Familial Resemblance in Humoral Immune Response to Defined and Crude Schistosoma mansoni Antigens in an Endemic Area in Brazil

Transcription

1 1665 Familial Resemblance in Humoral Immune Response to Defined and Crude Schistosoma mansoni Antigens in an Endemic Area in Brazil J. Bethony, 1 A. M. S. Silveira, 3 L. F. Alves-Oliveira, 3 A. Thakur, 2 G. Gazzinelli, 4 R. Correa-Oliveira, 4 and P. T. LoVerde 2 Departments of 1 Social and Preventive Medicine and 2 Microbiology, State University of New York at Buffalo; 3 CENBIOS, Universidade Vale do Rio Doce, Governador Valadares, and 4 Centro de Pesquisas, Rene Rachou-FIOCRUZ, Belo Horizonte, Brazil This study addressed whether the humoral immune response to crude and defined Schistosoma mansoni antigens aggregates within families. The sample included 155 siblings from 42 nuclear families in Brazil. Sera examined by ELISA for antibody isotypes reactive to defined schistosome antigens and crude schistosome antigens (soluble adult worm antigen preparation and soluble egg antigen) demonstrated that there was a difference in sibling-pair correlations between defined and crude S. mansoni antigens. In contrast to the finding with crude antigens, egg-positive sibling pairs showed significant familial resemblance for all IgG subclasses and IgE to adult-stage antigens Smp20.8 and Smp50. Only the IgE and IgG4 isotypes showed familial resemblance to the egg-stage antigen, Smp40. Egg-negative sibling pairs showed significant familial resemblance only for IgE and IgG4 to Smp40. That both the IgE and IgG4 response to defined S. mansoni antigens showed familial resemblance is interesting in light of the converging evidence for the role of IgE and IgG4 in human susceptibility and resistance to reinfection. Schistosomiasis affects 1200 million people throughout the world and remains a major public health problem in many developing countries [1]. Host immune responses are critical to strategies for the control of both infection and pathology [2]. Parasite egg production, granuloma formation, disease severity, resistance to reinfection, and drug efficacy have been linked to immunologic factors. Antibody-dependent mechanisms may play a major role in the expression of acquired resistance to schistosome infection [2] and appear to be an important factor in killing parasites in rat, primate, and human schistosomiasis [2 5]. Antibody-dependent, cell-mediated cytotoxic mechanisms involving proinflammatory cell populations (macrophages, eosinophils, and platelets) as cellular partners and antibody isotypes, such as IgE or subclasses of IgG with anaphylactic properties, have been described [2, 4, 5]. Several immunoepidemiologic studies have provided evidence for the protective role of the IgE antibody in human infection [3, 6 9]. Several studies have provided evidence for at least two types of genetic control over IgE responses in humans, namely, major Received 9 February 1999; revised 16 June 1999; electronically published 8 October Informed consent was obtained from all subjects and their parents. The human subject studies were approved by the Ministry of Health of Brazil and the Institutional Review Board of the State University of New York at Buffalo. Grant support: NIH (International Center for Infectious DiseaseResearch grant AI-26505) and FAPEMIG (1168/95). Reprints or correspondence: Dr. Philip T. LoVerde, Dept. of Microbiology, 138 Farber Hall, School of Medicine and Biomedical Sciences, State University of New York, Buffalo, NY (loverde@buffalo.edu). The Journal of Infectious Diseases 1999;180: by the Infectious Diseases Society of America. All rights reserved /1999/ $02.00 histocompatibility complex (MHC) linked control of specific immune responses (cognate) and non MHC-linked control of the overall production of IgE (noncognate) [10 17]. While no studies have provided evidence for MHC-linked control of antibody responses in human schistosomiasis, several studies have found a higher relative risk of hepatosplenomegaly associated with certain HLA alleles. In Egyptian children, HLA-A1 and B4 were associated with a relative risk of 55.6 for hepatosplenomegaly when both alleles were present [18]. Secor et al. [19] showed that while there is no association between HLA class 2 alleles and egg excretion in a Brazilian population, an association was found between HLA-DQB and hepatosplenic disease. A major locus outside the HLA region has been identified, and this locus may play a role in controlling the humoral response to schistosomiasis. Abel et al. [20, 21] showed a low susceptibility phenotype and a major gene (SM1) controlling the intensity of infection. Marquet et al. [22] localized SM1 to chromosome 5q31 q33, an area that contains various interleukin (IL) genes (e.g., IL-4, -5, and -13). Muller-Myhsok et al. [23] confirmed the finding of a locus on chromosome 5 in susceptibility to infection with Schistosoma mansoni in a Senegalese population. Of interest, Marsh et al. [14] suggested that IL-4 or a nearby gene on 5q31.1 may regulate IgE production in a noncognate fashion. The production of IgE by both cognate and noncognate pathways is reflected in specific and total IgE concentrations, respectively, which can readily be measured as a quantitative trait. In quantitative genetics, phenotypes are measured on a continuous scale, and the observed distribution in a sample is frequently unimodal. Because the phenotypes in quantitative

2 1666 Bethony et al. JID 1999;180 (November) Table 1. Descriptive statistics for each of the sex and generation groups in a study of familial resemblance in humoral immune response to defined and crude Schistosoma mansoni antigens. Variable Father (n = 50) Mother (n = 56) Parents (n = 106) Son (n = 74) Daughter (n = 81) Offspring (n = 155) Age (years) Egg positive (%) 35 (18) 30 (17) 32 (35) 48 (35) 48 (40) 47 (75) Eggs per gram NOTE. Unless otherwise indicated, values are mean SD. Values for eggs per gram of feces are mean SE (no. of individuals). Sample does not include parents or offspring when sibships have only 1 member. Significant differences were not found between groups after Bonferroni multiple comparison tests. genetics are continuous, summary statistics (means and variances) must be used to describe a particular sample, with the analytical focus on the correlation or covariance among relatives. In this regard, we have initiated a quantitative genetic analysis of the humoral immune response to defined S. mansoni antigens. As a first step in this study, we address the question of whether the antibody response to defined schistosome antigens aggregates within families. This is best assessed by correlations among sibling pairs (sibships) stratified by infection status into egg-negative, egg-positive, and mixed (egg-positive and egg-negative) sibships [24 26]. Materials and Methods Study design. A family study was done in Alto Santa Helena and São Jose do Goiabal, villages in the prefecture of Governador Valadares in the state of Minas Gerais, Brazil. A population-based genealogic sample of the village was administered. The requirements of the population-based genealogic sampling strategy included recruitment of all persons living in both villages, identification of genealogic links among all individuals living in the village, and placement of each person into one or more multigenerational pedigrees. These data were entered into the pedigree-based data management system PEDSYS [27]. All participants were surveyed as to previous antischistosomicidal treatment within the last 5 years through a 10-item questionnaire. Participating family members received a medical examination and ultrasonography; all infected patients were determined to have the intestinal form of schistosomiasis. All infected individuals were treated with oxamniquine according to the standard Brazilian dose of 15 mg/kg for adults and 20 mg/kg for children. Pregnant women were not eligible for participation; women who reported that they were pregnant were recontacted after their pregnancy. Collection and storage of plasma samples. Plasma samples were obtained from whole blood collected into heparinized tubes. Plasma and cells were separated by centrifugation at 800 g for 10 min; the resulting plasma supernatant was transferred to sterile 1- ml tubes and stored at 20 C. Parasitologic data. Infection intensities were measured by individual fecal egg counts, expressed in eggs per gram of feces, performed by the Kato-Katz [28] method on stool samples obtained on 3 different days. An individual s fecal egg count was the arithmetic mean of these three determinations. Preparation of recombinant and crude S. mansoni antigens. Sera were examined by ELISA for antibody isotypes reactive to the defined schistosome antigens Smp20.8, Smp40, and Smp50 and to the crude antigens, soluble worm antigen preparation (SWAP) and soluble egg antigen (SEA). SWAP was prepared by homogenization of adult worms, and SEA was prepared by standard methods from eggs obtained from livers of hamsters that had been infected for 8 weeks [29 31]. We prepared 60 mg of each at a time to remove any variables that might occur in different preparations. The supernatant was dialyzed against PBS overnight, sterile-filtered, and stored in aliquots at 70 C. The concentration of each protein was determined by use of the BCA kit (Pierce, Rockford, IL). The following defined antigens were selected on the basis of previous studies in humans [7] and experimental animal models [32, 33]: Smp20, Smp40, and Smp50. Smp20.8 is a 20.8-kDa molecule that belongs to a family of tegument-associated proteins [34] that includes Smp22.6 [7] and Smp21.7 [35]. The molecule is expressed in all stages except the egg stage. Smp40 is a 40-kDa molecule that is a major component of SEA [32]. It displays homology to small heat shock proteins and alpha crystallins. It is expressed throughout development. Smp50 is a 50-kDa molecule that shows homology and biologic activity with immunophilins [36]. The antigen is expressed throughout development. Each of the recombinant antigens were expressed in pgex expression vectors as Schistosome japonicum glutathione-s-transferase (GST) fusion proteins, purified over a glutathione affinity column to obtain the fusion protein, cleaved with thrombin to remove the GST, and repurified over a glutathione column to obtain the purified recombinant protein [37]. Purity was confirmed by visualization of each step of the purification on SDS polyacrylamide gels. As a control against reactivity with the GST fusion partner, human sera were screened against GST. All human sera were negative for reactivity against the S. japonicum GST. ELISA. An ELISA was used to study isotype responses of participants to defined and crude schistosome antigens. Flexible assay plates (Falcon Microtest III; Becton Dickinson, Fullerton, CA) were coated with 0.5 mg/well defined or crude schistosome antigen in carbonate bicarbonate buffer and stored overnight at 4 C. Plates were washed 5 times with PBS and then blocked for 1 h with PBS containing 1% fetal calf serum. Plates were washed 5 times with PBS containing 0.05% Tween 20 (PBST; Sigma, St Louis). Serum samples were diluted 1 : 100 in PBST, and 100 ml/ well was added in duplicate to a plate. Plates were incubated overnight at 4 C and then washed 5 times with PBST as before. One hundred microliters of the following dilutions of alkaline phosphatase conjugated anti-human antibodies (Zymed, San Francisco) were added to each well: 1 : 5000 of IgM; 1 : 1000 of IgG1, IgG2, IgG3, and IgG4; and 1 : 2500 of IgE. The plates were incubated for 1 h at 37 C and then washed 10 times with PBS, and 100 ml of 1 M diethanolamine buffer containing 1 mg of p-nitro-

3 JID 1999;180 (November) Familial Responses to S. mansoni Antigens 1667 Table 2. Descriptive statistics for sibling pairs (sibships) stratified by infection status in a study of familial resemblance in humoral immune response to defined and crude Schistosoma mansoni antigens. Infection status No. of males No. of females Total No. of families Average sibship size No. of pairs Age range (years) Egg-negative sibships 27 (0.46) 32 (0.54) Egg-positive sibships 30 (0.43) 39 (0.57) Mixed sibships 68 (0.48) 73 (0.52) Total 74 (0.48) 81 (0.52) NOTE. Nos. in parentheses indicate proportion of total for each stratum. Does not include parents or sibships with only 1 member. phenylphosphate (Kirkegaard & Perry, Gaithersburg, MD) was added to each well. After developing the plates for 1 h in the dark, we measured the optical density (OD), using an automated ELISA reader (Bio-Rad Laboratories, Richmond, CA) at 405 nm. Data were saved into the Microplate manager system (Bio-Rad) and transferred to SPSS 8.0 (SPSS, Chicago) for statistical analysis. Standardization of assay conditions. Two negative controls were added to each plate. The first consisted of the combined sera of 6 uninfected patients from the United States. The second consisted of the combined sera of 3 egg-negative patients without a history S. mansoni infection from Belo Horizonte, Minas Gerais, Brazil. Negative control sera were diluted 1 : 100 in PBST, and 100 ml were added to each plate in duplicate. Two positive controls were added to each plate: the first consisted of the combined sera of 27 egg-positive individuals from a nearby endemic area (Corrego do Bernardo, Governador Valadares Prefecture), and the second consisted of the serum from 1 high responder to a specific antigen from the egg-positive patients from the Corrego do Bernardo sample. Positive control sera were diluted 1 : 100 in PBST, and 100 ml was added to each plate in duplicate. Assay runs were repeated if negative or positive control sera had a coefficient of variation110% between plates. Duplicates on a single plate were rejected if they differed by a coefficient of variation of 110% from each other. Assays for a specific isotype were conducted during one run, by use of the same stock of PBS, PBST, PBS with 1% fetal calf serum, and alkaline phosphatase developer to ensure standard conditions across plates. Each plate was tested for correlations in OD reading by well position. Assay plates in a single run were also tested for correlations in OD reading by the serial position of the plate during reading. Student s t test, analysis of variance (ANOVA), and Bonferroni post hoc tests. Student s t test was used to determine whether there were significant differences for prevalence and intensity of infection between individuals who were included and those who were excluded (because of missing demographic or serologic data) from the study. In addition, Student s t test was used to determine whether there were significant differences between sexes for infection status and infection intensity. A standard one-way ANOVA was used to determine whether there were significant differences among familial positions (e.g., fathers, mothers, sons, and daughters) for infection status and infection intensity. A Bonferroni multiple comparison tests was then used to determine whether there were significant differences in these measures between each family position. Data adjustment. To reduce heteroscedasticity, we performed all analyses on the log-transformed OD values of the isotype responses to crude and recombinant S. mansoni antigens. We then adjusted each measure for the effects of age, using a stepwise regression procedure, with isotype response as the dependent variable (y) and age as the independent variable (x); this was done separately for each sex by generation group (e.g., fathers, mothers, sons, and daughters). The standardized residual or the interval from each individual data point to the regression line was used in the subsequent analysis of familial aggregation and was defined as the age-adjusted and standardized residual score from regression analysis. Intraclass correlation coefficients. To assess whether the immunologic phenotypes (antibody response to schistosome antigens) may be influenced by genetic factors, we measured the familial resemblance of the trait by estimating intraclass correlation coefficients. Details of the methods for measuring intraclass correlations among siblings are found in [25, 26]. In brief, for the estimation of intraclass correlations (R cc ), individuals were first classified into families. Families with sibships having only 1 member were excluded from the analysis. The phenotype, defined above as the age-adjusted and standardized residual score from regression analysis, was then used to compute a standard one-way ANOVA. In the ANOVA, the phenotypes were grouped by family to provide estimates for the mean squared within family variance (MSW) and the mean squared among family variance (MSA). The ratio of these variance components formed the measure of sibling resemblance: if the siblings were similar with respect to the specified phenotype, the MSW was!msa. The intraclass correlation coefficient was computed from the ANOVA, by use of the methods of Donner [25, 26], where F refers to the ratio MSB/MSW and is the significance testing procedure for the ANOVA, N is the number of families studied, K is the total number of observations, k i is the number of siblings in the ith family, and ( ) N i k = K k. N 1 k i=1 The intraclass correlation R cc is defined [25, 26] as Rcc = S A/ (SA S W) = (MSA MSW)/ (MSA MSW k MSW) 0 = (F 1)/ (k F 1). 0

4 1668 Bethony et al. JID 1999;180 (November) Figure 1. Prevalence and intensity of Schistosoma mansoni infection in study sample; epg, eggs per gram of feces. Ninety-five percent confidence intervals (95% CIs) were given by [ ] (1 L ), (1 R ), where L = (F/FN 1,K N;a 1) /k 0 and R = FFK N,N 1;a/2 1/k 0. Interclass correlation coefficients. A second measure of familial resemblance of an immunologic trait was the degree of resemblance between parents and their offspring or between spouses (interclass correlation coefficients). We estimated the parent-offspring correlations using the pairwise method [25, 26]. For the parent-offspring correlations (R po ), the mean of the offsprings scores was paired with the score of their fathers or mothers, and a Pearson product moment correlation was computed over the resulting pairs. For the midparent-offspring scores (R mp ), the mean of the parents value (midparent value) was paired with the mean of the offsprings value, and a Pearson product-moment correlation was computed over the resulting pairs. Denoting the sibling s scores by y ij,i = 1,2,...N; J = 1,2,...,k and the parent s score by x i,i = 1,2,...,N, the pairwise estimator (20) of a parent-offspring correlation is given by Results N ki ( ) ( i=1 xi x j=1 yij y) r p =. N 2 N ki 2 [ ( ) ][ ( i=1 xi x i=1 j=1 yij y) ] Study sample. The total study sample consisted of 299 individuals in 90 nuclear families. The analysis of each phenotype was restricted to the 264 individuals (88%) for whom complete demographic, genealogic, parasitologic, and serologic data were available. The differences in age, sex, prevalence, and intensity of infection between included and excluded individuals was not significant (data not shown). The study sample was further restricted to individuals whose sibships had 11 member (multiplex sibships). Table 1 shows the sample sizes and infection status by sex and familial position (father, mother, parents, son, daughter, and offspring) for families of multiplex sibships. Significant differences were not found among the familial positions for infection status or infection intensity by use of Bonferonni multiple comparison tests. Table 2 shows the sample sizes for siblings by infection status (egg-negative, egg-positive, and mixed sibships). The egg-negative sibship category contained 59 individuals, with 27 males (46%) and 32 females (54%) from 20 families. The average size of the sibships was 3 individuals (range, 2 4), and there were 132 sibling pairs. Egg-positive sibships contained only siblings with current infection. The egg-positive sibship category contained 69 individuals, with 30 males (43%) and 39 females (57%) from 31 families. The average size of the sibships was 2 individuals (range, 2 5), and there were 104 sibling pairs. The term mixed sibship refers to sibships with both egg-positive and egg-negative siblings. The mixed sibship category contained 141 individuals, with 68 males (48%) and 73 females (52%) from 43 families. The average size of the sibships was 3 individuals (range, 2 7), and there were 396 sibling pairs. Figure 1 shows the distribution of the prevalence and intensity of S. mansoni infection (as determined by geometric mean) in the study population. Infection with S. mansoni was defined as having 1 egg per gram of feces. The highest prevalence and intensity of infection were found within the second and third decades. There was no statistically significant difference in the prevalence and intensity of infection by sex (data not shown). None of the study participants had received schistosomicidal drugs within the last 5 years, and they most likely had never been treated. Strong familial resemblance for the IgE and IgG4 response to defined S. mansoni antigens. Figure 2 presents the 95% CIs for sibling-pair correlations among egg-negative (A), egg-positive (B), and mixed sibships (C) for the antibody response to defined S. mansoni antigens. The upper and lower boundaries of the 95% CIs are represented as bars. Among egg-negative sibships (figure 2A), positive correlations were found for the IgG1 response to Smp20.8 (0.33), the IgG4 (0.85) and IgE (0.47) responses to Smp40, and the IgG4 response (0.33) to Smp50. For egg-positive sibships (figure 2B), positive correlations were found for the IgG1 (0.79), IgG2 (0.47), IgG4 (0.86), and IgE (0.54) responses to Smp20.8; the IgG4 (0.84) and IgE (0.85) response to Smp40; and the IgG1 (0.68), IgG2 (0.90), IgG4 (0.79), and IgE (0.91) responses to Smp50. Among mixed sibships (figure 2C), positive correlations were demonstrated for the IgG1 (0.55) and IgG4 (0.40) responses to Smp20.8, the IgG4 (0.82) and IgE (0.54) response to Smp40, and the IgG2 (0.37) and IgG4 (0.31) responses to Smp50. Lack of familial resemblance for the antibody isotype response

5 JID 1999;180 (November) Familial Responses to S. mansoni Antigens 1669 Figure 2. Intraclass correlation coefficients for antibody isotype response to defined S. mansoni antigens stratified by infection status. Bars represent 95% confidence intervals of intraclass correlation coefficients (R cc ) for antibody isotype response to recombinant antigens Smp20.8, Smp40, and Smp50 in egg-negative siblings from 20 families (A), egg-positive siblings from 31 families (B), and mixed siblings (egg negative and egg positive) from 43 families (C). to crude S. mansoni antigens. No positive correlations were found for the antibody isotype responses against the crude antigen extract SWAP (figure 3). Among egg-positive sibling pairs, there were two positive correlations to the crude antigen extract SEA: IgG4 (0.68) and IgE (0.71) No apparent pattern to the parent-offspring and spouse-spouse correlations in antibody isotype response to defined or crude S. mansoni antigens. Table 3 shows the father-offspring, mother-offspring, and spouse-spouse correlations. Statistically significant ( P!.05) Pearson product moment correlation coefficients are presented below. There was a positive correlation for father-offspring pairs for the IgG2 response to Smp40 (0.49). Several positive correlations were found for the motheroffspring pairs: the IgG4 (0.44) response to Smp50 and the IgM (0.42), IgG3 (0.36), and IgG4 (0.28) responses to SWAP. Two positive correlations were found for spouse-spouse pairs: the IgM (0.45) and IgG2 (0.41) responses to Smp40. Few correlations were found for midparent-offspring pairs. Table 4 shows the midparent-offspring correlations stratified by the infection status of the offspring. Statistically significant ( P!.05) coefficients are presented below. The only consistent midparent-offspring correlation was found for the IgG4 response to SWAP, which included egg-positive (0.39), egg-negative (0.41), and mixed offspring (0.42) sibships. No other consistent patterns were observed. Discussion This study is novel in three respects: the level of antibody response is treated as a quantitative trait, we estimated the degree of resemblance among family members with respect to antibody isotype level, and the antibody isotype level was measured to defined and crude S. mansoni antigens. The basic idea is that the resemblance in antibody isotype level among family members reflects the degree to which this phenotype may be influenced by genetic factors. The principal measure of familial resemblance was the positive correlation among siblings in nuclear families or the intraclass resemblance.

6 1670 Bethony et al. JID 1999;180 (November) Figure 3. Intraclass correlation coefficients for isotype response to crude antigen extracts of S. mansoni stratified by infection status. Bars represent 95% confidence intervals of intraclass correlation coefficients (R cc ) for antibody isotype response to crude antigen extracts (soluble egg antigen [SEA] and soluble worm antigen preparation [SWAP]) in egg-negative siblings from 20 families (A), egg-positive siblings from 31 families (B), and mixed siblings (egg negative and egg positive) from 43 families (C). The simplest analytical approach for the estimation of intraclass resemblance is the ANOVA on the age-adjusted and standardized residual scores of offspring, with the F ratio transformed into the intraclass correlation coefficient by the method of Donner [25, 26]. A positive intraclass correlation coefficient indicates that offspring of the same family tend to be more alike with respect to antibody isotype level than are offspring from different families. A second measure of familial resemblance is the degree of resemblance between parents and offspring, referred to as the interclass resemblance. The pairwise method, as used in this study, has been used by a number of investigators to estimate the interclass resemblance, despite the fact that the pairs of observations are not independent [25, 26]. The lack of independence arises as the score of each parent is paired with the score of each offspring, with the result that a parent appears as many times as he or she has children. The third measure of familial resemblance is the degree of resemblance between spouses residing in the same household or spouse-spouse resemblance. The spouse-spouse resemblance was estimated in the same manner as the interclass resemblance, with the age-adjusted and standardized residual score of the spouses paired and the Pearson product-moment correlation computed over each pair. The measure of spouse-spouse resemblance is often used to estimate the influence of shared environment on a phenotype. The fourth measure is the midparent-offspring level in which the mean of the parent s ageadjusted and standardized residual score is paired with the score of their offspring, and a Pearson product-moment correlation is computed over the pairs. Several interesting patterns emerged from the present study. The most pronounced pattern was the difference in sibling correlations between defined and crude S. mansoni antigens. For example, egg-positive sibships showed correlations for the defined antigens, but, with only one exception, no correlations were observed for the crude antigen extracts, SWAP and SEA (figure 3). The difference may be related to the fact that crude antigen extracts of the parasite contain a large population of proteins, which present numerous and diverse epitopes to the immune system. In comparison, the defined antigens, which consist of a single polypeptide antigen, present the immune system with a more limited set of epitopes. As such, one might anticipate that immune responsiveness to a crude S. mansoni antigen preparation would be more complex and involve interactions of antigen peptides with several class II molecules. Conversely, responsiveness to defined antigens would involve fewer major T cell epitopes, with fewer class II molecules. In studies of highly polymorphic human populations, it has been suggested that more complex protein mixtures may not yield informative data [12, 13]. Thus, in the analysis of the genetic influence on immune responsiveness, the defined antigens used herein may prove to be more informative than are the crude preparations, indicating a novel use for defined S. mansoni antigens. However, it is possible that the failure to detect positive correlations for the isotype responses to crude antigens results from the large number of families that are re-

7 JID 1999;180 (November) Familial Responses to S. mansoni Antigens 1671 Table 3. Interclass correlation coefficients between siblings with mixed infection status and their parents for the isotype response to defined and crude Schistosoma mansoni antigens. Isotype Interclass correlations Father Mother Spouse Smp20.8 IgM IgG IgG IgG IgG IgE Smp40 IgM a IgG IgG2.49 a b IgG IgG IgE Smp50 IgM IgG IgG IgG IgG b.10 IgE SEA IgM b.03 IgG IgG IgG IgG IgE SWAP IgM b.18 IgG IgG a IgG a.18 IgG IgE NOTE. Columns for father and mother represent the Pearson product moment correlation between mean of offspring s standardized residuals and standardized residuals of the parent. Column for spouse represents the Pearson product moment correlation for standardized residuals between spouses. Both parents and siblings are of mixed infection status. SEA, soluble worm antigen; SWAP, soluble worm antigen preparation. a P!.05. b P!.01. quired to provide reasonable levels of power for detecting low correlations (see [24] for power analysis of intraclass correlation). Another important observation is that the intraclass resemblance for defined S. mansoni antigens was not a matter of the presence or absence of antibody isotype. Family members tended to share similar age-adjusted and standardized isotype levels across a range of values. This points to the possible role of factors encoded by genes outside the HLA class II region in the regulation of human immune response against S. mansoni [22, 23, 38]. Such loci may control B cell function, antigen processing, and expression of cytokines, such as IL-4 and IL- 5 [39 44]. There is considerable evidence to indicate that production of IgE, which consistently showed high intraclass resemblance to defined antigens, is regulated by such genetic factors. The IgE control locus has been mapped to chromosome 5q31 q33 [14]. This locus has a cluster of several cytokine genes involved in immunoglobulin heavy chain switching to the epsilon isotype [14]. Most family studies favor the possibility of a major gene controlling total serum IgE level, but the mode of inheritance is still a matter of controversy [14 16]. Other determinants, such as sex, age, race, occupational exposures, and exposure to environmental antigens, such as tobacco smoke, also appear to influence IgE levels. The observation that both IgE and IgG4 produce similar correlations among siblings to defined S. mansoni antigens is noteworthy in light of evidence supporting the role of these isotypes in human resistance to schistosomiasis. Hagan et al. Table 4. Interclass correlation coefficients between the mean of the siblings isotype response and the mean of the parents isotype response (midparent) for egg-positive, egg-negative, and mixed infection status. Isotype Correlation coefficient by infection status of offspring Egg positive Egg negative Mixed Smp20.8 IgM IgG a.09 IgG IgG IgG IgE Smp40 IgM IgG a.11 IgG2.55 a IgG IgG IgE Smp50 IgM IgG a.26 IgG IgG IgG a.26 IgE SEA IgM IgG IgG IgG IgG4.39 a IgE SWAP IgM IgG IgG IgG a.32 IgG4.39 a b IgE NOTE. Pearson product moment correlation was computed over the mean of standardized residuals for siblings and those for parents (midparent). Eggnegative, egg-positive, and mixed refer to the infection status of the offspring. SEA, soluble worm antigen; SWAP, soluble worm antigen preparation. a P!.05. b P!.01.

8 1672 Bethony et al. JID 1999;180 (November) [8] demonstrated a significant correlation between the production of anti-schistosome IgE antibodies and the acquisition of immunity to schistosomiasis. They also showed that anti-schistosome IgG4 antibodies, which potentially block IgE effector pathways, were correlated with susceptibility to reinfection, supporting the idea that they might be responsible for delaying the development of protective immunity. These findings have been confirmed by subsequent field studies [6, 8, 9, 45]. Our data add a new dimension to previous studies on the role of IgE and IgG4 in human resistance to schistosomes by observing the familial aggregation of these isotype responses to defined antigens. Two limitations of the study need to be mentioned. First, the indirect ELISA measures antibody binding not antibody quantity [46], with antibody binding being a function of antibody affinity and antibody concentration. Therefore, whereas it is possible to treat the data from the indirect ELISA as continuous variables, we emphasize that it is much more problematic to conclude that family members share a similar amount of antibody isotype to a given antigen. Second, the strength of the sibling correlations to defined S. mansoni antigens may reflect shared household and shared genetic effects. While the low correlation among spouses (unrelated) argues against the influence of shared environmental effect, household membership and sibship overlap as members of nuclear families tend to occupy the same residence. An optimal study design would consider large extended pedigrees crossing multiple households [47]. Future studies following this approach are planned. As a prelude to identifying and localizing specific genes influencing infection and disease in schistosomiasis, we first had to establish that genetic factors were indeed involved. An analysis of familial aggregation provides the means for this first step by allowing the quantification of the degree of resemblance in isotype response among siblings in nuclear families. The identification of host genes regulating schistosome host-parasite interactions will provide new ways to understand the immune mechanisms in humans that are critical for determining susceptibility and resistance to reinfection and disease development. Acknowledgments We thank the residents of Alto Santa Helena and São Jose do Goiabal for participating in this study; Marcia Cordeiro, Marluce Rodrigues Lima, Maria Fatima da Silva, and Lillian Cardoso Moreira for assistance throughout the study; and Ann Frank (American Airlines) for facilitating transport of supplies from Buffalo, New York, to Belo Horizonte, Brazil. References 1. World Health Organization. WHO Fact Sheet. May Capron A. Schistosomiasis forty years war on the worm. Parasitol Today 1998;14: Dunne DW, Butterworth AE, Fulford AJ, et al. Immunity after treatment of human schistosomiasis: association between IgE antibodies to adult worm antigens and resistance to reinfection. Eur J Immunol 1992;22: Capron A, Dessaint JP. Immunologic aspects of schistosomiasis. Annu Rev Med 1992;43: Capron AR. Immunity to schistosomes. Curr Opin Immunol 1992;4: Dunne DW, Butterworth AE, Fulford AJ, Ouma JH, Sturrock RF. Human IgE responses to Schistosoma mansoni and resistance to reinfection. Mem Inst Oswaldo Cruz 1992;87: Dunne DW, Webster M, Smith P, et al. The isolation of a 22 kda band after SDS-PAGE of Schistosoma mansoni adult worms and its use to demonstrate that IgE responses against the antigen(s) it contains are associated with human resistance to reinfection. Parasite Immunol 1997;19: Hagan P, Blumenthal UJ, Dunn D, Simpson AJ, Wilkins HA. Human IgE, IgG4 and resistance to reinfection with Schistosoma haematobium. Nature 1991;349: Rihet P, Demeure CE, Bourgois A, Prata A, Dessein AJ. Evidence for an association between human resistance to Schistosoma mansoni and high anti-larval IgE levels. Eur J Immunol 1991;21: Marsh DG, Zwollo P, Huang SK, Ansari AA. Molecular genetics of human immune responsiveness to allergens. Ciba Found Symp 1989;147: Marsh DG, Zwollo P, Huang SK, Ghosh B, Ansari AA. Molecular studies of human response to allergens. Cold Spring Harb Symp Quant Biol 1989;54: Marsh DG, Zwollo P, Huang SK. Molecular and cellular studies of human immune responsiveness to the short ragweed allergen, Amb a V. Eur Respir J Suppl 1991;13:60s 7s. 13. Marsh DG, Huang SK. Molecular genetics of human immune responsiveness to pollen allergens. Clin Exp Allergy 1991;21(Suppl 1): Marsh DG, Neely JD, Breazeale DR, et al. Linkage analysis of IL4 and other chromosome 5q31.1 markers and total serum immunoglobulin E concentrations. Science 1994;264: Marsh DG, Neely JD, Breazeale DR, et al. Total serum IgE levels and chromosome 5q. Clin Exp Allergy 1995;25(Suppl 2):79 83; Marsh DG, Neely JD, Breazeale DR, et al. Genetic basis of IgE responsiveness: relevance to the atopic diseases. Int Arch Allergy Immunol 1995;107: Marsh DG. Mapping the genes for IgE production and allergy. Adv Exp Med Biol 1996;409: Abdel-Salam E, Abdel Khalik A, Abdel-Meguid A, Barakat W, Mahmoud AA. Association of HLA class I antigens (A1, B5, B8 and CW2) with disease manifestations and infection in human schistosomiasis mansoni in Egypt. Tissue Antigens 1986;27: Secor WE, del Corral H, dos Reis MG, et al. Association of hepatosplenic schistosomiasis with HLA-DQB1*0201. J Infect Dis 1996;174: Abel L, Demenais F, Prata A, Souza AE, Dessein A. Evidence for the segregation of a major gene in human susceptibility/resistance to infection by Schistosoma mansoni. Am J Hum Genet 1991;48: Abel L, Dessein A. Genetic predisposition to high infections in an endemic area of Schistosoma mansoni. Rev Soc Bras Med Trop 1991;24: Marquet S, Abel L, Hillaire D, et al. Genetic localization of a locus controlling the intensity of infection by Schistosoma mansoni on chromosome 5q31- q33. Nat Genet 1996;14: Muller-Myhsok B, Stelma FF, Guisse-Sow F, et al. Further evidence suggesting the presence of a locus, on human chromosome 5q31-q33, influencing the intensity of infection with Schistosoma mansoni. Am J Hum Genet 1997;61: Donner A. The number of families required for detecting the familial aggregation of a continuous attribute. Am J Epidemiol 1978;108: Donner A. The use of correlation and regression in the analysis of family resemblance. Am J Epidemiol 1979;110: Donner A, Eliasziw M. Methodology for inferences concerning familial correlations: a review. J Clin Epidemiol 1991;44:

9 JID 1999;180 (November) Familial Responses to S. mansoni Antigens Dyke B. PEDSYS. PGL technical report no. 2. San Antonio: Southwest Foundation for Biomedical Research, Katz N, Chaves A, Pellegrino J. A simple device for quantitative stool thicksmear technique in schistosomiasis mansoni. Rev Inst Med Trop São Paulo 1972;14: Carter CE, Colley DG. Partial purification and characterization of Schistosoma mansoni soluble egg antigen with Con A sepharose chromatography. J Immunol 1979;122: Boros DL, Warren KS. Delayed hypersensitivity type granuloma formation and dermal reaction induced and elicited by a soluble factor isolated from Schistosoma mansoni eggs. J Exp Med 1970;132: Gazzinelli G, Katz N, Rocha RS, Colley DG. Immune responses during human schistosomiasis mansoni. VIII. Differential in vitro cellular responsiveness to adult worm and schistosomular tegumental preparations. Am J Trop Med Hyg 1983;32: Stadecker MJ, Hernandez HJ. The immune response and immunopathology in infection with Schistosoma mansoni: a key role of major egg antigen Sm-p40. Parasite Immunol 1998;20: Waine GJ, Mazzer DR, McManus DP. Production of IgE antibodies against the 22-kDa tegumental membrane associated antigen of schistosomes is directed by the antigen itself. Parasite Immunol 1997;19: Mohamed MM, Shalaby KA, LoVerde PT, Karim AM. Characterization of Sm20.8, a member of a family of schistosome tegumental antigens. Mol Biochem Parasitol 1998;96: Francis P, Bickle Q. Cloning of a 21.7-kDa vaccine-dominant antigen gene of Schistosoma mansoni reveals an EF hand like motif. Mol Biochem Parasitol 1992;50: Osman A, Kiang D, Lo Verde PT, Karim AM. Schistosoma mansoni: characterization of p50, an immunophilin. Exp Parasitol 1995;80: Smith DB, Johnson KS. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 1988; 67: Chiarella JM, Goldberg AC, Abel L, Carvalho EM, Kalil J, Dessein A. Absence of linkage between MHC and a gene involved in susceptibility to human schistosomiasis. Braz J Med Biol Res 1998;31: Troye-Blomberg M, Riley EM, Kabilan L, et al. Production by activated human T cells of interleukin 4 but not interferon-gamma is associated with elevated levels of serum antibodies to activating malaria antigens. Proc Natl Acad Sci USA 1990;87: Erard F, Wild MT, Garcia-Sanz JA, Le Gros G. Switch of CD8 T cells to noncytolytic CD8 CD4 cells that make Th2 cytokines and help B cells. Science 1993;260: Gollob KJ, Coffman RL. A minority subpopulation of CD4 T cells directs the development of naive CD4 T cells into IL-4 secreting cells. J Immunol 1994;152: Gollob KJ, Dutra WO, Coffman RL. Early message expression of interleukin-4 and interferon-gamma, but not of interleukin-2 and interleukin-10, reflects later polarization of primary CD4 T cell cultures. Eur J Immunol 1996;26: Mosmann TR, Coffman RL. Th1 and Th2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu Rev Immunol 1989;7: Mosmann TR, Coffman RL. Heterogeneity of cytokine secretion patterns and functions of helper T cells. Adv Immunol 1989;46: Demeure CE, Rihet P, Abel L, Ouattara M, Bourgois A, Dessein AJ. Resistance to Schistosoma mansoni in humans: influence of the IgE/IgG4 balance and IgG2 in immunity to reinfection after chemotherapy. J Infect Dis 1993;168: Vankatesan P, Wakelin D, ELISAs for parasitologists: or lies, damn lies, and ELISAs. Parasitol Today 1993;9: Williams-Blangero S, Blangero J, Bradley M. Quantitative genetic analysis of susceptibility to hookworm infection in a population from rural Zimbabwe. Hum Biol 1997;69:201 8.