Dysregulation of the NLRP3 Inflammasome in Diabetic Retinopathy and Potential Therapeutic Targets

ABSTRACT

Diabetic Retinopathy (DR) is an insidious neurovascular disorder secondary to chronic glycemic dysregulation in elderly diabetic patients. In the later stages of DR, the disease manifests as fluid infiltrating the macula, culminating in the leading cause of irreversible visual impairment in working age adults. With the current mainstay treatments preoccupied with slowing down the progression of DR, this presents an unsustainable solution from both an economic and quality of life perspective. Although the exact mechanisms by which hyperglycemia leads to retinal tissue insult are unknown, the evidence suggests that chronic low-grade inflammation in diabetic eye is in part driving the constellation of symptoms present in DR. Of the innate immune system within the eye, the NLR Family Pyrin Domain Containing 3 Inflammasome (NLRP3) has been identified in retinal cells as a causal factor in the pathogenesis of DR. Multiple pathways appear to be present in the diabetic eye that instigate prolonged activation of the NLRP3 which subsequently exerts its deleterious effects by upregulating the release of Interleukin-1Beta and Interleukin-18. In this review, we highlight the current understanding of the pathophysiology of DR, the dysregulation of the NLRP3 secondary to hyperglycemic stress in retinal cells, and novel therapeutic targets to alleviate overactivation of the inflammasome.

KEYWORDS:

• Diabetic Retinopathy
• inflammasome
• NLRP3
• inflammation
• Interleukin-18
• Interleukin-1Beta

Background

Introduction

Secondary to a vast array of economic and cultural developments, certain health complications are becoming increasingly prevalent, namely diabetes. It is the complications from the prolonged hyperglycemia (HG) that leads to a deterioration in an individual’s quality of life. Currently, 2.6% of global blindness and the leading cause of blindness in working age adults can be directly attributed to Diabetic Retinopathy (DR).¹ DR is a neurovascular condition that can be subclassified based on the presence of certain anatomical hallmarks. The early stage of the disease is termed Non-Proliferative Diabetic Retinopathy (NPDR), characterized by the presence of intra-retinal hemorrhages, microaneurysms, and cotton wool spots from infarcted nerve tissue.^2,3 NPDR has the potential to advance to a more severe stage of the disease called Proliferative Diabetic Retinopathy (PDR) which is distinguished by the presence of pathological retinal neovascularization.⁴ The novel vessels often lack the conventional integrity and if left untreated, may lead to hemorrhages within the vitreous and precipitate tractional retinal detachment.^5,6 At any point throughout the disease, an individual is also prone to developing Diabetic Macular Edema (DME) which is the leading cause of visual impairment from DR.⁷ DME manifests itself as retinal thickening due to fluid accumulation in the macula – the most light sensitive area of the retina consisting of densely packed photoreceptors.⁸ The built-up fluid within the neuronal tissue leads to the distortion and loss of high definition central vision.

Figure 1. Activation and complications of the NLRP3 Inflammasome in Diabetic Retinopathy.

Description: Figure 1 summarizes the upstream activators of the NLRP3 Inflammasome in Diabetic Retinopathy with a description of the subsequent complications with prolonged NLRP3 dysregulation.

Pathophysiology

Although the pathophysiology of DR is not entirely known, the sequence of certain fundamental complications is known. Initially, due to sustained HG, the capillaries in the retina begin to deteriorate. The pericytes composing the capillary wall begin to drop out and leukocytes infiltrate, occlude, and attach to damaged vascular endothelial cells with the help of various adhesion molecules.^9,10 This vascular degeneration induces hypoxia within the retina as a mismatch between the supply and demand of oxygen is now evident. To restore homeostasis in response to ischemic conditions, angiogenic growth factors are upregulated and secreted.^11,12 Vascular Endothelial Growth Factor A (VEGF) in particular stimulates the formation of new vessels to supply the oxygen-deprived tissue. However, the ensuing vessels tend to have faulty tight junctions amongst endothelial cells and grow into crucially non-perfused regions of the retina.^3,13 The complications of DR such as DME are instigated due to the hyperpermeable vessels and the subsequent breakdown of the blood-retina-barrier (BRB) resulting in fluid extravasation into the retina. With persistent hypoxia and edema, there is eventual apoptosis of retinal neurons and thus visual impairment. Of note, the visual complications from gross microvascular modifications tend to arise in the later stages of DR, despite subtle tissue insult from chronic HG being evident much earlier.¹⁴ This highlights the potential for novel early biomarkers and emphasizes the necessity to understand the chronology of events if we wish to cure, rather than symptomatically address the condition.

Despite glycemic control being the initial treatment for DR as the severity of the condition is initially a function of the degree of HG, glucose management tends to be ineffective in the later stages as chronic HG eventually produces irreversible damage.^2,11 The mainstay treatment for PDR and DME is anti-VEGF antibody injections into the vitreous chamber. Granted the antibodies are more effective than traditional laser retinal ablation therapy, the antibodies have their fair share of downfalls including a hefty price tag, a short half-life entailing frequent injections and consequential poor patient compliance.¹⁵ Not to mention, the antibodies are not particularly effective for all patients. For DME patients, in the DRCR Protocol Study 1, approximately 40% still had persistent edema after 6 months of treatment.² In fact, considering that VEGF is required for homeostatic maintenance of retinal vessels, prolonged suppression may inadvertently reduce choroidal thickness.¹⁶

For refractory cases of DME and PDR not responding to anti-VEGF antibodies, the previous gold standard, pan-retinal-photocoagulation is still employed sparsely. While the exact mechanism of action is unknown, it is thought that the lasers directly close microaneurysms, stimulate the retinal pigment epithelium (RPE), and ablation of peripheral neural tissue improves oxygenation of remaining tissue.^17,18 Unfortunately, the potential side effects include reduced night vision capacity due to destruction of the peripheral retina – the primary site of rod photoreceptors.

Additional treatments for refractory NPDR and DME include intraocular injections of anti-inflammatory corticosteroids; however, as a limited second-line therapy given the risk profile from prolonged usage.¹⁹ Risks include an increased incidence of cataracts, elevated intraocular pressure, vitreous hemorrhages, and glaucoma.^20,21 It is interesting to note that the success of anti-inflammatory therapy suggests that an inflammatory process is at least partially propelling DR. Given the increasing prevalence of diabetes, rising healthcare costs, and therapeutic efforts aimed at symptomatic management of DR, the current strategy does not represent a sustainable nor effective long-term action plan.

Inflammation in the Eye

Chronic low-grade inflammation has been implicated to be an ongoing process throughout the various stages of DR. Of particular importance is the inflammasome, a multiprotein molecular platform that modulates the innate immune system by regulating the processing and secretion of pro-inflammatory cytokines.^22,23 The inflammasome is assembled in response to a variety of pathogen- and danger-associated molecular patterns (PAMP and DAMP).^24,25 NOD-like receptors (NLRs), which are intracellular microbial and danger sensing proteins,²⁶ have a central nucleotide-binding and oligomerization (NACHT) domain, an effector domain at their amino terminus, and multiple leucine-rich repeats (LRRs) domain at their carboxyl terminus.²⁷ Amino terminus domains include either acidic transactivation domain, baculoviral inhibitor repeat-like domain, caspase activation and recruitment (CARD) domain, or the pyrin domain.²⁸ Upon detection of a PAMP or DAMP, the LRRs auto-phosphorylate and recruit ASC adaptor protein which in return recruits procaspase 1. The complex of NACHT, LRRs, apoptosis speck-like protein (ASC), and Procaspase 1 is collectively known as the inflammasome.²⁷ The CARD domain of ASC cleaves procaspase 1 into caspase 1 which is then able to catalyze the cleavage of pro-Interleukin-1Beta and pro-Interleukin-18 into active Interleukin-1Beta (IL-1β) and Interleukin-18 (IL-18), respectively.²⁷ Of the various inflammasomes, the NLR Family Pyrin Domain Containing 3 Inflammasome (referred to as NLRP3) is the most abundantly studied in relation to pathological conditions of the eye. When discussing the activation of the NLRP3, it is in fact a two-step process. The initial extracellular priming signal induced by the toll-like receptor (TLR)/nuclear factor (NF)-kB pathway upregulates cellular transcription of the NLRP3 Inflammasome component proteins.^29,30 The secondary activation signal – transduction of a PAMP or DAMP – promotes the assembly of the component proteins.^29,31 Common secondary signals include changes in potassium ion flux, reactive oxygen species (ROS), lysosomal destabilization, and certain post-translational protein modifications.²²

Evidence exists for chronic inflammation and the upregulated activity of the NLRP3 contributing to the pathogenesis of diabetes.^2,32–36 For example, in type 1 diabetes, the pancreatic expression of NLRP3 dependent IL-1β contributes to the autoimmune destruction of beta cells.^32,37 In type 2 diabetes, Lee et al. provide compelling evidence for the activation of the NLRP3 in the pathogenesis of DR. Macrophages derived from type 2 diabetics exhibited significantly increased expression of the NLRP3 at both the mRNA and protein level and unsurprisingly, an increased tendency to process downstream inflammatory cytokines IL-1β and IL-18 when exposed to DAMPs.³³ While IL-1β plays a similar role of pancreatic destruction in type 2 diabetics,³⁴ further evidence for inflammation is provided by the diabetes drug metformin alleviating both diabetic clinical findings and concurrently attenuating NLRP3 and IL-1β activity.³³

When considering DR in particular, Loukovaara et al. provide conclusive evidence of NLRP3 overactivity stemming from a study based on vitreous donations from NPDR and PDR patients. The authors demonstrate the presence of elevated inflammasome component proteins and their downstream inflammatory cytokines IL-1β and IL-18 in the vitreous of DR patients.³⁸ Interestingly, the measured levels of NLRP3 activity clearly correlated with disease severity. Patients with PDR exhibited significantly higher levels of NLRP3 proteins and IL-18 activity than patients with NPDR.³⁸ Within PDR, the presence of TRD exhibited similar results, as TRD patients tended to have higher levels of NLRP3 activity when compared to those with a fully intact retina.³⁸ In a study conducted by Chen et al., vitreous and fibrovascular membrane samples collected from NPDR and PDR patients feature congruent results at both the protein and mRNA level for the NLRP3 activity when compared to normal controls.³⁹ Similarly, the prognostic implications of NLRP3 activity are supported by Chen et al³⁹ who also demonstrate significantly higher expression of IL-1β and IL-18 in patients with PDR than those with NPDR. Moreover, in a study with PDR patients, Song et al. noted a correlation between IL-18 levels and PDR severity.³⁹ However, the study by Chen et al. presents conflicting results with Loukovaara et al. who suggest no difference in IL-1β between PDR and NPDR patients, only with IL-18. Considering the infancy of the research in NLRP3 activity and the presence of confounding variables such as diabetes type and temporal relations, more work is clearly required to elucidate accurate results. With the information at hand, it is possible that IL-18 provides a robust prognostic indicator. Loukovaara et al. suggest that IL-18 increases early on in the disease process and its measured levels correlate with disease severity as indicated by all three of Loukovaara et al., Chen et al., and Song et al.^38–40

Although the exact role of a dysregulated NLRP3 in DR is not entirely known, current attention to this area has produced numerous promising studies. For instance, human retinal microvascular endothelial cells (HRMEC) exposed to HG demonstrate a statistically significant spike in both the protein and mRNA of the NLRP3.³⁶ Similar results are evident for the NLRP3’s direct downstream inflammatory cytokines IL-1β and IL-18 and secondary mediators such as VEGF and Tumor Necrosis Factor Alpha.³⁶ In order to model type 1 diabetes, mice are often injected with streptozotocin (STZ) to destroy pancreatic beta cells. In such models with simultaneous blockade of the NLRP3 via shRNA, the models tend to exhibit reduced vascular permeability, pro-apoptotic caspase 3, caspase 1, IL-1β, IL-18, and VEGF.³⁶ Within the retina, the NLRP3 appears to be active in the inner and outer nuclear layers, within the resident immune cells, microglia.^11,41,42 Muller cells, the primary retinal glial type, are particularly important considering retinal caspase 1 is primarily located within Muller inner cell processes.⁴³ With HG, retinal Muller cells exhibit a spike in caspase 1, processing of IL-1β and IL-18, and prolonged gliosis.^43,44 While few studies exist exploring the effects of the NLRP3 activity on the RPE in DR patients, extrapolating related studies provides key insights. In a HG environment, RPE cells tend to display elevated NLRP3 and IL-1β activity.⁴⁵ The destruction of the RPE which represents the outer BRB may lead to various complications in the pathogenesis of DR. For instance, DR patient may experience vascular complications, edema, activation of compensatory angiogenic pathways⁴⁶ and eventual pyroptosis and apoptosis of the RPE cells as seen in studies evaluating NLRP3 activity in age-related macular degeneration.⁴⁷

Upstream Activators of the NLRP3

While the exact mechanism by which the NLRP3 becomes dysregulated in DR is not entirely known, it is proposed that prolonged HG increases the prevalence of prominent activating DAMPs, which are discussed below.

Adenosine Triphosphate

Adenosine triphosphate (ATP) is hypothesized to be working at both the NLRP3 priming and activating DAMP signals.^22,48 Extracellular ATP stimulates P2X7 ATP-gated ion channels to allow an efflux of potassium ions^49,50 and thereby stimulating the NLRP3. In particular, the literature focuses on the role of Connexin hemichannels, of which Connexin43 is the most prominent isotype in humans and upregulates early on in DR due to ischemia.⁵¹ Connexin hemichannels are six protein subunits which oligomerize to form a hexamer termed a Connexon or Hemichannel⁵¹ that may restrict the movement of anything larger than 1Kda⁴⁸; implying that ATP can freely pass through. The hemichannels are constitutively expressed in HRMECs, glial, RPE, and astrocytes⁵¹ within the retina; all of which are essential in the maintenance of the BRB.⁴⁸ Under hypoxic or ischemic conditions, the hemichannels open to form membranous pores^,52,53 which facilitates ATP outflow and thus an autocrine activation of the NLRP3 inflammasome.⁴⁸ The results from DR studies blocking the Connexin43 hemichannel by Peptide 5⁴⁸ or Xiflam⁵⁴ indicate a reduction in NLRP3 Inflammasome activity (fewer retinal aneurysms, reduced gliosis), pro-inflammatory cytokines, and VEGF to basal levels.^48,54 Interestingly, the Connexin43 hemichannels may be involved in a viscous positive feedback cycle. The presence of pro-inflammatory cytokines such as IL-1β processed by the NLRP3 trigger the opening,⁴⁸ increase the transcription, and availability of the hemichannels on the cell membrane.⁴⁸ As expected, this leads to an increase in available extracellular ATP for autocrine NLRP3 activation.

Reactive Oxygen Species

Reactive Oxygen Species (ROS) are a known DAMP which instigate the NLRP3.^45,55,56 In the diabetic retina, as the cells attempt to dispose of excess glucose, the upregulated mitochondrial activity is the prominent source of byproduct ROS.^57,58 It is proposed that ROS indirectly prompt the NLRP3 through dysregulation of the anti-oxidant Thioredoxin (TRX) system.⁵⁹ Under normoxidative conditions, the anti-oxidant TRX is negatively regulated by Thioredoxin Interacting Protein (TXNIP) physically creating a disulfide bond at the redox domain of TRX.^60,61 As such, the anti-oxidative interactions of TRX with other molecules are inhibited.^59,60 However, with excessive oxidative stress, ROS eventually oxidize the disulfide bond, liberating TXNIP.^60,61 Unlike bound TXNIP, free TXNIP is able to activate the NLRP3^{29,45,56,59,62} by binding its LRRs.²³

For diabetics, this activation of the NLRP3 by TXNIP may be further compounded as indicated by HG upregulating TXNIP expression in Muller and endothelial cells via genomic modifications.^36,44 For instance, in a HG environment, cells exhibit stimulatory histone acetylation at the histone activation marks H3K9AC at the TXNIP promoter region by P300 Histone Acetyltransferase.^63–65 Likewise, HG in the form of glucose-6-phosphate enhances the translocation of carbohydrate response element-binding protein (ChREBP) into the nucleus which promotes the expression of the carbohydrate response element in the TXNIP promoter region^59,64 consequently enhancing TXNIP transcription which is then able to subsequently activate the NLRP3 in diabetics.

The physiology of ROS stimulated TXNIP activation of the NLRP3 hinges on time-dependent attempts by homeostatic processes to prevent HG-induced mitochondrial dysregulation.^36,44 To dispose of the HG, there is an initial uptick in glucose processing by the cell resulting in increased production of ATP and ROS, which are swiftly neutralized by cellular antioxidants.⁴⁴ As antioxidants are overwhelmed by the persistent HG, ATP production drops as the cell switches to anaerobic glycolysis in an attempt to thwart ROS production, inadvertently generating a hypoxic environment.⁴⁴ Regardless of the cell’s homeostatic attempts, ROS continues to leak from mitochondria and the upregulated TXNIP due to persistent HG bind and inhibit the anti-oxidative capacity of TRX.⁴⁴ To preserve viability, the cell engages in autophagy and mitophagy to discard damaged organelles.^44,66 However, fragmentation of damaged mitochondria contributes continued DAMPs such as additional ROS leakage and mitochondrial DNA.⁶² While efficient autophagy and mitophagy are essential for homeostasis, in the diabetic retina, these cellular processes may be pathologically activated in part by the upregulated TXNIP. This is accomplished by TXNIP sustaining an environment with persistent oxidative stress⁴⁴ and culminating in inappropriate NLRP3 activity, pyroptosis, and thus progression of DR.

Lysosomal Destabilization

Lysosomal damage is a known activator of the NLRP3^,13,67; possibly through the release of Cathepsin B into the cytoplasm following lysosomal rupture.²³ Lysosomal stress can be precipitated by numerous means such as declining autophagy efficiency,⁶⁸ irritants such as uric acid (UA),⁶⁹ and aging.⁷⁰ UA is the byproduct of purine metabolism and at serum levels above 356 (μmol/L), it forms into the NLRP3 detectable DAMP, monosodium urate (MSU) crystals.⁶⁹ According to Thounaojam et al., DR patients tend to have serum UA levels exceeding the nucleation threshold and postmortem DR vitreous samples indicate a significant positive correlation between UA levels and IL-1β activity.⁶⁹ To ascertain the implications of UA on the retina, Thounaojam et al. exposed STZ mice to allopurinol, an inhibitor of UA production, and benzbromarone, which enhances UA excretion.⁶⁹ In conjugation with a HG environment, the mice exposed to both hypouricemic drugs exhibited lower levels of IL-1β, VEGF, and NLRP3 activity. It is hypothesized that MSU crystals lead to NLRP3 activity by either directly causing lysosomal rupture or UA alone is a pro-inflammatory DAMP,^23,69,71,72 possibly through the promotion of TXNIP dissociation from TRX in an ROS dependent manner. The preliminary results provide justification into exploring UA as a target, monitor, and early-stage prognostic biomarker in DR. Beyond UA, since age is the largest risk factor for DR, the deteriorating efficiency in general physiology may lead to lysosomal damage and successive NLRP3 activity. For example, the ability of the RPE to recycle photoreceptor tips declines with age and results in an accumulation of the lipid peroxidation byproduct lipofuscin, a lysosomal destabilizer resulting in enhanced NLRP3 activity.^70,73

Prostaglandin E2

Prostaglandin E2 (PGE2) is a proinflammatory molecule that has been noted to be statistically elevated by up to 53% in the vitreous of DR patients.^72,74,75 Likewise, its parent enzyme cyclooxygenase 2 (COX2) and its target receptor EP2R follow suit.^75,76 PGE2 is of importance because its inhibition has demonstrated an ability to suppress the progression of DR.⁷² This is partly due to its possible role as an upstream NLRP3 activator.⁷⁵ Additionally, in the diabetic retina, PGE2 exhibits a positive correlation with VEGF and promotes leukocyte adhesion, neovascularization, impairment of retinal vessels, retinal edema, and apoptosis.^{72,74,75,77,78} Upon application of PGE2 receptor antagonists, the effects of PGE2 are diminished possibly due to an inhibited association of the NLRP3 with ASC oligomers.⁷⁵

According to Wang et al., an understanding of how HG results in NLRP3 activation as mediated by PGE2 stems from its interaction with the EP2R.⁷⁵ Upon activation by PGE2, the EP2R coupled cAMP/PKA signaling results in PKA entering the cell nucleus and subsequent phosphorylation of the transcriptional factor cAMP response element-binding protein.⁷⁵ The genomic modifications are thought to upregulate transcription and translation of both the NLRP3 and pro-IL-1β. To compound the effects, prolonged HG is suggested to increase both the expression and activation of the COX2 enzyme, which may further provoke the PGE2/EP2R-cAMP/PKA pathway of NLRP3 activation.⁷⁵ In fact, EP2R is not the only receptor implicated in DR. Studies antagonizing the EP4 receptor demonstrate significantly reduced PGE2 induced VEGF expression in Muller cells.^77,78

Downstream Mediators of NLRP3

Interleukin-18

The retinal consequences of the dysregulated NLRP3 are primarily mediated by the proinflammatory cytokines IL-1β and IL-18. The ramifications of IL-18 are not particularly unified throughout the literature. Authors such as Shen et al. detail potential protective effects from IL-18 which are described as anti-angiogenic and anti-permeability⁷⁹; in essence, offsetting VEGF. Their studies exhibit a positive correlation between baseline levels of intraocular IL-18 and visual outcomes with VEGF antibodies, elevated IL-18 upon anti-VEGF treatment, and injections of IL-18 dampening the effects of VEGF such as vascular leakage and neovascularization.⁷⁹ It is proposed that IL-18 preserves the BRB by preventing a reduction in the tight junction protein, Claudin 5; the opposite is typically expected from dysregulated VEGF.⁷⁹ Nonetheless, it is worthwhile to note the previous study was not conducted on DR patients. Conversely, other studies with in vivo retinal injections of IL-18 did not statistically change choroidal neovascularization (CNV) volumes in age-related macular degeneration models.⁸⁰ What’s more, IL-18 may possibly be pro-angiogenic, as genetically deficient mice in IL-18 or its receptor Interleukin-18 Receptor 1 display a reduction in CNV volumes.⁸⁰ As such, the conflicting evidence does not paint a homogenous picture on the effects of IL-18 and epitomizes the need for further exploration.

Interleukin-1Beta

In contrast, the effects of IL-1β are congruent with the expectations from the inflammatory NLRP3. IL-1β and its parent enzyme caspase 1 are both heavily concentrated within Muller glia in the retina and affect neighboring cells in a paracrine manner.^42,43 Upon binding the specific Interleukin-1 Receptor Type 1 (IL-1R1), IL-1β is able to exert the deleterious constellation of effects expected in DR.³⁵ Specifically, IL-1β contributes to retinal vessel permeability,⁶² upregulation of adhesion molecules promoting occlusion of capillaries by leukocytes,^2,43 and apoptosis of endothelial cells.² What is more, the evidence suggests that HG-induced IL-1β participates in a positive feedback loop by increasing the presence of its parent enzyme Caspase 1 in an autocrine manner.³⁵ Lastly, IL-1β contributes to an increase in the presence of pro-apoptotic Semaphorin-3A and caspase 3, suggesting that HG-induced apoptosis may in part be driven by an inflammatory process.^35,42

Disrupting the NLRP3 Inflammasome Activation

Within the diabetic retina, there exist five material targets to avert the induction of the NLRP3. These include 1: blocking cell membrane receptors ex. Connexin43 hemichannels, 2: controlling cytoplasmic secondary messengers ex. ROS, 3: halting transcription ex. ChREBP, 4: preventing conjugation of the NLRP3 component proteins, and lastly, 5: antagonizing the Interleukins or their respective receptors.^70,81 Although preventing the activation of the NLRP3 seems like a plausible therapeutic for DR, it is cautionary to note that complete systemic inhibition would not be ideal. The NLRP3 is involved in a host of homeostatic processes such as the elimination of infectious agents.²⁹ An alternative approach may be to target the reduction of NLRP3 activity only in the vitreous chamber and for controlled periods of time. Throughout the literature, halting the NLRP3 in the diabetic retina distills into either direct upstream inhibitors or antioxidants to alleviate oxidative stress. What is more, a promising niche of gene therapy is also currently being explored which will not be discussed in depth in this paper. With recombinant adeno-associated viral vectors, researchers are able to target IL-1 directly, TXNIP, or the M103 gene, which prevents proteolytic processing of caspase 1.^82–84

Direct and Upstream Inhibitors

The small molecule, MCC950, is an NLRP3 inhibitor that prevents both canonical and non-canonical activation presumably²² by preventing the interaction of the NLRP3 with NEK7; a kinase NLRP3 activator that coordinates the oligomerization of the NLRP3 and ASC.^22,85 For DR, HG exposure studies exhibit MCC950 reducing the levels of IL-1β, IL-18, and cellular apoptosis.^22,86,87 Besides DR, remarkably the MCC950 yields additional aid to diabetics in the form of dampened diabetic encephalopathy and attenuated hippocampal NLRP3 activity manifesting as reduced anxiety and depression.⁸⁸ Lastly, the MCC950 is structurally a sulfonylurea molecule, a traditional diabetes pharmaceutical employed to enhance insulin tolerance and secretion from pancreatic beta cells.^88,89 This dual nature proposes the possibility of a uniquely engineered molecule that may inhibit both the NLRP3 and act as a potent insulin secretagogue. For instance, the conventional diabetes therapeutic, Glyburide, can also halt the induction of the NLRP3.⁶¹

Additional upstream inhibitors include Compound 49B, a beta adrenergic receptor agonist which exerts its effects by reducing TLR4 signaling.⁹⁰ Specifically, it is proposed that beta adrenergic receptors signal through cAMP leading to a spike in exchange protein activated by cAMP 1 (EPAC1) which can subsequently regulate the levels of pro-inflammatory cytokines.⁹¹ Jiang et al. propose that Compound 49B reduces TLR4 signaling which in turn signals via HMGB1 to initiate the NLRP3.⁹² However, this is in stark contrast to both Kim et al. and Van Beijnum JR et al., who propose instead it is HMGB1 signaling via TLR4 to activate the NLRP3.^93,94 Moreover, Resolvin D1 acts as an upstream mediator of the NLRP3 likely by preventing the initial transcription of the component proteins.⁴¹ Lastly, Peptide 7, a Connexin43 hemichannel blockade, prevents HG and cytokine driven ATP efflux and thus, autocrine NLRP3 activation.⁴⁸

Antioxidants

An emerging theme in the inhibition of the NLRP3 is the use of antioxidants to reduce ROS. The average diabetic patient, especially with PDR, tends to be deficient in Vitamin D3,⁹⁵ a multifunctional hormone possessing potent anti-oxidative and anti-inflammatory capabilities.⁹⁶ The evidence implies Vitamin D3 dampens NLRP3 activity by reducing ROS and thus preventing the dissociation of TXNIP from TRX.⁹⁶ In the diabetic retina, Vitamin D3 is able to inhibit the elevation of TXNIP due to ROS at a level comparable to ROS scavengers, reduce VEGF secretion, and thwart the change in density and disorganization of the retina.⁹⁶ Additionally, the application of Vitamin D3 attenuates the apoptosis of various cells exposed to HG likely as byproduct of preventing IL-1β induced mitochondrial dysfunction^96,97 or possibly due to reduced ROS inflicted mitochondrial damage and subsequent homeostatic apoptosis.⁴⁴

Nuclear Erythroid Type 2 Factor (NRF2) is a transcriptional factor that binds the antioxidant response element to regulate the transcription of anti-oxidants such as HO1 and NQO1.^43,98 The antioxidants subsequently reduce ROS and thus, NLRP3 activation.⁹⁸ For DR patients, there exist two molecules that mediate their effects by upregulating NRF2. Firstly, Sulforaphane (SFN), an isothiocyanate found in cruciferous vegetables demonstrates reduced retinal thickness and pro-inflammatory cytokines such IL-1β in a dose-dependent manner.⁹⁸ Similarly, the widely employed triglyceride lowering drug, Fenofibrate, attenuates the generation ROS via stimulating NRF2 signaling.⁴³ This peroxisome proliferator-activated receptor alpha agonist has affirmed its ability to delay the progression of DR in both the ACCORD and FIELD trials, reducing the need for PRP by up to 32% in diabetics.^99,100

Other antioxidants include Methylene Blue, which reduces canonical NLRP3 activation with secondary effects of reducing glycated hemoglobin.¹⁰¹ Methylene Blue is assumed to be either directly inhibiting TXNIP or preventing its association with the NLRP3.¹⁰¹ Lastly, Minocycline, a tetracycline antibiotic, possesses similar anti-oxidative effects conceivably through the reduction of ROS induced TXNIP activation of the NLRP3.³⁶ Additional effects include a reduction in cellular apoptosis in HG environments,³⁵ dampened diabetic retinal microglia activation,¹⁰² and halting of the transcription of NLRP3 proteins.³⁶ Nonetheless, an obvious impediment exists in that systemic application of a tetracycline antibiotic is not warranted as a long-term solution; rather, controlled doses within the vitreous may be of relevance to DR patients.

Conclusion

Diabetic Retinopathy is a neurovascular retinal disease which in it’s late stage is characterized by compromised visual quality secondary macular edema driven by chronic HG insult. With the accelerating global prevalence of diabetes, exploration into the complex pathophysiology of diabetic retinopathy is absolutely necessary in order to promote primary prevention and relieve the costly burden of chronic management. In this review, we present evidence for the dysregulation of the NLRP3 Inflammasome as a contributing factor to the constellation of tissue insults evident in the diabetic retina. As suggested by experimental studies, multiple concurrent instigators such as reactive oxygen species, adenosine triphosphate, lysosomal damage, and prostaglandins may provoke the dysregulated activation of the NLRP3 Inflammasome and it’s damaging pro-inflammatory cytokines IL-1β and IL-18. While this understanding presents various potential checkpoints to control inflammation, it also suggests that a more efficient approach is to focus on the common denominator with direct inhibition of the NLRP3 Inflammasome. Though direct inhibition of the NLRP3 Inflammasome is feasible, we are currently lacking substantial in vivo studies. We, therefore, advocate for a heightened focus in this particular area of inflammation and are hopeful that local inhibition of the NLRP3 Inflammasome may reduce the disease burden of Diabetic Retinopathy.

Abbreviations

BRB	=	Blood-retina-barrier
ChREBP	=	Carbohydrate response element binding
CARDS	=	Caspase activation and recruitment domain
CNV	=	Choroidal neovascularization
COX2	=	Cyclooxygenase 2
DME	=	Diabetic Macular Edema
HRMEC	=	Human retinal microvascular endothelial cells
IL-1	=	Interleukin-1
IL-18	=	Interleukin-18
IL-1β	=	Interleukin-1Beta
IL-1R1	=	Interleukin-1 Receptor Ty[e 1
LRRs	=	Leucine rich repeats
MSU	=	Monosodium urate
NLRP3	=	NLR Family Pyrin Domain Containing 3 Inflammasome
NLRs	=	Nod like receptors
NPDR	=	Non-Proliferative Diabetic Retinopathy
NACHT	=	Nucleotide binding and oligomerization domain
PDR	=	Proliferative Diabetic Retinopathy
PGE2	=	Prostaglandin E2
ROS	=	Reactive oxygen species
RPE	=	Retinal pigment epithelium
STZ	=	Streptozotocin
SFN	=	Sulforaphane
TRX	=	Thioredoxin
TXNIP	=	Thioredoxin Interacting Protein
VEGF	=	Vascular Endothelial Growth Factor A

Author’s contributions

KSR sourced the referenced literature, analyzed the literature, and subsequently wrote the manuscript. JAM was a significant contributor in refining the topic of the review and in reviewing the drafts of the manuscript. All authors read and approved the final manuscript.

Availability of data and material

All data generated or analyzed in this review are publicly available and referenced within the text.

Declaration of interests

The authors declare that they have no competing interests.

Additional information

Funding

This study was funded by CIHR (to JAM) and NSERC (to JAM).