European Review for Medical and Pharmacological Sciences 2007; 11: 155-163 Advances in biochemical mechanisms of diabetic retinopathy

C. GIUSTI, P. GARGIULO*

Department of Ophthalmology, University “Campus Bio-Medico” – Rome (Italy) *Department of Internal Medicine, University “La Sapienza” – Rome (Italy)

Abstract. – mellitus is a major lar subretinal fibrosis or proliferative retinopathy cause of blindness in the working population (PDR) with neovascular glaucoma and tractional of the Western World. Numerous large, retinal detachment (RD). prospective, randomized clinical trials have delineated the current standard prevention and The medical, social and financial impact of treatment protocols including intensive diabetic retinopathy (DR) is substantial: in the glycemic and blood pressure control as well as USA, diabetes affects over eighteen million laser photocoagulation for clinically significant people, being DR present in nearly all persons macular edema and/or proliferative retinopathy with duration of the disease of 20 years or at a high risk for tractional retinal detachment. more. If this complication remains untreated, However, despite all these interventions, vi- about 60% of subjects with PDR are expected sion loss from diabetic retinopathy still occurs at an alarming rate and no data provide an ade- to become blind within five years in one or quate explanation for the serious and rapid in- both eyes. Blindness is estimated to occur year- volvement of the retinal microcirculation that ly in over 10,000 patients affected by diabetes may be observed in the disease despite a good in the USA2-5. metabolic control. In fact, there is now ample Despite the growing concern of DR, its ae- of evidence that the development of diabetic tiopathogenesis is still not completely under- retinopathy is a multifactorial process where genetic, metabolic and growth factors play an stood. Most retinal cells are affected by the important role. Some biochemical mecha- metabolic abnormalities of diabetes, but the nisms, supposed to be involved in the patho- sight-threatening manifestations of DR are ulti- genesis of diabetic retinopathy, have been mately attributable to capillary damage (macu- highlighted in this review. lar edema due to abnormal permeability of bar- rier capillaries, and ischemia with unregulated Key Words: angiogenesis due to capillary closure). There is Biochemical mechanisms, Diabetes, Diabetic no doubt that timely tight glycemic control retinopathy, Retina. with glycosylated hemoglobin (HbA1c) close to the normal range (4.0-6.0%) should lead to a significant decrease in incidence and progres- sion of retinopathy4,5. However, only a minori- ty of diabetics achieve near-normal glycemia soon and on a long-term basis and no data pro- vide an adequate explanation for the serious Introduction and rapid involvement of the retinal microcir- culation that may be observed in the disease Diabetes mellitus is a major cause of blindness despite a good metabolic control. Moreover, in among young adults in economically developed a subgroup of patients, proliferation may per- societies1-5. Its most frequent and potentially sist even after well done full scatter laser ap- blinding complications are represented by exuda- plications (still the first-line treatment of tive maculopathy with clinically significant mac- choice for high risk retinal microangiopathy), ular edema (DME), intravitreal or preretinal he- resulting in a 30-50% risk of severe visual loss morrhages, macular pigmentary changes, macu- within the next 5 years2,3.

Corresponding Author: Cristiano Giusti, MD; e-mail: [email protected] 155 C. Giusti, P. Gargiulo

Pathogenesis formation of advanced glycation end products (AGEs); increased de-novo synthesis of diacyl- Hyperglycemia glycerol from , causing -kinase C At present, the most effective medical treatment (PKC) activation; oxidative-nitrosative stress for DR is represented by glycemic control2-6. Two with overproduction of major trials have demonstrated the effect of “in- (ROS)2,3. More recently, it has been established tensive” blood glucose in reducing the incidence that reactive oxygen and nitrogen species trigger and progression of DR. The Diabetes Control activation of mitogen-activated protein kinases and Complications Trial (DCCT) has implicated (MAPKs) and poly(ADP-ribose) polymerase hyperglycemia as a major pathogenetic factor in (PARP), as well as the inflammatorycascade, and type 1 patients and a strong correlation has been these downstream mechanisms are also involved observed between the glycemic control and the in the pathogenesis of diabetes complications. incidence and progression of diabetic microvas- The interactions among various hyperglycemia- cular complications. In particular, the adjusted initiated mechanisms are not completely under- mean risk for development of any retinopathy stood, and the relationship between increased al- was reduced by 76% in the intensive therapy, dose reductase activity and the oxidative-ni- compared with the conventional group. For those trosative stress/PARP activation has recently be- with some retinopathy already, the intensive come a focus of interest. According to several group had a higher incidence of progression dur- studies performed in the diabetic lens, nerve, reti- ing the first year whereas from 3 years onwards, na and high-glucose–exposed endothelial cells, the progression of retinopathy was reduced in the increased activity leads to ox- intensive group by 54%. An adverse effect shown idative stress9. However, it has also been reported by the DCCT is that intensive diabetes therapy that increased aldose reductase activity is a con- reduces plasma levels of LDL cholesterol and sequence rather than a cause of triglycerides but increases the risk of major (in particular, mitochondrial superoxide produc- weight gain, which might adversely affect the tion) and PARP activation in the pathogenesis of risk of cardiovascular disease7. diabetes complications9. Similar results were observed for type 2 dia- As a consequence of all this, beside the opti- betics in the United Kingdom Prospective Dia- mal glycemic control that always needs to be betes Study (UKPDS) group8. In this study, pa- achieved in each patient, pharmacologic inhibi- tients who were assigned to intensive glucose tion of the above indicated pathways might pre- control had a 25% risk reduction in microvascu- vent some of the characteristic lesions of DR, lar endpoints, including the need for retinal pho- such as loss of retinal pericytes and microa- tocoagulation. Both studies showed that neurysm formation, changes in retinal hemody- glycemic control is protective for all levels of namics, and aberrant neovascularization. control: there is no glycemic threshold below which a reduction in microvascular complica- Polyol Accumulation tions is not observed. The current recommenda- Aldose reductase, the first and rate-limiting tion is for maintaining the glucose levels as near enzyme in the polyol pathway, reduces glucose normal as possible. However, because of some to using NADPH as a cofactor; sorbitol risks associated to hypoglycemia (such as hospi- is then metabolized to by sorbitol de- talizations and possibly deaths while operating hydrogenase, which uses NAD+ as a cofactor9. motor vehicles), the glycemic control targets The polyol (sorbitol) pathway of glucose me- should be individualized. Therapy should be di- tabolism is activated in many cell types when rected toward achieving the lowest glycemic lev- intracellular glucose concentrations are very el that is the safest in terms of hypoglycemic risk high, and it can generate cellular oxidative for each patient. stress through a variety of biochemical abnor- Hyperglycemia is involved in the pathogenesis malities, including myo- depletion and of , retinopathy, nephropathy, downregulation of Na/K ATP-ase activity, and macrovascular disease via multiple mecha- NAD+/NADH and NADP+/NADPH im- nisms, the best studied of which are the follow- balances, changes in fatty acid metabolism, im- ing: increased flow through the aldose-reductase paired neurotrophic support, and upregulation pathway (increased aldose reductase activity); of vascular endothelial growth factor (VEGF)10. nonenzymatic glycation and glycoxidation with The polyol pathway appears to be both a

156 Advances in biochemical mechanisms of diabetic retinopathy

“dream” and a “dread” target by devising venting retinal and neural damage in diabetes strategies to prevent DR. The pathway is a have been highlighted, thus justifying further dream target because its activation is immedi- clinical trials of specific, potent, and low-toxic ately linked to hyperglycemia, generates vari- ARI9. ous types of cellular stress, and occurs promi- nently in the tissues that develop complications, AGEs Accumulation thus promising returns beyond retinopathy. In Chronically increased amounts of glucose am- addition, polymorphisms of the aldose reduc- plify the physiological process of nonenzymatic tase gene may help in predicting individual sus- protein glycosylation (glycation). For example, ceptibility to retinopathy and other microvascu- (HbA1c) is an acknowl- lar complications, and the enzymatic function edged indicator of time-integreted glycemia4. of aldose reductase can be specifically inhibit- Glucose forms labile links with the NH2-termi- ed. However, the polyol pathway has become a nal and side-chain lysine radicals within , dread target because aldose reductase inhibitors which, after a cascade of molecular rearrange- (ARIs) have yielded inconsistent results in the ments (Maillard reaction), result in molecules of diabetic or diabetic-like retinopathy of experi- brown color and specific fluorescence (Amadori mental animals and only minor benefits in hu- products), leading to degradation of both struc- man DR10. tural and functional proteins and accelerated ag- However, the polyol pathway seem to be real- ing. While most glycated proteins are eliminated ly a rational candidate mechanism for the gan- in physiological conditions, they accumulate, on glion cell apoptosis and Müller glial cell activa- the contrary, in the presence of diabetes and un- tion10,11. Ganglion and Müller cells are the retinal dergo further structural arrangements with the cells most consistently found to contain aldose formation of AGEs, which, in turn, are implicat- reductase in all species studied, including hu- ed in the development of vascular lesions having mans. Since neuroglial changes may cause vas- a proven effect in determining a significant loss cular changes, and given the general agreement of mural pericytes. Moreover, it is also clear that that at least the pericytes of retinal capillaries AGEs formation mechanisms are diverse and contain aldose reductase, the inhibition of the complex, encompassing both non-oxidative (gly- polyol pathway could also prevent the vascular cation) and oxidative (glycoxidation) pathways. abnormalities of DR. In fact, inhibition of aldose These reactions, together with intra- and inter- reductase was also able to prevent the early acti- molecular cross-link formation, are able to modi- vation of complement in the retinal vessel wall fy structure and function of target molecules in as well as the apoptosis of vascular pericytes and such a way that they do not respond anymore to endothelial cells and the development of acellu- biological signals12,13. Interaction of AGEs with lar capillaries. Moreover, retinal endothelial cells their receptors (RAGE) has also been implicated showed aldose reductase immunoreactivity, and in enhanced ROS formation and inflammatory human retinas exposed to high glucose in organ vascular complications. For example, N(car- culture increased the production of sorbitol. Fi- boxymethyl)lysine-protein (CML-protein), nally, experimental evidence exists that defects macrophage colony stimulating factors (M-CSF) in the polyol pathway may produce thickening and soluble vascular cell adhesion molecule-1 of the capillary basement membrane, loss of (sVCAM-1) have been found to be increased in mural pericytes and microaneurysm formation, patients with diabetic micro-angiopathy and the earliest vascular features of diabetic mi- CML-human serum protein (CML-HSP) levels, croangiopathy. In fact, high glucose levels in- which are at variance with the HbA1c index for crease flux through the polyol pathway with the blood glucose, have been proposed as a good enzymatic activity of aldose reductase, thus de- biomarker both for glycoxidation and for the de- termining a build-up of intracellular sorbitol velopment of microvascular complications in concentrations and, consequently, an osmotic type II diabetes. damage to the vascular cells. Therefore, it seems The use of compounds that inhibit AGE for- possible to conclude that excess of aldose reduc- mation (such as pimagedine and aminoguani- tase activity might be a mechanism in the patho- dine) has been investigated as a possible thera- genesis of DR9-11. peutic intervention, highlighting promising re- Positive preliminary results of some ARIs sults in the prevention of DR in animal models. (such as sorbinil, zenarestat or fidarestat) in pre- Preliminary results for human diabetic nephropa-

157 C. Giusti, P. Gargiulo thy have been published and other clinical trials may result from various mechanisms, including are under way to confirm the efficacy of this new glucose auto-oxidation, protein glycation, in- kind of treatment in preventing the onset of dia- creased flux through the polyol pathway, and betic microvascular complications14,15. prostanoid production. These high ROS levels are thought to determine structural and functional PKC Inhibitors changes in all cellular components, leading to There is a large body of evidence to support DNA and protein modification and lipid peroxi- the hypothesis that PKC plays a major role in hy- dation. In particular, pericytes are highly sensi- perglycemia-induced microvascular dysfunction tive to the oxidative stress not only directly but in diabetes16-18. also indirectly, due to significantly decreased lev- In fact, hyperglycemia leads to persistent de els of scavenging enzymes and increased rate of novo synthesis and activation of PKC, induced apoptosis. Pericyte loss or functional deficiency by increased glucose availability through de novo have been found to reduce the inhibition of en- synthesis of diacyglycerol (DAG), which, in turn, dothelial proliferation in vivo. As damage pro- is associated with a number of biochemical and gresses, the blood vessel wall becomes more metabolic abnormalities, including increased ex- porous, letting proteins and other matherials leak pression of matrix proteins, such as and out abnormally, thus determining the typical fea- fibronectin, and increased expression of vasoac- tures of nonproliferative DR (e.g. hard exudates tive mediators, such as endothelin. The net effect and clinically significant DME). of these changes may be manifested as basement Furthermore, animal studies suggest that an- membrane thickening and changes in vessel per- tioxidants such as vitamin E may prevent some meability and/or blood flow. Although the activi- of the vascular dysfunction associated with dia- α β β δ ε ty of multiple PKC isoforms ( , 1, 2, and ) betes by means of several different mechanisms: is increased in vascular diabetic tissues, studies reduced retinal DAG levels; normalized PKC-β β suggest that the PKC- 2 isoform is preferentially activation; normalized retinal blood flow; re- activated18. Moreover, PKC-β has been shown to stored -mediated endothelium-depen- be an integral component of cellular signaling by dent relaxation. The use of anti-oxidants is vascular endothelial growth factor (VEGF), an promising, but further studies are needed to de- important mediator of ocular neovascularization, termine appropriate doses, and/or whether this secondary to retinal ischemia and DME, thus approach will translate into long-term benefits of stimulating retinal pericyte proliferation18. reduced DR and DME5. A selective inhibitor of PKC-β, ruboxistaurin mesylate (LY333531), was initially reported to Endothelium-Related Dysfunction of the prevent the increase in leukostasis and the de- Coagulant and Anticoagulant Pathways crease in blood flow in the retinas of transgenic Besides but strictly associated with these diabetic rats. However, multicenter clinical trials metabolic vascular changes, there are many re- in humans did not show a same effectiveness on ports that extensively document the presence of a DR progression17,18. A possible reason of that remarkable endothelium-related dysfunction of might be the observation that PKC inhibition the coagulant and anticoagulant pathways in dia- augments pro-apoptotic effects of high glucose betics, but it is not yet clear if this condition is on cultured pericytes. Therefore, it seem possible due to hyperglycemia only24-26. Abnormal coagu- to conclude that the potential effectiveness of lation is manifested by enhanced prothrombin PKC-β inhibitors on DR progression is at this conversion to thrombin – as demonstrated both time still controversial19,20. by increased activated factor VII (FVII:c) and prothrombin degradation products (F1 + 2) plas- Oxidative Damage ma levels and downregulation of the anticoagu- Diabetes and hyperglycemia are associated lant pathway, caused by reduced antithrombin III with increase in oxidative stress, and overproduc- activity and thrombomodulin endothelial recep- tion of ROS (free radicals) are thought to be re- tors. In particular, it has been suggested that in sponsible for microvascular damage, being con- patients without retinal lesions, despite the hy- sistent with increased malonyldialdehyde, iso- perglycemic condition, the vascular endothelium prostanes, nitrotyrosine or 8-hydroxy-2’de- preserves its physiological thromboresistance. oxyguanosine levels as well as an overall de- On the contrary, in the cases of PDR a hyperco- creased antioxidant status21-23. Production of ROS agulable state is present as a consequence of a

158 Advances in biochemical mechanisms of diabetic retinopathy functional conversion of the endothelium from a thy and nephropathy in animals with experi- thromboresistant to a thrombogenic surface with mental diabetes27-30. In particular, it has been re- activation of the extrinsic haemostatic pathway. cently reported that thiamine and benfotiamine In fact, a higher platelet and erythrocyte sedi- reduce aldose reductase mRNA expression and mentation rate, with a elevated erythrocyte sodi- activity, sorbitol concentrations, and intracellu- um-lithium countertrasport activity, has also been lar glucose while increasing the expression and observed in diabetics, as well as: increased plas- activity of transketolase in human endothelial ma concentrations of tissue plasminogen activa- cells and bovine retinal pericytes cultured in tor (tPA) inhibitor (PAI-1), von Willebrand-anti- high glucose31. Therefore, thiamine and ben- haemophilic factor A and fibrinogen; decreased fothiamine correct the polyol pathway activa- concentrations of endothelium-derived relaxing tion induced by high glucose in vascular cells: factor, prostacyclin and tPA; increased fi- in fact, the activation of transketolase may shift bronectin and thrombomodulin levels and re- excess glycolytic metabolites into the pentose duced fibrinolytic potential of vascular endothe- phosphate cycle, accelerate the glycolytic flux, lia25. and reduce intracellular free glucose, thereby The endothelin system is probably of great im- preventing its conversion to sorbitol. These ef- portance in mediating the diabetes-induced reti- fects on the polyol pathway, together with other nal vascular dysfunction as suggested by in- beneficial effects reported for thiamine in high creased endothelin-1 levels (see below). More- glucose, could justify testing thiamine as a po- over, the finding of antipericyte and phospho- tential approach to the prevention and/or treat- lipid-binding autoantibodies (e.g. Lupus Antico- ment of diabetic complications. agulant)26, as well as the presence of several im- munocompetent cells (activated T lymphocytes, Hypercholesterolemia Lipoproteins B lymphocytes, macrophages, HLA DR/DQ ex- The beneficial effect on DR of a medical treat- pressing cells and immunoglobulin deposits) in ment for hypercholesterolemia is still unclear: the vitreous and in preretinal membranes of the Early Treatment Diabetic Retinopathy Study PDR, suggests that the immune system might (ETDRS) and the Wisconsin Epidemiologic play an important, but still unkown, role in the Study of Diabetic Retinopathy (WESDR) pointed early pathophysiology of DR25. out a significant association between elevated serum lipid levels and increased risk of retinal Thiamine and Benfotiamine hard exudates2-4,25. However, it was also reported Blockade of the unifying mechanism of the that lowering cholesterol by therapeutic means pathogenesis of diabetic microangiopathy may may not be indicated for the sole purpose of de- be also obtained through another, more radical, creasing the incidence and progression of these approach in order to correct the metabolic imbal- diabetic retinal lesions. In one recently published ances induced by hyperglycemia. Thiamine (vita- report only, the severity of retinopathy was posi- min B1) is the cofactor of three enzymes in- tively associated with triglycerides and negatively volved in and Kreb’s cycle: transketo- associated with HDL cholesterol, being retinopa- lase, which shifts 2,3-diphosphoglycerate (2,3- thy positively associated with small and medium DPG) into the pentose phosphate shunt; pyruvate VLDL and negatively with VLDL size32. In men dehydrogenase, which links pyruvate to acetyl- only, DR was positively associated with small CoA and channels them into the Kreb’s cycle; α- LDL, LDL particle concentration, apolipoprotein ketoglutarate dehydrogenase, which accelerates B concentration, small HDL and was negatively the Kreb’s cycle by turning α-ketoglutarate into associated with large LDL, LDL size, large HDL, succinyl-CoA. The net result is facilitation of and HDL size. No associations were found with glucose metabolism, with accumulation of highly apolipoprotein A1, Lipoprotein(a), or susceptibil- reactive metabolites that, like 2,3-DPG, play a ity of LDL to oxidation33. Elevated apolipoprotein key role in the synthesis of AGE and the activa- C-III (apoC-III) levels have been reported to be tion of DAG/PKC. associated with increased macrovascular disease It has been demonstrated that thiamine and risk and an independent positive association of benfotiamine (a lipophilic analogue that can be apoC-III level with microvascular complications administered orally) are able to correct in- of type 1 diabetes was documented34. However, creased lactate and AGE formation in cultured neither gender nor glycemia influence LDL oxi- endothelial cells and found to prevent retinopa- dation in vitro35.

159 C. Giusti, P. Gargiulo

In conclusion, the exact role of serum lipopro- Vitreal elevation of matrix metalloproteinases tein concentrations in the development of DR (MMPs)-2,-9 and their tissue inhibitors (TIMPs)- needs to be elucidated by further investigations. 1,-4 has also been reported in diabetic patients and correlated to the severity of retinopathy44,45. Cytokines Interleukines MMPs are a tightly regulated family of zinc-de- Functional damage and necrosis of the retinal pendent endopeptidase, capable of degrading all capillaries are probably the main contributing components of the extracellular matrix and basa- factors to the breakdown of the blood-retina bar- ment membranes and affecting the cell-cell and rier and the onset of nonproliferative DR and cell-matrix interactions. Given the early DME in diabetic patients2-4,36. However, high lev- histopathological features of DR (thickening of els of serum interleukine (IL)-6 seem also to in- the capillary basament membrane and loss of fluence the development of DME while other pericytes with microaneurysm formation), these measured serum cytokines (e.g. TGF-β, AGE, matrix degrading enzymes might play a role in TNF-α) have not been correlated to the severity the pathogenesis of DR. In fact, MMP activity of DR36. In particular, Funatsu et al. observed not seems to represent a “final common pathway” in only that a significant relationship is present be- the process of retinal neovascularization, from tween VEGF and IL-6 but also that their aqueous whatever cause44 and a therapeutical inhibition at levels are significantly correlated with the severi- this level could be potentially more attractive ty of fundus findings. In these cases, both con- than targeting individual systems such as VEGF. centrations were found to be higher in the aque- In fact, even if this could be effectively achieved, ous than in the plasma37. there might be some other “escapes” by different In different studies, these Authors observed inducers of neovascularization employing this fi- that the IL-6 plasma levels and the degree of nal common pathway. MMP-9 might also be in- the posterior vitreous detachment (PVD) may volved in the hemorrhagic transformation of the be significantly correlated with the severity of vitreous in patients with PDR46. macular edema whereas the aqueous levels of Finally, increased levels of macrophage migra- VEGF and IL-6 may predict the postoperative tion inhibitory factor (MIF) have been observed exacerbation of macular edema in patients with in the vitreous of patients with PDR together NPDR after phacoemulsification for cataract with a significant association between MIF levels surgery38,39. On the contrary, the vitreous levels and grades of fibrous proliferation47,48. MIF is the of angiotensin II (AII), VEGF and IL-6 were first T cell-derived soluble lymphokine reported found to be elevated in DME patients irrespec- to prevent random migration of macrophages, to tive of the status of PVD, suggesting that they recruit them at inflammatory loci and to enhance may promote an increase of vascular perme- their activity such as adherence, motility and ability without PVD40,41. However, the vitreous phagocytosis. Aqueous MIF levels are signifi- itself should play a still unknown role, as sub- cantly correlated with aqueous monocyte chemo- jects who have posterior vitreous detachment tactic protein-1 (MCP-1) levels and the clinical rarely develop DME or PDR25. stage of DR, suggesting that both MIF and MCP- A significant relationship between VEGF and 1 might have a co-operative role in the progres- IL-6 levels in aqueous humor and in vitreous flu- sion of DR48. id has been recently reported42, thus concluding that these measurements may be useful to ana- Endothelin-1 lyze the pathogenesis of DR and to predict the Endothelin-1 (ET-1) is a peptide produced by progression of the retinal disease. An increased endothelial cells that induces vasoconstriction by expression of IL-6 has also been observed in sur- interacting with endothelin A receptors (ETA) on gically removed epiretinal membranes41. the vascular smooth muscle cells and vasodilata- What’s more, serum and vitreous leptin, a tion by interacting with the endothelin B recep- pleiotropic cytokine with reported angiogenic ac- tors (ETB) on the vascular endothelial cells, re- tivity, have been found to be higher in patients sulting in the release of endothelium derived ni- with diabetes than in those without, and vitreal tric oxide and prostacyclin. It has been shown concentrations were especially elevated in pa- that hypoxia induces the ET-1 gene expression in tients with PDR. Leptin and leptin receptors were endothelial cells49. The presence of an ET-1 sys- detected in fibrovascular epiretinal membranes of tem in the eye is well established and plays a role diabetic subjects43. on ocular blood flow, glial proliferation, and col-

160 Advances in biochemical mechanisms of diabetic retinopathy lagen matrix contraction by the retinal pigmented tion and tube formation. Therefore, because of epithelial cells. Immunoreactive ET-1 in human this complex interplay, targeting a single growth vitreous has been found to be elevated in the factor will be unlikely to result in therapeutic in- presence of proliferative vitreoretinopathy (PVR) hibition of angiogenesis. However, this important (such as PDR), RD, and idiopathic epiretinal topic will be specifically analysed in a further re- membranes49. ET-1 and its receptors ETA and view. ETB are present in epiretinal tissue of both idio- Within the near future, pharmacologic treat- pathic and PVR membranes, thus suggesting an ment will probably be available for treating and involvement of ET-1 in these retinal diseases. preventing the progression of DR. Antioxidant Moreover, as a result of the demonstrated associa- administration such as high-dose vitamin E may tion between enhanced ET-1 expression and PKC help to reduce the oxidative stress from diabetes activation in early diabetes, a PKC inhibitor could and hyperglycemia, but further studies are need- be able to reverse the upregulation of ET-150. The ed to determine whether retinopathy progression therapeutical effect of a long-term selective can be reduced. Aldose reductase inhibitors have blockade of the ETA receptor has also been re- been developed for the prevention of retinal and cently evaluated in a genetic mouse model of neural damage in diabetes, but additional clinical nonobese type 1 diabetes (NOD), in order to sug- trials are needed to assess whether they can delay gest a new strategy for preventing the develop- or stop the progression of DR and macular ede- ment of retinopathy51. In this study, an associated ma effectively. A long-term selective blockade of upregulation of ET-1 and adrenomedullin (an an- the ET-1 receptors, once tested in humans, could giogenic factor) was documented51. This novel also disclose a novel therapeutical approach of approach to DR has also been confirmed by an- DR in the future. other study, in which an ETA/ETB dual receptor The increasing use of medical therapies such antagonist was found to reverse the VEGF ex- as pharmacologic agents55 will also require pression levels in the diabetic rat retina52. greater communication between ophthalmolo- gists, diabetologists, and primary care physi- cians. This interdisciplinary cooperation will aid in identifying those patients most at risk for vi- Conclusion sion loss and those most likely to benefit from new treatments once they become available. DR is the most severe ocular complication of diabetes, its earliest clinical signs being repre- sented by microaneurysms and haemorrhages. Later signs include dilated, tortuous irregular veins and retinal non-profusion, leading to retinal References ischaemia that ultimately results in neovasculari- sation. DME, which is caused by the breakdown 1) GARDNER TW, ANTONETTI DA, BARBER AJ, LANOUE KF, LEVISON SW. Diabetic retinopathy: more than of the blood-retinal barrier, also occurs and is re- meets the eye. Surv Ophthalmol 2002; 47: S253- sponsible for a major part of vision loss, particu- S262. larly in type 2 diabetes. The pathogenesis of DR 2) AIELLO LP. Perspectives on diabetic retinopathy. is very complex. Many biochemical mechanisms Am J Ophthalmol 2003; 136: 122-135. have been proposed as potential explanations for 53 3) FONG DS, AIELLO LP, FERRIS FL 3RD, KLEIN R. Diabetic the development and progression of DR . retinopathy. Diabetes Care 2004; 27: 2540-2553. Chronic hyperglycaemia leads to oxidative in- jury, microthrombi formation, cell adhesion, 4) FONG DS, AIELLO L, GARDNER TW, KING GL, BLANKEN- SHIP G, CAVALLERANO JD, FERRIS FL 3RD, KLEIN R. molecule activation, leukostasis and cytokine ac- American Diabetes Association. Retinopathy in tivation. Last but not least, the retinal ischaemia diabetes. Diabetes Care 2004; 27(Suppl 1): S84- induces a demonstrated overexpression of growth S87. factors (VEGF, IGF-1, angiopoetin-1 and -2, 5) PORTA M, ALLIONE A. Current approaches and per- stromal-derived factor-1, fibroblast growth fac- spectives in the medical treatment of diabetic tor-2, tumour necrosis factor and several retinopathy. Pharmacol Ther 2004; 103: 167-177. 54 others) , which act in synergy in mediating the 6) SCHAUMBERG DA, GLYNN RJ, JENKINS AJ, LYONS TJ, RI- steps toward angiogenesis, including protease FAI N, MANSON JE, RIDKER PM, NATHAN DM. Effect of production, endothelial cell proliferation, migra- intensive glycemic control on levels of markers of

161 C. Giusti, P. Gargiulo

inflammation in type 1 diabetes mellitus in the di- 18) YAMAGISHI S, TAKEUCHI M. Inhibition of protein ki- abetes control and complications trial. Circulation nase C might be harmful to diabetic retinopathy. 2005; 111: 2446-2453. Med Hypotheses 2004; 63: 135-137.

7) KILPATRICK ES, RIGBY AS, ATKIN SL. The effect of glu- 19) WEGEWITZ U, GOHRING I, SPRANGER J. Novel ap- cose variability on the risk of microvascular com- proaches in the treatment of angiogenic eye dis- plications in type 1 diabetes. Diabetes Care 2006; ease. Curr Pharm Des 2005; 11: 2311-2330. 29: 1486-1490. 20) COMER GM, CIULLA TA. Pharmacotherapy for dia- 8) STRATTON IM, CULL CA, ADLER AI, MATTHEWS DR, NEIL betic retinopathy. Curr Opin Ophthalmol 2004; 15: HA, HOLMAN RR. Additive effects of glycaemia and 508-518. blood pressure exposure on risk of complications 21) MATTEUCCI E, GIAMPIETRO O. Oxidative stress in in type 2 diabetes: a prospective observational families of type 1 diabetic patients. Diabetes Care study (UKPDS 75). Diabetologia 2006; 49: 1761- 2001; 24: 167-168. 1769. 22) YOKOI M, YAMAGISHI SI, TAKEUCHI M, OHGAMI K, 9) OBROSOVA IG, PACHER P, S ZABO C, ZSENGELLER Z, HI- OKAMOTO T, S AITO W, M URAMATSU M, IMAIZUMI T, ROOKA H, STEVENS MJ, YOREK MA. Aldose reductase OHNO S. Elevations of AGE and vascular en- inhibition counteracts oxidative-nitrosative stress dothelial growth factor with decreased total an- and poly(ADP-ribose) polymerase activation in tioxidant status in the vitreous fluid of diabetic pa- tissue sites for diabetes complications. Diabetes tients with retinopathy. Br J Ophthalmol 2005; 89: 2005; 54: 234-242. 673-675. 10) DAGHER Z, PARK YS, ASNAGHI V, H OEHN T, G ER- 23) AMANO S, YAMAGISHI S, INAGAKI Y, N AKAMURA K, HARDINGER C, LORENZI M. Studies of rat and human TAKEUCHI M, INOUE H, IMAIZUMI T. Pigment epitheli- retinas predict a role for the polyol pathway in hu- um-derived factor inhibits oxidative stress-in- man diabetic retinopathy. Diabetes 2004; 53: duced apoptosis and dysfunction of cultured reti- 2404-2411. nal pericytes. Microvasc Res 2005; 69: 45-55. 11) OYAMA T, M IYASITA Y, W ATANABE H, SHIRAI K. The role 24) GIUSTI C, SCHIAFFINI R, BRUFANI C, PANTALEO A, VINGO- of polyol pathway in high glucose-induced en- LO EM, GARGIULO P. Coagulation pathways and di- dothelial cell damages. Diabetes Res Clin Pract abetic retinopathy: abnormal modulation in a se- 2006 (in press). lected group of -dependent diabetic pa- 12) WAUTIER MP, MASSIN P, G UILLAUSSEAU PJ, HUIJBERTS M, tients. Br J Ophthalmol 2000; 84: 591-595. LEVY B, BOULANGER E, LALOI-MICHELIN M, WAUTIER JL. 25) GIUSTI C. Retinopathy in juvenile diabetes: a 10- N(carboxymethyl)lysine as a biomarker for mi- year (1990-2000) review. Pediatr Diabetes 2001; crovascular complications in type 2 diabetic pa- 2: 83-93. tients. Diabetes Metab 2003; 29: 44-52. 26) GIUSTI C. Are phospholipid-binding antibodies im- 13) WAUTIER MP, BOULANGER E, GUILLAUSSEAU PJ, MASSIN plicated in the pathogenesis of diabetic microan- P, W AUTIER JL. AGEs, macrophage colony stimulat- giopathy? Med Hypotheses 2004; 63: 235-238. ing factor and vascular adhesion molecule blood levels are increased in patients with diabetic mi- 27) BABAEI-JADIDI R, KARACHALIAS N, AHMED N, BATTAH S, croangiopathy. Thromb Haemost 2004; 91: 879- THORNALLEY PJ. Prevention of incipient diabetic 885. nephropathy by high-dose thiamine and benfoti- amine. Diabetes 2003; 52: 2110-2120. 14) BOLTON WK, CATTRAN DC, WILLIAMS ME, ADLER SG, APPEL GB, CARTWRIGHT K, FOILES PG, FREEDMAN BI, 28) HAMMES HP, DU X, EDELSTEIN D, TAGUCHI T, M AT- RASKIN P, R ATNER RE, SPINOWITZ BS, WHITTIER FC, SUMURA T, J U Q, LIN J, BIERHAUS A, NAWROTH P, WUERTH JP; ACTION I INVESTIGATOR GROUP. Random- HANNAK D, NEUMAIER M, BERGFELD R, GIARDINO I, ized trial of an inhibitor of formation of advanced BROWNLEE M. Benfotiamine blocks three major glycation end products in diabetic nephropathy. pathways of hyperglycemic damage and pre- Am J Nephrol 2004; 24: 32-40. vents experimental diabetic retinopathy. Nat Med 2003; 9: 294-299. 15) YATOH S, MIZUTANI M, YOKOO T, K OZAWA T, S ONE H, TOYOSHIMA H, SUZUKI S, SHIMANO H, KAWAKAMI Y, O KU- 29) KARACHALIAS N, BABAEI-JADIDI R, KUPICH C, AHMED N, DA Y, Y AMADA N. Antioxidants and an inhibitor of THORNALLEY PJ. High-dose thiamine therapy coun- advanced glycation ameliorate death of retinal ters dyslipidaemia and advanced glycation in microvascular cells in diabetic retinopathy. Dia- streptozotocin-induced diabetic rats. Ann NY betes Metab Res Rev 2006; 22: 38-45. Acad Sci 2005; 1043: 777-783.

16) AIELLO LP. The potential role of PKC b in diabetic 30) BELTRAMO E, BERRONE E, BUTTIGLIERI S, PORTA M. Thi- retinopathy and macular edema. Surv Ophthal- amine and benfotiamine prevent increased apop- mol 2002; 47: S263-S269. tosis in endothelial cells and pericytes cultured in high glucose. Diabetes Metab Res Rev 2004; 20: 17) AIELLO LP, CAHILL MT, CAVALLERANO JD. Growth fac- 330-336 tors and protein kinase C inhibitors as novel ther- apies for the medical management diabetic 31) BERRONE E, BELTRAMO E, SOLIMINE C, APE AU, PORTA retinopathy. Eye 2004; 18: 117-125. M. Regulation of intracellular glucose and polyol

162 Advances in biochemical mechanisms of diabetic retinopathy

pathway by thiamine and benfotiamine in vascu- 43) MABERLEY D, CUI JZ, MATSUBARA JA. Vitreous leptin lar cells cultured in high glucose. J Biol Chem levels in retinal disease. Eye 2006; 20: 801-804. 2006; 281: 9307-9313. 44) ISHIZAKI E, TAKAI S, UEKI M, MAENO T, M ARUICHI M, 32) Lyons TJ, Jenkins AJ, Zheng D, Lackland DT, SUGIYAMA T, O KU H, IKEDA T, M IYAZAKI M. Correlation McGee D, Garvey WT, Klein RL. Diabetic between angiotensin-converting enzyme, vascular retinopathy and serum lipoprotein subclasses in endothelial growth factor, and matrix metallopro- the DCCT/EDIC cohort. Invest Ophthalmol Vis teinase-9 in the vitreous of eyes with diabetic Sci 2004; 45: 910-918. retinopathy. Am J Ophthalmol 2006; 141: 129-134.

33) JENKINS AJ, ROWLEY KG, LYONS TJ, BEST JD, HILL MA, 45) JACQUEMINET S, BEN ABDESSELAM O, CHAPMAN MJ, KLEIN RL. Lipoproteins and diabetic microvascular NICOLAY N, FOGLIETTI MJ, GRIMALDI A, BEAUDEUX JL. complications. Curr Pharm Des 2004; 10: 3395- Elevated circulating levels of matrix metallopro- 3418. teinase-9 in type 1 diabetic patients with and without retinopathy. Clin Chim Acta 2006; 367; 34) KLEIN RL, MCHENRY MB, LOK KH, HUNTER SJ, LE NA, 103-107 JENKINS AJ, ZHENG D, SEMLER A, PAGE G, BROWN WV, LYONS TJ, GARVEY WT; DCCT/EDIC RESEARCH GROUP. 46) DESCAMPS FJ, MARTENS E, KANGAVE D, STRUYF S, GEBOES Apolipoprotein C-III protein concentrations and K, VAN DAMME J, OPDENAKKER G, ABU EL-ASRAR AM. gene polymorphisms in Type 1 diabetes: associa- The activated form of gelatinase B/matrix metallo- tions with microvascular disease complications in proteinase-9 is associated with diabetic vitreous the DCCT/EDIC cohort. J Diabet Complications hemorrhage. Exp Eye Res 2006; 83: 401-407. 2005; 19: 18-25. 47) MITAMURA Y, T AKEUCHI S, MATSUDA A, TAGAWA Y, M IZUE 35) JENKINS AJ, LYONS TJ, ZHENG D, OTVOS JD, LACKLAND Y, N ISHIHIRA J. Macrophage migration inhibitory fac- DT, MCGEE D, GARVEY WT, KLEIN RL; DCC/EDIC RE- tor levels in the vitreous of patients with prolifera- SEARCH GROUP. Serum lipoproteins in the diabetes tive diabetic retinopathy. Br J Ophthalmol 2000; control and complications trial/epidemiology of di- 84: 636-639. abetes intervention and complications: associa- tions with gender and glycemia. Diabetes Care 48) TASHIMO A, MITAMURA Y, N AGAI S, NAKAMURA Y, O HTSU- 2003; 26: 810-818. KA K, MIZUE Y, N ISHIHIRA J. Aqueous levels of macrophage migration inhibitory factor and mono- 36) AROCA PR, SALVAT M, FERNANDEZ J, MENDEZ I. Risk cyte chemotactic protein-1 in patients with diabetic factors for diffuse and focal macular edema. J Di- retinopathy. Diabet Med 2004; 21: 1292-1297. abetes Complications 2004; 18: 211-215. 49) ROLDAN-PALLARES M, ROLLIN R, MEDIERO A, MARTINEZ- 37) FUNATSU H, YAMASHITA H, SHIMIZU E, KOJIMA R, HORI MONTERO JC, FERNANDEZ-CRUZ A, BRAVO-LLATA C, FER- S. Relationship between vascular endothelial NANDEZ-DURANGO R. Immunoreactive ET-1 in the growth factor and interleukin-6 in diabetic vitreous humor and epiretinal membranes of pa- retinopathy. Retina 2001; 21: 469-477. tients with proliferative vitreoretinopathy. Mol Vis 38) SHIMIZU E, FUNATSU H, YAMASHITA H, YAMASHITA T, 2005; 11: 461-471. HORI S. Plasma level of interleukin-6 is an indica- 50) ZHU Q, XU X, XIA X, GU Q, HO PC. Role of protein tor for predicting diabetic macular edema. Jpn J kinase C on the alteration of retinal endothelin-1 Ophthalmol 2002; 46: 78-83. in streptozotocin-induced diabetic rats. Exp Eye 39) FUNATSU H, YAMASHITA H, NOMA H, SHIMIZU E, MIMU- Res 2005; 81: 200-206. RA T, H ORI S. Prediction of macular edema exacer- 51) Shaw SG, Boden JP, Biecker E, Reichen J, bation after phacoemulsification in patients with Rothen B. Endothelin antagonism prevents dia- nonproliferative diabetic retinopathy. J Cataract betic retinopathy in NOD mice: a potential role of Refract Surg 2002; 28: 1355. the angiogenic factor adrenomedullin. Exp Biol 40) FUNATSU H, YAMASHITA H, IKEDA T, M IMURA T, S HIMIZU Med 2006; 231; 1101-1105. E, HORI S. Relation of diabetic macular edema to 52) MASUZAWA K, JESMIN S, MAEDA S, ZAEDI S, SHIMOJO N, cytokines and posterior vitreous detachment. Am MIYAUCHI T, G OTO K. Effect of endothelin dual re- J Ophthalmol 2003; 135: 321-327. ceptor antagonist on VEGF levels in streptozo- 41) FUNATSU H, YAMASHITA H, IKEDA T, M IMURA T, E GUCHI tocin-induced diabetic rat retina. Exp Biol Med S, HORI S. Vitreous levels of interleukin-6 and vas- 2006; 231; 1090-1094. cular endothelial growth factor are related to dia- 53) AIELLO LP. Perspectives on diabetic retinopathy. betic macular edema. Ophthalmology 2003; New Engl J Med 2005; 353: 839-841. 110:1690-1696. 54) GARGIULO P, G IUSTI C, PIETROBONO D, LA TORRE D, DI- 42) FUNATSU H, YAMASHITA H, NOMA H, MIMURA T, N AKA- ACONO D, TAMBURRANO G. Diabetes mellitus and MURA S, SAKATA K, HORI S. Aqueous humor levels of retinopathy. Dig Liver Dis 2004; 36: S101-S105. cytokines are related to vitreous levels and pro- gression of diabetic retinopathy in diabetic pa- 55) GIUSTI C. Novel diagnostic and therapeutic ap- tients. Graefes Arch Clin Exp Ophthalmol 2005; proaches to the diabetic retinopathy. Eur Rev 243: 3-8. Med Pharmacol Sci 2001; 5: 155-166.

163