Bovine milk intake has been associated with various disease outcomes, with modulation of the gastro-intestinal microbiome being suggested as one potential mechanism. The aim of the present study was to explore the oral microbiota in relation to variation in self-reported milk intake. Saliva and tooth biofilm microbiota was characterized by 16S rDNA sequencing, PCR and cultivation in 154 Swedish adolescents, and information on diet and other lifestyle markers were obtained from a questionnaire, and dental caries from clinical examination. A replication cohort of 31,571 adults with similar information on diet intake, other lifestyle markers and caries was also studied. Multivariate partial least squares (PLS) modelling separated adolescents with low milk intake (lowest tertile with <0.4 servings/day) apart from those with high intake of milk (≥3.7 servings/day) based on saliva and tooth biofilm, respectively. Taxa in several genera contributed to this separation, and milk intake was inversely associated with the caries causing Streptococcus mutans in saliva and tooth biofilm samples by sequencing, PCR and cultivation. Despite the difference in S. mutans colonization, caries prevalence did not differ between milk consumption groups in the adolescents or the adults in the replication cohort, which may reflect that a significant positive association between intake of milk and sweet products was present in both the study and replication group. It was concluded that high milk intake correlates with different oral microbiota and it is hypothesized that milk may confer similar effects in the gut. The study also illustrated that reduction of one single disease associated bacterial species, such as S. mutans by milk intake, may modulate but not prevent development of complex diseases, such as caries, due to adverse effects from other causal factors, such as sugar intake in the present study.
Citation: Johansson I, Esberg A, Eriksson L, Haworth S, Lif Holgerson P (2018) Self-reported bovine milk intake is associated with oral microbiota composition. PLoS ONE 13(3): e0193504. https://doi.org/10.1371/journal.pone.0193504
Editor: Peter M. Milgrom, University of Washington, UNITED STATES
Received: November 13, 2017; Accepted: February 12, 2018; Published: March 21, 2018
Copyright: © 2018 Johansson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The sequence data underlying this study have been uploaded to figshare and are accessible using the following DOI: 10.6084/m9.figshare.5794989. Additional data underlying this study contain sensitive information and are restricted by the Swedish law. Interested researchers can send data access requests to the ethical review board at Umea University, using the following email address: email@example.com.
Funding: The study was supported by the Swedish Research Council for Health, Working Life and Welfare (FORTES), the Swedish National Research Council (VR), the County Council of Västerbotten (TUA), and the Patent Revenue Fund for Research in Preventive Odontology (PMF).
Competing interests: The authors have declared that no competing interests exist.
Milk and dairy products are important sources of energy and macro- and micronutrients, but intake varies between and within populations due to traditions and ability to digest lactose . Sweden is a country with a high prevalence of lactose tolerance  and has among the highest intake of bovine milk in the world [3, 4].
Bovine milk intake has been studied in relation to various diseases, including dental diseases, cardiovascular diseases, type 2 diabetes, and cancer [5–7]. Three independent studies from Sweden have recently reported increased mortality risk among subjects with the highest intake of non-fermented bovine milk [8, 9, 10]. Causality remains unclear but lactose exposure and modulation of microbiota have been suggested as potential mechanisms [8, 9]. The microorganisms in the gastro-intestinal (GI) tract are organised in habitat specific, resilient communities regulated by habitat and exposure influences, including pH, oxygen, nutrients, neighbouring species, anti- and probiotics and transmission and availability of attachment sites [11, 12]. In addition to nutritional components, human and bovine milk contain, a large number of bioactive proteins, glycoproteins, peptides and glycolipids with effects on inflammatory responses and cell signalling in the host and metabolism and adhesion of microorganisms in the GI canal . Human milk (breastfeeding) is known to affect establishment of the gut [13,14] and oral [15, 16] microbiota of the infant. Similarly, bovine milk and milk components have been shown to affect the gut microbiome in animals [17, 18], and reduce adhesion and metabolism of the tooth colonizing, caries causing Steptococcus mutans in vitro [19–21]. Such effects have not been evaluated in vivo in man, but it may be hypothesized that bovine milk modulates the microbiota in the mouth and other parts of the GI tract in later child- and adulthood. The primary aim of the present study was to explore the oral microbiota by DNA based characterization and culturing in relation to variations in milk intake, and, following our findings, a secondary aim was to evaluate caries status by milk intake and potential confounding factors. We found that milk intake levels were associated with distinct microbiota in saliva and tooth biofilms with strikingly less cariogenic S. mutans by increasing milk intake but found no strong association with caries status, a finding potentially confounded by sugar intake.
Subjects and methods
The study group involved 17-year-old adolescents (n = 154), who were recruited from public dental health care clinics in the city of Umeå, Sweden. The study group has been described previously . Adolescents who consented to participate, were healthy, had not taken antibiotics during the latest 6 months, and were able to communicate in Swedish or English. Study participants abstained from oral hygiene on the morning of their visit to the dental clinic. Whole chewing stimulated saliva was collected for 3 minutes, and supragingival biofilm was collected from all accessible tooth surfaces using sterile wooden toothpicks. Samples from the different tooth surfaces were pooled for the subject into 100 μl of TE buffer (10 mM Tris, 1 mM EDTA, pH 7.6). All samples were stored at -80°C.
To evaluate and possibly replicate potential findings from the study group, we included participants in the Gene-Lifestyle Interaction in Dental endpoints (GLIDE) database with information on caries prevalence and lifestyle variables from participating in a health screening program, the Västerbotten Intervention Program (VIP) . VIP is a population-based health screening and intervention program which includes anthropometric measures and questionnaire information on diet, tobacco use and education. The average recruitment rate has been around 60% of potential participants, and no indication of a systematic selection bias has been found [24, 25]. Participants aged 30–64 years at time of participation in VIP, who had completed diet questionnaires to an adequate level and had data on caries status were eligible (n = 43,788). Of these participants, those with diet information 0 to 5 years before the dental examination were included in analysis (n = 31,571).
Information on diet and other lifestyle variables
The adolescents answered a food frequency questionnaire (FFQ) on dietary habits, tobacco and alcohol use, and oral hygiene habits at their visit to the dental clinic. For the adults, diet information was obtained from the FFQ included in the questionnaire at the health screening . In both cohorts, the FFQ provided 9 intake frequency options. The FFQ used in the replication cohort has been validated , and the FFQ used for the adolescents were identical to that used in the replication cohort with addition of questions on vegetarian foods and non-sugar sweetened products. Quality filtering of the FFQ data were as described . In brief, participants were excluded from analysis if FFQ answers were incomplete, if the reported Food Intake Level (FIL) was extreme with an unrealistic ratio between estimated energy intake and need. Tobacco use [smoking or Swedish snuff (snus)] was classified as current, former or never use, tooth brushing (adolescents only) as ≥1 a day or less, education (adults only) as having an academic education or not, BMI was calculated as weight (kg)/height (m)2.
Cultivation and PCR detection of S. mutans and S. sobrinus in the study cohort
Viable mutans streptococci (S. mutans and S. sobrinus) and lactobacilli were assessed in fresh saliva from the adolescents by cultivation on mitis salivarius sucrose agar supplemented with 0.2 U of bacitracin (MSB) and Rogosa agar (Merck, Darmstadt, Germany), respectively. The plates were incubated at 37°C in 5% CO2 for 48 h. Presence of S. mutans and S. sobrinus in tooth biofilm and saliva was evaluated by PCR using the KAPA2G Robust HotStart PCR Ready Mix (2´) kit (Kapa Biosystems, Boston, MA, USA) and the primers SmF5 (5′-ACTACACTTTCGGGTGGCTTGG-3′ and SmR4 (5′-CAGTATAAGCGCCAGTTTCATC-3′) primers for S. mutans and SsF3 (5′-GATAACTACCTGACAGCTGACT-3) and SsR1 (5′AAGCTGCCTTAAGGTAATCACT-3′) primers for S. sobrinus as described previously .
DNA extraction and Illumina MISeq sequencing in the study cohort
DNA was extracted from 154 saliva and 139 tooth biofilm samples from the adolescents. The procedure for DNA extraction and sequencing, and overall findings from sequencing have been presented previously . Briefly, genomic DNA was extracted from tooth biofilm and saliva samples using the GenElute™ Bacterial Genomic DNA Kit (Sigma-Aldrich, St. Louis, MO, USA), and samples with a 260/280 ratio of the extracted DNA of ≥1.8 was used, and adjusted to 20 ng/μl by lyophilization of a fixed amount of DNA and resolving in a defined volume. Multiplex 16S rDNA amplicon sequencing was performed using the Illumina MiSeq platform (http://www.illumina.com) at the Forsyth Research Institute (Cambridge, MA, USA) using the HOMINGS protocol . The V3-V4 hypervariable regions of the 16S rRNA gene were PCR amplified using the 341F forward primer AATGATACGGCGACCACCGAGATCTACACTATGGTAATTGTCCTACGGGAGGCAGCAG and the 806R reverse primer CAAGCAGAAGACGGCATACGAGATNNNNNNNNNNNNAGTCAGTCAGCCGGACTACHVGGGTWTCTAAT. The sequences in bold-italic represent the primers and the 12 N stretch designate individual barcode sequences (http://homings.forsyth.org/protocol.html). Ultrapure water served as the negative control and three mock communities as positive controls as described previously .
Pair-end reads were combined with an overlap of 70 bp and mismatch ratio of 0.25, and barcodes, primers, and ambiguous and chimeric sequences were removed. Taxa were identified by in silico recognition of 17–40 bp species- and genus-level 16S rRNA-based oligonucleotide “probes” using ProbeSeq (a custom designed BLAST program); 638 species- and or 129 genus-level probes for closely related species . Taxa targeted by the species and genus probes (gp) are described in S1 Table and S2 Table and at http://homings.forsyth.org/index2.html.
As described previously , 9,744,159 reads with a mean read length of 426 bp were retained for 152 saliva samples (2 samples failed), and 4,829,569 reads (with an average read length of 424 bp) for the 139 tooth biofilm samples. The average (min-max) number of included sequences per sample was 64,106 (21,951–157,126) for saliva (n = 152 samples) and 34,745 (8,237–109,373) for tooth biofilm samples (n = 139). Taxa were found in 85 genera (9 phyla) in saliva and 81 genera (10 phyla) in tooth biofilm. In total, 311 and 323 species were identified in saliva and tooth biofilm, respectively. All sequences can be found at 10.6084/m9.figshare.5794989.
In the adolescents, dental caries was scored from visual and radiographic examinations . The sum of manifest, cavitated lesions and initial, non-cavitated caries in the enamel (D1) and filled (F) tooth surfaces was calculated (D1FS scores). No participants had lost any tooth due to caries, i.e. the D1FS scores corresponded to a DMFS score with M = 0.
For the adults in the replication cohort, information on caries prevalence (DMFS scores, i.e. the sum of decayed (manifest cavitated lesions), missing, and filled tooth surfaces for 32 teeth) was retrieved from electronic dental records and linked to questionnaire data using personal reference numbers, which uniquely identify Swedish citizens.
The adolescents in the study group were ranked into tertile groups and the adults in the replication group into quintile groups based on their self-reported total milk (non-fermented+fermented milk), non-fermented and fermented milk intakes, respectively. Ranking was done within sex subgroups for the adolescents and sex and 10-year age subgroups for the adults.
Continuous, normally distributed variables are presented as means with 95% confidence limits (95% CI) adjusted by general linear modelling (GLM) for potential confounders as specified in the legend or footnotes of the tables. Categorical variables are presented as percentages. Differences between group means were tested with ANOVA in GLM after appropriate covariate standardization; differences in group distributions were tested using Chi2 tests. Trends across the strata of milk consumption were assessed using the non-parametric Jonckheere-Terpstra trend test. These analyses were performed using SPSS version 24 (IBM Corporation, Armonk, NY, USA) with p-values <0.01 considered statistically significant.
For the Illumina MiSeq sequences each taxa was scored for detection (yes/no) and relative abundance (percentage of all sequences in the individual). Univariate comparisons of relative taxa abundances were done with non-parametric Mann-Whitney test and p-values were corrected for multiple comparisons by the false discovery rate (FDR). Association between bacterial taxa (in saliva and tooth biofilm) and tertiles of milk consumption was assessed using multivariate partial least squares (PLS) regression (SIMCA P+ version 12.0, Umetrics AB, Umeå, Sweden). The PLS models included the milk tertile allocation as a block of dependent variables and all taxa identified by sequencing as the independent blocks. The results are presented in PLS scatter plots for subject clustering and PLS correlation coefficients from PLS column loading plots for being in the highest versus lowest milk tertile group. Taxa with a statistically significant correlation or a correlation coefficient ≥0.1 are presented. PLS modelling was also done with lifestyle variables as the independent block. The relative importance of each x-variable in PLS scatter loading plots is expressed by variable importance in the projection (VIP) values. VIP-value ≥1.0 are considered influential and ≥1.5 as highly influential. Variables utilized for PLS regression were auto-scaled and logarithmically transformed as needed to improve normality.
Ethical approval and consent to participate
Both studies have been approved by the Regional Ethical Board, at Umeå university, Sweden with registration number: Dnr 2012-111-31M with an addendum (Dnr 2015-389-32M) for the adolescents, and Dnr 2010-387-31M and 2011-74-32M for the adults. All participants signed an informed consent to participate and that the information could be used for research. The guardians of the adolescents´ approved their participation.
Study group characteristics
The 154 participating adolescents were allocated into tertile groups based on their reported total milk, non-fermented and fermented milk intake, respectively. The mean intakes for the tertile groups are shown in Table 1 for total milk, and S3 Table for non-fermented and fermented milk. Other characteristics such as sex, tobacco use, BMI and oral hygiene behaviours were similar in the three tertiles, with no statistically significant difference between the strata of milk consumption Table 1.
Milk intake and tooth biofilm microbiome
PLS multivariate regression was used to identify if milk intake was associated with participant clustering based oral bacteria profiles. As a first step, PLS regression of bacterial taxa identified in saliva and tooth biofilm, respectively, was performed with all three tertiles of total milk intake as dependent variables (S1 Fig and S2 Fig). These models indicated that milk consumption was associated with different taxa profiles in saliva and tooth biofilms, respectively, with pronounced differences between the highest and lowest tertiles.
Taxa with the highest relative abundance in the low milk tertile and the difference against the high tertile are illustrated in Bland-Altman plots for saliva (Fig 1A) and tooth biofilm (Fig 1B). Streptococcus gp 4 recognised taxa dominated in both sample types, followed by Rothia mucilaginosa and Prevotella melaninogenica in saliva and Rothia dentocariosa in tooth biofilm. Among the indicated taxa, only S. mutans in tooth biofilm differed significantly after FDR correction. i.e., p<0.008, Less abundant taxa that differed significantly between the low and high milk tertiles in the univariate comparisons were Actinomyces sp. HOT448 for saliva, and Butyrivibrio sp. HOT080, Peptostreptococcaceae[XI][G-7] sp. HOT081, Prevotella sp. HOT472 and TM7[G-1] sp. HOT348 for tooth biofilm.
The plots show mean taxa abundances for the lowest milk intake group on the x-axis and the mean abundance difference between the lowest and highest tertile groups on the y-axis for A) saliva and B) tooth biofilm.
A PLS model restricted to adolescents in the highest and lowest tertiles, performed to search for hidden structures among taxa, separated the adolescents with the highest milk intake from those with the lowest intake by saliva (Fig 2A) and tooth biofilm (Fig 3A) taxa.
(A) PLS separation of participants based on saliva microbiota characterization by Illumina MiSeq sequencing displayed in a PLS score scatter plot. The scores t1 and t2 on the x- and y-axes are the new created variables summarizing the x variables. Red dots indicate adolescents in the lowest tertile of total milk intake, and blue dots those in the highest tertile. The two lower bar graphs list taxa with statistically significant correlation or a PLS correlation coefficient ≥0.1 for (B) being in the highest tertile of total milk intake and (C) in the lowest tertile. Colored bars are for taxa with statistically significant PLS correlation coefficients, i.e. the 95% CI does not include zero.
(A) PLS separation of participants based on tooth biofilm microbiota characterization by Illumina MiSeq sequencing displayed in a PLS score scatter plot. The scores t1 and t2 on the x- and y-axes are the new created variables summarizing the x variables. Red dots indicate adolescents in the lowest tertile of total milk intake, and blue dots those in the highest tertile. The two lower bar graphs list taxa with statistically significant correlation or a PLS correlation coefficient ≥0.1 for (B) being in the highest tertile of total milk intake and (C) in the lowest tertile. Colored bars are for taxa with statistically significant PLS correlation coefficients, i.e. the 95% CI does not include zero.
For saliva (Fig 2B and 2C), PLS revealed high and low intake of milk as associated with different profiles of species in the genera Actinomyces, Aggregatibacter, Capnocytophaga, Lachnoanaerobaculum, Leptotrichia, Prevotella, Selenomonas and Treponema. Low intake was specifically associated with more prevalent detection of single species in Alloprevotella, Campylobacter, Cardiobacterium, Centipedia, Fusobacterium, Johnsonella, Neisseria and Streptococcus (S. mutans), compared to species in the candidate division TM7 [G-1] and Bergeyella in high intake.
For tooth biofilm (Fig 3B and 3C) high and low intake of milk was associated with different profiles of species in the genera Actinomyces, Capnocytophaga, Peptostreptococcaceae [XI] [G-7], Prevotella, Selenomonas, Streptococcus, Tannerella and Treponema. S. mutans correlated significantly with being in the lowest milk tertile also in tooth biofilms. Further, low intake was specifically associated with more prevalent detection of single species in the genera Bulleida, Butyrivibrio, Fretibacterium, and Stomatobaculum compared to species in Bergeyella, Cardiobacterium, Corynebacterium, Mogibacterium, Porphyromonas, Shuttleworthia, and candidate division TM7 [G-1].
Univariate follow-up revealed that PCR analyses detected S. mutans more frequently in both saliva and tooth biofilm from adolescents with low milk intake than adolescents with higher milk intake (Table 2). PCR detected S. mutans in 20% of the tooth biofilm samples from the adolescents with the highest total milk intake compared to 45% in the middle tertile and 51% of those with the lowest intake (ptrend = 0.002; Table 2). In agreement with this, increasing milk intake was associated with decreasing trends of viable mutans streptococci in saliva (ptrend = 0.001; Fig 4A), and decreasing relative abundance of S. mutans by Illumina sequencing of both saliva and tooth biofilm samples (ptrend = 0.01, ptrend = 0.002, respectively) (Table 2). The trends were similar for non-fermented and fermented milk (S3 Table). Further, the numbers of viable lactobacilli decreased with increasing total milk intake (Fig 4B).
The plot illustrates colony forming units (CFU) per ml of whole chewing-stimulated saliva of (A) mutans streptococci, and (B) lactobacilli. Data are from culturing on MSB and Rogosa agar, respectively and are expressed as 10log counts.
Milk intake and caries prevalence in 154 adolescents
The finding that the caries related S. mutans was associated with the level of milk intake urged us to evaluate caries prevalence by milk intake. The sex and tooth brushing adjusted mean caries prevalence (D1FS) was numerically lower in the adolescents with the highest milk intake but the difference was not statistically significant (p = 0.460; Table 1). Additional adjustment for BMI did not alter the difference and the result was robust in sensitivity analyses by sex, and S. mutans infection.
Sweets snack intake by milk intake in 154 adolescents
Intake of sweet products was compared between the milk tertile groups in order to search for a potential explanation for the similar caries levels among the milk groups in spite of similar tooth brushing frequency and significantly less S. mutans infection in the highest consumption group. Higher milk intake was paralleled by higher intake of several sweet products, such as frequency of sugar containing snacks, cookies, sweet rolls, and jam/marmalade (Fig 5, Table 3). Adjustment for reported intake frequency of sweet snacking frequency attenuated the difference among D1FS scores across strata of milk consumption slightly and the difference remained non-significant with no evidence for interaction between sweet snack frequency and milk intake (Table 1 footnote).
The plot illustrates the association of self-reported sweet and non-sweet snack intake from the questionnaire, BMI, and makers of S. mutans and lactobacilli in tooth and saliva samples in a PLS model where tertile group allocation from total milk intake were the dependent variables. Variables to the left (blue) were associated with being in the highest tertile and those to the right (red) with being in the lowest tertile. The relative importance of each x-variable in the projection is expressed by a variable importance in the projection (VIP). VIP-value ≥1.0 are considered influential and ≥1.5 as highly influential. VIP-values are listed to the right.
Data are presented as means with 95% confidence limits adjusted for sex and BMI in adolescents and sex, age, BMI, and education in adults.
The interaction pattern between milk intake level, total snacking frequency and microbial taxa that differed significantly in PLS are illustrated in Circos plots (http://circos.ca/) for saliva and tooth biofilms, respectively (Figs 6 and 7).
The plot illustrates the associations between milk intake levels (low, middle, and high tertiles) and saliva taxa differing significantly by milk intake in PLS (cf Figs 1 and 2), frequency of sweet snacks, and PCR detection of S. mutans (S. m) and cultivation of mutans streptococci (log CFU ms). Taxa are scaled and listed by relative abundance with those associated with low milk intake highlighted in red and those associated with high intake in blue. The relative abundancies are illustrated by the sizes of each colour segment in the outer circle. The percentages outside the milk group segments refer to the contribution from each of the variables but variable labels are omitted.
The plot illustrates the associations between milk intake levels (low, middle, and high tertiles) and for tooth biofilm taxa differing significantly by milk intake in PLS (cf Figs 1 and 2), frequency of sweet snacks, and PCR detection of S. mutans (S. m) and cultivation of mutans streptococci (log CFU ms). Taxa are scaled and listed by relative abundance with those associated with low milk intake highlighted in red and those associated with high intake in blue. The relative abundancies are illustrated by the sizes of each colour segment in the outer circle. The percentages outside the milk group segments refer to the contribution from each of the variables but variable labels are omitted.
Characteristics of 31,571 adults in the replication cohort
To evaluate a) the lack of association between milk intake and dental caries and b) the positive association between intake of milk and sweet products we searched for replication in a larger cohort of adults with information on dental caries and diet intake available. The adults were ranked into quintile groups based on their reported total milk intake while accounting for sex and age (Table 1). Similarly to the adolescents, there was a large difference in milk consumption between these groups; the quintile with the lowest intake of total milk consumed 0.4 servings per day and the quintile group with the highest intake 3.7 servings per day (Table 1). The difference of other characteristics, such as smoking status, BMI and educational level were statistically significant (p<0.01) across the quintile groups reflecting the large size of these groups rather than biological relevance.
Sweets snack intake by milk intake in 31,571 adults
The trend of increasing sweet product intake paralleling increasing milk consumption was replicated in the independent adult cohort. Intake frequencies of sugar containing snacks, cookies, sweet rolls, sugar/honey/jam, ice cream and as well as total sugar intake (g/day) increased consecutively with increasing milk quintiles (Table 3).
Caries prevalence by milk intake in 31,571 adults
In the adults, the DMFS scores (adjusted for sex, age, education, smoking, and examination year) tended to increase by increasing milk intake (pbetween groups<0.001), but with a non-significant trend (ptrend = 0.082). The difference in DMFS scores between milk consumption were minimally affected by adjustment for BMI (p<0.001), and sugar intake adjusted for energy (p<0.001) but was somewhat attenuated by adjustment for intake frequency of sweet snacks (p = 0.020, footnote Table 1). The number of missing teeth (out of 32 teeth) was slightly higher in the group with the highest milk intake (Table 1).
Dietary bovine milk consumption has been discussed in relation to health with disparate opinions on health promoting or detrimental effects [5, 30, 31]. The primary aim of the present study was to evaluate the association between bovine milk intake and oral microbiota profile, with the main finding that participants with low milk consumption (less than one serving every second to third day) had distinctly different microbiome profiles in saliva and tooth biofilms compared to those that consumed milk more frequently. For example, species in the Neisseria genus were more prevalent in saliva from low milk consumers and the caries associated S. mutans more prevalent in both saliva and tooth biofilm in the same group. Notably, the variation in milk intake had no penetrance on dental caries in the present study group, which may reflect the complex nature of the disease with S. mutans being one of several recognized risk factors including exposure to fermentable sugars.
Besides studies on the effects of probiotics in non-fermented/fermented milk  we are not aware of any in vivo study in man that describes bovine milk modulating the GI microbiome. The oral cavity is the first part of the GI canal and though there are significant differences between different parts with niche specific microbiomes , there are also similarities, such as exposure to components from the diet and the innate immune system and flushing by host secretions [34, 35]. Given that the mouth may, at least partly, serve as a model for the GI tract, it is hypothesized that bovine milk modulates the gut microbiota in line with what is shown for the mouth in the present study and for the mouth and gut for human milk [15, 36]. Such modulations may affect local metabolic events, epithelial barrier leakage and trigger inflammatory responses with further effects on health and disease [11, 37].
Similar to other parts of the GI tract, the mouth is characterized by niche specific bacterial societies determined by the availability of host receptors for bacteria attachment, nutrients, pH and oxygen conditions and competition between bacterial species . Taxa identified in tooth biofilms represent bacteria that are selected by the ability to attach to saliva coated tooth surfaces (acquired pellicle), whereas saliva mirrors bacteria that colonize epithelial surfaces in the mouth as well as teeth. These differences contribute to the taxa profiles for the two sample types studied here, with some species prevalent enough to reach detection level in both sample types, such as S. mutans, and others in one type only.
The present study found an inverse association between milk intake and S. mutans colonization, which is in line with previous publications demonstrating S. mutans adhesion and metabolism blocking effects of bovine milk proteins and peptides [19–22, 38]. In those studies, secretory-IgA, β-casein and lactoferrin (with mapped binding epitopes in the C-terminal section of β-casein and central section of lactoferrin) were responsible for blocking the attachment of S. mutans cells to saliva coated hydroxyapatite beads (a model for tooth enamel in the mouth) . Milk is a complex fluid with a panel of milk-specific bioactive components, such as the milk fat globular membrane proteins, and others in common for secretions in the mouth and gut, such as secretory-IgA, mucins, lactoferrin, lysozyme and other innate immunity peptides. Proteins and lipids in the fluid phase an on epithelial membranes provide peptide and carbohydrate epitopes for targeted attachment or blocking of free floating bacteria. Other components are bacteriostatic or even bacteriocidal. Thus, there are numerous actions by which bovine milk, similar to human milk, can affect both the oral and gut microbiome, but with the possible exception of S. mutans, we cannot speculate on what specific components may affect the prevalence of single taxa in high and low milk consumers. One emerging health associated aspect of milk is the A1 versus A2 polymorphism, with A1, with its release of ß-casomorphin opioid peptides, being considered less healthy . In the study area, milk is dominated by A1 but contains both types. It is possible that variation in milk polymorphism contributes to GI microbiome ecology but not colonization of S. mutans since the binding epitope for S. mutans in ß-casein was mapped to the C-terminal section distant from the mutation in A1 versus A2 . Thus, there is experimental support for a specific milk effect on S. mutans colonization, and it may be anticipated that this might be the case for other species too, but we cannot exclude contribution from confounding factors like other dietary components or host genetics.
Though effects of milk on dental caries was not a primary aim of the present study, the inverse association between milk intake and S. mutans colonization triggered the evaluation of caries by milk intake. Experimental studies have suggested that milk and milk components have cariostatic properties through pH neutralization , supersaturation of calcium and phosphate in the tooth interphase [41, 42, 43], and effects on bacteria, but observational studies are inconsistent . Based on these traits combined with our finding of less S. mutans by increasing milk intake, we expected high milk consumption to be associated with less caries, which was not found. Dental caries is a complex disease with the clinical symptom being tooth tissue demineralization following bacterial secretion of organic acids from sugar fermentation and a drop in local pH . We do not have an immediate explanation for the apparent lack of association between milk intake and caries development, but suggest that any potential protective effect from milk was attenuated by the higher sugar intake. This view was supported by some effect modulation when adjusting for sugar intake. Following that observational studies, such as the present, are influenced by confounding from both socio-economic and host genetic factors, the result should be followed up in a larger study group with greater power and broader monitoring of confounding factors. An alternative would be to use confounding free, milk intake associated genetic variation in lactase persistence in Mendelian randomization to avoid the obstacles in observational studies . Taken together, the present finding of a seemingly lack of effect of milk on caries status illustrates that the role of single risk factors (here, a bacterial species) cannot be easily understood for complex diseases, such as caries, obesity and cardio-metabolic conditions, without also understanding complementary causal or confounding factors .
The strengths of the present study include that microbiota sequencing was done in a comparably large group, that the variation in milk intake was wide, that the FFQ monitored the habitual intake during the latest year, and that all participants had been included in the same dental care program from 2–3 years of age. This program includes regular check-ups every 1–1.5 years and preventive measures, including oral hygiene and sugar restriction advices and topical fluoride treatments, based on the clinical status. A further strength was that part of the results was confirmed by independent methods (cultivation and PCR) and other parts were replicated in an independent study group from the same population. Still, there are some weaknesses that should be considered, including recall bias (under-reporting) in the dietary information, that dental data for the adults were available only as an aggregate score across all 32 teeth, and that the mutans streptococci levels were low in the population with approximately 30% being non-colonized and 10% with ≥105 CFU/mL saliva by cultivation. This calls for the study to be repeated in a population with higher colonization variation and then preferentially as a randomized controlled study if a suitable study group can be identified.
It was concluded that milk intake may modulate the saliva and tooth biofilm microbiota. Low intake was associated with higher prevalence of several opportunistic species, including the caries associated S. mutans. It is hypothesized that milk confers parallel effects in the gut. The study also illustrated that development of complex diseases, such as dental caries, is more elaborate than single bacteria-effects even if they are considered highly influential for the disease, as demonstrated in this study by the lack of association with dental caries due to an assumed confounding by sugar intake.
S1 Table. List of the 638 species-level targets recognizing 538 species in ProbeSeq.
Each identified species is accompanied by an Oral Taxon designation (HOT) as defined in the Human Oral Microbiome Database (HOMD) where detailed information is available. The information can also be found at http://homings.forsyth.org/bacterialtaxa.html.
S2 Table. List of the 129 genus-level targets for recognition of taxa where identification is limited to a group of several closely related species rather than individual species in the targeted region (E. coli positions ~341 to 806).
The information can also be found at http://homings.forsyth.org/bacterialtaxa.html.
S3 Table. Detection of mutans streptococci and lactobacilli in adolescents classified into tertiles based on their reported intake of total milk, non-fermented milk and fermented milk, respectively.
S1 Fig. PLS loading plot illustrating separation of the tree milk tertiles groups based on bacterial taxa identified by Illumina 16S rDNA sequencing in saliva samples.
Personnel at the Department of Biobank Research, Umeå university are acknowledged for data maintenance and the organization team for VIP for data collection.
- 1. Gerosa S, Skoet J. Milk availability: Current production and demand and medium-term outlook. In: Muehlhoff E, Bennett A, McMahon D, eds. Milk and Dairy Products in Human Nutrition. Rome: Food and Agriculture Organisation of the United Nations (FAO), 2013:11–40.
- 2. Torniainen S, Hedelin M, Autio V, Rasinpera H, Balter KA, Klint A, et al. Lactase persistence, dietary intake of milk, and the risk for prostate cancer in Sweden and Finland. Cancer Epidemiol Biomarkers Prev. 2007;16:956–61. pmid:17507622
- 3. Almon R, Sjöström M, Nilsson TK. Lactase non-persistence as a determinant of milk avoidance and calcium intake in children and adolescents. J Nutr Sci. 2013;2:e26. pmid:25191575
- 4. Hjärtåker A, Lagiou A, Slimani N, Lund E, Chirlaque MD, Vasilopoulou E, et al. Consumption of dairy products in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort: data from 35 955 24-hour dietary recalls in 10 European countries. Public Health Nutr. 2002;5:1259–71. pmid:12639231
- 5. Larsson SC, Crippa A, Orsini N, Wolk A, Michaëlsson K. Milk Consumption and Mortality from All Causes, Cardiovascular Disease, and Cancer: A Systematic Review and Meta-Analysis. Nutrients. 2015;7:7749–63. pmid:26378576
- 6. Rice BH. Dairy and Cardiovascular Disease: A Review of Recent Observational Research. Curr Nutr Rep. 2014;3:130–8. pmid:24818071
- 7. Kroenke CH, Kwan ML, Sweeney C, Castillo A, Caan BJ. High- and low-fat dairy intake, recurrence, and mortality after breast cancer diagnosis. J Natl Cancer Inst. 2013;105:616–23. pmid:23492346
- 8. Michaëlsson K, Wolk A, Langenskiöld S, Basu S, Warensjö Lemming E, et al. Milk intake and risk of mortality and fractures in women and men: cohort studies. BMJ. 2014;349:g6015. pmid:25352269
- 9. Tognon G, Nilsson LM, Shungin D, Lissner L, Jansson JH, Renström F, et al. Nonfermented milk and other dairy products: associations with all-cause mortality. Am J Clin Nutr. 2017;105(6):1502–11. pmid:28490510
- 10. Tognon G, Rothenberg E, Petrolo M, Sundh V, Lissner L. Dairy product intake and mortality in a cohort of 70-year-old Swedes: a contribution to the Nordic diet discussion. Eur J Nutr. 2017. [Epub ahead of print]
- 11. Blander JM, Longman RS, Iliev ID, Sonnenberg GF, Artis D. Regulation of inflammation by microbiota interactions with the host. Nat Immunol. 2017;18(8):851–60. pmid:28722709
- 12. Palmer J, Flint S, Brooks J. Bacterial cell attachment, the beginning of a biofilm. J Ind Microbiol Biotechnol. 2007;34(9):577–88. pmid:17619090
- 13. Pannaraj PS, Li F, Cerini C, Bender JM, Yang S, Rollie A, et al. Association Between Breast Milk Bacterial Communities and Establishment and Development of the Infant Gut Microbiome. JAMA Pediatr. 2017;171:647–54. pmid:28492938
- 14. Milani C, Duranti S, Bottacini F, Casey E, Turroni F, Mahony J, et al. The First Microbial Colonizers of the Human Gut: Composition, Activities, and Health Implications of the Infant Gut Microbiota. Microbiol Mol Biol Rev. 2017;8:81(4).
- 15. Holgerson PL, Vestman NR, Claesson R, Ohman C, Domellöf M, Tanner AC, et al. Oral microbial profile discriminates breast-fed from formula-fed infants. J Pediatr Gastroenterol Nutr. 2013;56(2):127–36. pmid:22955450
- 16. Timby N, Domellöf M, Holgerson PL, West CE, Lönnerdal B, Hernell O, Johansson I. Oral Microbiota in Infants Fed a Formula Supplemented with Bovine Milk Fat Globule Membranes—A Randomized Controlled Trial. PLoS One. 2017;18:12(1):e0169831.
- 17. Poulsen AR, de Jonge N, Sugiharto S, Nielsen JL, Lauridsen C, Canibe N. The microbial community of the gut differs between piglets fed sow milk, milk replacer or bovine colostrum. Gut Microbes. 2013;4:136–9.
- 18. Maga EA, Weimer BC, Murray JD. Dissecting the role of milk components on gut microbiota composition. Br J Nutr. 2017;117:964–78.
- 19. Vacca-Smith AM, Van Wuyckhuyse BC, Tabak LA, Bowen WH. The effect of milk and casein proteins on the adherence of Streptococcus mutans to saliva-coated hydroxyapatite. Arch Oral Biol. 1994;39(12):1063–9. pmid:7717888
- 20. Danielsson Niemi L, Hernell O, Johansson I. Human milk compounds inhibiting adhesion of mutans streptococci to host ligand-coated hydroxyapatite in vitro. Caries Res. 2009;43(3):171–8. pmid:19390191
- 21. Wernersson J, Danielsson Niemi L, Einarson S, Hernell O, Johansson I. Effects of human milk on adhesion of Streptococcus mutans to saliva-coated hydroxyapatite in vitro. Caries Res. 2006;40(5):412–7. pmid:16946610
- 22. Eriksson L, Lif Holgerson P, Johansson I. Saliva and tooth biofilm bacterial microbiota in adolescents in a low caries community. Sci Rep. 2017;7:5861. pmid:28724921
- 23. Norberg M, Wall S, Boman K, Weinehall L. 2010. The Västerbotten intervention programme: Background, design and implications. Global Health Action. 3:15.
- 24. Weinehall L, Hallgren CG, Westman G, Janlert U, Wall S. Reduction of selection bias in primary prevention of cardiovascular disease through involvement of primary health care. Scand J Prim Health Care. 1998;16:171–76. pmid:9800231
- 25. Pukkala E, Andersen A, Berglund G, Gislefoss R, Gudnason V, Hallmans G, et al. Nordic biological specimen banks as basis for studies of cancer causes and control—more than 2 million sample donors, 25 million person years and 100,000 prospective cancers. Acta Oncol. 2007;46:286–307. pmid:17450464
- 26. Johansson I, Hallmans G, Wikman A, Biessy C, Riboli E, Kaaks R. Validation and calibration of food-frequency questionnaire measurements in the Northern Sweden Health and Disease cohort. Public Health Nutr. 2002;5:487–96. pmid:12003662
- 27. Yano A, Kaneko N, Ida H, Yamaguchi T, Hanada N. Real-time PCR for quantification of Streptococcus mutans. FEMS Microbiol Lett. 2002;217:23–30. pmid:12445641
- 28. Gomes BP, Berber VB, Kokaras AS, Chen T, Paster BJ. Microbiomes of Endodontic-Periodontal Lesions before and after Chemomechanical Preparation. J Endod. 2015;41:1975–84. pmid:26521147
- 29. Chen T, Yu WH, Izard J, Baranova OV, Lakshmanan A, Dewhirst FE. The Human Oral Microbiome Database: a web accessible resource for investigating oral microbe taxonomic and genomic information. Database (Oxford). 2010 Jul 6;2010:baq013. pmid:20624719
- 30. Visioli F, Strata A. Milk, dairy products, and their functional effects in humans: a narrative review of recent evidence. Adv Nutr. 2014;5:131–43. pmid:24618755
- 31. Lovegrove JA, Hobbs DA. New perspectives on dairy and cardiovascular health. Proc Nutr Soc. 2016;75:247–58. pmid:26907978
- 32. Ceapa C, Wopereis H, Rezaiki L, Kleerebezem M, Knol J, Oozeer R. Influence of fermented milk products, prebiotics and probiotics on microbiota composition and health. Best Pract Res Clin Gastroenterol. 2013;27:139–55. pmid:23768559
- 33. Shafquat A, Joice R, Simmons SL, Huttenhower C. Functional and phylogenetic assembly of microbial communities in the human microbiome. Trends Microbiol. 2014;22:261–6. pmid:24618403
- 34. Bashan A, Gibson TE, Friedman J, Carey VJ, Weiss ST, Hohmann EL, et al. Universality of human microbial dynamics. Nature. 2016;534:259–62. pmid:27279224
- 35. Kodukula K, Faller DV, Harpp DN, Kanara I, Pernokas J, Pernokas M, et al. Gut Microbiota and Salivary Diagnostics: The Mouth Is Salivating to Tell Us Something. Biores Open Access. 2017;6:123–32. pmid:29098118
- 36. Biagi E, Quercia S, Aceti A, Beghetti I, Rampelli S, Turroni S, et al. The Bacterial Ecosystem of Mother's Milk and Infant's Mouth and Gut. Front Microbiol. 2017;8:1214. pmid:28713343
- 37. Charbonneau MR, Blanton LV, DiGiulio DB, Relman DA, Lebrilla CB, Mills DA, et al. A microbial perspective of human developmental biology. Nature. 2016;535:48–55. pmid:27383979
- 38. Drago-Serrano ME, Campos-Rodríguez R, Carrero JC, de la Garza M. Lactoferrin: Balancing Ups and Downs of Inflammation Due to Microbial Infections. Int J Mol Sci. 2017;18(3).
- 39. Jianqin S, Leiming X, Lu X, Yelland GW, Ni J, Clarke AJ. Effects of milk containing only A2 beta casein versus milk containing both A1 and A2 beta casein proteins on gastrointestinal physiology, symptoms of discomfort, and cognitive behavior of people with self-reported intolerance to traditional cows' milk. Nutr J. 2016;15:35. pmid:27039383
- 40. Dashper SG, Saion BN, Stacey MA, Manton DJ, Cochrane NJ, Stanton DP, et al. Acidogenic potential of soy and bovine milk beverages. J Dent. 2012;40:736–41. pmid:22609610
- 41. Aimutis WR. Bioactive properties of milk proteins with particular focus on anticariogenesis. J Nutr. 2004;134:989S–95S. pmid:15051859
- 42. Hegde AM, Naik N, Kumari S. Comparison of salivary calcium, phosphate and alkaline phosphatase levels in children with early childhood caries after administration of milk, cheese and GC tooth mousse: an in vivo study. J Clin Pediatr Dent. 2014;38(4):318–25. pmid:25571682
- 43. Ravishankar TL, Yadav V, Tangade PS, Tirth A, Chaitra TR. Effect of consuming different dairy products on calcium, phosphorus and pH levels of human dental plaque: a comparative study. Eur Arch Paediatr Dent. 2012;13(3):144–8. pmid:22652212
- 44. Tanaka K, Miyake Y, Sasaki S. Intake of dairy products and the prevalence of dental caries in young children. J Dent. 2010;38(7):579–83. pmid:20433890
- 45. Pitts NB, Zero DT, Marsh PD, Ekstrand K, Weintraub JA, Ramos-Gomez F, Tagami J, Twetman S, Tsakos G, Ismail A. Dental caries. Nat Rev Dis Primers. 2017 25;3:17030. pmid:28540937
- 46. Timpson NJ, Brennan P, Gaborieau V, Moore L, Zaridze D, Matveev V, Szeszenia-Dabrowska N, Lissowska J, Mates D, Bencko V, Foretova L, Janout V, Chow WH, Rothman N, Boffetta P, Harbord RM, Smith GD. Can lactase persistence genotype be used to reassess the relationship between renal cell carcinoma and milk drinking? Potentials and problems in the application of Mendelian randomization. Cancer Epidemiol Biomarkers Prev. 2010;19:1341–8. pmid:20447925
- 47. Ylöstalo PV, Järvelin MR, Laitinen J, Knuuttila ML. Gingivitis, dental caries and tooth loss: risk factors for cardiovascular diseases or indicators of elevated health risks. J Clin Periodontol. 2006;33:92–101. pmid:16441731