Journal of Clinical Laboratory Analysis 6:216-218 (1992)
Enzyme-Linked lmmunoadsorbent Assay for the Detection of Cytomegalovirus-lgM: Comparison Between Eight Commercial Kits, Immunofluorescence, and lmmunoblotting T. Lazzarotto, B. Dalla Casa, B. Campisi, and M.P. Landini Institute of Microbiology, Faculty of Medicine, University of Bologna, Bologna, Italy Eight commercially available enzymelinked immunoadsorbent assays (ELISA) for the detection of cytomegalovirus (CMV)specific IgM were used in parallel to determine the presence of CMV-IgM in 123 serum samples from pregnant women. The results obtained with the eight kits were compared. Based on concordance of six or more of the eight kits, we assessed sensitivity, specificity, and overall agreement, as well as incidence of false-positive and -negative results for each kit. The results obtained by ELSA were then compared with those obKey words:
tained by immunofluorescence (IF) and immunoblotting (IB). Our study did not single out one outstanding ELISA kit among the eight evaluated, nor did it suggest that IF or IB are better than ELISA. Furthermore our results indicate that IB might be useful in several cases as, beside its good sensitivity, most IB-false-positivesera are easily recognized as reacting exclusively with ppl50, the unique reactivity to ppl50 not being among the IB profiles of IB-true-positive sera. Nevertheless 14.6% of sera remained ss, CMV-lgM-indeterminate. o 1 9 9 2 ~ i 1 e y - ~ iInc.
commercial ELSA kits, CMV, IF, IB
MATERIALS AND METHODS
Cytomegalovirus (CMV) is known to be the most common agent causing congenital infection in humans. Intrauterine primary infections are second only to Down’s syndrome as a known cause of mental retardation (1). CMV may cause less severe complications as a result of secondary exogenous infection or endogenous virus reactivation (1). Mothers of infected infants seldom have a history of infectious disease during pregnancy: therefore, laboratory techniques ascertaining CMV infection are the sole means of diagnosing acute CMV infection in these subjects. Diagnosis of primary CMV infection is exclusively accomplished by serological methods, i.e., demonstration of the appearance of antibodies to CMV in previously seronegative subjects. CMV-specific IgM is a sensitive and specific indicator of primary CMV infection. In fact IgM are detected in 55-75% of sera from pregnant women undergoing primary infection and in approximately 70% of sera from congenitally infected babies (2-4). A number of serologic procedures have been adopted for the detection of IgM (2,5,6-8), although enzyme-linked immunadsorbent assay (ELISA) is the most practical. However, in our experience discordant results are often obtained when a serum sample is tested with different ELISA kits. The present study aimed to compare eight commercial ELISA kits, indirect immunofluorescence (IF), and immunoblotting (IB) for the detection of anti-CMV IgM and evaluate the percentage of discordant results.
0 1992 Wiley-Liss, Inc.
A total of 123 serum samples obtained from pregnant women were used in the present study. Serum samples were sent to our diagnostic center from other Italian laboratories during the third trimester of 1991 for confirmation of the presence of CMV-specific IgM. The samples were stored sterile at 4°C until tested. Sera were tested for the presence of rheumatoid factor (RF) by latex agglutination (Rheuma-Wellco test; Wellcome, Dartford, England) and only RF-negative sera were included in the present study.
ELSA The following kits commercially available during the second semester of 1991 were used: CMV-IgM ELA Test (Technogenetics, Hamburg, Germany: TECHNO); ELISA CMV IgM (Sclavo, Siena, Italy: SCLAVO); BEIA CMV-IgM (Bouty diagnostici, Milano, Italy: BOUTY); CMV IgM ELISA (Eurogenetics N.V., Tessenderlo, Belgium: EURO); CMV STATM (M.A. Bioproducts, Walkerville, MD, U.S.A.: M.A.);
Received February 1, 1992; accepted February 13, 1992 Address reprint requests to M.P. Landini, M.D., Institute of Microbiology, St. Orsola General Hospital, Via Massarenti 9, 40138 Bologna, Italy.
Detection of Anti-CMV IgM
Cytomegaly-virus Elisa IgM antibody test (Human Gesellschaft fur Biochemica und Diagnostica, Taunsstein, Germany: BCI); ETI-CYTOK-M reverse P3238 (Sorin Biomedica, Vercelli, Italy: SORIN); Enzygnost anti-Cytomegalovirus for IgM-antibody determinations (Behringwerke AG, Marburg, Germany: BEHRING). TECHNO, SCLAVO, BOUTY, EURO, and SORIN are IgM-capture ELISA, while M.A., BCI, and BEHRING are traditional indirect ELISA. The tests were performed and the results interpreted carefully following manufacturer's instructions.
lmmunoblotting For IB, protein extracts from infected cells were run in a 9% acrylamide gel and electrophoretically separated polypeptides were then transferred to nitrocellulose paper. Infection, protein extraction, blotting, and immune reaction with sera were done as previously described in detail (9,lO).
sensitivity, specificity, and overall agreement as well as falsepositive and -negative results rates for each kit determined on the basis of the agreement among at least 6 kits. Sensitivity ranges from 28.4 to 95.2%, specificity from 44.2 to loo%, and overall agreement from 44.7 to 65.8%. All the sera were then tested by IB and IF. The results obtained testing the 81 ELISA-determined sera are shown in Table 2 and were based on concordance of at least two tests. The results obtained by IB are more similar to those obtained by ELISA than those we had obtained by IF. In this regard it is interesting that 4 out of 5 IB false-positive sera were found to react exclusively with pp150. The other sera reacted with pp150 and faintly with pp65. The reactivity to pp150 alone was never found among the 36 IB true-positive sera (data not shown). The results obtained testing the 42 ELISA-indeterminate sera by IB and IF are shown in Table 3. Twenty-six percent of the sera were IB-IF positive, 26% were IB-IF negative, and 48% remained indeterminate.
Human embryo fibroblasts (Flow 2002) were grown in coverslips with Minimal Essential Medium and 10% foetal calf serum. Three days after seeding, cells were infected with 0.2 pfukell of the Towne strain of human CMV. Five days later infected cells with evident cytopathic effect were fixed with methano1:acetone (3: 1) and used for IF. Sera were used at a 1:20 dilution; anti-p, chain conjugate (Dakopatts, Glostrup, Denmark) was used at a 1:30 dilution.
RESULTS A total of 123 serum specimens were tested by eight different ELISA kits for the detection of CMV-IgM. Only 43 sera (34.96%) gave identical (negative or positive) results with all ELISA kits. Based on concordance of at least six of the eight kits 81 of the 123 sera (65.85%) gave similar results and 42 were ELISA indeterminate. Table 1 summarizes the
The present study aimed to evaluate the percentage of discordant results that can be obtained in CMV-IgM detection when different ELISA kits or IB and IF are used. The results obtained show a very poor agreement among the different ELISA kits (44.7-65.8%). This is probably due to the fact that the antigen used in the assays, which actually consists of a mixture of 20- 100 different antigenic proteins is very difficult to standardize. In fact its quantitative and qualitative composition depends on many variables in individual laboratories such as the type of cells used to grow the virus and their age, the multiplicity of viral infection, the day of harvest, the extraction procedure and, in the case of purified viral preparations, the procedure used for virus purification. Hopefully at least some of the problems encountered will be solved using single well-Characterized viral proteins or por-
TABLE 1. Comparison of the Results Obtained With the Eight Commercial ELISA Kits for the Detection of CMV-IgM" Total
IgM-capture kits TECHNO SCLAVO BOUTY EURO SORIN
80 27 29 69 69
43 96 94 54 54
40 27 29 40 40
41 28 30 41 41
40 0 0 29 29
2 68 64 13 13
95.2 28.4 31.2 75.5 75.5
44.2 100 100 58.5 58.6
65.8 44.7 48.0 65.8 65.8
Indirect IgM detection BEHRING M.A. BC1
41 80 62
82 43 61
40 40 40
41 41 41
I 40 22
41 2 20
49.3 95.2 66.6
97.6 50.6 65.1
65.8 65.8 65.8
"SNC, sensitivity;SPE, specificity. SNC = (true-positives)/(tre-positives + false-negatives) X 100. SPE = (true-negatives)/(true-negatives + false-positives) X 100. Overall agreement (true positives + true negatives)/(totalnumber tested) X 100.
Lazzarotto et al.
This work was partially supported by the Italian National Research Council, Ministry of Education (40% and 60%) and Ministry of Public Health (I. S.S . , AIDS Projects).
TABLE 2. Results Obtained Testing 81 ELISA-Determined Sera and Based on Concordanceof at Least Two Tests Total
40 41 34
41 40 47
39 36 34
41 37 42
tions of them, obtained by biotechnology or peptide chemistry. In this respect pp150 (the basic phosphoprotein coded by UL 32), p65 (coded by UL83), and nonstructural p52 (coded by UL44) are the three best IgM-binding antigens (9-14). Several fusion proteins containing portions of pl50 have been obtained and have been shown to react efficiently with human IgM (12- 14). These results are encouraging as well as those obtained with synthetic peptides (15,16) and we forecast that a new generation of ELISA kits for the detection of CMVIgM will soon be available. This is strengthened by the fact that the detection of CMV-IgG by ELISA using recombinant polypeptides of pp150 recently was carried out successfully (17). Our results also show that 34.1 % of the sera are ELISAIgM undetermined and 48% of them remain IgM-indeterminate after IB and IIF testing. This stresses the need for more reliable serological tests. IB can be of some help in several cases as, beside its good sensitivity, the great majority of IB-false-positives can be easily recognized because they show a reactivity exclusively to pp150, which is not comprised among those of true-positive sera. This finding can be explained by the presence of IgM reacting with other members of the Herpesviridae family. In fact, there are two proteins of MW 150 kD in the viral structure, the basic phosphoprotein and the major capsid protein (coded by UL86). While the first is known to be the strongest viral immunogen, the second has cross reacting regions with other herpes viruses (18). In conclusion CMV-IgM detection is still a serious problem which all should be aware of. The most urgent challenge is an improvement in the production and characterization of specific IgM-reacting antigens which should be easily standardizable. TABLE 3. Results Obtained Testing 42 ELISA-Undetermined Sera by IB and IF IB Positive
IF positive IF negative
ACKNOWLEDGMENTS The authors are grateful to Prof. La Placa for continuous support.
REFERENCES I . Ho M: Cytomegalovirus: Biology and Infection, Second Edition. In Current Topics in InfectiousDisease. WB Greenough 111, TC Merigan, eds. Raven Press, New York, 1991. 2. Griffiths PD, Stagno S , Pass RF, Smith RJ, Alford CH: Infection with Cytomegalovirus during pregnancy: specific IgM antibodies as a marker of recent primary infection. JInfect Dis 145:647-653, 1982. 3. Schmitz H, Kampa D, Doerr HW, Luthardt T, Hillemanns HG, Wurtele A: IgM antibodies to cytomegalovirus during pregnancy. Arch Virol 53: 177- 184, 1977. 4. Stagno S, Tinker MK, Elrod C, Fuccillo DA, Cloud G, O'Beime AJ: Immunoglobulin M antibodies detected by enzyme-linkedimmunosorbent assay and radioimmunoassay in the diagnosis of cytomegalovirus infections in pregnant women and newborn infants. J Clin Microbiol 21: 930-935, 1985. 5. Griffiths PD: The presumptive diagnosis of primary cytomegalovirus infection in early pregnancy by means of a radioimmunoassay for specific IgM-antibodies. Br J Obstet Gynaeco/88:582-587, 1981. 6. Knez V, Stewart JA, Ziegler DW: Cytomegalovirus specific IgM and IgG response in humans studied by radioimmunoassay. J Immunol 117:2006-2013, 1976. 7. Rasmussen L, Kelsall D, Nelson R, Carney W, Hirsch M, Winston D, Preiksatis J , Merigan TC: Virus specific IgG and IgM antibodies in normal and immunocompromisedsubjects with cytomegalovirus.J Infect D ~ 145~191-201, s 1982. 8. Re MC, Landini MP: IgM to human cytomegalovirus: Comparison of two enzyme immunoassays and IgM reactivity to viral polypeptides detected by immunoblotting. JClin Lab Anal 3: 169-173, 1989. 9. Landini MP, Mirolo G, Baldassarri B, La Placa M: Human immune response to cytomegalovirus structural polypeptides studied by immunoblotting. JMed Virol 17:303-311, 1985. 10. Landini MP, Mirolo G, Coppolecchia P, La Placa M: Serum antibodies to individual cytomegalovirus structural polypeptides in renal transplant recipients during viral infection. Microbiol Immunol30:683-694 1986. 11. Jahn G, Scholl VC, Traupe B, Fleckenstein B: The two major phosphoproteins (pp65 and pp150) of human cytomegalovirus and their antigenic properties. JGen Virol68:1327-1337, 1987. 12. Landini MP, Guan MX, Jahn G , Lindenmaier W, Mach M, Ripalti A, Necker A, Lazzarotto T, Plachter B: Large scale screening of human sera with Cytornegalovirus recombinant antigens. J CIin Microbiol 28:1375-1379, 1990. 13 Landini MP, Lazzarotto T, Ripalti A, Guan MX, La Placa M: Antibody response to recombinant Lambda gtl 1 fusion proteins in Cytomegalovirus infection. J Clin Microbio/27:2324-2327, 1989. 14 Ripalti A, Landini MP, Mocarski E, La Placa M: Identification and preliminary use of recombinant lambda gtl 1 fusion proteins in Cytomegalovirus diagnosis. J Gen Virol70: 1247-1251, 1989. 15 Landini MP, Ripalti A, Sra K, Pouletty P: Human cytomegalovirus structural proteins: Immune reaction against pp150 synthetic peptides. J Clin Microbiol29:1868-1872, 1991. 16 Novak J, Sova P, Krchnak V. Hamsikova E, Zavadova H: Mapping of serologically relevant regions of human cytomegalovirus phosphoprotein pp150 using synthetic peptides. J Gen Virol72:1409-1413, 1991. 17 Plachter B, Wieczorek L, Scholl BC, Ziegelmaier R, Jahn G: Detection of Cytomegalovirusantibodies by an enzyme-linked Immunosorbent assay using recombinant polypeptides of the large phosphorylated tegument protein pp150. J Clin Microbiol30:201-206, 1992. 18 Rudolph SA, Kuhn JE, Kom K, Braun RW, Jahn G: Procaryotic expression of the major capsid protein of human cytomegalovims and antigenic cmssreactions with herpes simplex virus. J Gen Viro/71:2023-2031, 1990.