Mitochondria Damage in the Pathogenesis of Diabetic Retinopathy and in the Metabolic Memory Associated with its Continued Progression

Abstract

Diabetic retinopathy is the leading cause of blindness in young adults, and with the incidence of diabetes increasing at a frightening rate, retinopathy is estimated to threaten vision for almost 51 million patients worldwide. In diabetes, mitochondria structure, function and DNA (mtDNA) are damaged in the retina and its vasculature, and the mtDNA repair machinery and biogenesis are compromised. Proteins encoded by mtDNA become subnormal contributing to dysfunctional electron transport system, and the transport of proteins that are important in mtDNA biogenesis and function, but are encoded by nuclear DNA, is impaired. These diabetes-induced abnormalities in mitochondria continue even when hyperglycemic insult is terminated, and are implicated in the metabolic memory phenomenon associated with the continued progression of diabetic retinopathy. Diabetes also facilitates epigenetic modifications-the changes in histones and DNA methylation in response to cells changing environmental stimuli, which the cell can memorize and pass to the next generation. Epigenetic modifications contribute to the mitochondria damage, and are postulated in the development of diabetic retinopathy, and also to the metabolic memory phenomenon. Thus, strategies targeting mitochondria homeostasis and/or enzymes important for histone and DNA methylation could serve as potential therapies to halt the development and progression of diabetic retinopathy.

DIABETIC RETINOPATHY

Diabetes has become an epidemic of the twentieth century, and with this comes an increased risk of long-term complications, including cardiomyopathy, nephropathy and retinopathy. Control of blood sugar in patients reduces the risk of development and worsening of diabetic complications, but the maintenance of tight glycemic control for the entire life for most patients, however, is a daunting task which requires modification of behavior and extensive family support [1–5]. This leaves patient with this life-long disease to strive for the best possible, sensible glycemic control, and likelihood of developing complications.

Diabetic retinopathy, one of the major microvascular complications of diabetes, is considered as the most acquired ocular disease affecting individuals in their productive years [5]. It is the leading cause of blindness in young adult, and is responsible for 1.8 million of the 37 million cases of blindness throughout the world. People with diabetes are 25 times more likely to become blind than the general population, and approximately 33% of patients with diabetes have some signs of retinopathy. Recent estimates suggest that ~100 million people have diabetic retinopathy worldwide [6], but with the incidence of diabetes increasing at an alarming rate, the projections for 2030 are startling-almost 155 million patients with retinopathy, among those over 51 million with vision-threatening retinopathy [7, 8]. The pivotal, Diabetes Control and Complications Trial (DCCT), which compared the effects of conventional glucose control versus intensive control on diabetic complications, clearly demonstrated that the maintenance of tight glycemic control reduces/delays the risk of diabetic complications, including retinopathy [2, 9, 10]. Even after decades of cutting edge research, tight glycemic control remains one of the most viable options in preventing/slowing diabetic retinopathy, but this tight glycemic control for most of the patients is difficult to maintain for long durations, and in addition to requiring modification of behavior, it could increase the risk of severe hypoglycemia. This raises the importance of adjunct therapies for diabetic patients to halt/slow these devastating complications.

Hyperglycemia remains the major factor implicated in the development and progression of diabetic retinopathy [3, 4, 9]. Another important factor that influences its development is the duration of diabetes; the Wisconsin epidemiologic study of diabetic retinopathy revealed that 98% individuals with type 1 and 78% of individuals with type 2 diabetes developed diabetic retinopathy within 15 years of diagnosis [4, 11]. In addition, other risk factors are hypertension and hyperlipidemia, and although with improved management of overall hyperglycemia, hypertension and hyperlipidemia over the few decades has resulted in a reduction in the development and the progression of diabetic retinopathy [12, 13], it still remains one of the most devastating diseases that a diabetic patient fears the most.

Diabetic retinopathy, as stated above, is a slow progressing, duration-dependent disease. In the early stage of the disease, vascular permeability is increased and microanurysums start to appear in the retina, but as the disease progresses, blood vessels are blocked. In more advanced stages, these new fragile blood vessels start to form along the retina and on posterior surface of the vitreous, and this could result in the retinal detachment [3]. Animal studies have demonstrated that retinal cells, including Muller and glial cells, pericytes and endothelial cells undergo accelerated apoptosis before histopathology characteristic of diabetic retinopathy can be seen and retinal neurodegeneration is an early event in diabetic retinopathy [3, 14–20]. Taken together, these studies have suggested that the clinically silent initial phase of diabetic retinopathy consists of irreversible cellular events with late structural consequences [16].

MITOCHONDRIA AND DIABETIC RETINOPATHY

Circulating high glucose initiates dysmetabolism in the retina and its capillary cells, and as previously reviewed, several molecular and biochemical mechanisms are implicated in the development of diabetic retinopathy, including polyol pathway, advanced glycation end-products (AGEs), protein kinase C (PKC) activation, oxidative stress, inflammation [3, 19, 21–23]. Among these, oxidative stress is considered as one of the critical metabolic abnormalities implicating the development of diabetes retinopathy [21, 24–26]. Mitochondria, the powerhouse of the cell, use approximately 90% of the consumed O₂, but their electron transport chain (ETC) system also generates superoxide as by-product making mitochondria vulnerable to the damaging effect of free radicals [27, 28]. These toxic radicals can induce oxidative modification of cellular molecules activating several pathways leading to cell death, and free radicals have been implicated in a number of chronic diseases such as cancer, diabetes, Parkinson’s, Alzheimer’s [29, 30]. In contrast, research has also shown the importance of free radicals in many cellular signaling pathways triggering complex cellular events [31]. Thus, free radicals are not always deleterious, but can be important for cell survival as well, and how they influence the development of diabetic retinopathy is being investigated in many laboratories.

In the development of diabetic retinopathy, mitochondria become dysfunctional in the retina and its capillary cells, mitochondrial morphology and membrane potential are altered, superoxide radicals are elevated and oxygen consumption is impaired [32–35]. Superoxide radicals act as causal links between elevated glucose and the major abnormalities associated with the vascular complications of diabetes [25]. We have shown that mitochondrial dysfunction has a significant role in accelerating the apoptosis of retinal vascular cells, a phenomenon that precedes the development of histopathology characteristic of diabetic retinopathy [16, 33]. The importance of mitochondria in the pathogenesis of diabetic retinopathy is further confirmed by our studies showing that the regulation of mitochondrial superoxide by maintaining mitochondrial superoxide scavenging enzyme, MnSOD, protects capillary cell apoptosis and the development of diabetic retinopathy in mice [34].

Mitochondria DNA

Superoxide radicals, in addition to damaging protein and lipids, can also damage DNA by combining with guanine and forming oxidatively modified guanine bases (8-OHdG) resulting in irreversible genotoxic alterations [36]. Mitochondria are equipped with their own DNA (mtDNA), and this circular 16.5kb DNA encodes only 13 proteins, all from the ETC system and essential for normal mitochondrial function. Unlike nuclear DNA, which has introns and is packaged into nucleosomes, mtDNA does not have histones and it is packed as nucleoid-like structures [37–39]. Due to the close proximity of mtDNA to the ROS-generating ETC system and lack of protective histones, mtDNA is particularly prone to oxidative damage [40]. Oxidative damage to mtDNA initiates a vicious cycle leading to decreased transcription and protein synthesis, and this is further exacerbated due to subsequent decreased electron transport and increased ROS [41–43]. Diabetes is shown to damage mtDNA in the retina and its capillary cells, and the levels of 8-OHdG are increased. The damaged mtDNA, by compromising the transcription of mtDNA-encoded proteins important in ETC system, fuels into a vicious cycle of free radicals [19, 43]. In addition to damage to mtDNA, diabetes also compromises the DNA repair machinery in the retina [44, 45], further contributing to mitochondrial dysfunction.

Although mitochondria have their own DNA, they are under dual genetic control of both nuclear DNA and the mitochondrial genome. Biogenesis is critical for normal cellular function, and mtDNA biogenesis is tightly regulated by nucleus-mitochondria signaling pathway. Peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC1), a nuclear DNA-encoded protein, initiates mtDNA biogenesis and activation of PGC1 targets nuclear respiratory factor, NRF-1, a factor which regulates the expression of mitochondrial transcription factor A (TFAM) and cytochrome oxidase IV, Cox IV [46–49]. TFAM is a key activator of mitochondrial transcription, and it plays a central role in the expression, maintenance and organization of the mitochondrial genome. In addition to its transcriptional activities, TFAM also organizes the mitochondrial genome into nucleoids and has a responsibility for stability and maintenance of mtDNA, and it is encoded by nuclear DNA and translocated to the mitochondria to initiate mtDNA transcription and replication [48, 49]. In the pathogenesis of diabetic retinopathy, mitochondria copy number and mass is decreased in the retina, and mitochondrial functional integrity is compromised. However, despite decreased copy number, gene transcripts of PGC1, NRF1 and TFAM are increased, but the transport of TFAM from nucleus to the mitochondria is compromised. In addition, mitochondrial membrane transport system is impaired, and the binding of TFAM with the membrane transport proteins is decreased. To add insult to the misery, the binding of TFAM at the promoter region of the mtDNA and at other non-specific regions is also attenuated resulting in subnormal transcription and compromised stability [43, 46, 47, 50, 51].

Mitochondrial DNA contains a non-coding region, the displacement loop (D-loop). The D-loop has essential transcription and replication elements; it serves as a promoter for both the heavy and light strands to facilitate mtDNA replication, and provides control sites for mtDNA transcription [52, 53]. Mutations in the D-Loop regions are associated with various chronic diseases including cancers [54, 55]. For replication of mtDNA, DNA polymerase gamma (with two subunits-POLG1 and POLG2), and mtDNA helicase (Twinkle) are the two important enzymes. These nuclear encoded proteins form the minimal replisome which is required for mtDNA replication [39, 56, 57]. In diabetes, D-loop region encounters more damage than other regions of the retinal mtDNA, and the copy number of the D-loop region is decreased, possibly due to hypermethylation of the CpG sites at the regulatory region of POLG, affecting its binding to the mtDNA. This is also accompanied by suboptimal repair/replication process with decreased levels of POLG1, POLG2 and Twinkle in the mitochondria, the site of their action. Damage of the D-loop region and the replication of mtDNA are under the control of mitochondrial superoxide as mice overexpressing MnSOD are protected from such damage [50, 51]. Thus, there is a wealth of evidence suggesting important role of mitochondria in the development of diabetic retinopathy.

METABOLIC MEMORY

Convincing DCCT studies had shown beneficial effects of tight glycemic control on diabetic complications [2]. At the end of DCCT in 1993, the patients were followed in a long-term observational study, the Epidemiology of Diabetes Interventions and Complications (EDIC), to assess the incidence and worsening of diabetic retinopathy and other complications. Despite similar hemoglobin A1C levels during the first four years of EDIC, the ongoing risk of retinopathy significantly reduced in the former intensive group patients compared with the conventional group patients, and in contrast, it remained unchanged in the former conventional therapy group patients, suggesting a ‘metabolic memory’ phenomenon [58]. These differences between former conventional and the intensive groups have now been reported even after ten years of EDIC study suggesting that metabolic memories are stored early in the course of diabetes, and glycemic control initiated prior to the onset of overt pathology has the most profound long-term impact [59, 60]. The metabolic memory phenomenon was also confirmed in type 2 diabetic patients, the United Kingdom Prospective Diabetes Study showed that patients treated intensively for a period had a lower risk of diabetic retinopathy during their subsequent follow-up, despite the fact that these patients had relaxed back to glycated hemoglobin levels that were not different from those in conventional therapy group [61].

Metabolic memory phenomenon in the continued progression of diabetic retinopathy is duplicated in animal models [46, 62–69]. Dog model of diabetic retinopathy has demonstrated that re-institution of tight glycemic control for 2.5 years after 2.5 years of poor glycemic control does not arrest the progression of diabetic retinopathy [62, 70]. In rats, intervention with islet transplantation after several months of diabetes arrests the progression of retinopathy less effectively than if the intervention occurs after only a few weeks of diabetes [63]. Re-institution of tight glycemic control after a period of poor glycemic control in rats does not protect the retina from increased oxidative stress, mitochondria dysfunction and histone modifications, and the capillary cells continue to undergo apoptosis [64–66]. These studies have suggested that some submicroscopic processes already begun in retina during the initial period of high circulating glucose, and chronic elevation of blood glucose results in metabolic or physiological abnormalities that are not readily corrected by reestablishment of normal glycemia.

In addition to in vivo models, in vitro models have also documented metabolic memory phenomenon. A pivotal study published over two decades ago showed that when endothelial cells are transferred to normal glucose after prolonged exposure to high glucose, overexpression of fibronectin and collagen IV remained detectable, possibly via introducing self-perpetuating changes in gene expression [71]. Glucose-induced alterations in the parameters of oxidative stress, inflammatory mediators, and mitochondria dysfunction and DNA damage do not benefit from the normal glucose exposure that has followed a period of high glucose [46, 47, 51, 67, 68, 72–75]. This model is opening up avenues to understand the molecular mechanism of the metabolic memory phenomenon.

MOLECULAR MECHANISM(S) OF METABOLIC MEMORY

As with the development of diabetic retinopathy, many potential molecular mechanisms have been suggested in the metabolic phenomenon. Increased level of AGEs, that the diabetic environment favors, is considered as one of the plausible factor, possibly because AGEs can cross-link proteins to induce stable molecular changes. This is supported by a report showing that subsequent progression of diabetic retinopathy in patients enrolled in EDIC is correlated with their skin levels of glycated collagen and carboxymethyllysine [76]. Increase in inflammatory mediators and oxidative stress are also some of the metabolic abnormalities associated with the metabolic memory phenomenon [64, 65, 72, 77]. Increased production of ROS is considered as a key effector in diabetic complications as they can modulate several biological pathways including, polyol pathway flux, AGEs formation and activation of PKC, and these metabolic abnormalities can further produce ROS [25], thus making them a plausible targets for metabolic memory.

Reinstitution of tight glycemic control for six months, after six months of poor glycemic control in streptozotocin-induced diabetic rats, does not provide any benefit to the retinal oxidative stress and the capillary cell apoptosis with activated caspase-3 and retinal histopathology continues to progress [64]. Inhibition of glyceraldehyde dehydrogenase (GAPDH), a classic glycolytic enzyme, implicated in diverse cytoplasmic, membrane and nuclear activities [78], is considered to activate major pathways of endothelial cell damage including activation of PKC, hexosamine pathway flux, and AGEs formation [79], and this enzyme continues to be inhibited in the retina, and its downstream and upstream signaling pathways remain altered even after poor glycemic insult is removed [45, 77, 80]. Furthermore, 23 retinal encoding apoptosis genes, and nuclear transcriptional factor kB (NF-kB) continue to be up-regulated in the retina, suggesting that the signaling cascade does not turn off by the reinstitution of normal glycemia [77]. However, if normal glycemic control for six months is initiated just after two months of poor glycemic control, retinal oxidative stress and activation of caspase-3 and NF-kB are partially reversed. In contrast, if the tight glycemic control is initiated soon after the induction of diabetes in rats, these biochemical abnormalities, capillary cell apoptosis and the development of diabetic retinopathy can be prevented [66, 77]. These results have strongly suggested the importance of duration of tight glycemic control.

Metabolic memory is somewhat a ‘stubborn’ phenomenon that persists even after removal of the insult suggesting that that the changes associated with this phenomenon should be long-term with very slow, or even no, turnover. However, the release of free radicals is a quick response to an insult. As stated above, in the pathogenesis of diabetic retinopathy, retinal mitochondria become dysfunctional, their membrane potentials are impaired, and complex III activity is inhibited. As complex I and complex III are the major sources of superoxide production in the mitochondria, and complex III transfers electrons from reduced ubiquinone to cytochrome c, subnormal activity of complex III in diabetes helps further in the accumulation of free radicals [34]. Damaged mtDNA and compromised repair system results in attenuation of the transcripts of mtDNA-encoded proteins that are critical in the integrity of ETC system, and the biogenesis of mtDNA is compromised [44–46, 51]. Decreased synthesis of mtDNA-encoded subunits impairs the ETC system further augmenting the generation of superoxide. This vicious cycle continues to self-propagate and does not turn off even after the hyperglycemic insult is terminated, and the retinopathy continues to progress. In addition, in the early stages of diabetes, to overcome increased ROS, genes important in mtDNA biogenesis and repair machinery are upregulated, but with sustained insult, this mechanism is overwhelmed, and mtDNA and ETC are damaged [43]. The compromised ETC keeps on propagating a vicious cycle of ROS and the dysfunctional mitochondria fuels loss of capillary cells by initiating their apoptosis [44, 45]. Thus, in the early stages, before mtDNA is damaged, there appears to be a window where the cell has an opportunity to protect the mitochondria, but once they are damaged, tight glycemic control fails to provide any benefit. The plausible reason could be that because of the damaged mtDNA, the ETC system becomes overwhelmed and superoxide radicals continue to accumulate. Thus there is compelling evidence suggesting a major role of mitochondria in the metabolic memory phenomenon associated with the continued progression of diabetic retinopathy (Fig. 1).

Fig. (1).

Diabetes increases oxidative stress in the retina and its capillary cells and the levels of reactive oxygen species (ROS) are elevated. ROS damage mitochondria structure, function, electron transport (ETC) system and DNA. Due to damaged mtDNA, the transcription of genes encoded by mtDNA become subnormal, and the biogenesis of mtDNA is attenuated. The ETC system is compromised and the mitochondria copy number is decreased. Suboptimal ETC system further fuels into a vicious cycle of free radicals, and the retinopathy continues even after termination of the hyperglycemic insult.

Role of Epigenetic Modifications in the Metabolic Memory Phenomenon

Epigenetic changes, by modulating chromatin access to the cellular machinery for transcription, regulate gene expression and allow the cells to respond to changing environmental stimuli [81, 82]. The cell “memorizes” these challenges, and epigenetic modifications have potential to persist across generations [83]. Nutrition affects the severity of diabetic complications, and recent studies have shown that it can also bring out epigenetic modifications altering the transcription of genes [84, 85]. Diabetic environment facilitates epigenetic changes in the DNA including histone modifications and these modifications do not terminate even after normal glycemia is re-instituted suggesting their role in the metabolic memory phenomenon associated with the continued progression of microvascular complications of diabetes including diabetic retinopathy [67, 68, 86–89]. Glucose-induced persistent transcriptional activation of p65 in vascular cells even after termination of high glucose insult is associated with methylation of lysine 4 of histone 3 (H3K4) and hypomethylation of lysine 9 of histone 3 (H3K9) [86, 87]. The global acetylation of retinal histones, a process generally associated with gene activation, is increased in diabetes, and reversal of hyperglycemia does not provide any benefit, and Sod2 continues to be epigenetically modified with increased H4K20me3, decreased H3K4me1 and H3K4me2 [67, 68, 74].

Another major epigenetic modification is the methylation of DNA; methylation of the genomic region rich in cytosine and guanine, the CpG island, changes the appearance and structure of DNA without changing its sequence, and the methylation process is carried out by DNA methyltransferases (Dnmts) [90]. Protein-DNA interactions are altered, and due to alterations in the recognition sites involving cytosine or CpG, the binding of transcriptional machinery is impaired [91].

Using Zebra fish as a model of metabolic memory, hyperglycemia has been shown to induce global DNA hypomethylation and aberrant gene expression, and these changes are inherited by daughter cells even after reestablishment of a euglycemic state, suggesting the role of DNA methylation in the metabolic memory phenomenon [92]. In the retina, diabetes is shown to activate Dnmts and CpG islands at the regulatory region of POLG are hypermethylated decreasing its binding with the D-loop region of mtDNA. Reversal of hyperglycemic insult does not provide any benefit to these abnormalities, and DNA continues to be hypermethylated suggesting the role of DNA methylation in the metabolic memory phenomenon [50, 51]. These studies have clearly suggested the importance of epigenetic modification in the could identify novel potential therapeutic targets to prevent/slow down diabetic retinopathy.

THERAPEUTICS TO REGULATE MITOCHONDRIAL DAMAGE IN DIABETIC RETINOPATHY

The evidence compiled herein suggests that the selective targeting of specific molecules to mitochondria could serve as an effective strategy for maintaining mitochondria homeostasis and ameliorating the development of diabetic retinopathy and the metabolic memory associated with its continued progression. However, the complexity of mitochondria structure and function makes them challenging targets, and maintenance of their homeostasis becomes difficult. Mitochondria can be targeted via various routes, e.g., detoxifying the dangerous superoxide radicals generated by the mitochondria using scavengers, maintaining the carriers utilized by various proteins to move into the mitochondria and regulating mitochondrial protein transcription and biogenesis [93, 94]. Although the potential use of mitochondria-targeted therapies in the regulation of diabetic retinopathy remains unclear, experimental studies have shown that targeting of the mitochondrial superoxide scavenging by overexpression of MnSOD prevents the development of retinopathy in diabetic mice [34]. Retinal mitochondria of these diabetic mice also have normal function and mtDNA biogenesis [47]. Furthermore, administration of lipoic acid or benfotiamine, the compounds known to act as antioxidants and protect mitochondria damage, is shown to prevent the development of diabetic retinopathy in rodent models [95, 96]. Lipoic acid, in addition to ameliorating accelerated retinal capillary cell apoptosis in diabetic rats, also prevents abnormalities in mtDNA biogenesis [46]. As mentioned in the previous section, hypertension is also one of the factors in the pathogenesis of diabetic retinopathy. Using spontaneously hypertensive rats Silva and associates have shown that the co-presence of diabetes and hypertension increases retinal neurodegeneration and oxidative stress, and also augments mitochondrial dysfunction that can be blocked by angiotensin receptor blocker [13]. This raises the possibility of the use of angiotensin receptor blockers to protect retinal mitochondria dysfunction and inhibit the development/progression of this blinding disease faced by diabetic patients.

In addition to the development of diabetic retinopathy, mitochondrial dysfunction appears to have a significant role in the metabolic memory phenomenon associated with the continued progression of diabetic retinopathy. Supplementation with lipoic acid during the good glycemic control period in rats, which has followed poor glycemic control, provides significant benefit to the impaired mitochondria biogenesis, and also on the continued progression of diabetic retinopathy [46]. Consistent with this, inclusion of a superoxide dismutase mimetic during the normal glucose exposure of retinal endothelial cells, which has followed high glucose exposure, helps prevent dysregulated mitochondria biogenesis [46]. These results have clearly suggested that supplementation with a therapy targeting mitochondrial damage during the normal glycemic phase which has followed a period of poor glycemia has potential to better ameliorate the progression of diabetic retinopathy than the normal glycemia alone.

Current evidence clearly shows that diabetes-mediated epigenetic modifications are also important in mitochondrial damage. In diabetes retinal Sod2 is epigenetically regulated with altered histone methylation, and lysine specific demethylase 1 (LSD1), the enzyme responsible for regulating methylation, is activated and its binding with Sod2 is increased. In addition to the possible role of these abnormalities in the development of diabetic retinopathy, experimental data have clearly implicated their role in the metabolic memory phenomenon associated with its continued progression. This clearly suggests that targeting enzymes important for histone methylation could serve as a potential therapy to halt the development of diabetic retinopathy. Fortunately, LSD1 inhibitors have shown promising results for other chronic diseases such as carcinoma, leukemia and hypertension [97–99], and evidence reviewed here paves a path for their possible use for the treatment of diabetic retinopathy.