There exists a newly discovered, well defined, acellular, strong layer, termed pre-Descemets layer or Dua’s layer (PDL), in the cornea just anterior to the Descemets membrane. This, with the Descemets membrane, separates along the last row of keratocytes in most cases of deep anterior lamellar keratoplasty with the big bubble technique. Recognition of this layer has considerable impact on lamellar corneal surgery, understanding of posterior corneal biomechanics and posterior corneal pathology, such as descemetocele, acute hydrops and pre-Descemets dystrophies. The aim of this work was to understand the dynamics of big bubble formation in the context of the known architecture of the cornea stroma, ascertain how type 1 (air between deep stroma and PDL), type 2 (air between PDL and Descemets membrane) and mixed bubbles (combination of type 1 and type 2) form and measure the pressure and volume of air required to produce big bubbles in vitro, including the intra-bubble pressure and volume for the different types of big bubbles. We also aimed to characterise the optical coherence tomography characteristics of the different layers in the wall of the big bubbles to help surgeons identify bubbles and understand the structures seen by intra-operative OCT. Finally we evaluated the endothelial cell density and viability in tissue samples obtained for Descemets membrane endothelial keratoplasty (DMEK) and pre-Descemets endothelial keratoplasty (PDEK) by the pneumodissection technique. Air was injected in 145 corneo-scleral samples, which were unsuitable for transplantation. Samples were obtained in organ culture medium from the UK eye banks and transferred to balanced salt solution ready for injection. Different types of big bubble formed were ascertained. Air pressure and volume required to create the big bubble in simulated deep anterior lamellar keratoplasty were measured. It was found that PDL could withstand a high pressure before bursting at around 700 mm of Hg. Accurate measurements of type-2 big bubble proved challenging. The volume of the type-1 BB was fairly consistent at 0.1ml. The movement of air injected in the corneal stroma was studied from the point of exit from the needle tip to complete aeration of the stroma and formation of a BB. This was video recorded and analysed. A very consistent pattern of air movement was observed. The initial movement was predominantly radial from the needle tip to the limbus, then circular in a clock-wise and counter clock-wise direction circumferentially along the limbus, then centripetally to fill the stroma. All type 1 BB started in the centre as multiple small bubbles which coalesced to form a BB. Almost all type 2 BB started at the periphery near the limbus. Ultrastructural examination of the point of commencement of type 2 BB revealed the presence of clusters of fenestrations, which most likely allow air to escape from the otherwise impervious PDL to access the plane between PDL and DM. This was a novel discovery and explained how type 2 BB formed and why they almost always start at the periphery. The consistent pattern of passage of air was in concordance with the known microarchitecture of the central and peripheral corneal stroma. Optical coherence tomography (OCT) characteristics of different types of big bubbles were studied. Samples obtained from the UK eye banks were scanned with Fourier-domain (FD-OCT), while that obtained from Canada eye bank were scanned with Time-domain (TD-OCT). A special clamp was used to affix the corneo-scleral sample on the OCT table with its posterior surface face the machine and mounted on artificial anterior chamber. It was found that FD-OCT could demonstrate type 1 BB wall as two parallel, double contour, hyper-reflective lines with hypo-reflective space in between. It also revealed that in type-2 BB, the posterior wall showed a parallel, double-contour curved hyper-reflective line with a dark space in between. This probably corresponds to the banded and non-banded zones of DM. Dua’s layer presents as a single hyper-reflective line. In TD-OCT, the posterior wall of type-1 and type-2 BB showed a single hyper-reflective curved line rather than the double-contour line. This finding will help cornea surgeons to identify and interpret different layers of big bubble intra-operatively with high resolution OCT devices. Endothelial cell density of PDEK and DMEK tissue were calculated. Endothelial cells were counted using light microscope before pneumodissection. Air was then injected to ascertain the creation of type-1 and type-2 BB. Tissue was then harvested by trephination and endothelial cell density of both types were calculated again. It was found that the corneal endothelial cell count in PEDK tissue preparation is no worse, if not slightly better than, in DMEK tissue prepared by pneumodissection. Therefore, PDEK preparation represents a viable graft preparation technique.

Although there are several publications on the prevalence of AMD in the UK, there remains a paucity of data from population studies from this country. This study provides the most detailed exploration of the epidemiology of AMD within a UK community based population to date. It provides contemporary prevalence rates for different stages of AMD in a UK population, and indicates the prevalence of advanced AMD is more common than previously thought (4.3%), reaching a maximum of 21.2% for the over 90 year age group. Despite this high prevalence of advanced AMD, approximately 50% of the over 65-year old population have no or minimal signs of ARM. This study confirms that persons with the earlier and intermediate stages of AMD are asymptomatic. The prevalence of GA is more common than that of nAMD, with GA being 1.7 and 1.6 times more prevalent in the right and left eyes respectively. This difference remains when analysed for the worse eye, with GA still being 1.3 times more common than nAMD. For the first time in a large population based study, this thesis reports self- satisfaction with participant’s vision and association with AMD severity/grade. It shows that a large proportion of individuals (61.0%) with GA in their worse eye still remain satisfied with their vision. When AMD grade in the worse eye was the variable, nAMD is the only stage at which the majority of participants are dissatisfied (59.4%) with the level of their vision. When the better eye was evaluated, there was a significant increase in subject dissatisfaction with vision, particularly for individuals with bilateral nAMD, with almost 90% of subjects being dissatisfied with the quality of their vision. We describe in detail the features of GA within this UK population. The information will be useful in the design of future clinical trials that evaluate the development or progression of GA or its treatment, as well as modelling the impact of GA on sight loss in the UK. In summary, GA has a prevalence of 2.5% for the worse eye in the over 65 year Bridlington population, increasing with age to a maximum of 9.84% for those aged 85 to 90 years. Bilateral GA is not an infrequent finding, occurring in 34% of subjects with known GA. In the mainstay it is largely an eccentric condition, with only 31.5% of involved eyes having subfoveal GA. The majority of GA appears to occur in the perifoveal region, particularly GA that is associated with conventional drusen. Only a minority eyes with GA in the community have large areas of atrophy over 17.5mm2; the mean area of GA was 4.51mm2. This thesis, in chapter 4, demonstrates there is a clear association between GA phenotype and the phenotype of drusen identified within the remaining fundus. In chapter 4 it is demonstrated that RPD are, in the mainstay, associated with horseshoe shaped GA, and have a high prevalence of 55.9% within this group. This is a previously unreported finding. The prevalence of RPD in eyes with round GA is considerably less (8.3%). There is a suggestion that GA in eyes with RPD may have larger areas of atrophy and are less likely to have foveal involvement. However, the mean VA remains the same secondary to a diffuse, central retinal pigment epitheliopathy that appears to be associated with RPD horseshoe shaped GA. This again, is a previously unreported finding. Unlike in the previous hospital based literature, we demonstrate most multifocal GA is still associated with conventional drusen. In chapter 5, the prevalence of RPD in the over 65 year Bridlington population is estimated as 5.06% for either eye. This is the highest population based prevalence of RPD reported to date, and represents the only detailed report of the epidemiology of RPD within a UK population. Out of all the population based prevalence studies performed to date, the BEAP Study is the only one to utilise and analyse the red-free channel of the FP, which enhanced RPD detection within the study by 13%. As such, this study possibly represents the most accurate measure of RPD prevalence to date. In the present study, the prevalence of RPD increases significantly with age, reaching a maximum of 27% in the over 90-year age group. The BEAP study confirms that RPD occur most frequently in the upper outer subfield. Unlike previously reported, however, RPD do occur within the central subfield, albeit in a form that differs slightly in its appearance, and significantly reduced size. The BEAP study confirms that RPD have a female preponderance, with a higher gender specific prevalence rate in women. They are commonly found in association with other signs of ARM, including drusen over 125μm (50 % of the time) and pigmentary changes. Isolated RPD, in the absence of conventional drusen, is an uncommon finding but does occur. Approximately 1 in 4 subjects with advanced AMD will have evidence of RPD in either eye. This thesis also reports, for the first time, the prevalence of RPD in eyes with PPCNVs. There is a suggestion that RPD are associated with visual dissatisfaction, and this may be associated with a central pigmentary epitheliopathy that is sometimes seen within RPD eyes. We specifically report the population prevalence of PPCNV, as all previous publications on the subject have arisen from hospital based populations, and therefore included predominantly symptomatic individuals, and therefore carried inherent selection bias. PPCNVs (grade 4c AMD) were an infrequent finding, with a prevalence of 0.29% for individuals over 65-years of age. This is considerable lower than nAMD (grade 4b AMD), which within the same population had a prevalence of 1.8% for the worse eye. There was a female preponderance, with 70% of PPCNVs occurring in females. This difference was maintained in gender specific prevalence rates of 0.36% and 0.19% for females and males respectively. Previous publications have reported that PPCNVs accounted for less than 10% of all CNVM. This figure is confirmed in the BEAP, in which PPCNV accounted for 12 out of a total of 90 cases of CNVMs, representing 13.3% of all prevalent CNVs identified. This study reports that the majority of cases of PPCNV are unilateral. There is a myriad of published associations between PPCNV and other conditions. These are mainly single case reports or small case series. In addition, the larger hospital based studies may also give a poor representation of the true associations between PPCNV and other conditions as small, nasal or age-related membranes may remain asymptomatic. In Chapter 6, 90% of eyes with PPCNV had evidence of drusen ≥63μm within the macula, an association which is far higher than previously published. Furthermore, in the current series, 30% of PPCNV were associated with RPD, a finding which had previously been unreported. Previous studies on PPCNV’s have graded the macula for age-related change, but not reported or graded the peripapillary area for degenerative changes. In this report, these features have been investigated for the first time, and find that all (100%) of our subjects had RPE pigmentary changes within half a disc diameter of the disc margin (in both affected and contralateral eyes). As drusen and pigmentary changes within the macula are the known hallmarks of both GA and CNV, it seems logical to consider these changes in the peripapillary area as potentially pathological for PPCNV and for them not to be overlooked. The association of PPCNV with angioid streaks is well established, but in chapter 6 it is reported that in 10% of subjects with PPCNV, there were identifiable angioid streaks. This is a more frequent association than previously published (although the numbers are small), and highlights an overlap between the aetiologies of PPCNV in patients with angioid streaks within the elderly, and suggests a possible tendency for an association with small membranes that remain asymptomatic. / In the present study of asymptomatic individuals, a large proportion of PPCNVs (42%) were nasal to the disc margin and no individual in this series was thought to have developed direct visual loss from PPCNV. This finding is very distinct from that in some previous hospital studies of symptomatic patients. This study suggests, by inference that up to two thirds of PPCVN may remain asymptomatic.

The Draize eye irritation test has been widely adopted as the "gold standard" to evaluate the potential eye irritation of a wide range of chemicals and formulations, including; pharmaceuticals, cosmetics, and their raw ingredients. The rationale for pursuing the development of human ocular based in vitro alternatives is to provide greater confidence in the prediction of human reactions to mild and moderate chemicals. This is particularly important with the implementation of the EC White Paper, "Strategy for a future chemicals policy" (2001) that is estimated to require the testing of approximately 30,000 'existing' chemicals by 2012. The development of in vitro alternatives for toxicity testing has mainly focused upon tests for quantitative measurement of acute toxicity following a single high-dose exposure. However, the degree of toxicity of any exposure is a function of; the dose which target cells receive, the duration of the exposure and the ability of the exposed cells to recover from the exposure. However, little account is taken of the potential role of long-term effects in modulating the toxic response. This study aims to generate an in vitro model utilizing a human corneal cell line monolayer to investigate the effects chronic exposure to exogenous chemicals has upon toxicity of a subsequent acute challenge. Surfactants are ubiquitous within our daily environment, being significant active components in both household and personal care products, cosmetics and pharmaceuticals. Initially the effects of four representative surfactants (sodium dodecyl sulphate; anionic, tween 20; non-ionic, cocamidopropylbetaine; amphoteric and benzalkonium chloride; cationic) were examined following chronic exposure. Although the measured endpoints (neutral red uptake, resazurin reduction, fluorescein leakage and total protein content) revealed no alterations in J-HCET morphology, barrier function or biochemistry as a consequence of chronic exposure, it was determined that pre-exposure modulated the toxicity of subsequent acute exposures. The observed modulation in toxicity could have significant health implications for personal care products, cosmetics and pharmaceuticals intended for use near or within the eye. However, the mechanism(s) by which the toxicity of subsequent surfactant exposures was modulated remains to be elucidated. Whilst standard surfactants are good indicators of the effects following chronic exposure, there are pharmaceuticals designed for repeat use in the eye that have been associated with long term ocular irritancy, and a discontinuation of use i.e. timolol maleate. J-HCET cultures exposed to BSO in vitro confirm that the toxicity of timolol to the human corneal cells line was enhanced by suppressing the activity of γ-glutamylcysteine synthetase through irreversible inhibition, resulting in a decrease in intracellular glutathione (GSH) levels. In addition, chronic exposures to timolol maleate were also associated with an increase in toxicity of subsequent acute challenges. The long term use of Timolol maleate in the treatment of glaucoma in vivo may result in similar alterations in the intracellular GSH concentrations, resulting in discontinuation of treatment as consequence of ocular irritation through the generation of reactive oxidation species beyond the threshold that depleted intracellular GSH can respond. Since in vitro methods have been, and are being developed as alternatives to animal experiments, the use of bovine serum as a source of growth factors can seem to be contradictory to the purposes of the Three R's concept of Russell and Burch (1959). This study was conducted using culture media that contain animal derived growth supplements and a media where these had been substituted for plant derived materials. Comparisons were made between the effects these two supplements had upon a number of biological factors including morphology, biochemistry, barrier function and the response to exogenous chemicals. No alterations were observed in these parameters as a consequence of using culture medium containing plant derived materials compared to those containing animal derived growth supplements. This study has demonstrated that the development of a reliable and reproducible in vitro assay in keeping with the principles of the Three R's for modeling chronic – repeat ocular irritation is possible. However, the mechanistic relevance of the endpoints chosen and cell layer ultrastructure is considered to be an essential component. The further development of cell based in vitro systems to predict human responses to chronic/repeat ocular irritation is required.

Infectious keratitis is a sight threatening disease which can cause permanent visual loss if not diagnosed and treated at an early stage. It can be caused by different types of microbes which are either commensals or transferred from the environment. To fight against these threats, the ocular surface (OS) has developed innate and adaptive immune mechanisms. Antimicrobial peptides (AMPs) are natural effectors on the OS with actions that range from microbicidal effects to cell signalling. Human beta defensin (hBD) 1-3 and 9, Liver expressed antimicrobial peptide (LEAP) 1 and 2, human cathelicidin (LL37), ribonuclease7 (RNase7) are the main AMPs on the OS. In this work, the pattern of ocular AMPs gene expression in human OS cells treated with Acanthamoeba castellanii, Pseudomonas aeruginosa (P. aeruginosa) and Staphylococcus aureus (S. aureus) was studied and established. This was examined by quantitative real-time PCR (RT-PCR) using the Taqman assay. Among the studied AMPs, hBD3 gene showed the most significant increase in human OS cells infected with Acanthamoeba. LL37 demonstrated the highest level of gene expression in the samples infected with bacteria. In a different study, the gene expression of two AMPs (hBD3 and 9) was studied in OS samples taken from patients with different types of infectious keratitis both during and after the infection. This was compared with the expression in healthy subjects. Impression cytology (IC) was used to obtain samples of OS epithelium from recruited subjects. An optimized method for RNA extraction of IC samples was developed. Corresponding to the results of the in vitro study, hBD3 showed an overall up-regulation in all categories whereas hBD9 was down-regulated. These changes were most significant in patients with acute Acanthamoeba keratitis. The gene expression of both hBD3 and 9 showed a tendency towards returning to the levels found in healthy subjects when healing of the corneal infection was complete. In another study carried out to examine the antimicrobial activity of hBD3 we were surprised to find that we could not replicate this. We were unable to reproduce the previously reported antimicrobial activity of hBD3 but were able to demonstrate that the antimicrobial effect could be attributed to the acidic solvent used in preparing the hBD3 protein. The clinical significance of application of corneal densitometry as measured by the Pentacam system was assessed for the first time in patients with infectious keratitis. We demonstrated that corneal densitometry varied with levels of inflammation and was not confined to the site of infection only. It affected the whole cornea and reverted towards normal values as the inflammation settled when the infection was brought under control. We were able to demonstrate that densitometry can be used as a measure of the corneal response to infection and inflammation and could be used to monitor response to therapy. Finally, separate comprehensive prospective and retrospective studies of the clinical profile of severe infectious keratitis in Nottinghamshire were conducted. These two studies covered a total period of 7 years. The results of both studies were similar. Indeed, OS disease, CL wear and previous ocular surgery were found to be the most common risk factors. Positive results of corneal scraping were obtained only in about 40% of cases. Acanthamoeba, S. aureus, and P. aeruginosa were the most frequent causative organisms. Fortified topical antibiotics were effective in treating most cases. Therapeutic corneal grafting was found to be an effective and safe procedure in refractive infectious keratitis.

Primary open angle glaucoma is mostly prevalent in patients over 60 years. However, studies examining the role of laser scanning tomography in detecting glaucoma hitherto have employed non-population based data drawn from younger subjects. This study employed an elderly, Caucasian, population-based cohort (minimum age 65 years, n=721). All subjects underwent ophthalmological examination including Goldmann applanation tonometry, suprathreshold automated visual field test and Heidelberg retina tomography. Normality was defined as normal visual acuity, visual field and intraocular pressure. Perimetrically normal males were found to have significantly larger optic cups than females. Normative data were used to construct three diagnostic tests: Linear regression of rim area, RADAAR (rim area/disc area asymmetry ratio) and the optic disc hemifield test (comparison of the superior and inferior rim/disc area ratios in the temporal and nasal sectors). Specificities at the 99th limit of normality were 91.4%, 95.1% and 98.3% respectively. Sensitivity was assessed using a cohort of patients (n=58) with a new diagnosis of open angle glaucoma at the Queen's Medical Centre, Nottingham within the last two years. Sensitivities at the 99th limit of normality were 72.4%, 55.6% and 27.6% respectively. Applying the regression analysis bilaterally and RADAAR simultaneously generated a specificity of 83.0% and sensitivity of 88.9%. Inclusion of the optic disc hemifield test did not further increase sensitivity. Linear regression of rim area/disc area was found to be susceptible to non-linearity and heteroscedasticity, causing reduced specificity in bigger optic discs. Modeling the relationship separately for each disc area quartile overcame this limitation to produce a diagnostic test with constant accuracy. Whilst laser scanning ophthalmoscopy can discriminate glaucoma with reasonable accuracy, it is limited by disagreement in contour placement and poor image quality in the elderly. The combination of multiple statistical tests is not ideal - application of statistical shape analysis techniques may better employ the normative data.

Intraocular angiogenesis is associated with a number of common, blinding conditions including wet age-related macular degeneration, proliferative diabetic retinopathy, retinopathy of prematurity and rubeotic glaucoma. The pathogenesis of these disorders is centered on choroidal, retinal and iris microvascular endothelial cells (ECs) respectively. When studying these conditions, workers have applied conclusions from in vitro studies of human umbilical vein EC (HUVEC) and microvascular endothelial cells, derived from different species or organs. However it is now widely accepted that endothelial cells are very heterogeneous and .extrapolation of results from these cells may not provide reliable data applicable to human eye disease. I have successfully isolated and cultured matched human retinal, choroidal and iris endothelial cells (and sourced HUVECs) and have examined their gene expression profiles. I found wide differences in gene expression between HUVEC and ocular ECs and in addition, between matched human retinal and choroidal endothelial cells. Taken together, these results suggest that HUVECs are not suitable surrogates for studying human ocular disease and secondly. It has implications for our understanding and treatment of retinal and choroidal vascular diseases. To further define the heterogeneity of endothelial cells within a vascular bed, I developed a technique for the isolation of human inner choroidal ECs. After the cells were characterized by surface marker expression, the response of these unique cells to various growth factors was defined and gene expression microarray analysis was performed and compared with matching outer choroidal ECs. This demonstrated differences in genes involved in fenestration formation and growth factor expression. These findings may have implications for our understanding of macular diseases. Overall, this body of work provides an insight into endothelial cell heterogeneity within the eye and may provide clues as to the reasons for ocular vascular bed disease susceptibility and helps identify potential future selective anti-angiogenic treatments.

In vivo confocal microscopy (IVCM) offers a unique real time non-invasive imaging method to explore live tissues at cellular and subcellular levels of histological detail with magnifications very much comparable to conventional ‘ex vivo’ light microscopy. Therefore it has been widely used over the past two decades to investigate the ocular surface and cornea in health, disease and following surgical procedures. One of the main challenges in understanding IVCM is to get a proper interpretation of the images that present various figures and patterns of tissue structures all in black and white with variable degree of reflectivity, being whiter (or brighter) when they are more (or hyper) reflective. The lack of good correlation between IVCM and corresponding light microscopy for the same tissue samples has lead to speculative interpretations of IVCM images in the literature. In this work we tried to fill that gap through performing IVCM to patients with various corneal and ocular surface disorders just few days prior to obtaining the tissue samples (corneal graft, excisional biopsy or alcohol delamination) for histopathological examination to make sure that IVCM images were truly representative to the tissue details that might change if more time was left to elapse between IVCM and tissue sampling. In ocular surface disease group (chapter three) we contrasted IVCM criteria in conjunctival epithelial overgrowth onto the cornea in limbal stem cell deficiency and Pterygium-like disorders with those seen in Corneal/Conjunctival Intraepithelial neoplasia (CIN) to confirm reliable diagnostic criteria of CIN with IVCM. In corneal oedema (chapter four) IVCM viewing of (histologically confirmed) subepithelial fibroblasts without clinically visible corneal scarring has been reported for the first time in IVCM literature particularly in Fuchs endothelial dystrophy. Sub-basal corneal nerve reduction as well as stromal keratocytes and endothelial changes were clearly illustrated with IVCM. In keratoconus cases (chapter five) morphological epithelial changes related to regenerative atypia have been studied for the first time and compared to those seen in CIN through light microscopy and IVCM in addition to other epithelial changes associated with keratoconus. Novel IVCM criteria for Bowman’s zone breaks have been described and compared carefully with histological sections. In corneal dystrophies (chapter six) IVCM criteria in Thiel-Behnke’s corneal dystrophy (CDBII) as well as in macular, granular and lattice dystrophy correlated well with light microscopic findings with a novel IVCM pattern in one Macular dystrophy case. In Acanthamoeba keratitis (chapter seven) a new form of Acanthamoeba cysts has been described for the first time through IVCM. Moreover this study presented for the first time IVCM diagnostic criteria in corneal intraepithelial neoplasia, keratoconus, Acanthamoeba keratitis and some corneal dystrophies. A comprehensive IVCM illustration of various types and stages of corneal fibrosis has been achieved as well.

The use of amniotic membrane (AM) as a dressing for ocular surface injuries has attracted the interest of the military ophthalmological community. First applied in the 1930s, the tissue is widely used today, although clinical indications for treatment are incompletely defined. While AM is most commonly stored frozen and thawed before use, dried AM is preferred for logistical reasons. Optimal preservation of the tissue is necessary to preserve its quality. The effect of drying on the physical and biological properties of the tissue are unknown. A systematic review of the evidence of AM treatment of acute chemical injuries was conducted. A framework was proposed for optimising the dried tissue through thermal, moisture sorption and surface analytical techniques. The physical properties of AM preparations were compared by mechanical testing and mathematical modelling, and an attempt was made to cross-link the AM collagen. Inflammatory aspects of the tissue were assessed by immunological techniques, zymography and macrophage assays. There is a lack of high quality evidence to support the clinical application of AM for acute burns. Complex interactions were demonstrated between the dried tissue, its excipients and moisture, suggesting novel ways of optimising the product. The mechanical properties of the dried membrane indicated that the process adversely affected the tissue, and artificial cross-linking could not be achieved. While the presence of antimicrobial peptides was not clearly established, the elution of collagenolytic enzymes was shown in therapeutic preparations of AM. The production of tumour necrosis factor by macrophages, which adhere to the spongy layer of AM, was suppressed. This project makes original contributions relevant to the use of dried AM as a biomaterial in ophthalmic surgery. Further refinements of this work, animal model experimentation and clinical trials may support its future acceptance as a clinical application.