The human gut contains a complex and diverse ecosystem consisting of hundreds of microbial species that are acquired during the first years of life (van Best et al., 2015). Although a myriad of bacterial species have been studied within the human infant gut, important colonizers that are often disregarded are the methanogenic archaea (Horz, 2015). At present, five methanogenic archaea species and two halophilic archaea have been isolated from human feces from which only the Methanobrevibacter smithii (M. smithii), Methanosphaera stadtmanae (M. stadtmanae), and Methanomassiliicoccus luminyensis (M. luminyensis) have been detected more than once (Miller et al., 1982; Miller and Wolin, 1985; Dridi et al., 2012a; Khelaifia et al., 2013, 2017; Khelaifia and Raoult, 2016). A previous study from Dridi et al. showed that the most dominant archaeal gut inhabitant is M. smithii with a prevalence of 88% in children of 0–10 years of age. In contrast, M. stadtmanae and M. luminyensis tend to colonize the child's gut less frequently with a prevalence of 11 and 1%, respectively (Dridi et al., 2012b).

Upon colonization, methanogenic archaea are responsible for producing the majority of methane in the gut by reducing carbon dioxide into methane in the presence of hydrogen (Roccarina et al., 2010). The hydrogen in the gut is mainly the result of bacterial fermentation, and accumulation of hydrogen subsequently inhibits this process of breaking down food components for energy. Therefore, reduction of hydrogen levels by methanogens stimulates food fermentation by saccarolytic bacteria (Horz and Conrads, 2010; Gaci et al., 2014). On the other hand, colonization by archaea has been suggested to be potentially detrimental for host health due to alterations in gut metabolism and syntrophic interactions with other microbes (Cavicchioli et al., 2003; Conway de Macario and Macario, 2009; Nakamura et al., 2010; Roccarina et al., 2010; Gill and Brinkman, 2011). For instance, in previous studies higher levels of archaea and excreted methane were found in patients with gastrointestinal and metabolic diseases such as colorectal cancer, inflammatory bowel disease, irritable bowel syndrome, constipation, and obesity (Haines et al., 1977; Pimentel et al., 2003; Kim et al., 2012; Blais Lecours et al., 2014; Triantafyllou et al., 2014; Mbakwa et al., 2015; Vandeputte et al., 2016).

Methanogenic archaea have been shown to be absent during infancy while omnipresent in school-aged children and their presence seems to increase with age (Dridi et al., 2012b). The latter suggests that colonization may result either through exposure to sources of archaea during childhood, or through factors shaping the gastrointestinal ecophysiology to make the gut more favorable for archaeal colonization during this time period. The factors that determine the acquisition of methanogenic archaea, however, have remained undefined. Although the rumen of beef cattle have been shown to be a carrier for M. smithii and M. stadtmanae (Carberry et al., 2014), these human gut colonizers have not been identified in selected food products so far (Brusa et al., 1998). Moreover, no study has conducted a comprehensive analysis on potential lifestyle and dietary determinants of human gut colonization by these methanogenic archaea. Therefore, we aimed to explore a wide variety of potential determinants associated with the acquisition of the two main archaeal species, M. smithii and M. stadtmanae, in children. To this end, we used extensive data on determinants prospectively gathered through repeated questionnaires within the KOALA Birth Cohort Study in combination with data on presence and abundance of M. smithii and M. stadtmanae obtained from fecal samples in children of 6–10 years.

Fecal samples from 472 children with a median age of 7.2 (6.0–12.0) years were analyzed. Of these, 369 (78.2%) out of 472 were colonized by M. smithii, while 39 (8.3%) out of 472 children were colonized by M. stadtmanae. Due to the low colonization rate of M. stadtmanae, analyses on its abundance were not performed in this study. Supplementary Table 2 shows participant characteristics of the KOALA Birth Cohort Study (n = 2834) and the study population (n = 472) for all selected determinants. In general, the distribution of participant characteristics of the study population was similar to the population of the entire KOALA Birth Cohort Study.

To the best of our knowledge, this is the first large-scale explorative study investigating the associations between a wide range of potential determinants and intestinal colonization by methanogenic archaea in school-aged children. We found that consumption of organic yogurt and milk products is related to colonization of M. smithii in particular. Moreover, we identified M. smithii in raw and processed milk, which could therefore be considered as potential sources for archaeal colonization.

The difference in metabolic activity might be a possible explanation for the large difference in the prevalence of the two main intestinal archaea. M. smithii is responsible for producing the majority of methane in the gut by reducing carbon dioxide using hydrogen as the primary electron donor. Alternatively, M. smithii can use formate as a direct substrate to produce methane (Roccarina et al., 2010). In contrast, M. stadtmanae is only able to produce methane from hydrogen (Gaci et al., 2014). Moreover, colonization by M. smithii and M. stadtmanae might originate from different environmental sources, which could explain their different prevalence. Factors that play a role in the acquisition of methanogenic archaea are largely unknown. However, the environment is assumed to play a crucial role in gut colonization. This view is strengthened by a twin study that indicated that not genetic factors, but shared and unique environmental factors were of importance for the occurrence of methanogens in humans (Florin et al., 2000). The prospective design and long-term follow-up allowed us to examine a wide range of environmental, dietary and life-style associated determinants throughout childhood. Out of the determinants present during infancy, caesarean section compared to vaginal delivery at home, first exposure to antibiotics in the 2nd year of life compared to no antibiotic treatment during the first 2 years and organic fruit consumption were associated with increased odds of colonization by M. stadtmanae. These findings on M. stadtmanae in the present study need however to be interpreted with great care. Due to the small group of children colonized by M. stadtmanae (n = 39), models for M. stadtmanae were prone to instability. Therefore, these factors need to be verified in other studies with, preferably, larger sample sizes. In particular, the association with birth mode is in contrast with previous findings that intestinal colonization by methanogenic archaea in neonates and infants is very rare (Palmer et al., 2007). However, it could be postulated that the pioneer species associated with caesarian section delivery drive a subsequent colonization pattern that is more favorable for the establishment of M. stadtmanae at a later stage. As such, it is important to study the co-occurrence and co-exclusion of archaea and bacterial taxa in future studies, in order to identify microbial networks that might favor archaeal colonization.

Although one previous study found females to be more often colonized by methanogenic archaea than males (Florin et al., 2000), the current view is that there is no association between methanogenic archaea and gender (Dridi et al., 2011b). In this study, we assessed fecal samples of 244 (51.7%) males and 228 (48.3%) females. Of all males under study 187 (76.6%) were colonized by M. smithii, while 182 (79.8%) females were colonized by M. smithii. Furthermore, 19 (7.8%) males were colonized by M. stadtmanae. This was similar to the colonization by M. stadtmanae in females, as 20 (9.6%) females were colonized. Moreover, we also did not find an association between gender and either M. smithii or M. stadtmanae.

So far, only one study assessed potential carriers of archaea in food, but yogurt and milk have not been assessed herein (Brusa et al., 1998). The present study suggests an association between dietary factors and archaeal colonization, as organic yogurt and milk were significantly associated with the presence of M. smithii. Due to the high correlation between organic milk and yogurt intake, we were however not able to disentangle whether only organic milk or organic yogurt or both were truly responsible for the association with M. smithii during multivariable analyses. Cheese was however not associated with M. smithii colonization, which is in line with previous findings where no M. smithii or M. stadtmanae were detected in different types of cheese (Brusa et al., 1998). Analyses on the number of dairy products consumed, used to address multicollinearity, indicated that the consumption of multiple different dairy products might be associated with an increased M. smithii presence.

A well-known carrier of M. smithii and M. stadtmanae is the rumen of beef and cows (Carberry et al., 2014; Cersosimo et al., 2016). Therefore, it is likely that products derived from cows, such as dairy products, may contain some of these taxa, which was reflected in our results. Moreover, these specific methanogenic archaea have also been found in soil which could be the route of origin to cows as revealed from the Integrated Microbial Next Generation Sequencing database (Lagkouvardos et al., 2016). The archaeal presence in soil might be the underlying reason that especially organic products were associated with colonization. Organic cows have more outdoor access compared to conventionally farmed cows and could therefore have increased archaea uptake. Additionally, drugs for organic cows are less prescribed whereas conventional cows more often receive antibiotics to prevent microbial infections (Zwald et al., 2004). Methanogens are susceptible to antibiotics such as bacitracin, a commonly used antibiotic in cattle, which might eliminate them (Dridi et al., 2011a). Although the screening of multiple dairy products indicated the presence of M. smithii in both conventional and organic dairy products, the latter could alternatively explain the lack of association of consumption of conventional dairy products with M. smithii colonization within our cohort study.

As the culture-independent techniques applied in the present study do not distinguish viable microbes from cell-free DNA originating from lysed microorganisms, we cannot completely exclude that dairy products only contain archaeal DNA instead of viable archaea. However, we used the concentrated pellet of the microbial cells for downstream analyses of milk products, thereby minimizing the detection of circulating cell-free DNA. It is therefore likely that a living fraction of archaeal cells has been measured.

For milk, the typical neutral pH is viable for both M. smithii and M. stadtmanae, which favor an optimal pH of 6.9–7.4 whereas yogurt has a more acidic environment (Miller and Lin, 2002; Ledenbach and Marshall, 2009; Dridi et al., 2011b), which might explain why we could not detect methanogenic archaea in yogurt samples. In addition, methanogens have been discovered in a wide variety of extreme environments with temperatures until at least 100°C (Elias et al., 1999; Tung et al., 2005). The latter might indicate that these microbes could resist heat treatments and pasteurization performed for milk products. All in all, this strengthens the possibility that viable methanogenic archaea could indeed be present in milk products, even after thermal processing.

To get definite prove that milk products are a source of archaea, future studies using either culture-based methods or molecular methods that enable differentiation between intracellular and cell-free DNA, such as the use of propidium monoazide (PMA) as a membrane impermeable DNA intercalating dye (Janssen et al., 2016), are warranted. Several studies already identified living bacteria with an implemented PMA-assay in both human feces and processed milk (Bae and Wuertz, 2009; Soejima et al., 2012; Quigley et al., 2013; Cangelosi and Meschke, 2014).

In conclusion, dairy products, in particular organic milk products, may play an influential role in the colonization of the gut by M. smithii in children. Moreover, M. smithii seems to be present in both raw and commonly consumed milk products. For the first time, a large variety of determinants have been explored in association with gut colonization by methanogenic archaea. Although more information is needed to confirm and unravel the mechanisms of archaeal colonization in more detail, it may provide new targets for prevention of diseases associated with the presence of methanogenic archaea.

Jv performed the analyses on the cohort, and wrote partly the manuscript together with Nv. Nv designed the laboratory experiments, performed laboratory isolation, and qPCR techniques on milk products, analyzed the genomic data, and wrote partly the final manuscript as submitted. CM performed the laboratory isolation of fecal DNA and qPCR techniques, data processing, and approved the final manuscript as submitted. JP contributed to the design of the study and collection of the data, supervised the laboratory work, interpreted the data, critically reviewed the manuscript, and approved the final manuscript as submitted. MM contributed to the design of the study and collection of the data, interpreted the data, critically reviewed the manuscript and approved the final manuscript as submitted. CT and IA contributed to the design of the study and collection of the data, critically reviewed and revised the manuscript and approved the final manuscript as submitted. PS and MH critically reviewed and revised the manuscript and approved the final manuscript as submitted.

This work was supported by TI Food and Nutrition, public private partnership in food and nutrition research. The mission is to contribute to optimum human nutrition, food safety, and sustainable food production and to increase the competitiveness of the food industry. Partners are key players in the global food industry, leading research institutes, universities, and medical centers. Additional funding for data collection was received from: Netherlands Organization for Health Research and Development (grant 2100.0090), Netherlands Asthma Foundation (grants 3.2.03.48, 3.2.07.022), Netherlands Heart Foundation (grant 2008B112), Triodos Foundation, Phoenix Foundation, Raphaël Foundation, Iona Foundation, Foundation for the Advancement of Heilpedagogiek, Royal Friesland Foods (currently FrieslandCampina), Netherlands Sugar Foundation, and the Ministry of Economic Affairs. The sponsors had no influence on the analysis and reporting of this study. This work has also been supported by the Deutsche Forschungsgemeinschaft (Ho2236/8-1 and Priority Program 1656 to MH) and by ZonMW/JPI HDHL Intestinal Microbiomics (grant 50-52905-98-599 to JP).