Search

Humans use visual features of the environment (landmarks) to allow them to navigate: finding new routes and following those learnt previously. An important subset oflandmarks is objects. Despite knowing this, we still do not have a complete picture of how landmarks are used during human navigation. The work presented here assesses how the location of an object along a route influences its use as a landmark. Thus, the current work provides a theory of landmark use in human navigation, defining first how objects become encoded as landmarks and secondly how those objects are used to guide directional decisions. The first five experiments test recognition of objects placed along a learnt route, to discover how objects become landmarks. These experiments reveal an important and novel finding of a difference in recognition for objects placed at decision points at which a turn occurs compared to decision points at which no turn occurs. By including a temporary turning motion at decision points where no permanent turn occurs, it is shown that the presence of a turn is the key factor driving the difference in object recognition at the two decision point locations. The change in optic flow that occurs at such turns is put forward as the basis for this difference. The change in optic flow triggers the encoding of objects at turns as key landmarks. In the final five experiments, participants make directional decisions when presented with images of decision point junctions at which a turn occurs, after having learnt the route. These experiments were designed to investigate how people use landmark objects while retracing a learnt route. These experiments demonstrated that participants use the objects as beacons (selecting the directions based on the location of the object at the junction). This was shown by changing the location of half of the obj ects from one side of a junction between the learning and testing phases of the experiments. During the trials where the obj ects had changed position, participants made systematic errors due to the change. Thus, this thesis provides a theory of the use of landmarks in humans and increases our understanding of human navigation.

One of the principal steps in the formation of a calcium sulphate deposit on a surface under isothermal conditions is shown to be the nucleation of crystals in the bulk of the solution followed by their attachment to the surface. The ease of attachment of the' crystals depends on the supersaturation and agitation of the solution as well as on the smoothness of the surface, but it appears to be independent of the surface material. It is shown that by controlling the number of crystals in the bulk of the: solution or the access of these crystals to. the surface, the number of crystals attached can be. controlled. The rate of nucleation of hemihydrate and dihydrate has been measured under isothermal conditions. The induction time for the hemihydrate is shorter than that for the dihydrate in simple calcium sulphate solutions; the induction time for the hemihydrate is reduced even further in mixed calcium chloride and sodium sulphate solutions. The growth rates at controlled levels of supersaturation and stirring velocity both of single hemihydrate and dihydrate crystals in the bulk of the solution and of massive deposits of these crystals on a metal surface have been measured, and a direct relationship is indicated. From the effect of stirring on the growth rate of single crystals, the relative contributions of mass transfer and surface reaction to the overall growth rate of the dihydrate, are found, and it is shcma that the growth rate is controlled to some. extent, by the rate of mass transfer except at low supersaturations and high stirring rates. When mass transfer effects are eliminated, the surface reaction rate is found to be approximately proportional to the second power of the supersat— uration. Also, it is shown that the growth of the hemihydrate can be inhibited by a coating of glass dust on the crystal surface.

Bacteriophages are natural viruses that attack bacteria and are abundant in all environments, including air, water, and soil- following and co-existing with their hosts. Unlike antibiotics, bacteriophages are specific to their bacterial hosts without affecting other microflora. The use of bacteriophages and phage endolysin based therapies to kill pathogens without harming the majority of harmless bacteria has received growing attention during the past decade, especially in the era where bacterial resistance to antibiotic treatment is increasing. The aims of this study were to characterise a putative prophage (phiFNP1) residing in the genome of the periodontal pathogens Fusobacterium nucleatum polymorphum ATCC 10953, investigating the possibility of prophage induction, examining its presence in clinical plaque samples taken from patients with chronic periodontitis cases, and purify its putative lysis module genes to examine their potential antimicrobial activity. In addition, attempts will be made to catalogue phages that are presents in samples from the oral cavity of patients with chronic periodontitis and in wastewater, followed by isolation of lytic phages targeting endodontic and oral associated pathogens, characterisation of the isolated phages, evaluation of the efficiency of phage towards biofilm elimination, and established an animal model to study phage-bacterial interaction in order to develop them for treatments. The results revealed that phiFNP1 prophage are common in subgingival plaque of patients suffering from chronic periodontal disease but attempts to induce phiFNP1 using Mitomycin C were inconclusive, indicating that it might be defective. Bioinformatics revealed that the genome of phiFNP1 contained potential lysin genes, with one being cloned and purified successfully in soluble form, though its activity as antibacterial was not confirmed, and needs to be further explored as phage lytic enzymes against Fusobacterium have not been identified yet. Various phage like particles with different morphology were visualised by direct electron microscopy from oral and wastewater samples, which reflect the richness and diversity of bacteriophage within those samples. Several phages were isolated namely phiSHEF 2,3,4,5,6,7 which belong to the Siphoviridae family which are specific to Enterococcus faecalis and the full chromosome sequence comparisons for three of the isolated phages (phiSHEF 2,4 and 5) revealed that they are lytic in nature reflected by absence of genes associated with lysogenic cycle, therefore place them as suitable candidate for therapy. They exhibit genomic variations corresponded to their difference in phage-host range especially in the tail region. In addition, phiSHEF 2 was capable of complete biofilm eradication on abiotic surfaces and significantly reduced biofilms formed on the surfaces of tooth root slices. Most importantly phiSHEF 2 recovered a Zebrafish larvae from the deadly infection caused by an oral clinical isolate of E. faecalis, indicating the establishment of a successful animal model for testing the efficiency of phage therapy towards E. faecalis infection. Finally, Exopolysaccharide mutants of E. faecalis were not infected by three of our phiSHEF phages tested, indicating that E. faecalis exopolysaccharide capsule and proper bacterial membrane integrity play an important role during the initial stages of phage infection. In conclusion, bacteriophages exhibit an apparent superiority to fight hard to eradicate pathogens such as those associated with recalcitrant endodontic infections thus "bacteriophage based therapy" could represent an innovative alternative to control and eliminate oral infections caused by biofilm forming and antibiotic resistant pathogens.

The keystone periodontal pathogen, P. gingivalis, is a Gram-negative oral anaerobe that is strongly implicated as the prime etiological agent in the initiation and progression of periodontal disease. P. gingivalis contains a plethora of virulence factors, including fimbriae, proteases, lipopolysaccharides and outer membrane proteins that contribute to the pathogenesis of the disease. The bacterium also displays the ability to interact with host cells through the adherence and invasion both in vitro and in vivo. In this thesis new light has been shed on the role of the heterotrimeric outer membrane protein A (OmpA). Through the creation of ompA mutants and recombinant complementation plasmids, alongside the use of standard antibiotic protection assays in an oral epithelial cell line, this research has demonstrated the importance of OmpA, specifically the OmpA2 subunit, in the invasion of the host and in the ability of the bacteria to form biofilms. Structural analysis of the protein identified extracellular loops, which when synthetic versions were applied to host cells, demonstrated successful interruption of wild-type P. gingivalis adherence and invasion of the host, indicating a direct interaction of OmpA2 with oral epithelial cells. In particular, this research demonstrates that OmpA2-loop 4 plays an important role in the interaction with the host through significantly increased binding to host cells when applied to fluorescent latex beads. This work also characterised the putative periplasmic chaperone protein OmpH, identifying potential proteins that act as clients to this chaperone, including OmpA. Through the creation of an ompH1H2 knock out mutant, enzymatic analysis of various virulence factors was assessed, demonstrating a loss of gingipain activity, whilst also identifying a loss of membrane stability through the observation of an increase in outer membrane vesicle formation. Evidence is also presented for the role of OmpH protein in chaperoning OmpA across the periplasm to the outer membrane and this work has opened up the potential for novel work identifying other clients of this chaperone system. Overall, the work in this thesis has demonstrated that the OmpA protein, specifically OmpA2, is directly involved in the interaction with host oral epithelial cells and in biofilm formation, specifically demonstrating the direct role of the extracellular surface loops in this interaction. This work also hypothesises a role for the OmpH protein in chaperoning outer membrane proteins, with the identification of potential clients, such as OmpA and the gingipains. Finally, this work has developed an improved mutagenesis technique for the creation of P. gingivalis mutants based on a modified natural competence method.

Various biomarkers in gingival crevicular fluid (GCF) have been examined in an attempt to obtain a specific and sensitive marker for periodontal disease progression and response to treatment. To date none have been found to be particularly reliable on their own. Therefore, the aim of this longitudinal clinical study was to determine whether key host enzymes (Matrix metalloproteinase-8, Cathepsin G, Elastase) and bacterial enzymes (Trypsin-like activity, Sialidase) detectable in GCF plus levels of key bacteria (Porphyromonas gingivalis, Tannerella forsythia and Fusobacterium nucleatum) detectable in subgingival plaque can be used in combination to provide an improved prognostic “finger print” for the outcome of treatment in patients with chronic periodontitis. Methods: 89 subjects were recruited to a longitudinal study. At baseline, 3-month and 6-month samples of GCF and subgingival plaque were collected from 3 representative sites: healthy (≤ 3mm), deep non-bleeding (DNB) (≥ 6mm) and deep bleeding (DB) (≥ 6mm) using a Periopaper® strips and a periodontal curette. In addition full mouth clinical data (pocket probing depth, clinical attachment loss, plaque index and bleeding on probing) were also recorded and patients received standard non-surgical periodontal treatment. GCF samples were assayed for each of the above enzyme activities using colourimetric/fluorometric substrates and subgingival plaque samples were assayed for the levels of each of the above bacterial species by qPCR. Data were analysed on a site-by- site basis using logistic regression for enzyme and bacterial profiles predictive of ≥ 2mm improvement in pocket probing depth (PPD) 6 months after treatment. Results: 77 individuals completed the 6-month interval phase. Full mouth clinical data showed statistically significant reduction in response to treatment, however, one third of DNB and DB sites showed less than 2mm improvement in pocket depth. The average levels of all biomarkers (enzymes and bacteria) were significantly higher in diseased sites than healthy sites and overall they decreased through the course of the study except Fusobacterium nucleatum. Matrix metalloproteinase-8 (MMP8), elastase, sialidase, Porphyromonas gingivalis and Tannerella forsythia showed the greatest reductions. The levels of MMP8, elastase and sialidase at baseline significantly correlated with the initial PPD as follows: MMP8 (r= 0.58), elastase (r=0.51) and sialidase (r=0.5). Using threshold enzyme levels that were the values with the highest sensitivity and specificity MMP8 (94ng/μl), elastase (33ng/μl) sialidase (2.3ng/μl), Porphyromonas gingivalis (0.23%) and Tannerella forsythia (0.35%), ROC curve analysis demonstrated that the baseline levels of these five biomarkers at sites are reliable diagnostic biomarkers as they differentiated healthy sites from diseased sites with degree of sensitivity and specificity above 77%. Furthermore, logistic regression showed that the combination of MMP8, elastase and sialidase provided accurate predictions of treatment outcome (81.3% for DNB, 80.3% for DB), which was significantly better than each enzyme alone (62.5%). When combined with the levels of Porphyromonas gingivalis and Tannerella forsythia the prediction value increased to 92% for DNB sites and 93.3% for DB sites. The biomarker values of MMP8, elastase and sialidase were also shown to be reliable after validating with an independent cohort. Preliminary tests showed that these three enzymes can be translated in to a simple chair-side test but this requires further development. Conclusion: Combined profiles of the above biomarkers offer a significantly improved indication of a site’s likely response to non-surgical periodontal treatment.

Background: Adverse drug reactions are responsible for a significant proportion of hospitalisations. This PhD aimed to develop and optimise methods for detecting serious adverse drug reactions in databases of electronic health records for use in pharmacoepidemiology and genetic epidemiology, with a focus on cholestatic liver injury. Methods: A systematic review was performed before developing a multiple database source (“multisource”) algorithm for identifying cholestatic liver injury. Multisource algorithm case status was used to guide the development of another algorithm using data from a standard UK Clinical Practice Research Datalink (CPRD) record only (the CPRD algorithm). Testing of the CPRD algorithm was performed within a cohort analysis of an established cause of the injury (flucloxacillin), before carrying out a casecontrol study investigating a number of putative associations (drug exposures carbamazepine, celecoxib, duloxetine, ramipril and risperidone). Results: The majority of reviewed studies lacked a reproducible case definition, and case assignment generally required information external to database records. Secondary care (HES) data provided little additional information than that found in primary care (CPRD), meaning that the CPRD algorithm had a very good ability to discriminate between multisource algorithm cases statuses (ROC area under the curve 0.95). The flucloxacillin 45-day risk estimate obtained from the cohort study using the highest specificity CPRD algorithm (6.15 per 100 000 users, 95% CI 4.61 – 8.04) was very similar to previous studies. Celecoxib and risperidone were associated with cholestatic liver injury (celecoxib multivariablemultivariable RR recent vs. current users low specificity CPRD algorithm 1.89, 95% CI 1.11 – 3.22, risperidone multivariablemultivariable RR high specificity CPRD algorithm 2.59, 95% CI 1.41 – 4.75). Conclusions: The CPRD algorithm detected similar flucloxacillin effects as (1) the multisource algorithm and (2) previous studies. Associations with risperidone and celecoxib were also detected. Algorithm characteristics that could facilitate (1) pharmacovigilance and (2) recruitment to genetic association studies include the ability to (a) detect cases without using information external to the EHR and (b) apply varying levels of specificity and sensitivity.