressed as copies per milliliter and in PBLs determined by cpcr a (copies/ml) b ( 150) (150 15,999) (16, ,999) ( 800,000) ( 8) 13 2 (8 100) (200 1,000) ( 1,000) 1 5 a Percent correlation 85.0% (n 60 pairs). Spearman r (P was approved by the nstitutional Review Board of the Children s Hospital of Pittsburgh. A total of 60 representative pairs of plasma and whole-blood samples were selected for testing. However, the volume of one of the plasma samples was less than the minimum needed for DNA isolation, and this sample was excluded. The samples had been obtained from 35 abdominal (i.e., kidney, liver, and/or intestine), 5 thoracic (i.e., heart and/or lung), and 4 abdominal and thoracic transplant recipients who were 1.6 to 25.4 years of age. A single set of plasma and whole-blood samples was obtained from 33 recipients, and two to four samples of each type were obtained at different time points from 11 recipients. The group-categorized results from the testing of whole blood by TaqMan PCR, expressed either as copies per milliliter, copies per microgram of DNA, or copies per 10 5 PBLs, correlated strongly with the group-categorized results from the testing of PBLs by the cpcr assay (Spearman r ) (Tables 2, 3, and 4). However, a much weaker correlation was observed between the group-categorized results from the testing of plasma by TaqMan PCR assay and the testing of PBLs by the cpcr assay (r ) (Table 5). None of the results comparing EBV viral loads determined on whole blood and PBLs differed by more than one load group (Tables 2 to 4). However, 15.3% of the results from the plasma-pbl pairs differed by two or more load groups (Table 5). Furthermore, the EBV DNA load, as measured by TaqMan PCR, was often dramatically higher in whole blood than in plasma (Fig. 1). To determine if the strength of the correlations could be affected by the magnitude of the PBL count, cross-tabulations of the group-categorized results from the TaqMan and cpcr TABLE 3. Correlation of EBV DNA load in whole blood determined by TaqMan PCR and expressed as copies per microgram of DNA and in PBLs determined by cpcr a (copies/ g of DNA) b ( 10) (10 3,999) (4,000 29,999) ( 30,000) ( 8) 14 1 (8 100) (200 1,000) 3 10 ( 1,000) 1 5 a Percent correlation 86.7% (n 60 pairs). Spearman r (P

3 VOL. 41, 2003 NOTES 5247 TABLE 4. Correlation of EBV DNA load in whole blood determined by TaqMan PCR and expressed as copies per 10 5 PBLs and in PBLs determined by cpcr a (copies/ b ( 10) (10 1,999) (2,000 16,999) ( 17,000) ( 8) 11 (8 100) (200 1,000) ( 1,000) 4 a Percent correlation 86.4% (n 44 pairs). Spearman r (P assays were determined in the subset of samples from which an absolute lymphocyte count had been determined on the same day that the PCR samples had been obtained. For the testing of whole-blood samples by TaqMan PCR and expression of the results as either copies per milliliter, copies per microgram of DNA, or copies per 10 5 PBLs versus the testing of PBLs by cpcr, the correlation coefficients for the group-categorized results were similar at the low- and high-lymphocyte strata (i.e., PBLs/liter and PBLs/liter, respectively) (Table 6). n contrast, the correlation between the viral load in plasma determined by TaqMan PCR versus the viral load in PBLs determined by cpcr was appreciably higher at the higher lymphocyte level. None of the samples derived from whole blood (n 60) or plasma (n 59) inhibited the amplification of the internal positive control in the TaqMan PCR assay. dentical results were also obtained with another set of 40 whole-blood samples obtained from transplant patients. These results confirm that the Qiagen method is effective in providing an inhibitor-free target for quantitative assessment by PCR (4). This cross-sectional, interlaboratory study identified a high correlation between the EBV DNA load measured in whole blood by TaqMan PCR and in PBLs by cpcr, indicating that the former method appears to be an acceptable alternative to the latter. We recommend expressing the load in whole blood as copies per milliliter rather than as copies per microgram of DNA or copies per 10 5 PBLs, since all three units of measurement correlate equally well with the load in PBLs determined TABLE 5. Correlation of EBV DNA load in plasma determined by TaqMan PCR and expressed as copies per milliliter and in PBLs determined by cpcr a ( 50) Plasma viral load ml) b (50 799) (800 20,999) ( 21,000) ( 8) (8 100) (200 1,000) ( 1,000) 2 3 a Percent correlation 42.4% (n 59 pairs). Spearman r (P by cpcr, and expression as copies per milliliter does not require a spectrophotometric analysis of the DNA concentration or a lymphocyte count and is therefore simpler than the other forms of measurement. Unexpectedly, the viral loads in whole blood measured by TaqMan PCR and expressed as copies per 10 5 PBLs were as much as 20 times higher than the values measured in PBLs by the cpcr assay. Although the reason for this difference is not known, it could be related to technical differences between the TaqMan and cpcr assays (e.g., PCR efficiency, purity of DNA samples, etc.). By necessity, transplant patients are sometimes monitored for EBV DNA load during their hospital stay and in the postdischarge period by different PCR n situations in which some of the testing is done in whole blood by TaqMan PCR and other testing is done in PBLs by cpcr, we believe that health care providers can use the EBV DNA load groups, described in Table 1, as a general comparative guide. n contrast to our results, Stevens et al. (11) recently reported a relatively low but a significant correlation between EBV DNA loads in whole blood determined by a real-time LightCycler-based PCR assay and by a cpcr assay, but observed poor results with a TaqMan PCR assay compared to the cpcr assay. Because the reasons for the apparent differences between the two studies are not known, but may be related to differences in a variety of techniques, including DNA purification, quality of standards, percentage of target volume in PCR mixture, etc., as well as selection of cutoff values, laboratories should proceed carefully in adopting or converting to a particular PCR system for quantification of EBV DNA load. Furthermore, efforts should be made to develop national standards that will allow for comparison of results of EBV viral load testing between laboratories regardless of the method used. The current literature does not identify the best specimen type to sample for EBV viral load monitoring (4, 5, 8 11). Despite a number of publications suggesting the potential utility of plasma specimens (1, 6 8), our results suggest that it may be advantageous to monitor the EBV DNA load in whole blood rather than plasma. Whole blood contains all EBV DNA (i.e., cell free and cell associated), whereas plasma contains EBV DNA from free virus as well as any episomal DNA resulting from the lysis of lymphocytes. Because the degree of lysis may vary depending on the EBV-specific cytotoxic T-cell response, the action of anti-cd20 monoclonal antibodies administered for treatment of PTLD, and the manipulation and storage of the sample, the testing of plasma can yield an underestimate of the total EBV DNA and may also lead to a higher degree of variability in DNA load measurement compared to the testing of whole blood. ndeed, our study showed that the viral load in whole blood is often much higher than it is in plasma (Fig. 1). However, our study did not clinically assess EBV illness in the study group of patients. n an earlier study (8), high EBV DNA loads were identified in plasma samples from several transplant patients with PTLD. Therefore, additional studies are needed to clarify the clinical significance of elevated EBV DNA loads in different blood compartments, in different types of transplant recipients, and in relation to the onset, treatment, and recurrence of EBV-associated PTLD. A potential problem may exist in interpreting results from

4 5248 NOTES J. CLN. MCROBOL. Downloaded from FG. 1. Correlation of EBV DNA load in plasma and whole-blood samples (n 59 pairs) measured by TaqMan PCR. Values less than 50 copies/ml were coded as 50 copies/ml. The diagonal line represents theoretical, perfect agreement between viral loads in plasma and whole blood. EBV DNA load assays using whole blood or plasma in patients with a very low or absent PBL count. Assays that test PBLs inherently control for this situation, whereas assays that test whole blood or plasma cannot easily do so. Accordingly, evaluation of a peripheral leukocyte count and differential would TABLE 6. Effect of the lymphocyte count in peripheral blood on the correlation between the TaqMan and cpcr assays Sample type and unit of measurement Spearman r 2 for group-categorized crosstabulation between TaqMan and cpcr assays by lymphocyte count: PBLs/liter (n 17 pairs) PBLs/liter (n 27 pairs) Whole blood Copies/ml Copies/ g of DNA Copies/10 5 PBLs Plasma (copies/ml) a be needed to confirm that a given patient had an adequate number of lymphocytes to validate a given result. t appears that this potential problem may be more of a concern in testing of plasma samples than in the testing of whole-blood samples, because our study observed that the correlation between viral loads in plasma and PBLs clearly improves when the lymphocyte count is PBLs/liter, whereas the magnitude of the lymphocyte count did not affect the correlations between the viral loads measured in whole blood and PBLs. However, our findings on the effect of the lymphocyte count are based on a small sample size, and additional confirmatory studies are needed. n conclusion, EBV DNA load values measured in whole blood by the TaqMan PCR assay correlate highly with the values measured in PBLs by the cpcr assay (9). Furthermore, the relative ease of use of TaqMan PCR and the minimal requirement of technical time for the assay should make it an attractive alternative to cpcr and may facilitate interlaboratory standardization. on January 29, 2019 by guest a This calculation was derived from 16 sample pairs, since 1 of the original 60 plasma samples was excluded, as described in the text. We thank Ronald Jaffe and Jorge Reyes for encouragement.

5 VOL. 41, 2003 NOTES 5249 REFERENCES 1. Fan, H., S. A. Schichman, L. J. Swinnen, J. M. Nicholls, P. A. Eagan, M. Luther, and M. L. Gulley Analytic validation of a competitive polymerase chain reaction assay for measuring Epstein-Barr viral load. Diagn. Mol. Pathol. 10: Green, M., J. Bueno, D. Rowe, G. Mazariegos, L. Qu, K. Abu-Almagd, and J. Reyes Predictive negative value of persistent low Epstein-Barr virus viral load after intestinal transplantation in children. Transplantation 70: Green, M., T. V. Cacciarelli, G. B. Mazariegos, L. Sigurdsson, L. Qu, D. T. Rowe, and J. Reyes Serial measurement of Epstein-Barr viral load in peripheral blood in pediatric liver transplant recipients during treatment for posttransplant lymphoproliferative disease. Transplantation 66: Kimura, H., M. Morita, Y. Yabuta, K. Kuzushima, K. Kato, S. Kojima, T. Matsuyama, and T. Morishima Quantitative analysis of Epstein-Barr virus load by using a real-time PCR assay. J. Clin. Microbiol. 37: Leung, E., B. K. Shenton, G. Jackson, F. K. Gould, C. Yap, and D. Talbot Use of real-time PCR to measure Epstein-Barr virus genomes in whole blood. J. mmunol. Methods 270: Limaye, A., M.-L. Huang, E. E. Atienza, J. M. Ferrenberg, and L. Corey Detection of Epstein-Barr virus DNA in sera from transplant recipients with lymphoproliferative disorders. J. Clin. Microbiol. 37: Lo, Y. M. D., L. Y. S. Chan, K. Lo, S. Leung, J. Zhang, A. T. C. Chan, J. C. K. Lee, N. M. Hjelm, P. J. Johnson, and D. P. Huang Quantitative analysis of cell-free Epstein-Barr virus DNA in plasma of patients with nasopharyngeal carcinoma. Cancer Res. 59: Niesters, H. G. M., J. van Esser, E. Fries, K. C. Wolthers, J. Cornelissen, and A. D. M. E. Osterhaus Development of a real-time quantitative assay for detection of Epstein-Barr virus. J. Clin. Microbiol. 38: Rowe, D. T., L. Qu, J. Reyes, N. Jabbour, E. Yunis, P. Putnam, S. Todo, and M. Green Use of quantitative competitive PCR to measure Epstein- Barr virus genome load in the peripheral blood of pediatric transplant patients with lymphoproliferative disorders. J. Clin. Microbiol. 35: Stevens, S. J. C.,. Pronk, and J. M. Middeldorp Toward standardization of Epstein-Barr virus DNA load monitoring: unfractionated whole blood as preferred clinical specimen. J. Clin. Microbiol. 39: Stevens, S. J. C., S. A. W. M. Verkuijlen, A. J. C. van den Brule, and J. M. Middeldorp Comparison of quantitative competitive PCR with Light- Cycler-based PCR for measuring Epstein-Barr virus DNA load in clinical specimens. J. Clin. Microbiol. 40: Webber, S., D. Rowe, G. Boyle, M. Michaels, A. Fitzsimmons, Y. Law, S. Miller, M. Green, G. Kurland, C. Bentlejeweski, A. Zeevi, and L. Qu Natural history of Epstein-Barr virus load in pediatric thoracic recipients with post transplant lymphoproliferative disorders and other primary EBV infections. Transplantation 67:215. Downloaded from on January 29, 2019 by guest

Comparison of Quantitative Competitive PCR with LightCycler-Based PCR for Measuring Epstein-Barr Virus DNA Load in Clinical Specimens

Comparison of Quantitative Competitive PCR with LightCycler-Based PCR for Measuring Epstein-Barr Virus DNA Load in Clinical Specimens JOURNAL OF CLINICAL MICROBIOLOGY, Nov. 2002, p. 3986 3992 Vol. 40, No. 11 0095-1137/02/$04.00 0 DOI: 40.11.3986 3992.2002 Copyright 2002, American Society for Microbiology. All Rights Reserved. Comparison