Phenotypes and biomarkers of diabetic retinopathy.

Progress in Retinal and Eye Research xxx (2014) 1e22

Contents lists available at ScienceDirect

Progress in Retinal and Eye Research journal homepage: www.elsevier.com/locate/prer

Phenotypes and biomarkers of diabetic retinopathy José Cunha-Vaz a, b, *,1, Luisa Ribeiro a, b,1, Conceição Lobo a, b,1 a b

Association for Innovation and Biomedical Research on Light and Image (AIBILI), Azinhaga de Santa Comba, 3000-548 Coimbra, Portugal Faculty of Medicine, University of Coimbra, Polo III, Azinhaga de Santa Comba, 3000-548 Coimbra, Portugal

a r t i c l e i n f o

a b s t r a c t

Article history: Available online xxx

Diabetic retinopathy (DR) remains a major cause of blindness as the prevalence of diabetes is expected to approximately double globally between 2000 and 2030. DR progresses over time at different rates in different individuals with only a limited number developing significant vision loss due to the two major vision-threatening complications, clinically significant macular edema and proliferative retinopathy. Good metabolic control is important to prevent and delay progression, but whereas some patients escape vision loss even with poor control, others develop vision loss despite good metabolic control. Our research group has been able to identify three different DR phenotypes characterized by different dominant retinal alterations and different risks of progression to vision-threatening complications. Microaneurysm turnover has been validated as a prognostic biomarker of development of clinically significant macular edema, whereas subclinical macular edema identified by OCT and mfERG appear to be also good candidates as organ-specific biomarkers of DR. Hemoglobin A1c remains the only confirmed systemic prognostic biomarker of DR progression. The availability of biomarkers of DR progression and the identification of different phenotypes of DR with different risks for development of vision-threatening complications offers new perspectives for understanding DR and for its personalized management. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Diabetic retinopathy Macular edema Phenotypes Biomarkers Microaneurysm Optical coherence tomography

Contents 1. 2. 3.

4. 5.

6.

Epidemiology and socio-economic impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pathogenesis of diabetic retinopathy: an overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clinical diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Nonproliferative diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Vision threatening complications of diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Proliferative diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Clinically significant diabetic macular edema . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Progression. Natural history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Complexity and specificity of retinal damage in diabetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Human macular region. Relevance for diabetic retinopathy and vision loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Retinal neurovascular damage in diabetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1. The retinal neurovascular unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2. Neuropathy. Increased apoptosis and degeneration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.3. Blood-retinal barrier and angiogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.4. Inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Phenotypes of diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Characterization of phenotypes of diabetic retinopathy progression using non-invasive methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

Abbreviations: DR, diabetic retinopathy; ETDRS, early treatment diabetic retinopathy study; HbA1c, hemoglobin A1c; MA, microaneurysm; mfERG, multifocal electroretinography; OCT, optical coherence tomography. * Corresponding author. AIBILI, Azinhaga de Santa Comba, 3000-548 Coimbra, Portugal. Tel.: þ351 239480136; fax: þ351 239480117. E-mail addresses: [email protected] (J. Cunha-Vaz), [email protected] (L. Ribeiro), [email protected] (C. Lobo). 1 Percentage of work contributed by each author in the production of the manuscript is as follows: José Cunha-Vaz: 80%; Luisa Ribeiro: 10%; Conceição Lobo: 10. http://dx.doi.org/10.1016/j.preteyeres.2014.03.003 1350-9462/Ó 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Cunha-Vaz, J., et al., Phenotypes and biomarkers of diabetic retinopathy, Progress in Retinal and Eye Research (2014), http://dx.doi.org/10.1016/j.preteyeres.2014.03.003

2

J. Cunha-Vaz et al. / Progress in Retinal and Eye Research xxx (2014) 1e22

6.2.

7.

8.

Suggested predominant disease mechanisms in different phenotypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.1. Accelerated Apoptosis. Vasoregression and pericyte loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.2. Breakdown of the blood-retinal barrier and inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.3. Hypoxia and angiogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3. Integrated perspective of diabetic retinopathy progression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomarkers of diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1. Definitions of biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2. Prognostic biomarkers of diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.1. Systemic disease-related biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.2. Genetic disease-related biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.3. Organ-specific disease-related biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3. Predictive biomarkers of therapeutic response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.1. Degree of decrease of OCT central retinal thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4. Prognostic and predictive biomarkers of diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflicts of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Epidemiology and socio-economic impact Diabetes according to most recent estimates affects 8.3% of adults, 382 million people, and the number of people with the disease is set to rise beyond 592 million in less than 25 years. Yet, with 175 million of cases currently undiagnosed, a vast amount of people with diabetes are progressing towards complications unaware of them. Diabetic retinopathy (DR) is a frequent complication of diabetes and may lead to blindness, making it one of the most feared complications of diabetes. Indeed, DR is the leading cause of vision loss in working age adults (International Diabetes Federation, 2013). The incidence of DR increases with the duration of diabetes, and after 20 years, nearly all patients with type 1 diabetes and more than 60% of those with type 2 diabetes will develop DR (Fong et al., 2004a,b). Vision-threatening complications of DR, proliferative retinopathy and clinically significant macular edema are, however, much less frequent. DR prevalence in the diabetic population is around one-third with one-tenth presenting vision-threatening states such as diabetic macular edema or proliferative DR (Yau et al., 2012). Recently, data collected from 22.896 individuals from 35 studies in the US, Australia, Europe and Asia, showed that the overall agestandardized prevalence of any DR was 34,6%, proliferative DR was 7.0%, diabetic macular edema 6.8% and vision threatening DR was 10.2%, thus confirming the most recent data available (Yau et al., 2012). The impact of visual impairment goes beyond the individual; communities and economics lose caring capacity and productivity and a need is created for increased social support. Data for the year of 2008 state that the economic costs of partial sight and blindness on the United Kingdom total £22 billion, with direct health care costs amounting to £2.14 billion. The research estimated that there were a total of 1.8 million people with partial sight and blindness in the UK adult population in 2008, with 3.5% from DR (Royal National Institute of Blind People. http/www.mib. org.uk). Early detection and treatment of DR could potentially avoid or reduce these costs. In a study to estimate the prevalence of healthcare costs of DR in Sweden, the prevalence of any DR was 41.8% for patients with diabetes type 1 and 27.9% for patients with diabetes type 2. The annual cost of DR was calculated as 106.000V per 100.000 inhabitants (Heintz et al., 2010). In the United States, screening and treatment of eye diseases in patients with diabetes mellitus costs $3190 per quality-adjusted life-year (QALY) saved.

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

This cost is a weighted average (based on prevalence of different subtypes diabetic patients) of the cost-effectiveness of detecting and treating diabetic eye disease in those with insulin dependent diabetes mellitus ($1996 per QALY), those with noninsulin dependent diabetes mellitus who use insulin for glycemic control ($2933 per QALY), and those with noninsulin diabetes mellitus who do not use insulin for glycemic control ($3530 per QALY) (Javitt and Aiello, 1996). Comparing resource utilization and medical costs over a 12month period diagnosed with macrovascular and microvascular complications in diabetes in a US study (Pelletier et al., 2009) the average total costs per patient over 12 months were US$14,414 where complications were present versus US$8.609 without complications. Among the 15.326 patients included in this study, 61% had peripheral neuropathy, 28% diabetic retinopathy and 19% nephropathy. DR is a frequent complication of diabetes, and, therefore, has tremendous impact on society. DR is a particularly good example of a disease that needs to be addressed on a perspective that includes planning for prospective health care (Snyderman and Williams, 2003). In summary, diabetes is clearly a major health problem which is affecting progressively more people all over the world as a result of the progressive aging of the population and the spread of obesity due to an increased sedentary lifestyle and unhealthy eating habits. This is even more relevant when it is realized that DR, now the most frequent cause of blindness in working age adults in the western world will progressively involve similarly the entire world as the economic conditions improve in poorer parts of the world. Diabetic macular edema is the most frequent complication of DR and the most common cause of vision loss due to diabetes. Identifying the eyes/patients at risk to develop clinically significant macular edema and visual loss and understanding its causes and development is fundamental for its appropriate treatment and, finally, to avoid vision loss due to diabetes. 2. Pathogenesis of diabetic retinopathy: an overview Diabetic retinopathy is said to be present when microaneurysms and small hemorrhages appear on ophthalmoscopic examination. On histopathological examination, the vascular changes are initiated in the small vessels in the form of endothelial proliferation, pericyte damage and microaneurysms formation (Ashton, 1958).



These initial lesions are focal and located at the posterior pole of the retina. At first the endothelial proliferation and microaneurysms appear to be confined to the venous side of the retinal circulation, whereas at this stage endothelial degeneration appear to be limited to capillaries on the arterial side. The pericytes appear selectively involved, but their association with microvascular changes is highly irregular (Cunha-Vaz, 1978). With progression of disease the capillaries of the arterial side of the retinal circulation show increased vasoregression, with cell loss and closure. Simultaneously, the number of microaneurysms increase and the areas of capillary closure enlarge. As they enlarge they are seen to be crossed by remaining enlarged capillaries, which appear to act as arteriovenous shunts, receiving the blood directed from the surrounding closed capillary net. These microvascular lesions, microaneurysms, capillary closure, basement membrane thickening and pericytic damage are different from other vascular retinopathies only in their intensity and distribution. While the vascular involvement in other retinopathies is localized initially either in the arterial or in the venous side of the retinal circulation, in diabetic retinopathy both sides of the posterior pole retinal vascular tree are involved, although focally, from the beginning (Cunha-Vaz, 1978). What remains to be fully understood is the mechanism involved in triggering these microvascular lesions. Hyperglycemia appears to be sufficient to initiate the development of DR as revealed by the development of retinopathy in animals experimentally made hyperglycemic (Engerman and Kern, 1984; Kador et al., 1990; Kern and Engerman, 1996). Consistently, a number of experimental studies have shown that intensive therapy sufficient to minimize hyperglycemia inhibits the development of retinopathy (Engerman et al., 1977; Engerman and Kern, 1993). Excessive transport of glucose or concentration of glucose within cells of the retina is a common thread underlying most of the biochemical and molecular mechanisms that have been postulated to play a role in the pathogenesis of diabetic retinopathy. However, the observation that not all patients with poor metabolic control develop advanced stages of diabetic retinopathy suggests that other factors, such as genetic predispositions, are likely to determine individual susceptibility to the disease. It is relevant that diabetes is associated with generalized endothelial dysfunction and inflammation as manifested by a universal increased transvascular passage of macromolecules (Zeeuw, 2007; Feldt-Rasmussen, 2007). The retina consists of three major types of cells: neurons, glial cells and blood vessels and most, if not all, of these cell types are affected to some degree in diabetic retinopathy. The retina is primarily a neuronal tissue. Indeed, neurons and glial cells comprise about 95% of the retinal mass. The glial cells of the retina, Muller cells and astrocytes serve as support cells for the neurons and blood vessels. The inner blood-retinal barrier, which is a specific and a unique structure in the retina, and is affected early in diabetes, is formed by the retinal neurovascular unit, assembling in a functional unit the endothelial cells, the pericytes and the glial cells immediately surrounding the retinal vessels (Antonetti et al., 2006; Cunha-Vaz, 1979). The earliest alterations that may be detected clinically in the retina in diabetes are the breakdown of the blood-retinal barrier and alterations in the neurosensory retinal function. Both these alterations can be detected before ophthalmoscopic signs of diabetic retinopathy are visible, in preclinical retinopathy (Cunha-Vaz et al., 1975; Daley et al., 1987). These and other observations indicate that dysfunction of the inner retina may affect neurons and glial cells and induce changes in the neurovascular unit and blood-retinal barrier before obvious

3

signs of vascular lesions (Antonetti et al., 2012). A specific role for the pericyte damage resulting in loss of regulation of retinal vascular tone has been also proposed (Ejaz et al., 2008; Hammes, 2005). Oxygen consumption and metabolic activity in the retina is one of the highest in the body. The neuroretina is nourished by transport of glucose across the endothelial cells across the inner bloodretinal barrier and from the choroidal vessels across the retinal pigment epithelium of the outer blood-retinal barrier (Kumagai, 1999). Mechanisms that may concur in diabetic retinopathy to explain toxicity of glucose in diabetes include: activation of protein kinase C (PKC), activation of aldose reductase, formation of advanced glycation end products and the hexosamine pathway. These different mechanisms share a common thread. Hyperglycaemia induces overproduction of superoxide by the mitochondrial electron transport chain (Du et al., 2000; Nishikawa et al., 2000; Brownlee, 2001) and this excess production of reactive oxygen species (ROS) can be the upstream event leading to retinal damage. This unifying hypothesis is consistent with the four pathways suggested to be involved in the development of diabetic complications (activation of aldose reductase, increased formation of AGEs, activation of protein kinase C and increased hexosamine pathway flux). This hypothesis further accounts for the relevance of increased production of reactive oxygen species in diabetes and also provides a unifying hypothesis regarding the effects of hyperglycaemia on cellular dysfunction (Nishikawa et al., 2000; Brownlee, 2001). Increased production of superoxide may further interfere with NO function in the retina and, as mentioned before, may even result in an increased production of superoxide by a mechanism that does not involve the mitochondria, thus amplifying the production of reactive oxygen species. NO is generated from the metabolism of Larginine by the enzyme nitric oxide synthase (NOS). Three known isoforms of the enzyme, the constitutive brain (bNOS) and endothelial (eNOS) isoforms and an inducible isoform (iNOS). A number of stimuli were shown to induce iNOS, including hyperglycaemia (Baek et al., 1993; Leal et al., 2012). In addition, several growth factors including VEGF.A and PDGF are induced by the diabetic milieu and may play a relevant role in the early stages of DR (Cui et al., 2007; Van Gueest et al., 2010). VEGF-A is the best studied inducer of retinal vascular permeability (Klaassen et al., 2013; Schlingemann and van Hinsbergh, 1997). It is also of major relevance that retinal neurodegeneration participates in the development of the early microvascular changes that occur in DR such as the breakdown of the blooderetinal barrier (BRB) (Kusari et al., 2007; Liu et al., 2012; Silva et al., 2009), vasoregression (Feng et al., 2009), and the impairment of neurovascular coupling (Lecleire-Collet et al., 2011; Luu et al., 2010). The relationship between the excitotoxicity mediated by glutamate and the breakdown of the BRB induced by VEGF is one of the most interesting pathways linking neurodegeneration with vascular impairment. In this regard, it has been demonstrated that hyperglycemia induces an increase in extracellular glutamate, and the subsequent overactivation of NMDA receptors mediates VEGF production and BRB breakdown (Imai et al., 2009; Kusari et al., 2010). Glial dysfunction also has an essential role in this pathophysiological event (Shen et al., 2010) and it has been shown that it can contribute to the disruption of the BRB directly or through the upregulation of VEGF (Cervia et al., 2012; Chen et al., 2006; Hellström et al., 1999; Mei et al., 2012; Mori et al., 2002; Shen et al., 2011; Wilson et al., 2001; Zhang et al., 2006). The impairment of neurovascular coupling appears, therefore, to be an early event in DR.


4


Finally a large body of evidence supports the role of inflammation in the pathogenesis of DR (Blom et al., 2011; Joussen et al., 2004; Tang and Kern, 2011). It has been associated with the universal endothelial dysfunction occurring in diabetes as a repair process (Feldt-Rasmussen, 2007). An emerging issue in DR research is the focus on the mechanistic link between the activation of subclinical inflammation and neurodegeneration. In this regard, it has been shown that Müller cells show inflammation-linked responses when exposed to the diabetic milieu (Gerhardinger et al., 2005; Kern, 2007). In addition, it has recently been demonstrated that upregulation of the receptor for AGEs (RAGE) may have a key role in the hyperglycemia-induced activation of Müller glia and downstream cytokine production in the context of DR (Zhong et al., 2012; Zong et al., 2010). Microglia activation and macrophage migration may also play an active role in bringing to the picture more cytokines (Omri et al., 2011). The mechanism by which these cytokines contribute to neural and vascular apoptosis is not clear but may involve the induction of excitotoxicity, oxidative stress, or mitochondrial dysfunction (Barber et al., 2011). 3. Clinical diabetic retinopathy Clinical DR can be identified as any retinal changes occurring in the retina in diabetes, nonproliferative DR, without manifest visual loss, and its vision-threatening complications, macular edema and proliferative retinopathy (Fig. 1). 3.1. Nonproliferative diabetic retinopathy On ophthalmoscopic examination, the characteristic features of nonproliferative DR are: microaneurysms (MAs), intrarretinal hemorrhages and hard exudates. Retinal MAs are usually the first ophthalmoscopic sign of DR. They are located predominantly within the inner nuclear layer of the retina and in the deep retinal capillary network (Ashton, 1958). On ophthalmoscopy, fresh MAs appear as small red dots (Fig. 2). Microaneurysms may become later yellowish due to increased thickening of basement membrane. Finally, they occlude. Fluorescein angiography demonstrates particularly well the MAs as they become hyperfluorescent and leak profusely (Kohner and Henkind, 1970). However, later in the disease process as they become occluded, fluorescein angiography is not able to identify them. MA count on fluorescein angiography is, therefore, only reliable as an indicator of diabetic retinal disease in the very initial stages of DR. Intraretinal hemorrhages are another ophthalmoscopical feature of nonproliferative DR and result from ruptured small

Fig. 1. DR stages. Clinical DR initiates as nonproliferative retinopathy and may develop vision-threatening complications, macular edema and/or proliferative retinopathy.

Fig. 2. DR. Fundus photography of the posterior pole showing typical alterations, predominantly MAs and hemorrhages.

vessels and MAs, and are mostly located deep in the retina. In the diabetic eye, retinal intraretinal hemorrhages are characteristically most numerous in the posterior pole. Numerous peripheral hemorrhages should lead one to suspect of another concomitant disease process. Hard exudates are another ophthalmoscopic feature of background DR. They are extracellular accumulations of lipids, proteins and lipoproteins derived from leakage from abnormal vessels. Clinically, these yellowish deposits vary in size from small dots to a confluent arrangement that may even cover most of the posterior pole. Capillary closure is a major feature of the vascular damage in diabetic patients (Ashton, 1958). It has become recognized that areas of retinal capillary closure, or capillary nonperfusion, as demonstrated by histology and fluorescein angiography, are a frequent feature of advanced DR. The areas of capillary nonperfusion enlarge as the disease progresses. Vasoregression may be the initial step in this process (Ashton, 1953; Gardiner et al., 2007; Hammes et al., 2011). Retinal edema is defined as thickening of the macula and is mainly due to accumulation of fluid in the central macular area resulting mainly from fluid leakage due to the alteration of the blood-retinal barrier. Subclinical macular edema is detected in diabetic eyes using optical coherence tomography (OCT). Severe DR is an advanced stage of background retinopathy and indicates increased risk of progression to proliferative retinopathy. The presence of soft exudates, venous beading and loops, intraretinal microvascular abnormalities and widespread areas of capillary nonperfusion suggests progression of the retinopathy to a more severe stage (Benson et al., 1988). Fluorescein angiography confirmed the histopathological findings of diabetic retinal disease in a dynamic manner. It was the first technique to document in vivo the alteration of the blood-retinal barrier, well demonstrated by abnormal leakage of intravenously administered fluorescein through the walls of the retinal vessels (Cunha-Vaz and Maurice, 1967). Vitreous fluorophotometry was developed as a sensitive clinical method to quantify the alteration of the blood-retinal barrier (Cunha-Vaz et al., 1975). This method confirmed that alteration of the blood-retinal barrier (fluorescein leakage) as one of the earliest changes to occur in the diabetic retina, calling attention to the importance of the retinal endothelial



damage. The alteration of the blood-retinal barrier (fluorescein leakage) was detected by vitreous fluorophotometry, with values higher than 2 SD above the normal values in approximately 30% of the eyes of patients with diabetes that, otherwise, did not show any ophthalmoscopical changes (Cunha-Vaz et al., 1979). This method has now been replaced by OCT, a non-invasive procedure. When evaluating neuronal and glial damage in the ophthalmoscopic retinopathy stage, it is important to consider changes in retinal function and in particular macular function since diabetes retinal disease affects primarily the central posterior pole of the eye. Prospective analysis of focal mfERG has, indeed, been able to identify functional abnormalities in eyes of diabetic patients in areas without signs of vascular disease (Bearse et al., 2004, 2006). MfERGs may identify, therefore, changes in the neuronal component of the retinal neurovascular unit in the preclinical stages of diabetic macular disease (Han et al., 2004a,b; Harrison et al., 2011a,b). In this respect it must be kept in mind that assessment of standard visual acuity is not expected to be very rewarding since visual acuity remains stationary until w50% of the neuroretinal pathways are affected (Frisen, 1976) and the foveal avascular zone is frequently enlarged in diabetes without any sign of change in visual acuity (Arden et al., 1986). Another problem associated with detection of functional changes in the retina is that diabetic microvascular retinal disease is initially focal in nature, which renders most electrophysiological methods that measure the global response of photoreceptors and have relatively low spatial resolution rather unpromising approaches.

3.2. Vision threatening complications of diabetic retinopathy 3.2.1. Proliferative diabetic retinopathy The exact cause of new vessel formation is not known. It is, however, always secondary to the presence of large areas of capillary nonperfusion. It is, therefore, not specific to DR, as it occurs also in a number of other retinal vascular diseases characterized by marked ischemia, such as sickle cell disease and retinal vein occlusion (Cunha-Vaz, 1986; Valone et al., 1981). An imbalance between angiogenic and antiangiogenic factors appears to be crucial to its development (Michaelson, 1948; Simó et al., 2006; Wise, 1956). New vessels arise from the optic disk or from the retina (Fig. 3). Retinal new vessels lie initially in the plane of the retina but soon pierce the internal limiting membrane and become preretinal, forming adhesions with the overlying vitreous. While the vitreous is attached to the retina the new vessels are symptomless. However, the presence of the new vessels leads to retraction of the

5

vitreous. It is this pulling effect that leads to the progressive complications associated with retinal neovascularization, such as vitreous hemorrhage and progressive visual distortion. Characteristically, retinal neovascularization leaks fluorescein profusely demonstrating an abnormal blood-retinal barrier. Proliferative DR is usually a bilateral disease. Approximately 90% of the persons who present with proliferative DR have it in both eyes at the time of initial examination (Valone et al., 1981). Neovascularization elsewhere in the retina, originates from the remaining perfused vessels, almost exclusively venules, next to areas of capillary nonperfusion. Neovascularization elsewhere in diabetes involves mainly the posterior pole and mid periphery in contrast to other systemic blood disorders that lead characteristically to formation of peripheral new vessels. As the vitreous shrinks, possibly due to the abnormal leakage associated with the abnormal new vessels, it gradually pulls the neovascular fronds, causing preretinal and intravitreal bleading, a frequent cause of acute vision loss in diabetes. Proliferative retinopathy responds well to photocoagulation, but it is essential that it be treated early and adequately, when it is symptomless, i.e., before tractional complications have developed. 3.2.2. Clinically significant diabetic macular edema Clinically significant diabetic macular edema is the largest cause of visual acuity reduction in diabetes (Aiello et al., 1998). It may affect central vision from the early stages of retinopathy and is a frequent complication of DR, particularly in older type 2 diabetic patients. Its role in the process of vision loss in diabetic patients and its occurrence in the evolution of the retinopathy is being increasingly recognized. Diabetic macular edema is mainly extracellular and is generally the first alteration occurring in the retina that causes visual loss. Retinal edema can be intracellular or extracellular but intracellular edema causing increase in retinal thickness is probably rare and its best example is Berlin’s edema due to acute light toxicity. However, recently, Muller cell swelling has been proposed as a relevant factor in the diabetic retina (Wurm et al., 2011). Extracellular edema is directly associated with a situation of breakdown of the inner blood-retinal barrier, one of the earliest alterations occurring in the diabetic retina. The increase in tissue volume is due to an increase in the retinal extracellular space which may be detected using OCT, whereas the breakdown of the bloodretinal barrier is better identified by fluorescein leakage, using fluorescein angiography or vitreous fluorometry. It is important to recall that such direct correlation has been questioned (Bolz et al., 2009; Soliman et al., 2008). However using the more precise scanning laser identification of leakage with Retinal Leakage Analizer the correlation was more apparent (Lobo et al., 2004). This type of edema is characterized by its reversibility if addressed in its initial stages. When there is a situation of breakdown of the bloodretinal barrier, the Starling law governing the movements of fluids applies as we have suggested for the first time (Cunha-Vaz and Travassos, 1984). With an open blood-retinal barrier, any change in the equilibrium between hydrostatic, oncotic and osmotic pressure gradients across the retinal vessels contribute to further water movements and may result in increased edema formation. In this situation, the “force” driving water across the capillary wall is represented by the result of a hydrostatic pressure DP and an effective osmotic pressure difference Dps. The equation regulating movements across the blood-retinal barrier is, therefore:

ðdriving forceÞ ¼ Lp; ½ðPplasma PtissueÞ sðpplasma ptissueÞ; Fig. 3. Proliferative DR. Neovascularization in the optic disk.


6


where Lp is the hydraulic conductivity or membrane permeability of the blood-retinal barrier and s, an osmotic reflection coefficient, Pplasma, the blood pressure, Ptissue, the retinal tissue pressure, p plasma, blood omostic pressure and p tissue, the tissue osmotic pressure. An increase in DP, contributing to increased movements of fluids into the retinal tissue and retinal extracellular edema, may be due to an increase in Pplasma or a decrease in Ptissue or both. An increase in Pplasma due to increased systemic blood pressure does contribute to retinal edema formation only when there is breakdown of the blood-retinal barrier. A decrease in Ptissue is also an important component that is now being given attention. Any alteration in the cohesion of the retinal tissue due to pathologies, such as localized cell loss, retinal fluid, cyst formation and vitreous traction with pulling on the inner limiting membrane of the retina will lead to a decrease in Ptissue thus facilitating fluid accumulation in the retina and an increase in retinal thickness, i.e., retinal edema. Vitreous traction is now accepted as playing a major role in avoiding edema reabsorption and recovery (Ghazi et al., 2007). Similarly, a decrease in Dp contributes to retinal edema due to protein accumulation in the retina associated with the breakdown of the blood-retinal barrier. Extravasation of proteins and lipoproteins, such as in hard exudates, increase the osmotic pressure in the retinal tissue and draw more water into the retinal extracellular space contributing to retinal edema. Increased leakage of proteins is a characteristic feature of the diabetic vasculature (FeldtRasmussen, 2007). This is the main factor associated with oncotic-driven fluid movements in the retina, considering that reduction in plasma osmolarity high enough to contribute to edema formation is an extremely rare event. When there is a breakdown of the blood-retinal barrier the progression of the edema depends directly on the gradient induced by differences between blood pressure and tissue pressure and the oncotic gradient induced by protein accumulation in the retina. A relevant role in this process is also played by the retinal pigment epithelium, which in the early stages of edema may compensate with its pump mechanisms the breakdown of the inner bloodretinal barrier (Sander et al., 2001). Alterations of the retinal pigment epithelium have also been identified in diabetes, mostly in experimental models but their relevance in the initial stages remains controversial (Carmo et al., 1998; Tso et al., 1980; Xu and Le, 2011). The ETDRS, made an effort to establish some guidelines to define “clinically significant macular edema” in order to establish an

Fig. 4. ETDRS classification of clinically significant macular edema.

outcome when designing clinical trials to test the efficacy of treatment for diabetic macular edema (Fig. 4). They paid special attention to the involvement of the center of the macula taking into the consideration the associated visual loss, with its clinical significance. The ETDRS classification of clinically significant macular edema is a follows: 1. Thickening of the retina (as seen either by slit lamp biomicroscopy or by stereo fundus photography) at or within 500 microns of the center of the macula. 2. Hard exudates at or within 500 microns of the center of the macula, associated into the thickening of the adjacent retina (but not residual exudates remaining after disappearance of retinal thickening); 3. A zone, or zones, of retinal thickening one disc area or larger size, any part of which is within one disc diameter of the center of the macula. The problems associated with these guidelines are self-evident, taking into account the subjectivity of the evaluation regarding “abnormal” thickening, the presence of “hard exudates” which are not “residual” and the relative involvement of the central 500 microns circle of the macula. A methodology capable of measuring objectively changes in retinal thickness is OCT (Puliafito et al., 1995). The advent and development of OCT changed dramatically our understanding of the incidence, evolution and rates of progression of diabetic macular edema. We are now able, in the routine clinical setting, to measure changes in retinal thickness and identify the presence of macular edema, using OCT, a non-invasive instrumentation. Definitions need to be agreed upon (Bressler et al., 2012; Browning and Fraser, 2008; Pires et al., 2013). Diabetic macular edema can be identified, using OCT, regarding its type and distribution, its evolution, its pathophysiology, and degree of involvement of the central macular area. 4. Progression. Natural history The initial stages of nonproliferative DR are characterized by the presence of MAs, small hemorrhages, and indirect signs of vascular hyperpermeability and capillary closure, i.e., respectively, hard and soft exudates. These alterations dominate the fundus picture in the initial stages of DR and are the only ones used for characterization of the first four levels (10e43) of the ETDRS classification of DR. An abnormality of the blood-retinal barrier, demonstrated both by vitreous fluorometry and fluorescein angiography, is also an early finding both in human and experimental diabetes (Cunha-Vaz et al., 1975; Waltman et al., 1978). The alteration of the bloodretinal barrier is well demonstrated by fluorescein leakage and it is accepted that it leads to development of retinal edema. On ophthalmoscopic examination and fundus photography, the formation of MAs, small hemorrhages and hard exudates are the initial changes that are identified. MAs may be counted and MA counting has been suggested as an appropriate marker of retinopathy progression (Klein et al., 1989). It must be realized that formation of new MA and their disappearance are dynamic processes. During a 2-year follow-up study of 24 type 1 diabetic patients with mild background DR, using fluorescein angiography, Hellstedt and Immonen (1996) observed 395 new MAs and disappearance of 258 previously identified MAs. DR has been considered a microvascular complication of diabetes clinically identified by changes produced through either due to progressive cell degeneration and vasoregression or



abnormalities of the blood-retinal barrier, limiting the diagnostic and therapeutic focus to the vascular system. However, it is now accepted that DR involves the neuronal as well as the vascular compartments (Antonetti et al., 2012). Attempts have been made to identify functional changes of the retina that may precede micoaneurysms, such as blood flow changes but the results have been contradictory mainly because of technical problems and lack of reliable methodology. Subtle changes in microvascular hemodynamics are certainly one of the earliest changes to occur in DR. There is evidence from a number of researchers suggesting its relevance (Dorner et al., 2003; Kur et al., 2012; Metea and Newman, 2007; Riva et al., 2005). It is noteworthy that aside from the technical difficulties associated with measuring retinal blood flow in humans, the changes detected appear to be different in eyes of patients with the same level of retinopathy, metabolic control and duration of the disease, thus further confounding the results obtained (Ludovico et al., 2003). High resolution imaging with spectral domain OCT and multifocal electroretinography (mfERG) are other promising procedures that are offering new perspectives for the evaluation of the initial stages of diabetic retinal disease. Validation of biomarkers of DR progression must involve demonstration that these potential biomarkers are associated with vision loss, the most generally accepted clinical outcome. This is a major problem as it is well recognized that vision loss only occurs when approximately 50% of macula neuronal component is damaged (Frisen, 1976). Vision loss is clearly a late development in retinal disease and what we need is to identify outcomes that can be recognized before vision less is present. Vision loss is associated with the two major complications of DR, clinically significant macular edema and proliferative DR, and does not occur before these complications develop. This concept is crucial. The clinically meaningful outcomes are clinically significant macular edema and proliferative DR. There is, therefore, a clear need to identify biomarkers of diabetic retinal disease progression that predict development of these late clinically significant outcomes directly associated with vision loss. In summary, the fundus abnormalities seen in DR can conceptually be split into three categories- those findings resulting from leaking microvasculature (hemorrhages, lipid exudates, retinal edema; those findings resulting from structural damage to the microvasculature wall (MAs); and those findings resulting from ischemia with a subsequent overproduction of vascular growth factors (cotton-wool patches, intraretinal neovascular abnormalities, preretinal neovascularization, fibrous proliferation and vitreous hemorrhage) (Wong, 2010). The severity of each of these findings has been classified and quantified based on the degree of retina involvement. The Airlie House classification used initially only five photographic fields. The Diabetic Retinopathy Study (DRS) created the modified Airlie House classification system and added more gradations of severity and used seven fields. This classification was developed to classify DR progression to proliferative DR. Similarly, the ETDRS designed a classification to answer whether a strategy of earlier treatment with scatter photocoagulation to proliferative retinopathy was beneficial. Although complicated the ETDRS severity scale has become the de facto gold standard for grading retinopathy severity in clinical trials. Based on a patient’s current ETDRS retinopathy level, one can predict the chance of developing high-risk proliferative retinopathy. However, limitations of the ETDRS classification become clear as it indicates that moderate nonproliferative DR (level 47) carries an 8.6% 1-year risk of developing high-risk proliferative retinopathy whereas, the next step (level 53) carries a 45% of developing high-risk proliferative retinopathy. Therefore, a jump in a single numerical level of severity (47e53) results in a big jump in risk (8.6e45%). When reviewing the ETDRS data the Global DR Project Group realized that

7

the presence of intraretinal neovascular abnormalities and venous beading was more predictive of the risk of progressing to proliferative retinopathy, whereas MAs and hemorrhages were poorly predictive. Because macular edema was not considered in the ETDRS classification they proposed another classification that was easier to use and included macular edema. However, the final result did not translate into clinical practice, lost the advantages of the ETDRS classification to predict progression to proliferative retinopathy and the inclusion of macular edema was not helpful due to the progressively increased availability of OCT, offering an objective evaluation of macular edema. 5. Complexity and specificity of retinal damage in diabetes DR is a disease of the retinal neurovascular unit occurring in the well-recognized multifactorial environment of diabetes. In order to understand the features of DR and its progression it is important to review the specific characteristics of the retina and how they may be affected by diabetes. Of particular interest is to address the specificity of the human macular region and of the retinal neurovascular unit as diabetic retinal disease initiates at the posterior pole of the eye in the central retina between the main vascular arcades. 5.1. Human macular region. Relevance for diabetic retinopathy and vision loss The fovea centralis is located at the center of the macular region and is a characteristic feature of all primate retinas. The fovea lies in the visual axis of the eye such that a beam of light passing perpendicularly through the center of the lens will fall in the fovea. It is recognized as the site of maximal visual acuity (Helmholtz, 1924). Although the fovea only contains 0.3% of the total cones, it contains 25% of the ganglion cells (Curcio and Allen, 1990), illustrating its importance for vision. A large portion of the brain, 40% of the primary visual cortex processes the central 5 degrees of the visual field (Curcio et al., 1990). The central retina or macula is divided into four concentric zones with the foveola in the center surrounded in turn by the fovea, parafovea and perifovea. The foveola is 250e350 microns in diameter, and is formed by cone cell bodies surrounded by Muller glial cell processes (Burris et al., 2002). The fovea includes the adjacent 750 microns around the foveola making it 1.85 mm wide. This region is characterized histologically by having a thick layer of ganglion cells up to eight deep and a very thick outer plexiform layer. Capillaries are present in the inner retina up to the foveal slope, where they form a foveal avascular zone about 400e500 microns wide. The central foveola contains the highest cone density in the retina and is responsible for the highest visual acuity, emphasizing how critical optimal health of the fovea is for good vision throughout life. Muller cells are particularly important in the fovea because the foveola is formed only by cones and Muller cells. They are found throughout the retina and are the only cell type which spans all retinal layers. Involvement of the posterior pole and the central macula characterize the pathology of DR and explain the evolution of the disease to vision loss. 5.2. Retinal neurovascular damage in diabetes 5.2.1. The retinal neurovascular unit It is becoming apparent that the retinal dysfunction associated with diabetes should be considered as affecting initially different components of the retinal neurovascular unit in different patients.


8


The retinal neurovascular unit refers to the physical and biochemical relationship among neurons, glia and specialized vasculature and the close interdependency of these tissues occurring in the central nervous system and retina (Fig. 5) (Antonetti et al., 2012). This intimate association of glia, neurons and blood vessels allows for energy homeostasis and neurotransmitter regulation. Furthermore, its close signaling interdependence manifests itself in the blood-retinal barrier which controls the flux of fluids and blood-borne metabolites into the retinal tissue (Cunha-Vaz, 1979). There is evidence that the different components of the retinal neurovascular unit are at some time affected in diabetes. They appear also to be involved to a different degree in different patients. Accumulating evidence show that not all patients show a detectable alteration of the blood-retinal barrier in the early stages of the disease (Cunha-Vaz et al., 1975) and not all patients present electrophysiological changes in the same stage of the disease (Bearse et al., 2006; Bronson-Castain et al., 2009). Alterations of the blood-retinal barrier, neuronal damage and inflammation appear to be the most relevant changes occurring in the initial stages of diabetic retinal disease. 5.2.2. Neuropathy. Increased apoptosis and degeneration The neurosensory retina is altered in diabetes. There is indication of early nerve fiber loss and alterations have been detected using psycophysical and electrophysiological techniques in the early stages of diabetic retinal disease, even before detection of microvascular changes on ophthalmoscopy (Bearse et al., 2006; Bronson-Castain et al., 2009; Han et al., 2004a). These alterations have been detected in variable degree in diabetic patients again suggesting that different patients are affected differently. Biochemical defects, such as impaired control of glutamate metabolism (Gowda et al., 2011) as well as loss of synaptic activity and dendrites (VanGuilder et al., 2008), apoptosis of neurons primarily in the ganglion cell and inner nuclear layers (Barber et al., 1998), have all been incriminated in the diabetic disease process (Simó and Hernández, 2013; Stem and Gardner, 2013). Together with reduced corneal nerve sensation and impaired autonomic innervation of the pupil, altered function of the retinal sensory nerve indicates that diabetes causes denervation of multiple

Fig. 5. Functional anatomy of the retina. The close interconnection of the different components of the retinal neurovascular unit is well represented in the diagram. (Adapted from Antonetti et al. (2006).

sensory inputs to the eye. Thus, although the retinal neural structure differs from the peripheral sensory system, DR resembles diabetic peripheral sensory neuropathy and could be considered that there is a neuropathy component of diabetic retinal disease. 5.2.3. Blood-retinal barrier and angiogenesis The barrier system that separates the intraocular structures and fluids from the blood should be viewed as formed by two different boundaries (Cunha-Vaz and Maurice, 1965; Cunha-Vaz, 1979). One regulating exchanges between the blood and aqueous humor, the blood-aqueous barrier, and the other, regulating exchanges between the blood and the extravascular spaces of the retina, the blood-retinal barrier. The blood-retinal barrier forming a particularly tight barrier due to the presence of specialized junctions between the endothelial cells of the retinal vessels and retinal pigment epithelial cells (Shakib and Cunha-Vaz, 1966). DR involves an alteration of the blood-retinal barrier and vascular leakage very early in the disease process, leading in certain patients to macular edema and later in the disease process, to neovascularization with tufts of new abnormally permeable vessels in the proliferative stage of the disease. Macular edema remains the clinical feature most frequently associated with vision loss, with evidence of leakage on fluorescein angiography and thickening of the retina on OCT. The efficacy of treatment with anti-VEGF agents indicates that VEGF contributes to the pathogenesis of diabetic macular edema (DRCR.net, 2010; Nguyen et al., 2010). VEGF may be even the major factor in eyes that are optimal responders to anti-VEGF intravitreal injections. VEGF promotes leakage of plasma proteins from blood-vessels and it has been shown to induce increased vascular permeability (Kaur et al., 2008; Klaassen et al., 2013; Mark et al., 2004; Mayhan, 1999). Conditional deletion of the VEGF gene from Muller cells reveals the importance of glial cells for VEGF production in retinopathy models of angiogenesis, and this finding underscores the consequences of altered glial-vascular communications (Bai et al., 2009). The mechanism of VEGF induced vascular permeability involves activation of classical protein kinase C isoforms, particularly protein kinase C beta (Harhaj et al., 2006; Ishii et al., 1996). Recently, the tight junction protein occludin was identified as a target of protein-kinase C beta, leading to ubiquitin-mediated endocytosis of tight junction components and increased vascular permeability (Murakami et al., 2009) and providing a molecular mechanism for the regulation of the properties of the blood-retinal barrier in response to VEGF. An altered blood-retinal barrier implies also an inflammatory response and may serve as a link between different disease mechanisms (Joussen et al., 2004; Kaur et al., 2008). Other signaling pathways such as kallikrein inhibitors (Clermont et al., 2011) and matrix metalloproteases (Navaratna et al., 2007; Yang et al., 2010) may also contribute to increased vascular permeability in DR. 5.2.4. Inflammation The concept of retinal neurovascular unit should also include the potential involvement of activated microglia in DR and bloodborne inflammatory mediators. Systemic inflammation is, furthermore, an intrinsic response to overfeeding, obesity and diabetes. Diabetes increases the release of retinal inflammatory mediators (interleukin 1b, tumor necrosis factor a -TNFa, intercellular adhesion molecule-ICAM1’ and angiotensin ii) and activation of microglial cells have been shown to occur in early DR (Wilkinson-Berka et al., 2009).



The role of iNOS in the development of the early stages of DR recently has been demonstrated directly using mice genetically deficient in iNOS (Leal et al., 2007; Zheng et al., 2007). Increased retinal leukostasis has been correlated with increased expression of intercellular adhesion molecule-1 (ICAM-1) and vascular leakage, since the blockage of ICAM-1 with a monoclonal antibody prevented both leukostasis and vascular leakage (Joussen et al., 2004). More recently, evidences have demonstrated that nitric oxide produced by inducible (iNOS) appears to have a predominant role in leukostasis and blood-retinal barrier breakdown (Leal et al., 2007). Diabetic endothelial dysfunction and associated leukostasis and inflammation may, indeed, play a central role in the pathogenesis of DR, explaining the initial alteration of the blood-retinal barrier (Tang and Kern, 2011). Recent findings suggest also that TNF and cyclooxygenase-2 contribute to DR, perhaps by preventing endothelial cell damage from adhering leukocytes (Joussen et al., 2002). Other inflammatory mediators that have been incriminated are angiotensin II (Ghattas et al., 2011) and pigment epithelium-derived factor (PEDF) (Yoshida et al., 2009). Muller cells and microglia have also been shown to release proinflammatory mediators in DR (Ibrahim et al., 2011; Wang et al., 2010). The alteration of the blood-retinal barrier may lead to macrophage migration into the neurosensory retina or increased adherence to the vasculature, as well as accumulation of inflammatory and angiogenic mediators in the vitreous cavity, contributing to microthrombosis formation in the retinal vessels and progressive capillary closure. The available data indicates that inflammation functioning as a repair process contributes to the progression of DR, particularly by creating conditions for the development of its vision threatening complications. Anti-inflammatory treatments with intravitreal glucocorticoids have been shown to reduce macular edema and may prevent progression to proliferative retinopathy by restoring the blood-retinal barrier (Campochiaro et al., 2011). 6. Phenotypes of diabetic retinopathy It is recognized that the duration of diabetes and the level of metabolic control condition the development of DR. However, these risk factors do not explain the great variability that characterizes the evolution and rate of progression of the retinopathy in different diabetic individuals. There is clearly great individual variation in the presentation and course of DR. There are many diabetic patients who after many years with diabetes never develop sight-threatening retinal changes, maintaining good visual acuity. However, there are also other patients that even after only a few years of diabetes show a retinopathy that progresses rapidly and may not even respond to available treatments. In a prospective 3-year follow-up study of the macular region in 14 patients with type 2 diabetes mellitus and mild nonproliferative retinopathy, using multimodal macula mapping we found that there is marked individual variations in the progression of DR and activity of the retinal disease (Lobo et al., 2004). Multimodal macula mapping involved a combination of imaging examinations used in a research environment, including scanning laser leakage analysis, retinal thickness analysis and microperimetry (Bernardes et al., 2002). In a span of 3 years, eyes with minimal changes at the start of the study (levels 20 and 35 of the ETDRS-Wisconsin grading) were followed at 6-month intervals in order to monitor progression of the retinal changes. The most frequent alterations observed, by decreasing order of frequency, were leaking sites (Lobo et al., 1999), areas of increased retinal thickness and MAs/ hemorrhages.

9

The number of MA and small hemorrhages increased in most eyes during the 3-year follow-up period. This was particularly well demonstrated when the location of each MA was taken into consideration. This increase in the number of microaneuryms was considered to have the potential to be an indicator of retinal vascular damage and remodeling of the retinal circulation, particularly in the initial stages of DR. Three patterns of DR progression were identified. Pattern A included eyes with reversible and relatively little abnormal fluorescein leakage, a slow rate of microaneurysm formation and a normal foveal avascular zone. This group appeared to represent eyes presenting slowly progressing retinal disease. Pattern B included eyes with persistently high leakage values, indicating a dominant alteration in the blood-retinal barrier, relatively higher rates of microaneurysm accumulation and a normal foveal avascular zone. This group appeared to identify a ’wet’ form of diabetic retinopathy. Pattern C included eyes with variable and reversible leakage, increased microaneurysm formation and an abnormal foveal avascular zone (Lobo et al., 2004). We then extended our observations by following 57 patients with type 2 diabetes for 7 years; at the time of enrollment, all eyes presenting mild nonproliferative DR (Cunha-Vaz and Bernardes, 2005). In this larger study, with longer follow-up, the three previously identified different phenotypes were again identified after an initial 2-year follow-up period. After an average of 7 years of follow-up, 10 of these 57 eyes had developed clinically significant macular edema. In this series of patients, after the initial 2-year follow-up period, 35 eyes (61% of the total) were identified as showing the characteristics of pattern A, i.e. slow progression, 12 (21%) were classified as presenting pattern B, and the other 10 (18%) had the characteristics of pattern C. Severe macular edema needing treatment developed after 7 years of follow-up only in those eyes classified as belonging to patterns B and C. Of the 12 eyes classified as having pattern B, 5 (42%) developed severe macular edema. Similarly, of the 10 eyes identified with pattern C, 5 (50%) developed severe macular edema. In contrast, none of the eyes classified as belonging to pattern A developed severe macular edema in the 7-year follow-up period. In summary, the slow progression type, pattern A, appears to take longer than 7 years to develop severe macular edema, one of the main complications of DR, confirming that this subtype of DR has a good prognosis. On the other hand, both other types of DR progression, the leaky type, or pattern B, characterized initially by particularly high levels of leakage, i.e. alteration in the blood-retinal barrier, and the ischemic type, or pattern C, characterized by signs of capillary closure, lead much more frequently to the development of severe macular edema, with incidences at 7 years of 42% and 50%, respectively. If DR is a multifactorial disease in the sense that different factors or different pathways may predominate in different groups of cases with DR - then it is crucial that these differences and the possible different phenotypes be identified (Grange, 1995). The characterization of different phenotypes of DR, with different progression patterns, opened particularly interesting perspectives to gain more insight into the understanding and management of DR (Cunha-Vaz, 2007). 6.1. Characterization of phenotypes of diabetic retinopathy progression using non-invasive methods The initial identification of 3 major patterns of progression of nonproliferative DR was made using elaborated multimodal imaging methodology involving new examination technologies, some


10


still in the research domain (Lobo et al., 2004). Furthermore, the sample size was small. In a more recent study performed by our research group we were able to follow a much larger population of eyes/patients with well characterized mild nonproliferative retinopathy in patients with diabetes type 2, using only non-invasive procedures easy to repeat, such as digital color fundus photography and OCT (Nunes et al., 2013). This study was a two-year prospective, observational study, designed to follow eyes/patients with mild non-proliferative DR (grades 20 and 35) for a period of two years or until the time of development of clinically significant macular edema (clinically significant macular edema ETDRS classification). Four hundred ten patients (410) were included, diagnosed with adult-onset type 2 diabetes, age 40e75 years, mild nonproliferative DR (20 and 35 of ETDRS classification), best corrected visual acuity 95 ETDRS letters (20/25) (Fig. 6). At the three study visits, V0, V6 and V24 (or pre-treatment visit), the study eyes underwent a complete eye examination, which included best corrected visual acuity, as tested in the ETDRS, slitlamp examination, intraocular pressure measurements, fundus photography and OCT. The automated computer-aided diagnostic system, RetmarkerDR (Critical Health SA, Coimbra, Portugal), was used to automatically detect MA on the field-2 color fundus images. MA counting on fundus photographies and MA counting on fluorescein angiography have been proposed as predictive indicators for progression of DR (Klein et al., 1989; Kohner et al., 1986). A recently developed software, the RetmarkerDR (Bernardes et al., 2009; Nunes et al., 2009), allows the identification of the exact location of each red-dot in successive fundus photographs performed in each eye (Fig. 7). Identification of the exact location of an individual red-dot is considered particularly important because a new MA is considered to develop only once in a

Fig. 6. CONSORT flowchart (Schulz et al., 2010). (NPDR: Nonproliferative DR; CSME: Clinically Significant Macular Edema needing photocoagulation treatment).

Fig. 7. The RetmarkerDR software automatically calculates MA formation and disappearance rates and as a result, MA turnover.

specific location, its disappearance leaving in its place mainly remnants of basement membrane (Ashton, 1974; Cunha-Vaz, 1978). MA turnover, a dynamic process over time, can, therefore, be represented either by the formation of new MAs in a set period of time (Haritoglou et al., 2014; Nunes et al., 2009) or by both formation and disappearance of MAs (Ribeiro et al., 2013a). To identify phenotypes of progression of mild nonproliferative DR to clinically significant macular edema using non-invasive procedures a cluster analysis was performed based on the following parameters, at baseline: mean retinal thickness in the central 1500 mm in diameter macular area, macular area with increased retinal thickness, maximal increased retinal thickness, number of MA and MA turnover (computed for the first 6 months of follow-up). Cluster analyses are unsupervised segmentation techniques which build models of the observed data in order to identify and create homogeneous groups (Everitt, 1995). These techniques group data sharing some similarity measure or feature (Kaufman and Rouseeuw, 1990), being it either of the hierarchical or nonhierarchical type, respectively when the number of underlying clusters is unknown or known a priori. In this work the first type was used as an exploratory tool to establish the number of homogeneous groups in the data set. The Ward’s method used for the hierarchical clustering, confirmed the existence of three clusters. The three phenotypes result from the statically significant differences for the MA and retinal thickness parameters (p < 0.001). The first cluster/phenotype was composed of 181 eyes/patients (48.1%), the second cluster/phenotype was composed of 87 eyes/patients (23.2%), and the third cluster/phenotype was composed of 108 eyes/patients (28.7%). Phenotype B is characterized by a higher central subfield retinal thickness (p < 0.001). Phenotype C is characterized by higher MA parameters, number, and turnover (p < 0.001). Threshold values were identified using a decision and classification tree and the following rules were used to classify eyes/patients into one of the three phenotypes of DR progression: Phenotype A, MA turnover< 6 and normal retinal thickness values (central subfield < 220 microns, i.e. normal mean þ 1SD); Phenotype B, MA turnover 220 microns); Phenotype C, MA turnover > 6 (Fig. 8). From the 133 eyes/patients with an MA turnover less than 6 and a central thickness less than 220 microns, 1 eye/patient developed clinically significant macular edema (0.7%); from the 94 eyes/patients with an MA turnover less than 6 and a central retinal



11

Fig. 8. Examples from individual patients belonging to the three different phenotypes. Phenotype A - MA Turnover 6 and area with retinal thickness values in the normal range (5% increased retinal thickness over normal range; Phenotype C - MA Turnover > 6 and variable retinal thickness. MA T e microaneurysm turnover, calculated over a 6-month interval.

thickness greater than or equal to 220 microns, 8 eyes/patients developed clinically significant macular edema (8.5%) and from the 121 eyes/patients with an MA turnover greater than or equal to 6, seventeen eyes/patients developed clinically significant macular edema (14.5%). Using these rules to estimate the risk for clinically significant macular edema development, phenotype B shows a sensitivity and a specificity of 88.9% and 60.5%, respectively (when compared to phenotype A), and phenotype C shows a sensitivity and specificity of 94.4% and 55.9%, respectively (when compared to phenotype A). Phenotype A shows a negative predictive value for developing clinically significant macular edema of 99.2%. This study showed that using the mathematical model of hierarchical cluster analysis and only noninvasive procedures, fundus photography and OCT, three different phenotypes of DR can be identified, which show different risks of progression to clinically significant macular edema. These three phenotypes are in agreement with the number of patterns of DR progression proposed previously by Lobo et al. (2004) in a different and smaller sample. In this original study, three phenotypes of progression of mild DR were characterized: phenotype A as slow progression, phenotype B as “leaky” and phenotype C as ischemic. Similar categorization was found in this study as a result of hierarchical analysis, phenotype B with predominance of edema, and phenotype C with predominance of MA turnover (i.e. with increased rates of MA formation and disappearance). It was striking to find that 348 eyes/patients with similar levels of ETDRS classification (20e35) showed a large interquartile range both in MA turnover and retinal thickness values, demonstrating the wide range of values for each of these alterations. Hierarchical cluster analysis showed that these initial changes occur in different levels of intensity in different groups of patients and that different groups of patients can be characterized by levels of intensity of these retinal changes. Our findings show that increased activity of microvascular disease in the macular region (field 2), demonstrated by increased

rates of MA turnover that characterize phenotype C, are associated with higher risk for development of clinically significant macular edema in the relatively short period of 2 years. This phenotype represents approximately 30% of the patients. Of relevance is also the finding that on the other hand, phenotype A, which is characterized by low MA turnover and also no signs of macular edema (representing approximately 50% of the patients) has the lowest risk for development of clinically significant macular edema, predictive negative value of 99.2%. This observation might have significant implications for management of DR. Furthermore, this observation indicates that a large proportion of the eyes with mild DR will progress very slowly, suggesting that these eyes/patients should not be included in clinical trials because of their slow rates of progression. The identification of different DR phenotypes characterized by different dominant retinal alterations and different risks of progression to clinically significant macular edema opens new perspectives for personalized management of DR, although it is not known if they are also predictive for proliferative DR. It must be realized that further research will be needed to verify if one phenotype can in the course of the disease change to another phenotype and if these phenotypes are also predictive for proliferative DR. 6.2. Suggested predominant disease mechanisms in different phenotypes 6.2.1. Accelerated Apoptosis. Vasoregression and pericyte loss There is in the retinal tissue of humans with diabetes and in diabetic animal models a situation of increased apoptosis of a variety of neural and vascular cells (Barber, 2003). There is also a widespread presence of cell apoptosis in the retinal vessels manifested in the streptozotocin diabetic rat model by areas acellular capillaries (Curtis et al., 2009; Gardiner et al., 2007). The generalized apoptosis occurring in diabetes suggests a generalized damage to the retinal and vascular tissue occurring


12


in response to chronic hyperglycemia and may, indeed, be the basis for the pericyte loss and vascular vasoregression (Hammes, 2005). It is possible that this generalized degenerative process corresponds to a situation of neuropathy similar to the one occurring in the other parts of the body as a result of diabetes and may underly the disease process occurring in the slowly progression phenotype A. 6.2.2. Breakdown of the blood-retinal barrier and inflammation Breakdown of the blood-retinal barrier is one of the earliest alterations occurring in the diabetic retina and in 30e40% of diabetic patients may be present before fundoscopic signs of DR (Cunha-Vaz et al., 1975, 1979; Waltman et al., 1978). This breakdown of the blood-retinal barrier has been shown to lead to extracellular expansion and edema of the retina. Inflammation, associated increased leukostasis and proinflammatory cytokines may be particularly relevant in this phenotype (Adamis, 2002; Joussen et al., 2004). Abnormal leakage through the blood-retinal barrier and the resulting edema is the dominant alteration in phenotype B, which is well identified by increased thickening of the retina in the OCT examination (Nunes et al., 2013). 6.2.3. Hypoxia and angiogenesis Capillary nonperfusion and capillary closure occur frequently in the diabetic retina and are other characteristic features of DR. They are indicators of retinopathy progression and direct signs of presence of ischemia. These features further aggravate the reduced oxygen delivery occurring in diabetes (Ditzel, 1976). Both capillary nonperfusion and capillary closure are, most probably, the result of retinal microthrombosis. Increased platelet adhesiveness and aggregation has been documented in diabetic patients since at least three decades (Heath et al., 1971). Microthrombosis have been demonstrated by electron microscopy in experimental diabetes, these microthrombi are mainly composed of aggregated platelets and fibrin strands (Ishibashi et al., 1981). Boeri et al. (2001) demonstrated in post-mortem specimens from diabetic and nondiabetic donors, that diabetic retinas present an increased number of platelet-fibrin thrombi in the retinal capillaries. Our results using the RetmarkerDR and evaluating MA turnover suggest increased microvascular disease activity and remodeling of the retinal circulation and disappearing MAs may indicate increased microthrombotic activity. Thrombosis is a dynamic process which is initiated when the hemostatic system is perturbed and may be associated with vessel wall damage and inflammation. These features are the central ones in endothelial dysfunction in diabetes (Feldt-Rasmussen, 2007; Parving et al., 1982). Alterations in endothelial integrity, increase in inflammatory cytochines, disturbances in laminar blood flow, thrombin generation, and inhibition of endogenous fibrinolytic activities contribute to thrombus formation. This phenomenon may, finally, stimulate angiogenesis and vascular remodeling. This pathological mechanism appears to have a predominant role in phenotype C. The increase in hypoxia due to vascular closure creates an imbalance between angiogenic and antiangiogenic factors creating further stimuli for vascular leakage. The process becomes more dynamic and the retinal vascular disease more active with increased retinal vascular modeling and more rapid progression of the disease (Bai et al., 2009; Nguyen et al., 2010). 6.3. Integrated perspective of diabetic retinopathy progression The results of our research group on the characterization of different phenotypes of DR confirm that there are distinct

morphological manifestations in DR with different subjects, with similar metabolic control, presenting different rates of progression and different risks for the development of vision-threatening complications (Lobo et al., 2004; Nunes et al., 2013). There is also evidence indicating that susceptibility to the late vascular complications of diabetes, such as retinopathy, depend, at least partly, on genetic factors (Simó-Servat et al., 2013). Looking at DR and the identification of three major phenotypes allows an integrated perspective of DR progression. Chronic hyperglycemia induces generalized cell damage to the retina involving the entire retinal neurovascular unit, but causing different degrees of damage in different cells in different individuals. Some patients develop generalized low-grade vascular, neuronal and glial damage which manifests as a slow progressing neuropathy with slow progressing vascular damage (phenotype A). Other patients develop breakdown of the blood-retinal barrier with resulting retinal edema, possibly associated with the neuroglial damage and an active inflammatory repair process (phenotype B). Finally, another group of patients either due to an abnormal accumulation of VEGF and other angiogenic factors associated with rapidly developing hypoxia induce an abnormal interaction between endothelial and blood cells, possibly due to specific genetic characteristics, shows signs of active microvascular disease, and more rapid progression to vision-threatening complications (phenotype C). A classification of DR based on both relevant genotypes and disease phenotypes is an ambitious goal. We believe that this route may help identify the particular form that threatens an individual patient and consequently offer an opportunity for specific and more effective therapies. 7. Biomarkers of diabetic retinopathy The Eye Diseases Prevalence Research Group classified DR into major outcomes: any DR, as any DR consisting of mild, moderate, or severe DR; and vision-threatening DR (VTDR), as DR likely to result in vision loss on the absence of treatment, consisting of proliferative DR, clinically significant diabetic macular edema, or both (Kempen et al., 2004). This concept is crucial and a good way to address the issue of management of DR in order to prevent vision loss and to identify which patients will progress to VTDR (i.e., to clinically significant diabetic macular edema and/or proliferative DR). It is now clear that systemic markers of diabetes such as duration of the disease, poor glycemic control, increased blood pressure, and lipid levels are relevant factors, but they do not identify DR worsening (Hove et al., 2006). It is a well-established fact that patients under good metabolic control may worsen rapidly and develop vision-threatening DR complications before other patients with poor metabolic control. These observations led to the identification of different phenotypes of progression (Nunes et al., 2013) based on the characteristics of the retinal lesions. It is, therefore, fundamental to be able to identify the retinal lesions, their number, and dynamics in the earlier stages of DR and correlate their occurrence to the progression of any stage of DR to VTDR (Hove et al., 2006). 7.1. Definitions of biomarkers Biomarkers have become the basis for preventive medicine, meaning medicine that recognizes diseases or the risk of disease early, and takes specific countermeasures to prevent the development of disease. Biomarkers are also seen as the key to personalized medicine, treatments individually tailored to specific patients for highly efficient intervention in disease processes.



It is necessary to distinguish between prognostic, diseaserelated, and predictive, drug-related, biomarkers (Shapiro et al., 2009; Tevak et al., 2010). Prognostic markers show the progression of disease with or without treatment whereas predictive biomarkers help to assess the most likely response to a particular treatment type. Chronic diseases often begin with an early, symptom-free phase. In such symptom-free patients there may be more or less probability of actually developing symptoms. In these cases, biomarkers help to identify high-risk individuals reliably and in a timely manner so that they can either be treated before onset of the disease or as soon as possible thereafter. In order to use a biomarker for diagnostics, the procedure involved should be as easy to obtain as possible and well accepted by the patient. Also, a rapid test which delivers a result after only a few minutes is optimal. This makes it possible for the physician to discuss with the patient how to proceed and if necessary to make management decisions immediately after the test. 7.2. Prognostic biomarkers of diabetic retinopathy 7.2.1. Systemic disease-related biomarkers 7.2.1.1. Serum disease-related biomarkers. The cornerstone of treating diabetes is to bring blood glucose levels to the lowest possible range possible without severe side effects. In 1993, the Diabetes Control and Complications Trial demonstrated that intensive therapy lowered time-averaged blood glucose value (measured as hemoglobin A1c) and significantly reduced development of microvascular complications in type 1 diabetes. However, only 6.6% of the variation in risk of retinopathy was explained by the differences in treatment groups (Hirsch and Bownlee, 2010). For type 1 diabetes, despite the fact that glycemia is the major systemic risk factor for developing retinopathy, its overall contribution is only 11%, i.e., 89% of the risk must be explained by other unknown factors (Lachin et al., 2008). Similarly, intensified glucose control has also been studied either alone or as a part of a multifactorial intervention in type 2 diabetic patients. The first study to demonstrate an effect of glycemia on retinopathy progression was the UK Prospective Diabetes Study (UKPDS) (Stratton et al., 2001). This study showed that the overall effect of intensified treatment was modest, and that it took 6 years to see a difference between the conventional (mean HbA1c 7.9%) and the intensive group (mean HbA1c 7.0%). In our clinical studies when considering the usual systemic parameters, such as lipid levels and blood pressure, only HbA1c values at baseline were consistently correlated with the development of the vision-threatening complications of diabetes such as clinically significant diabetic macular edema. Observational and epidemiological studies have inconsistently reported an association between DR and elevated serum lipids (UKPDS, 2004). Reduction of blood pressure appears to be particularly beneficial for type 2 diabetes but their effect has been attributed primarily in relation to the rates of development and progression of diabetic macular edema (Cikamatana et al., 2007; Chowdhury et al., 2002). A variety of serum markers of inflammation and endothelial dysfunction have been proposed as biomarkers of DR. Candidates proposed include retinol-binding protein 4 (Li et al., 2010; Takebayashi et al., 2007), advanced glycation end products (Boehm et al., 2004; Ghanem et al., 2011; Salman et al., 2009; Wautier et al., 2003), homocysteine (Aydin et al., 2008; Brazionis et al., 2008; Goldstein et al., 2004; Van Hecke et al., 2008), laminin (Masmiquel et al., 2000; Simó et al., 1996), markers of inflammation and endothelial dysfunction (Costagliola et al., 2013; Klein et al., 2009; Roy et al., 2013; Van Hecke et al., 2005), vascular

13

cells adhesion molecules (Matsumoto et al., 2002; Nowak et al., 2008; Soedamah-Muthu et al., 2006; Yoshizawa et al., 1998), rhodopsin or retina-specific m-RNA (Butt et al., 2012; Hamaoui et al., 2004; Shalchi et al., 2008). Associations between serum apolipoproteins levels measurements of microvascular function have also been proposed (Sasongko et al., 2012). Recently, Muni et al. (2013) reanalyzed the Diabetes Control and Complications Trial cohort looking for increased levels of serum inflammatory markers and their correlation with DR progression. Increased levels of high-sensitivity C-reactive protein was the only serum marker showing an association with higher risk for the development of clinically significant macular edema. In another recent publication, Laursen et al. (2013) looked at associations between C-reactive protein and von-Willebrand factor and DR in type 1 diabetic patients. Their results were generally negative, with evidence of some degree of correlation between higher levels of high sensitive C-reactive protein and development of proliferative DR. No association was found between serum levels of vonWillebrand factor and DR. In summary, at present, the only validated systemic biomarker for development of DR is HbA1c. 7.2.2. Genetic disease-related biomarkers The onset, intensity and progression of diabetic complications show large interindividual variations (Lobo et al., 2004). There is evidence from aggregation in families and specific ethnic groups, together with lack of serious complications in some diabetic patients with poor metabolic control that there is a genetic predisposition to develop some diabetic complications such as retinopathy (Leslie and Pyke, 1982; Warpeha and Chakravarthy, 2003). Heritability has been estimated to be as high as 27% for DR and 52% for proliferative DR (Arar et al., 2008; Hietala et al., 2008). Efforts to unravel the human genetics of DR have been undertaken using the candidate gene linkage approaches, and more recently genome wide association studies. A large number of putative genes and genetic variants have been reported in the literature and some of them exhibit consistent associations with DR (ALR2, VEGF and RAGE genes). However, these results have not been replicated in multiple populations, and, therefore, no genes have achieved widespread acceptance as conferring a high risk for DR (SimóServat et al., 2013). One of the major problems is associated with poor characterization of different retinopathy phenotypes. It is fundamental before embarking in a search for candidate genes to define clinical phenotypes characterized by specific patterns of severity and progression of DR. It is clear that it is necessary to identify first and well the DR phenotypes that are associated with rapid progression of the retinopathy to severe forms of the disease, such as macular edema and proliferative retinopathy. Only then, studies on candidate genes are worth pursuing, involving appropriately welldefined subgroups of patients (Warpeha and Chakravarthy, 2003). Recently, our research group performed a caseecontrol association study in type 2 diabetic patients looking for genetic biomarkers that may predict DR progression. A population of 307 type 2 diabetic patients was classified in three different phenotypes of DR progression according to the findings of Nunes et al. (2013). Clinically significant diabetic macular edema occurred after twoyears of follow-up in 26 eyes/patients. This population was examined for genetic association with the three different phenotypes identified in the clinical examination. Eleven candidate genes were chosen based on literature searches. SNPs described for these genes were filtered through bioinformatics tools and 177 were then genotyped in the 307 patients. Statistically significant association (p 0.05) among DR progression phenotypes were found for 11


14


genetic variants in ACE, AKR1B1, ICAM1, NOS, PPARGC1A and VEGFA. Different genetic variants were associated with phenotype B and with phenotype C when compared to patients with phenotype A. Comparison between phenotypes B and C confirmed the findings. These preliminary results suggest that different phenotypes of DR progression are associated with different gene variants. Associations with development of clinically significant macular edema were also analyzed in this study. A statistically significant association was found for two variants of ICAM1 suggesting that an abnormal inflammatory response may be the basis for the development of macular edema. Studies on epigenetic modifications in DR are also expected to add important insights in our understanding of diabetic retinal disease (Kowluru et al., 2013; Tang et al., 2013). Studies involving larger numbers of patients with DR, well phenotyped, are now needed to establish the set of genetic markers that may help predict DR progression to vision-threatening complications in type 2 diabetic patients. 7.2.3. Organ-specific disease-related biomarkers 7.2.3.1. Microaneurysm turnover. Microaneurysms and hemorrhages identified as red-dots are the initial changes seen on ophthalmoscopic examination. They may be counted on fundus photography and red-dot counting has been suggested as an appropriate marker of retinopathy progression (Klein et al., 1995a, b). It must be realized that MA formation and their disappearance are dynamic processes. The disappearance of an MA is not a reversible process and indicates vessel closure and progressive vascular damage. Therefore, to assess progression of retinopathy, MA counting should take into account not only every newly developed MAs identified in a new location but also the disappearing ones. The presence and number of MAs and their rates of formation and disappearance are, therefore, good candidates as markers of retinal vascular damage and may be good indicators of retinopathy progression. Using the RetmarkerDR automated methodology, we analyzed data from a group of 113 type 2 diabetic patients with mild-tomoderate nonproliferative DR, followed up for 2 years as controls in DR clinical trials, and thereafter by usual care at the same institution (Nunes et al., 2009). MA turnover from the initial 2 years and the occurrence of clinically significant macular edema during the following 8 years were analyzed in this retrospective 10-year follow-up study. At the end of the 10-year follow-up period, 17 out of the 113 patients developed clinically significant macular edema needing photocoagulation. An MA formation rate of at least 2 MAs/year was found in 12 of the 17 eyes that developed clinically significant macular edema (70.6%), whereas this was only found in 8 of the 96 eyes that did not develop clinically significant macular edema during the 10-year follow-up period (8.3%). This study showed that in the initial stages of DR higher MA turnover obtained from color fundus photography is a good indicator of retinopathy activity and development of clinically significant macular edema needing photocoagulation. These results have since been confirmed by another research group in Munich using also the RetmarkerDR (Haritoglou et al., 2014). In their study they have analyzed a group of 287 eyes that were followed by fundus photography during a period of 5 years (CALDIRET study). Comparing 47 eyes that did develop clinically significant macular edema over the period of the study with 240 eyes that did not develop clinically significant macular edema they were able to show that an increased MA formation rate was associated with development of clinically significant macular edema. Values of MA formation rate greater than 2 per year in this early

stage of retinopathy are present in 70.2% of the eyes that developed clinically significant macular edema. In clear contrast, the eyes that did not develop clinically significant macular edema during the period of the study showed an MA formation rate less that 2 per year in 71.7% of the cases. This study, using also the RetmarkerDR software confirms the studies by our group in which 70.6% of the eyes that developed clinically significant macular edema showed an MA formation rate greater than 2. More recently, our research group performed a prospective, observational study designed to follow eyes/patients with mild nonproliferative DR (grades 20 and 35) for a period of two years or until the time of development of a vision threatening DR complication, clinically significant macular edema needing laser photocoagulation (Ribeiro et al., 2013b), using RetmarkerDR. MA turnover was 11.2 11.2 in the 26 eyes/patients that developed clinically significant macular edema and 5.0 5.2 in the remaining 322 eyes (p < 0.001). The MA turnover showed a high predictiveness for clinically significant macular edema with a ROC area ¼ 0.695. for an MA turnover cut-off of 9 or more, a sensitivity of 57.7% and a specificity of 81.2% was achieved (i.e. 79.4% of the eyes are correctly classified). Eyes with an MA turnover higher than 9 during the initial 6 month period showed a higher risk for clinically significant macular edema development than eyes with a lower MA turnover (OR ¼ 5.886; 95% CE ¼ (2.503e13.844). The MA turnover predictive values for clinically significant macular edema development were, for the period of two years of follow-up a positive predictive value of 20%, and a negative predictive value of 96%, showing that a low MA turnover value predicts slow disease progression and indicates that development of clinically significant macular edema is unlikely. Furthermore, eyes that developed clinically significant macular edema before the 24 months visit presented a higher MA turnover (26.6 15.9) when compared to the eyes in which clinically significant macular edema was detected only at month 24 (12.8 3.6 e p ¼ 0.018), reinforcing the correlation between high turnover values and risk for the development of clinically significant macular edema for eyes with same ETDRS retinopathy level. Multivariate analysis showed also that MA turnover is predictive of clinically significant macular edema independently of the HbA1c values. MA turnover has, therefore, been validated in both retrospective and prospective studies performed in different centers as a prognostic biomarker of DR progression and development of its most frequent VTDR complication, clinically significant diabetic macular edema. 7.2.3.2. Subclinical macular edema. Clinical evaluation of macular edema before the availability of OCT was characterized by its subjectivity. Slit lamp biomicroscopy demonstrate changes in retinal thickness in the macular area but it is dependent on the observer. Measurement of retinal thickness by OCT is reliable and an increase in retinal thickness defines macular edema (Hee et al., 1995; Pires et al., 2013). A quantitative characterization of macular edema became feasible, as determined by measurements of retinal thickness and volume (Hee et al., 1995; Lang, 2007; Massin et al., 2006; Yang et al., 2001). OCT images of diabetic macular edema depict the presence of low intraretinal reflectivity, due to fluid accumulation in the extracellular space of the retina. The process begins as increased thickening with sponge-like appearance of the retinal layers, showing increase in the extracellular spaces, later advancing to the typical image of cystoid spaces (Alkuraya et al., 2005; Otani et al., 1999). Our research group has now shown that it is possible to identify the blood-retinal barrier alteration with OCT, a noninvasive procedure, in the initial stages of retinal edema, without



the need for the intravenous injection of fluorescein (Bernardes et al., 2011). The term subclinical diabetic macular edema has been proposed to describe the early stages of macular edema (Bressler et al., 2012; Browning and Fraser, 2008; Sng et al., 2012). Nevertheless, there are few data in the literature regarding the natural history of eyes with subclinical macular edema. Browning and Fraser (2008) found progression to clinically significant macular edema in 48 of 153 eyes with subclinical macular edema, over a median follow-up period of 14 months. More recently, the DRCR-net (Bressler et al., 2012) showed that one-quarter to one-half of these eyes will progress to clinically significant diabetic macular edema within 2 years after its identification. Our group followed prospectively a large group of patients with diabetes type 2 and mild nonproliferative DR, during 2 years, with repeated clinical and OCT examinations. Eyes with subclinical macular edema at baseline were identified and rate of progression to clinically significant diabetic macular edema investigated (Nunes et al., 2013). The definition of subclinical macular edema proposed by the DRCR.net (Bressler et al., 2012) consisting in absence of edema involving the center of the fovea on slit-lamp examination, and a center point thickness (CPT) measurement on Stratus OCT of 225 mm and

Phenotypes and biomarkers of diabetic retinopathy. Personalized medicine for diabetic retinopathy: the Weisenfeld award.

HLA phenotypes and diabetic retinopathy.

Biomarkers in Diabetic Retinopathy.

Haptoglobin phenotypes in diabetes mellitus and diabetic retinopathy.

Circulating Biomarkers of Diabetic Retinopathy: An Overview Based on Physiopathology.

Genetic variants in ICAM1, PPARGC1A and MTHFR are potentially associated with different phenotypes of diabetic retinopathy.

Hypomagnesaemia and diabetic retinopathy.

Smoking and diabetic retinopathy.

AIDS and diabetic retinopathy.

Diabetic retinopathy.

Photocoagulation and diabetic retinopathy.

Dyslipidemia and diabetic retinopathy.

Diabetic retinopathy and pregnancy.

Clinical biomarkers and molecular basis for optimized treatment of diabetic retinopathy: current status and future prospects.

Diabetic retinopathy and maculopathy.

Photocoagulation and diabetic retinopathy.

Photocoagulation and diabetic retinopathy.

Diabetic retinopathy.

Diabetic Retinopathy and Diabetic Macular Edema.

Genetics of diabetic retinopathy.

Pathophysiology of diabetic retinopathy.

Management of diabetic retinopathy.

Combined Methods for Diabetic Retinopathy Screening, Using Retina Photographs and Tear Fluid Proteomics Biomarkers.

Cigarette smoking and diabetic retinopathy.