Dietary Supplement Enriched in Antioxidants and Omega-3 Promotes Glutamine Synthesis in Müller Cells: A Key Process against Oxidative Stress in Retina

To prevent ocular pathologies, new generation of dietary supplements have been commercially available. They consist of nutritional supplement mixing components known to provide antioxidative properties, such as unsaturated fatty acid, resveratrol or flavonoids. However, to date, only one preclinical study has evaluated the impact of a mixture mainly composed of those components (Nutrof Total®) on the retina and demonstrated that in vivo supplementation prevents the retina from structural and functional injuries induced by light. Considering the crucial role played by the glial Müller cells in the retina, particularly to regulate the glutamate cycle to prevent damage in oxidative stress conditions, we questioned the impact of this ocular supplement on the glutamate metabolic cycle. To this end, various molecular aspects associated with the glutamate/glutamine metabolism cycle in Müller cells were investigated on primary Müller cells cultures incubated, or not, with the commercially mix supplement before being subjected, or not, to oxidative conditions. Our results demonstrated that in vitro supplementation provides guidance of the glutamate/glutamine cycle in favor of glutamine synthesis. These results suggest that glutamine synthesis is a crucial cellular process of retinal protection against oxidative damages and could be a key step in the previous in vivo beneficial results provided by the dietary supplementation.


Introduction
Müller cells play a crucial role in maintaining the complex architecture and function of the retina. They are considered as the principal glial cells of the retina since they perform many of the functions carried out by astrocytes, oligodendrocytes and ependymal cells in other regions of the central nervous system [1]. Müller cells are radially oriented, spanning the entire retina from the inner to the distal end of the outer nuclear layer (ONL) and presenting close contacts with photoreceptors and bipolar cells. These strong interactions with retinal neurons ensure numerous metabolic exchanges maintaining retinal integrity,

Experimental Design
After 3 days of culture corresponding to 80% confluence, cells were incubated with (+SC) or without (−SC) dietary supplementation. Four days later (T0) media was replaced with fresh culture media containing 1mM of glutamate and cells were untreated (vehicle) or treated with 100 µM H2O2 or with 200 µM H2O2. The different assays were performed at T0, T24 (+24 h) and T48 (+24 h) ( Figure 1D).
Extra-and intra-cellular glutamate and glutamine assays as well as glucose consumption experiments were performed at T0, T24 and T48. mRNA expression, protein expression and cell viability were performed at T48. Lipid composition and transcriptomic profiles were evaluated at T0 ( Figure 1C). Total amount of proteins was determined for each well and used to normalize results.
Each experiment was performed at least 5 times and had been carried out by 3 different experimenters.

Glutamate and Glutamine Dosages
Glutamate and glutamine concentrations were determined by modified GLN1 Kit (Sigma, Saint-Quentin-Fallavier, France). Glutamate concentration was directly evaluated in the sample by dehydrogenation of the L-glutamate to α-ketoglutarate accompanied by reduction of NAD+ to NADH. The conversion of NAD+ to NADH was measured spectrophotometrically and was proportional to the amount of oxidized glutamate. De-

Glutamate and Glutamine Dosages
Glutamate and glutamine concentrations were determined by modified GLN1 Kit (Sigma, Saint-Quentin-Fallavier, France). Glutamate concentration was directly evaluated in the sample by dehydrogenation of the L-glutamate to α-ketoglutarate accompanied by reduction of NAD+ to NADH. The conversion of NAD+ to NADH was measured spectrophotometrically and was proportional to the amount of oxidized glutamate. Determination of L-glutamine was a two-step reaction: (A) deamination of L-glutamine into L-glutamate by glutaminase and (B) dehydrogenation of the L-glutamate to α-ketoglutarate by glutamate dehydrogenase. Hence, the amount of glutamine converted into glutamate in the samples was obtained by the difference between glutamate levels without and with glutaminase. Therefore, two mixes were prepared: sodium acetate buffer (0.4 M, pH 4.8) with 25% of glutaminase (from Escherichia coli, Sigma, Saint-Quentin-Fallavier, France) and the same sodium acetate buffer without glutaminase. To determine the extracellular glutamate and glutamine concentrations, 90 µL of assessed medium were mixed to 60 µL of each mix. After incubation at 37 • C for 1 h, glutamate concentration was dosed in L-GLDH Nutrients 2021, 13, 3216 5 of 18 buffer composed of Tris-EDTA-hydrazine buffer (0.33 M Tris-HCl pH 9.0, 1.33 mM EDTA, 2% hydrazine pH 9.0) with 2 mM β-NAD+, and 6.67 mM ADP, in UV microplates. Thirty percent L-GLDH (L-glutamic dehydrogenase from bovine liver, Sigma, Saint-Quentin-Fallavier, France) enzyme in H 2 O was added to samples, and plates were left for 1 h in the dark at room temperature. Absorbances at 340 nm were then measured (CLARIOstar imager, BMG LABTECH, Ortenberg, Germany), and values recorded before the addition of L-GLDH was subtracted. Glutamate concentration was reported to sample protein concentration. Results were expressed as percent of medium glutamate concentration used for the stimulation. In addition, Aera Under the Curve (AUC) over the 48 h of test were calculated (Arbitrary Unit-AU) using MicroCal software. For intracellular glutamate and glutamine concentrations, cell suspensions were centrifuged for 30 min, 13,000× g, at 4 • C, and cellular pellets were homogenized in sodium acetate buffer (0.05 M, pH 5.0) with sonification. Supernatants were collected and underwent the standard dosage as described previously.

Cell Viability Test
Cell viability was evaluated using MTT colorimetric assays as described previously [22]. Briefly, medium was removed, 500 µL of RPMI medium containing MTT (0.5 mg/mL) were added to each well and incubated for 4 h at 37 • C. After incubation, 50 µL of SDS solution (10% of SDS in DMSO added with pure acetic acid 0.6% final) were added to each well to solubilize the formazan product. The plate was left in the dark overnight at room temperature. Absorbance of each well at 610nm was measured using CLARIOstar imager (BMG LABTECH, Ortenberg, Germany). Results were expressed as percent of unsupplemented and untreated cells.

Glucose Consumption
Glucose concentration in culture media was determined by spectrophotometric methods using the Glucose GOD-PAD kit (Prolabo, Paris, France) according to the manufacturer's instructions. Absorbance at 500 nm was measured for each well using the CLARIOstar imager (BMG LABTECH, Ortenberg, Germany). Glucose consumption was calculated as the difference between the glucose concentration at T48 and the initial glucose concentration at T0 (1 g/L). Results were expressed as percent of untreated cells.

Quantitative RT-PCR
Quantitative RT-PCR was performed using Taqman technology (Applied technologies, Cleveland, OH, USA) as described previously [23]. Briefly, total RNA was isolated from Müller cells culture using TRIzol reagent (Ambion, Life Technologies, Carlsbad, CA, USA), quantified and reverse transcripted (Superscript iii reverse transcriptase, Invitrogen, Life Technologies). Real-time PCR reactions were performed in the Mx3005P Agilent (Applied, Life Technologies) with fivefold dilution of cDNA, 200 nM of each Taqman primer using the Expression Master Mix (Applied, Life Technologies). Data were normalized to the reference standard RNA 18S, analyzed by ∆∆Ct method and expressed as LogFC (Fold Change) compared to basal condition (−SC). Each measurement was performed three times.

Expression Arrays
Total RNA was extracted using TRIzol reagent (Invitrogen, Waltham, MA, USA) according to the manufacturer's instruction. Genome-wide transcriptional profiling was performed by Mice Exon 1.0 ST arrays (Affymetrix, Santa Clara, CA, USA) following manufacturer's instructions and as described previously [24,25]. Each condition was performed in duplicate. GeneChip CEL files were analyzed using GeneSpring (Agilent Technologies, Santa Clara, CA, USA). The log2 values were obtained and the microarray data were normalized using RMA16 algorithm. Statistical data analyses were done on all samples for each group. Similarly, enrichment rankings were based on all samples per group. Data were filtered with GeneSpring analysis using a cutoff of at least 1.5 times up or down (Fold Change 1.5, FC1.5). Subsequently, one-way ANOVA was performed on a filtered list. The Benjamini-Hochberg false discovery rate (FDR) was set to 5%, and the P value was set to <0.05 (ANOVA test). The Database for Annotation, Visualization, and Integrated Discovery (DAVID) was used to find gene ontology-enriched terms. Data were confirmed by quantitative RT-PCR. Quantitative RT-PCR were performed using Taqman technology (Applied Technologies) as described previously [25]. Data were analyzed by ∆∆Ct method and normalized to the reference standard RNA 18S. Each measurement was performed three times.

Fatty Acid Composition
Müller cells were homogenized in PBS-1X after scraping, centrifuged at 10,000× g for 10 min at 4 • C and conserved at −80 • C until lipid extraction.
Total lipids extraction: Total lipids were extracted from cell homogenates following Folch's method [26].
Determination of lipid class distribution: The distribution of lipids into phospholipids, triglycerides, free fatty acids, free cholesterol, and cholesteryl esters in cells was determined using a combination of thin-layer chromatography on silica gel-coated quartz rods and flame ionization detection (Iatroscan ® system, Iatron, Tokyo, Japan), according to Ackman's technique [27] (Ackman, 1981). The values obtained for each compound were corrected according to their response factor using specific calibration curves, as published previously [28]. Data were reported as a percentage of the total lipids in the sample.
Determination of fatty acid composition: Lipids were transmethylated using boron trifluoride in methanol according to Morrison and Smith (1964) [29]. The fatty acid methyl esters formed by transmethylation were analyzed on a Trace 1310 gas chromatograph (Thermofisher, Waltham, MA, USA) fitted with a CPSIL-88 column (100 m × 0.25 mm i.d., film thickness 0.20 µm; Varian, Les Ulis, France). Hydrogen was used as a carrier gas (inlet pressure 210 kPa). The oven temperature was held at 60 • C for 5 min, increased to 165 • C at 15 • C/min and held for 1min, then increased to 225 • C at 2 • C/min, and finally held at 225 • C for 7 min. The injector and the flame ionization detector were maintained at 250 • C. Fatty acid methyl esters were identified by comparison with commercial and synthetic standards. The data were computed using the EZChrom software (Agilent Technologies, Massy, France) and reported as a percentage of the total fatty acids.

Statistical Analysis
All results are expressed as mean ± SEM. Data analysis was performed using Graph-Pad Prism version 8.0.0 for Windows, GraphPad Software, San Diego, California, USA. For supplementation effect, we performed Mann-Whitney's non-parametric test, whereas the oxidative stress effect was investigated using a one-way ANOVA followed by a post-hoc Holm-Sidak test. Statistical significance was defined as p < 0.05. Significant differences between groups are noted by * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.
All together, these results showed the nutritional supplement leaded to modifications of the glutamate/glutamine metabolic cycle in Müller cells, mainly characterized by an increase of glutamine synthesis independently from glutamate uptake.
In +SC, glutamine concentration in culture medium increased over time with or without sub-lethal ( Figure 4G). Therefore, the decrease in extracellular glutamine was not correlated neither to an increase of intracellular glutamine concentration nor to a decrease of SN1 expression despite an upregulation of GS protein.
These results suggest that in +SC, in oxidative condition, the level of intracellular glutamine is maintained by an increase of its synthesis by GS from a glutamate source other than the one added in the medium. Therefore, nutritional supplementation (+SC) seems to promote maintenance of a high level of glutamine in Müller cells by enhancing glutamine synthesis.
Therefore, it appears that supplementation reduced Müller cells reactivity to oxidative stress thus preventing cell death in lethal oxidative stress conditions. Nutrients 2021, 13, x FOR PEER REVIEW 13 of 19

Discussion
Ocular supplements are routinely recommended for retinal degenerative diseases such as Age-related Macular Degeneration (AMD). However, little is known about their preventive mechanisms against retinal damage. Considering the importance of Müller cells in the retina, especially in photoreceptors cell neuroprotection [6], we raised the question of their involvement in the preventive effect of ocular supplementation. We particularly focused on the glutamate/glutamine metabolic cycle which is a major actor of Müller cell-associated neuroprotection [6] in both normal and oxidative stress conditions.

Dietary Supplementation Modifies Glutamate/Glutamine Metabolic Cycle in Far of Glutamine Synthesis
Important metabolic couplings exist among various cells through the use of common substrates and the exchange of several metabolic intermediates such as glutamate or glutamine [30,31]. In the retinal tripartite glutamatergic synapse, the fine control of glutamate level is ensured by Müller cells controlling the glutamate/glutamine metabolic cycle. The metabolic cycle aims to recapture glutamate from the synaptic cleft to maintain homeostasis, and to provide glutamine to replenish neurotransmitter pools in neurons, and thus avoid their excitotoxicity [6].
In this study, as expected, Müller cells were able to uptake glutamate, which could then be converted into glutamine by the glutamine synthetase (GS). Then, glutamine was released in culture medium, and acted as the driving force for glutamate uptake and GS activity to avoid any glutamate toxicity [32][33][34]. Interestingly, in high glutamate concentration [19] as used in this study to mimic glutamate toxicity involved in retinal degeneration [20,21], the main striking effect of dietary supplementation on Müller cells is the modulation of the glutamate/glutamine cycle in favor of glutamine release. The increase in GS, responsible for conversion of glutamate into glutamine [35], leads to an increase in intracellular glutamine as well as in glutamine release from Müller cells. This was correlated to an up-regulation of the membrane externalizing transporter SN1. The direct relationship between the glutamine synthesis, glutamine release and extracellular glutamine level is in accordance with the literature [36]. However, while glutamine increase is usually linked to an upregulation of glutamate internalizing transporter (EAAT1) Glucose consumptions are expressed as the percentage of glucose of the initial amount (known at T0) which had been consumed. For each group, n = 6 mice. All results are presented as mean ± SEM. Significant differences between groups are noted by ** p < 0.01.

Discussion
Ocular supplements are routinely recommended for retinal degenerative diseases such as Age-related Macular Degeneration (AMD). However, little is known about their preventive mechanisms against retinal damage. Considering the importance of Müller cells in the retina, especially in photoreceptors cell neuroprotection [6], we raised the question of their involvement in the preventive effect of ocular supplementation. We particularly focused on the glutamate/glutamine metabolic cycle which is a major actor of Müller cell-associated neuroprotection [6] in both normal and oxidative stress conditions.

Dietary Supplementation Modifies Glutamate/Glutamine Metabolic Cycle in Far of Glutamine Synthesis
Important metabolic couplings exist among various cells through the use of common substrates and the exchange of several metabolic intermediates such as glutamate or glutamine [30,31]. In the retinal tripartite glutamatergic synapse, the fine control of glutamate level is ensured by Müller cells controlling the glutamate/glutamine metabolic cycle. The metabolic cycle aims to recapture glutamate from the synaptic cleft to maintain homeostasis, and to provide glutamine to replenish neurotransmitter pools in neurons, and thus avoid their excitotoxicity [6].
In this study, as expected, Müller cells were able to uptake glutamate, which could then be converted into glutamine by the glutamine synthetase (GS). Then, glutamine was released in culture medium, and acted as the driving force for glutamate uptake and GS activity to avoid any glutamate toxicity [32][33][34]. Interestingly, in high glutamate concentration [19] as used in this study to mimic glutamate toxicity involved in retinal degeneration [20,21], the main striking effect of dietary supplementation on Müller cells is the modulation of the glutamate/glutamine cycle in favor of glutamine release. The increase in GS, responsible for conversion of glutamate into glutamine [35], leads to an increase in intracellular glutamine as well as in glutamine release from Müller cells. This was correlated to an up-regulation of the membrane externalizing transporter SN1. The direct relationship between the glutamine synthesis, glutamine release and extracellular glutamine level is in accordance with the literature [36]. However, while glutamine increase is usually linked to an upregulation of glutamate internalizing transporter (EAAT1) [6,37,38] leading to an increase in glutamate uptake [39] and enhancing GS activity [34], we did not observe any effect of the supplementation on extracellular glutamate or EAAT1. Therefore, intracellular glutamate does not origin from an extracellular compartment, but from a regulation of intracellular metabolic pathway involving a new carbonate source for glutamate synthesis. Glutamate can be produced 1/from glucose through a metabolic pathway that begins with conversion of glucose to pyruvate which enter the Krebs cycle (TCA cycle) [40,41], 2/from protein catabolism producing amino acid free, such as glutamine leading to glutamate by deamidation [40,42]. Since we did not observe any change in glucose consumption with the supplementation it is unlikely for glutamate to originate from the Krebs cycle. On the other hand, the observed enrichment of GO term (GO:0032435) named "negative regulation of proteasomal ubiquitin-dependent protein catabolic process" (Figure 2) supports an upregulation of protein breakdown. So, supplementation would enhance protein breakdown leading to an increase in glutamate which in turn would have led to an increased glutamine synthesis by GS. This would have been be balanced by an increase in GLS retroconverting glutamine to glutamate, in order to maintain stoichiometric amounts of glutamate and ammonium ions [43] and/or to the bioenergetics needs of Müller cells [44,45].

Synergic Effect of the Nutritional Supplementation Compounds on Glutamine Enhancement
A seven day supplementation was shown to induce changes in lipid composition such as DHA, EPA and AA in vivo [16] in rats as well as in vitro in retinal pigment epithelial cells (ARPE19) (personal communication, Ramchani-Ben Othman, K). However, in the present study, we did not notice any significant modification in Müller cells' lipid composition, and especially their fatty acid content, after 4 days of supplementation. Therefore, the observed impact of our short-term exposure on the glutamate/glutamine metabolic cycle is independent from cells' lipid composition, in opposition to a long-term exposure to PUFAs [46][47][48]. However, the participation of DHA from supplements to the reduction of glutamate uptake without modifications in the expression of membraneassociated astroglial glutamate transporters (EAAT1 and GLT-1) [49] but with a decreased GS expression [48] cannot be excluded. Therefore, the supplement could have a direct cellular signaling effect through the PUFAs it contains.
Among the other compounds, resveratrol can significantly interfere with the glutamate/glutamine cycle. It has been shown that short-term resveratrol exposure from one to 100 µM induced a linear increase in glutamate uptake and glutamine synthase activity [50]. In addition, in primary cortical astrocytes, 25 µM resveratrol increased glutamate uptake and glutamine content, whereas at 250 µM resveratrol decreased glutamate uptake but increase GS activity [50]. Similarly, resveratrol improved the glutamate metabolism in Müller cells via up-regulation of GS protein expression and activity in diabetic retinas, providing protective effect against retinal degeneration [51]. Even if our experimental conditions are quite different in terms of cellular model (Müller cell), concentration used (113 µM), exposure timing and nutritional supplementation type (complex formulation), we could nevertheless suggest a participation of resveratrol on the glutamate/glutamine cycle modulation. More recently, it was shown that resveratrol administration enhances the activity of the ubiquitin-proteasome system [52] and promotes proteasomal degradation [53]. This mechanism could explain the modification of proteasomal ubiquitin-dependent GO term observed in our model, suggesting that resveratrol is involved in providing carbonate source for the new glutamate synthesis.
Resveratrol and PUFAs are the major components of the supplementation mixture, and the literature has brought evidence of several effects of these substances on glutamatergic homeostasis modulation. Consequently, even if the dietary supplement contains other compounds, we could hypothesize that the observed effect on the glutamate/glutamine metabolic cycle in our experimental conditions might originate mainly from the synergic effect of both PUFAs and resveratrol.

Dietary Supplementation Induced a Glutamine Buffering to Delay Oxidative Stress Impact: High Glutamine Level as a Key Target for Retinal Neuroprotection
Glutamate is also important for maintaining levels of antioxidant glutathione GSH [54]. Transcriptomic analysis demonstrates that nutritional supplementation impacts glutamate homeostasis GO terms as discussed above, but also classical pathways for oxidative stress defense. Indeed, "GO:0004364 Glutathione transferase activity" and "GO:0006749 Glutathione metabolic process" were enriched, and the associated genes such as GSTA1, GSTA2, GM10639 and GM3776 were upregulated. Based on these observations, we raised the question about the effect of oxidative stress on the glutamate/glutamine metabolic cycle of Müller cells and the protective effect of nutritional supplementation against this oxidative stress. Herein, we have chosen sub-lethal oxidative stress conditions to investigate the early oxidative stress response.
As previously described in several different experimental models [55][56][57][58], we showed that oxidative stress affects the glutamate/glutamine metabolic cycle of the Müller cells by decreasing the glutamate uptake, the glutamine synthesis, and thus the glutamine release. Supplemented Müller cells exposed to oxidative stress had a glutamate/glutamine metabolic cycle reaching the level of untreated cells. This is particularly interesting for glutamine levels, which are essential for neuron function. Therefore, we hypothesize that the high glutamine level induced by nutritional supplementation protects Müller cells by creating a buffering against oxidative stress impact. To go further, increasing the oxidative stress up to 800 µM H 2 O 2 clearly showed an offset of cell death in the supplemented conditions compared to the unsupplemented ones. Consequently, by increasing the level of glutamine, the nutritional compound protects Müller cells against oxidative mechanism. Therefore, we can assume that in vivo, the supplement offered retinal protection against oxidative stress induced by light [16] partly by acting on glutamine level. Our data and conclusions are in accordance with literature, since impairments of glutamine level associated to GS defect are clearly highlighted in many diseases leading to neuronal glutamate depletion [59]. Indeed, in photoreceptor cells death (related to inherited photoreceptor degeneration, retinal light injury, or retinal detachment), Müller cells display a drop in their GS expression and glutamine level [60][61][62]. Similarly, a decline in GS expression and activity was also observed under ischemic, inflammatory, and traumatic conditions or in glaucoma [7,[63][64][65]. These observations are in accordance with experimental data using pharmacological inhibitor of the GS activity in the retina. In substance, in vivo retinal inhibition of GS led to a loss of free glutamate content in bipolar and ganglion cells, and thus to rapid blindness in animals [8,66]. Moreover, downregulation of GS using siRNA results in the breakdown of the blood-retinal barrier [67], suggesting that glutamate homeostasis defect is also associated to the integrity of the blood-retinal barrier. Consequently, we could raise the hypothesis that the control of glutamine level, especially its high concentration, is a key point for neuronal retinal cells' protection, and thus the photoreceptor cells' survival ( Figure 6).
In conclusion, our data demonstrate that enrichment of Müller cell environment with omega-3, resveratrol and minerals contained in a dietary supplement offers a preventive effect against oxidative stress by promoting glutamine synthesis. This increase of intracellular glutamine synthesis results from intracellular metabolic pathway deviation and probably protein breakdown enhancement. Such Müller cell mechanisms and their consequences on the entire retina could explain the in vivo protective effect of dietary supplement in light-induced damage [16] and its beneficial effect in human patients affected by Age-related Macular Degeneration (AMD) [68]. Further studies will be conducted in order to evaluate the effect of sub-lethal oxidative stress on the glutamate/glutamine cycle in Müller cells cultured with lower or no glutamate in media and to quantify the variations of glutamate/glutamine cycles in retina from in vivo supplemented animals subjected to oxidative stress induced by light exposure. Figure 6. Schematic representation of nutritional supplementation impact on Müller cells' glutamine homeostasis wit without oxidative stress. The upper panels represent the glutamate/glutamine cycle without (−SC) and with (+SC) n tional supplementation in absence of oxidative stress. The lower panels present the glutamate/glutamine cycle with (−SC) and with (+SC) nutritional supplementation in presence of oxidative stress. EAAT1: glutamate transporter; S glutamine transporter; GLS: glutaminase; GS: glutamine synthetase. The asterisks (*) represent the new synthesis of tamate or glutamine induced by action of SC on intracellular pathways. The size of the characters or arrows in the figu images represent the significance of the process that they illustrate or the quantity of a protein or enzyme: the bigger size, the more significant the biological item is.  The lower panels present the glutamate/glutamine cycle without (−SC) and with (+SC) nutritional supplementation in presence of oxidative stress. EAAT1: glutamate transporter; SN1: glutamine transporter; GLS: glutaminase; GS: glutamine synthetase. The asterisks (*) represent the new synthesis of glutamate or glutamine induced by action of SC on intracellular pathways. The size of the characters or arrows in the figure's images represent the significance of the process that they illustrate or the quantity of a protein or enzyme: the bigger the size, the more significant the biological item is.