Immunoglobulin M Antibody Capture Enzyme Immunoassay With

Rapid Diagnosis of Acute Mumps Infection by a Direct Immunoglobulin M Antibody Capture Enzyme Immunoassay with Labeled Antigen JEAN-PIERRE GUT,* CLAUDE SPIESS, SYLVIE SCHMITT, AND ANDRÉ KIRN Laboratoire de Virologie and Groupe de Recherche de l'Institut National de la Santé et de la Recherche Médicale U74 Faculty of Medicine, University Louis Pasteur, 67000 Strasbourg, France Received 2 August 1984/Accepted 23 October 1984

A new immunoglobulin M (IgM) antibody capture enzyme immunoassay with peroxidase-labeled mumps antigen (dMACEIA) is described, and its suitability for practical diagnosis of acute mumps infection is evaluated. All 54 patients with proven mumps infection that were tested showed mumps-specific IgM antibodies. On the other hand, no specific IgM antibodies were present in 16 cases of suspected mumps that could not be confirmed by classical complement fixation serology, and IgM mumps virus antibodies could be detected neither in the sera of 100 healthy individuals nor in those of 16 patients positive for rheumatoid factor. In ali, 22 children with acute respiratory illness caused by parainfluenza virus and 44 patients with infections due to other viruses showed no IgM response in mumps dMACEIA. The particular characteristic in which complement fixation antibodies against mumps nucleocapsids appear before and disappear earlier than antibodies to the enveloped mumps virus could not be demonstrated in the dMACEIA. In an extensive epidemic of mumps virus infection, the dMACEIA gave a clear diagnosis of mumps infection in 200 out of 371 suspected cases. By day 2 of the illness, 71% of the patients had detectable IgM, and by day 3, all of them had detectable IgM. In 99% of the cases, dMACEIA gave a positive result in the first available serum specimens, most of which were negative for complement fixation antibodies. A positive but only moderate correlation was thus observed between the two serological procedures. IgM antibodies persisted for at least 6 weeks. The dMACEIA, performed in 3 h, offers a reliable, simple, and rapid alternative to routine methods for detection of acute mumps infection.

A rapid and reliable laboratory diagnosis of mumps virus disadvantage of the first technique, M-antibody capture infection is needed since the clinical symptomatology is hemadsorption, is the difficulty encountered in reading the often atypical, especially in cases of meningitis, orchitis, or test, and that of the second technique, M-antibody capture pancreatitis without evidence of parotitis. Conventional lab- radioimmunoassay, is the use of a radioisotope-labeled rea- oratory methods, such as virus isolation from saliva or spinal gent. In this report we describe a new direct IgM antibody fluid, and serological procedures, such as complement fixa- capture enzyme immunoassay (dMACEIA) with a peroxi- tion (CF) or hemagglutination inhibition, which require dase-labeled mumps antigen. A relatively large number of paired sera, delay the results. The demonstration of specific sera collected during an extensive epidemic of mumps that immunoglobulin M (IgM) class antibody is considered to be occurred in the eastern area of France during the winter of a useful tool for the rapid diagnosis of acute virus infections 1984 allowed us to assess the methodology described. since, as a rule, this antibody can be detected only for a limited period. The preliminary physical separation of IgM from IgG by either sucrose density gradient centrifugation MATERIALS AND METHODS (27) or exclusion chromatography (12), followed by testing the IgM-containing fractions by a conventional serological Sera. Group 1 consisted of acute- or convalescent-phase method, is generally reliable but laborious. Indirect tests, sera or both from patients with clinical mumps. Diagnoses including the immunofluorescence test (2), a radioimmunoas- were established by the presence of a significant (fourfold or say (6), or an enzyme-linked immunosorbent assay (ELISA) greater) increase in CF antibodies to mumps (40 patients) or (9, 16, 18-21, 23, 24) have been used to detect mumps-spe- by virus isolation from the throat or cerebrospinal fluid cific IgM. The problem with these indirect tests is that IgG (CSF) (14 patients). Group 2 consisted of single sera from 317 may inhibit the binding of IgM antibodies, particularly when patients whose clinical data suggested a mumps infection. high IgG antibody titers are present, thus giving false-nega- Group 3 consisted of 14 single and 16 paired sera from 22 tive results. Furthermore, rheumatoid factor of the IgM class patients with acute respiratory illnesses caused by parain- may cause false-positive results by reacting with antigen-IgG fluenza viruses; diagnoses were established through demon- complexes (26). These problems may be overcome by an stration of a significant increase in CF antibodies to parain- antibody capture assay in which human IgM is selectively fluenza viruses (6 cases), through a CF antibody titer higher attached to the solid phase by an anti-human IgM antibody. than 64 in a single serum (11 cases), or by virus isolation of Mumps-specific IgM is then detected by adding either virus parainfluenza virus type 1 (2 cases) or type 3 (3 cases); 9 of hemagglutinin and guinea pig erythrocytes (7, 25) or mumps these patients were 4 to 9 years old, and 13 were less than 4 antigen and its radioisotope-labeled antibody (3). The main years old. Group 4 consisted of 11 sera with rubella-specific IgM antibodies (titers varied between 800 and 12,800 as determined by M-antibody capture hemadsorption), 11 sera * Corresponding author. with hepatitis A IgM antibodies, 6 sera with herpes simplex

346 VOL. 21, 1985 DIRECT ENZYME IMMUNOASSAY FOR MUMPS IgM ANTIBODY 347 virus IgM antibodies, and 16 sera with cytomegalovirus IgM phere and washed five times with PBS-T. A 0.1-ml volume of antibodies (all determined by dMACEIA with titers between labeled antigen dilution in PBS-T was then added to each 200 and 25,600). Group 5 consisted of sera from 100 healthy well and incubated for 2 h at 37°C in a humidified atmos- individuals between 40 and 50 years of age. Group 6 con- phere. The optimal dilution of each kind and batch of antigen sisted of 16 single sera with IgG-ELISA mumps virus anti- was determined by box titration with a known positive bodies from patients for whom clinical data suggestive of a serum and a known negative one. recent mumps infection was lacking. These sera had varying After five washes with PBS-T, substrate solution (0.2 ml levels of rheumatoid factor. per well) was added. The substrate solution was prepared dMACEIA antigens. Allantoic fluid from mumps virus-in- immediately before use by dissolving 10 mg of orthophenyl- fected embryos was used as a source of antigen. The Enders enediamine in 25 ml of phosphate-citrate buffer (pH 5) strains of mumps virus (kindly provided by G. A. Denoyel, followed by the addition of 0.038 ml of 30% H202. After a Pasteur Institute, Lyons, France) was propagated by allan- 15-min incubation in the dark at room temperature, the toic inoculation in 10-day-old embryos at a dose of 1 reaction was stopped by adding 0.05 ml of 2 N H2SO4. The hemagglutinin unit per 0.2 ml. After 7 days at 35°C, allantoic absorbance was measured at a wavelength of 486 nm in a fluid and membranes were harvested. Three preparations of vertically measuring photometer (Kontron Analytique). The allantoic fluid antigen containing the virus-bound (V-) anti- mean absorbance of eight blank wells each containing 0.2 ml gen were prepared: (i) crude virus-containing allantoic fluid of substrate solution and 0.05 ml of 2 N H2SO4 was auto- clarified by low-speed centrifugation and dialyzed overnight matically subtracted from the absorbance of the test wells. against 20 volumes of 0.01 M phosphate-buffered saline Since comparison of the results obtained with the three (PBS); (ii) pelleted virus obtained by centrifugation of the preparations of V-antigen showed equal performances for latter suspension at 100,000 x g for 1.5 h at 4°C in an SW27 pelleted and purified viruses and insufficient ones with crude rotor and suspension by sonication to ca. 1:30 of the original virus, we used only pelleted V- or S-antigen (see below). The volume in 0.01 M carbonate buffer (pH 9.5); (iii) virus absorbance values of IgM antibody-positive and -negative purified in a discontinuous sucrose gradient as previously sera with control antigen were ca. 10 times lower than those described (14). Allantoic fluid from uninfected eggs treated of IgM antibody-negative sera with V-antigen (0.01 and 0.1, in the same ways served as control antigens. Ribonucleopro- respectively). The cutoff level was therefore defined by tein (S-) antigen was extracted from the membranes of the calculating the absorbance ratio of the specimen serum same eggs from which the pelleted virus was obtained. After compared with that of the negative control serum (S/N washing twice in PBS, a 20% suspension of the membranes ratio). An S/N ratio of 2.0, which generally lay between 3 in PBS was prepared by homogenization. The homogenate and 6 standard deviations (SD) above the mean value was first clarified by low-speed centrifugation and then obtained from four negative sera, was arbitrarily considered centrifuged at 100,000 x g for 1.5 h at 4°C in an SW27 rotor to indicate a specific reaction. Only nine serum samples to sediment most of the mumps virus nucleocapsids. These collected from patients clinically suspected of having mumps were finally suspended by sonication to 1:20 of the original virus infection showed an S/N ratio between 1.6 and 1.9 at a volume in 0.01 M carbonate buffer (pH 9.5). The protein dilution of 1:100 and were considered as negative. The concentrations of the different batches of antigen were endpoint titer was regarded as the highest serum dilution in determined by the Lowry method. which the S/N ratio was equal to or greater than 2.0. The peroxidase labeling of the different antigen prepara- Indirect ELISA for IgG antibody. Mumps virus IgG anti- tions was carried out by the periodate method in two steps bodies were determined by a classical indirect ELISA (21) (28). Horseradish peroxidase type IV (4 mg; Sigma Chemical with some differences: the technical conditions were the Co., St. Louis, Mo.) was dissolved in 1 ml of distilled water same as those of the dMACEIA except for the use of and oxidized by adding 0.2 ml of freshly prepared 0.1 M pelleted V-antigen without peroxidase labeling; IgG anti- NaIO4. After stirring for 2 h at room temperature in the dark, body was detected by peroxidase-conjugated rabbit anti-hu- the solution was dialyzed overnight at 4°C against 1 mM man IgG (-y-chain specific; Dakopatts) diluted at 1:2,000. sodium acetate buffer (pH 4.4). A solution of mumps virus Other methods. A standard CF test in microplates was antigen containing 4 mg of protein in 3 ml of 0.01 M used (4). Mumps CF antibodies were determined by using carbonate buffer (pH 9.5) was added to 1 ml of the activated pelleted V- and S-antigen-nonlabeled dMACEIA antigens. enzyme. The mixture was held for 2 h at room temperature Parainfluenza CF antigens were purchased from MA Bio- in the dark before 0.1 ml of an NaBH4 solution (4 mg/ml) was products (Walkersville, Mass.). Mumps virus isolation from added. The labeled antigen was kept at 4°C for 5 h, then the throat or from spinal fluid and parainfluenza virus dialyzed against PBS overnight at 4°C, and finally stored in isolation from the throat were carried out in an established small volumes at -70°C. monkey kidney cell line (MK2 cells). The cytopathic effect dMACEIA procedure. The procedure used was that pre- was generally noted within 5 to 7 days. The neutralization viously described for detection of cytomegalovirus IgM test with specific reference sera (MA Bioproducts) permitted antibodies (22) but with some modifications. Polystyrene the identification of the agents. Rheumatoid factor was flat-bottom microtiter plates (Kontron Analytique, Trappes, determined by a commercially available latex test (Hoechst- France) were coated with 0.1 ml per well of rabbit anti-IgM Behring, Rueil Malmaison, France) and by an enzyme immunoglobulin (,u-chain specific; Dakopatts, Copenhagen, immunoassay, previously described (11), with heat-aggre- Denmark) at a dilution of 1:500 in 0.1 M bicarbonate buffer gated IgG. To evaluate the antibody class, sera were frac- (pH 9.6) and sealed with tape. After overnight incubation at tionated by sucrose density gradient centrifugation (27). The 4°C in a moist chamber, the plates were washed five times effective separation of IgM and IgG antibodies was demon- with PBS containing 0.05% Tween 20 (PBS-T) and were strated by the immunodiffusion test of Ouchterlony with shaken dry. On the same day, 0.1 ml of serum specimen at rabbit anti-human,u- and y-chain sera (Dakopatts). IgM and an initial dilution of 1:100 in PBS-T with 0.4% bovine serum IgG antibodies were found in fractions 1 to 10 and 8 to 20, albumin was added to each of two wells. The sealed plates respectively. Each fraction was diluted 1:50 before its use in were incubated at 370C for 30 min in a humidified atmos- the dMACEIA. 348 GUT ET AL. J. CLIN. MICROBIOL. 0.8 E 0.7 c Co 0.6 w D 0.5 -J 0.4 0.3 0.2

0 ------Co - CUTOFF LEVEL ------0.1 00w0'o^0@@90"'°"'0'

lig+I +II + I+I-I- I+1+1+ +- +1-1-1- 1+-+-+- 1 2 3 4 5 6 7 8 91011121314151617181920 FRACTION NUMBER FIG. 1. dMACEIA of a fractionated serum positive (O) and negative (O) for mumps IgM antibodies. The two sera, positive for mumps-specific IgG antibodies, were fractionated in a sucrose gradient, and 1:50 dilutions of the fractions were tested by dMACEIA. The presence or absence of IgM and IgG was measured by the technique of Ouchterlony.

RESULTS tionated sera each from groups 1 and 5 are shown in Fig. 1. These data indicate that the specific IgG antibody contained Reliability of dMACEIA for detecting acute mumps infec- in specific IgM-positive or IgM-negative sera did not give tion. To assess the specificity of dMACEIA for the IgM false-positive reactions. antibody class, sera from five patients with clinical mumps The intraassay variability in the absorbance values of the and a significant increase in CF antibody titers (group I) dMACEIA was estimated in three different sera (intermedi- were fractionated by sucrose density gradient centrifugation. ate positive, low positive, and negative in mumps IgM The IgM fractions ofthe five sera were positive in dMACEIA antibodies) which were diluted 1:100, pipetted into 30 wells, with antibody titers up to 12,800, whereas the IgG fractions and tested with pelleted V-antigen. The coefficients of vari- were negative. Moreover, the IgM and IgG fractions of five ation (SD/mean) werç 0.8, 1.2, and 2.0, respectively, in the sera from individuals with remote mumps infection showing three sera (Table 1). To investigate the interassay variability a mean IgG ELISA titer of 3,200 (group 5) were negative in of the dMACEIA, the same intermediate positive and nega- the dMACEIA. The results obtained with one of the frac- tive sera were assayed on 20 different days with the same batch of pelleted V-antigen. The coefficients of variation were 13.8 and 23.2, respectively, in the two sera (Table 1). Sixty convalescent-phase sera of 54 patients with clinical TABLE 1. Intra- and interassay variability of the dMACEIA' mumps and a significant increase in CF antibodies to mumps Absorbance at 486 nm Coefficient or a virus isolation (group 1) tested by dMACEIA showed of varia- positive reactions with IgM-antibody titers ranging from 800 Parameter (type of serum) M Maxi- Mini- SD tion (SDi to 102,400. Conversely, 16 patients with a mumps-like illness mum mum mean %) who had paired serum samples that remained negative or Intraassay variability showed low stable titers in the CF test did not show positive Intermediate positive 0.822 0.950 0.710 0.006 0.8 reactions in dMACEIA. For comparison, dMACEIA was Low positive 0.237 0.280 0.180 0.003 1.2 carried out on sera of 100 healthy individuals (group 5) Negative 0.108 0.150 0.070 0.002 2.0 whose ages varied from 40 to 50 years (Table 2). Although 97% of these had a remote mumps virus infection (IgG- Interassay variability ELISA antibody titer ranging from 800 to 102,400), none Intermediate positive 0.860 1.130 0.690 0.118 13.8 showed IgM antibodies. Negative 0.149 0.190 0.090 0.035 23.2 To explore the possibility of cross-reactivity with parain- a The sera were tested at a dilution of 1:100 with the same preparation of fluenza and other viruses, sera from 22 patients with acute pelleted V-antigen. respiratory illnesses caused by parainfluenza viruses (group b An intermediate positive serum (titer, 12,800), a low positive serum (titer, 200), and a negative serum were tested 30 times in the same test. 3) and from 44 patients with hepatitis A, rubella, herpes C An intermediate positive serum (titer, 12,800) and a negative serum were simplex virus, or cytomegalovirus primary infection (group tested on 20 different days. 4) were tested by the dMACEIA. Mumps CF antibodies VOL. 21, 1985 DIRECT ENZYME IMMUNOASSAY FOR MUMPS IgM ANTIBODY 349

TABLE 2. Distribution of mumps IgG and IgM antibody titers in sera from 100 healthy individuals Antibody No. of sera based on antibody titer: class Negative 800 1,600 3,200 6,400 12,800 25,600 51,200 102,400 IgGb 3 6 31 30 15 4 5 4 2 IgMc 1Od 0 0 0 0 0 0 0 O a Aged 40 to 50 years. b Assayed by ELISA with pelleted V-antigen. C Assayed by dMACEIA with pelleted V-antigen. d Mean S/N ratio ± SD, 0.85 ± 0.24. were also assayed in sera from group 3. None of the sera of Of eight sera sampled more than 28 days after the onset of groups 3 and 4 exhibited mumps IgM antibodies (Table 3). mumps, four showed a negative IgM result against S-antigen Rheumatoid factor did not give false-positive results in the and one showed a negative result against V-antigen (Fig. dMACEIA. In group 6, 16 sera with IgG-ELISA titers 2B). Late sera always showed a higher S/N ratio for V-an- against mumps virus of 3,200 to 51,200 and with various tigen as compared with that for S-antigen, respective values levels of rheumatoid factor ranging from 20 to 1,280 and 100 being 4.7 and 1.8. Based on these data, we chose to use to 12,800 as determined, respectively, by a latex fixation test V-antigen only to diagnose acute mumps infection by IgM and ELISA, were negative in the dMACEIA (S/N ratio ± antibody assays. SD, 0.99 ± 0.16). Rheumatoid factor was also determined by Diagnosis of current mumps infection by dMACEIA. The ELISA in the 200 patients with acute mumps infection development of the IgM response was studied in the sera of (group 1 and part of group 2) and 20 showed positive 371 patients whose clinical symptoms could be attributed to reactions with a mean titer ± SD of 225 ± 337. With approximately the same 50% male/female ratio, this group comprised 7 children and 13 adults. B Comparative reactivity of S-antigen and pelleted V-antigen 10.2i A in dMACEIA. Most of the sera sampled on days 1, 2, or 3 after the onset of illness showed IgM antibody against pelleted V- and S-antigens (Fig. 2A). In four cases S/N ratios were lower than 2.0 (1.6 to 1.9) and higher than 2.0 (2.3 to 2.6) against S- and V-antigens, respectively. The reverse 8 was never observed. In particular, all 171 sera from patients whose clinical data were suggestive of a mumps virus infection (group 2), which were IgM negative in the first screening with V-antigen were also negative with S-antigen. 7

TABLE 3. IgM antibodies and increases or high levels in CF mumps virus antibodies in patients with parainfluenza virus O e infection, hepatitis A, rubella, herpes simplex virus, and cytomegalovirus primary infections z No. of pa- 5 tients with No. of pa- Co No. of increased tients with Diagnosis patients or high mumps tested titer in IgM anti- mumps CF bodies 4 antibodies Parainfluenza CF titer increase 6 2 0 CF titer higher than 64 il 2 0 3 Isolation of parainfluenza 2 1 0 type 1 Isolation of parinfluenza 3 0 0 type 3 2 Other viruses Hepatitis A il NDd 0 Rubella il ND 0 1 Herpes simplex virus 6 ND 0 Cytomegalovirus 16 ND 0 S Ag V Ag S Ag V Ag a Fourfold increase or titer higher than 64. b FIG. 2. Correlation between the use of pelleted enveloped virus Assayed by dMACEIA with pelleted V-antigen; mean S/N ratio + SD, (V-antigen) and nucleocapsids (S-antigen) in dMACEIA in sera 0.99 + 0.25. the first 3 or 28 after the onset of ' Sera of these patients showed specific IgM antibodies (titers between 200 sampled during days (A) days (B) and 25,600) determined by IgM antibody capture hemagglutination assay or mumps virus infection. A ratio of the specimen serum absorbance dMACEIA. value to that of the negative control serum (S/N ratio) equal or d ND, Not done. greater than 2.0 correlates with the presence of IgM antibodies. 350 GUT ET AL. J. CLIN. MICROBIOL. mumps virus infection (groups 1 and 2). In 200 patients 102400 oe e (54%), IgM antibodies were found. These appeared on day 1 or 2 after the onset of the disease. Indeed, all except two sera sampled on day 1 and all sera collected on days 2 or 3 51200 . m e contained IgM antibodies (Fig. 3). Also, 84 of 86 acute-phase sera in which CF antibodies were still absent exhibited IgM 25600 e - m e antibodies (Fig. 4). A moderate but statistically significant positive correlation (r = 0.17; P < 0.01) between dMACEIA and CF is shown in Fig. 4. In effect, antibodies revealed by 12800 * _ m the first test appeared earlier than those revealed by the second (Fig. 3). The mean IgM titer was 117 times that of CF 6400 m A _ _ _ (4,802 and 41, respectively). When CF antibodies deter- w and were first mined with S-antigen pelleted V-antigen _ detected on days 2 and 3 after the onset of the symptoms, 3200 a ou f a g dMACEIA with pelleted V-antigen already showed mean titers of 663 and 1,280 on days 2 and 3, respectively. Study > 1600 * n of the temporal course of the antibody titers also revealed CD *

800 m is 256 . 400 m 128` 3 t t . n & 64 o 2001 i 8 *o 18 *° r.O.17 t 1§ g a. . 100 c,LaL l6 *° *8 la *e *§ 38 I<10

8 8 8 O8 *8 18 .8 ° <8 8 16 32 64 128 256 512 I <81 l--j0Io -oo CF TITRE FIG. 4. Correlation of mumps IgM and CF antibody titers determined with pelleted enveloped virus in 260 serum samples of 200 I 2 3 4 5 B 7-14 15-2122-2829536-43 patients with acute mumps infection. The correlation coefficient was calculated by linear regression analysis. 25600f S

12800 that the peak was reached on weeks 1 and 3 for IgM and CF antibodies, respectively. Mumps IgM persisted for at least 41 days in one patient but disappeared by day 64 in another 6M00 . patient. . CSF from 16 meningitis patients who showed mumps IgM 3200 1 antibodies in their sera with titers ranging from 20 to 12,800 were also tested dMACEIA and w by IgG-ELISA. Although 1600 t 11 of these CSF samples yielded mumps virus on culture and 5 did not, none showed specific IgM or IgG antibodies. F 800 * * * DISCUSSION 400 t S The aim of this study was to develop a new assay for 200 S demonstration of mumps IgM antibodies, to analyze it, and to evaluate its suitability for practical serological diagnosis of mumps infection as compared with that of the commonly 100...... used techniques. The dMACEIA is easy to perform. Antigen preparation is <100: relatively simple. Since sensitized microtiter plates can be stored at -80°C for several weeks, the test can be carried 1 2 3 4 5 6 7-1415-2122-2829-3536-1 out in 3 h and needs only a small amount of serum. We also demonstrated that dMACEIA is specific for the IgM class of DAYS AFTER ONSET OF ILLNESS antibody to mumps (Table 1). All sera from patients with FIG. 3. Development of mumps CF antibodies determined with clinical mumps and a significant increase in CF antibodies or nucleocapsid antigen (O) and with pelleted enveloped virus (0) and from which the virus was isolated showed specific IgM IgM antibodies in 74 serum samples of 55 patients with acute mumps (group 1). Conversely, patients with an illness suggestive of infection...... mumps whose paired sera remained antibody negative or VOL. 21, 1985 DIRECT ENZYME IMMUNOASSAY FOR MUMPS IgM ANTIBODY 351 displayed a low stationary titer of CF antibodies did not symptoms were suggestive of mumps infection and did so show a positive reaction in dMACEIA. Sera were also with the first available serum specimen in 99% of these negative in 100 healthy individuals aged 40 to 50 years (group cases. Despite a statistically positive correlation between the 5). These results would be expected since only the IgM of results obtained with CF and dMACEIA, the latter assay the sera from these patients attached to the wells coated with was more sensitive. This was indicated, as previously shown anti-human IgM. Other studies with the same highly specific (23), by the ability to demonstrate relatively high dMACEIA procedure for capturing IgM antibody were used in detecting absorbance values in the early-phase sera which were neg- IgM antibodies to hepatitis A virus (8), rubella virus (15), ative or had only a low level of CF antibodies (Fig. 3). Peak and cytomegalovirus (22). For diagnosis of mumps infection, titers of IgM were reached in week 1 after the onset of the we consider dMACEIA preferable to M-antibody capture illness. IgM persisted in a patient 41 days after the onset of radioimmunoassay and M-antibody capture hemadsorption infection but was absent at day 64 in another patient. (7, 25), which require radio-labeled antibody or sometimes Adequate convalescent-phase sera were not available in our present difficulties in interpretation of the test results. study to determine how long these antibodies actually per- The dMACEIA was reproducible since the intraassay sist in the course of infection. The possibility that prolonged variability of absorbance values in the same test showed IgM responses may occur (2, 25) emphasizes the importance coefficients of variation ranging from 0.8 to 2.0. The inter- of interpreting laboratory results in the light of clinical and assay variability of the test system assayed on 20 different epidemiological details. days exhibited higher coefficients of variation (13.8 and 23.2) Contrary to what has been observed in neurological dis- but was still considered acceptable. We chose a cutoff level eases due to herpes simplex virus or cytomegalovirus, between negative and positive scoring, defined as 3 to 6 SD mumps IgM and IgG antibodies could not be found in the above the mean absorbance of the negative reference serum, CSF of 16 patients with proven mumps meningitis. Of these keeping in mind the possibility that such a level might give patients, 11 had CSF which contained mumps virus, but 5 false-negative results. did not. The reason for the absence of specific IgM and IgG To find out which kind of mumps antigen was preferable antibodies did not seem to be a technical one attributable to for the diagnosis of acute mumps, we compared results the occurrence of immune complexes. Our data indicate that obtained with pelleted enveloped virus (V-antigen) and nu- detection of IgM in serum may be preferable to virus cleocapsids (S-antigen). It is well known that the S-antibody isolation from CSF, in view of the greater sensitivity of the frequently appears in the course of infection before the former. V-antibody, but the latter persists longer (13). This was also The dMACEIA is a simple, rapid, and reliable test for observed in our results for CF antibodies but could not be mumps-specific IgM, and it is particularly valuable for the confirmed for IgM by using pelleted V- and S-antigens in the rapid diagnosis of atypical mumps infection and for con- dMACEIA (Fig. 2). Thus, the diagnosis of acute mumps firming clinically suspected cases of the disease. The could not be made earlier by demonstrating S-IgM. On the dMACEIA is also useful when sera are taken too late for contrary some acute-phase sera showed IgM against V-an- diagnosis based on a significant increase in CF titer or when tigen but not against S-antigen. A possible explanation for the only serum collected is from the convalescent phase. this may be the difficulty of obtaining high specific activity with membrane-extracted nucleocapsids as compared with ACKNOWLEDGMENTS pelleted virus particles from allantoic fluid. Mumps virus has This work was supported by Direction de la Recherche des Etudes an antigenic relationship with other viruses of the paramyxo- et des Techniques grant no. 82.34.333.01.470.75.01, under the col- virus group (5) as evidenced by serological cross-reactions laboration agreement between the Institut National de la Santé et de (17). In accordance with certain other procedures for detect- la Recherche Médicale and Caisse Nationale d'Assurance Maladie ing mumps IgM (18, 24) but contrary to the M-antibody des Travailleurs Salariés. capture hemadsorption (25), our data showed that none of We thank T. Kirn for assistance with the English translation. the patients with acute respiratory illnesses caused by par- LITERATURE CITED ainfluenza viruses (group 3) had IgM antibodies to mumps 1. Bartfeld, H. 1960. Incidence and significance of seropositive virus (Table 3). We could not determine the reason for this tests for rheumatoid factor in non-rheumatoid diseases. Ann. discrepancy. Cross-reactions were also not observed in 44 Intern. Med. 52:1059-1066. sera containing IgM antibodies against hepatitis A virus, 2. Brown, G. C., J. V. Baublis, and T. P. O'Leary. 1970. Devel- rubella virus, herpes simplex virus, or cytomegalovirus opment and duration of mumps fluorescent antibodies in various (Table 3). immunoglobulin fractions of human serum. J. Immunol. 104:86-94. Rheumatoid factor often gives false-positive results in 3. Burkhardt, F., U. Schilt, and R. Y. Andres. 1982. Virologische indirect assays where the antibodies of different classes are Diagnostik der Mumps-Infection: ein Festphasen-Radioim- not separated before testing (26). It is known that the muntest zum Nachweiss mumpsspezifischer IgM-Antikorper immunocapture of IgM reduces or eliminates this problem (MACRIA). Schweiz. Med. Wochenschr. 122:638-643. (10), as we observed in our results. The use of the latter 4. Casey, H. L. 1965. Adaptation of LBCF-method microtech- procedure is highly desirable since virus-infected individuals nique, p. 1-34. In Standardized diagnostic complement fixation frequently exhibit rheumatoid factor in their serum (1), as and adaptation to microtest. U.S. Public Health Service mono- was the case in our study, in which 10% of mumps-infected graph no. 74. U.S. Government Printing Office, Washington, patients showed rheumatoid factor. D.C. a sensitive method for the 5. Cook, M. K., B. E. Andrews, H. H. Fox, H. C. Turner, W. D. The dMACEIA is early diag- James, and R. M. Chanock. 1959. Antigenic relationship among nosis of acute mumps infection. Mumps IgM antibodies the "newer" myxoviruses (parainfluenza). Am. J. Hyg. appeared soon after the onset of clinical symptoms. By day 69:250-263. 2 of the illness, 71% of the patients had detectable IgM in 6. Daugharty, H., D. T. Warfield, W. D. Hemingway, and H. L. their serum, and by day 3, all of the patients had detectable Casey. 1973. Mumps class-specific immunoglobulin in radioim- IgM (Fig. 3). The dMACEIA thus gave a clear diagnosis of munoassay and conventional serology. Infect. Immun. mumps infection in 200 out of 371 patients whose clinical 7:380-385. 352 GUT ET AL. J. CLIN. MICROBIOL.

7. Denoyel, G. A., A. Gaspar, and D. Peyramond. 1982. A solid- virus by solid-phase enzyme immunoassay. J. Virol. Methods phase reverse immunosorbent test (SPRIST) for demonstration 4:249-257. of specific-mumps virus IgM-class antibody. Arch. Virol. 19. Nicolai-Scholten, M. E., R. Ziegelmeyer, F. Behrens, and W. 71:349-352. Hopken. 1980. The enzyme-Iinked immunosorbent assay 8. Duermeyer, W., and J. van der Ween. 1978. Specific detection of (ELISA) for determination of lgG and IgM antibodies after IgM antibodies by ELISA, applied in hepatitis A. Lancet infection with mumps virus. Med. Microbiol. Immunol. ii:684-685. 168:81-90. 9. Engel, S., V. Goldstein, and I. Sarov. 1980. Enzyme-linked 20. Popow-Kraupp, T. 1981. Enzyme-linked immunosorbent assay immunosorbent assay for detection of antibodies to mumps (ELISA) for mumps virus antibodies. J. Med. Virol. 8:79-88. virus. Isr. J. Med. Sci. 16:461-464. 21. Sakata, H., M. Hishiyama, and A. Sugiura. 1984. Enzyme-linked 10. Gispen, R., J. Nagel, B. Brand-Saathof, and S. de Graaf. 1975. immunosorbent assay compared with neutralization tests for Immunofluorescence test for IgM rubella antibodies in whole evaluation of live mumps vaccines. J. Clin. Microbiol. 19:21-25. serum after absorption with anti-FC. Clin. Exp. Immunol. 22. Schmitz, H., U. von Deimling, and B. Flehming. 1980. Dectec- 22:431-437. tion of lgM antibodies to cytomegalovirus (CMV) using enzyme- 11. Gripenberg, M., F. Wafin, H. Isomaki, and E. Linder. 1979. A labelled antigen (ELA). J. Gen. Virol. 50:59-68. simple enzyme immunoassay for the demonstration of rhuma- 23. Ukkonen, P., M. L. Granstrom, and K. Penttinen. 1981. Mumps- toid factor. J. Immunol. Methods 31:109-118. specific immunoglobulin M and G antibodies in natural mumps 12. Gupta, J. D., V. Peterson, M. Stout, and A. M. Murphy. 1971. infection as measured by enzyme-linked immunosorbent assay. Single-sample diagnosis of recent rubella by fractionation of J. Med. Virol. 8:131-142. antibody on Sephadex G-200 column. J. Clin. Pathol. 24:547-550. 24. Ukkonen, P., O. Vaisanen, and K. Penttinen. 1980. Enzyme- 13. Henle, G., S. Harris, and W. Henle. 1948. The reactivity of linked immunosorbent assay for mumps and parainfluenza type various human sera with mumps complement fixation antigens. 1 immunoglobulin G and immunoglobulin M antibodies. J. Clin. J. Exp. Med. 188:133-147. Microbiol. 11:319-323. 14. Jensik, S. C., and S. Silver. 1976. Polypeptides of mumps virus. 25. Van der Logt, J. M., F. W. A. Heessen, A. M. van Loon, and J. J. Virol. 17:363-373. van der Veen. 1982. Hemadsorption immunosorbent technique 15. Krech, U., and J. A. Wilhelm. 1979. A solid-phase immunosor- for determination of mumps immunoglobulin M antibody. J. bent technique for the rapid detection of rubella IgM by Clin. Microbiol. 15:82-86. haemagglutination inhibition. J. Gen. Virol. 44:281-286. 26. Vejtorp, M. 1980. The interference of IgM rhumatoid factor in 16. Leinikki, P. O., I. Shekarchi, N. Tzan, D. L. Madden, and J. L. enzyme-linked immunosorbent assays of rubella IgM and IgG Sever. 1979. Evaluation of enzyme-linked immunosorbent assay antibodies. J. Virol. Methods 1:1-9. (ELISA) for mumps virus antibodies. Proc. Soc. Exp. Biol. 27. Vesikari, T., and A. Vaheri. 1968. Rubella: a method for rapid Med. 160:363-367. diagnosis of a recent infection by demonstration of the IgM 17. Lennette, E. H., F. W. Jensen, R. W. Guenther, and R. L. antibodies. Br. Med. J. i:221-223. Magoffin. 1963. Serologic responses to para-influenza viruses in 28. Wilson, M. B., and P. K. Nakane. 1978. Recent developments in patients with mumps virus infection. J. Lab. Clin. Med. the peroxidase method of conjugating horseradish peroxidase 61:780-788. (HRPO) to antibodies, p. 215-224. In W. Knapp (ed.), Im- 18. Meurmann, O., P. Hanninen, R. V. Krishna, and T. Ziegler. munofluorescence and related techniques. Elsevier Biomedical 1982. Determination of IgG- and IgM-class antibodies to mumps Press, Amsterdam.