- Email: [email protected]
Combination of native and recombinant cytomegalovirus antigens in a new ELISA for detection of CMV-specific antibodies
Journal of Clinical Virology 43 (2008) 137–141
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Combination of native and recombinant cytomegalovirus antigens in a new ELISA for detection of CMV-specific antibodies Clemens Busse a , Andreas Strubel b , Paul Schnitzler c,∗ a
Applied Tumor Virology, German Cancer Research Center, University of Heidelberg, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany Genzyme Virotech, Löwenplatz 5, 65428 Rüsselsheim, Germany c Hygiene Institute, Department of Virology, University of Heidelberg, Im Neuenheimer Feld 324, 69120 Heidelberg, Germany b
a r t i c l e
i n f o
Article history: Received 18 October 2007 Received in revised form 20 March 2008 Accepted 1 May 2008 Keywords: Cytomegalovirus CMV ELISA Native antigen Recombinant antigen Antibody
a b s t r a c t Background: Serological detection of cytomegalovirus (CMV)-specific antibodies varies greatly due to antigen composition and the lack of antigen standardization. Objectives: To develop and evaluate a new ELISA with native and/or recombinant cytomegalovirus antigens for the detection of anti-CMV IgG and IgM antibodies. Results: The diagnostic performance of three anti-CMV ELISAs coated with different CMV antigen preparations, (i) native CMV antigen, (ii) a mixture of recombinant CMV peptides pp150, pp28, gB2 and pp52 and (iii) a combination of native CMV antigens and recombinant CMV IE1 antigen applied in the new Genzyme Virotech CMV ELISA, were compared. All tested sera were derived from patients or healthy blood donors and were predefined with the Dade Behring Enzygnost® CMV ELISA as well as by CMV PCR analysis. Additionally, official well-characterized serum panels were also tested. The new Genzyme Virotech CMV ELISA IgG/IgM test applying a combination of native antigens and recombinant IE1 antigen was evaluated and the performance was compared to the Dade Behring Enzygnost® CMV ELISA. The sensitivities were 98.9% (IgG) and 98.2% (IgM), the specificities were 98.8% (IgG) and 98.9% (IgM) for the Genzyme Virotech CMV ELISA. Furthermore all sera of the BBI mixed titer performance panel as well as the BBI seroconversion panel were identified 100% correctly with the new Genzyme Virotech ELISA. Conclusions: These data suggest that the new Genzyme Virotech CMV ELISA has higher sensitivity and specificity than ELISAs based on native antigens or recombinant peptides only. Specific combinations of native and recombinant antigens increase the serological detection of CMV infections and may add to further standardization of CMV serology. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Cytomegalovirus (CMV) is a widely distributed herpesvirus of humans most commonly transmitted via saliva, urine, or breast milk. It can also be transmitted sexually, in donated blood and organ transplants. Infection of immuncompetent hosts with CMV rarely causes clinical symptoms, whereas in patients with suppressed cellular immune functions or infected in utero, CMV may cause a variety of clinical symptoms (Gozlan et al., 1996; Daiminger et al., 1999; Ho, 1991; Ison and Fishman, 2005; Steininger, 2007). Serology is an economical method to diagnose CMV and can be automated for routine application. Detection of CMV-specific IgM antibodies is a sensitive and specific indicator of primary infection in immunocompetent subjects and is often produced during viral
∗ Corresponding author. Tel.: +49 6221 56 50 16; fax: +49 6221 56 50 03. E-mail address: Paul [email protected] (P. Schnitzler). 1386-6532/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2008.05.011
reactivation in transplant recipients (Landini, 1993; Landini et al., 1995). Antigenic materials composed of single well-characterized viral proteins or fragments of those expressed in procaryotic or eucaryotic systems have proven to be promising tools for improving CMV serology (Landini, 1993; Ohln et al., 1995; Plachter et al., 1992; Lazzarotto et al., 1996). It has been suggested that viral lysates used in diagnostic assays should be replaced by defined antigen preparation, preferably consisting of a combination of CMV specific and immunodominant antigens, in order to achieve the highest sensitivity and specificity. The aim of the present study was the development and largescale evaluation of a new ELISA for detection of CMV-specific IgG and IgM antibodies using different antigen preparations. Native CMV antigen (i), a mixture of defined recombinant CMV peptides (ii), and a combination of native CMV antigen and recombinant CMV IE1 antigen, which are applied in the new Genzyme Virotech CMV ELISA (iii), were compared. The combination of native antigen and recombinant antigen IE1 in the Genzyme Virotech CMV ELISA test
138
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141
revealed the highest sensitivity and specificity and will improve serological CMV diagnosis. 2. Materials and methods 2.1. Antigens Microtiter plates were coated with three different CMV antigen preparations, (i) native CMV antigen (Microbix, Toronto, Canada), (ii) a mixture of recombinant peptides pp150, pp28, gB2 and pp52 (recMix; University Leiden, Netherlands, Greijer et al., 1999), and (iii) native CMV antigen combined with recombinant IE1 antigen. Recombinant CMV IE1 antigen was kindly provided by Dr. Mach, University Erlangen, Germany. The concentrations of the antigens applied in the ELISAs were 35 g/ml for the native antigen, 1.3 g/ml of each recombinant peptide pp150, pp28, gB2 and pp52, and 0.5 g/ml for recombinant antigen IE1. Microtiter plates with different CMV antigens were used for detection of anti-CMV antibodies in predefined serum samples. 2.2. Sera 2.2.1. Patient sera The sera used in this study consisted of samples from graft recipients, from pregnant women, and from healthy blood donors. These sera were predefined with a Dade Behring Enzygnost® ELISA and all serum samples had been additionally characterized and defined by CMV PCR analysis and/or CMV-pp65 antigenemia testing. Negative samples were negative for CMV antibodies in the Dade Behring test, negative for CMV DNA in the PCR test and negative in the CMV pp65antigen test. In positive samples, CMV antibodies were detectable with the Dade Behring test, CMV DNA was positive by PCR and some of these samples were also positive for CMV pp65-antigen. 2.2.2. Sera panels In addition to sera from patients and blood donors, official wellcharacterized serum panels were also tested, e.g. the mixed titer performance panel (BBI Diagnostics), the Seroconversion Panel (BBI Diagnostics), and predefined sera from German proficiency panels (INSTAND, Düsseldorf, Germany). The two BBI panels consist of reference sera for detection of antibodies against CMV and are used by many manufacturers for evaluating new ELISAs. Sera from patients with a confirmed IgM-positive status for other viral infections, e.g. measles, chickenpox, that might cause cross-reactive results in a CMV antibody test were also used. The provisional cut-off values for IgG and IgM used in the beginning of the development were determined by adding three standard deviations to the mean value of CMV-negative sera. Applied to the BBI mixed titer performance panel these cut-off values proved to provide correct interpretations. In the later phase of the development the cut-off was slightly raised according to the results of the ROC curve analysis. 2.3. ELISAs Enzygnost CMV IgG and IgM ELISA (Dade Behring, Marburg, Germany) was applied according to the instructions of the manufacturer. In order to develop a new ELISA for the detection of anti-CMV antibodies, different CMV antigen preparations were applied, i.e. (i) native CMV antigen, (ii) mixture of recombinant CMV peptides or (iii) combination of native CMV antigen with recombinant IE1 antigen. The combination of native antigen and recombinant IE1 antigen is now commercially available in the newly developed Genzyme Virotech CMV ELISA. About 500 serum samples were tested in parallel with Dade Behring Enzygnost® and Genzyme Virotech CMV ELISA, and results were compared.
The ELISAs under evaluation were processed according to the following procedure. Ninety-six-well microtiter plates (Nunc, Roskilde, Denmark) were coated with 100 l antigen dilution per well, i.e. (i) native CMV antigen, (ii) mixture of recombinant CMV peptides or (iii) combination of native CMV antigen with recombinant IE1 antigen. These microtiter plates were kept overnight at room temperature in a humid chamber. The native antigen was diluted in a carbonate–bicarbonate buffer, pH 9.6, the recombinant antigens in PBS, respectively. The plates were washed four times with PBS and incubated with 200 l sheep serum in PBS per well for 1 h at room temperature, the blocking buffer was then discarded. A volume of 100 l human serum samples (1:101 in PBS with 0.001% Tween-20) were added to each well and allowed to react for 30 min at 37 ◦ C. After three washing steps, peroxidase labelled anti-human IgG or IgM antibody (in Tris with 10% fetal calf serum) was incubated for 30 min at 37 ◦ C. After repetition of the washing steps, 100 l of 3,3 ,5,5 -tetramethyl-benzidine substrate (KPL, Gaithersburg, MD, USA) were added and incubated for 30 min at 37 ◦ C in the dark. Reactions were stopped by the addition of 50 l 1 M citrate. The optical density was determined photometrically at 450 nm with 620 nm as reference filter. 3. Results 3.1. Comparison of native CMV antigen ELISA and recombinant CMV peptides ELISA In a first approach, the diagnostic performance of two different ELISAs with (i) native CMV antigens and (ii) recombinant CMV antigens were evaluated by testing predefined sera—165 sera for IgG and 164 sera for IgM. The native CMV antigen assay has a sensitivity of 100% and 94.3% for detection of IgG and IgM, respectively, and a specificity of 98.4% and 100%, respectively. In contrast, the recombinant CMV peptide mixture ELISA revealed lower sensitivities of 82.5% and 62.6% for detection of IgG and IgM, respectively, with specificities of 86.7% and 100%, respectively. The receiver operating characteristics (ROC) curves are shown in Fig. 1A for IgG and Fig. 1B for IgM. The areas under the ROC curves and their 95% confidence intervals were calculated to compare the results (Fig. 1C). The area under the curve for IgG reaches 0.9931 for the native antigen ELISA, and 0.9248 for the recombinant peptide mixture ELISA. The area under the curve for IgM reaches a value of 0.9409 for the native antigen ELISA, but only 0.8798 for the recombinant antigen mixture ELISA. The native CMV antigen ELISA seems to be superior for the detection of IgG and IgM to the recombinant peptide ELISA. 3.2. Combination of native and recombinant IE1 CMV antigens (Genzyme Virotech CMV ELISA) On the basis of the previous results a new ELISA was developed. To improve the sensitivity, in particular for the detection of anti-CMV IgM, the native CMV antigen was employed in combination with recombinant IE1 antigen. It had been shown that IgG directed against IE1 can be detected early after infection with CMV (Schoppel et al., 1997). Recombinant CMV IE1 proved to be a sensitive supplement (data not shown) and was added as an additional antigen component to the ELISA, which is commercially available from Genzyme Virotech. 3.3. Comparison of Dade Behring and Genzyme Virotech CMV ELISA Using predefined positive and negative CMV sera, a large-scale evaluation was performed with Genzyme Virotech CMV ELISA. All antibody-positive and negative samples had been predefined with
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141
139
Fig. 1. Receiver operating characteristics (ROC) curves for native CMV antigen ELISA and the mixture of recombinant peptides. Predefined sera were tested for the presence of anti-CMV-IgG (A) and IgM (B). The area under the ROC curve for each antigen preparation is shown in (C) for the detection of anti-CMV IgG and IgM antibodies, the 95% confidence interval is indicated by black bars. Table 1 Comparison between IgG and IgM detection by the new Genzyme Virotech CMV ELISA and the Dade Behring Enzygnost® Anti-CMV ELISA for the sera of the evaluation process Virotech ELISA
IgG (n = 356), Dade Behring Enzygnost®
Positive Equivocal Negative
266 2 3
Positive
Sensitivity (%) Specificity (%)
Equivocal 0 0 1
IgM (n = 513), Dade Behring Enzygnost® Negative 1 0 83
98.9 98.8
the Dade Behring CMV Enzygnost ELISA, by CMV PCR and CMV pp65-antigen assay. The Genzyme Virotech IgG assay achieves a sensitivity of 98.9% and a specificity of 98.8%, the IgM assay achieves 98.2% and 98.9%, respectively (Table 1).
Positive 224 2 4
Equivocal 2 2 5
Negative 3 2 269
98.2 98.9
Ehrlich Institute (Langen, Germany) and is shown in Fig. 3. The detection limit of the Genzyme Virotech CMV ELISA is comparable to that of the Dade Behring CMV Enzygnost ELISA. 3.5. Reproducibility of Genzyme Virotech CMV ELISA
3.4. Evaluation of Genzyme Virotech CMV ELISA with different characterized serum panels Samples from the German proficiency panel and two BBI panels were tested in the new Genzyme Virotech CMV ELISA and yielded a correct classification rate of 100% for detection of IgG as well as IgM for these panels. Altogether, 41 IgG positive, 12 IgG negative, 14 IgM positive and 39 IgM negative samples were tested. The results of the BBI seroconversion panel consisting of nine sera in the newly developed ELISA are shown in Fig. 2. A sample to cut-off ratio over 1.1 is considered a positive result. Anti-CMV IgM antibodies could be detected in the BBI seroconversion panel as early as 17 days after CMV infection. Five days later, IgG antibodies were detected. The IgM titer declined within 1 month after infection, whereas the amount of IgG antibodies was still increasing 2 months after infection. These results correspond exactly to the specifications of the BBI panel. Sera from patients with acute infections with other viral pathogens that might induce cross-reactive antibodies were further tested to challenge the specificity of the Genzyme Virotech ELISA. These sera had been predefined with Dade Behring Enzygnost® assays. When tested with predefined 22 sera positive for anti-EBV IgM, 22 sera positive for anti-VZV IgM, 20 sera positive for antiparvovirus B19 IgM, 13 sera positive for anti-measles virus IgM and 20 sera positive for anti-HIV the ELISA achieved a specificity of 87.6%. The detection limit of the Genzyme Virotech anti-CMV ELISA was analysed with a commercially available standard of the Paul
Reproducibility of test results was compared using a serum panel consisting of negative and positive samples by several individual assays performed in different laboratories by different technicians. Inter-assay variability was determined by testing 10
Fig. 2. Performance of the new Genzyme Virotech CMV ELISA with sera of the BBI seroconversion panel. A sample to cut-off ratio over 1.1 is considered as positive result.
140
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141
Fig. 3. Determination of the detection limit of the Genzyme Virotech ELISA with a commercially available standard for anti-CMV IgG (Paul-Ehrlich-Institute, Germany). OD, optical density; AU, arbitrary units.
sera for IgG as well as for IgM in 10 individual experiments. The inter-assay coefficients of variation ranged from 4.3% to 8.2% for the detection of IgG and from 4.9% to 15.3% for IgM, respectively (Table 2). Although test results varied slightly between different assays, no influence on the correct interpretation could be observed. The calculated values for the coefficient of variation are similar to those of others commercially available test kits. 4. Discussion Many recombinant antigens have been produced and characterized for their ability to bind CMV specific antibodies (Vornhagen et al., 1994; Van Zanten et al., 1993; Sipewa et al., 2002). However, the exact composition of the antigenic mixture representative of the entire complex of CMV antigens to be used for detecting CMV antibodies is still a matter of study. When only native antigens were used in the ELISA, a high sensitivity and specificity for the detection of IgG and IgM antibodies was demonstrated. In contrast, the Table 2 Reproducibility and inter-assay variability of the Genzyme Virotech CMV ELISA was determined by testing 10 sera for IgG as well as for IgM in 10 individual experiments in different laboratories Mean OD
Standard deviation
IgG
0.364 0.382 0.979 1.029 1.055 1.179 1.207 1.552 1.611 1.707
0.020 0.031 0.057 0.048 0.050 0.062 0.061 0.084 0.088 0.074
5.6 8.2 5.8 4.7 4.8 5.2 5.0 5.4 5.5 4.3
IgM
0.048 0.062 0.144 0.244 0.285 0.352 0.408 0.559 0.883 1.263
0.005 0.009 0.008 0.020 0.015 0.017 0.023 0.029 0.044 0.080
11.1 15.3 5.8 8.3 5.2 4.9 5.5 5.2 5.0 6.3
recombinant peptide mixture ELISA displayed lower sensitivity and specificity, which might be due to expression of peptides in procaryotic systems lacking glycosylation and incorrect protein folding (Scholl et al., 1988). The native antigen ELISA seems to be superior in sensitivity and specificity in our assays when compared to the recombinant peptide ELISA. The specificity of the native antigen CMV ELISA was improved by adding recombinant IE1 protein to the solid phase in the new Genzyme Virotech CMV ELISA and a large-scale screening of different serum samples was performed with this newly developed ELISA. Sensitivity and specificity reached nearly 99% for IgG. For the detection of IgM antibodies, an increased sensitivity of 98.2% and specificity of 98.9% were determined. Other commercially available assays reveal sensitivities and specificities of 96% and higher. Medac states a sensitivity of 98.8% for the detection of IgG and 99.4% for IgM and specificities of 98.2% (IgG) and 99.7% (IgM), respectively. Dade Behring declares sensitivities of 99.3% (IgG) and 95.0% (IgM) and specificities of 99.2% (IgG) and 100% (IgM). The comparison of these values with the data produced with the Genzyme Virotech CMV ELISA reveals the high diagnostic performance of the newly developed ELISA. Detection of IgM antibodies plays an important role for diagnosis of primary infections, especially during pregnancy, where positive results are usually further analysed by avidity assays (Eggers et al., 1998, 2000; Dangel et al., 2006). When the BBI seroconversion panel was applied to the new Genzyme Virotech CMV ELISA, anti-CMV IgM antibodies could be detected 17 days after infection. Five days later, IgG antibodies were detected. The IgM titer declines within 1 month after infection, whereas the amount of IgG antibodies still increases 2 months after infection. These results correspond exactly to the specifications of the BBI panel, which were defined by several commercially available tests. In conclusion, three different ELISAs for detection of CMVspecific IgG and IgM were developed and compared. Microtiter plates were coated with (i) native CMV antigen, (ii) recombinant peptide mixture, and (iii) a combination of native CMV antigens and recombinant IE1. The combination of native CMV antigens and IE1 protein is used in the new Genzyme Virotech CMV ELISA. This test revealed the highest sensitivity and specificity for detection of anti-CMV antibodies and a very good interlaboratory performance, which will improve serological CMV diagnostics. Acknowledgement We would like to thank Dr. M. Mach, University of Erlangen, Germany, for providing recombinant antigen IE1.
CV (%)
References Daiminger A, Bäder U, Eggers M, Lazzarotto T, Enders G. Evaluation of two novel enzyme immunoassays using recombinant antigens to detect cytomegalovirusspecific immunoglobulin M in sera from pregnant women. J Clin Virol 1999;13:161–71. Dangel V, Bäder U, Enders G. Improvement of cytomegalovirus avidity testing by adjusting the concentration of CMV-specific IgG in test samples. J Clin Virol 2006;35:300–9. Eggers M, Bäder U, Enders G. Combination of microneutralization and avidity assays: improved diagnosis of recent primary human cytomegalovirus infection in single serum sample of second trimester pregnancy. J Med Virol 2000;60:324–30. Eggers M, Metzger C, Enders G. Differentiation between acute primary and recurrent human cytomegalovirus infection in pregnancy, using a microneutralisation assay. J Med Virol 1998;56:352–8. Gozlan J, Laporte JP, Lesage S, Labopin M, Majman A, Gorin NC, et al. Monitoring of cytomegalovirus infection and disease in bone marrow recipients by reverse transcription PCR and comparison with PCR and blood and urine cultures. J Clin Microbiol 1996;34:2085–8. Greijer AE, van de Crommert JMG, Stevens SJC, Middeldorp JM. Molecular finespecificity analysis of antibody responses to human cytomegalovirus and design of novel synthetic-peptide-based serodiagnostic assays. J Clin Microbiol 1999;37:179–88.
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141 Ho M. Cytomeglovirus biology and infection. 2nd ed. New York: Plenum Medical Press; 1991. Ison MG, Fishman JA. Cytomegalovirus pneumonia in transplant recipients. Clin Chest Med 2005;26:691–705. Landini MP. New approaches and perspectives in cytomegalovirus diagnosis. Prog Med Virol 1993;40:157–77. Landini MP, Lazzarotto T, Maine GT, Ripalti A, Flanders R. Recombinant mono- and polyantigens to detect cytomegalovirus-specific immunoglobulin M in human sera by enzyme immunoassay. J Clin Microbiol 1995;33:2535–42. Lazzarotto T, Maine GT, Dal Monte P, Frush H, Shi K, Landini MP. Detection of serum immunoglobulin M to human cytomegalovirus by Western blotting correlates better with virological data than detection by conventional enzyme immunoassay. Clin Diagn Lab Immunol 1996;3:597–600. Ohln M, Plachter B, Sandqvist VA, Steenhakkers PG, Middeldorn JM, Borreback CA. Human antibody reactivity against the lower matrix protein (pp65) produced by cytomegalovirus. Clin Diagn Lab Immunol 1995;2:325–9. Plachter B, Wiecdzorek L, Scholl C, Ziegelmaier R, Jahn G. Detection of cytomegalovirus antibodies by an enzyme-linked immunosorbent assay using recombinant polypeptides of the large phosphorylated tegument protein pp150. J Clin Microbiol 1992;30:206–91.
141
Scholl BC, Von Hintzestein B, Borisch B, Traupe B, Broker M, Jahn G. Procaryotic expression of immunogenic polypeptides of the large phosphoprotein (pp150) of human cytomegalovirus. J Gen Virol 1988;69:1195–204. Schoppel K, Kropff B, Schmidt C, Vornhagen R, Mach M. The humoral immune response against human cytomegalovirus is characterized by a delayed synthesis of glycoprotein specific antibodies. J Infect Dis 1997;187:633–44. Sipewa MJ, Goubau P, Bodéus M. Evaluation of a cytomegalovirus glycoprotein B recombinant enzyme immunoassay to discriminate between a recent and a past infection. J Clin Microbiol 2002;40:3689–93. Steininger C. Clinical relevance of cytomegalovirus infection in patients with disorders of the immune system. Clin Microbiol Infect 2007;13: 953–63. Van Zanten J, Van der Giessen M, Van Son WJ, The TH. Antibody responses to human cytomegalovirus-specific polypeptides studied by immunoblotting in relation to viral load during cytomegalovirus infection. J Med Virol 1993;39:80–7. Vornhagen R, Plachter B, Hinderer W, The TH, Van Zaten J, Matter L, et al. Early serodiagnosis of acute human cytomegalovirus infection by enzymelinked immunosorbent assay using recombinant antigens. J Clin Microbiol 1994;32:981–6.
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Combination of native and recombinant cytomegalovirus antigens in a new ELISA for detection of CMV-specific antibodies Clemens Busse a , Andreas Strubel b , Paul Schnitzler c,∗ a
Applied Tumor Virology, German Cancer Research Center, University of Heidelberg, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany Genzyme Virotech, Löwenplatz 5, 65428 Rüsselsheim, Germany c Hygiene Institute, Department of Virology, University of Heidelberg, Im Neuenheimer Feld 324, 69120 Heidelberg, Germany b
a r t i c l e
i n f o
Article history: Received 18 October 2007 Received in revised form 20 March 2008 Accepted 1 May 2008 Keywords: Cytomegalovirus CMV ELISA Native antigen Recombinant antigen Antibody
a b s t r a c t Background: Serological detection of cytomegalovirus (CMV)-specific antibodies varies greatly due to antigen composition and the lack of antigen standardization. Objectives: To develop and evaluate a new ELISA with native and/or recombinant cytomegalovirus antigens for the detection of anti-CMV IgG and IgM antibodies. Results: The diagnostic performance of three anti-CMV ELISAs coated with different CMV antigen preparations, (i) native CMV antigen, (ii) a mixture of recombinant CMV peptides pp150, pp28, gB2 and pp52 and (iii) a combination of native CMV antigens and recombinant CMV IE1 antigen applied in the new Genzyme Virotech CMV ELISA, were compared. All tested sera were derived from patients or healthy blood donors and were predefined with the Dade Behring Enzygnost® CMV ELISA as well as by CMV PCR analysis. Additionally, official well-characterized serum panels were also tested. The new Genzyme Virotech CMV ELISA IgG/IgM test applying a combination of native antigens and recombinant IE1 antigen was evaluated and the performance was compared to the Dade Behring Enzygnost® CMV ELISA. The sensitivities were 98.9% (IgG) and 98.2% (IgM), the specificities were 98.8% (IgG) and 98.9% (IgM) for the Genzyme Virotech CMV ELISA. Furthermore all sera of the BBI mixed titer performance panel as well as the BBI seroconversion panel were identified 100% correctly with the new Genzyme Virotech ELISA. Conclusions: These data suggest that the new Genzyme Virotech CMV ELISA has higher sensitivity and specificity than ELISAs based on native antigens or recombinant peptides only. Specific combinations of native and recombinant antigens increase the serological detection of CMV infections and may add to further standardization of CMV serology. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Cytomegalovirus (CMV) is a widely distributed herpesvirus of humans most commonly transmitted via saliva, urine, or breast milk. It can also be transmitted sexually, in donated blood and organ transplants. Infection of immuncompetent hosts with CMV rarely causes clinical symptoms, whereas in patients with suppressed cellular immune functions or infected in utero, CMV may cause a variety of clinical symptoms (Gozlan et al., 1996; Daiminger et al., 1999; Ho, 1991; Ison and Fishman, 2005; Steininger, 2007). Serology is an economical method to diagnose CMV and can be automated for routine application. Detection of CMV-specific IgM antibodies is a sensitive and specific indicator of primary infection in immunocompetent subjects and is often produced during viral
∗ Corresponding author. Tel.: +49 6221 56 50 16; fax: +49 6221 56 50 03. E-mail address: Paul [email protected] (P. Schnitzler). 1386-6532/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2008.05.011
reactivation in transplant recipients (Landini, 1993; Landini et al., 1995). Antigenic materials composed of single well-characterized viral proteins or fragments of those expressed in procaryotic or eucaryotic systems have proven to be promising tools for improving CMV serology (Landini, 1993; Ohln et al., 1995; Plachter et al., 1992; Lazzarotto et al., 1996). It has been suggested that viral lysates used in diagnostic assays should be replaced by defined antigen preparation, preferably consisting of a combination of CMV specific and immunodominant antigens, in order to achieve the highest sensitivity and specificity. The aim of the present study was the development and largescale evaluation of a new ELISA for detection of CMV-specific IgG and IgM antibodies using different antigen preparations. Native CMV antigen (i), a mixture of defined recombinant CMV peptides (ii), and a combination of native CMV antigen and recombinant CMV IE1 antigen, which are applied in the new Genzyme Virotech CMV ELISA (iii), were compared. The combination of native antigen and recombinant antigen IE1 in the Genzyme Virotech CMV ELISA test
138
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141
revealed the highest sensitivity and specificity and will improve serological CMV diagnosis. 2. Materials and methods 2.1. Antigens Microtiter plates were coated with three different CMV antigen preparations, (i) native CMV antigen (Microbix, Toronto, Canada), (ii) a mixture of recombinant peptides pp150, pp28, gB2 and pp52 (recMix; University Leiden, Netherlands, Greijer et al., 1999), and (iii) native CMV antigen combined with recombinant IE1 antigen. Recombinant CMV IE1 antigen was kindly provided by Dr. Mach, University Erlangen, Germany. The concentrations of the antigens applied in the ELISAs were 35 g/ml for the native antigen, 1.3 g/ml of each recombinant peptide pp150, pp28, gB2 and pp52, and 0.5 g/ml for recombinant antigen IE1. Microtiter plates with different CMV antigens were used for detection of anti-CMV antibodies in predefined serum samples. 2.2. Sera 2.2.1. Patient sera The sera used in this study consisted of samples from graft recipients, from pregnant women, and from healthy blood donors. These sera were predefined with a Dade Behring Enzygnost® ELISA and all serum samples had been additionally characterized and defined by CMV PCR analysis and/or CMV-pp65 antigenemia testing. Negative samples were negative for CMV antibodies in the Dade Behring test, negative for CMV DNA in the PCR test and negative in the CMV pp65antigen test. In positive samples, CMV antibodies were detectable with the Dade Behring test, CMV DNA was positive by PCR and some of these samples were also positive for CMV pp65-antigen. 2.2.2. Sera panels In addition to sera from patients and blood donors, official wellcharacterized serum panels were also tested, e.g. the mixed titer performance panel (BBI Diagnostics), the Seroconversion Panel (BBI Diagnostics), and predefined sera from German proficiency panels (INSTAND, Düsseldorf, Germany). The two BBI panels consist of reference sera for detection of antibodies against CMV and are used by many manufacturers for evaluating new ELISAs. Sera from patients with a confirmed IgM-positive status for other viral infections, e.g. measles, chickenpox, that might cause cross-reactive results in a CMV antibody test were also used. The provisional cut-off values for IgG and IgM used in the beginning of the development were determined by adding three standard deviations to the mean value of CMV-negative sera. Applied to the BBI mixed titer performance panel these cut-off values proved to provide correct interpretations. In the later phase of the development the cut-off was slightly raised according to the results of the ROC curve analysis. 2.3. ELISAs Enzygnost CMV IgG and IgM ELISA (Dade Behring, Marburg, Germany) was applied according to the instructions of the manufacturer. In order to develop a new ELISA for the detection of anti-CMV antibodies, different CMV antigen preparations were applied, i.e. (i) native CMV antigen, (ii) mixture of recombinant CMV peptides or (iii) combination of native CMV antigen with recombinant IE1 antigen. The combination of native antigen and recombinant IE1 antigen is now commercially available in the newly developed Genzyme Virotech CMV ELISA. About 500 serum samples were tested in parallel with Dade Behring Enzygnost® and Genzyme Virotech CMV ELISA, and results were compared.
The ELISAs under evaluation were processed according to the following procedure. Ninety-six-well microtiter plates (Nunc, Roskilde, Denmark) were coated with 100 l antigen dilution per well, i.e. (i) native CMV antigen, (ii) mixture of recombinant CMV peptides or (iii) combination of native CMV antigen with recombinant IE1 antigen. These microtiter plates were kept overnight at room temperature in a humid chamber. The native antigen was diluted in a carbonate–bicarbonate buffer, pH 9.6, the recombinant antigens in PBS, respectively. The plates were washed four times with PBS and incubated with 200 l sheep serum in PBS per well for 1 h at room temperature, the blocking buffer was then discarded. A volume of 100 l human serum samples (1:101 in PBS with 0.001% Tween-20) were added to each well and allowed to react for 30 min at 37 ◦ C. After three washing steps, peroxidase labelled anti-human IgG or IgM antibody (in Tris with 10% fetal calf serum) was incubated for 30 min at 37 ◦ C. After repetition of the washing steps, 100 l of 3,3 ,5,5 -tetramethyl-benzidine substrate (KPL, Gaithersburg, MD, USA) were added and incubated for 30 min at 37 ◦ C in the dark. Reactions were stopped by the addition of 50 l 1 M citrate. The optical density was determined photometrically at 450 nm with 620 nm as reference filter. 3. Results 3.1. Comparison of native CMV antigen ELISA and recombinant CMV peptides ELISA In a first approach, the diagnostic performance of two different ELISAs with (i) native CMV antigens and (ii) recombinant CMV antigens were evaluated by testing predefined sera—165 sera for IgG and 164 sera for IgM. The native CMV antigen assay has a sensitivity of 100% and 94.3% for detection of IgG and IgM, respectively, and a specificity of 98.4% and 100%, respectively. In contrast, the recombinant CMV peptide mixture ELISA revealed lower sensitivities of 82.5% and 62.6% for detection of IgG and IgM, respectively, with specificities of 86.7% and 100%, respectively. The receiver operating characteristics (ROC) curves are shown in Fig. 1A for IgG and Fig. 1B for IgM. The areas under the ROC curves and their 95% confidence intervals were calculated to compare the results (Fig. 1C). The area under the curve for IgG reaches 0.9931 for the native antigen ELISA, and 0.9248 for the recombinant peptide mixture ELISA. The area under the curve for IgM reaches a value of 0.9409 for the native antigen ELISA, but only 0.8798 for the recombinant antigen mixture ELISA. The native CMV antigen ELISA seems to be superior for the detection of IgG and IgM to the recombinant peptide ELISA. 3.2. Combination of native and recombinant IE1 CMV antigens (Genzyme Virotech CMV ELISA) On the basis of the previous results a new ELISA was developed. To improve the sensitivity, in particular for the detection of anti-CMV IgM, the native CMV antigen was employed in combination with recombinant IE1 antigen. It had been shown that IgG directed against IE1 can be detected early after infection with CMV (Schoppel et al., 1997). Recombinant CMV IE1 proved to be a sensitive supplement (data not shown) and was added as an additional antigen component to the ELISA, which is commercially available from Genzyme Virotech. 3.3. Comparison of Dade Behring and Genzyme Virotech CMV ELISA Using predefined positive and negative CMV sera, a large-scale evaluation was performed with Genzyme Virotech CMV ELISA. All antibody-positive and negative samples had been predefined with
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141
139
Fig. 1. Receiver operating characteristics (ROC) curves for native CMV antigen ELISA and the mixture of recombinant peptides. Predefined sera were tested for the presence of anti-CMV-IgG (A) and IgM (B). The area under the ROC curve for each antigen preparation is shown in (C) for the detection of anti-CMV IgG and IgM antibodies, the 95% confidence interval is indicated by black bars. Table 1 Comparison between IgG and IgM detection by the new Genzyme Virotech CMV ELISA and the Dade Behring Enzygnost® Anti-CMV ELISA for the sera of the evaluation process Virotech ELISA
IgG (n = 356), Dade Behring Enzygnost®
Positive Equivocal Negative
266 2 3
Positive
Sensitivity (%) Specificity (%)
Equivocal 0 0 1
IgM (n = 513), Dade Behring Enzygnost® Negative 1 0 83
98.9 98.8
the Dade Behring CMV Enzygnost ELISA, by CMV PCR and CMV pp65-antigen assay. The Genzyme Virotech IgG assay achieves a sensitivity of 98.9% and a specificity of 98.8%, the IgM assay achieves 98.2% and 98.9%, respectively (Table 1).
Positive 224 2 4
Equivocal 2 2 5
Negative 3 2 269
98.2 98.9
Ehrlich Institute (Langen, Germany) and is shown in Fig. 3. The detection limit of the Genzyme Virotech CMV ELISA is comparable to that of the Dade Behring CMV Enzygnost ELISA. 3.5. Reproducibility of Genzyme Virotech CMV ELISA
3.4. Evaluation of Genzyme Virotech CMV ELISA with different characterized serum panels Samples from the German proficiency panel and two BBI panels were tested in the new Genzyme Virotech CMV ELISA and yielded a correct classification rate of 100% for detection of IgG as well as IgM for these panels. Altogether, 41 IgG positive, 12 IgG negative, 14 IgM positive and 39 IgM negative samples were tested. The results of the BBI seroconversion panel consisting of nine sera in the newly developed ELISA are shown in Fig. 2. A sample to cut-off ratio over 1.1 is considered a positive result. Anti-CMV IgM antibodies could be detected in the BBI seroconversion panel as early as 17 days after CMV infection. Five days later, IgG antibodies were detected. The IgM titer declined within 1 month after infection, whereas the amount of IgG antibodies was still increasing 2 months after infection. These results correspond exactly to the specifications of the BBI panel. Sera from patients with acute infections with other viral pathogens that might induce cross-reactive antibodies were further tested to challenge the specificity of the Genzyme Virotech ELISA. These sera had been predefined with Dade Behring Enzygnost® assays. When tested with predefined 22 sera positive for anti-EBV IgM, 22 sera positive for anti-VZV IgM, 20 sera positive for antiparvovirus B19 IgM, 13 sera positive for anti-measles virus IgM and 20 sera positive for anti-HIV the ELISA achieved a specificity of 87.6%. The detection limit of the Genzyme Virotech anti-CMV ELISA was analysed with a commercially available standard of the Paul
Reproducibility of test results was compared using a serum panel consisting of negative and positive samples by several individual assays performed in different laboratories by different technicians. Inter-assay variability was determined by testing 10
Fig. 2. Performance of the new Genzyme Virotech CMV ELISA with sera of the BBI seroconversion panel. A sample to cut-off ratio over 1.1 is considered as positive result.
140
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141
Fig. 3. Determination of the detection limit of the Genzyme Virotech ELISA with a commercially available standard for anti-CMV IgG (Paul-Ehrlich-Institute, Germany). OD, optical density; AU, arbitrary units.
sera for IgG as well as for IgM in 10 individual experiments. The inter-assay coefficients of variation ranged from 4.3% to 8.2% for the detection of IgG and from 4.9% to 15.3% for IgM, respectively (Table 2). Although test results varied slightly between different assays, no influence on the correct interpretation could be observed. The calculated values for the coefficient of variation are similar to those of others commercially available test kits. 4. Discussion Many recombinant antigens have been produced and characterized for their ability to bind CMV specific antibodies (Vornhagen et al., 1994; Van Zanten et al., 1993; Sipewa et al., 2002). However, the exact composition of the antigenic mixture representative of the entire complex of CMV antigens to be used for detecting CMV antibodies is still a matter of study. When only native antigens were used in the ELISA, a high sensitivity and specificity for the detection of IgG and IgM antibodies was demonstrated. In contrast, the Table 2 Reproducibility and inter-assay variability of the Genzyme Virotech CMV ELISA was determined by testing 10 sera for IgG as well as for IgM in 10 individual experiments in different laboratories Mean OD
Standard deviation
IgG
0.364 0.382 0.979 1.029 1.055 1.179 1.207 1.552 1.611 1.707
0.020 0.031 0.057 0.048 0.050 0.062 0.061 0.084 0.088 0.074
5.6 8.2 5.8 4.7 4.8 5.2 5.0 5.4 5.5 4.3
IgM
0.048 0.062 0.144 0.244 0.285 0.352 0.408 0.559 0.883 1.263
0.005 0.009 0.008 0.020 0.015 0.017 0.023 0.029 0.044 0.080
11.1 15.3 5.8 8.3 5.2 4.9 5.5 5.2 5.0 6.3
recombinant peptide mixture ELISA displayed lower sensitivity and specificity, which might be due to expression of peptides in procaryotic systems lacking glycosylation and incorrect protein folding (Scholl et al., 1988). The native antigen ELISA seems to be superior in sensitivity and specificity in our assays when compared to the recombinant peptide ELISA. The specificity of the native antigen CMV ELISA was improved by adding recombinant IE1 protein to the solid phase in the new Genzyme Virotech CMV ELISA and a large-scale screening of different serum samples was performed with this newly developed ELISA. Sensitivity and specificity reached nearly 99% for IgG. For the detection of IgM antibodies, an increased sensitivity of 98.2% and specificity of 98.9% were determined. Other commercially available assays reveal sensitivities and specificities of 96% and higher. Medac states a sensitivity of 98.8% for the detection of IgG and 99.4% for IgM and specificities of 98.2% (IgG) and 99.7% (IgM), respectively. Dade Behring declares sensitivities of 99.3% (IgG) and 95.0% (IgM) and specificities of 99.2% (IgG) and 100% (IgM). The comparison of these values with the data produced with the Genzyme Virotech CMV ELISA reveals the high diagnostic performance of the newly developed ELISA. Detection of IgM antibodies plays an important role for diagnosis of primary infections, especially during pregnancy, where positive results are usually further analysed by avidity assays (Eggers et al., 1998, 2000; Dangel et al., 2006). When the BBI seroconversion panel was applied to the new Genzyme Virotech CMV ELISA, anti-CMV IgM antibodies could be detected 17 days after infection. Five days later, IgG antibodies were detected. The IgM titer declines within 1 month after infection, whereas the amount of IgG antibodies still increases 2 months after infection. These results correspond exactly to the specifications of the BBI panel, which were defined by several commercially available tests. In conclusion, three different ELISAs for detection of CMVspecific IgG and IgM were developed and compared. Microtiter plates were coated with (i) native CMV antigen, (ii) recombinant peptide mixture, and (iii) a combination of native CMV antigens and recombinant IE1. The combination of native CMV antigens and IE1 protein is used in the new Genzyme Virotech CMV ELISA. This test revealed the highest sensitivity and specificity for detection of anti-CMV antibodies and a very good interlaboratory performance, which will improve serological CMV diagnostics. Acknowledgement We would like to thank Dr. M. Mach, University of Erlangen, Germany, for providing recombinant antigen IE1.
CV (%)
References Daiminger A, Bäder U, Eggers M, Lazzarotto T, Enders G. Evaluation of two novel enzyme immunoassays using recombinant antigens to detect cytomegalovirusspecific immunoglobulin M in sera from pregnant women. J Clin Virol 1999;13:161–71. Dangel V, Bäder U, Enders G. Improvement of cytomegalovirus avidity testing by adjusting the concentration of CMV-specific IgG in test samples. J Clin Virol 2006;35:300–9. Eggers M, Bäder U, Enders G. Combination of microneutralization and avidity assays: improved diagnosis of recent primary human cytomegalovirus infection in single serum sample of second trimester pregnancy. J Med Virol 2000;60:324–30. Eggers M, Metzger C, Enders G. Differentiation between acute primary and recurrent human cytomegalovirus infection in pregnancy, using a microneutralisation assay. J Med Virol 1998;56:352–8. Gozlan J, Laporte JP, Lesage S, Labopin M, Majman A, Gorin NC, et al. Monitoring of cytomegalovirus infection and disease in bone marrow recipients by reverse transcription PCR and comparison with PCR and blood and urine cultures. J Clin Microbiol 1996;34:2085–8. Greijer AE, van de Crommert JMG, Stevens SJC, Middeldorp JM. Molecular finespecificity analysis of antibody responses to human cytomegalovirus and design of novel synthetic-peptide-based serodiagnostic assays. J Clin Microbiol 1999;37:179–88.
C. Busse et al. / Journal of Clinical Virology 43 (2008) 137–141 Ho M. Cytomeglovirus biology and infection. 2nd ed. New York: Plenum Medical Press; 1991. Ison MG, Fishman JA. Cytomegalovirus pneumonia in transplant recipients. Clin Chest Med 2005;26:691–705. Landini MP. New approaches and perspectives in cytomegalovirus diagnosis. Prog Med Virol 1993;40:157–77. Landini MP, Lazzarotto T, Maine GT, Ripalti A, Flanders R. Recombinant mono- and polyantigens to detect cytomegalovirus-specific immunoglobulin M in human sera by enzyme immunoassay. J Clin Microbiol 1995;33:2535–42. Lazzarotto T, Maine GT, Dal Monte P, Frush H, Shi K, Landini MP. Detection of serum immunoglobulin M to human cytomegalovirus by Western blotting correlates better with virological data than detection by conventional enzyme immunoassay. Clin Diagn Lab Immunol 1996;3:597–600. Ohln M, Plachter B, Sandqvist VA, Steenhakkers PG, Middeldorn JM, Borreback CA. Human antibody reactivity against the lower matrix protein (pp65) produced by cytomegalovirus. Clin Diagn Lab Immunol 1995;2:325–9. Plachter B, Wiecdzorek L, Scholl C, Ziegelmaier R, Jahn G. Detection of cytomegalovirus antibodies by an enzyme-linked immunosorbent assay using recombinant polypeptides of the large phosphorylated tegument protein pp150. J Clin Microbiol 1992;30:206–91.
141
Scholl BC, Von Hintzestein B, Borisch B, Traupe B, Broker M, Jahn G. Procaryotic expression of immunogenic polypeptides of the large phosphoprotein (pp150) of human cytomegalovirus. J Gen Virol 1988;69:1195–204. Schoppel K, Kropff B, Schmidt C, Vornhagen R, Mach M. The humoral immune response against human cytomegalovirus is characterized by a delayed synthesis of glycoprotein specific antibodies. J Infect Dis 1997;187:633–44. Sipewa MJ, Goubau P, Bodéus M. Evaluation of a cytomegalovirus glycoprotein B recombinant enzyme immunoassay to discriminate between a recent and a past infection. J Clin Microbiol 2002;40:3689–93. Steininger C. Clinical relevance of cytomegalovirus infection in patients with disorders of the immune system. Clin Microbiol Infect 2007;13: 953–63. Van Zanten J, Van der Giessen M, Van Son WJ, The TH. Antibody responses to human cytomegalovirus-specific polypeptides studied by immunoblotting in relation to viral load during cytomegalovirus infection. J Med Virol 1993;39:80–7. Vornhagen R, Plachter B, Hinderer W, The TH, Van Zaten J, Matter L, et al. Early serodiagnosis of acute human cytomegalovirus infection by enzymelinked immunosorbent assay using recombinant antigens. J Clin Microbiol 1994;32:981–6.