Infant gut microbiota and the hygiene hypothesis of allergic disease: impact of household pets and siblings on microbiota composition and diversity

In recent decades, dozens of studies from around the world have reported that early-life exposure to pets or older siblings is protective against the development of allergic disease [1–4], and these associations are commonly attributed to the “hygiene hypothesis”. In its original form, this hypothesis claims that limiting early-life infection impedes natural immune system development and causes predisposition to allergic disease [5]. The modified “microflora hypothesis” proposes that, rather than specifically limiting infection, the overly hygienic Western lifestyle limits general microbial exposure and alters the colonization of the infant gut, which in turn disrupts development of the immune system and ultimately leads to allergic disease [6].

In support of the microflora hypothesis, reduced gut microbiota diversity during infancy has been associated with allergic disease later in childhood [7–13]. One could speculate that this is a function of fewer microbes in the home environment; however, human and piglet studies indicate that greater microbe diversity in the environment actually leads to reduced diversity of the gut microbiota [14, 15]. In some cases, the prevalence or relative abundance of specific organisms has been associated with atopic outcomes. For example, early-life colonization by Clostridium difficile reportedly increases risk for childhood wheeze, eczema and asthma [16], whereas certain Bifidobacteria and Lactobacilli are considered protective [8, 17–19]. It therefore remains unclear whether diversity is paramount, or whether gut microbiota composition is more important for immune system development and avoidance of atopic disease.

Household pets and siblings are thought to protect against atopy by increasing contact with environmental microbes during early life, thereby promoting a diverse and healthy gut microbiota. While associations of pets and siblings with atopy have been reported (and debated) for decades [1–5, 20, 21], few studies have directly evaluated the gut microbiota as a mediator. There is only limited evidence that contact with siblings influences the gut microbiota [22, 23], and no studies have evaluated the role of pets.

In this study, we aimed to determine the impact of two early-life exposures traditionally associated with the hygiene hypothesis, pets and siblings, on the diversity and composition of infant gut microbiota.

Using a comprehensive and culture-independent approach, we have characterized the impact of household pets and siblings on the composition and diversity of infant gut microbiota. Our findings suggest that these two traditionally protective “hygiene hypothesis” factors exert distinct effects on microbiota diversity, and select for different assemblages of gut microbes (Figure 3). These results imply that the “microflora hypothesis” of allergic disease is likely a function of complex changes in microbiota composition, rather than simplified variations in overall diversity.

Although no clear consensus has been reached, many studies have reported inverse associations between early-life pet exposure and allergic disease. In a questionnaire study of over 2000 Swedish children, exposure to pets during the first year of life was associated with a lower prevalence of allergic rhinitis and asthma at school age [2]. Similarly in an American prospective birth cohort, exposure to 2 or more dogs or cats in the first year of life was found to reduce subsequent risk of allergic sensitization to multiple allergens during childhood [1]. Whether these associations can be attributed to changes in the infant gut microbiota remains to be seen. Sharing of specific microbes between pet and owner has been occasionally reported in adults [30–32], but colonization of the infant gut has rarely been studied in relation to pets. Using targeted methods to identify specific organisms, it has been reported that the presence of household pets does not influence colonization by Lactobacillus, E. coli, or C. difficile[23, 33, 34]. Ours is the first comprehensive profiling study to address the impact of pets on the infant gut microbiome as a whole.

Similar to household pets, the presence of older siblings has been associated with a reduced risk of atopic disease. Indeed, it was David Strachan’s landmark paper on the inverse association between hay fever and household size (especially older siblings) that inspired the hygiene hypothesis [5]. Similar associations have since been reported around the world for both asthma [2] and food allergy [21]. Although frequently suggested as the underlying mechanism for these associations, there is only limited evidence that contact with siblings alters the infant gut microbiota. Using targeted or culture-based methods, others have shown that infants with older siblings have slightly higher numbers of Bifidobacteria [23], delayed colonization by Clostridium species [22], and increased colonization by Staphylococcus aureus[35], with no difference in C. difficile colonization [28]. Until now, no study has reported the impact of siblings on the overall community structure of the gut microbiota.

Thus, despite the vast amount of evidence that household pets and siblings protect against allergic disease, and the widespread consensus that microbial exposure underlies these associations, ours is the first study to directly assess the impact of pets and siblings on the community composition and diversity of infant gut microbiota. We observed that while microbiota richness and diversity tended to be increased in infants living with pets, the opposite was true for infants with older siblings. Pets and siblings also appeared to select for distinct assemblages of gut microbes, possibly because they harbor and transmit different organisms (e.g. microbes from the outdoor environment, schools or daycares, and organisms from the normal animal or human microbiota). Transmission of microbes to the infant could occur through direct contact, or through intermediate reservoirs in the home environment (such as house dust or the maternal microbiome) [36, 37]. Also, in contrast to a previous study [28], we found that C. difficile tended to be more prevalent among infants living with pets, and less prevalent among infants with siblings. Despite these apparently opposing influences, both pets and siblings are presumably protective against allergic disease (although this remains to be confirmed in our cohort, once the participants are old enough to evaluate atopic outcomes).

Collectively, this evidence suggests that simplified measures of microbiota diversity may be insufficient or even misleading, for the purposes of evaluating atopic disease risk and related biological mechanisms. Indeed, conflicting reports have indicated that both low and high infant gut microbiota diversity are associated with atopic eczema [7, 38], illustrating the limited clinical value of these measures. The observation that two protective exposures apparently select for different combinations of microbes indicates that multiple “healthy” microbiota profiles are possible, perhaps due to functional redundancy among different organisms. As others have suggested, this functional redundancy may obviate the need for “keystone species” and high microbial diversity [15, 39].

As the microbiota of domestic dogs, cats and other companion animals continue to become better defined by modern methods [40–42], the ability to resolve their microbial contributions to other microbial communities will improve. Because of the great degree of overlap at the genus-level between human- and pet-associated microbiota, sequencing approaches that yield species-level or even strain-level resolution are ultimately needed to differentiate these communities. In the meantime, algorithms to establish robust associations have already been demonstrated using sequence data of coarser taxonomic resolution [43–45]).

The major strengths of this study are the application of modern high-throughput sequencing technology to profile the gut microbiota of healthy infants, and the prospective collection of detailed exposure data. Our combined use of 16S rRNA sequencing and targeted qPCR analysis permitted the integration of reliable species-level data into the framework of comprehensive, culture-independent microbiome analysis. In addition, our finding that cats and dogs seem to favour enrichment of different organisms in the gut microbiota indicates that “pet exposure” is complex and should be defined as specifically as possible in population health studies in order to support future analyses.

The main limitation of this study is sample size; despite being relatively large among infant gut microbiota studies (many of which have comprised fewer than 15 infants [46–52]), we had insufficient power to detect differences according to type of pet or number of siblings. Also, we could not adjust for the simultaneous effects of multiple exposures; since the majority (>80%) of infants without siblings in this cohort happened to live with pets, it is possible that pet exposure may be driving the sibling-associated differences observed in the current study. Further research is required to establish the independent and combined effects of these and other exposures, including potential confounders such as antibiotic use, caesarean delivery or breastfeeding [53, 54]. These deficiencies will be addressed in our ongoing research within the full CHILD cohort, where longitudinal sampling and clinical evaluation of over 2000 infants is anticipated, and multivariate analyses will be possible.

In summary, this study provides new evidence that early-life exposure to pets and siblings may influence the overall diversity and composition of the infant gut microbiota, with potential implications for the development of atopy (Figure 3). Allergic diseases have reached epidemic proportions in recent decades, and our research supports the hypothesis that modifiable environmental exposures may be responsible, through their impact on the gut microbiota. Our findings also emphasize the complexity of these associations, and underscore the need for further research.