Role of HMGB1 signaling in the inflammatory process in diabetic retinopathy

Abstract

High mobility group box 1 (HMGB1) is a key player in retinal inflammation. HMGB1 is a danger associated protein pattern receptor which can sense high glucose as a stressor. Increased HMGB1 levels have been found in patients with late stage diabetic retinopathy. HMGB1 can bind toll-like receptor 4 (TLR4) and the receptor for advanced glycation end-products (RAGE), leading to increased inflammation commonly through nuclear factor kappa beta (NFkB). Because diabetic patients have been found to have increased HMGB1 and RAGE levels, as well as polymorphisms of TLR4, a number of investigations have focused on inhibition of these pathways in the diabetic retina. Work in diabetic animal models and cell culture have demonstrated a number of factors that can inhibit HMGB1/TLR4/RAGE signaling. This regulation offers potential new avenues for therapeutic development. This review is focused on HMGB1 signaling and downstream pathways leading to inflammation in the diabetic retina.

Introduction:

Diabetic retinopathy remains the leading cause of blindness in working age adults. With rates of obesity ever increasing, there remains a real need to better understand the retinal damage associated with diabetes. In the past decade, the role for retinal inflammation in the diabetic-induced retina damage has become increasingly important ^{[1, 2]}. In addition to specific cytokine actions in the retina, sterile inflammation has also recently been observed to cause retinal damage ^{[3, 4]}. Sterile inflammation is a form of pathogen-free inflammation that can be caused by trauma, ischemia, or other stressors. High glucose due to diabetes can activate a form of sterile inflammation^[5]. High mobility group box 1 (HMGB1) is a key mediator of sterile inflammation in the retina ^{[6, 7]}.

Role of HMGB1 in retinal inflammation.

HMGB1 is secreted by immune cells ^[8]. In response to high glucose or other stressors, HMGB1 will bind toll-like receptor 4 (TLR4) and the receptor for advanced glycation end-products (RAGE) (Figure 1) ^[9]. HMGB1 is a nuclear protein that leads to inflammation once translocated to the cytoplasm. Studies in diabetic patients with proliferative retinal disease showed a significant link of disease severity and increased HMGB1 levels ^{[10, 11]}. Based upon these clinical findings, researchers have investigated the role of HMGB1 in the diabetic retina in rodent models. HMGB1 appears to mediate diabetes-induced damage in retinal pericytes ^[12] and Müller cells ^[13]. Others reported a significant role for HMGB1 in retinal inflammation in retinal ganglion cells (RGC) ^[14] and retinal pigmented epithelial cells (RPE) ^[15]. Using diabetic rats, authors demonstrated that diabetes increased HMGB1 levels, which was reduced when a HMGB1 inhibitor, glycyrrhizin, was used ^[16]. Similarly, another groups showed that diabetes and high glucose culturing conditions in retinal cells significantly increased HMGB1 levels ^[17]. Silencing of HMGB1 in the cells reduced apoptosis and inflammatory mediators ^[17]. Another group reported that HMGB1 caused retinal neuropathy in response to diabetes which was mitigated by glycyrrhizin ^[18]. These findings agree with our recent data showing that glycyrrhizin reduced loss of retinal thickness and loss of RGC numbers due to streptozotocin ^[19].

Figure 1.

Schematic of the signaling of HMGB1 through RAGE/TLR4 to inflammation in diabetes.

In addition to inhibition of HMGB1 by siRNA or glycyrrhizin, increasing sirtuin 1 (SIRT1) levels appears key to blocking HMGB1 actions ^[20]. We found that both protein kinase A (PKA) and exchange protein for cAMP1 (Epac1) could increase SIRT1 to reduce HMGB1 actions in retinal cells ^{[21, 22]}. Another option to reduce HMGB1 actions in the diabetic retina is to block its interaction with TLR4 and RAGE ^[23]. Table 1 provides a list of HMGB1 inhibitors and their link to specific inflammatory diseases.

Table 1.

HMGB1 inhibitors

HMGB1 inhibitors	Inflammatory disease	Reference-PMID
Glycyrrhizin	Diabetic retinopathy	23261684
	Endothelial progenitor cells	26798412
	Diabetic Kidney	29241180
Box A	Retinal ischemia/reperfusion	28542588
Cilostazol	Rheumatoid Arthritis	25126750
Zeaxanthin	Diabetic retinopathy	18385086
Resveratrol	Diabetic CV disease	22768138
	Hypoxic Brain injury	31362164
MnTBAP	Diabetic Oxidative Stress	19833897
Astilbin	Diabetic CV disease	24211745
N-acetyl cysteine (NAC)	Diabetic Kidney disease	25896065
Empagliflozin	Diabetic Kidney disease	31168342
CaMKIV	Diabetic Neuropathic Pain	27216039
Peptide P5779	Sepsis, pulmonary hypertension	Reviewed in Yang et al. 2020 DOI 10.3389
M2G7	Sepsis, arthritis	Reviewed in Yang et al. 2020 DOI 10.3389
Dexmedetomide	Endotoxemia	24803295
Haptoglobin	Anti-inflammatory	30568040
Metformin	LPS-induced inflammation	31772144

TLR4 actions in the diabetic retina.

TLR4 is a transmembrane protein that serves as a pattern recognition receptor, and its activation leads to inflammatory cytokine production ^[24]. TLR4 is typically activated by lipopolysaccharide (LPS) ^[25]. TLR4 can signal through a MyD88-dependent or -independent pathway leading to NFkB and activation of inflammatory cytokines ^[26]. TLR4 has a large number of agonists and antagonists, due to its role in opioid analgesia and sepsis ^{[27, 28]}. Focusing on diabetic retinopathy, studies have shown that TLR4 polymorphisms are linked to non-proliferative diabetic retinopathy in type 2 diabetic patients of Romanian descent ^[29]. Additionally, studies in monocytes freshly isolated from type 2 diabetic patients demonstrated increased TLR4 activation and downstream signaling ^[30]. Another group found that TLR4 polymorphisms were not linked to type 2 diabetes but that this polymorphism did contribute to progression of diabetic retinopathy ^[31]. Taken together, the exact role of TLR4 in diabetic retinopathy is not clear, but a link to diabetic retinopathy is established potentially through its activation by HMGB1 [7].

Based upon the work in humans, investigators initiated studies to better understand TLR4 in mouse and cell models of diabetic retinopathy. Work in human retinal endothelial cells showed that high glucose increased TLR2 and TLR4, leading to increased MyD88-dependent and independent signaling ^[32]. Similarly, others showed that increased TLR4 levels led to retinol-binding protein 4 (RBP4)-mediated inflammation ^[33]. In RGC grown in high glucose, inhibition of TLR4 reduced apoptosis and inflammatory markers ^[34]. Using diabetic TLR4 knockout mice, authors showed that bone marrow transplantation from wild-type mice to mutant mice led to increased vascular endothelial cell grown factor (VEGF), tumor necrosis factor alpha (TNFα) and interleukin-1β (IL-β) levels in the retina, suggesting that TLR4 is key to inflammation in diabetic retinopathy ^[35]. Additional studies on diabetic TLR4 knockout mice also showed that TLR4 drives inflammation in the retina ^[36].

Inhibition of TLR4 offers a novel avenue for therapeutic development. Using diabetic mice treated with miR-499 had increased progression of diabetic retinopathy markers through TLR4 actions ^[37]. In contrast, studies showed that miR-145 reduced TLR4/nuclear factor kappa beta (NFkB)-induced inflammation in REC ^[38]. Similarly, we found that miR-146 reduced TLR4mediated inflammation in the retina ^[39]. In addition to miRNA, we previously demonstrated that Compound 49b, a β-adrenergic receptor agonist, reduced TLR4 actions in REC, retinal Müller cells, and in the diabetic retina ^[40]. In a follow-up study, we showed that Epac1 inhibited TLR4 signaling in REC and in whole retina using Epac1 endothelial cell specific knockout mice ^[41]. Using whole retinal lysates from Müller cell specific TLR4 knockout mice, we demonstrated reduced MyD88-dependent and -independent signaling, leading to reduced inflammatory mediator levels ^[42]. The studies in cell-specific Epac1 KO and TLR4 KO mice were done in non-diabetic animals; however, the data suggest that TLR4 can be inhibited by a number of pathways that could lead to novel therapeutics for diabetic retinopathy.

RAGE actions in the diabetic retina.

RAGE is a transmembrane receptor in the immunoglobulin family, named for its primary agonist, advanced glycation products (AGEs) formed following chronic exposure to hyperglycemia ^{[43, 44]}. When RAGE interacts with its ligands, AGEs or HMGB1, activation of NFkB leads to a pro-inflammatory state ^[45]. AGE/RAGE have been associated with most of the complications of diabetic retinopathy, including reactive oxygen species (ROS), permeability changes, vascular damage, and angiogenic changes ^{[46, 47]}. RAGE has been localized in retinal glial cells, Müller cells, and REC ^[48]. Studies of patients with proliferative retinal disease found evidence of increased levels of RAGE and its ligands in the vitreous and epiretinal membranes ^[49]. The RAGE gene promoter is hypomethylated in patients with type 2 diabetes ^[50].

In support of the studies of RAGE actions in humans, a study using human samples, diabetic mice, and endothelial and Müller cells in culture demonstrated that both TLR4- and AGE-induced retinal inflammation through an upregulation of galectin-1 ^[51]. Studies using streptozotocin-induced diabetic mice and Müller glia showed that RAGE was significantly upregulated and primarily localized in Müller cells. When cells were cultured in high glucose, inhibition of RAGE abrogated the inflammatory response ^[52]. Use of a RAGE fusion protein in streptozotocin-induced diabetes showed that inhibition of RAGE prevented the vascular and permeability damage commonly associated with diabetes ^[53]. Studies using natural products demonstrated that an extract of Polygonum cuspidatum was able to block the HMGB1/RAGE pathway in diabetic rats ^[54]. An additional study showed that the class A scavenger receptor (SR-A) antagonized RAGE actions to reduce the inflammatory actions in the diabetic retina ^[55]. Somewhat similarly, studies also demonstrated that olmesartan, a drug altering the reninangiotensin pathway, blocked inflammatory actions and reactive oxygen species in REC mediated through RAGE ^[56]. Thus, the data suggest that HMGB1/RAGE/TLR4 activities lead to inflammatory-induced damage in the diabetic retina, and its inhibition may serve as a novel arena for therapeutic development (Figure 1).

NFkB is a key point for retinal inflammation in diabetes.

A key action of HMGB1/TLR4/RAGE is to phosphorylate NFkB, leading to activation of a number of inflammatory cytokines ^{[57, 58]}. NFkB is a protein complex controlling DNA transcription and production of a large number of cytokines ^[59]. Due to its role in regulating transcription and activation of cytokines, there are a plethora of agonists and antagonists for NFkB to treat a number of diseases, including cancers and inflammatory diseases ^{[60, 61]}. NFkB can also be regulated via acetylation/deacetylation, which can be regulated by epigenetic factors common to diabetic retinopathy ^{[62, 63]}. Because of its role in activation of a number of key cytokines observed in diabetic retinopathy, a large number of studies have investigated pathways that can inhibit/activate NFkB. For example, one study showed that resolvin D1 inhibited NFkB actions in STZ diabetic rats ^[64]. Similarly, additional studies have shown that selenium or coconut kernel protein can also block RAGE and NFkB in STZ-treated rats ^{[58, 65]}. These studies are representative of a larger number of studies showing that inhibition of NFkB can reduce retinal damage in response to diabetes.

Conclusions:

Data demonstrate that HMGB1 levels are increased in patients with later stage diabetic retinopathy. Similarly, RAGE levels and polymorphisms of TLR4 are increased in diabetic patients. Studies in diabetic animal models and cells demonstrate that the HMGB1/TLR4/RAGE axis is linked to increased levels of retinal damage and inflammatory mediators in the retina. Taken together, HMGB1 signaling can activate both TLR4 and RAGE pathways, leading to increased NFkB actions and retinal inflammation in the diabetic retina.

Highlights.

HMGB1 mediates retinal inflammation in diabetic retinopathy

HMGB1 binds TLR4 and RAGE

Inhibition of HMGB1 offers new therapeutic development