The gut microbiome and asthma in a Swedish twin study.

To the editor, Asthma studies have naturally focused on the airway microbiome, however, some studies support the hypothesis that the onset of allergic disease, lies partly in the microbial communities in the gastrointestinal tract, with asthma and taxa-specific shifts in the abundance of specific bacteria being interrelated.1 Our aim was to study the association between the gut microbiome and allergic asthma in a well-characterized Swedish cohort of twin children aged from 9 to 14. We conducted a population-based study of Swedish twin children recruited from the Swedish Twin study On Prediction and Prevention of Asthma (STOPPA) cohort.2 Per protocol, twins completed written standardized questionnaires and participated in clinical examinations where blood and faecal samples were collected.2 Recruitment was based on informed consent received from both children and their parents/guardians and approval from the Swedish Ethical Review Authority. We defined allergic asthma as a composite outcome based on parental response to the question of current asthma (yes/no) and/or wheezing within the last 12 months (yes/no)3 and a raised serum IgE level of ≥0.35 kUA/L.4 Those with a raised IgE level but no asthma were classified as IgE+ sensitized, while children with neither asthma nor raised IgE levels (IgE+) were included as healthy controls. In total, 355 individual children provided faecal samples, and these were processed using shotgun sequencing. Quality filtering of samples and taxa yielded a total of >8.6 × 109 reads, representing 339 participants and 4461 distinct taxa. 43 children (13.6%) suffered from allergic asthma, 18 (5.7%) from asthma only, and 91 (28.7%) had raised IgE levels but no asthma (IgE+ allergic); in comparison, 165 (52%) had neither asthma nor elevated IgE (healthy controls). More boys than girls had raised IgE (69.2% vs. 30.8%) and allergic asthma (65.1% vs. 34.9%), respectively. After rarefaction to 1 × 106 reads for each of the 339 samples, alpha- and beta-diversity measures were calculated and the n = 317 samples with non-missing exposure information were used for comparing diversity between exposure groups.5 We found no statistically significant evidence for differences in within-sample alpha diversity between exposure groups, either before (Shannon: p = .39, Simpson: p = .29) or after adjustment for shared twin effects (Shannon: p = .59, Simpson: p = .33). Furthermore, there was no statistically significant evidence for differences in overall species composition between exposure groups in unadjusted analyses, for either phylogenetically informed measures (weighted and unweighted UniFrac distance6) or the Bray–Curtis dissimilarity: asthma/allergy phenotypes explained approximately 1% of variability in microbial community composition for all three beta-diversity measures. After adjustment for sequencing run, geographical sampling site and shared twin effects, the amount of variability was reduced to 0.5–0.8%, with marginal significance for the Bray–Curtis dissimilarity, but no statistical significance for the UniFrac distances, providing little or no evidence for systematic differences in species composition between asthma phenotypes. The largest explanatory factor for microbial composition, regardless of diversity measure, was twin pair (R2 = 66–71%, all p = .001). In contrast, geographical sampling site and sequencing run jointly explained less than 5% of variability. Running ANCOMBC7 on the quality-filtered data with an extra 50% prevalence filter provided inference for 1969 species, corresponding to 93% of all quality-filtered reads. In total, 201 of the 1969 assessed taxa (10%) were statistically significantly differentially abundant (DA) between the healthy control group and at least one of the three exposure levels at an FDR level ≤1%. The majority of DA species were specific to a single exposure group (n = 175; 87%), i.e. asthma only (n = 68; 34%), IgE+ allergic (n = 34; 17%) and allergic asthma groups (n = 73; 36%). Little overlap was found between the asthma-only and IgE+ allergic groups (n = 4; 2%), while allergic asthma shared slightly more DA taxa with both the asthma-only and IgE+ allergic groups (n = 9; 4.5% and n = 13; 6.5%, respectively). Most of the overlap between all three exposure groups was concentrated in the family Streptococcaceae, Figure 1A. We found that the eight most common microbial families among all DA taxa covered 77% of DA taxa, compared to 70% for taxa DA for asthma alone, and 82% for IgE+ allergy and allergic asthma DA taxa. Similarly, the four most common families (Streptococcaceae, Lachnospiraceae, Ruminococcoceae and Bacteroidaceae) alone already accounted for 68% (pooled across exposures), 56% (asthma alone), 75% and 77% (IgE+ allergic and allergic asthma, respectively) of all taxa DA in their category. However, the actual distribution of families within each exposure level differed widely, with the proportions of DA Streptococcaceae and Lachnospiraceae taxa more similar between the IgE+ allergic- and allergic asthma groups, and the proportion of DA Ruminococcoceae and Bacteroidaceae more similar between the asthma-alone- and IgE+ allergic groups. The overall most common family of DA taxa, Streptococcaceae, was very low-abundant in the raw quality-filtered data, accounting for only 0.13% of reads across samples, whereas the other three common families of DA taxa represent the three most common taxa in the raw quality-filtered data (Figure 1B). We mapped microbial families and statistical significances of the 201 DA taxa to their phylogenetic tree, Figure 1C. DA Streptococcacea taxa were phylogenetically closely related, with most taxa statistically significantly DA for allergic asthma, followed by the IgE+ allergic group and the asthma-only group. In contrast, DA Lachnospiraceae taxa showed more phylogenetic diversity, and were mostly DA for asthma-only, followed by allergic asthma and the IgE+ −allergic group. Ruminococcoceae taxa were mostly DA for asthma-only and the IgE+ allergic group, with some phylogenetic separation between the two groups of taxa. Species from Bacteroidaceae were mostly DA for allergic asthma, followed by asthma-only and IgE+ allergic. We found little systematic evidence for non-additive contributions from raised IgE and non-allergic asthma to DA in allergic asthma. In conclusion, we found no evidence for differences in the overall gut microbial richness or composition between healthy controls and children diagnosed with allergic asthma, asthma only or IgE sensitization only. The combined activity of raised IgE and asthma on the differential abundance in allergic asthma was almost exclusively additive, with no systematic evidence for synergistic or antagonistic effects. Additional information on this study is published on Zenodo: https://doi.org/10.5281/zenodo.7998496. AP, CA, JWD, PKEM and MM were involved in the conception and design of the study. ISK, FB and JWD were involved in the sequencing of faecal samples and bioinformatics processing of the data. AP, CA, JWD, PKEM and MM were involved in planning, and ISK, CA and AP supervised the work. JWD, FB and AP aided in interpreting the results. CA obtained funding for the project. AP performed the analysis and designed the figures. MM drafted the manuscript and all co-authors reviewed and revised the manuscript. All authors commented on the manuscript. We acknowledge The Swedish Twin Registry for access to the data used in this research project. The Swedish Twin Registry is managed by Karolinska Institutet and receives funding through the Swedish Research Council under the grant no 2017-00641. We thank the Karolinska Institutet Biobank for sample storage and collaboration. We are also thankful to Marica Hamsten and Alexandra Pennhag at the Centre for Translational Medicine and Research. The Center for Translational Microbiome Research, Department of Microbiology, Tumour and Cell Biology, Karolinska Institutet, is partly funded by Ferring Pharmaceuticals. The other authors report no conflict of interest. Obtained from the Swedish Research Council (number 2017-00641; 2018-02640), the Heart-Lung Foundation, the Swedish Asthma and Allergy research foundation and Stiftelsen Frimurare Barnhuset in Stockholm, Sweden. All data were pseudonymised prior to analysis. Study participants were not involved in the planning or writing of this research project. Ethical approval for this project was obtained from the Swedish Ethical Review Authority with diary number 2010/1336-31/3. The authors assert that all the procedures contributing to this work comply with the ethical standards of the relevant National and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Informed consent for the study was obtained from all the participants and their parents. Due to ethical and data integrity concerns, individual-level data cannot be made openly available. Further information about the data and conditions for access are available at the Swedish Twin Registry, Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Sweden. https://ki.se/en/research/the-swedish-twin-registry