点击显示 收起
1Laboratory for Health Effects Research, 2Research Laboratory for Infectious Diseases, and 3Computerization and Methodological Consultancy Unit, National Institute of Public Health and the Environment, Bilthoven, 4Wilhelmina Children's Hospital, University Medical Center Utrecht, Utrecht, and 5Department of Pediatrics, Erasmus MC/Sophia Children's Hospital, Rotterdam, The Netherlands
Received 14 June 2002; revised 6 September 2002; electronically published 13 December 2002.
The association of variants of genes encoding interleukin (IL)4 and the IL-4 receptor chain (IL-4R) with respiratory syncytial virus (RSV) bronchiolitis was examined in hospitalized infants. Polymorphisms in IL-4 (C-590T) and IL-4R (I50V and Q551R) were genotyped by restriction fragmentlength polymorphism analysis. Control subjects included parents of the hospitalized children (for the transmission/disequilibrium test), and a random population sample (for the case-control study). Results were also analyzed in a combination of these 2 tests, using Fisher's method. The IL-4 590T allele was found more frequently among children hospitalized with RSV than expected in the case-control (odds ratio [OR], 1.43; P = .04) and combination (OR, 1.41; P = .02) tests. Among children who were >6 months old when they were hospitalized, compared with the control group or with the <6 months old who were hospitalized for RSV infection, higher frequencies of both the IL-4 590T allele and the IL-4R R551 allele were found. These results indicate that gain-of-function variants of T helper type 2 cytokine genes may play a role in increasing the severity of RSV disease, which appears more pronounced after the first half-year of life.
Informed consent was obtained from the parents of each study subject in accordance with the guidelines of the medical ethical committees of both the Wilhelmina Children's Hospital and the Sophia Children's Hospital, and these committees approved the study. For the control population, informed consent was obtained in accordance with the guidelines of the medical ethical committee of Netherlands Organization for Applied Scientific Research (TNO) Leiden.
Reprints or correspondence: Dr. Barbara Hoebee, Laboratory for Health Effects Research, National Institute of Public Health and the Environment, PO Box 1, 3720 BA Bilthoven, The Netherlands .
Each year during the winter season, respiratory syncytial virus (RSV) causes severe respiratory infections in infants and young children. Almost all children are infected with RSV in their first or second year of life. Reinfection may occur throughout life, but the first infection often is the most severe and may result in bronchiolitis and pneumonia [1]. The severity of disease in these young children varies from no symptoms at all or a mild upper respiratory tract infection to a more or less severe lower respiratory tract infection (found in 40% of infected children). This may lead to respiratory failure, which is the reason for hospitalization in 0.5%2% of the infected children, and occasionally to death (in 1% of hospitalized children) [2, 3]. The reasons for this variation in disease severity are not completely clear. Some children (e.g., children born prematurely, children with congenital heart or lung disease, and children with immune deficiencies) are at higher risk of developing more-serious symptoms of the disease [4 7]. We hypothesized that genetic heterogeneity in the immune response of infected children explains some of these differences. To test our hypothesis, we investigated the role of immune response gene variants (polymorphisms) in an association study in which both the classic case-control design (comparing individuals with the severe form of RSV infection and a control group) and a transmission/disequilibrium test (TDT; comparing alleles that were transmitted from the parents to the affected child and alleles that were not transmitted).
Several authors have suggested that T cellmediated immunopathology, in particular, a more pronounced Th2 response, may play an important role in the development of severe RSV disease [8, 9]. This has been suggested because children vaccinated with a formalin-inactivated vaccine developed a more severe course of disease when natural infection occurred [10]. Later, it was shown in mice that this enhanced form of disease was accompanied by a shift in the immune response toward a Th2 response [11 13]. However, the significance of Th2 responses in natural RSV infections in humans remained unclear. Some studies demonstrated Th2 cytokine responses when natural infection occurred in children who went on to develop severe RSV disease [9, 14, 15], whereas, in other studies, such a shift toward a Th2 response was not found [16, 17]. It should be mentioned that the induction of a Th1 response does not necessarily exclude the presence of a concomitant Th2 response and its associated pathologic effects, because Th2 cytokines and Th1 cytokines can be detected simultaneously [18, 19]. In addition, Aberle et al. [20] and Bont et al. [21] showed that children with severe RSV disease had a lower level of expression of the Th1 cytokine interferon (IFN) in peripheral blood mononuclear cells than did children with mild forms of the disease, which suggests that Th1 cytokines may diminish the severity of RSV disease.
Another reason to study genetic predisposition to Th2 responses in RSV disease is the suggested relationship between the occurrence of severe RSV disease and the development of asthma later in life. In the Tucson cohort study, it was shown that patients who experienced RSV bronchiolitis had recurrent wheezing that lasted up to 11 years, but RSV bronchiolitis did not increase the risk of allergy sensitization [22]. In contrast, Sigurs et al. [23] presented evidence that RSV infections may increase allergy. Although RSV might affect the development of asthma in several ways, such findings could suggest that individuals with a genetic predisposition toward strong Th2 responses are at risk for development of more-severe RSV disease as well as asthma. RSV might affect the development of asthma by allergic processes or by nonallergic processes, such as bronchial hyperresponsiveness, wheezing, and bronchus obstruction [24, 25]. These findings prompted us to study the association between severe RSV disease and gain-of-function mutations in Th2 cytokine genes.
The Th2 cytokine interleukin (IL)4 is one of the cytokines that may play a role in the pathogenesis of asthma. The gene is located in the Th2 cytokine cluster on human chromosome 5q31 and encodes for IL-3, -4, -5, and -13, among others [26]. In genetic studies, this chromosomal area is linked with total IgE concentration, asthma, and bronchial hyperreactivity [27 29]. IL-4 plays an important role in the stimulation and differentiation of Th2 cells. It promotes the proliferation and differentiation of activated B cells, promotes the recruitment of circulating cells by up-regulation of vascular cell adhesion molecule 1 on endothelial cells, and suppresses IFN- production by Th1 cells. That last quality makes the study of IL-4 in the context of severe RSV disease particularly interesting, because results from studies in both human and mice indicate that IFN- protects against severe RSV disease [20, 21, 30]. IL-4 mediates its biological effects by binding to the IL-4 receptor, which is located on almost all hematopoietic cells and a large number of nonhematopoietic cells. After binding of IL-4 to its receptor, the kinases Jak-1 and Jak-3 become phosphorylated, and in turn phosphorylate the receptor, the insulin receptorlike substrates, and the transcription factor STAT6 [31, 32]. The IL-4 receptor consists of 2 chains: the IL-4 receptor chain (IL-4R) and the common cytokine receptor chain. The latter chain is present in a large number of cytokine receptors (e.g., IL-2, -7, -9, and -15) [33 36], whereas IL-4R is also a part of the IL-13 receptor [37]. The IL-4R gene is located on chromosome 16p21 [38]. We studied the genetic association of severe RSV disease with 1 polymorphism located in the promoter region of the IL-4 gene and 2 polymorphisms located in the IL-4R gene.
SUBJECTS, MATERIALS, AND METHODS
Study design. Children included in the study were hospitalized because of RSV bronchiolitis during the period 1992 2000. RSV infection was confirmed by direct immunofluorescent assay of nasopharyngeal cells. Blood samples or buccal swabs were collected from 200 children and from both of the parents of each child for DNA isolation. In 7 cases, samples were obtained from the child and the mother only. All parents completed a questionnaire that gathered medical data and information about pregnancy and ethnic origin. As far as possible, the medical data were verified against clinical records. A subset of the children (110 children) from the Wilhelmina Children's Hospital took part in a follow-up study that examined recurrent wheezing that occurred as a result of RSV infection [39, 40]. An unselected control population of 447 persons born in The Netherlands (37% of whom were women) was randomly taken from the REGENBOOG study, a large Dutch population health examination survey [41]. In this survey, a random sample of the Dutch population was interviewed, and 30% of those individuals participated in an additional health examination at a municipal health center. There were no major differences, with respect to many background and health-related variables, between participants interviewed at home and those who underwent physical examination.
The mean age of the children at the time of hospitalization was 115 days (SD, 111 days). Of the 207 children included in the study, 148 were native Dutch (their parents and grandparents were born in The Netherlands and consider themselves to be native Dutch); the mean age of this group at the time of hospitalization was 113 days (SD, 96 days). At the time of inclusion in the present study, the mean age for all children was 3.4 years (SD, 1.1 year).
DNA isolation. DNA was isolated from blood samples or, when blood was not available, from buccal swabs, using the QIAamp DNA Blood Mini or Midi Kit (Qiagen). DNA from blood samples from parents was obtained using an isolation kit for mammalian blood (Roche). For the control population (REGENBOOG samples), genomic DNA was extracted from buffy coats by digestion with proteinase K, followed by salting out with potassium acetate and chloroform/isoamyl alcohol extraction [42]. The DNA concentration was determined using PicoGreen (Molecular Probes).
Polymerase chain reaction (PCR) amplification and genotyping. All hospitalized children, parents, and adult control subjects were genotyped for polymorphisms in the IL-4 gene (C-590T) and the IL-4R gene (I50V and Q551R) by PCRrestriction fragmentlength polymorphism analysis. Primers and experimental conditions are listed in . All PCRs were performed on a GeneAmp PCR system 9700 (Applied Biosystems), with an initial denaturation step of 6 min at 95°C and then 35 cycles of the specific program. The program ended with an additional 10 min at 72°C. PCR products were digested, electrophoresed in 3% agarose gel with ethidium bromide, and visualized by UV transillumination. Two different investigators independently evaluated the gels.
fig.ommitted
Table 1. Polymerase chain reaction (PCR) conditions, primers, and restriction enzymes used to determine genetic variations at polymorphic sites in the interleukin (IL)4 and IL-4 receptor (IL-4R) genes.
Statistical analysis. The data were analyzed for the total group of children, for children with different severities of RSV bronchiolitis (children receiving and those not receiving mechanical ventilatory support), children with known risk factors (children born prematurely, children who had RSV infection at age <6 months, and children with cardiac or lung disease), and allergic children (children with eczema or food allergy).
To analyze the role of different alleles at specific loci, 2 different but complementary tests were used: (1) The genotyping results for all hospitalized children and parents were used for analysis in the TDT [46]. This test determines whether the "risk" allele is transmitted at greater frequency than the "normal" allele from a heterozygous parent to an affected offspring. All parents can be included in this analysis, regardless of their ethnic origins; differences in allele distribution in different ethnic populations are independent of the transmission of the risk allele. (2) A case-control approach to analysis of data from native Dutch children was used (to prevent bias by population admixture). In this analysis, allele frequencies (frequency with which particular allele is found 0, 1, or 2 times in different groups) among Dutch case patients and Dutch population controls were compared as risk factors for RSV infection. Results from case and control samples were compared using a Mantel-Haenszel trend test for the number of susceptibility alleles of interest (0, 1, or 2), essentially assuming a codominant penetrance model. That is, the more copies of the candidate gene are present, the higher the risk of disease. Because both tests make use of data from the same children (case patients), results from these 2 tests are not statistically independent.
Whenever one of these tests resulted in a (borderline) significant association, we combined all available information, making use of the fact that a case-control comparison of parents of case patients with control subjects is statistically independent of the TDT [47, 48] (N.J.D.N., T.G.K., and B.H., unpublished data). This comparison could be used to explore the same hypothesis, that an association exists between disease and a putative risk allele: if affected children are selected for the presence of specific alleles, then, by implication, their parents are also selected for these alleles (although less so). We thus combined evidence from the TDT with a Mantel-Haenszel trend test, comparing allele frequencies at the loci of interest between Dutch parents and control subjects. Combining evidence from the 2 tests increases the overall power to detect an association between severe disease and the putative risk allele. For this combination of tests, we used the generally applicable Fisher's method [49] for combining multiple independent tests of the same hypothesis. Under the null hypothesis, the 1-sided P values from the 2 tests are both uniformly (0 and 1) distributed. Thus, we can test the null hypothesis, that no association exists between disease and a candidate allele, by comparing -2{ln (pTDT ) + ln (ppc)} with a 2 statistic with 4 degrees of freedom, where pTDT is the (1-sided) P value of the TDT and ppc is the same, for the comparison of parents and control subjects under the codominant penetrance model. Genetic interaction between the different polymorphisms was tested using logistic regression.
RESULTS
IL-4 C-590T polymorphism. The results of the TDT and the case-control analysis are given in . In the case-control study, the T allele was found in a significantly higher frequency among children hospitalized with RSV bronchiolitis than in the control group (odds ratio [OR], 1.43; P = .04). Although the result was not statistically significant, the TDT analysis also showed that the level of transmission of the T allele was higher than the expected 50% (OR, 1.33; P = .13). When all available information was used in the combination test, the association was still significant, which confirmed that a genetic association exists between the occurrence of the T allele and RSV bronchiolitis (OR, 1.41; P = .02). When specific features among the children were examined (e.g., known risk factors or other clinical manifestations), different levels of association were found in both the TDT and the case-control analysis. In the TDT, the T allele had been transmitted significantly more frequently among children who were not receiving mechanical ventilatory support (OR, 1.68; P = .02) and among children who became infected at an age >6 months (OR, 2.22; P = .04). A significant association was also found for these 2 groups of children in the case-control study; the T allele was found at a higher frequency in both groups than it was in the control group (OR, 1.63; P = .01, and OR, 1.83; P = .04, respectively). In addition, in the case-control study, a significant association was found for children who had no recurrent wheezing (OR, 1.95; P = .01), for those who had no cardiac or lung disease (OR, 1.49; P = .03), and for those who did not have eczema (OR, 1.48; P = .04). In the combination test, all of these associations remained significant or became more significant, and in all cases, the T allele was found more frequently than expected by chance. Although significant associations were not found for prematurely born children and children without food allergy in the TDT and case-control study, significant results were found for these groups in the combination test (P = .02 and P = .05, respectively). This discrepancy is probably explained by the higher power of the combination test.
fig.ommitted
Table 2. Results of the transmission/disequilibrium test, case-control study, and combination test for the interleukin-4 C-590T polymorphism among children hospitalized for respiratory syncytial virus (RSV) infection.
IL-4R I50V and Q551R polymorphisms. The results of the TDT and case-control studies of the IL-4R polymorphisms I50V and Q551R are given in and . No significant association between the I50V polymorphism and RSV bronchiolitis was found in any of the tests. Remarkably, no homozygous II genotype was found among children with cardiac or lung disease (P = .06 for genotype distribution). Although only 13 children with such diseases were included in our study, we would expect, on basis of the findings for the control group (in which 29% of subjects were homozygous for this allele), 34 of these children to be homozygous for this allele.
fig.ommitted
Table 3. Results of the transmission/disequilibrium test, case-control study, and combination test for the interleukin-4 receptor I50V polymorphism among children hospitalized for respiratory syncytial virus (RSV) infection.
fig.ommitted
Table 4. Results of the transmission/disequilibrium test, case-control study, and Fisher's combination test for the interleukin-4 receptor Q551R polymorphism among children hospitalized for respiratory syncytial virus (RSV) infection.
Transmission of the R551 allele was significantly more frequent among children who were hospitalized for RSV infection at an age >6 months, (OR, 2.75; P = .01; TDT). Although no significant association with this allele was found in the case-control study (P = .26), the results of the combination test were significant (OR, 1.66; P = .03).
Hardy-Weinberg equilibrium and genetic interactions. For all 3 polymorphisms studied, the genotype distributions in the Dutch population (both RSV-infected children and control subjects) were in Hardy-Weinberg equilibrium. We could not demonstrate genetic interaction between the different polymorphisms (data not shown).
DISCUSSION
Some indications for the involvement of genetic heterogeneity in RSV-induced bronchiolitis were obtained in 2 recent studies: Hull and colleagues [50, 51] showed genetic association with the chemokine IL-8 gene region, which is involved in the neutrophil response, and Löfgren et al. [52] found an association with the surfactant protein A locus, which is involved in the innate immune response. We focused our study on genetic heterogeneity in the Th2 cytokines, in particular IL-4 and the IL-4 receptor, expecting that the frequency of gain-of-function variants of these Th2 cytokine genes would be higher among children with severe RSV bronchiolitis. If so, this would provide further evidence that the Th2 pathway plays a role in natural RSV disease.
A number of polymorphisms have been described in the promoter area of the IL-4 gene, of which the C-590T polymorphism is best studied. In vitro experiments have shown that the T allele has a higher level of promoter activity, probably as a result of the introduction of a site resembling the recognition site for the nuclear factor of activated T cell family of transcription factors [53, 54]. In association studies, the T allele was positively associated with asthma [43, 55 57], but in other studies, this association could not be confirmed [58, 59].
Other polymorphisms described in the IL-4 gene are a C-34T polymorphism [60], a C+33T polymorphism [61], a 70-bp repeat in intron 3 [62], and a dinucleotide repeat in intron 2 [27]. Although the 5 polymorphisms have never been tested together for linkage disequilibrium, strong linkage disequilibrium exists between different individual polymorphisms, at least in Japanese [60, 64, 65], Australian [66], and French populations [63].
In our study, we found a genetic association between the IL-4 locus and RSV bronchiolitis: the -590T allele was more frequently found among children hospitalized for RSV bronchiolitis than in the control group (P = .04; case-control study) and was more often transmitted from the parents of RSV-infected children than would be expected to occur by chance (although this difference was not significant). Combination of all available information in the combination test resulted in a P value of .02 and an OR of 1.41.
We found no association with the IL-4 locus in the group of children with most severe disease (those who required mechanical ventilatory support). This could mean that IL-4 does not play an important role in disease in these children and that other biological processes probably are involved in the development of respiratory failure. However, this group of children might have been too small for detection of a significant association.
Very interesting is our finding of a significant association between the IL-4 locus and severe RSV disease among children who were hospitalized because of RSV bronchiolitis at age <6 months or age >6 months (OR, 2.09; P = .02; combination test), whereas no association could be found among children <6 months old. This fits quite well with the current understanding of maturation of the immune system in infants [67]. Thus, it is easily conceivable that gain-of-function variation in IL-4 and the IL-4 receptor may be more relevant at age 6 months and after. In addition to maturation of the immune system, the loss of maternal antibodies could be relevant.
A subset of the children took part in a follow-up study in which postRSV infection recurrent wheezing was examined. In children who did not experience recurrent wheezing during or after hospitalization, a higher frequency of the IL-4 -590T allele was observed (OR, 1.89; P = .02; combination test), in contrast to the control group and to children with recurrent wheezing. Van Schaik et al. [68] suggested that the release of leukotrienes might be involved in the acute (wheezing) phase of RSV-induced bronchiolitis, probably as a result of enhanced IFN- production. In this context, it would be very interesting to study the genetic association between RSV disease, recurrent wheezing, and IFN- gene variants.
The IL-4R locus is characterized by a large number of polymorphic sites (almost 30). Ten polymorphic sites result in an amino acid substitution, and 9 of these are located in the intracellular domain [69]. Using 7 of these amino acid substitutions, 11 different putative haplotypes were calculated, of which 4 have a cumulative frequency of 90% [70].
In our study, we investigated the association of severe RSV disease with 2 polymorphic sites of IL-4R: the I50V and the Q551R polymorphisms. The I50V polymorphism is located in the extracellular domain of the IL-4 receptor. Different approaches have been used to demonstrate that the 50I variant has the ability to up-regulate the receptor response to IL-4 than does the 50V variant [32, 71, 72]. The I allele was associated with atopic asthma in 2 Japanese association studies [71, 73], but the association was not found in a third Japanese study [44]. The I50V polymorphism is not in linkage disequilibrium with the Q551R polymorphism, as was shown in the present study and elsewhere [74, 75].
The Q551R polymorphism is located in the intracellular domain of the receptor in a region known to play an important role in IL-4induced activation of STAT6 DNA-binding activity [76]. Hershey et al. [77] found that the level of expression of CD23 is higher after stimulation with IL-4 in peripheral blood mononuclear cells from R551 homozygotes and R551 heterozygotes, which suggests that the signaling ability of the R551 allele is greater. The level of binding of STAT6 was similar for both alleles. The same group reported a genetic association of the R551 allele with severe atopic eczema [77] and severe asthma [78]. Mitsuyasu et al. [72] found no association between this allele and atopic asthma. Finally, Kruse et al. [79] found lower IgE levels in individuals bearing the R allele.
In our study, we did not find any indication of a genetic association between the I50V polymorphism and RSV bronchiolitis. In contrast, we found a significant association between the R551 allele and severe RSV bronchiolitis in children who were hospitalized at an age 6 months (OR, 1.66; P = .03; combination test). This further supports the idea that Th2-mediated cell response is more important after the first half-year of life. Because both the IL-4 receptor and the IL-13 receptor contain IL-4R, we cannot determine to what extent the positive association may result from binding of IL-4, IL-13, or both to their receptors. Interestingly, IL-13 polymorphisms were found to be associated with asthma and atopy [80, 81]. Moreover, a significant gene-gene interaction between the S478P polymorphism in IL-4R and the -1111 promoter variation in IL-13 was associated with a risk of developing asthma that was 5 times greater among individuals who carry these risk alleles [82]. These findings and the observation that IL-13 induces airway hyperreactivity during RSV infection of mice [83 85] make IL-13 an important candidate gene for future studies on RSV disease in children.
In conclusion, our results demonstrate positive associations of gain-of-function mutations in 2 genes involved in Th2-type immune response with severe RSV bronchiolitis. Our results support the idea that Th2 responses may contribute to severe RSV disease. Our study also supports the idea that genetic differences may, to a certain extent, explain the differences in disease severity among RSV-infected children. In theory, positive associations do not prove causality. Another locus in close linkage disequilibrium with the examined one may be responsible (i.e., confounding). On the other hand, we have chosen markers with proven functional significance, which supports the suggestion that they play a causative role in the RSV disease process. Our results reveal interesting aspects of the pathogenesis of RSV disease. Because there were discrepant results for children younger than and children older than 6 months of age, it appears that Th2-mediated pathology is relatively more important at older ages, after the loss of maternal antibodies. Further studies should elucidate the relative role of these and other cytokine polymorphisms. In particular, whether children with a Th2 genotype have higher levels of Th2 cytokines in their lungs should be examined. In addition, it would be interesting to examine the extent to which genetic predisposition to severe RSV disease contributes to the development of asthma.
Acknowledgments
We thank Anita Boelen and Marie-Louise Heijnen, for their help in initiating the project; Lodewijk Sandkuijl, for advice and stimulating discussions; Marijke Steijn, for collecting a large number of the samples; and Wendy Swelsen, Petra van Impelen, and Mirjam Schaap, for technical assistance.
References
1.
2.
Simoes EAF. Respiratory syncytial virus infection. Lancet 1999; 354:84752.
3.
Hall CB. Respiratory syncytial virus and parainfluenza virus. N Engl J Med 2001; 344:191728.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
Viet AL, van Gils HWV, Melker H de, et al. Risk factors and health in the Netherlands, a survey by municipal health centres; annual report 1998. Centraal Bureau voor de Statistiek (CBS-KPE); GGD Nederland RIVM Rapport 266807003. Available at: . Accessed 25 November 2002.
42.
43.
44.
45.
46.
Sham PC. Statistics in human genetics. London: Arnold, 1998.
47.
48.
49.
Fisher RA. Statistical methods for research workers. 4th ed. London: Oliver and Boyd, 1932.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.