Oral microbiota reveals signs of acculturation in Mexican American women

Abstract

The oral microbiome has been linked to a number of chronic inflammatory conditions, including obesity, diabetes, periodontitis, and cancers of the stomach and liver. These conditions disproportionately affect Mexican American women, yet few studies have examined the oral microbiota in this at-risk group. We characterized the 16S rDNA oral microbiome in 369 non-smoking women enrolled in the MD Anderson Mano a Mano Mexican American Cohort Study. Lower bacterial diversity, a potential indicator of oral health, was associated with increased age and length of US residency among recent immigrants. Grouping women by overarching bacterial community type (e.g., "Streptococcus," "Fusobacterium," and "Prevotella" clusters), we observed differences across a number of acculturation-related variables, including nativity, age at immigration, time in the US, country of longest residence, and a multi-dimensional acculturation scale. Participants in the cluster typified by higher abundance of Streptococcus spp. exhibited the lowest bacterial diversity and appeared the most acculturated as compared to women in the "Prevotella" group. Computationally-predicted functional analysis suggested the Streptococcus-dominated bacterial community had greater potential for carbohydrate metabolism while biosynthesis of essential amino acids and nitrogen metabolism prevailed among the Prevotella-high group. Findings suggest immigration and adaption to life in the US, a well-established mediator of disease risk, is associated with differences in oral microbial profiles in Mexican American women. These results warrant further investigation into the joint and modifying effects of acculturation and oral bacteria on the health of Mexican American women and other immigrant populations. The oral microbiome presents an easily accessible biomarker of disease risk, spanning biological, behavioral, and environmental factors.

Fig 1. Core oral genera of Mexican American women (Houston, TX, 2004–2011) exhibit wide variation in relative abundance.

(A) Box plots indicate log relative abundance of 18 core taxa, defined as those genera detected in ≥ 98% of study subjects. (B) Stacked bar plot shows the contribution of all core genera across each study participant.

Fig 2. Beta diversity analysis indicates overlapping yet distinct bacterial community clusters as determined by DMM modeling.

Bray-Curtis dissimilarity distance varied significantly by DMM cluster (P<0.01) and was visualized using principal coordinates analysis. Lines connect samples to cluster centroids. DMM, Dirichlet multinomial mixtures.

Fig 3

LEfSe analysis indicates differentially abundant taxa (A) and Tax4Fun-imputed functional pathways (B) by DMM cluster. Clusters were named for the single OTU with the greatest effect size. LEfSe analyses were conducted using α = 1E-05 and minimum LDA score = 2.5. Due to space limitations, only taxa with a minimum LDA score of 3.0 are shown. A complete list of taxa with minimum LDA score of 2.5 is included in supporting information. DMM, Dirichlet multinomial mixtures; LDA, linear discriminant analysis; LEfSe, LDA effect size.