Nepali oral microbiomes reflect a gradient of lifestyles from traditional to industrialized

Abstract

Background: Lifestyle plays an important role in shaping the gut microbiome. However, its contributions to the oral microbiome remains less clear, due to the confounding effects of geography and methodology in investigations of populations studied to date. Furthermore, while the oral microbiome seems to differ between foraging and industrialized populations, we lack insight into whether transitions to and away from agrarian lifestyles shape the oral microbiota. Given the growing interest in so-called 'vanishing microbiomes' potentially being a risk factor for increased disease prevalence in industrialized populations, it is important that we distinguish lifestyle from geography in the study of microbiomes across populations.

Results: Here, we investigate salivary microbiomes of 63 Nepali individuals representing a spectrum of lifestyles: foraging, subsistence farming (individuals that transitioned from foraging to farming within the last 50 years), agriculturalists (individuals that have transitioned to farming for at least 300 years), and industrialists (expatriates that immigrated to the United States within the last 20 years). We characterize the role of lifestyle in microbial diversity, identify microbes that differ between lifestyles, and pinpoint specific lifestyle factors that may be contributing to differences in the microbiomes across populations. Contrary to prevailing views, when geography is controlled for, oral microbiome alpha diversity does not differ significantly across lifestyles. Microbiome composition, however, follows the gradient of lifestyles from foraging through agrarianism to industrialism, supporting the notion that lifestyle indeed plays a role in the oral microbiome. Relative abundances of several individual taxa, including Streptobacillus and an unclassified Porphyromonadaceae genus, also mirror lifestyle. Finally, we identify specific lifestyle factors associated with microbiome composition across the gradient of lifestyles, including smoking and grain source.

Conclusion: Our findings demonstrate that by controlling for geography, we can isolate an important role for lifestyle in determining oral microbiome composition. In doing so, we highlight the potential contributions of several lifestyle factors, underlining the importance of carefully examining the oral microbiome across lifestyles to improve our understanding of global microbiomes.

Keywords: Nepali populations; Oral microbiome; lifestyle; oral microbiota; salivary microbiome.

Figure 1:. Sampling locations of all Nepal- and US-based Populations

A) Locations of the populations sampled in Nepal and the US. The US populations are specifically from the Northern California region. Location of Kathmandu is indicated in red on the Nepal map. Colors correspond to lifestyle groupings as described in (B). B) Oral microbiome samples were collected from individuals that span a spectrum of lifestyles, from nomadic foraging populations (dark blue), to populations that recently transitioned from foraging to small scale agriculture (teal), to established small scale agriculturalists (sky blue), to Nepali expats residing in the US practicing an industrialized lifestyle (peach), to American Industrialists (red). Sample sizes for each lifestyle category are indicated.

Figure 2:. Alpha diversity does not significantly differ by lifestyle in Nepal.

Faith’s phylogenetic diversity (Faith’s PD - left) and Shannon alpha diversity (Shannon Diversity - right) shown for all individuals, grouped by lifestyle. Lifestyles are ordered from most traditional (Foragers) to most industrialized (American Industrialists), left to right. No significant difference detected across lifestyles for Shannon alpha diversity (p = 0.8, Kruskal-Wallis), but a marginally significant difference detected for Faith’s phylogenetic diversity (p = 0.028, Kruskal-Wallis). Notably, those significant differences only occur between the American Industrialists and other lifestyle groups, not between Nepali individuals residing in Nepal or the US. Significant differences (p < 0.05) between specific populations are indicated (*).

Figure 3:. Oral microbiome composition significantly differs based on lifestyle.

A) Microbiome composition varies significantly with lifestyle (p = 2.3×10⁻⁴, PERMANOVA). The PCoA plot shows individuals ordinated based on Bray-Curtis distance and colored by lifestyle. B) The distribution of individuals along PCoA axis 1 follows the lifestyle gradient, from traditional to Industrial (p = 0.0014, Jonckheere-Terpstra test). Lifestyles are ordered from most traditional (Foragers) to most industrialized (American Industrialists), left to right.

Figure 4:. Abundances of nine genera significantly follow the lifestyle gradient.

The relative abundances of nine genera significantly follow the lifestyle gradient via a Jonckheere-Terpstra test followed by Benjamini-Hochberg correction (adjusted p < 0.05). Lifestyles are ordered from most traditional (Foragers) to most industrialized (American Industrialists), left to right. All taxa have been log10-transformed for visualization purposes. Taxa marked with * are also significantly differentially abundant across lifestyles based on ALDEx2. Most taxa tend to decrease in relative abundance as the lifestyles transition from more traditional to industrial, while the abundance of Atopobium increases.

Figure 5:. Alcohol, smoking, location, sisnu, and grain type are associated with the oral microbiome

A) There are significant associations between lifestyle factors and the microbiome as observed via CCA. Points represent individuals and color represents corresponding lifestyle, with ellipses around each population. Red arrows represent the lifestyle factors significantly associated with the microbiome. A total of 15 lifestyle factors were inputted into the CCA model based on contribution to each CA axis. B) Brachymonas is significantly associated with grain type consumed (padj = 0.0219). Specifically, relative abundance is higher in individuals who report barley and maize consumption compared to rice and wheat. Taxa were log10 transformed for visualization. C) Several specific lifestyle factors are associated with individual oral genera. Significant associations based on a nominal p-value threshold are indicated with *. The significant association based on an adjusted p-value threshold is indicated with **.

Figure 6:. Differentially abundant taxa are highly connected in the oral microbiome cooccurrence network

A) The SparCC module in the SpiecEasi package was used to generate a network from 111 genera. Network of 37 nodes with at least one significant edge is shown, with 6 co-abundance groups (CAGs) indicated by node color. Labeled nodes indicate genera that were identified as significantly differentially abundant across lifestyles. B) Proportions of CAGs vary across lifestyle. Specifically, CAG1 decreases with industrialization, whereas CAG2 increases with industrialization.

Figure 7:. Correlation between the oral and gut microbiomes within an individual strengthens with agrarianism

Microbiome dissimilarity (as measured by Bray Curtis dissimilarity) between the gut and oral microbiomes within an individual decreases across the gradient of lifestyles from traditional to agrarian for individuals residing in Nepal, albeit not significantly (p = 0.11; Jonckheere-Terpstra test).