Portal de Periódicos da CAPES

Studies of cord blood mononuclear cell responses and allergy: still in their infancy?

2002; Wiley; Volume: 32; Issue: 3 Linguagem: Inglês

10.1046/j.1365-2222.2002.01322.x

1365-2222

Graham Devereux, Robert N. Barker,

IL-33, ST2, and ILC Pathways

It is nearly 10 years since Kondo et al. [1] first demonstrated that cord blood lymphocytes can proliferate after in vitro stimulation with allergens. It was also shown that these responses were more common in samples from neonates who subsequently developed atopic disease (atopic asthma, atopic dermatitis, allergic rhinitis, food allergy). Since then there has been a steady flow of publications describing such responses of cord blood mononuclear cells (CBMC), their associations with subsequent childhood atopic disease and their determinants. The study by Marks and colleagues [2] in this issue of Clinical and Experimental Allergy provides a timely opportunity to review issues relating to in vitro CBMC responses to allergens. There is increasing recognition that factors present during fetal development can influence the risk of developing postnatal disease [3]. Atopic diseases are characterized by chronic immune-mediated inflammation, initiated and perpetuated by CD4+ T helper (Th) cells. Th1 and Th2 phenotypes are characterized by the secretion of IFN-γ and IL-4, respectively [4], and in atopic diseases Th2 responses predominate [5,6]. If the phenomenon of fetal programming influences the susceptibility to atopic diseases, then it would be predicted to do so via the developing fetal immune system. Indeed, several lines of evidence suggest that in utero influences are important in the immunopathogenesis of postnatal atopic diseases. First, observational epidemiological studies have demonstrated that gestational age and anthropometric measurements at birth, such as weight and head circumference are associated with serum IgE and atopic disease [7–12]. Secondly, differences have been reported between the proliferative and cytokine responses in vitro of CBMC from neonates who subsequently develop atopic disease and healthy controls. The results of these studies are summarized in Tables 1 and 2[1,13–20]. Initial reports uniformly demonstrated that increased CBMC proliferative responses are associated with an increased likelihood of childhood atopic disease [1,13,14,17], but two more recent studies have failed to demonstrate such a relationship [15,16]. The CBMC cytokine responses of neonates who subsequently develop atopic disease exhibit a reduced IFN-γ secretory response [13,15,17,18,20] when compared with healthy controls, with Prescott et al. [15] also demonstrating reduced IL-4, IL-6, IL-10 and IL-13 responses. On balance it appears that there is an underlying association between CBMC responses and the development of atopic disease. Despite this overall trend, there are clear inconsistencies between studies. Furthermore, within studies there are different associations with individual polyclonal and allergenic stimuli [13–15,17]. Several factors could contribute to this inter- and intrastudy variability. The numbers of neonates included have generally been small and inconsistently selected by parental atopic disease. In addition, variations between culture conditions such as cell number, culture volume, media, serum supplementation and culture duration have all been shown to influence measured in vitro Th cell responses [21,22]. Such methodological considerations are likely to have contributed to the inconsistencies between studies because different investigators have used different culture parameters and methods to quantify CBMC responses. This lack of standardization is compounded further by a lack of consensus in the interpretation of CBMC responses. For example, Marks et al. [2] use several parameters to express a single cytokine concentration and other authors have defined significant proliferative responses as stimulation indices ranging from 1.5 to 3 [1,14,17]. Another limitation to the interpretation of results is that statistical analysis of CBMC responses has been limited to the use of conservative non-parametric methods: the population frequency distributions of CBMC proliferative and cytokine responses are unknown because study populations have been relatively small and heavily biased by parental atopy. Given that, overall, in vitro CBMC proliferative and cytokine responses are associated with subsequent atopic disease, a number of studies have attempted to identify influential antenatal factors. Observational epidemiological studies have established that family history, birth order, maternal smoking and diet influence the risk of atopic disease. It is surprising therefore, that only two factors (family history and in utero sensitization) have been investigated as determinants of CBMC responses. The focus on these two factors arose from the expectation that the analysis of CBMC responses at birth would identify infants at risk of atopic disease, enabling primary preventative measures in the form of pre and/or postnatal allergen avoidance to be targeted. Unfortunately this hope has not been realized [23], but measurement of CBMC responses does present an opportunity to identify the antenatal genetic and environmental factors that modulate the postnatal risk of atopic disease. Several studies have demonstrated that a family history of atopic disease is associated with increased in vitro CBMC proliferative responses [15, 24, 25]. However, an equal number have also failed to demonstrate any significant association [1, 26, 27] and a more recent study demonstrated that maternal asthma was associated with reduced CBMC proliferative responses [16]. Reports of the association between CBMC cytokine responses and a family history of atopic disease are similarly inconsistent, with a family history of atopic disease being associated with reduced IFN-γ CBMC responses [28,29], increased IL-4 and IL-5 responses [30] or showing no association [19,24,27]. Any influence of genetic factors on the fetal immune system and the subsequent risk of atopic disease is important, but environmental factors that have changed in the last 30 years must underlie the recent increase in atopic disease [31,32]. Physical exposure of the fetus to cat and house dust mite allergen can occur [33,34], although the frequency of such exposure, the range of allergens involved, and the consequences need to be established. An important question is whether antenatal allergen exposure specifically stimulates the fetal immune system and thereby influences the postnatal risk of atopic disease. The demonstration of CBMC responses after in vitro stimulation with allergens and their association with atopic disease has been widely cited as reliable evidence that antenatal sensitization occurs and is a risk factor for atopic disease [13–15,27]. Early work relating CBMC proliferative responses with maternal exposure to seasonal allergens supports this concept. These studies reported that the frequency and magnitude of CBMC proliferative responses after stimulation with birch pollen and timothy grass pollen were greater if maternal exposure had occurred during specified periods of gestation [26,35]. Recent work has cast doubt on this association between maternal allergen exposure and CBMC proliferative responses. Szepfalusi et al. [36] demonstrated that the likelihood of CBMC proliferative responses did not significantly increase during the course of gestation and were unrelated to maternal exposure to timothy grass pollen and birch pollen. Similarly, Chen-Yeung et al. [16] were also unable to demonstrate any association between maternal house dust mite exposure and CBMC proliferative responses after in vitro stimulation with the allergen. In accordance with this, in the current issue of Clinical and Experimental Allergy, Marks and colleagues [2] report that the concentration of house dust mite allergen in the mother's bed at 36 weeks gestation does not correlate with CBMC cytokine responses. The studies with seasonal allergens such as birch pollen clearly demonstrated that they failed to elicit proliferative responses by CBMC samples from some neonates whose mothers had been exposed to the allergen. In these cases, the lack of responsiveness could be explained if the allergen did not cross the placenta, or, alternatively, the methodological considerations discussed above may have limited assay sensitivity. More intriguingly, and seldom commented on, are the CBMC samples that respond in vitro to seasonal allergen despite the lack of exposure of the mothers to that allergen during pregnancy [26,35,36]. It seems unlikely that this common phenomenon is a consequence of antigen cross-reactivity, or persistence of the allergen within the environment or the mother. An alternative, but much ignored explanation, is that some of the observed CBMC proliferative responses are primary responses mediated by Th cells that have not been previously stimulated in utero. It is well recognized in other areas of immunology that certain culture conditions can support primary proliferative responses, which are slow to develop, peaking 6 days or later after stimulation with non-recall antigens such as keyhole limpet haemocyanin [21,37,38]. Therefore, the prolonged culture periods (6+ days) typically used in CBMC proliferation studies raise the possibility of sustaining and detecting primary responses. Such primary in vitro Th cell responses may well have contributed to the inconsistencies between CBMC proliferation studies. Our own work analysing the activation status of Th cells has demonstrated that, in equal numbers of neonates tested, CBMC proliferating in response to in vitro stimulation with timothy grass pollen were drawn from the previously activated CD45ROhigh or the naïve CD45RAhigh subsets [39]. Thus, in utero sensitization by the allergen is common, but in vitro proliferative responses by CBMC are not alone reliable indicators of such priming. Although in vitro activation of naïve cells may have weakened associations between atopic diseases and CBMC proliferative responses to seasonal allergens, it is less clear whether this phenomenon contributes to the negative results from studies of perennial allergens such as that by such as Marks et al. [2]. These authors report no association between CBMC cytokine responses 24 and 48 h after stimulation with house dust mite and maternal exposure to this allergen. It may be argued that, at such early time-points, the observed CBMC responses are likely to be mediated by Th cells that have been sensitized in utero, and this contention could be confirmed by characterization of the sensitive methodology used., Assuming that all of the CBMC cytokine responses reported by Marks et al. [2] are a consequence of antenatal sensitization, why was there no association demonstrated with maternal exposure? Our own study of 223 neonates representative of the general population [40] suggests that parental atopy and in utero allergen sensitization are weak influences on in vitro CBMC proliferative and cytokine responses, whereas maternal smoking, birth order and maternal dietary vitamin E intake exert more powerful influences. The finding that the magnitude of CBMC responses is influenced by these multiple factors has a number of consequences. As in any observational epidemiological study, small studies of selected neonates that attempt to relate CBMC responses to parental atopic disease and/or allergen exposure are likely to be inconsistent because of uncontrolled variation in the distribution of the other, more potent, influences. For example, in the study of Marks et al. [2], if the distribution of more potent influential factors is randomly associated with maternal allergen exposure, simple univariate analyses without adjustment for these factors may fail to demonstrate a weak association between allergen exposure and CBMC response. The relationships between the development of atopic disease and in utero allergen sensitization, and other influential factors such as family history, birth order, maternal smoking and diet are unknown. At least three models can be proposed. First, in utero allergen priming may be a necessary step, but insufficient alone, for the development of atopic disease, with the other factors determining the risk of disease by influencing subsequent differentiation of the activated Th cells. Alternatively, sensitization of the fetal immune system by allergen may be the final critical event that determines the subsequent risk of atopic disease, and the other factors interact to modulate the likelihood of Th cell activation by in utero allergen exposure. The third possibility is that in utero allergen sensitization is only one of several independent influences that together modulate immune development and the risk of atopic disease. In this last scenario, the risk of atopic disease can be increased by in utero sensitization, but also by the other factors in the absence of such priming. Furthermore, fetal allergen sensitization may be a relatively weak or unimportant influence on the risk of atopic disease. Testing these three models, which is dependent on resolving the methodological issues discussed here, will be crucial for the design of primary preventative measures. In particular, it will be necessary to determine whether in utero allergen sensitization is a prerequisite for childhood atopy and to establish the nature of any interactions between sensitization and other antenatal risk factors such as family history, birth order, maternal smoking and diet. Once this is achieved, studies of CBMC responses will be seen to have reached maturity.