Dietary nitrate intake and 16S rRNA inferred nitrite-generating capacity of the subgingival microbiome may influence glucose metabolism: Results from the Oral Infections Glucose Intolerance and Insulin Resistance Study (ORIGINS)

Abstract

Aims

We investigated if the association between the nitrite-generating capacity of the subgingival microbiome and early cardiometabolic risk biomarkers varied by dietary nitrate intake.

Materials and Methods

Cross-sectional data from 668 participants (mean age 31±9 years, 73% women) were analyzed. Dietary nitrate intake was calculated from food frequency questionnaires. Subgingival 16S rRNA sequencing (Illumina, MiSeq) and PICRUSt2 estimated microbial genes. The Microbiome Induced Nitric oxide Enrichment Score (MINES) was calculated as a ratio of microbial gene abundances representing enhanced net capacity for NO generation. Adjusted multivariable linear models regressed cardiometabolic risk biomarkers (HbA1c, glucose, insulin, insulin resistance (HOMA-IR), blood pressure) on nitrate intake and MINES together with a MINES*nitrate intake interaction term.

Results

Mean nitrate intake was 190±171 mg/day. Significant interactions of MINES and nitrate intake were observed for insulin and HOMA-IR (p<0.05). Among participants with a low MINES, higher nitrate intake was associated with lower HOMA-IR [1.2(1.1, 1.4) vs. 1.5(1.3,1.6);(p=0.002)], but levels were similar in those with high MINES (p=0.84).

Conclusions

A biomarker of higher microbial NO-generating capacity in subgingival plaque is associated with lower insulin and insulin resistance among individuals with lower dietary nitrate intake. Future trials evaluating the cardiometabolic benefits of nitrate-rich diets should incorporate measures of the entire oral microbiome.

INTRODUCTION

Emerging evidence highlights the complex interplay between the microbiome and diet in shaping health. For example, the conversion of dietary choline and carnitine into trimethylamine-N-oxide (TMAO), a metabolite strongly linked to increased risk of cardiovascular disease, is dependent on the gut microbiome(Chakaroun et al., 2023). Similarly, in the oral cavity, the oral microbiome modulates host physiology through its interaction with dietary components, particularly via the enterosalivary nitrate-nitrite-nitric oxide (NO₃⁻ NO₂⁻ NO) pathway(Govoni et al., 2008; Kapil et al., 2020; Lundberg et al., 2018).

This pathway enables oral microbes to convert dietary nitrate to nitrite (common in green leafy vegetables, beetroot, radishes, etc.) (Figure 1), which then becomes bioavailable for NO production. NO bioavailability improves hemodynamics, insulin-mediated glucose uptake, and blood pressure(Kapil et al., 2013; Vincent et al., 2004). Thus, the oral microbiome and dietary nitrate intake potentially play an important role in maintaining systemic NO levels and cardiometabolic health. While increased dietary nitrate consumption is associated with improved cardiometabolic health in some studies (Bahadoran et al., 2017; Blekkenhorst et al., 2018; Bondonno et al., 2021; Jackson et al., 2019; Srour et al., 2023), substantial heterogeneity in results across studies has limited translation of these potential benefits into clear nutritional recommendations.

Figure 1. Conceptual overview of the process by which the oral microbiome facilitates the generation of bioavailable nitric oxide.

Nitrates consumed via the diet (e.g., green leafy vegetables), or endogenously generated, are concentrated in saliva. Oral bacteria reduce salivary nitrate to nitrite in the mouth and nitrite is in turn swallowed and converted to systemically available NO via enzymatic and non-enzymatic downstream pathways. *A necessary step in this chain of events is bacterial-dependent conversion of nitrate to nitrite. This system is viewed as a “backup system” ensuring a reservoir of NO is available when NO generation from other pathways is compromised (e.g., dysfunction in endothelium-dependent vasodilation).

Prior studies’ lack of accounting for the role of the microbiome could represent one potential source of heterogenous findings. Directly measuring the microbiome’s nitrite-producing capacity is resource-intensive, yet assessing it is likely important given the reported variation in nitrite production (Goh et al., 2019; Rosier, Moya-Gonzalvez, et al., 2020). Accordingly, biomarkers of oral microbiota-dependent nitrate reduction may have translational value by reflecting biologically relevant exposures such as nitrite production, enabling population-level risk stratification, and serving as treatment targets and/or eligibility criteria for clinical trials. In a previous population-based study, we observed a potential biomarker of oral microbial net nitrite-generating capacity, using inferred metagenomic content, to be strongly associated with lower cardiometabolic risk (Goh et al., 2022). However, the interaction between dietary nitrate intake and the oral microbiome’s net nitrite-generating capacity in relation to cardiometabolic outcomes remains unexplored.

The need for more epidemiological research on the cardiometabolic benefits of nitrate consumption and synergy between the microbiome and dietary patterns has been emphasized by the 2016 National Heart, Lung, and Blood Institute Workshop on Dietary Nitrate (Ahluwalia et al., 2016) and the 2020–2030 Strategic Plan for NIH Nutrition Research (National Institutes of Health, 2020). Such studies can inform the real-world effectiveness of habitual dietary nitrate intake, which occurs at doses lower than those used in most experimental trials. Understanding the interaction of the oral microbiome and dietary nitrate intake may provide insights into which individuals will benefit from increased nitrate consumption. These studies can also inform whether oral microbiome assessments, and/or adjunctive therapies (e.g. prebiotic or probiotic supplementation (Rosier, Buetas, et al., 2020; Rosier, Moya-Gonzalvez, et al., 2020)) targeting microbial net nitrite-generating capacity, should be incorporated into future dietary nitrate supplementation studies to improve interventional precision. Similarly, dietary adjunctive therapies to periodontal interventions might enhance cardiometabolic and/or periodontal outcomes. However, population-level knowledge needed to justify and design clinical trials is lacking. For example, it is currently unknown whether microbiome-based biomarkers capture biologically relevant microbiome functional potential, what level of variation in nitrite generating capacity has systemic implications, if host nitrate consumption is important in this context, and whether modulating the microbiome and/or nitrate intake has clinically relevant cardiometabolic effects.

The aim of this investigation was to conduct hypothesis-generating analyses examining the effects of the interaction between the predicted net nitrite-generating capacity of the subgingival microbiome and self-reported dietary nitrate intake on cardiometabolic risk biomarkers. Gaining insights into these associations will be critical for establishing the scientific rationale and guiding the methodological design of future observational studies and dietary nitrate interventions.

MATERIALS AND METHODS

Description of ORIGINS

ORIGINS is a prospective cohort study investigating the relationship between the subgingival microbiome and impaired glucose metabolism(Demmer et al., 2015). The cross-sectional data used in this study are from a subset of n=800 Wave 2 participants with standardized oral microbiota and dietary data. Inclusion criteria were: i) aged 20–55 years; ii) no diagnosis of diabetes mellitus via participant self-report, HbA1c values <6.5% and fasting plasma glucose<126 mg/dl; iii) no self-reported history of myocardial infarction, heart failure, stroke, or chronic inflammatory conditions; iv) no antibiotic use in the past 30 days. The Columbia University Institutional Review Board approved the protocol. All participants provided informed consent.

Operationalization of Dietary Nitrate Intake

Baseline food and nutrient consumption were assessed via the National Cancer Institute Diet History Questionnaire 1 (DHQ-1) which queries consumption frequency and portion size for 124 food items over the past 12 months. Dietary nitrate intake was calculated from a US nitrate food composition database developed and validated utilizing an earlier version of the DHQ-1(Inoue-Choi et al., 2016), or obtained from other best-available sources, related foods, or calculated using standardized recipes. The residual method created energy-adjusted nitrate intake values for each participant. For further details see Supplemental Materials (Section1; Table S1).

Oral bacterial nitrite generating versus depletion gene abundance

Previously published methods detail sample collection, bacterial assessment, taxa identification, and estimation of functional gene profiles from 16S rRNA marker gene sequences(Goh et al., 2022) (See Supplemental Materials Section 2). Briefly, subgingival plaque samples were collected from index teeth, pooled by shallow and deep (≥4 mm) probing depths, microbial DNA extracted for the 16S rRNA gene (V3–V4 region), and underwent sequencing on a MiSeq (Illumina), generating 44,776,283 sequences (median of 37,067 sequences/sample). PICRUSt2 estimated microbial gene abundances from the 16S rRNA sequences(Douglas et al., 2020). Gene abundances were weighted-averaged across deep and shallow sites.

We previously observed the ratio of bacterial reductase genes that convert nitrite to nitric oxide (NO) versus ammonium (NH₄⁺) (specifically nirK & nirS vs. nirA, nirB, nirD, nrfA & nrfH) was associated with lower cardiometabolic risk(Goh et al., 2022), and may serve as a biomarker for net nitrite-generating capacity of the microbiome. These findings support the idea that not only nitrate-reduction capacity, but the specific nitrite reduction pathways (toward NO or NH₄⁺) critically influence how much salivary nitrite is swallowed (Morou-Bermúdez et al., 2022). Therefore, we focus on this a priori-determined gene abundance ratio, representing net capacity for nitrite generation; referred to hereafter as the Microbiome Induced Nitric oxide Enrichment Score (MINES).

To assess the robustness of the inferred MINES score, we compared scores derived from subgingival plaque 16S rRNA-based PICRUSt2 predictions, against scores calculated from saliva-based metagenomic sequences. A correlation (0.34, p<0.0001) was found (Figure 2; Supplemental Methods). Additionally, we again (Goh et al., 2022) show via ANCOM-BC, that nitrate-reducing bacteria (e.g. Actinomyces, Neisseria, and Rothia(Morou-Bermúdez et al., 2022; Rosier et al., 2022)) correlate with MINES (Figure S1).

Figure 2.

MINES gene abundance ratio predicted from subgingival plaque 16S rRNA gene amplicon sequencing data was significantly correlated with MINES ratio derived from saliva shotgun metagenomic sequencing from a subset of participants (n=128)

(Spearman ρ = 0.340, p-value = <0.0001).

Cardiometabolic risk biomarkers

Standard methods assessed plasma glucose, hemoglobin A1c (HbA1c), and insulin following an overnight fast(Demmer et al., 2015, 2017). The Homeostasis Model Assessment for Insulin Resistance (HOMA-IR) quantified insulin resistance, calculated from fasting insulin and glucose levels(Matthews et al., 1985). Seated resting systolic and diastolic blood pressures were measured in triplicate, and the last two measurements were averaged. We created normalized z-scores for systolic and diastolic blood pressure, HbA1c, fasting plasma glucose, log-transformed insulin, and log-transformed HOMA-IR, then averaged them to create a composite cardiometabolic z-score (CMZ) where higher values indicates worse cardiometabolic health.

Demographic, Anthropometric and Behavioral Risk Factors

Trained research assistants collected data on additional cardiometabolic risk factors as previously described(Demmer et al., 2015). Questionnaires assessed age, sex, race/ethnicity (non-Hispanic Black, non-Hispanic White, Hispanic, Other), education (<Bachelor’s degree, Bachelor’s degree, >Bachelor’s degree), and smoking status (current, former, or never). Body mass index (BMI) was calculated as weight(in kilograms)/height (meters²) from in-person measurements. We defined periodontal status using clinical oral examinations and the 2012 Centers for Disease Control and Prevention/American Academy of Periodontology (CDC/AAP) classification (None/Mild, Moderate/Severe). The Alternative Healthy Eating Index-2010 (AHEI) (Chiuve et al., 2012) was calculated from food frequency data and detailed in the Supplemental Materials.

Statistical Analysis

All analyses were conducted in SAS version 9.4 (SAS Institute Inc, Cary, North Carolina, USA). Multivariable generalized linear regression regressed CMZ and individual cardiometabolic biomarkers on tertiles of energy-adjusted nitrate intake. All models adjusted for total energy intake (kcal), sex, age, race/ethnicity, education, smoking status, BMI, periodontal status, AHEI. In interaction analyses, the MINES and nitrate intake variables were dichotomised via median split and used in the multivariable models, along with a MINES*nitrate intake interaction term.

Motivated by evidence that periodontitis impairs the nitrate-reducing capacity of the oral microbiome(Rosier et al., 2024) and evidence that females exhibit increased nitrite-producing capacity (Kapil et al., 2018), sensitivity analyses were conducted by periodontal status and sex (Supplemental Materials Section 4).

RESULTS

Participant characteristics and dietary nitrate intake

Participants who completed a food frequency questionnaire (FFQ), had plausible food energy levels [females with food energy values ≤600 or ≥6,000 and males with values ≤800 or ≥8,000 were excluded(Choi et al., 2020)], had available 16S rRNA data, and were not missing important baseline risk factors or outcome data were included presently, yielding n=668 participants. Demographic characteristics of the cohort are shown in Table 1. Participants were mean age 31±9 yrs and predominantly female (73%). Moderate and severe periodontitis were present in 26% and 2% of participants respectively. The majority (79%) of participants had ≥bachelor’s degree and were overwhelmingly never smokers (87%). The mean AHEI diet score was 46±12 and comparable to previous reports(Chiuve et al., 2012). Median[IQR] and mean±SD dietary nitrate intakes were 143[72, 245] mg/day and 190±171 mg/day, respectively. The major contributors to mean daily nitrate intake included cooked spinach/greens (37%), lettuce (19%), and raw spinach/greens (9%), with the remainder from other plant foods (Table S2). These top contributing items to dietary nitrate intake did not differ across dietary nitrate intake tertiles.

Table 1.

Characteristics of the study participants by tertile of dietary nitrate intake.

	Total (N=668) 143[72,245]† mg/day	Tertile 1 (n=222) 55[36,72]† mg/day	Tertile 2 (n=223) 143[120,170]† mg/day	Tertile 3 (n=223) 310[245, 436]† mg/day	p-value‡
Age (years)	31.3 ±9.3	32.3 ±9.5	31.1±9.0	30.5 ±9.1	0.09
Sex
Male	27%	32%	28%	22%	0.03*
Female	73%	68%	72%	78%
Race/ethnicity					<0.0001*
Hispanic	28%	36%	26%	22%
Non-Hispanic White	31%	23%	40%	31%
Black	14%	18%	10%	13%
Other	27%	24%	23%	34%
Education					<0.0001*
< Bachelor’s Degree	21%	28%	17%	18%
Bachelor’s Degree	51%	54%	46%	54%
> Bachelor’s Degree	28%	18%	37%	27%
Smoking Status					0.18
Never	87%	85%	89%	86%
Former	6%	5%	5%	9%
Current	7%	9%	6%	5%
Periodontitis					0.05
None/Mild	72%	67%	77%	73%
Moderate/Severe	28%	33%	23%	27%
BMI (kg/m2)	25.4 ±5.9	26.0 ±6.3	25.4±5.7	24.9± 5.6	0.12
Total energy intake (kcal)	1760±967	1918 ±1165	1790 ±877	1574 ±793	0.001*
Dietary nitrate intake (mg/day)	190±171	54 ±21	143 ±31	371 ±182	<0.0001*
Alternate Healthy Eating Index (AHEI)	46.4±12.1	41.1 ±10.3	47.2 ±11.9	50.9 ±12.1	<0.0001*
Cardiometabolic Risk Biomarkers
CMZ	0.01±4.2	0.75 ±4.1	−0.06±4.2	−0.7±3.9	0.001*
Systolic Blood Pressure, mmHg	116 ±12	117 ±12	117±12	116±12	0.48
Diastolic Blood Pressure, mmHg	72.0 ±8.8	72.7 ±9.1	71.7±8.9	71.5±8.4	0.34
HbA1c, %	5.25 ±0.3	5.30 ±0.3	5.24 ±0.3	5.21±0.3	0.02*
Fasting Plasma Glucose, mg/dL	84.4 ±7.5	85.8 ±8.0	84.2±7.2	83.1±7.2	0.001*
Insulin, μU/mL	6.2 [4.4, 9.3]	6.9 [4.9, 9.8]	6.0[4.4, 9.7]	5.8[4.3, 8.8]	0.04*
HOMA-IR§	1.3 [0.9, 2.0]	1.4[1.0, 2.1]	1.2[0.9, 2.0]	1.2[0.8, 1.8]	0.01
MINES bacterial gene abundance ratio ¶	0.57 ±0.30	0.52 ±0.28	0.60 ±0.32	0.59±0.29	0.004

Table values are mean ± SD or median [IQR] for continuous variables, and percentages % for categorical variables. p values <0.05 are indicated with a *.

^†Median [IQR] dietary nitrate intake.

^‡P values are for ANOVA F statistics or X² tests for any differences in the level of covariates between participants across the tertiles of dietary nitrate

Missing values: n=13 for smoking status; n=7 for BMI

^§Insulin resistance as measured by the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR); BMI indicates body mass index; HbA1c indicates haemoglobin A1c.

^¶MINES bacterial gene abundance ratio: Microbiome Induced Nitric oxide Enrichment Score was calculated as a ratio of microbial gene abundances necessary for NO (nitric oxide) vs. NH₄⁺ (ammonium) production. This score is a surrogate for enhanced net capacity for nitrite and nitric oxide generation.

Participants in the two highest tertiles of dietary nitrate intake had significantly higher MINES and AHEI scores. MINES increased significantly across higher tertiles of dietary nitrate intake levels (Table 1), was slightly higher in males (mean±SE: 0.56±0.28 in females vs. 0.60±0.34 in males, p=0.10) and those with none/mild periodontal status (0.58±0.30 for none/mild vs. 0.54±0.29 for moderate/severe; p=0.09) though this was not statistically significant. Higher MINES was significantly associated with lower cardiometabolic risk biomarkers overall as previously shown (Goh et al., 2022) (Table S3), with consistent trends in sex-specific analyses (Table S4).

Interaction between microbiome induced nitric oxide enrichment score (MINES) and dietary nitrate intake on cardiometabolic risk

Our findings showed that higher dietary nitrate intake is significantly associated with lower plasma glucose, insulin, and HOMA-IR in multivariable adjusted models (Table S5, Model 2). Similar patterns were observed in stratified analyses by periodontitis status (Supplementary Tables S6 and S7) and sex (Tables S4).

Figure 3 and Table 2 present mean cardiometabolic risk biomarker levels across high vs. low dietary nitrate intake, stratified by high vs. low MINES. Among participants with low MINES, higher nitrate intake was associated with lower glucose, insulin and HOMA-IR, with significant interactions observed for insulin and insulin resistance (interaction term p<0.05(Figure 2). For example, in the low MINES group, mean insulin [95%CI] levels were lower for high nitrate intake compared to low nitrate intake (5.8[5.3,6.5] vs. 6.8[6.2,7.5]; p=0.004), while in the high MINES group, levels were similar (6.0[5.4,6.6] vs. 5.9[5.3,6.5]; p=0.76) (interaction p=0.02) (Table 2). Table 3 presents the same interaction analyses from a complementary perspective, comparing high vs. low MINES within dietary nitrate strata. In participants with low dietary nitrate intake, high MINES was associated with a significantly lower mean HOMA-IR [95% CI] levels, 1.2 [1.1,1.4] vs. 1.5[1.3,1.6] (p=0.004), an association not seen in those with high dietary nitrate intake 1.3[1.1,1.4] vs. 1.2[1.1, 1.4](p=0.62)(interaction p=0.02).

Figure 3. Mean cardiometabolic biomarker levels by low vs. high dietary nitrate intake, stratified by low vs. high MINES.

(The corresponding interaction analyses results used to generate this figure are available in Table 2)

MINES: Microbiome Induced Nitric oxide Enrichment Score was calculated as a ratio of microbial gene abundances necessary for NO vs. NH₄⁺ production. This score is a surrogate for enhanced net capacity for nitrite generation. These data support the concept that among individuals with low dietary nitrate intake, an oral microbiome enriched for nitrite generation can ensure an adequate reservoir of NO-generating capacity and promote cardiometabolic health (i.e., the microbiome buffers against a low nitrate diet). Or, conversely, among individuals with an oral microbiome that has limited nitrite-generating capacity (e.g., in oral dysbiosis), a diet high in nitrates might ensure an adequate reservoir of NO-generating capacity and promote cardiometabolic health (i.e., the diet buffers against a dysbiotic microbiome).

Table 2.

Estimated Mean Cardiometabolic Risk Biomarker (95% CI) Values across Low vs. High Dietary Nitrate Intake, Stratified by Low vs. High MINES – As illustrated in Figure 3

Low MINES*				High MINES*			P-value for interaction
	Low Nitrate Intake ** n=178	High Nitrate Intake** n=156	p ‡	Low Nitrate Intake ** n=156	High Nitrate Intake** n=178	p ‡
CMZ	0.99 (0.33, 1.63)	0.13 (−0.54, 0.80)	0.02	−0.12 (−0.79, 0.56)	−0.07 (−0.71, 0.57)	0.90	0.07
Systolic Blood Pressure (mmHg)	119 (118, 121)	119 (117, 121)	0.91	117 (115, 119)	118 (116, 120)	0.41	0.49
Diastolic Blood Pressure (mmHg)	74 (72, 75)	74 (72, 76)	0.81	72 (70, 74)	73 (71, 74)	0.48	0.73
HbA1c (%)	5.25 (5.19, 5.31)	5.23 (5.16, 5.29)	0.52	5.23 (5.17, 5.29)	5.20 (5.14, 5.26)	0.33	0.81
Glucose (mg/dL)	87 (85, 88)	85 (83, 86)	0.04	86 (84, 87)	85 (84, 87)	0.60	0.26
Insulin (μU/mL)	6.8 (6.2, 7.5)	5.8 (5.3, 6.5)	0.004	5.9 (5.3, 6.5)	6.0 (5.4, 6.6)	0.76	0.02
HOMA-IR	1.5 (1.3, 1.6)	1.2 (1.1, 1.4)	0.002	1.2 (1.1, 1.4)	1.3 (1.1, 1.4)	0.84	0.02

Multivariate linear regression models adjusted for total energy intake, sex, age, race/ethnicity, education, smoking status, body mass index, periodontal status and Alternative Healthy Eating Index

^*Median nitrate intake [25^th,75^th percentile] = 143 [72, 245] mg/day.

^**MINES: Microbiome Induced Nitric oxide Enrichment Score was calculated as a ratio of microbial gene abundances necessary for NO vs NH₄⁺ production. This score is a surrogate for enhanced net capacity for nitrite and nitric oxide generation. Low vs. High MINES was defined by overall median split of MINES in the full population. Median MINES [25^th,75^th percentile]= 0.54 [0.37, 0.74].

Table 3.

Estimated Mean Cardiometabolic Risk Biomarker (95% CI) Values among Low vs. High MINES, Stratified by Low vs. High Dietary Nitrate Intake

	Low Dietary Nitrate intake†			High Dietary Nitrate Intake†			P-value for interaction
	Low MINES‡ (n=178)	High MINES‡(n=156)	p §	Low MINES‡ (n=156)	High MINES‡ (n=178)	p §
CMZ	0.99 (0.33, 1.63)	−0.12 (−0.79, 0.56)	0.002	0.13 (−0.54, 0.80)	−0.07 (−0.71, 0.57)	0.56	0.07
Systolic Blood Pressure (mmHg)	119 (118, 121)	117 (115, 119)	0.02	119 (117, 121)	118 (116, 120)	0.19	0.49
Diastolic Blood Pressure (mmHg)	74 (72, 75)	72 (70, 74)	0.07	74 (72, 76)	73 (71, 74)	0.19	0.73
HbA1c (%)	5.25 (5.19, 5.31)	5.23 (5.17, 5.29)	0.60	5.23 (5.16, 5.29)	5.20 (5.14, 5.26)	0.39	0.81
Glucose (mg/dL)	87 (85, 88)	86 (84, 87)	0.23	85 (83, 86)	85 (84, 87)	0.69	0.26
Insulin (μU/mL)	6.8 (6.2, 7.5)	5.9 (5.3, 6.5)	0.005	5.8 (5.3, 6.5)	6.0 (5.4, 6.6)	0.64	0.02
HOMA-IR	1.5 (1.3, 1.6)	1.2 (1.1, 1.4)	0.004	1.2 (1.1, 1.4)	1.3 (1.1, 1.4)	0.62	0.02

Multivariate linear regression models adjusted for total energy intake, sex, age, race/ethnicity, education, smoking status, body mass index, periodontal status and Alternative Healthy Eating Index

^†Median nitrate intake [25^th,75^th percentile] = 143 [72, 245] mg/day.

^‡MINES: Microbiome Induced Nitric oxide Enrichment Score was calculated as a ratio of microbial gene abundances necessary for NO vs NH₄⁺ production. This score is a surrogate for enhanced net capacity for nitrite and nitric oxide generation. Low vs. High MINES was defined by overall median split of MINES in the full population. Median MINES [25^th,75^th percentile]= 0.54 [0.37, 0.74].

^§P-value for difference between groups with above vs. below median MINES within strata of dietary nitrate intake.

DISCUSSION

Higher nitrate intake was associated with lower insulin and HOMA-IR levels, but only among participants with low MINES. Similarly, higher MINES was associated with lower cardiometabolic risk, but only in individuals with low dietary nitrate intake. These findings may inform precision-nutrition research and/or support the development of probiotics designed to enhance the oral microbiome’s nitrite-generating potential.

Both dietary nitrate supplementation and nitrate-reducing probiotics have been suggested as strategies to leverage the enterosalivary pathway of NO generation(Bahadoran, Mirmiran, Carlström, et al., 2021; Rosier, Buetas, et al., 2020; Rosier, Moya-Gonzalvez, et al., 2020). However, the evidence guiding when and for whom these interventions are most effective remains limited and inconsistent. For example, several trials have tested the impact of nitrate supplementation on cardiometabolic risk, yielding mixed results. Among studies reporting benefit, the effects often plateau for several cardiovascular outcomes (Blekkenhorst et al., 2017, 2018; Bondonno et al., 2017). One possible explanation is that prior studies did not account for the oral microbiome’s role in modulating the conversion of dietary nitrate into bioactive NO.

Our analysis provides valuable population-based evidence supporting the hypothesis that the cardiometabolic benefits of dietary nitrate may depend, in part, on the oral microbiome. Notably, and somewhat counter-intuitively, increased nitrate intake appeared beneficial only among participants with low MINES, a potential marker of oral nitrite-generating capacity. While these findings could be due to chance, some biologically plausible explanations support them.

First, previous experimental studies have shown that dietary nitrate can act as a prebiotic, promoting the growth of oral nitrite-generating bacteria(Rosier et al., 2022; Vanhatalo et al., 2018; Velmurugan et al., 2016). Correspondingly, we observed a positive association between nitrate intake and MINES (Table 1), suggesting that habitual nitrate-rich diets may exert selective pressures favoring nitrate-utilizing oral microbes.

Second, it is possible that nitrite generation has a threshold above which additional production yields diminishing cardiometabolic returns. This threshold might be reached either through high nitrate intake in individuals with low oral nitrate-reducing capacity, or through high bacterial nitrate-reducing activity even with modest nitrate intake. This hypothesis presents interesting research directions. Future clinical trials could stratify participants by baseline oral nitrate-reducing capacity or selectively enroll individuals with low nitrate reducing capacity to test whether they derive greater benefit from nitrate supplementation. Alternatively, as our findings show that elevated MINES was only related to cardiometabolic biomarkers among those with low nitrate intake (Table 3), oral microbiome centered approaches (e.g., pre/probiotic therapies) could be implemented in people unable to maintain nitrate-rich diets. Supporting this notion, a recent study reported that supplementing low-dose dietary nitrate with a Rothia aeria probiotic significantly increased nitrite production – comparable to levels achieved with a 10-fold higher nitrate dose (Mazurel et al., 2023). In our present analysis, Rothia aeria abundance was higher among participants with elevated MINES and those with no/mild periodontitis (Supplementary Methods Section 4 & Figure S1). Higher Rothia aeria abundance among individuals with no/mild periodontitis underscores the relevance of oral health status in shaping nitrite-generating capacity. While periodontitis adjustments did not impact the observed findings, we did find that dietary nitrate intake was more strongly associated with insulin and HOMA-IR among participants with no/mild periodontitis vs. those with moderate/severe periodontitis (Supplemental Tables 6 & 7). This suggests that periodontal interventions might indirectly boost oral nitrate reduction(Rosier et al., 2024; Simpson et al., 2024).

This study also advances the growing body of research on dietary nitrate intake and cardiometabolic health outcomes by extending beyond blood pressure to include biomarkers of glucose regulation. While experimental animal trials have demonstrated beneficial metabolic effects of dietary nitrate intake, human studies have been less conclusive(Bahadoran, Mirmiran, Kashfi, et al., 2021; Lundberg et al., 2018; Srour et al., 2023), potentially due to differences in population characteristics. For example, Bahodoran et. al. 2017 (Bahadoran et al., 2017) reported substantially higher nitrate intakes (mean±SD=434±147 vs. 190±171 mg/day) than in our study, but found no association with incident diabetes. The elevated smoking rates in Bahodoran et. al. vs. ORIGINS (22% vs. 7%) may have contributed to this lack of an association. Smoking may blunt dietary nitrate’s effects by increasing thiocyanate levels and impairing enterosalivary nitrate to nitrite conversion inhibiting the uptake of NO₃⁻, and increasing NO scavenging by reactive oxygen species, thereby lowering systemic NO bioavailability (Shannon et al., 2021; Tsuchiya et al., 2002). In addition, although the lack of association between dietary nitrate and lower blood pressure in our study is inconsistent with some prior studies (Blekkenhorst et al., 2017; Bondonno et al., 2017, 2021), it is important to note that our sample is ~20 years younger than previous studies. Older adults who might have diminished capacity for NO generation through the endogenous L-arginine-NO pathway, show greater blood pressure response to nitrate supplementation(Vanhatalo et al., 2018).

Our findings were specific to insulin and insulin resistance levels, and we did not observe statistically significant associations between MINES and dietary nitrate intake on glucose or HbA1c in this analysis (although our prior publication did observe a MINES-HbA1c association)(Goh et al., 2022). This is not completely unexpected in our younger, generally healthy sample. Rising insulin as a compensatory response to insulin resistance, is known to precede rising glucose or HbA1c and transition to overt prediabetes/diabetes by many years. We have shown similar findings for periodontitis/glucose regulation associations previously(Demmer et al., 2012). Moreover, even modest longitudinal increases in insulin resistance occurring before overt impairment in glycemia, have been shown to predict progression to prediabetes and diabetes, underscoring the relevance of early small metabolic alterations(Kim et al., 2018).

Limitations and implications for future research

Our predominantly female population (73% female) limits generalizability. Sex-steroids might shift the microbiome(Cornejo Ulloa et al., 2021) towards increased abundance of Prevotella spp.(Gürsoy et al., 2009), a genus also observed to be differentially higher in women in our study (Figure S1), and increased nitrate reduction (Kapil et al., 2018). Alternatively, hormone-microbiome effects could create effect modification, altering the potential benefits of nitrate consumption via suppression or enrichment of nitrate/nitrite reductase functionality (Kapil et al., 2018), or increasing NO bioavailability as has been observed during the late follicular phase of the menstrual cycle(Rosselli et al., 1994). We did not collect data on hormonal levels, hormone modulating therapies, or common physiological states linked to hormonal state. Additionally, sex differences in platelet NO production following dietary nitrate supplementation have been observed and attributed to differential processing by the gut flora, but could potentially involve the oral microbiome as well(Godwin et al., 2021). Nevertheless, while we did observe a trend towards modestly larger associations between MINES and cardiometabolic biomarkers in women than men, the associations were present in both sexes (Table S4). Larger studies with balanced sex distributions are necessary.

Microbes in multiple anatomical sites, particularly on the tongue(Doel et al., 2005) and in the gut(Koch et al., 2017), are likely relevant to nitrate reduction, but we only analysed subgingival plaque. These sites are compositionally distinct, but from the oral perspective, the tongue and subgingival sites exhibit substantial overlap in key nitrite-generating genera (e.g. Rothia, Nesseria) (Rosier, Buetas, et al., 2020). While the tongue is likely the dominant site for nitrate reduction, nitrate-reducing bacteria are enriched in the subgingival plaque following periodontal treatment and correlate with salivary nitrate-reducing capacity (Rosier et al., 2024; Simpson et al., 2024), suggesting that the subgingival microbiome may be relevant to nitrate reduction. Moreover, the proximity of the subgingival communities to the vasculature and the demonstrated ability of disease-relevant molecules to enter systemic circulation from the gingival crevice, support the notion that subgingival nitrate reduction could contribute to bioactive nitric oxide(Schreiber et al., 2010). Although we did not assess the gut microbiome presently, it is noteworthy that studies using oral chlorhexidine to blunt the oral microbiome’s role in nitrate reduction show substantial reductions in plasma nitrite following nitrate consumption(Govoni et al., 2008). This suggests that the oral, rather than the gut, microbiome is the major contributor to circulating nitrite.

Our study measured taxonomic composition via 16S rRNA DNA next-generation sequencing methods and operationalized MINES using predicted gene abundances as opposed to direct measures of microbial functional capacity. Therefore, MINES is only a biomarker of potential underlying functional capacity for nitrate metabolism. In our cohort, MINES was correlated with salivary metagenomic MINES score (Figure 2), which supports the premise that this score might serve as a valid biomarker of true metagenomic capacity. Nevertheless, the current MINES index does not fully capture nitrite-generating capacity of the oral cavity across oral niches, and we did not validate that MINES correlates with directly measured nitrite enrichment capacity in subgingival plaque. Future longitudinal studies using multi-site microbiome (saliva, tongue, plaque, and gut), metagenomic/metatranscriptomic sequencing, complemented by preplanned surrogate measures of NO production (e.g. plasma or salivary nitrate and nitrite), will be needed for external validation of MINES and to precisely explore the optimal metagenomically determined biomarkers for net NO generation.

Future studies should also account for oral environmental factors, including pH, oral diseases aside from periodontitis or antimicrobial mouthwash use, as well as specific sources of dietary nitrate (i.e., water vs. plant vs. animal) and dietary antioxidant intake, or systemic factors such as exercise, all of which can impact endogenous oral nitrate reduction, NO production or NO bioactivity (Arefirad et al., 2022; Morou-Bermúdez et al., 2022; Rosier et al., 2025; Zhong et al., 2024), potentially confounding associations. Finally, we utilized cross-sectional data, with diet and oral microbiome measured at a single time point precluding temporal assessments.

Strengths

This study enrolled a diverse study population with >1,000 subgingival samples and robust data collection for key potential confounders. The young, diabetes-free study population allowed us to examine the interplay between dietary nitrate intake, the subgingival microbiome, and cardiometabolic biomarkers before the development of clinical disease and related pharmacological interventions and behavioral changes; thus reducing the potential for reverse causality.

Our findings inform the amount of nitrate consumption potentially relevant to cardiometabolic health. The median split of low and high nitrate intake used in the interaction analyses of 143 mg is on the low side of nitrate doses used in intervention trials, which ranged from 155 mg to as high as 1100 mg (Blekkenhorst et al., 2018). This suggests that modest, real-world nitrate consumption levels may confer benefits, with prior observational studies reporting increases in dietary nitrate intake as low 30 mg/day to have cardiometabolic benefits(Blekkenhorst et al., 2017).

If our observations reflect a causal diet-microbiome interaction on cardiometabolic health, biomarkers such as MINES could become valuable for informing inclusion criteria and treatment targets for future larger human studies where direct lab-based validations and assessments of microbial functional capacity are resource-demanding.

CONCLUSIONS

In summary, our study offers observational evidence from a young, generally healthy cohort, that higher nitrate intake is associated with lower insulin and insulin resistance – but only among individuals with low levels of an oral microbiome-derived nitrite enrichment biomarker. Strengths include a large sample and explicit modelling of diet × microbiome interactions. However, the cross-sectional design limits causal inference, and MINES is an inference-based proxy derived from subgingival 16S/PICRUSt2 data, potentially missing key nitrite-reducing niches such as the tongue/saliva/gut. These findings are hypothesis-generating and highlight the need for longitudinal studies, including randomized trials, to rigorously test causality.

Data Availability Statement

All sequencing data are available through Qiita50 under study ID 14375 for subgingival plaque samples. Raw sequence data is also available through EBI accession PRJEB50261 for subgingival plaque samples.