Quantification of Human Cytomegalovirus DNA in Bone Marrow ...

JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 2001, p. 4362–4369 0095-1137/01/$04.00⫹0 DOI: 10.1128/JCM.39.12.4362–4369.2001 Copyright © 2001, American Society for Microbiology. All Rights Reserved.

Vol. 39, No. 12

Quantification of Human Cytomegalovirus DNA in Bone Marrow Transplant Recipients by Real-Time PCR FRANK GRISCELLI,1* MICHEL BARROIS,2 SYLVIE CHAUVIN,1 STEPHANE LASTERE,1 DOMINIQUE BELLET,3 AND JEAN-HENRI BOURHIS4 Service de Microbiologie,1 Service de Ge´ne´tique Mole´culaire,2 Service d’Immunologie,3 and Service d’He´matologie Oncologique,4 Institut Gustave-Roussy, 94805 Villejuif Cedex, France Received 24 May 2001/Returned for modification 22 August 2001/Accepted 18 September 2001

A real-time PCR assay was developed to quantify human cytomegalovirus (CMV) DNA in peripheral blood leukocytes (PBLs) of bone marrow transplantation patients. Unlike other teams, we quantified CMV and the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene using a plasmid containing both sequences as an external standard. Tenfold serial dilutions of this plasmid yielded overlapping standard curves that allowed the quantification of CMV and GAPDH gene copies in an efficient and accurate manner. Sequential blood samples (164 specimens) were collected from 16 patients. PBLs were tested by the pp65 antigenemia assay and quantitative CMV and GAPDH gene PCRs. CMV DNA was detected by PCR in 13 patients a mean of 15 days prior to the appearance of antigenemia. The administration of anti-CMV drugs led to a rapid decrease in the numbers of viral copies and positive nuclei. Real-time PCR assay results correlated with those of the CMV pp65 antigenemia assay (P < 0.00001). The TaqMan assay may be a useful tool for rapid quantification of CMV infection and for monitoring of CMV reactivation in bone marrow transplantation recipients. clinical symptoms, this method (6, 20) poses a number of problems. It is difficult to perform before engraftment because the number of leukocytes is limited and false-negative results are obtained due to the poor sensitivity of the technique and the weak expression of the pp65 antigen in white blood cells in some patients who develop CMV disease (18). Alternatively, quantitative PCRs based on TaqMan technologies for detection of CMV reactivation after BMT have been investigated (14, 22), but to date, none of the techniques described have been adequately standardized since the numbers of CMV copies were never normalized by quantification of a housekeeping gene. The aim of our study was to design a quantitative PCRbased assay capable of quantifying the CMV load in PBLs by two independent, quantitative PCR methods. The first PCR technique measured the CMV genome copy number by using a target sequence located in the UL83 gene, which codes for pp65. The second PCR technique quantified the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene in order to normalize the CMV DNA loads in the samples. The originality of our study is that we used a plasmid (Pic19Rpp65/GAPDH) that contains both CMV and GAPDH DNA fragments to quantify CMV DNA. We used these tools to assess the usefulness of real-time automated PCR as a quantitative, highly reproducible, and sensitive method for the detection of CMV DNA in PBLs and to evaluate the extent to which it was correlated with the CMV antigenemia assay in BMT recipients.

Human cytomegalovirus (CMV) is a well-known cause of mortality in blood and bone marrow transplantation (BMT) patients. Monitoring of CMV reactivation from latency is critical for these patients. Prophylaxis against CMV disease with ganciclovir (8, 9) or foscarnet (16) in asymptomatic BMT recipients has been shown to dramatically reduce the incidence of CMV disease (8, 17). As ganciclovir and foscarnet are myelotoxic and nephrotoxic, respectively, full treatment is often started at the time of documented CMV reactivation. The key to efficient and effective management of CMV infection in these patients is a test capable of rapidly monitoring and quantifying the presence of CMV in the blood. This is particularly essential for the identification of subjects at high risk of developing CMV disease, e.g., patients receiving steroid or immunosuppressive compounds for accelerated graft-versus-host disease and also for the application and monitoring of preemptive antiviral therapeutic strategies. The CMV assays presently available and frequently used in this setting include shell vial culture (7), the CMV antigenemia assay (2), PCR for CMV DNA (3), hybrid capture assay for quantitation of CMV DNA (10), and detection of CMV RNA by nucleic acid sequencebased amplification (1). Among white blood cells, peripheral blood leukocytes (PBLs) are the main CMV carriers during active CMV infection. The detection of CMV antigenemia in PBLs has been shown to be an early marker of CMV infection. A monoclonal antibody is used to detect pp65, the CMV lower matrix phosphoprotein in PBLs, and this test is widely used to monitor BMT recipients. Although a correlation was found between the number of pp65-positive PBLs and the development of

MATERIALS AND METHODS Patients and samples. Sixteen patients who underwent BMT between April 1998 and March 2000 were examined. Antigenemia assays were performed in real time, and the results were used to determine each patient’s treatment. CMV DNA was retrospectively quantified in PBLs in patients with either symptomatic or asymptomatic infection. In the present study, 164 blood samples for the CMV antigenemia assay and CMV DNA detection were drawn twice a week from the onset of aplasia until the third month posttransplantation. Thirty-three additional samples from BMT recipients were used to compare both methods. Thus,

* Corresponding author. Mailing address: Laboratoire de Microbiologie Me´dicale, Institut Gustave-Roussy, 39 rue Camille Desmoulins, 94805 Villejuif Cedex, France. Phone: (33-1) 42 11 51 93. Fax: (33-1) 42 11 53 13. E-mail: [email protected]. 4362

VOL. 39, 2001

CMV DNA QUANTIFICATION IN BMT PATIENTS BY PCR

the comparison of real-time PCR and the pp65 antigenemia assay was performed with 128 PBL samples which were positive by PCR and/or by the antigenemia assay. According to our present CMV prevention strategy, between 25 and 35 days post-BMT, patients underwent a systematic bronchoscopy to obtain bronchoalveolar lavage (BAL) fluid samples for CMV cultures (4, 11). All patients received prophylactic acyclovir at a dose of 10 mg/kg of body weight every 8 h from day ⫺1 until discharge from the BMT unit or the beginning of anti-CMV therapy, initiated at the time of documented CMV reactivation on the basis of the results of the pp65 antigenemia assay. For all patients, prophylaxis against graft-versus-host disease consisted of cyclosporine A on the day before the graft and a short course of methotrexate on days 1, 3, and 6 posttransplantation. Sample preparation. To prepare leukocytes, cells were separated by sedimentation of 3 to 5 ml of a heparinized blood sample in a 6% dextran solution over 20 min, centrifuged at 220 ⫻ g for 10 min at 20°C, and washed once with phosphate-buffered saline (PBS). The resulting pellet was washed with 2 ml of erythrocyte lysing solution for 2 min and washed once with PBS. Aliquots of 200,000 leukocytes were used for the pp65 antigen test, and 106 cells were used for DNA extraction, performed with a QIAamp Blood mini-kit (Qiagen, Valencia, Calif.). The DNA absorbed on the spin column was eluted with 50 ␮l of DNase-free distilled water and was then submitted to PCR. The concentration of the extracted DNA was quantified by spectrophotometric measurement at a wavelength of 260 nm. CMV antigenemia assays. The CMV antigenemia assays (15) were performed by indirect immunofluorescence detection of pp65 (65 to 68 kDa), the human CMV internal matrix phosphoprotein in PBLs, by standard procedures (CINA kit; Argene Biosoft, Varilhes, France). Briefly, for each patient three cytospin slides with 200,000 cells per glass slide were prepared. The PBLs were fixed and permeabilized to allow subsequent detection of the CMV pp65 antigen. The presence of the CMV pp65 antigen was detected with the 1C3/AYM-1 antibody cocktail and was visualized with a specific secondary antibody. The number of CMV antigen-positive cells was counted on duplicate stained slides, and the results were reported as the number of positively staining cells per 200,000 leukocytes. Real-time quantification PCR. The sequences of the PCR primers and that of the probe used to quantify CMV were selected from the phosphorylated matrix protein (pp65) gene (locus HSPPBC; GenBank) with Primer Express software (Perkin-Elmer Biosystems, Foster City, Calif.). The sequences of the forward and reverse primers were 5⬘-GCAGCCACGGGATCGTACT and 5⬘-GGCTTT TACCTCACACGAGCATT, respectively. The TaqMan probe selected between the primers was fluorescence labeled at the 5⬘ end with 6-carboxyfluorescein (FAM) as the reporter dye and at the 3⬘ end with 6-carboxytetramethylrhodamine (TAMRA) as the quencher (5⬘-FAM-CGCGAGACCGTGGAACTGCGTAMRA). The PCR product was detected as an increase in fluorescence with the ABI PRISM 7700 instrument (Perkin-Elmer Biosystems). PCR was performed with 25 ␮l of TaqMan Universal PCR master mixture (Perkin-Elmer Biosystems), each of the primers at a concentration of 400 nM, 100 nM TaqMan probe, and 100 ng of DNA in a total volume of 50 ␮l. PCR was performed in 96-well microtiter plates under the following conditions: after 2 min at 50°C and 10 min at 95°C, the samples were submitted to 45 cycles, with each cycle consisting of a step at 95°C for 15 s, followed by a step at 60°C for 1 min, for both CMV and GAPDH amplification. The human genomic sequence was quantified with PCR primers and the TaqMan probe that recognized the GAPDH gene. The upstream and downstream primer sequences were 5⬘-CTCCCCACACACATGCA CTTA and 5⬘-CCTAGTCCCAGGGCTTTGATT, respectively, and the fluorogenic probe located between the PCR primers and labeled with VIC and TAMRA was synthesized by PE Biosystems (5⬘-VIC-AAAAGAGCTAGGAAG GACAGGCAACTTGGC-TAMRA). The GAPDH gene PCR was performed under the same PCR conditions described above for the CMV PCR. A plasmid containing both targeted sequences was used as a standard in the present study. This plasmid (Pic19Rpp65/GAPDH) contained a 448-bp DNA fragment derived from the pp65 gene and a 538-bp fragment derived from the human GAPDH gene, which were generated by conventional PCR with Pfu polymerase (Stratagene, Amsterdam, The Netherlands). A standard graph of the cycle threshold (CT) values obtained from serial dilutions (10 to 104 copies/well) of the plasmid was constructed for both CMV and the GAPDH gene. The 10-fold dilutions of the plasmid were concocted with a solution of salmon sperm DNA as a carrier at a final concentration of 100 ng of DNA per sample. The CT values from unknown samples were plotted on the respective standard curves, and the number of CMV genome copies per 200,000 leukocytes was calculated with Sequence Detection System software (version 1.6.3; Perkin-Elmer Biosystems). Each sample of DNA extract and serial dilutions of the plasmid were analyzed in duplicate. To control for cross-contamination, a sample consisting of distilled water was also submitted

4363

to the DNA extraction procedure, and the resulting extract was amplified in duplicate. Samples were considered negative if the CT values exceeded 45 cycles.

RESULTS TaqMan PCR design: PCR specificity, sensitivity, and reproducibility. Our goal was to design a quantitative PCRbased assay capable of quantifying the CMV load in human cells. Two independent, quantitative PCR methods were used and performed with two separate aliquots of the same DNA extract. The CMV TaqMan PCR is based on the amplification of a 159-bp region of a sequence located in the UL83 gene, which codes for the lower matrix protein detected in the pp65 antigenemia test. A nonvariable region of the pp65 gene, common to strains AD149 and Towne, was chosen. The GAPDH gene PCR is based on the amplification of a 99-bp region of human genomic DNA. In order to confirm the specificity of this assay, several virus strains were tested. No cross-reactivity between CMV and herpes simplex virus types 1 and 2, varicella-zoster virus, Epstein-Barr virus, and human herpesviruses 6 and 8 was observed (data not shown). The sensitivities of both PCR techniques were then compared by using for each sample 10 to 104 copies of the control plasmid diluted in salmon sperm DNA. Each sample was submitted to the CMV and GAPDH gene real-time PCRs, and amplification was repeated four times for each dilution. A standard curve of the CT values plotted against the logarithm of the copy number was constructed. CMV and GAPDH gene quantification was linear over a wide range (from 10 to 104 copies per well). The detection rates were 100% for both PCRs when the copy number was ⱖ10 copies per well; and the detection rates were 60 and 40% for the GAPDH gene and CMV PCRs, respectively, when the copy number was 1 copy per well, which is in agreement with the values that can be estimated from the Poisson probabilities. As shown in Fig. 1A, B, and D, the amplification yield and detection rates were comparable when plasmid dilutions were submitted to both PCR techniques, with CT values of 38.9 ⫾ 0.67 (values are means ⫾ standard deviations) and 38.71 ⫾ 0.69 for the amplification of 10 copies of the GAPDH gene and CMV, respectively. In this test of intra-assay variation, the coefficients of variation (CVs) were 0.17, 0.13, 0.32, and 1.71% for 104, 103, 102, and 10 copies/well, respectively, for the GAPDH gene PCR and 0.25, 0.55, 1.03, and 1.79% for 104, 103, 102, and 10 copies/well, respectively, for the CMV PCR. As shown in Fig. 1C, the linear correlations between the CT value and the logarithm of the DNA copy number were identical for both PCRs. The slopes were ⫺3.70 and ⫺3.62 for the GAPDH gene and CMV PCRs, respectively, and the correlation coefficient was identical for both PCRs (R2 ⫽ 0.9968). To estimate interexperiment variability and to measure the accuracies of both the CMV and the GAPDH gene PCRs, each plasmid dilution (104 to 10 copies) was submitted in duplicate to both PCRs in 10 distinct experiments. The CVs were less than 2% for DNA inputs of 104, 103, and 102 copies/well and less than 3.3% for DNA inputs of 10 copies/well for both the CMV and the GAPDH gene TaqMan PCRs. Thus, the CT values were similar for the CMV and GAPDH gene PCRs: 27.15 ⫾ 0.34 and 27.79 ⫾ 0.55, respectively, for DNA inputs of 104 copies/well, 30.56 ⫾ 0.38 and 30.9 ⫾ 0.32, respectively, for

4364

GRISCELLI ET AL.

J. CLIN. MICROBIOL.

FIG. 1. Amplification plots obtained with the control plasmid for GAPDH gene (A) and CMV (B) PCRs (⌬Rn, normalized reporter fluorescence signal). Serial 10-fold dilutions with 104 to 10 copies per reaction well were amplified for 45 cycles, and amplifications were repeated four times for each dilution. (C) Standard curves for CMV and GAPDH gene real-time PCRs. CT values were plotted against the normalized fluorescence signal. The correlation coefficient was 0.9968 for both PCRs, and the slopes were ⫺3.7 and ⫺3.6 for the GAPDH gene and CMV PCRs, respectively. (D) After amplification of 104 to 0 copies of the plasmid, the PCR products were loaded onto an ethidium bromide-stained 2% agarose gel.

DNA inputs of 103 copies/well, 34 ⫾ 0.42 and 34.68 ⫾ 0.62, respectively, for DNA inputs of 102 copies/well, and 37.06 ⫾ 0.95 and 37.88 ⫾ 1.20, respectively, for DNA inputs of 10 copies/well. Patient characteristics. Patient characteristics are summarized in Table 1. There were 10 females and 6 males, with a median age of 34 years (age range, 8 to 52 years). Conditioning for allogeneic BMT consisted of total body irradiation (n ⫽ 12) or busulfan (n ⫽ 3) combined with endoxan, melphalan, etoposide, and/or cytarabine. One patient received a reduced conditioning regimen with busulfan, endoxan, anti-T globulin, and fludarabine. Fifteen patients were seropositive for CMV before BMT, and they received a graft either from seronegative donors (n ⫽ 4) or from seropositive donors (n ⫽ 11). Patient 16 and his donor were seronegative. The BAL fluid samples of two patients were positive for CMV, and the patients were immediately prescribed a 14-day course of intravenous ganciclovir at a dose of 10/mg/kg/day. PCR quantification of CMV DNA and comparison with pp65 antigenemia assay. We tested 197 PBL samples from BMT recipients to evaluate the CMV DNA TaqMan assay with clinical specimens. CMV DNA was quantified in parallel

with the GAPDH gene in PBL samples in order to determine the amount of cellular DNA input into each reaction mixture. In our study, the concentration of the extracted DNA was quantified by spectrophotometric measurement at a wavelength of 260 nm; this measurement was only an approximation of the DNA concentration in each sample, i.e., 330 copies per ng of genomic DNA, in theory. The mean value of the GAPDH gene copy number for the 197 PBL samples tested was 87,974 copies (range, 24,779 to 296,295 copies), which corresponded to 2.6 times (range, 0.75 to 8.9 times) the GAPDH gene copy number expected for the amount of cellular DNA input into the reaction mixture. Of the 197 samples tested, 69 were negative by both techniques. The 128 samples that were PCR positive were classified into three groups according to the results of the pp65 antigenemia assay. Samples in group 1 (n ⫽ 66) were negative for pp65 antigenemia, samples in group 2 (n ⫽ 37) had ⬍10 pp65-positive cells, and samples in group 3 (n ⫽ 25) had ⱖ10 pp65-positive cells. As shown in Fig. 2, the samples in group 3 had significantly higher CMV DNA copy numbers than the numbers in the pp65positive samples in group 2 (P ⫽ 0.001 by the Student test) and the pp65-negative samples in group 1 (P ⬍ 0.0001 by the

VOL. 39, 2001


4365

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Patient no.

41/F 15/M 33/M 41/F 40/F 27/F 49/F 15/M 47/F 52/F 28/F 38/M 44/M 26/F 42/F 8/M

Age (yr)/ Sex

RAEB SAA CML AML5 Myeloma ALL Myeloma ALL Myeloma AML CML ALL NHL ALL AML ALL

Disease Status

Transformation

CP1 CR1 Postautograft relapse Belated relapse CR1 Induction refractory PR1 CR2 CP1 CR1 CR2 CR1 CR1 CR2

Conditioning

sibling sibling sibling sibling sibling sibling sibling sibling sibling sibling sibling sibling sibling sibling

Donor type

MUD HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical HLA-identical MUD

TABLE 1. Patient characteristicsa

Busulfan, EDX EDX, ATG TBI, EDX TBI, EDX TBI, melphalan TBI, melphalan Busulfan, Fluda/ATG TBI, EDX TBI, melphalan TBI, EDX TBI, EDX TBI, EDX, VP16 TBI, EDX TBI, EDX Busulfan, EDX TBI, CYT, melphalan Bone marrow Bone marrow Bone marrow Bone marrow Blood cells Bone marrow Blood cells Bone marrow Blood cells Blood cells Bone marrow Bone marrow Bone marrow Bone marrow Blood cells Bone marrow

Graft type

⫺/⫹ ⫺/⫹ ⫹/⫹ ⫺/⫹ ⫹/⫹ ⫹/⫹ ⫹/⫹ ⫺/⫹ ⫹/⫹ ⫹/⫹ ⫹/⫹ ⫹/⫹ ⫹/⫹ ⫹/⫹ ⫹/⫹ ⫺/⫺

D/R status

No No No No No Lung No No No No No No No Lung No No

CMV disease Day p.g. of CMV disease

29 (BAL)

34 (BAL)

Time of anti-CMV therapy S/D

37/59 70/46 84/24 149/14 50/14 28/36 220/26 106/14 66/14 44/23 34/50 55/14 59/25 35/14 43/14

Outcome

Dead Alive Alive Dead Dead Dead Alive Alive Alive Dead Alive Alive Alive Dead Alive Alive

a Abbreviations: F, female; M, male; CML, chronic myelocytic leukemia; ALL, acute lymphoblastic leukemia; SAA, severe aplastic anemia; AML, acute myeloid leukemia; RAEB, refractory anemia with excess blasts; NHL, non-Hodgkin’s lymphomas; CP1, first chronic phase; CR1, first complete response; PR1, first partial response; CR2, second complete response; TBI, total body irradiation; EDX, endoxan; VP16, etoposide; CYT, cytarabine; ATG, anti-T globulin; Fluda, fludarabine; MUD, matched unrelated donor; D/R, donor CMV serology/recipient CMV serology; S/D, day of beginning of treatment/number of days of treatment; p.g., postgrafting.

FIG. 2. Comparison of CMV DNA loads by the pp65 antigenemia assay in PBL samples. PCR-positive samples (n ⫽ 128) were classified into three groups according to the results of the pp65 antigenemia assay. Samples in group 1 (n ⫽ 66) were negative for pp65 antigenemia, samples in group 2 (n ⫽ 37) had ⬍10 pp65-positive cells, and samples in group 3 (n ⫽ 25) had ⱖ10 pp65-positive cells. The samples in group 3 had significantly higher CMV DNA copy numbers than the pp65-positive samples in group 2 (P ⫽ 0.001 by the Student test) and the pp65-negative samples in group 1 (P ⬍ 0.0001 by the Student test). Bars show the mean CMV DNA copy numbers.

Student test). The mean CMV DNA copy number per 2 ⫻ 105 PBLs was 138 (range, 4 to 3,444) for samples in group 1, 296 (range, 7 to 1,770) for samples in group 2, and 2,164 (range, 36 to 14,446) for samples in group 3. Furthermore, a statistically significant correlation was observed between the CMV DNA copy number and the number of pp65-positive cells (n ⫽ 128; r ⫽ 0.425; P ⬍ 0.00001 by the Spearman rank test). Figure 3 shows on a logarithmic graph the results of a comparison of both techniques in which samples which were negative for CMV DNA or pp65 antigenemia (n ⫽ 66 samples from group 1) were not included. Analysis of BMT patients. The quantification of CMV DNA was performed retrospectively with 164 samples from 16 BMT recipients tested for the pp65 antigen, with a median number of 10 samples (range, 7 to 14 samples) available from each patient. Table 2 shows the results of monitoring of CMV reactivation by the pp65 antigen test and quantification of CMV DNA by PCR. As shown in Table 2, 42 of 164 samples (25.3%) were positive for CMV DNA and pp65 antigen detection, while 69 of 164 samples (41.5%) were negative by both tests. Fifty-three pp65 antigen-negative samples (32%) were PCR positive. After BMT, the CMV DNA PCR was positive for 13 of 15 patients a mean of 15 days before the pp65 antigenemia assay was positive (84 ⫾ 69 and 69 ⫾ 65 days for the pp65 antigenemia assay and PCR, respectively). For two patients

4366

GRISCELLI ET AL.

J. CLIN. MICROBIOL.

DISCUSSION

FIG. 3. Correlation between CMV DNA copy number and the number of pp65-positive cells in PBLs on the basis of the results for 62 samples which were positive by both the antigenemia assay and the PCR assay. The CMV DNA copy number was plotted on a logarithmic graph against the number of pp65-positive cells detected by the antigenemia assay. The correlation between the CMV DNA copy number and the number of pp65-positive cells was examined by the Spearman rank test and was found to be highly significant, with a correlation coefficient of 0.425 (P ⬍ 0.00001).

(patients 3 and 11) the two tests were positive simultaneously. Patient 16, who had a negative CMV serology, as did his donor, was negative by both techniques. All 15 patients received antiviral therapy, which was started on the day that positivity for pp65 antigenemia was documented. A single course of ganciclovir therapy led to marked decreases in CMV DNA loads and levels of antigenemia, but significant differences were observed among patients. Eight patients (seven CMV-positive patients with CMV-positive donors and one CMV-negative patient with a CMV-positive donor) responded promptly to a single 14- to 59-day course of therapy, and both the pp65 antigenemia assay and the CMV DNA PCR became negative after ganciclovir therapy. For six patients (patients 3, 6, 7, 8, 9, and 11), both techniques became negative on the same day posttreatment. For patients 10 and 12, the pp65 antigenemia assay was negative before negativity by the CMV DNA PCR, with CMV DNA still being detectable 14 and 21 days posttreatment, respectively. Seven patients exhibited asymptomatic CMV reactivation after the first course of ganciclovir therapy, and a second pp65 antigenemia and/or CMV DNA peak was observed. Among these seven patients, the second peak was observed by both techniques in five patients (patients 1, 2, 4, 5, and 14) but only by PCR in the other two patients (patients 13 and 15). In five patients exhibiting CMV reactivation (patients 1, 2, 5, 13, and 15), CMV DNA was still detectable 111, 151, 88, 73, and 69 days posttransplantation, respectively.

The development of real-time PCR technology is a promising advance in the quantification of CMV DNA in clinical samples and will be useful for the monitoring of CMV reactivation in patients at high risk of developing CMV disease. In our study, we assessed PBLs for increased levels of CMV DNA replication, and an increased level of CMV DNA replication in PBLs was shown to be the first manifestation of CMV reactivation, which is consistent with previous findings (12). As PBLs were tested, the variations in the efficiency of the DNA extraction step and in the measurement of DNA levels by spectrophotometry could have major effects on the reproducibility of the results. Consequently, and as proposed by others, quantification of a housekeeping gene to normalize CMV loads is essential to guarantee highly reproducible PCR results (5, 14). Unlike other studies, we used a plasmid that harbors both sequences as an external standard to quantify CMV DNA and genomic DNA. We have shown that serial dilutions of this plasmid allowed the quantification of CMV and GAPDH gene copy numbers in an efficient and accurate manner, since the standard curves for CMV and GAPDH gene copy numbers were identical (Fig. 1C). Furthermore, our assay was highly reproducible, as indicated by the intra-assay and interassay CV values obtained with the standard plasmid. In our study, the intra-assay variation values were 0.17, 0.13, 0.32, and 1.71% for samples containing 104, 103, 102, and 10 copies, respectively, for the GAPDH gene PCR and 0.25, 0.55, 1.03, and 1.79% for samples containing 104, 103, 102, and 10 copies, respectively, for the CMV PCR. The interassay variations were less than 2% for DNA inputs of 104, 103, and 102 copies/well and less than 3.3% for DNA inputs of 10 copies/well for both the CMV and the GAPDH gene TaqMan PCRs. We consider that these variations will be negligible in clinical use. The real-time PCR technique used in the present study allowed the quantification of CMV DNA over a wide dynamic range for both CMV and GAPDH gene amplification (10 to 104 and 10 to 105 copies of plasmid, respectively). Furthermore, we found a good positive correlation between the number of copies of CMV DNA and the GAPDH gene present and the number of cycles at which the amplification curves became steep using serial dilutions of plasmid Pic19Rpp65/GAPDH. In this 10-fold dilution range, we were able to accurately determine the number of copies of viral and genomic DNAs present in all samples used in our study. In agreement with the results of other studies that used other TaqMan-based assays (5, 13, 19), we found a significant correlation between the results of the antigenemia assay and DNA copy numbers in PBL samples, but we nevertheless observed inconsistencies. Interestingly, 32% of the samples were pp65 antigenemia assay negative and PCR positive. Furthermore, samples with low levels of antigenemia and a large number of copies of CMV DNA were observed. Similar differences between the results of an antigenemia assay and quantitative PCR have been reported by others (5, 14, 21), but the clinical significance of large CMV DNA copy numbers with a low or a negative antigenemia assay result remains to be elucidated. We have shown that PCR quantification of CMV DNA is more sensitive than the antigenemia assay for the monitoring

VOL. 39, 2001


4367

TABLE 2. Comparison of CMV antigenemia performed by indirect immunofluorescence detection of the 65- to 68-kDa internal matrix phosphoprotein and the CMV TaqMan quantitative PCR results for 16 patients who underwent BMT Patient

1

2

3

4

5

6

Day postgrafting

No. of pp65-staining cells/ 200,000 leukocytes

No. of CMV copies/ 200,000 leukocytes

16 19 26 37 40 48 51 56 58 69 97 100 104 111

0 0 0 7 30 15 25 0 0 10 180 36 0 0

0 0 12 722 149 36 284 95 127 373 11,040 552 360 149

56 59 63 67 70 81 88 95 102 109 130 133 140 151

0 0 0 0 11 4 4 0 1 1 2 1 0 0

0 0 52 218 1206 138 324 13 295 52 776 69 18 6

42 59 73 84 87 91 94 101 119 129 134

0 0 0 1 4 5 0 0 0 0 0

0 0 0 272 491 1,670 0 0 0 0 0

105 119 129 138 149 152 170 209 213 216 242

0 0 0 0 6 0 0 2 0 0 0

0 0 51 77 78 14 0 10 0 0 0

⫺6 ⫺3 39 50 53 64 71 74 81 88 17 21

0 0 0 3 7 0 1 4 8 0 0 0

0 0 6 285 17 345 401 342 7 2 0 40

Day postgrafting



28 38 42 45 49 52 56 59

2 3 2 0 2 0 0 0

53 110 381 3,444 34 0 0 0

7

181 198 206 213 220 227 231 234 238 241

0 0 0 0 3 0 1 0 0 0

4 12 0 59 17 0 6 0 0 0

8

78 84 98 101 106 109 113 127 130 137

0 0 0 0 1 0 0 0 0 0

0 0 16 11 22 0 0 0 0 0

9

34 38 41 66 76 79 83 90 93

0 0 0 5 0 0 0 0 0

0 11 6 3 0 0 0 0 0

10

23 44 47 51 58 72 82

0 25 2 0 0 0 0

26 51 91 255 662 0 0

11

⫺1 34 42 45 49 52 56 66 70 73

0 3 0 0 0 2 0 0 0 0

0 159 0 0 0 8 0 0 0 0

12

27 34 41 44 48

0 0 0 0 0

0 0 0 116 200

Patient

Continued on following page

4368

J. CLIN. MICROBIOL.

GRISCELLI ET AL. TABLE 2—Continued Day postgrafting



55 63 76 83 90

4 0 0 0 0

1,770 31 20 0 0

13

38 42 45 52 56 59 66 70 73

0 0 0 0 0 14 0 0 0

20 27 34 137 376 748 42 207 39

14

20 36 63 68 78 89 92 99 103 110 113 120

0 0 1 1 0 0 0 0 0 2 0 0

0 198 158 126 90 38 126 257 69 42 0 0

15

22 26 29 40 43 48 51 55 65 69

0 0 0 0 3 0 0 0 0 0

0 0 6 107 178 19 21 11 178 30

16

48 79 87 101 122 182 207

0 0 0 0 0 0 0

0 0 0 0 0 0 0

Patient

of CMV reactivation in BMT patients. As shown in Table 2, CMV DNA was detected by PCR in 13 of 15 patients a mean of 15 days prior to the detection of antigenemia. All 15 patients were administered anti-CMV drugs when the antigenemia assay was positive. The antigenemia assay and the real-time automated PCR both showed rapid decreases in the numbers of viral copies and pp65-positive nuclei once treatment with ganciclovir was initiated. In our opinion, real-time PCR showed greater sensitivity than the antigenemia assay in reflecting the effect of ganciclovir since the samples became negative by the antigenemia assay before the quantification of CMV DNA in more than 50% of the patients. Furthermore, real-time PCR will permit more reliable screening than the antigenemia assay for patients who develop ganciclovir-resistant CMV, in whom CMV DNA replication is not inhibited.

All but two patients had detectable CMV DNA and pp65positive nuclei in the absence of clinical manifestations of the disease. A factor that compounds difficulties in correlating the amounts of CMV DNA with the development of active disease is the widespread use of preemptive therapy. Two patients in whom CMV DNA was detected before antigenemia (by 7 and 37 days, respectively) developed active CMV-induced lung disease 29 and 34 days posttransplantation, respectively. Furthermore, in one patient (patient 6), the presence of infectious CMV in BAL fluid was detected at day 29 posttransplantation, whereas CMV DNA was quantified 7 days before the BAL. These findings support the use of quantitative PCR to assess when preemptive therapy should be initiated. Furthermore, the early detection of CMV DNA in PBLs should spare these patients from undergoing BAL, which is generally carried out between 25 and 35 days posttransplantation. On the basis of our findings, we believe that real-time PCR promises to be an interesting alternative test to the antigenemia assay presently in use for the monitoring of CMV disease in BMT patients. We are therefore conducting a prospective study with a large patient population in which half of the dose of preemptive antiviral therapy is administered when CMV reactivation is documented by the CMV TaqMan PCR and until CMV DNA is no longer detectable. This approach will avoid the myelotoxicity and the nephrotoxicity caused by the drugs usually prescribed (ganciclovir and foscarnet, respectively) and should therefore significantly improve the clinical conditions of these patients. In summary, we have presented an accurate and rapid PCR assay for the quantification of CMV DNA which may prove useful for routine clinical testing. We have shown that the standardization of the technique requires the amplification of a cellular gene to monitor the efficiency of the reaction. Since both CMV- and GAPDH gene-specific probes are labeled with two different fluorogenic dyes (FAM and VIC, respectively), a multiplex PCR technique will be developed in order to decrease the cost of the TaqMan PCR. ACKNOWLEDGMENTS We are grateful to all the staff of the Microbiology Laboratory of the Institut Gustave-Roussy and to L. Saint Ange and M. Mackenthun for editing. We thank E. Dussaix and A.-M. Roque for providing us with herpes simplex virus type 1 and 2, varicella-zoster virus, Epstein-Barr virus, and human herpesvirus 6 and 8 DNAs. This study was supported by a CRC (Contrat de Recherche Clinique) grant from Institut Gustave-Roussy. REFERENCES 1. Blank, B. S., P. L. Meenhorst, J. W. Mulder, G. J. Weverling, H. Putter, W. Pauw, W. C. Van Dijk, P. Smits, S. Lie-A-Ling, P. Reiss, and J. M. Lange. 2000. Value of different assays for detection of human cytomegalovirus (HCMV) in predicting the development of HCMV disease in human immunodeficiency virus-infected patients. J. Clin. Microbiol. 38:563–569. 2. Boeckh, M., T. A. Gooley, D. Myerson, T. Cunningham, G. Schoch, and R. A. Bowden. 1996. Cytomegalovirus pp65 antigenemia-guided early treatment with ganciclovir at engraftment after allogeneic marrow transplantation: a randomized double-blind study. Blood 10:4063–4071. 3. Einsele, H., M. Steidle, A. Vallbracht, J. G. Saal, G. Ehninger, and C. A. Muller. 1991. Early occurrence of human cytomegalovirus infection after bone marrow transplantation as demonstrated by the polymerase chain reaction technique. Blood 77:1104–1110. 4. Fajac, A., F. Stephan, A. Ibrahim, E. Gautier, J. F. Bernaudin, and J. L. Pico. 1997. Value of cytomegalovirus detection by PCR in bronchoalveolar lavage routinely performed in asymptomatic bone marrow recipients. Bone Marrow Transplant. 20:581–585. 5. Gault, E., Y. Michel, A. Dehee, C. Belabani, J. Nicolas, and A. Garbarg-

VOL. 39, 2001

6.

7.

8.

9. 10.

11.

12.

13.


Chenon. 2001. Quantification of human cytomegalovirus DNA by real-time PCR. J. Clin. Microbiol. 39:772–775. Gerna, G., D. Zipeto, E. Percivalle, M. Parea, M. G. Revello, R. Maccario, G. Peri, and G. Milanesi. 1992. Human cytomegalovirus infection of the major leukocyte subpopulations and evidence for initial viral replication in polymorphonuclear leukocytes from viremic patients. J. Infect. Dis. 166:1236– 1244. Gleaves, C. A., T. H. Smith, E. A. Shuster, and G. R. Person. 1984. Rapid detection of cytomegalovirus in MRC-5 cells inoculated with urine specimens by using low-speed centrifugation and monoclonal antibody to an early antigen. J. Clin. Microbiol. 19:917–919. Goodrich, J. M., M. Mori, C. A. Gleaves, C. Du Mond, M. Cays, D. F. Ebeling, W. C. Buhles, B. DeArmond, and J. D. Meyers. 1991. Early treatment with ganciclovir to prevent cytomegalovirus disease after allogeneic bone marrow transplantation. N. Engl. J. Med. 325:1601–1607. Goodrich, J. M., R. A. Bowden, L. Fisher, C. Keller, G. Schoch, and J. D. Meyers. 1993. Ganciclovir prophylaxis to prevent cytomegalovirus disease after allogeneic marrow transplant. Ann. Intern. Med. 118:173–178. Hebart, H., D. Gamer, J. Loeffer, C. Mueller, C. Sinzger, G. Jahn, P. Bader, T. Klingebiel, L. Kanz, and H. Einsele. 1998. Evaluation of Murex CMV DNA hybrid capture assay for the detection and quantification of cytomegalovirus infection in patients following allogeneic stem cell transplantation. J. Clin. Microbiol. 36:1333–1337. Ibrahim, A., E. Gautier, S. Roittmann, J. H. Bourhis, A. Fajac, I. Charnoz, P. Terrier, J. M. Salord, C. Tancrede, M. Hayat, J. F. Bernaudin, and J. L. Pico. 1997. Should cytomegalovirus be tested for in both blood and bronchoalveolar lavage fluid of patients at a high risk of CMV pneumonia after bone marrow transplantation? Br J. Haematol. 98:222–227. Ljungman, P., K. Lore, J. Aschan, S. Klaesson, I. Lewensohn-Fuchs, B. Lonnqvist, O. Ringden, J. Winiarski, and A. Ehrnst. 1996. Use of a semiquantitative PCR for cytomegalovirus DNA as a basis for pre-emptive antiviral therapy in allogeneic bone marrow transplant patients. Bone Marrow Transplant. 17:583–587. Machida, U., M. Kami, T. Fukui, Y. Kazuyama, M. Kinoshita, Y. Tanaka, Y. Kanda, S. Ogawa, H. Honda, S. Chiba, K. Mitani, Y. Muto, K. Osumi, S. Kimura, and H. Hirai. 2000. Real-time automated PCR for early diagnosis

14. 15. 16.

17.

18.

19. 20. 21.

22.

4369

and monitoring of cytomegalovirus infection after bone marrow transplantation. J. Clin. Microbiol. 38:2536–2542. Nazzari, C., A. Gaeta, M. Lazzarini, T. D. Castelli, and C. Mancini. 2000. Multiplex polymerase chain reaction for the evaluation of cytomegalovirus DNA load in organ transplant recipients. J. Med. Virol. 61:251–258. Perol, Y., V. Caro, and M. C. Mazeron. 1993. Cytomegalovirus antigenemia assay: therapeutic usefulness and biological significance. Nouv. Rev. Fr. Hematol. 35:95–98. Reusser, P., J. G. Gambertoglio, K. Lilleby, and J. D. Meyers. 1992. Phase I-II trial of foscarnet for prevention of cytomegalovirus infection in autologous and allogeneic marrow transplant recipients. J. Infect. Dis. 166:473– 479. Schmidt, G. M., D. A. Horak, J. C. Niland, S. R. Duncan, S. J. Forman, and J. A. Zaia. 1991. A randomized, controlled trial of prophylactic ganciclovir for cytomegalovirus pulmonary infection in recipients of allogeneic bone marrow transplants. The City of Hope-Stanford-Syntex CMV Study Group. N. Engl. J. Med. 324:1005–1011. Seropian, S., D. Ferguson, E. Salloum, D. Cooper, and M. L. Landry. 1998. Lack of reactivity to CMV pp65 antigenemia testing in a patient with CMV disease following allogeneic bone marrow transplant. Bone Marrow Transplant. 22:507–509. Tanaka, N., H. Kimura, K. Iida, Y. Saito, I. Tsuge, A. Yoshimi, T. Matsuyama, and T. Morishima. 2000. Quantitative analysis of cytomegalovirus load using a real-time PCR assay. J. Med Virol. 60:455–462. The, T. H., W. Van der Bij, A. P. Van den Berg, M. Van der Giessen, J. Weits, H. G. Sprenger, and W. J. Van Son. 1990. Cytomegalovirus antigenemia. Rev. Infect. Dis. 12(Suppl. 7):S734–S744. Weber, B., U. Nestler, W. Ernst, H. Rabenau, J. Braner, A. Birkenbach, E. H. Scheuermann, W. Schoeppe, and H. W. Doerr. 1994. Low correlation of human cytomegalovirus DNA amplification by polymerase chain reaction with cytomegalovirus disease in organ transplant recipients. J. Med. Virol. 43:187–193. Yun, Z., I. Lewensohn-Fuchs, P. Ljungman, and A. Vahlne. 2000. Real-time monitoring of cytomegalovirus infections after stem cell transplantation using the TaqMan polymerase chain reaction assays. Transplantation 69:1733– 1736.