Allergen skin test reactivity and asthma are inversely associated with ratios of IgG4/IgE and total IgE/allergen‐specific IgE in Ugandan communities

Summary Background Serum inhibition of allergen‐specific IgE has been associated with competing IgG4 and non‐specific polyclonal IgE. In allergen immunotherapy, beneficial responses have been associated with high IgG4/IgE ratios. Helminths potentiate antibody class switching to IgG4 and stimulate polyclonal IgE synthesis; therefore, we hypothesized a role for helminth‐associated IgG4 and total IgE in protection against atopic sensitization and clinical allergy (asthma) in tropical low‐income countries. Methods Among community residents of Ugandan rural Schistosoma mansoni (Sm)–endemic islands and a mainland urban setting with lower helminth exposure, and among urban asthmatic schoolchildren and non‐asthmatic controls, we measured total, Schistosoma adult worm antigen (SWA)–specific, Schistosoma egg antigen (SEA)–specific and allergen (house dust mite [HDM] and German cockroach)–specific IgE and IgG4 by ImmunoCAP® and/or ELISA. We assessed associations between these antibody profiles and current Sm infection, the rural‐urban environment, HDM and cockroach skin prick test (SPT) reactivity, and asthma. Results Total IgE, total IgG4 and SWA‐, SEA‐ and allergen‐specific IgE and IgG4 levels were significantly higher in the rural, compared to the urban setting. In both community settings, both Sm infection and SPT reactivity were positively associated with allergen‐specific and total IgE responses. SPT reactivity was inversely associated with Schistosoma‐specific IgG4, allergen‐specific IgG4/IgE ratios and total IgE/allergen‐specific IgE ratios. Asthmatic schoolchildren, compared with non‐asthmatic controls, had significantly higher levels of total and allergen‐specific IgE, but lower ratios of allergen‐specific IgG4/IgE and total IgE/allergen‐specific IgE. Conclusions and clinical relevance Our immuno‐epidemiological data support the hypothesis that the IgG4–IgE balance and the total IgE–allergen‐specific IgE balance are more important than absolute total, helminth‐ or allergen‐specific antibody levels in inhibition of allergies in the tropics.


| INTRODUC TI ON
Several immuno-epidemiological studies have shown that helminth infections are associated with protection against allergy-related conditions, [1][2][3][4][5][6][7][8][9] and highlighted the potential role of helminths in rural-urban differences in prevalence of allergy-related diseases in the tropics. 10,11 We have previously shown that maternal hookworm infection modifies risk factors for childhood eczema, implying that early-life exposure to helminths may also establish protection against allergy-related diseases. 12 Experimental human and animal studies have demonstrated that the mechanisms through which helminths may down-modulate allergic responses are extensive, 13,14 covering almost the entire range of the allergy-related immunological pathway. Antibody-mediated immune mechanisms of protection are less understood compared with cell-mediated mechanisms: current theories involve helminth-induced immunoglobulin (Ig) G4 and polyclonally stimulated IgE.
By inducing high levels of interleukin (IL)-10, helminths can promote immunoglobulin class switching to IgG4. Moreover, chronic helminth infection is associated with elevated serum IgG4 levels, 15 and serum inhibition of helminth-specific IgE has been associated with competing IgG4 in a Schistosoma mansoni (Sm)-endemic setting. 16 Several human studies have suggested that helminth-induced IgG4 is important in protection against allergy: among school-age children in rural Ecuador, Ascaris-specific IgG4 was inversely associated with allergen skin prick test (SPT) reactivity 2 ; in a Sm-endemic Ugandan rural setting, we observed an inverse association between house dust mite-specific IgG4/IgE ratios and reported recent wheeze. 17 Studies conducted four decades ago 18,19 provided initial evidence that parasitic helminths mediate production of high levels of IgE that is not specific to the parasite, or to inhalant allergens.
Non-specific polyclonally stimulated IgE has been proposed to inhibit allergic responses by competing with allergen-specific IgE to saturate IgE receptors, 20 reducing the chances that an allergen will result in cross-linking of FcεRI-bound IgE and hence effector cell degranulation. 14,21 However, there is little evidence for polyclonal IgE-mediated protection against allergic inflammation, and high IgE titres have previously been linked to increased expression of IgE receptors on human basophils, 22 signifying potential for polyclonally stimulated IgE in increased effector cell degranulation. Therefore, the question of whether polyclonal IgE mitigates allergic responses remains unresolved.
Large, well-defined immuno-epidemiological studies in helminth-endemic settings are required to better understand population-level interactions between allergy-related disease and helminth-and allergen-associated IgG4 and polyclonal IgE profiles.
This will contribute to bridging the gap between understanding basic antibody mechanisms and clinical applications. We used the opportunity presented by studies designed to assess the epidemiology of allergy-related disease in (1) Sm-endemic Ugandan rural fishing villages, 17,[23][24][25] (2) proximate urban communities with lower helminth exposure 11 and (3) urban asthmatic schoolchildren and nonasthmatic controls. 26,27 Samples collected enabled us to measure total IgE (as a proxy for polyclonally stimulated IgE), total IgG4 and Sm-and allergen-specific IgE and IgG4 profiles, and to analyse their associations with current Sm infection, the rural-urban environment, allergic sensitization and asthma.

| Study design and population
The current investigation was conducted using samples from participants of cross-sectional community surveys in rural 24 and urban 11 Uganda, and from a case-control study investigating asthma risk factors among 5-to 17-year-old schoolchildren in urban Uganda. 26 The rural survey was the three-year outcome survey (September community settings, both Sm infection and SPT reactivity were positively associated with allergen-specific and total IgE responses. SPT reactivity was inversely associated with Schistosoma-specific IgG4, allergen-specific IgG4/IgE ratios and total IgE/ allergen-specific IgE ratios. Asthmatic schoolchildren, compared with non-asthmatic controls, had significantly higher levels of total and allergen-specific IgE, but lower ratios of allergen-specific IgG4/IgE and total IgE/allergen-specific IgE.

Conclusions and clinical relevance:
Our immuno-epidemiological data support the hypothesis that the IgG4-IgE balance and the total IgE-allergen-specific IgE balance are more important than absolute total, helminth-or allergen-specific antibody levels in inhibition of allergies in the tropics. The asthma case-control study (May 2015-July 2017) 26 enrolled children with doctor-diagnosed asthma ("cases") and non-asthmatic controls from primary and secondary schools in Entebbe municipality and Katabi sub-county in Wakiso, Uganda. The International Study on Allergy and Asthma in Children (ISAAC) questionnaire 28 was used for screening, to identify participants who reported wheezing in the previous 12 months. These children then underwent a comprehensive clinical evaluation of medical and treatment history, and examination for asthma signs and symptoms by study clinicians in order to make an asthma diagnosis. Lung function tests and asthma control tests 26,27 were conducted to assess asthma control. As reported elsewhere, 27 only three cases had abnormal lung function tests, 85% had well-or partly controlled asthma and 15% had poorly controlled asthma. All cases were seen at one time-point, started on recommended treatment and referred for further management; therefore, asthma severity or response to treatment could not be assessed. 27 Non-asthmatic controls were children in the same class as cases, with no history of wheezing or any asthma symptoms, randomly selected using a Stata program (StataCorp, College Station, TX, USA) to obtain a control/case ratio of 2:1.

| Parasitological examinations
In all studies, we used the stool Kato-Katz (KK) technique 29 for diagnosis of Sm, hookworm (Necator americanus), Ascaris lumbricoides and Trichuris trichiura infections. For each participant, one stool sample (prepared on two slides) was examined under a microscope by two laboratory technologists blinded to each other's result. In the rural and urban surveys, stool was further suspended in 70% ethanol, stored at −80°C and later examined for Sm, Strongyloides stercoralis and hookworm infections using multiplex real-time PCR. 30,31 Infection with Schistosoma haematobium has not been documented in our study areas. 32 Plasma samples were assessed for Schistosoma adult worm antigen (SWA)-and Schistosoma egg antigen (SEA)specific IgE and IgG4 by ELISA (details in Appendix S1).

| Assessment of allergy-related outcomes
Wheezing in the previous 12 months was also assessed in the rural and urban surveys using interviewer-administered standardized paper and video ISAAC questionnaires. Skin prick test (SPT) reactivity to crude extracts from Dermatophagoides mix (D. pteronyssinus and D. farinae), Blomia tropicalis and Blattella germanica (ALK-Abelló; supplied by Laboratory Specialities [Pty] Ltd., SA) was assessed using standard methods. 33 These allergens are common in the study settings. 34 We used an in-house ELISA to quantify crude house dust mite (D. pteronyssinus, HDM) and German cockroach (B. germanica) extract-specific plasma IgG4 and total IgG4. Total IgE and allergen extract-specific IgE were measured using the ImmunoCAP ® IgE test (Thermo Fisher Scientific, Uppsala, Sweden). 35 Allergen extractspecific IgE was additionally measured using an in-house ELISA, to facilitate determination of IgG4/IgE ratios, since, for logistical reasons, we could not measure allergen-specific IgG4 by ImmunoCAP ® .
The in-house assays are described in this article's supplementary information Appendix S1.
Graphs were drawn using GraphPad Prism (version 8.2.1; Fay Avenue, CA, USA). We accounted for study design in all analyses involving the rural and urban community surveys: Stata "svy" commands were used to allow for the non-self-weighting clustering by village in the rural survey and for clustering by sub-ward in the urban survey. 11 Initial analyses in the rural survey assessed the impact of intensive versus standard MDA on IgE and IgG4 profiles. For this, we used a cluster-level approach 36 (previously described in other analyses for this study 24 ) that involved comparing cluster-specific means of antibody concentrations between trial arms, with 95% confidence intervals and p-values calculated using the t distribution. For all subsequent analyses, standard and intensive trial arm participants in the rural survey were grouped together.
Next, participant characteristics were tabulated and compared between urban and rural settings separately, between asthma cases and controls, using logistic and linear regression for categorical and continuous variables, respectively.
Independently for each of the three studies, we conducted crude and age-and sex-adjusted cross-sectional analyses to investigate the direction and the strength of association between antibody concentrations or ratios (total, Schistosoma-specific and allergenspecific IgE and IgG4 concentrations; allergen-specific IgG4/IgE, total IgE/allergen-specific IgE, total IgG4/total IgE ratios) and Sm infection or SPT reactivity. In the asthma case-control study, we additionally investigated associations between asthma and antibody concentrations and ratios. Rural-urban comparisons of antibody profiles were conducted using data from the rural and urban surveys. Most raw antibody concentrations and ratios were skewed, so for all analyses, linear regression models of log-transformed data were used. ImmunoCAP ® allergen-specific IgE concentrations were log 10 (+0.001)-transformed, while all other antibody data were log 10 (+1)-transformed. The results were then back-transformed to obtain geometric mean ratios (GMRs) and 95% confidence intervals (CIs). Science and Technology (reference numbers: HS1183, HS2036 and HS1707). We obtained written informed consent from all participants and/or their legal guardians. Informed assent was obtained from children ≥8 years.

| Characteristics of study participants
Flow charts of the three studies are shown in Figure 1. Plasma samples were available for 2961, 1356 and 1685 participants of the rural survey, urban survey and asthma case-control study, respectively.
Antibody measurements were conducted in a subset of randomly selected samples per study: 791, 1320 and 406 plasma samples from the rural survey, urban survey and asthma case-control study, respectively, were included in assessment of at least one of total, Schistosoma-or allergen-specific IgE or IgG4. There was no effect of intensive versus standard anthelminthic MDA on antibody profiles (Table S1), hence data from both trial arms in the rural survey were combined for all subsequent analyses. Table 1 shows characteristics of participants for whom we obtained data on total, Schistosoma-or allergen-specific IgE or IgG4.
Participant characteristics were compared between the rural and urban surveys, and between asthma cases and controls. We have previously reported related comparisons in these studies, albeit in somewhat different subsets of participants. 37 Rural participants, compared with urban participants, were more likely to be older (p = .001) and male (p = 0.003). Prevalence of SPT reactivity to Dermatophagoides mix and B. tropicalis was higher among urban participants (p = .001). However, rural participants had higher median total IgE levels (p < .001), and higher prevalence of IgE sensitization to crude German cockroach (ImmunoCAP ® concentration ≥0.35 kU/L) [p = .001], ELISA-detectable cockroach-specific IgE (≥312.5 ng/ml) [p = .001], and infection with Sm (p < .001) and at least one nematode (p < .001). Reported wheeze was rare in both settings.
Compared to non-asthmatic controls, asthma cases were on average older, and more likely to be SPT positive, and to have higher total and allergen-specific IgE levels. Helminth infection prevalence was low in the asthma case-control study and comparable between cases and controls ( Table 1).

| Associations between antibody (IgE and IgG4) concentrations and current S. mansoni infection and allergen skin prick test reactivity
In the rural survey (Table 2), Sm infection was positively associated with total and SWA-and SEA-specific IgE and IgG4 In the urban survey (Table 3), Sm infection was positively associated with total and SWA-and SEA-specific IgE and IgG4 (p < .001), and ELISAdetermined HDM-specific IgE (p = .037), cockroach-specific IgE (p = .040) and cockroach-specific IgG4 concentrations (p = .005). Cockroach SPT reactivity was positively associated with cockroach-specific IgE respectively, among both asthma cases and non-asthmatic controls (Table S2).

| Associations between antibody ratios and current S. mansoni infection and skin prick test reactivity
Schistosoma mansoni infection was positively associated with total IgE/cockroach-specific IgE ratios in both community surveys  Figure 2A]. There were no significant associations between Sm infection and antibody ratios in the asthma case-control study.
In all three studies, cockroach SPT reactivity was inversely asso- [ Figure 2B].

F I G U R E 2
Associations between antibody ratios and Schistosoma mansoni infection and skin prick test reactivity. Forest plots show geometric mean ratios (GMRs) and 95% confidence intervals (95% CIs) for associations between antibody ratios and (A) current S. mansoni infection and (B, C) allergen skin prick test reactivity. Raw antibody responses were skewed, so log 10 -transformed antibody data were used in our linear regression models; we back-transformed the results to obtain GMRs and 95% CIs. All GMRs and 95% CIs were adjusted for age and sex, and additionally for survey design in the rural and urban surveys. GMRs and 95% CIs for associations with skin prick test reactivity were additionally adjusted for S. mansoni infection status. Red colour denotes associations where GMR and 95% CI >1; blue colour denotes associations where GMR and 95% CI <1, and black colour denotes lack of a significant association. * Antibody levels detected by ELISA.
Assessment of associations between antibody ratios and skin prick test reactivity, independently among asthma cases and nonasthmatic controls, showed observations akin to the above: cockroach and HDM SPT reactivity were inversely associated with allergen-specific IgG4/IgE ratios, total IgE/allergen-specific IgE ratios and total IgG4/total IgE ratios (Table S3).

| Associations between antibody profiles and asthma
Asthma was positively associated with total IgE (1. 53 Figure 3B]. Asthma is a syndrome of different phenotypes 38 and is sometimes broadly grouped into "allergic" and "non-allergic" phenotypes based on atopic sensitization. In our case-control study, "allergic" (SPT positive) asthma cases had lower HDM-specific IgG4/ IgE, total IgG4/total IgE and total IgE/allergen-specific IgE ratios, compared with "non-allergic" (SPT negative) asthma cases ( S3 and Figure S1A and S1B). Furthermore, HDM SPT-positive asthma cases had lower HDM-specific IgG4/IgE, total IgG4/total IgE and total IgE/HDM-specific IgE ratios, compared with both SPT-negative and SPT-positive controls (Table S3 and Figure S1A).
The same trend was observed for comparisons between cockroach SPT-positive asthma cases and both SPT-negative and SPTpositive controls, albeit without statistical significance ( Figure   S1B). Interestingly, HDM and cockroach SPT-negative asthma cases had lower geometric means of allergen-specific IgG4/IgE and total IgE/allergen-specific IgE ratios than SPT-negative controls, but higher levels compared with SPT-positive controls (Table   S3).

| DISCUSS ION
We assessed total, Schistosoma-and allergen-specific IgE and IgG4 concentrations among participants of three large population-based studies in Uganda and found strong inverse associations between allergen SPT reactivity and Schistosoma-specific IgG4, allergenspecific IgG4/IgE ratios and total IgE/allergen-specific IgE ratios.
Importantly, doctor-diagnosed asthma cases also had significantly lower allergen-specific IgG4/IgE ratios and total IgE/allergenspecific IgE ratios compared with non-asthmatic controls. Sm infection and the rural (vs. urban) environment were positively associated with total, Schistosoma-and allergen-specific IgE and IgG4, and with allergen-specific IgG4/IgE ratios and total IgE/allergen-specific IgE ratios.
Host responses to helminths exhibit significant similarities with responses to common allergens, owing to related molecular targets. 39  Mitre et al. 44 initially found that a high total IgE/allergen-specific IgE ratio did not inhibit basophil degranulation, but further experiments suggested that inhibition could occur at ratios exceeding 500:1. In the current analysis, we show that total IgE/allergen-specific IgE ratios were strongly inversely associated with allergen SPT reactivity (irrespective of study setting), and with asthma, providing strong support for the above hypotheses. Moreover, geometric means of total IgE/allergen-specific IgE ratios exceeded 700 among non-asthmatic controls in the case-control study, and 1000 among our Sm-infected and SPT-negative participants irrespective of study setting-in excess of Mitre's threshold for inhibition.
Our observations of inverse associations between allergenspecific IgG4/IgE ratios and SPT reactivity and asthma strengthen F I G U R E 3 Asthma case-control study: associations between asthma and antibody concentrations and ratios. Forest plots show geometric mean ratios (GMRs) and 95% confidence intervals (95% CIs) for associations between asthma and antibody concentrations (A) and ratios (B). Raw antibody responses were skewed, so log 10 -transformed antibody data were used in our linear regression models; we backtransformed the results to obtain GMRs and 95% CIs. All GMRs and 95% CIs were adjusted for age and sex. Red colour denotes associations where GMR and 95% CI >1; blue colour denotes associations where GMR and 95% CI <1, and black colour denotes lack of a significant association. * Antibody levels detected by ELISA. Concentrations are in ng/ml. ** Antibody levels detected by ImmunoCAP®. Concentrations are in kU/L. SEA, Schistosoma egg antigen; SWA, Schistosoma adult worm antigen the argument that the IgG4-IgE balance is also important in protection against allergies. In addition, SWA-and SEA-specific IgG4 concentrations were significantly lower among HDM SPT-reactive rural survey participants and cockroach SPT-reactive urban survey participants, respectively. Although IgG4 has been implicated in IgG4related disease, 45 it appears to have a more prominent role in immune regulation 16 and tolerance. 46 Besides, IgG4 does not activate the complement system or lead to formation of immune complexes, and is not associated with mast cell or basophil degranulation. 47 Its major mode of action against allergic inflammation seems to be blockage of allergen recognition by IgE, because both antibodies have similar antigenic specificity. 16 There is a counter-argument that IgG4 might not have a direct mechanistic role in protection against allergy-related outcomes.
Instead, the inverse association between IgG4 levels and IgE effector function may represent an 'epiphenomenon', merely reflective of parallel phenomena, such as abundance of IL-10 (and/or TGFβ)-producing T regulatory and B regulatory cells. 53 However, the latter hypothesis remains to be confirmed by experimental studies.
We measured allergen extract-specific IgE using the standard ImmunoCAP ® test, and later, using an in-house ELISA assay. The latter technique was used to facilitate determination of IgG4/IgE ratios, since, for logistical reasons, we could not measure allergen-specific IgG4 by ImmunoCAP ® . However, patterns of associations with ImmunoCAP ® -and ELISA-determined IgE were comparable (except for cockroach-specific IgE in a few instances). We did not formally adjust for multiple statistical testing. However, our interpretations rely on patterns and magnitude of associations, and consistency and biological credibility of our findings based on other published works.
It is noteworthy that our studies were cross-sectional; therefore, it was not possible to demonstrate causal relationships between antibody profiles and Sm infection, the rural-urban environment, allergic sensitization and asthma.
In conclusion, we underline the potential role of the total IgEallergen-specific IgE balance and the IgG4-IgE balance in mitigation of allergic responses, providing correlative evidence from three cross-sectional population-based studies. These results build upon previous findings that apportion high IgG4/IgE ratios an important role in allergen-specific immunotherapy. 46 Further studies in animal models would advance understanding of a potential role of IgG4 (and polyclonally stimulated IgE) in mitigation of allergy-related disease; however, they are hindered by the lack of IgG4 in mice. Therefore, future work should focus on in vitro experiments using human samples. For example, a microarray screening platform could be utilized to identify helminth antigens that provoke robust serum IgG4 and little or no IgE response. Such antigens may be valuable for therapeutic approaches against allergic conditions.

ACK N OWLED G EM ENTS
We thank Koome sub-county and Entebbe municipality community members for participating in the rural (LaVIISWA) study and the urban survey, respectively. We are grateful to schoolchildren in Entebbe municipality and the surrounding areas in Wakiso District who took part in the asthma case-control study. We also thank all study staff.

CO N FLI C T O F I NTE R E S T S
The authors declare no conflict of interest.