L’objectif de cette thèse est la validation fonctionnelle d’implants rétiniens pour la restauration fonctionnelle de la vision chez des patients non voyants suite à la perte de leurs photorécepteurs. Ce travail a été réalisé sur modèle animal et a évalué expérimentalement de nouveaux protocoles de stimulation. Tout d’abord nous avons utilisé la technique de spectroscopie d’impédance pour simuler mathématiquement l’interface tissu-implantafin de caractériser la présence d’un espace entre le tissu et l’implant. La seconde partie compare par imagerie optique (IO) les caractéristiques de la réponse corticale évoquée par stimulation visuelle ou électrique de la rétine par prothèse sous rétinienne. Nous avons retrouvé que la taille de l’activation par l’implant rétinien est beaucoup plus grande que son correspondant visuel. Dans une troisième partie, est réalisée une évaluation in vitro de la performance des stimulations sur rétine isolée pour définir comment les cellules ganglionnaires réagissent à différents modes de stimulations. Ce travail a permis d’établir la courbe des réponses en fonction de l’intensité des stimulations électriques. Enfin, la thèse décrit un modèle animal de dégénérescence rétinienne qui présente des désorganisations de la rétine externe. Une analyse en IO a été réalisée sur ce modèle afin d’évaluer la réponse corticale aux stimuli visuels et électriques. Ce travail de thèse, par des approches physiques et physiologiques complémentaires, apporte un certain nombre de réponses qui devraient permettre d’améliorer l’utilisation de futures prothèses rétiniennes par une adaptation physique des matrices d’électrodes ou des patrons de stimulations utilisées / The aim of this thesis is the functional validation of retinal implants used for vision restoration in blind patients due to the loss of photoreceptors. This work was designed to develop an animal model to experimentally validate prototypes of new implants and new stimulation protocols pattern. Firstly we used the technique of impedance spectroscopy to simulate mathematically the tissue/implant interface. These data confirm the importance of reducing the space between the stimulating electrodes and retinal tissue, as well as the importance of physical characteristics of the electrical stimulus used. In a second approach, we have compared responses of visual cortical neuronal population using optical imaging (OI), evoked either by visual or electric retinal stimulation through subretinal prosthesis. This approach has demonstrated that the stimulation of an electrode induces cortical activation that the size of the cortical response to the retinal implant stimulation is much larger than its corresponding visual stimulus. In the third part, I performed in vitro experiment to measure the performance of stimulation at the level of ganglion cells of isolated retina. We have quantified the response curve as a function of the intensity of the electrical stimulation. Finally, the thesis describes a new animal model of outter retinal degeneration. OI was also performed on this model to assess the response to the visual and retinal prosthesis stimulations. This thesis, through complementary physical and physiological approaches, provides a number of responses that can potentially improve the use of retinal prostheses through specification of their design or patterns of stimulation.

Diabetic retinopathy and age-related macular degeneration are two of the most common retinal diseases in these days, which can lead to partial or full loss of sight. Due to it, it is necessary to create new approaches enabling to detect these diseases and inform the patient about his condition in advance. The main objective of this work is to design and to implement an algorithm for retinal diseases classification based on images of the patient's retina of previously mentioned diseases. In the first part of this work, there is described in detail each stage of each disease and its the most frequent symptoms. In this thesis, there is also a chapter about fundus camera, which is a tool for image creation of human eye retina. In the second part of this thesis, there is proposed an approach for classification of diabetic retinopathy and age-related macular degeneration. There is also a chapter about algorithmic methods which can be used for image processing and object detection in image. The last part of this thesis contains the test results and their evaluation. Assessment of success of proposed and implemented methods is also part of this chapter.

In this work I present conclusions of clinical-laboratory research focused on the patients with diabetic macular edema (DME). We performed biochemical and immunochemical analyses of vitreous samples that were collected during the pars plana vitrectomy. Moreover, at patients with non-proliferative diabetic retinopathy (NPDR) we assessed morphological characteristics of DME using optical coherence tomography (OCT). According to our findings, the vitreous and serum concentrations of uric acid and glucose were significantly higher in patients with diabetic retinopathy and DME compared to controls. Also total ratio (serum/ vitreous concentration) of uric acid and glucose was in diabetics significantly higher than in controls. The most important determinant of increasing concentration of both uric acid and glucose in the vitreous was the grade of diabetic retinopathy. Moreover, we demonstrated significant correlation between vitreous concentration of uric acid and concentration of the vascular endothelial growth factor (VEGF) in patients with DME and NPDR. We found further, that the volume of the macula (cube volume - CV) computed with the software of Cirrus HD-OCT correlates in diabetics significantly with the vitreous VEGF concentration, but not with uric acid. This OCT parameter could be used to...

Le système kallicréine-kinines est un système peptidergique complexe impliqué dans les processus inflammatoires, le contrôle du tonus et de la perméabilité vasculaire. Les effets biologiques des kinines sont accomplis par l’intermédiaire de deux types de récepteurs couplés aux protéines G, soit le récepteur B1 (B1R) et le récepteur B2 (B2R). Alors que le B2R est un récepteur constitutif, le B1R est faiblement exprimé en situation physiologique; il est induit par le stress oxydatif, les cytokines pro-inflammatoires (interleukine-1β (IL-1β) et le facteur de nécrose tumorale-α (TNF-α)) ou par des endotoxines bactériennes à la fois au niveau systémique et local, notamment dans la rétine. Des études récentes de notre laboratoire ont montré l’implication du B1R dans la pathogenèse et la progression de la rétinopathie diabétique et de la dégénérescence maculaire liée à l’âge (DMLA). Les objectifs des travaux présentés dans cette thèse consistent à déterminer : 1) le mécanisme par lequel le B1R est impliqué dans la rétinopathie diabétique chez le rat; 2) l’implication de la iNOS en aval dans la cascade inflammatoire activée par le B1R; 3) l’expression et la localisation cellulaire du B1R dans les rétines humaines atteintes de DMLA exsudative et atrophique. Nos résultats ont permis de démontrer une implication du B1R dans la rétinopathie diabétique via l’activation de l’enzyme de synthèse du monoxyde d’azote inductible (iNOS) dans un modèle de diabète de type 1 induit par la streptozotocine (STZ) chez le rat. En plus de sa localisation généralisée dans toute la rétine, le B1R est exprimé dans la couche de l’épithélium pigmentaire qui forme la barrière hémato-rétinienne externe. Les taux d’expression (protéique et ARNm) du B1R, de la iNOS, de la carboxypeptidase M (impliquée dans la biosynthèse des agonistes B1R), de l'IL-1β, du TNF-α, du facteur de croissance de l'endothélium vasculaire A (VEGF-A) et de son récepteur, le VEGF-R2, ainsi que des protéines nitrosylées augmentent à deux semaines dans la rétine diabétique. Ces augmentations ainsi que l’hyperperméabilité vasculaire rétinienne induite par le diabète et par l’injection intravitréenne d’un agoniste du B1R (R-838) sont bloquées par un inhibiteur de la iNOS (1400W) appliqué topiquement à la surface de l’œil pendant 1 semaine (premier article). Les résultats du deuxième article montrent une augmentation significative de l'immunoréactivité du B1R dans les rétines humaines prélevées de patients atteints de DMLA exsudative. Toutefois, les changements d’immunoexpression du B1R ne sont pas significatifs dans les rétines des patients atteints de DMLA atrophique. La réactivité des cellules gliales est plus marquée dans la forme exsudative que dans la forme atrophique de DMLA. Une colocalisation du B1R est observée avec des marqueurs des cellules de Müller, des astrocytes, de la microglie, de la iNOS et de la fibrose, suggérant une implication du B1R dans le processus inflammatoire et la formation de fibrose dans la DMLA exsudative. En revanche, l’expression du B2R demeure stable dans les rétines de DMLA exsudative et atrophique par rapport aux rétines témoins; ce résultat ne supporte pas la possibilité que ce récepteur puisse être impliqué dans la DMLA chez l’humain. / The kallikrein-kinins system is a peptidergic system involved in inflammatory processes, the control of the vascular tone and permeability. These effects are mediated by two G proteincoupled receptors, the Bradykinin type 1 (B1R) and type 2 (B2R) receptors. While the B2R is a constitutive receptor, B1R is almost undetectable in physiological condition; it is, however, induced by oxidative stress, pro-inflammatory cytokines (interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α)) or by bacterial endotoxins at both systemic and local levels, notably in the retina. Recent studies from our laboratory supported an implication of B1R in the pathogenesis and progression of diabetic retinopathy and age-related macular degeneration (AMD). This thesis aims at unraveling: 1) the mechanism by which B1R is involved in diabetic retinopathy in rats; 2) the involvement of iNOS in the inflammatory cascade downstream to the B1R; and, 3) the expression and cellular localization of B1R in human retinae with exudative and atrophic AMD. Our results have shown the implication of B1R in diabetic retinopathy via the activation of the inducible nitric oxide synthase (iNOS) in a type 1 model of diabetes induced by streptozotocin (STZ) in rats. In addition to its generalized localization throughout the retina, B1R is expressed in the retinal pigment epithelium which forms the outer blood-retinal barrier. The protein and transcript expression of inflammatory markers; iNOS, carboxypeptidase M, IL-1β, TNF-α, vascular endothelium growth factor A (VEGF-A) and its receptor, VEGF-R2, including B1R as well as nitrosylated proteins are increased in the retina of diabetic rats at 2 weeks post-STZ. These upregulations, as well as the retinal vascular hyperpermeability induced by diabetes and by the intravitreal injection of an B1R agonist (R-838) are blocked by a topical one-week treatment by eye-drop with the selective iNOS inhibitor (1400W) (first manuscript). The results of the second manuscript show significant increases in the immunoreactivity of B1R in exudative AMD retinae. Despite a slight increase, B1R immunostaining does not reach statistical significance in the retina of donors with atrophic AMD. The reactivity of glial cells is more impressive in the exudative than in the atrophic form of AMD. B1R is co-expressed with markers of Müller cells, astrocytes, microglia, iNOS and fibrosis, suggesting an involvement of B1R in the inflammatory events and the formation of fibrosis in exudative AMD. On the other hand, the expression of B2R remains stable in the retinae of exudative and atrophic AMD, supporting a secondary role of this receptor in AMD in humans.

In this work I present conclusions of clinical-laboratory research focused on the patients with diabetic macular edema (DME). We performed biochemical and immunochemical analyses of vitreous samples that were collected during the pars plana vitrectomy. Moreover, at patients with non-proliferative diabetic retinopathy (NPDR) we assessed morphological characteristics of DME using optical coherence tomography (OCT). According to our findings, the vitreous and serum concentrations of uric acid and glucose were significantly higher in patients with diabetic retinopathy and DME compared to controls. Also total ratio (serum/ vitreous concentration) of uric acid and glucose was in diabetics significantly higher than in controls. The most important determinant of increasing concentration of both uric acid and glucose in the vitreous was the grade of diabetic retinopathy. Moreover, we demonstrated significant correlation between vitreous concentration of uric acid and concentration of the vascular endothelial growth factor (VEGF) in patients with DME and NPDR. We found further, that the volume of the macula (cube volume - CV) computed with the software of Cirrus HD-OCT correlates in diabetics significantly with the vitreous VEGF concentration, but not with uric acid. This OCT parameter could be used to...

La dégénérescence maculaire liée à l’âge (DMLA) est l’une des principales causes de la perte de vision chez les personnes âgées. Dans la DMLA de forme sèche, le stress oxydatif, l’inflammation et la dysfonction lipidique causent la perte des cellules de l’épithélium pigmentaire rétinien (RPE) qui sont essentielles au maintien des photorécepteurs. Nous avons précédemment démontré qu’un ligand sélectif du CD36 permet de préserver la fonction rétinienne dans un modèle murin de la DMLA. Cependant, l’effet des ligands synthétiques du CD36 dans la protection du RPE n’a jamais été vérifié. L’objectif de cette étude était de caractériser l’effet cytoprotecteur du ligand du CD36 MPE-001 sur le RPE contre le stress oxydatif. Le stress oxydatif a été induit chez la lignée humaine hTERT RPE-1 par le NaIO3. Il a été observé que le MPE-001 diminue la production de superoxydes mitochondriaux et l’apoptose des cellules du RPE sans toutefois affecter le système antioxydant au niveau transcriptionnel. Des essais par immunobuvardage et par immunocytochimie ont montré que le NaIO3 perturbe le flux autophagique, alors que le MPE-001 le rétablit. L’effet protecteur du MPE-001 était complètement aboli par les inhibiteurs de l’autophagie wortmannin et bafilomycine A1. En conclusion, nous avons démontré pour la première fois qu’un ligand du CD36 protège les cellules du RPE contre le stress oxydatif de façon autophagie-dépendante. Nous proposons la modulation du stress oxydatif chez le RPE par l’activation du CD36 comme stratégie potentielle pour traiter la DMLA de forme sèche. / Age-related macular degeneration (AMD) is a leading cause of irreversible blindness in the elderly population. The retinal pigment epithelium (RPE) is a monolayer of epithelial cells that are critical in maintaining retinal neurons. Oxidative stress, inflammation and defects in lipid clearance are thought to cause damages to the RPE and the eventual loss of photoreceptors in the dry form of AMD. CD36 is a scavenger receptor that plays an important role in inflammation and lipid homeostasis. We have previously shown that stimulation of CD36 by a selective ligand preserves photoreceptor function in high fat high cholesterol diet-fed Apoe-/- mice, a mouse model of AMD. The effect of CD36 ligands on RPE cells protection from oxidative insults, however, had yet to be investigated. In this study, we aimed to study the cytoprotective effect of the CD36 ligand MPE-001 in RPE cells exposed to oxidative stress. Oxidative stress was induced in the human RPE cell line hTERT RPE-1 using NaIO3. MPE-001 was observed to decrease NaIO3-induced mitochondrial superoxide production and apoptosis, with no transcriptional effect on antioxidant enzymes. Immunoblotting and immunostaining showed that NaIO3 disrupts autophagic flux while MPE-001 co-treatment restores it. The protective effect of MPE-001 was completely abolished by the autophagy inhibitors wortmannin and bafilomycin A1. In conclusion, we report for the first time that a CD36 ligand confers protection to RPE cells under oxidative stress through the improvement of autophagic process. We therefore propose modulation of oxidative stress by CD36 ligands as a potential strategy to treat dry AMD.

La pathologie numérique s’est développée ces dernières années grâce à l’avancée récente des algorithmes d’analyse d’images et de la puissance de calcul. Notamment, elle se base de plus en plus sur les images histologiques. Ce format de données a la particularité de révéler les objets biologiques recherchés par les experts en utilisant des marqueurs spécifiques tout en conservant la plus intacte possible l’architecture du tissu. De nombreuses méthodes d’aide au diagnostic à partir de ces images se sont récemment développées afin de guider les pathologistes avec des mesures quantitatives dans l’établissement d’un diagnostic. Les travaux présentés dans cette thèse visent à adresser les défis liés à l’analyse d’images histologiques, et à développer un modèle d’aide au diagnostic se basant principalement sur les relations spatiales, une information que les méthodes existantes n’exploitent que rarement. Une technique d’analyse de la texture à plusieurs échelles est tout d’abord proposée afin de détecter la présence de tissu malades dans les images. Un descripteur d’objets, baptisé Force Histogram Decomposition (FHD), est ensuite introduit dans le but d’extraire les formes et l’organisation spatiale des régions définissant un objet. Finalement, les images histologiques sont décrites par les FHD mesurées à partir de leurs différents types de tissus et des objets biologiques marqués qu’ils contiennent. Les expérimentations intermédiaires ont montré que les FHD parviennent à correctement reconnaitre des objets sur fonds uniformes y compris dans les cas où les relations spatiales ne contiennent à priori pas d’informations pertinentes. De même, la méthode d’analyse de la texture s’avère satisfaisante dans deux types d’applications médicales différents, les images histologiques et celles de fond d’œil, et ses performances sont mises en évidence au travers d’une comparaison avec les méthodes similaires classiquement utilisées pour l’aide au diagnostic. Enfin, la méthode dans son ensemble a été appliquée à l’aide au diagnostic pour établir la sévérité d’un cancer via deux ensembles d’images histologiques, un de foies métastasés de souris dans le contexte du projet ANR SPIRIT, et l’autre de seins humains dans le cadre du challenge CPR 2014 : Nuclear Atypia. L’analyse des relations spatiales et des formes à deux échelles parvient à correctement reconnaitre les grades du cancer métastasé dans 87, 0 % des cas et fourni des indications quant au degré d’atypie nucléaire. Ce qui prouve de fait l’efficacité de la méthode et l’intérêt d’encoder l’organisation spatiale dans ce type d’images particulier. / During the last decade, digital pathology has been improved thanks to the advance of image analysis algorithms and calculus power. Particularly, it is more and more based on histology images. This modality of images presents the advantage of showing only the biological objects targeted by the pathologists using specific stains while preserving as unharmed as possible the tissue structure. Numerous computer-aided diagnosis methods using these images have been developed this past few years in order to assist the medical experts with quantitative measurements. The studies presented in this thesis aim at adressing the challenges related to histology image analysis, as well as at developing an assisted diagnosis model mainly based on spatial relations, an information that currently used methods rarely use. A multiscale texture analysis is first proposed and applied to detect the presence of diseased tissue. A descriptor named Force Histogram Decomposition (FHD) is then introduced in order to extract the shapes and spatial organisation of regions within an object. Finally, histology images are described by the FHD measured on their different types of tissue and also on the stained biological objects inside every types of tissue. Preliminary studies showed that the FHD are able to accurately recognise objects on uniform backgrounds, including when spatial relations are supposed to hold no relevant information. Besides, the texture analysis method proved to be satisfactory in two different medical applications, namely histology images and fundus photographies. The performance of these methods are highlighted by a comparison with the usual approaches in their respectives fields. Finally, the complete method has been applied to assess the severity of cancers on two sets of histology images. The first one is given as part of the ANR project SPIRIT and presents metastatic mice livers. The other one comes from the challenge ICPR 2014 : Nuclear Atypia and contains human breast tissues. The analysis of spatial relations and shapes at two different scales achieves a correct recognition of metastatic cancer grades of 87.0 % and gives insight about the nuclear atypia grade. This proves the efficiency of the method as well as the relevance of measuring the spatial organisation in this particular type of images.

Indiana University-Purdue University Indianapolis (IUPUI) / Choroidal neovascularization (CNV) is the vision-threatening characteristic of wet age-related macular degeneration (AMD), a major cause of blindness affecting almost 2 million elderly Americans. The current approved treatments target the dominant angiogenic mediator, vascular endothelial growth factor (VEGF). However, repeated injections of anti-VEGF drugs can cause ocular and systemic side effects, and about 30% of wet AMD patients are non-responsive. There is thus an unmet need to develop VEGF-independent antiangiogenic molecules to complement or combine with existing medications. I studied SH-11037, a novel homoisoflavonoid with potent and selective antiangiogenic activity against human retinal endothelial cells. Intravitreal SH- 11037 dose-dependently suppressed angiogenesis in the laser-induced CNV (LCNV) mouse model. These effects were prominent as early as 7 days post-laser treatment as measured by a novel ellipsoid quantification method of optical coherence tomography images in vivo. A supratherapeutic dose of 100 μM SH- 11037 was not associated with signs of murine ocular toxicity, and did not interfere with pre-existing retinal vasculature or retinal function. SH-11037 synergized with anti-VEGF therapy in vitro and in vivo, suggesting a VEGFindependent mechanism. By photoaffinity pulldown, I identified soluble epoxide hydrolase (sEH) as an SH-11037-binding target. sEH is a key enzyme in ω-3 and ω-6 fatty acid metabolism. sEH levels were dramatically upregulated in retinal sections from L-CNV mice and a specific sEH inhibitor, t-AUCB, significantly suppressed L-CNV lesion volume. Additionally, SH-11037 inhibited sEH enzymatic activity in vitro and in vivo in L-CNV mice. Given the role of sEH in the metabolism of docosahexaenoic acids (DHA), inhibition of sEH using small molecules like SH-11037 would enhance ocular DHA levels, with beneficial antiangiogenic and anti-inflammatory effects. SH-11037 is thus a novel sEH inhibitor, which could make it an alternative or additive therapy to existing anti- VEGF drugs for treatment of neovascular diseases in the eye and other tissues.

No description available.

Optical Coherence Tomography (OCT) has been around for more than 30 years and is still being continuously improved. The department of ophthalmology is a part of Sahlgrenska Hospital that heavily uses OCT for helping people with the treatment of eye diseases. They are currently facing a problem where the time to go from an OCT scan to treatment is being increased due to having an overload of patient visits every day. Since it requires a trained expert to analyze each OCT scan, the increase of patients is too overwhelming for the few experts that the department has. It is believed that the next phase of this medical field will be through the adoption of machine learning technology. This thesis has been issued by Sahlgrenska University Hospital (SUH), and they want to address the problem that ophthalmology has by introducing the use of machine learning into their workflow. This thesis aims to determine the best suited CNN through training and testing of pre-trained models and to build a tool that a model can be integrated into for use in ophthalmology. Transfer learning was used to compare three different types of pre-trained models offered by Keras, namely VGG16, InceptionResNet50V2 and ResNet50V2. They were all trained on an open dataset containing 84495 OCT images categorized into four different classes. These include the three diseases Choroidal Neovascularization (CNV), Diabetic Macular Edema (DME), drusen and normal eyes. To further improve the accuracy of the models, oversampling, undersampling, and data augmentation were applied to the training set and then tested in different variations. A web application was built using Tensorflow.js and Node.js that the best-performed model later was integrated into. The VGG16 model performed the best with only oversampling applied out of the three. It yielded an average of 95% precision, 95% recall and got a 95% F1-score. The second was the Inception model with only oversampling applied that got an average of 93% precision, 93% recall and a 93% F1-score. Last came the ResNet model with an average of 93% precision, 92% recall and a 92% F1-score. The results suggest that oversampling is the overall best technique for this given dataset. The chosen data augmentation techniques only lead to models performing marginally worse in all cases. It also suggests that pre-trained models with more parameters, such as the VGG16 model, have more feature mappings and, therefore, achieve higher accuracy. On this basis, parameters and better mappings of features should be taken into account when using pre-trained models.