Apartado Postal 510-3, Colonia Miraval, Cuernavaca, Morelos 62271,Mexico,1 and Se,ao de Virologia,. Instituto Evandro Chagas, Fundaqdo Servicos de ...
Vol. 1, No. 1
CLINICAL AND DIAGNOSTIC LABORATORY IMMUNOLOGY, Jan. 1994, p. 89-94
1071-412X/94/$04.00+0 Copyright C) 1994, American Society for Microbiology
Neutralizing Antibody Immune Response in Children with Primary and Secondary Rotavirus Infections CARLOS F. ARIAS,'* SUSANA LOPEZ,' JOANA D'ARC P. MASCARENHAS,2 PEDRO ROMERO.' PAUL CANO,' YVONE B. GABBAY,2 RONALDO B. DE FREITAS,2 AND ALEXANDRE C. LINHARES2 Departamento de Biologia Molecular, Instituto de Biotecnologia, Universidad Nacional Aut6noma de Mexico, Apartado Postal 510-3, Colonia Miraval, Cuernavaca, Morelos 62271, Mexico,1 and Se,ao de Virologia, Instituto Evandro Chagas, Fundaqdo Servicos de Saude Publica, 492, 66.050 Belem, Pard, Brazil2 Received 10 May 1993/Returned for modification 7 September 1993/Accepted 23 September 1993
We have characterized the neutralizing antibody immune response to six human rotavirus serotypes (GI to G4, G8, and G9) in Brazilian children with primary and secondary rotavirus infections and correlated the response with the G serotype of the infecting rotavirus strain. Twenty-five children were studied: 17 had a single rotavirus infection, 4 were reinfected once, and 4 experienced three infections. Two of the reinfections were by non-group A rotaviruses. Among the 25 primary infections, we observed homotypic as well as heterotypic responses; the serotype Gl viruses, which accounted for 13 of these infections, induced mostly a homotypic response, while infections by serotype G2 and G4 viruses induced, in addition to the homotypic, a heterotypic response directed primarily to serotype Gl. Two of the primary infections induced heterotypic antibodies to 69M, a serotype G8 virus that by RNA electrophoresis analysis was found not to circulate in the population during the time of the study. The specificity of the neutralizing antibody immune response induced by a virus of a given serotype was the same in primary as well as secondary infections. These results indicate that the heterotypic immune response induced in a primary rotavirus infection is an intrinsic property of the virus strain, and although there seem to be general patterns of serotype-specific seroconversion, these may vary from serotype to serotype and from strain to strain within a serotype. type G8) was reported to be a new genomic P type (35) that might represent a fifth human P serotype, but this has not been
Group A rotaviruses are the leading cause of severe dehydrating gastroenteritis in children under 3 years of age, and there is considerable interest in developing an effective vaccine (22). The surface of group A rotaviruses is formed by two proteins, VP4 and VP7. VP4 forms spikes that extend from the surface of the virus particles and is involved in a variety of viral functions, including virulence, agglutination of erythrocytes, and trypsin-enhanced infectivity (13). VP7 is a glycoprotein that, in addition to VP4, has been proposed to be responsible for the initial attachment of rotaviruses to the target cells (13). The antibody response to these proteins has the ability to neutralize the infectivity of the virus in vitro as well as in vivo (28, 32, 36), and the specificity of these antibodies in neutralizing different rotavirus strains has been used to classify rotaviruses in various serotypes. Since VP7 and VP4 both induce neutralizing antibodies, the viruses can be classified by either of these two surface proteins (13, 14, 18, 21). Based on VP7, 14 different G serotypes (G for glycoprotein) have been identified among group A rotaviruses so far (2-4, 13, 16, 37). Nine of these serotypes (Gl to G4, G6, G8 to G10, and G12) infect humans, although four of them (Gl to G4) appear to account for the majority of isolates (23). The G serotypes have been defined by cross-neutralization assays with hyperimmune animal sera to the complete virus particles (21). Based on VP4, at least 11 P types (P for protease sensitivity) have been found by antigenic reactivity (18) and genomic (hybridization and amino acid sequence) analysis (13, 20, 35). Four of the genomic P types have been isolated from humans (13, 14), corresponding to serotypes PlA, PlB, P2, and P3, defined by neutralization with hyperimmune sera directed to the VP4 protein (18). Recently, the human rotavirus strain 69M (sero*
confirmed serologically. Both homotypic and heterotypic neutralizing antibody (NtAb) serological responses have been observed in rotavirus infections of seronegative children (8, 10, 11, 15, 34, 40). However, there is very limited information which correlates the serotype of the infecting virus with the specificity of the neutralizing immune response in primary rotavirus infections (5, 17). Determination of the specificity of the neutralizing immune response induced by the various rotavirus serotypes and study of the role of this response in protection against subsequent reinfections or disease would be most useful for designing rotavirus vaccines. In this work, we characterized the neutralizing immune response to rotavirus serotypes Gl to G4, G8, and G9 in children with serologically defined primary infections and reinfections and correlated this immune response with the G serotype of the infecting virus. MATERIALS AND METHODS
Patients and specimens. We studied the immune response to rotavirus infection in 25 children who were part of a longitudinal study carried out in Belem, Brazil (24). In that study, the children were monitored from birth to 3 years of age. While they were in the hospital, feces were collected daily from the newborn children; when the child went home, fecal samples were obtained on alternate days up to day 14, and feces samples were collected during any diarrheal episode. After day 14, feces samples were collected fortnightly or whenever signs of diarrhea were present. Serum samples were routinely collected every 6 months, and acute- and convalescent-phase serum samples were collected during any diarrheal episode. At
Corresponding author. Phone: (5273) 11-4900. Fax: (5273) 17-2388. 89
90
ARIAS ET AL.
the time of this study, about one-half of the acute- and convalescent-phase sera had been used up; therefore, in those cases, the routinely collected sera were used as preinfection and postinfection sera to study the immune response. The acute-phase serum samples were collected 1 to 2 days after onset of symptoms, and convalescent-phase serum was obtained 2 to 3 weeks later. The time that the preinfection serum samples were collected ranged from 4 days to 2.5 months (mean, 29 days) before the onset of symptoms, while the postinfection serum samples were collected from 1 to 6 months (mean, 3.2 months) after infection. Most children were between 6 and 24 months old (mean, 13 months) at the time of primary infection. IgM and IgG ELISA. Immulon II (Dynatech Laboratories, Inc.) microtiter plates were coated with a 1:5,000 dilution of goat anti-human rotavirus strain D (kindly provided by Harry B. Greenberg, Stanford University) in phosphate-buffered saline (PBS) for the enzyme-linked immunosorbent assay (ELISA). After overnight incubation at 4°C, the plates were washed twice with PBS and blocked with 3% fetal bovine serum in PBS for 1 h at 37°C. The plates were then washed twice with PBS and incubated for 2 h at 37°C with a Freonextracted MA104 cell lysate that had been infected with rotavirus SA114fM or mock infected. After the plates were washed four times, serial dilutions of the children's sera were added to duplicate wells and incubated for 1 h at 37°C. The plates were then washed four times and incubated with a 1:1,000 dilution in PBS of goat anti-human immunoglobulin G (IgG) or IgM conjugated to peroxidase (Kirkegaard & Perry Laboratories) and incubated for 1 h at 37°C. The plates were washed four times, and the presence of peroxidase activity was detected by incubation fot 30 min at room temperature with o-phenylenediamine hydrochloride (Sigma Chemical Co.) as the substrate. The reaction was stopped with 4 N H2SO4, and the optical density was read at 492 nm. The IgG and IgM antibody titers were defined as the highest serum dilution that gave an optical density of .0.2 and greater than twice the negative control value obtained when mock-infected cells were used as the antigen. Serotyping ELISA. The serotyping ELISA was done as described previously (33). NtAb assay. NtAb titers in the children's sera were measured by an immunochemical focus reduction neutralization test (1). The titer of NtAb in a serum sample was defined as the highest serum dilution at which a reduction of at least 60% in the number of infected cells was observed compared with controls for which PBS had been used instead of serum. RESULTS ELISA antibody response. The aim of this study was to determine the specificity of the NtAb immune response of children who had experienced primary and secondary infections with rotavirus and associate this response with the G serotype of the infecting rotavirus strain. Since the children included in this work were followed up from birth in a longitudinal study, it is reasonable to assume that the first rotavirus infections detected represented primary infections. To support this assumption, we analyzed the immunoglobulin class specificity of the children's serum antibody response, since it has been shown that the presence of virus-specific IgM is a reliable marker for primary rotavirus infections (17, 19). Of the 25 children studied, 17 had a single rotavirus infection and 8 were reinfected once or twice with rotavirus. As mentioned in Materials and Methods, some of the acuteand convalescent-phase serum samples collected from the
CLIN. DIAGN. LAB. IMMUNOL.
TABLE 1. ELISA IgM and IgG antibody response for serum samples from 25 children with presumed primary rotavirus infections No. of positive samples/total
(geometric mean titer)
Serum'
IgM
IgG
Acute phase Preinfection
8/14 (218) 0/9
2/14 (400) 0/9
Convalescent phase Postinfection
9/10 (216) 0/15
8/10 (566) 12/15 (599)
Total
15/25
20/25
U The acute-phase or preinfection sera were not available for two of the
children.
infected children included in this study were no longer available. In such cases, the preinfection and postinfection serum samples that had been obtained routinely were used instead. In 15 of the 25 (60%) presumed primary infections, we detected IgM in either the acute- or convalescent-phase serum or in both serum samples, suggesting that these were indeed primary infections. Eight of 14 acute-phase serum samples had detectable virus-specific IgM levels (titer, 1:100 to 1:400), while none of the 9 preinfection serum samples did (Table 1). In addition, 9 of 10 convalescent-phase serum samples had IgM (titer, 1:100 to 1:400), while none of the 15 postinfection serum samples had any detectable IgM. Only two of the children had IgM in both the acute- and convalescent-phase serum samples. We had an acute-phase serum sample for only 3 of the 10 IgM-negative children; for the other 7, we had only the preinfection serum. Furthermore, for all 10 IgM-negative cases, we had postinfection rather than convalescent-phase serum. Therefore, it is not surprising that we could not find IgM in these children, since this antibody class appears earlier than IgG and apparently disappears after 2 to 3 weeks (17). No IgM response was observed in the reinfections. We also analyzed the IgG response; 2 of the 14 acute-phase serum samples had detectable IgG (titer, 1:200 and 1:800), and in both samples, were also detected IgM. None of the 9 children whose preinfection sera were analyzed had preexisting IgG. For the convalescent-phase and postinfection serum samples, 8 of 10 and 12 of 15, respectively, had detectable IgG levels, giving a total of 20 of 25 children (80%) with detectable IgG levels (titer, 1:200 to 1:3,200 in both types of sera). In 18 of these cases, there was seroconversion (fourfold increases in the titer), and in 2 cases, the IgG titer did not change. Thus, the absence of rotavirus-specific IgG in the preinfection sera and the presence of IgM antibody in 90% of the convalescentphase sera from the children studied seem to indicate that the first rotavirus infections detected were primary infections. NtAb response in primary infections. The G serotype of 21 of the 25 rotavirus strains isolated from the primary infections had been reported previously (25). In this work, we determined the G serotype of two of the strains that could not be typed in the previous study (patients 24.075 and 24.145), both of which were shown to have serotype G3 specificity. Overall, 13 of the rotavirus strains isolated from primary infections were serotype Gl, 4 were serotype G2, 2 were serotype G3, and 4 were serotype G4. Two rotavirus strains were untypeable but had been shown before to belong to subgroup II (25). We analyzed the specificity of the NtAb response induced by these rotavirus strains; the acute-phase or preinfection sera and the convalescent-phase or postinfection sera were tested for their ability to neutralize the infectivity of human rotavi-
NEUTRALIZING IMMUNE RESPONSE TO ROTAVIRUS
VOL. 1, 1994 TABLE 2. NtAb seroconversion in children witi presumed primary rotavirus infections Patient no.'
23.979* 24.004* 24.013* 24.059 24.093 24.108* 24.175 24.191 24.376* 24.378* 24.383 24.384* 24.415
Infecting virus G serotype
1
1 1 1 1 1
1 1 1 1 1 1 1
Seroconversionb to G serotype
1 1 1 1 1 1,2 1
1,8 1 1 1 1 1
1,2
23.996* 24.072* 24.169* 24.195*
2 2 2 2
24.145* 24.075*
3 3
23.943* 24.002 24.166 24.333
4 4 4 4
1,4 1,4 1, 3, 4d 1, 4d
23.983 24.053*
? (II)y ? (II)
1,3 1
1, 2, 8 1, 2 1, 2
Age
(mo)
SymptomsC
15
S
15
11
1 12 6 6 10 11 3 7
19
10
SS A S S A S S A
S S A
23
S S A S
10 22
A S
24 15 26 10
S S A S
4 7
S
20 30
18
S
a *, IgM detected in either the acute- or convalescent b Seroconversion is defined as a fourfold increase in tlhe titer of NtAb. None of the children had any detectable NtAb in the preimmun e or preinfection serum sample. A, asymptomatic; S, symptomatic rotavirus infection d Twofold increase in NtAb titer but no seroconversio for strains whose The subgroup specificity is indicated in parenthe serotype could not be determined.
nses
91
rotavirus strains, one seroconverted to serotype Gl, and the other one seroconverted to serotypes Gi and G3. It is remarkable that all but the two serotype G3 viruses induced a neutralizing response to rotavirus strain Wa (serotype GI). The heterotypic NtAb response induced by serotype G2 and
G4 viruses to the serotype Gi virus Wa was strong, since it was equal to the homotypic response when the infecting virus was serotype G2 and even two- to fourfold higher than the homotypic response in serotype G4 infections. NtAb response in secondary infections. Of the 25 children analyzed in this study, 8 had more than one rotavirus infection: 4 had two and 4 had three infections. Five of the eight reinfected children were initially infected by a serotype Gl virus, and four of these five children were reinfected by a serotype G2 virus. As observed in the primary infections, the four secondary infections with serotype G2 strains induced homotypic antibodies (except in patient 24.059; see Table 3) and increased the titer of already existing NtAb to the serotype GI strain Wa. The one secondary infection caused by a serotype G3 virus induced homotypic antibodies as well as a heterotypic response to rotavirus strain Wa. The other three secondary infections were caused by subgroup II viruses that could not be typed; one of these infections induced NtAb to the serotype G8 virus 69M. None of the five serotyped
rotavirus strains isolated from secondary infections was of serotype Gl. Of the four strains isolated from children who had already experienced two rotavirus infections, two were non-group A rotaviruses, one was serotype G2, and one was serotype G3. The infection by the two non-group A rotaviruses did not modify the preexisting levels of NtAb to the group A viruses tested, as expected, since these two groups of rotaviruses are not antigenically related (30). The tertiary infection by the serotype G2 virus induced homotypic antibodies and increased the level of preexisting antibodies (already present after the primary infection) to serotypes Gi and G3 (patient 23.983). The serotype G3 tertiary infection induced homotypic antibodies and increased the level of preexisting NtAb to serotype Gi
(patient 24.059). belonging to G serotypes 1 (Wa), 2 (S2), 3 (P), 4 (ST3), 8 (69M), and 9 (WI61). The VP4-based P serotype has been reported to be 1A for rotavirus strains Wa, P, and WI61; PIB for strain S2; and P2 for strain ST3 (18). Rotavirus strain 69M may represent a new human P serotype (35). Since the P serotype of the infecting strains was not determined, whenever we refer to a heterotypic response, we are doing so only with regard to VP7, but it should be kept in mind that the response could be not heterotypic with regard to VP4. All children infected with serotype GI strains seroconverted to the serotype Gi virus Wa, and two seroconverted to either serotype G2 or G8 in addition to serotype Gl (Table 2). All children infected with serotype G2 viruses seroconverted to both serotypes Gi and G2, and one of these children also seroconverted to serotype G8. Neither of the two serotype G3-infected children had detectable neutralizing activity to any of the six strains tested. It is interesting that these two children had no detectable IgG in the postinfection sera, and one of these two infections was asymptomatic; however, in both cases, IgM was detectable in the acute-phase serum. The four serotype G4-infected children seroconverted to serotype Gl, and all four also had an increase in the level of NtAb to the serotype G4 strain ST3, but only two of them seroconverted; one serotype G4-infected child also seroconverted to serotype G3. Finally, of the two children infected with the untypeable ruses
DISCUSSION We have analyzed the serotype specificity of the NtAb response of children with primary and secondary rotavirus infections. The serological analysis supports the idea that the first rotavirus detected in the children studied represented a primary infection; the children did not have preexisting classspecific or neutralizing rotavirus antibodies in the preinfection serum samples, and 9 of 10 of the available convalescent-phase sera had detectable IgM, which has been shown to be a specific marker for primary rotavirus infections (17, 19). The child that did not have detectable IgM in the convalescent-phase serum instead had IgM on the acute-phase serum. In the first studies that characterized the NtAb serological response of children infected with rotavirus, there were conflicting results regarding the ability of a primary rotavirus infection to induce a heterotypic antibody response in humans. Some observed a homotypic response (40), while others found it to be heterotypic (10, 11, 15, 34); however, in most of these studies, the nature of the infection (primary or secondary) was not defined or the serotype of the infecting virus was not determined, complicating the interpretation of the heterotypic responses observed and the understanding of the relationship between such responses and the serotype of the infecting virus. Heterotypic antibody responses induced by single rotavirus infections have also been suggested from seroepidemiological
92
CLIN. DIAC,N. LAB. IMMUNC)I.
ARIAS ET AL.
TABLE 3. NtAb immune response in children with secondary rotavirus infections Patient
no.
Infecting virus G viuG serotype'
Age (mo)
or postinfection sera to indicated NtAb titer in acute-phase or preinfection/convalescent-phase rotavirus strain (G scrotype) Wa
(1)"
--"/400
S2 (2)
P (3)
ST3 (4)
69M (8)
Symptoms"
WI161 (9)
4 20 27
? (II) ? (II) 2
400/200
-1
