Müller Cells and Diabetic Retinopathy

Abstract

Müller cells are one of the primary glial cell types found in the retina and play a significant role in maintaining retinal function and health. Since Müller cells are the only cell type to span the entire width of the retina and have contact to almost every cell type in the retina they are uniquely positioned to perform a wide variety of functions necessary to maintaining retinal homeostasis. In the healthy retina, Müller cells recycle neurotransmitters, prevent glutamate toxicity, redistribute ions by spatial buffering, participate in the retinoid cycle, and regulate nutrient supplies by multiple mechanisms. Any disturbance to the retinal environment is going to influence proper Müller cell function and well being which in turn will affect the entire retina. This is evident in a disease like diabetic retinopathy where Müller cells contribute to neuronal dysfunction, the production of pro-angiogenic factors leading to neovascularization, the set up of a chronic inflammatory retinal environment, and eventual cell death. In this review, we highlight the importance of Müller cells in maintaining a healthy and functioning retina and discuss various pathological events of diabetic retinopathy in which Müller cells seem to play a crucial role. The beneficial and detrimental effects of cytokine and growth factor production by Müller cells on the microvasculature and retinal neuronal tissue will be outlined. Understanding Müller cell functions within the retina and restoring such function in diabetic retinopathy should become a cornerstone for developing effective therapies to treat diabetic retinopathy.

Introduction

Müller cells are the principle glia of the retina. They are the only cells to span the entire width of the retina and have intimate contact with both the retinal blood vessels and retinal neurons. Because of this arrangement, Müller cells have a variety of important functions in the healthy retina. Functions of Müller cells can be divided into 3 major categories: (1) Uptake and recycling of neurotransmitters, retinoic acid compounds, and ions (such as potassium K⁺), (2) control of metabolism and supply of nutrients for the retina, and (3) regulation of blood flow and maintenance of the blood retinal barrier.

The extensive contact of Müller cells with retinal neurons allows Müller cells to actively participate in proper neurotransmission. They rapidly take up and clear glutamate and γ-aminobutryic acid (GABA) in the inner plexiform layer[1–4]. Studies have shown that Müller cells take up extracellular glutamate through the Glutamate Aspartate Transporter (GLAST) and indicate that glutamate removal and prevention of neurotoxicity in the retina is achieved primarily by this mechanism[5,6]. Once taken up, glutamate is converted to glutamine by glutamine synthetase and released back to neurons for re-synthesis of glutamate and GABA[7]. This process provides substrate for neurotransmitter synthesis and also prevents glutamate toxicity. Müller cells further maintain proper retinal function by participating in a process known as “potassium spatial buffering”, a process that redistributes and normalizes K⁺ in the surrounding microenvironment to avoid prolonged accumulation of K⁺[8]. It has been shown that Müller cells can take up K⁺ from the inner and outer plexiform layers where neuronal synapses occur and release the K⁺ into the vitreous humor in an effort to redistribute K⁺ ions[9]. This process is also involved in retinal fluid removal. Müller cells act as potassium shuttle by taking up potassium from the extracellular fluid through Kir2.1 potassium channels and depositing the potassium into the vasculature using Kir4.1 channels that are found on the Müller cell processes that encompass the blood vessels[10,11]. This leads to osmotic fluid removal through aquaporin-4[11–14].

In addition to regulating neurotransmitters and ion levels within the retina, Müller cells also participate in the retinoid cycle with cone photoreceptors by taking up all-trans retinol from the subretinal space[15–18]. During the visual cycle, photons of light lead to isomerization of 11-cis retinal to all-trans retinal in the rod and cone photoreceptors. Once isomerized, all-trans retinal is expelled from the opsin protein to be reduced by retinol dehydrogenases to all-trans retinol[19]. The all-trans retinol from the cones is then released into the extracellular space where it is taken up by Müller cells, isomerized back to 11-cis retinol by all-trans retinol isomerase, and released back to the extracellular space to be taken up by the cone photoreceptors where it can finally be oxidized from 11-cis retinol back to original 11-cis retinal to restart the visual cycle[15–17,20].

Müller cells seem a primary site of nutrient storage for the retina. It has been shown that ATP production in Müller cells drastically declines when glycolysis is inhibited. However, ATP levels remained equal in aerobic versus anaerobic conditions as long as glucose was provided, indicating that Müller cells live primarily from glycolysis rather than oxidative phosphorylation[21]. This is important as it spares oxygen for retinal neurons and other cell types that use oxidative phosphorylation for ATP production. Furthermore, Müller cells are the primary site of glycogen storage in the retina[21,22]. When nutrient supplies are low Müller cells can utilize this glycogen storage to provide metabolites for other cell types. Furthermore, the large amounts of lactate they produce via glycolysis and irreversible conversion of pyruvate to lactate due to a specific lactate dehydrogenase isoform can be transported to photoreceptors to be used as a potential alternative source of energy in case of need[21,23,24]. Interestingly, studies suggest that the metabolism of glucose and glycogen by Müller cells is regulated by light being absorbed by the photoreceptors[7]. This means that as photoreceptors absorb light, the Müller cells respond by metabolizing more glucose in order to provide more lactate for photoreceptors as needed, indicating that Müller cells and photoreceptors are tightly coupled in their respective functions by metabolism. In addition to providing lactate as a fuel source for photoreceptors, Müller cells can also regulate nutrient supplies to the retina via regulation of retinal blood flow. In a healthy retina, increased light stimulation results in increased retinal blood flow, which is required to supply the activated neurons with oxygen and other nutrients, a process termed neurovascular coupling. Müller cells play a crucial role in neurovascular coupling as they release metabolites controlling vasoconstriction and vasodilation of retinal blood vessels[25,26].

One of the most important functions of Müller cells is their regulation of retinal blood flow and contribution to the blood retinal barrier. The blood retinal barrier is essential for preventing leakage of blood and other potentially harmful stimuli such as pathogens from entering the retinal tissue. It has been shown that Müller cells induce blood-barrier properties in retinal endothelial cells[27,28]. Studies using conditional ablation of Müller cells showed severe blood retinal barrier breakdown[29]. The exact mechanism of how Müller cells maintain the blood retinal barrier is debated but includes the secretion of factors such as pigment epithelium-derived factor (PEDF) and thrombospondin-1 which are anti-angiogenic and increase the tightness of the endothelial barrier[30,31].

It is clear that Müller cells are an integral part of a healthy and well functioning retina. Any disturbance to these cells certainly affects cellular cross-talk within the retina and its proper function. However, despite their importance Müller cells are still an under-studied cell type in the context of diseases such as diabetic retinopathy. The following aims to provide an overview about the effects of diabetes on Müller cells and the role Müller cells play in pathological events in the diabetic retina.

Influence of diabetes on neurotransmitter and potassium regulation in Müller cells

Functional changes that have been determined in Müller cells begin early in the disease, with significant decreases in glutamate transport via GLAST beginning after just 4 weeks of diabetes in rats[32]. This is consistent with reports showing significantly increased glutamate accumulation in the retinas of diabetic rats[33,34]. Furthermore, these studies have shown that there is decreased glutamine synthetase activity and a subsequent decrease in the conversion of glutamate to glutamine necessary for neurotransmitter regeneration[33,34]. These results are in line with reports demonstrating glutamate increases to a potentially neurotoxic level in the vitreous of diabetic patients[35]. However, in neurological diseases such as stroke, therapies targeting glutamate increase have been ineffective indicating that increased glutamate levels might not play a pathophysiological role[36,37]. Whether increased glutamate levels actually cause neurotoxicity over time in diabetic retinopathy has yet to be determined.

It seems that Müller cells not only contribute to glutamate toxicity directly by decreased glutamate uptake, but Müller cells also contribute indirectly via decreased K⁺ uptake during the progression of diabetic retinopathy. There is decreased K⁺ conductance on the plasma membrane of Müller cells isolated from rat retinas after 4 months of experimental diabetes[38]. Redistribution of the Kir4.1 K⁺ channel has been identified as the mechanism of decreased K⁺ conductance[38]. This decrease in K⁺ conductance was also observed in Müller cells of patients with proliferative diabetic retinopathy[39]. Alteration of the Kir4.1 K⁺ channel localization in Müller cells in the diabetic retina has been attributed to the accumulation of advanced glycation endproducts (AGEs)[40]. Together, this can lead to an imbalance in K⁺ concentrations and altered K⁺ homeostasis leading to neuronal excitation and subsequent glutamate toxicity.

In diabetes and diabetic macular edema, Müller cells have been shown to downregulate the Kir4.1 channels, but not Kir2.1, leading to continued potassium uptake with no release into the microvasculature[38,41,42]. This leads to subsequent swelling of Müller cells contributing to Müller cell dysfunction and decreased fluid removal contributing to diabetic macular edema. Diabetic macular edema leads to thickening of the macula due to fluid accumulation and can be observed by optical coherence tomography (OCT). The thickening of the macula due to fluid accumulation typically leads to disruption of the retinal structure and changes in visual acuity.

Release of growth factors and pro-/anti-inflammatory cytokines from Müller cells in response to hyperglycemia – the bad and the potentially good

As already stated above, Müller cell have contact with every cell in the retina. Müller cell ablation leads to photoreceptor degeneration, vascular leak, and intraretinal neovascularization demonstrating that Müller cells are necessary for both neuronal and vascular function and viability[29,43]. Changes to their environment by hyperglycemia alters functional interaction with pericytes[44]. Deletion of the dystrophin-Dp71 protein within Müller cells caused extensive vascular leakage and edema in the mouse retina. It was suggested that breakdown of the blood retinal barrier was initiated by improper localization of proteins in the endfeet of Müller cells that are necessary for establishing barrier function[45]. Other studies have shown that Müller cells participate in regulation of vascular tone in a process of neurovascular coupling[25,26]. They are also seemingly involved in lactate exchange with neurons, glia, and vascular cells[46]. Given the intricate contact Müller cells have with other retinal cell types it is easy to see that any disturbance to Müller cells will certainly affect proper function and viability of neurons as well as cell of the microvasculature.

In diabetes, it has been well established that Müller cells become activated[47–50]. One of the most prominent signs that Müller cells are activated in diabetic retinopathy is the increased expression of glial fibrillary acidic protein (GFAP), a common marker of reactive gliosis[33,48,51]. In healthy conditions, Müller cells generally do not express GFAP[47,52]. Interestingly, while Müller cells upregulate GFAP expression in the diabetic retina astrocytes seemingly downregulate GFAP expression[53]. Figure 1 demonstrates the high level of GFAP expression in Müller cells in the diabetic retina. It also highlights the extensive contact that Müller cells have with the retinal microvasculature making it easy to comprehend the influence activated Müller cells have on proper function of the microvasculature.

Figure 1.

picture (4×) taken of a retinal flatmount obtained from a STZ (streptozotocin) diabetic mouse that expressed a GFP (green fluorescence protein) tagged GFAP specifically in Müller cells[49]. Duration of diabetes: 10 weeks.

Purple: microvasculature

Green: Müller cells

Despite GFAP several other markers might be more useful to determine early glial activation such as phospho-ERK (extracellular signal-regulated kinase)[54]. Although elevated GFAP expression happens early and persists throughout the disease, no study to date has been able to connect increased levels of GFAP to any functional outcome.

However, hyperglycemia-induced gliosis goes hand in hand with stimulation of growth factor, cytokine, and chemokine release by Müller cells at least in vitro. Hyperglycemia promotes release of (1) growth factors, such as vascular endothelial growth factor (VEGF) and pigment epithelium-derived factor (PEDF), and (2) cytokines and chemokines including interleukin-1β (IL-β), interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), and chemokine ligand-2 (CCL2)[52,55–61] [62–64]. In vitro studies have provided ample evidence that Müller cells are a potential source for growth factors and cytokines when stimulated with elevated glucose levels. Considering that most of the growth factors, cytokines, and chemokines released by Müller cells have been identified in the vitreous of diabetic patients it is fair to assume that Müller cells contribute to the overall synthesis of these factors in vivo[65–68].

Growth factors – the bad

How much Müller cell derived growth factors really contribute to the pathology of diabetic retinopathy in vivo is still not fully understood. The first studies to understand the contribution and effect of Müller cell derived VEGF to the development and progression of diabetic retinopathy were done by the group of Y.Z. Le. This group disrupted VEGF in Müller cells with an inducible Cre/lox system and examined diabetes-induced retinal inflammation and vascular leakage in these conditional VEGF knockout (KO) mice. The diabetic conditional VEGF KO mice exhibited an overall decrease in parameters associated with the pathology of diabetic retinopathy such as leukostasis, expression of inflammatory biomarkers, depletion of tight junction proteins, numbers of acellular capillaries, and vascular leakage compared to diabetic control mice[59,69,70]. Additional studies focusing on altering known regulators of VEGF production such as HIF-1 (hypoxia inducible factor 1)[71] and the Wnt signaling pathway[72] specifically in Müller cells have supported the notion that Müller cell derived VEGF is actually a major component in the process of retinal angiogenesis and pathology in diabetic retinopathy. Besides VEGF, Müller cell derived PEDF has also been suggested to have its part in diabetes-induced retinal angiogenesis[30]. Taken together, it seems that Müller cell derived growth factors contribute heavily to pathological vascular events in diabetic retinopathy.

Growth factors – the potentially good

Although Müller cell derived VEGF contributes to detrimental effects on the microvasculature in the diabetic retina, the intent of such growth factor production by Müller cells in the first place might have been to protect itself and the retinal neurons from a diabetic insult. This idea is supported by a study using mice that carry a disrupted VEGFR2 specifically in Müller cells. Loss of VEGFR2 caused a gradual reduction in Müller glial density, decreased of scotopic and photopic electroretinography amplitudes, and accelerated loss of photoreceptors, ganglion cells, and inner nuclear layer neurons in the diabetic retina[73]. More studies are needed to fully explore and understand the beneficial effects of Müller cell derived growth factors on Müller cells itself and retinal neurons in the context of disease. This is especially important since long-term anti-VEGF treatment might hamper functional integrity of Müller cells and neurons causing unexpected additional problems in treating diabetic retinopathy.

Cytokines – the bad

Besides growth factors, Müller cells release a variety of cytokines and chemokines under hyperglycemic conditions. For example, Müller cells are a major source of retinal interleukin-1beta (IL-1β) production[63,74–77]. Caspase-1, originally named interleukin-1β converting enzyme (ICE), produces the active cytokines IL-1β and IL-18 by cleavage of their inactive proform[78–81]. In Müller cells, hyperglycemia strongly induces the activation of the caspase-1/IL-1β signaling pathway as we have previously shown[63,77]. Increased caspase-1 activation and elevated IL-1β levels have also been identified in the retinas of diabetic mice and retinal tissue and vitreous fluid of diabetic patients[63,75,82–84]. We have identified that targeting this pathway by knocking down caspase-1 or the IL-1 receptor (IL-1R1) or by pharmacological intervention protects against the development of diabetic retinopathy in diabetic rats and mice[76,85]. Prolonged IL-1β production by Müller cells has been shown to affect endothelial cell viability in a paracrine fashion[75]. Endothelial cells are extremely susceptible to IL-1β and rapidly progress to cell death in response to this pro-inflammatory cytokine[75]. Endothelial cell death is detectable in the retinal microvasculature of diabetic animals and isolated retinal blood vessels of diabetic donors and has been associated with the formation of acellular capillaries, a hallmark of retinal pathology in diabetic retinopathy[86]. Besides IL-1β, Müller cells produce other well-known pro-inflammatory cytokines such as tumor necrosis factor alpha (TNFα) and interleukin-6 (IL-6)[76,77,85,87–90]. Anti-TNFα therapy has been proposed as a strategy to treat diabetic retinopathy in diabetic animals[91–94]. Detrimental effects of IL-6 have been associated with vascular dysfunction and promotion of angiogenesis[95–97] which is why IL-6 recently has become a new therapeutical target of interest to prevent diabetes-induced vascular damage. The production and release of pro-inflammatory cytokines by Müller cells strongly contributes to the chronic inflammatory environment detected in the diabetic retina that over time promotes drop-out of a retinal cells.

Cytokines – the potentially good

From a vascular perspective, IL-6 has been solely associated with detrimental effects[95–97]. However, we have previously shown that IL-6 prevents hyperglycemia-induced Müller cell dysfunction and loss clearly supporting a beneficial and protective nature of IL-6[77]. This observation is well in line with reports that in the retina IL-6 is an important cytokine responsible for maintaining proper neuronal function as well as stimulating neuroprotective effects[77,98–101]. Treatment with IL-6 has been shown to protect retinal ganglion cells from pressure-induced cell death[98]. Additionally, in an experimental model of retinal detachment, genetic ablation or neutralization of IL-6 led to a significant increase in photoreceptor cell death. However, treatment with exogenous IL-6 resulted in a significant increase in photoreceptor density in the outer nuclear layer[101]. These different effects of IL-6 can potentially be attributed to the two distinct signaling pathways IL-6 acts through. Classical IL-6 signaling - thought to be the anti-inflammatory and protective pathway - is mediated by the membrane-bound form of the IL-6 receptor (IL-6R) and the ubiquitously expressed glycoprotein 130 (gp130). Only cells such as Müller cells (but not endothelial cells) that express IL-6R are able to signal through classical IL-6 signaling. Conversely, IL-6 trans-signaling, which is mediated by binding of IL-6 to the soluble form of the IL-6 receptor (sIL-6R) and gp130, is thought to be the more pro-inflammatory and pro-angiogenic pathway[77,96,99,102–112]. In diabetic patients, correlations between increased levels of IL-6 and the development of complications in the eye have been made[113–118]. However, whether IL-6 levels are increased in diabetes as an attempt to protect from a pro-inflammatory environment or whether high levels of IL-6 synergistically exaggerate diabetes-induced inflammation has yet to be determined.

Müller cell loss in diabetic retinopathy

Whether Müller cells die in diabetic retinopathy has long been a matter of debate. It is easy to see that Müller cells are “sturdy” cells taking into account how well equipped these cells are to produce fair amounts of protective factors that shield them at least in the beginning from a chronic diabetic insult as discussed above. However, newer studies indicate that over time Müller cells actually do begin to die the longer diabetic retinopathy progresses. Frequency of Müller cell death in the diabetic retina rapidly accelerates when protective growth factors are blunted[73].

Better understanding of types of cell deaths has furthered studies to look for mechanisms other than apoptosis by which Müller cells can die in a diabetic environment. We have identified one particular mechanism of cell death that stands out and can explain histological features described for Müller cells in the diabetic retina. Pyroptosis is an inflammatory driven type of cell death that depends on caspase-1 activation[119–121]. Müller cells show increased caspase-1 activity and IL-1β production following exposure to hyperglycemic conditions and cells die as a consequence[122,123]. While it is known that initiation of pyroptosis is caspase-1 and IL-1β driven, the execution phase of pyroptosis is not yet completely understood. Execution of pyroptosis shares traits with both apoptosis and necrosis[124,125]. Since execution of pyropototic cell death lacks specific marker, identifying retinal cells dying by pyroptosis in vivo is a difficult task. Markers such as TUNEL staining used to detect apoptotic cell death may not adequately detect pyroptosis. Therefore, we have performed a study actually counting Müller cells in the healthy and diabetic retina and determined roughly 15% cell death at 7 months of diabetes[88]. Even more important, inhibition of the caspase-1/IL-1β pathway inhibited diabetes-induced Müller cell death in vivo as we had previously shown in vitro[76,77,88]. Several other studies are in line with our observation that Müller cells die in a hyperglycemic environment. The first study to describe dying Müller cells in diabetic retinopathy was done using EM analysis[126]. Dying Müller cells are described as being hypertrophic consistent with the notion that during pyroptosis, cells swell rather than shrink as observed in apoptotic cell death[48]. To collect more evidence for Müller cells death in the diabetic retina we looked at earlier markers of cell death and we have identified that GAPDH (glyceraldehyde-3-phosphate dehydrogenase) accumulates in the nucleus of Müller cells in the retinas of diabetic rats[50]. Nuclear accumulation of GAPDH has been closely associated with cell death induction[127–129]. Consistent with our finding that Müller cells die by pyroptotic cell death, hyperglycemia-induced nuclear accumulation of GAPDH depends on the activation of the caspase-1/IL-1β pathway[52,130]. The consequences of dying Müller cells are multi faceted. On the bad side – Müller cell death will promote loss of retinal blood barrier integrity, increased vascular permeability, and loss of neuroprotection affecting both neurons and vascular cells. Loss of Müller cells in diabetes has also been associated with aneurysm formation, a clinical characteristic of diabetic retinopathy[126]. However, one can also argue that on the good side – removal of activated and pro-inflammatory Müller cells might be a “shut off” mechanism to deal with an increasing inflammatory environment in the diabetic retina. A lot more studies are needed to determine the full pathway of Müller cells death and to identify whether all Müller cells are equally affected by hyperglycemia.

Figure 2.

Timeline of caspase-1 activation, cytokine secretion, Müller cell death initiation and execution in comparison to other prominent events associated with diabetic retinopathy in retinas of STZ diabetic mice and rats.

Conclusion

Müller cells are a major component of a healthy retinal environment. Once chronic hyperglycemia disturbs their environment, Müller cells become dysfunctional and start activating pathways to counter-regulate and “repair” the environment.

In order to do so, Müller cells release a large variety of growth factors and cytokines in a diabetic environment. Most of the research to date has focused on the detrimental effects the release of these growth factors and cytokines causes to the retina. When taking a closer look most of these effects are associated with vascular dysfunction and angiogenesis. On the other hand, it seems that production of these growth factors and cytokines by Müller cells are primarily intended to protect Müller cells and consequently retinal neurons from diabetic insult and might only secondarily turn into the damaging components observed in diabetic retinopathy. Very few studies have started to consider the protective nature of Müller cell derived growth factors and cytokines in regards to the integrity of glia cells and neurons. A lot more studies are needed to understand the nature of Müller cells derived growth factors and cytokines. For a successful development of a new therapy targeting these factors both detrimental as well as beneficial effects need to be considered.

Understanding Müller cell functions within the retina and restoring such function in diabetic retinopathy should become a cornerstone for developing effective therapies to treat diabetic retinopathy. Some approaches have been tested to increase Müller cell function by stimulating the beta-adrenergic pathway[131,132]. Whether these studies materialize into effective therapy strategies has to be seen in the future.