A growing body of research has associated the nitrate (NO3-) -reducing bacteria which are involved in the NO3--nitrite (NO2-)-nitric oxide (NO) pathway with physiological and cognitive health during ageing. The primary aims of this thesis were: 1. to develop and evaluate robust bioinformatics methods for the analysis of large bacteria datasets and to explore the relationships between microbiome data and the physiological and cognitive characteristics of the host; 2. to determine the sampling location in the mouth that should be used to capture the greatest abundances of NO3--reducing bacteria; 3. to assess whether high-fat or high-carbohydrate diets impact the effects of dietary NO3- on the oral microbiome; 4. to assess whether there are consistencies in correlations between the oral microbiome and cognitive performance between young and older adults and between cohorts recruited from two discrete geographical locations; and 5. to apply network analysis in a novel context to assess relationships between the oral microbiome, genetic risk for cognitive decline, and cognitive health in older people with mild cognitive impairment.
The current work sampled the oral microbiome, nitric oxide (NO) biomarkers in plasma, saliva and fluid derived from mouth washing and tested cognitive performance in young and older adults to determine the relationship between the oral microbiome, NO bioavailability and cognitive function, employing streamlined statistical analysis methods to relate oral bacteria co-occurrence with physiological and cognitive function tests.
In chapter 4, statistical methods for analysing the oral microbiome were evaluated using data previously collected from 26 older adults. Two correlation network analysis approaches were compared using a subset of data to establish oral microbiome co-occurrence, and Weighted gene correlation network analysis (WGCNA) was used to correlate groups of highly positively correlated bacteria (modules) to functional markers of human health. WGCNA was selected as the optimal approach for network analysis in this thesis.
Chapter 5 used oral microbiome data mined from the Human Microbiome Project to establish where, amongst seven different oral niches, the highest abundance of NO3--reducing bacteria can be sampled. There were significant differences in Shannon H’ diversity between each oral site (p < 0.001) and 8 well-established NO3--reducing bacteria were found in greater proportion in saliva compared to the tongue dorsum (p < 0.01), including Campylobacter sp. Capnocytophaga sp. Corynebacterium sp. Eikenella sp. Porphyromonas sp. Prevotella sp. Propionibacterium sp. and Selenomonas sp. In contrast, there were higher proportions of Actinomyces sp. Granulicatella sp. Neisseria sp. Rothia sp. and Streptococcus sp. on the tongue dorsum. While saliva contained the highest proportion of NO3--reducing bacteria, it also exhibited the most variation between individuals. Therefore, the oral sampling site must be carefully considered in physiological studies.
In chapter 6, 13 participants were provided with a 4-day dietary intervention, followed by a 7 day washout in a 4-way randomised cross-over design. The diets included high-carbohydrate with 12 mmol potassium chloride (CHO-PL), high-carbohydrate with 12 mmol potassium nitrate (CHO-NIT), high-fat with 12 mmol potassium chloride (FAT-PL) and high-fat with 12 mmol potassium nitrate (FAT-NIT). After False Discovery Rate (FDR) correction, the CHO-NIT diet increased the relative abundances of Capnocytophaga sputgena (mean rank difference = 13.18; FDR q = 0.03) and Neisseria elongata (mean rank difference = 13.27; FDR q = 0.03) compared to the CHO-PL diet. Prevotella jejuni significantly decreased from pre to post dietary intervention in FAT-NIT compared to both CHO conditions (FDR q = 0.01). Acute KNO3 supplementation revealed that the change in plasma [NO3-] from baseline to 2 hours was significantly higher in CHO-NIT compared to FAT-NIT (mean difference, 595.93 µM; FDR q = 0.04).
In chapter 7, it was found that there were significant differences in oral microbiome composition between two European locations in young and older adults (n = 60, p = 0.03). Furthermore, co-occurrence relationships were not robust between the geographical locations. A WGCNA consensus analysis between UK young and older adults revealed that the Neisseria-Haemophilus module was significantly positively correlated with plasma [NO3-] (p = 0.02). In the Belgium cohort, the WGCNA consensus analysis between young and older adults identified 2 modules that were inversely correlated with cognitive function tests. These included a Prevotella sp. dominated module with risk-taking (p = 0.03) and a module containing several species of Capnocytophaga, Prevotella, Treponema and N. subflava with visual object learning (p = 0.03).
In chapter 8, oral bacteria, and [NO3-] and [NO2- ] in mouth rinse fluid, as well as cognitive function were measured in 60 healthy participants and 60 participants with mild cognitive impairment (MCI). Using WGCNA, the Neisseria-Haemophilus module was found to correlate with several cognitive tests in the MCI group. Porphyromonas gingivalis was identified as a potential biomarker in the MCI group (p = 0.03). Apolipoprotein E (APOE) data were available for a subset of participants (n = 33) and participants were grouped into APOE4 carriers and non-carriers. P. intermedia was identified as a potential biomarker in APOE4 carriers (p = 0.04).
The results of this thesis suggest that the oral microbiome plays a critical role in cognitive health, and that the oral microbiome changes throughout ageing and is modulated by diet, geographical location and dietary NO3- supplementation. The findings of this thesis are important because they highlight the importance of oral nitrate-reducing bacteria relationships and the association between co-occurring bacteria, NO biomarkers and cognitive health, and help to inform future studies that combine oral microbiome data with NO biomarkers and cognitive function testing.