Focusing on metabolomic dysregulation and modulation of retinal metabolism to develop novel therapeutic strategies for diabetic retinopathy

Abstract

Diabetic retinopathy (DR) is a neurovascular complication of diabetes mellitus and a leading cause of blindness in adults below the age of 65, in industrialized nations. Currently there are 126.6 million people with DR worldwide (34.6% of the total diabetic population) and it is estimated that this number will increase to 191.0 million by 2030. Generally, DR is divided into two stages: non-proliferative diabetic retinopathy (NPDR), an earlier phase characterized by appearance of microaneurysms, dot and blot hemorrhages, capillary occlusions and nerve fiber layer infarcts and; and proliferative diabetic retinopathy (PDR), the late-stage disease, which is diagnosed when pathological neovascular changes are identified on the retinal surface and/or vitreous. Diabetic macular edema (DME) can develop at any stage and reflects a pathological increase in retinal vascular permeability. Despite its high prevalence, current availability of preventive and therapeutic strategies is far from ideal. In fact, there are no reliable biomarkers to predict risk of developing DR and no effective, targeted and early-acting therapies to sustainably and safely prevent disease progression into its vision threatening stages: PDR and DME. The two main therapeutic options currently available for DR are laser photocoagulation and intravitreal injections of anti-vascular endothelial growth factor (VEGF) agents; these constitute “non-selective” and destructive (especially for laser therapy) approaches that are not only unable to effectively and sustainably arrest retinal disease progression in every case, but can also potentially induce a myriad of undesired off-target effects at the retinal level and even systemically, in the case of VEGF antagonists. Moreover, they act late in the disease course. The lack of reliable rodent models - there is no diabetic mouse model that spontaneously recapitulates the late stages of DR - has greatly hindered research progress and development of novel and effective drugs for PDR, further contributing to the present therapeutic scenario. In this dissertation I will introduce and present my experimental work in the context of the following concepts, which can potentially lead to development of targeted, earlier acting, less destructive and more effective future therapies for DR: (1) DR has long been regarded as a vascular disease and present-day DR management guidelines are still based on this assumption. However, a growing body of evidence shows that retinal neuronal function becomes impaired before vascular changes can be detected; these findings along with those showing that adequate retinal functioning depends on stable intercellular interactions within neurovascular units, suggest that disrupting retinal neurovascular crosstalk may play a critical role in promoting disease development and progression. (2) Metabolomic studies have been surprisingly neglected in the investigation of DR’s pathophysiology, and this is clearly reflected by the fact that the metabolome of human DR remains unknown. Furthermore, the neuroretina is one of the most metabolically demanding tissues in the body per unit weight, and diabetes is triggered by a metabolic defect that profoundly impairs cellular energy production. These features constitute a potentially disastrous combination in regard to retinal functioning and suggest that studying retinal energy metabolism in DR is critical. (3) Metabolic cycles of photoreceptors, interneurons and glial cells are still under debate and, even though it is known that intercellular communications within the neurovascular unit (NVU) are mediated by metabolites whose production becomes dysregulated under pathological conditions, the precise mechanisms underlying retinal neurovascular coupling are not fully identified. Gaining further insight into these interactions is pivotal because retinal NVUs are responsible for regulating blood flow for functionally dynamic retinal neuronal networks and, thus, for their proper functioning. Besides the points stated above, additional clinical clues were considered to guide the research plan presented in this dissertation. One the strongest came from a subset of long-term diabetic patients who appear to be protected from developing late-stage DR, by an unknown mechanism. Studying these patients provides an excellent opportunity to identify protective factors and to further understand the mechanisms involved in progression of DR. In order to better understand how metabolic dysregulation impacts development of DR, how neurovascular interactions become compromised in the diabetic retina, and to develop strategies to potentially restore homeostasis within the NVU, we decided to use metabolomic analyses. A highly sensitive metabolomics mass-spectrometry based approach was used in ocular and serum samples to identify the most prominent metabolic perturbations, to acquire a global overview of the metabolomic landscape in late-stage DR and to identify potentially protective circulating factors. At the ocular level, late-stage diabetic retinopathy was associated with severe dysregulation in amino acid levels; this was especially prominent in those generated during arginine metabolism, suggesting a preferential activity in the arginase pathway over the alternative Nitric Oxide Synthase (NOS) pathway; Studies in the Oxygen-Induced-Retinopathy (OIR) mouse, a non-diabetic model that develops features of ischemic retinopathy, revealed a very similar metabolic landscape and, in vivo global isotope analysis confirmed the presence of asymmetrical arginine metabolism by showing: (1) over-activity in the arginase pathway leading to enhanced proline production; and (2) reduced activity in the alternative NOS pathway, with potentially reduced NO production. Even though NO’s role in DR and other retinopathies is not clearly understood, NO is known to be an important modulator of cellular interactions within the NVU and its lower availability in specific locations and/or time-points in pathological conditions may significantly contribute to the disruption of retinal neurovascular crosstalk. The work presented in this dissertation has also described novel functions for interneurons (amacrine and horizontal cells) and photoreceptors within the NVU, by showing that these cells play an active role in regulating their primary vasculature and thus, their blood supply. Furthermore, it has also shown that dysfunction of retinal neurons in this capacity can directly alter their own blood supply, therefore providing additional clues for disrupted retinal neurovascular crosstalk. Metabolomic analyses comparing serum samples from diabetic patients with or without PDR (long-term diabetic patients “protected” from late-stage DR versus those who were non-protected) revealed that “protected” patients had higher circulating levels of a purine metabolite, inosine. To assess its therapeutic potential in conditions of retinal ischemia, inosine was delivered to the eye of the OIR mouse where it enhanced effective revascularization of ischemic retinal areas, thus significantly reducing pathological neovascularization. These effects were associated with a favorable modulation of the local pro-inflammatory response that could result from an improved overall retinal metabolic status. These beneficial effects on retinal metabolism induced by inosine injections were observed as: (1) a reduction in basal mitochondrial respiration in vaso-obliterated areas (i.e., ischemic areas), which can potentially increase retinal neuronal tolerance to hypoxia by reducing the metabolic mismatch created by scarce metabolic supply and high neuronal demand; and (2) a reduction in proline production, suggesting antagonism of the arginase pathway (which is hyperactive in oxygen-induced-retinopathy and potentially in human PDR). In summary, the work presented in this dissertation employed a metabolomic-focused approach with a strong focus on neurovascular crosstalk to answer two intriguing questions regarding DR: (1) What is the characteristic ocular metabolic landscape of severe DR? (2) Is the “protection” against severe DR (observed in some long-term diabetic patients) associated with differences in circulating metabolic factors? The answers to these questions, presented below, could serve as the basis of future targeted, more effective and earlier acting therapeutics that would revolutionize DR patient management. Identification of the most prominently affected metabolic pathways in eyes with severe DR has identified specific pathways of amino acid metabolism as potential targets for development of new drugs for DR. We have also identified a circulating protective factor, inosine, in “protected” patients and further investigated its ability to (1) prevent development of retinal ischemia and pathological neovascularization; (2) adjust retinal metabolism to the limited energy supplies in ischemic areas; and (3) counteract development of prominent metabolic dysregulation by potentially inhibiting the pathology-promoting arginase pathway. We believe that inosine can potentially become an effective, early-acting therapeutic agent to prevent progression of DR. In addition, these metabolites could potentially be used as reliable biomarkers for monitoring response to therapy and for predicting risk of developing or progressing DR. Finally, the work presented in this dissertation supports the concept that early intervention for treating DR will restore balance and stabilize cellular interactions within the NVU, thereby reversing the chronic stressors (e.g., extreme conditions of metabolic insufficiency in retinal ischemic areas) that ultimately drive development and progression of retinal pathology.