Consumption of a high fat diet promotes protein O-GlcNAcylation in mouse retina via NR4A1-dependent GFAT2 expression

Abstract

The incidence of type 2 diabetes, the most common cause of diabetic retinopathy (DR), is rapidly on the rise in developed countries due to overconsumption of calorie rich diets. Using an animal model of diet-induced obesity/pre-diabetes, we evaluated the impact of a diet high in saturated fat (HFD) on O-GlcNAcylation of retinal proteins, as dysregulated O-GlcNAcylation contributes to diabetic complications and evidence supports a role in DR. Protein O-GlcNAcylation was increased in the retina of mice fed a HFD as compared to littermates receiving control chow. Similarly, O-GlcNAcylation was elevated in retinal Müller cells in culture exposed to the saturated fatty acid palmitate or the ceramide analog Cer6. One potential mechanism responsible for elevated O-GlcNAcylation is increased flux through the hexosamine biosynthetic pathway (HBP). Indeed, inhibition of the pathway’s rate-limiting enzyme glutamine-fructose-6-phosphate amidotransferase (GFAT) prevented Cer6-induced O-GlcNAcylation. Importantly, expression of the mRNA encoding GFAT2, but not GFAT1 was elevated in both the retina of mice fed a HFD and in retinal cells in culture exposed to palmitate or Cer6. Notably, expression of nuclear receptor subfamily 4 group A member 1 (NR4A1) was increased in the retina of mice fed a HFD and NR4A1 expression was sufficient to promote GFAT2 mRNA expression and O-GlcNAcylation in retinal cells in culture. Whereas palmitate or Cer6 addition to culture medium enhanced NR4A1 and GFAT2 expression, chemical inhibition of NR4A1 transactivation repressed Cer6-induced GFAT2 mRNA expression. Overall, the results support a model wherein HFD increases retinal protein O-GlcNAcylation by promoting NR4A1-dependent GFAT2 expression.

1. Introduction

Diabetic retinopathy (DR) is a leading cause of vision impairment in developed countries, where the consumption of calorie rich diets is associated with the development of pre-/type 2 diabetes (1). Most people with type 2 diabetes develop some form of vision loss within two decades of disease onset (2). Moreover, the incidence of type 2 diabetes is rapidly on the rise, as it accounts for up to 95% of diabetes prevalence. Thus, type 2 diabetic patients comprise the largest proportion of the ~8 million Americans affected by DR. In humans with type 2 diabetes, early thinning of the inner retina and defects in vision occur before clinically visible signs of vascular pathology (3). Mirroring the pathophysiology in patients with type 2 diabetes, rodents fed a high fat diet (HFD) exhibit a slow onset of retinal complications, including retinal degeneration and impaired electroretinography (ERG) responses prior to the onset of microvascular disease (4–6). However, the mechanism underlying HFD-induced retinal dysfunction remains poorly understood (7,8). One potential mechanism whereby HFD contributes to retinal dysfunction is by promoting increased flux through the hexosamine biosynthetic pathway (HBP).

The HBP is a nutrient and stress sensing metabolic pathway responsible for production of uridine di phosphate-N-Acetylglucosamine (UDP-GlcNAc), which serves as the substrate for covalent addition of O-linked GlcNAc to the hydroxyl group of Ser/Thr residues (O-GlcNAcylation). Protein O-GlcNAcylation is one of the most abundant post-translational modifications within the nucleocytoplasmic compartment (9), however until recently its importance in regulating cell signaling pathways was largely unappreciated (10,11). The first and rate-determining step in the HBP involves the irreversible transfer of the amino group from glutamine and the isomerization of fructose-6-phosphate (F-6-P) to produce glucosamine-6-phosphate (GlcN-6-P) in a reaction that is catalyzed by the enzyme glutamine-fructose-6-phosphate amidotransferase (GFAT). Thus, GFAT isoform expression and activity represent a critical regulatory point in committing F-6-P to the HBP. Subsequent steps metabolize GlcN-6-P to UDP-GlcNAc, which serves as the donor for post-translational modification of proteins by O-GlcNAcylation. Protein O-GlcNAcylation is catalyzed by the enzyme O-GlcNAc transferase (OGT), whereas the enzyme O-GlcNAcase (OGA) catalyzes removal of the modification. These enzymes cycle GlcNAc residues on and off of proteins dynamically, in a manner that is more reminiscent of protein phosphorylation, rather than other more stable forms of protein glycosylation (12).

Augmented O-GlcNAcylation has been linked to the development of insulin resistance in type 2 diabetes (13,14), and contributes to the pathophysiology of diabetic complications (15). Moreover, evidence supports a role for defects in O-GlcNAc signaling in DR (16). O-GlcNAcylated proteins have been identified in the lens, cornea, retinal pigment epithelium, and neuroretina (13). Elevated O-GlcNAcylation of retinal proteins has been reported in diabetic Akita mice (17) and in mice with streptozotocin-induced diabetes (18). Similarly, the db/db type 2 diabetes mouse model also exhibits elevated O-GlcNAcylation of retinal proteins localized to the ganglion cell layer, retinal pigmented epithelium and inner pelxiform layers (19).

As with diabetes and hyperglycemia, HFD and increased oxidation of free fatty acids also promote flux through the HBP (20). In rats fed a HFD for 12-weeks, protein O-GlcNAcylation is elevated in the heart, liver (21), and cerebral arteries (22). Similarly, consumption of a Western diet high in saturated fat and sugar content enhances O-GlcNAcylation in rat heart in a manner that may contribute to the cardiac dysfunction associated with diabetes (23). Notably, these changes were not associated with a change in OGT or OGA. Alternatively, the saturated fatty acids palmitate and stearate have been found to increase GFAT1 mRNA and protein expression (23). This supports previous reports of increased GFAT activity and gene expression in the skeletal muscle of patients with type 2 diabetes (24). However, variation in relative tissue distribution suggests that GFAT2, rather than GFAT1, may be the primary GFAT isoform responsible for committing F-6-P to the HBP in retina. Whereas GFAT1 is more highly expressed in pancreas and liver, GFAT2 is the predominant isoform in the central nervous system (25).

The aim of the present study was to investigate the impact of a HFD on protein O-GlcNAcylation in the retina and evaluate the molecular mechanism responsible for the effect. In mice fed a HFD, retinal protein O-GlcNAcylation was elevated as compared to chow fed controls. Whereas GFAT1 mRNA expression was not different in the retina of mice fed a HFD as compared to control, GFAT2 mRNA expression was dramatically increased. Overall, the results support a model wherein HFD increases retinal protein O-GlcNAcylation through a mechanism that involves the lipotoxicity sensor nuclear receptor subfamily 4 group A member 1 (NR4A1, also known as NGF1B, TR3, and Nur77) and GFAT2.

2. Experimental procedures

2.1. Animals.

6-week-old C57BL/6J male mice were maintained on a 12:12-h reverse light dark cycle and fed ad libitum for 4 weeks with either a Teklad control chow (TD.08485) diet containing 13.0% kcal from fat, 67.9% from carbohydrates, and 19.1% from protein or a high fat diet (HFD, TD.95217) containing 39.7% kcal from fat, 41.4% kcal from carbohydrates, and 18.8% from protein (Envigo). Mice were fasted for 6 h prior to retinal extraction. Retinas were isolated and flash-frozen in liquid nitrogen, and lysates were prepared as previously described (26). Male Sprague-Dawley rats weighing ∼200 g were maintained on a 12:12-h light-dark cycle, with food (Harlan Teklad) and water provided ad libitum. Rat liver, kidney, retina, brain, and gastrocnemius were isolated and flash-frozen in liquid nitrogen, and lysates were prepared as previously described (26,27). Protein concentrations were assessed by DC™ Protein Assay (Biorad), and supernatants were combined with a 2×Laemmli buffer, boiled for 5 min, and analyzed via Western blotting. All experimental protocols used for the studies described herein were approved by the Institutional Animal Care and Use Committee (IACUC) of Penn State College of Medicine.

2.2. Cell Culture.

TR-MUL rat retinal Müller cells (provided by K. Hosoya, Toyama Medical and Pharmaceutical University) and MIO-M1 human Müller cells (obtained from the UCL Institute of Ophthalmology, London, UK) were maintained in DMEM (Gibco, #11885–084) containing 5.6 mM glucose, and supplemented with 10% heat inactivated (55°C, 30 min) fetal bovine serum (FBS) and 1% penicillin-streptomycin (Invitrogen). Human Umbilical Vein Endothelial Cells (HUVEC, Lonza) were maintained in Endothelial Growth Medium 2 (EGM-2, Lonza). HepG2 human liver cells (ATCC) were maintained in Eagle’s Minimum Essential Medium (EMEM) supplemented with 10% FBS and 1% penicillin-streptomycin (Invitrogen). Cells were maintained at either 33°C (TR-MUL) or 37°C (MIO-M1 and HepG2) and 5% CO₂ atmosphere. Transfections were performed using Lipofectamine 2000 (Life Technologies) according to the manufacturer’s instructions. Plasmids for expression of NR4A1 (Cat. # MR209316) and XBP1 (Myc-DDK-tagged ORF, Cat. # RR208206) were obtained from OriGene Technologies Inc (Rockville, MD, USA). For cell culture experiments, sodium palmitate (PAL) (P9767–5G; MilliporeSigma) was conjugated to 10% bovine serum albumin (BSA) (A7511–25G; Sigma-Aldrich) and added to cell culture medium at a final concentration of 0.5 mM. N-Hexanoyl-D-sphingosine (Cer6) (MilliporeSigma, H6524–5MG) was prepared in DMSO at 50 mM and added to cells at a final concentration of 60 μM. Where indicated cell culture medium was supplemented with GFAT inhibitors 6-diazo-5-oxonorleucine (DON, 50 μM) (D2141–5MG, MilliporeSigma) or Azaserine (Aza, 50 μM) (A4142, MilliporeSigma) for 1 h prior to Cer6 addition. The NR4A1 inhibitor DIM-C-pPhOH (DIM) (D7946–5MG, MilliporeSigma) was added to cell culture medium at a final concentration of 20 μM. Thapsigargin (TG) was used as an ER stress inducer at a concentration of 0.3 μM. At the end of the stimulation period, cells were carefully washed twice with cold PBS and harvested in 100 μl of lysis buffer containing 150 mmol/l NaCl, 10 mmol/l Tris, 1 mmol/l EGTA, 1 mmol/l EDTA (pH 7.4), 100 mmol/l NaF, 4 mmol/l sodium pyrophosphate, 2 mmol/l sodium orthovanadate, 1% Triton X-100, 0.5% NP-40-Igepal and a protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany) for Western blotting, or 500 uL Trizol (Invitrogen, Thermo Fisher Scientific, MA, USA) for RNA isolation. Lysates were centrifuged at 12,000×g for 15 min at 4 °C. The resulting supernatant fractions were recovered and stored at −80°C. Protein concentration was determined via DC™ Protein Assay and the lysates were combined with a 2×Laemmli buffer, boiled for 5 min, and analyzed via Western blotting.

2.3. Western Blot Analysis.

Lysates were fractionated using Criterion pre-cast 4–15% gels (Bio-Rad Laboratories, Inc; Hercules, CA, USA). Proteins were transferred to PVDF, blocked in 5% milk or 5% bovine serum albumin (BSA) in Tris buffered saline tween 20 (TBS-T) for 1 h, washed, and incubated overnight at 4 °C with the appropriate antibodies found in Table 1. Protein loading was assessed by MemCode Reversible Protein Staining. The antigen-antibody interaction was visualized with enhanced chemiluminescence (Clarity Reagent; Bio-Rad Laboratories, Inc; Hercules, CA) using a ProteinSimple Fluorochem E imaging system (Santa Clara, CA, USA). Blots were quantified using Image J software (NIH, Bethesda, MD, USA).

Table 1. Antibodies used for Western blotting.

Primary and secondary antibodies used for Western blotting are listed by manufacturer, catalogue number and lot number, molecular weight, and species of origin; antibody specificity to human (H), mouse (M), and rat (R) is also indicated.

Cell Signaling Technology	Cat #	Lot #	Molecular Mass (kDa)	Species of Origin	Specificity
O-GlcNAcylation	9875	4		Mouse	H, M, R
PERK-P	3179	19	170	Rabbit	M, R
PERK	3192	10	140	Rabbit	H, M, R
eIF2a-P	3398	6	38	Rabbit	H,M,R
eIF2a	9722	5	38	Rabbit	H, M, R
Protein Tech	Cat #	Lot #	Molecular Mass (kDa)	Species of Origin	Specificity
GFAT1	14132–1-AP	5163	77	Rabbit	H, M, R
Abcam	Cat #	Lot #	Molecular Mass(kDa)	Species of Origin	Specificity
GFAT2	ab190966	GR243495–2	77	Rabbit	H, M, R
Origene	Cat #	Lot #	Molecular Mass (kDa)	Species of Origin	Specificity
DDK (FLAG tag)	TA5011–100	A043		Mouse	M
Santa Cruz	Cat #	Lot #	Molecular Mass (kDa)	Species of Origin	Specificity
GAPDH	sc-32233	K0315	37	Mouse	H, M, R
α-tubulin	sc-32293	C0112	55	Mouse	H, M, R
Bethyl Laboratories	Cat #	Lot #	Molecular Mass (kDa)	Species of Origin	Specificity
Secondary antibody	A120–101P	40		Goat	R
Secondary antibody	A90–116P	38		Goat	M

2.4. RNA isolation and quantitative real-time PCR.

Total RNA was extracted with TRIzol reagent according to the manufacturer’s instructions (Invitrogen). RNA (1 μg) was reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) and subjected to quantitative real-time PCR (QuantStudio 12K Flex Real-Time PCR System) using QuantiTect SYBR Green master mix (Qiagen, Germantown, MD, USA). Mouse and rat primers used for amplification of GAPDH mRNA were previously described (28,29). The other primers used are listed in Table 2. Mean cycle threshold (C_T) values for REDD1 and GAPDH were determined for control and experimental samples. Changes in GFAT1, GFAT2, and NR4A1 mRNA expression, as well as XBP1 processing (XBP1s), were normalized to GAPDH mRNA expression using the 2^−∆∆CT calculations as previously described (30).

Table 2. Primers used for RT-PCR.

Primer specificity to human (H), mouse (M), and rat (R) is indicated.

Gene	Species		Primers
GFAT1	R	Forward	GCCAGCGACTTCTTGGATAG
	R	Reverse	AAGACCCATCAGGGTGTCAG
GFAT2	R	Forward	CCACAGCGTTCAGACAAAGA
	R	Reverse	CTCAAACTCGTAGCCCTTGC
GFAT1	M	Forward	TAAGGAGATCCAGCGGTGTC
	M	Reverse	CAGCTGTCTCGCCTGATTGA
GFAT2	M	Forward	CGCCTTTGCACTGGTTTTCA
	M	Reverse	TTCGTACCCCGATGAGCAAG
GFAT1	H	Forward	GCAAGCAGTTGGCACAAGG
	H	Reverse	CTCCACTGCTTTTTC TTCCAC
GFAT2	H	Forward	GGAGTCCGGAGCAAATACAAT
	H	Reverse	GACCCG GTGAATGGAGAG
GAPDH	H	Forward	GGTGGTCTCCTCTGACTTCAACA
	H	Reverse	GTTGCTGTAGCCAAATTCGTTGT
XBP1s	H/R/M	Forward	GGTCTGCTGAGTCCGCAGCAGG
	H/R/M	Reverse	GAAAGGGAGGCTGGTAAGGAAC
NR4A1	M	Forward	TCTGCCTTCCTGGAACTCTTCA
	M	Reverse	CAGGCCTGAGCAGAAGATGAG

2.5. Statistical Analysis

Data are expressed as Mean ± SEM. Student’s t-test or one-way ANOVA was used to analyze signaling data. Fischer’s LSD test was used to identify specific differences if a significant overall F-value was observed following one-way ANOVA. Relationships were determined by Pearson product moment correlation analysis. Significance was set at p<0.05 for all analyses.

3. Results

3.1. Relative GFAT isoform expression is tissue and cell type specific.

Relative expression of GFAT1 and GFAT2 mRNA was assessed in rat retina, liver, kidney, brain, and muscle (Fig. 1A). The ratio of GFAT1:GFAT2 mRNA was variable, as liver exhibited the highest and muscle the lowest relative expression of GFAT1 to GFAT2 (Fig. 1B). The ratio of GFAT1:GFAT2 mRNA in rat retina was approximately one-third of that observed in liver. We further compared relative GFAT isoform expression in three human cell lines: MIO-M1 retinal Müller cells, HUVEC endothelial cells, and HepG2 liver cells (Fig. 1C). In support of the observation in rat tissue, the ratio of GFAT1:GFAT2 mRNA in MIO-M1 and HUVEC cells was markedly lower compared to HepG2 cells (Fig. 1D).

Figure 1. Relative GFAT isoform expression is tissue and cell type specific.

A. GFAT1 and GFAT2 mRNA expression was assessed by RT-PCR in rat liver, kidney, retina, brain, and gastrocnemius muscle and is expressed relative to GAPDH (n=5). B. Relative ratios of GFAT1 to GFAT2 mRNA in various rat tissues from A. C. GFAT1 and GFAT2 mRNA expression was assessed by RT-PCR in human retinal MIO-M1, endothelial HUVEC, and liver HepG2 cells in culture (n=3). D. Relative ratios of GFAT1 to GFAT2 mRNA in human cell lines from C. Results are expressed as means ± SEM. Statistical significance is denoted in B. and D. by “*” (p<0.05) versus either liver or HepG2 cells, respectively.

3.2. Consumption of a high-fat diet promotes retinal protein O-GlcNAcylation and GFAT2 mRNA expression.

Protein O-GlcNAcylation was elevated in the retina of mice fed a HFD for 4 weeks as compared to the retina of mice fed a control chow diet (Fig. 2A, B). After 4-weeks of HFD, neither body weight (Fig 2C, p=0.068) nor blood glucose concentrations (Fig 2D) were significantly elevated. One potential mechanism whereby a HFD promotes retinal protein O-GlcNAcylation is by up-regulating GFAT expression. Whereas consumption of a HFD for 4 weeks did not alter GFAT1 mRNA expression in retina (Fig. 2E), GFAT2 mRNA expression was elevated as compared to mice receiving a control chow diet (Fig. 2F).

Figure 2. Protein O-GlcNAcylation is increased in the retina of mice fed a high-fat diet (HFD) concomitant with increased GFAT2 mRNA expression.

A. Protein O-GlcNAcylation was assessed by Western blotting of retinal lysates from mice fed either a control chow (Chow) or HFD for 4 weeks. Protein molecular mass (kDa) is indicated at the right of blots. B. Retinal protein O-GlcNAcylation in A. was quantified and is expressed relative to GAPDH. Representative blots are shown. Protein loading was assessed by protein staining (Protein S.). C. Change in body weight is expressed as a percentage body weight prior to initiating the 4-week diet. D. Nonfasting blood glucose concentrations were assessed the day before retinal extraction. GFAT1 (E.) and GFAT2 (F.) mRNA expression were assessed in the retina of mice fed a HFD for 4 weeks by RT-PCR. Results are expressed as means ± SEM (n=5). Statistical significance is denoted by “*” (p<0.05) or “**” (p<0.01), versus control chow diet.

3.3. Palmitate and Cer6 promote protein O-GlcNAcylation and upregulated GFAT2 expression in retinal Müller cells in culture.

To model HFD-induced effects in vitro, we exposed TR-MUL retinal Müller cells in culture to palmitate, the most common saturated fatty acid found in the HFD. Similar to the effect of a HFD on the retina, palmitate enhanced retinal protein O-GlcNAcylation (Fig. 3A). Whereas GFAT1 protein expression was similar in cells exposed to palmitate as compared to BSA vehicle alone, GFAT2 protein expression was increased (Fig. 3A). Following palmitate addition to culture medium, GFAT1 mRNA expression was decreased (Fig. 3B); however, GFAT2 mRNA expression was increased (Fig. 3C). Notably, GFAT1 mRNA expression was relatively low in Muller cells in culture as compared to GFAT2 (Fig. 1). Thus, the reduction in GFAT1 mRNA expression following palmitate exposure may represent a small variation in the absolute abundance. We recently demonstrated that diets rich in saturated fatty acids promote levels of retinal sphingolipids including ceramides (31). Ceramide is considered to be the central hub of sphingolipid metabolism as it can be converted to other interconnected bioactive lipid species (32). Thus, cells were treated with a cell-permeable and biologically active ceramide analog Cer6 to further evaluate the effects in vitro. In TR-MUL cells in culture, protein O-GlcNAcylation was enhanced by exposure to Cer6 for 16 h (Fig. 3D). Similarly, compared to DMSO vehicle, Cer6 exposure increased GFAT2 protein and mRNA expression without affecting GFAT1 (Fig. 3D-F). To provide further support for the role of GFAT2 in Cer6-induced protein O-GlcNAcylation, we employed two GFAT chemical inhibitors, DON and Azaserine. As expected, either DON or Azaserine abolished Cer6-induced protein O-GlcNAcylation (Fig. 3G, H), indicating that Cer6 enhanced protein O-GlcNAcylation through a GFAT-dependent mechanism.

Figure 3. Palmitate (PAL) and ceramide (Cer6) promote protein O-GlcNAcylation and increase GFAT2 expression in retinal cells in culture.

A-C. TR-MUL cells were exposed to culture medium containing either BSA or palmitate bound to BSA (PAL, 0.5 mM) for 8 h (n=3). Protein O-GlcNAcylation, GFAT1, GFAT2, and GAPDH protein expression were assessed by Western blotting in A. and D. Representative blots are shown. Protein molecular mass (kDa) is indicated at the right of blots. Protein Staining (Protein S.) is shown as a loading control. GFAT1 (B. & E.) and GFAT2 (C. & .F). mRNA expression were assessed by RT-PCR. D. TR-MUL cells were exposed to culture medium containing the cell permeable ceramide Cer6 or DMSO vehicle (Veh) for up to 16 h (n=3). E-H. TR-MUL cells were treated with either Cer6 or Veh for 16 h. TR-MUL cells were exposed to culture medium containing GFAT inhibitors 6-diazo-5-oxonorleucine (DON) or Azaserine (Aza) for 1 h prior to Cer6 addition in G. and H., respectively. Results are expressed as means ± SEM (n=3). Statistical significance is denoted by “*” (p<0.05), “**” (p<0.01), or “***” (p<0.001), versus control treatment. Protein O-GlcNAcylation and GAPDH protein expression were assessed by Western blotting in G-H.

3.4. Effect of ER stress on O-GlcNAcylation and GFAT isoform expression in retinal Müller cells in culture.

One potential mechanism whereby HFD might act to increase GFAT2 mRNA expression is by increasing the transcription factor spliced X box-binding protein 1 (XBP1s), which is involved in ER stress-induced regulation of HBP gene expression (33). Moreover, more long-term feeding of HFD was recently associated with enhanced ER stress markers in retina (34). Thus, we initially suspected that ER stress may act to promote GFAT2 mRNA expression. To determine if ER stress was sufficient to enhance GFAT2 mRNA expression, we employed the ER stress inducer thapsigargin (TG). In retinal Muller cells exposed to TG, the induction of ER stress was confirmed by enhanced phosphorylation of PERK (protein kinase RNA-like endoplasmic reticulum kinase) and eIF2α (eukaryotic initiation factor 2α), as well as an increase in XBP1 processing (Fig. 4A and B, respectively). Following exposure to TG, protein O-GlcNAcylation was enhanced in TR-MUL cells in culture (Fig. 4A). However, unlike with PAL/Cer6 (Fig. 3), TG-induced O-GlcNAcylation was associated with an increase in GFAT1 mRNA expression (Fig. 4C). Except for being slightly elevated after 4 h, GFAT2 mRNA expression was suppressed 8 to 16 h after TG addition to culture medium (Fig. 4D). Thus, the effect of TG was not consistent with the increase in GFAT2 mRNA expression that was observed the retina of mice fed a HFD and in retinal cells in culture exposed to PAL/Cer6.

Figure 4. ER stress promotes O-GlcNAcylation and GFAT1 but not GFAT2 mRNA expression.

TR-MUL cells were exposed to culture medium containing the ER stress inducer thapsigargin (TG) for up to 16 h (n=3). A. Protein O-GlcNAcylation and phosphorylation of PERK and eIF2α were evaluated by Western blotting. Representative blots are shown. Protein molecular mass (kDa) is indicated at the right of blots. Protein Staining (Protein S.) is shown as a loading control. XBP1 processing (B.) as well as GFAT1 (C.) and GFAT2 (D.) mRNA expression were assessed by RT-PCR. Results are expressed as means ± SEM (n=3). Statistical significance is denoted by the presence of different letters above bars on the graphs. Bars with different letters are statistically different; p<0.05. Results are representative of two experiments; within each experiment, three independent samples were analyzed.

3.5. Expression of the lipotoxicity sensor NR4A1 is increased in the retina of mice fed a HFD and in retinal Müller cells exposed to PAL or Cer6.

In a recent study (35), the lipotoxicity sensor NR4A1 was found to promote HBP flux via transcriptional upregulation of GFAT2. Thus, HFD, as well as PAL and Cer6 exposure, might act to increase GFAT2 mRNA expression via NR4A1. Indeed, NR4A1 mRNA expression was increased in retina of mice fed an HFD for 4 weeks as compared to control chow (Fig. 5A). Similarly, in TR-MUL cells in culture, NR4A1 mRNA expression was up-regulated by exposure to PAL (Fig. 5B) or Cer6 (Fig. 5C).

Figure 5. NR4A1 mRNA expression is increased in the retina of mice fed a HFD and in retinal cells exposed to PAL or Cer6.

A. NR4A1 mRNA expression was assessed in the retina of mice fed either a control chow (Chow) or HFD for 4 weeks (n=6). B. TR-MUL retinal cells were exposed to culture medium containing either BSA or palmitate bound to BSA (PAL) for 8 h (n=3). C. TR-MUL cells were exposed to culture medium containing the cell permeable ceramide Cer6 for up to 16 h (n=3). NR4A1 mRNA expression was assessed by RT-PCR. Results are expressed as means ± SEM. Statistical significance is denoted by the presence of different letters above bars on the graphs. Bars with different letters are statistically different; p<0.05. Results in B & C are representative of two experiments; within each experiment, three independent samples were analyzed.

3.6. NR4A1 promotes GFAT2 mRNA expression in retinal Müller cells in culture.

To test the hypothesis that increased NR4A1 is sufficient to promote GFAT2 mRNA expression and protein O-GlcNAcylation, we expressed NR4A1 in TR-MUL cells. Following transfection with an NR4A1 expression plasmid, NR4A1 protein (Fig. 6A) and mRNA expression (Fig. 6B) were markedly up-regulated as compared to an empty vector control. There was no change in GFAT1 protein (Fig 6A) or mRNA expression (Fig. 6C) in cells expressing NR4A1. However, GFAT2 protein and mRNA expression were significantly increased by NR4A1 (Fig. 6A and D, respectively). Furthermore, NR4A1 expression in TR-MUL cells enhanced protein O-GlcNAcylation, and chemical inhibition of GFAT by DON prevented the effect (Fig. 6E). To further evaluate the role of NR4A1 in regulating GFAT2 mRNA expression, we employed DIM, a NR4A1 antagonist that blocks transactivation of the receptor. As previously demonstrated, exposure to Cer6 increased NR4A1 and GFAT2 mRNA expression without affecting GFAT1 mRNA expression (Fig. 6F-H). As expected, DIM addition did not alter NR4A1 expression (Fig. 6F), as the inhibitor acts by preventing activation of the receptor. Importantly, DIM addition suppressed Cer6-induced GFAT2 mRNA expression (Fig. 6H). Thus, NR4A1 is not only sufficient to promote GFAT2 mRNA expression, but inhibition of NR4A1 transactivation also repressed Cer6-induced GFAT2 expression. This is consistent with a model wherein NR4A1 acts to promote GFAT2 expression, enhance HBP flux, and increase retinal protein O-GlcNAcylation (Fig. 6I).

Figure 6. NR4A1 promotes GFAT2 expression in retinal cells in culture.

A-E. TR-MUL cells were transfected with either an empty vector (EV) or NR4A1 expression plasmid for 48 h. A. FLAG-tagged NR4A1, GFAT1, GFAT2, and GAPDH protein expression were assessed by Western blotting. Protein molecular mass (kDa) is indicated at the right of blots. Protein Staining (Protein S.) is shown as a loading control. NR4A1 (B.), GFAT1 (C.), and GFAT2 (D.) mRNA expression were assessed using RT-PCR. E. TR-MUL cells were transfected with either EV or NR4A1 plasmid for 48 h prior to treatment with the GFAT inhibitor DON for 16 h. Protein O-GlcNAcylation and GAPDH expression were measured by Western blotting. F-H. TR-MUL cells were exposed to culture medium containing either DMSO (Veh) or Cer6 for 16 h in the presence or absence of the NR4A1 inhibitor DIM-C-pPhOH (DIM). NR4A1 (F.), GFAT1 (G.), and GFAT2 (H.) mRNA expression were assessed by RT-PCR. Results are expressed as means ± SEM (n=3). Statistical significance is denoted by the presence of different letters above bars on the graphs. Bars with different letters are statistically different; p<0.05. Results are representative of two experiments; within each experiment, three independent samples were analyzed. I. Working model for mechanism whereby high fat diet (HFD) promotes retinal O-GlcNAcylation via enhanced NR4A1 and GFAT2 expression.

4. Discussion

4.1. Consumption of a HFD promoted retinal protein O-GlcNAcylation and increased GFAT2 mRNA expression.

In the present study, we evaluated the impact of a diet high in saturated fats on O-GlcNAcylation of retinal proteins. In an attempt to focus on early biochemical changes in retina associated with consumption of a HFD, we have evaluated O-GlcNAcylation and gene expression changes prior to the development of obesity and type 2 diabetes. We found that in both the retina of mice fed a HFD and in retinal cells in culture exposed to the saturated fatty acid palmitate, protein O-GlcNAcylation was elevated concomitant with upregulated expression of GFAT2 mRNA. We recently demonstrated that in mice fed a diet high in saturated fatty acids, and specifically palmitate, the abundance of retinal sphingolipids, including ceramides, were elevated (31). In the present study, we found that addition of the cell permeable ceramide analog Cer6 to retinal cells in culture also promoted GFAT2 mRNA expression and increased O-GlcNAcylation, and that the effect of Cer6 on O-GlcNAcylation was dependent on GFAT activity. Importantly, in both the retina of mice fed a HFD and in retinal cells exposed to palmitate or Cer6, the increase in O-GlcNAcylation was not consistent with a change in GFAT1, but rather in GFAT2.

In agreement with a previous study (25), we found that relative GFAT isoform expression was tissue and cell type specific. In that study, GFAT1 was the dominant isoform expressed in the pancreas and liver, whereas GFAT2 was more highly expressed in the brain and spinal cord. In the present study, the relative ratio of GFAT1 to GFAT2 was lower in brain and retina as compared to liver. Similarly, the relative expression of GFAT2 as compared to GFAT1 in human MIO-M1 retinal Müller cells was dramatically higher than that observed in human HepG2 liver cells in culture. Importantly, the retina is composed of numerous unique cell types with widely distinct functions and corresponding gene expression profiles. Thus, it is not known if the HFD-induced GFAT2 gene expression changes observed in whole retinal lysates are reflective of a specific retinal cell type.

Notably, the enzymatic activity of GFAT1 and GFAT2 appear to be differentially regulated by protein phosphorylation. GFAT1 is phosphorylated by cyclic AMP dependent protein kinase (PKA) at Ser205 in a manner that blocks enzymatic activity (36). GFAT2 also has a PKA site at Ser202 that is homologous to Ser205 in GFAT1, however PKA phosphorylation of this site increases GFAT2 enzymatic activity (37). Similarly, GFAT1 is also more sensitive to feedback inhibition from UDP-GlcNAc (37). Thus, due to tissue specific distribution of GFAT1 and GFAT2 isoforms (25), the discrepancy in GFAT isoform regulation allows various tissues to metabolize glucose differently in response to hormonal stimulation. For example, following a meal the secretion of incretins like glucagon-like peptide-1 and enhanced sensitivity to product inhibition will lead to reduced GFAT1 activity in the pancreas and liver, whereas GFAT2 would continue to commit glucose to the HBP in the brain and retina.

4.2. ER stress promoted GFAT1 mRNA expression and increased O-GlcNAcylation

Wang et al (33) previously demonstrated that ER stress and the unfolded protein response promotes O-GlcNAcylation via Xbp1-dependent transcription of genes that control flux through the HBP. Thus, we initially suspected that HFD-induced ER stress may also regulate O-GlcNAcylation in the retina. Consistent with the previous report by Wang et al (33), ER stress induced O-GlcNAcylation was associated with increased GFAT1 mRNA expression. However, in the retina of mice fed a HFD and in retinal cells exposed to palmitate/Cer6 GFAT2 mRNA expression was increased, whereas GFAT1 mRNA levels were relatively unchanged. Thus, unlike HFD or with palmitate- and Cer6-induced O-GlcNAcylation, the effect of thapsigargin on O-GlcNAcylation in retinal cells in culture was not consistent with a GFAT2-dependent mechanism.

4.3. NR4A1 as a mechanism of HFD-induced retinal protein O-GlcNAcylation

In the present study, we found that expression of the lipotoxicity sensor NR4A1 was increased in the retina of mice fed a HFD. NR4A1 is an immediate-early response gene and orphan member of the nuclear receptor family that have no known endogenous ligands (38). The NR4A subfamily of nuclear receptors are important regulators of metabolic function and have been implicated in a variety of metabolic diseases including obesity and type 2 diabetes [reviewed in (39)]. NR4A1 is composed of an N-terminal transactivation domain, a zinc-finger DNA-binding domain, and a C-terminal ligand binding domain (40). Remarkably, NR4A1 has no apparent cavity in the canonical ligand-binding pocket, as it is filled with bulky hydrophobic residues (41). However, there is some evidence that unsaturated fatty acids associate with the ligand binding domain of NR4A1 to regulate conformation and oligomerization of the receptor (42). Nevertheless, NR4A1 is constitutively active and acts at least in part independent of ligand binding (38). Thus, it is believed that the activity of NR4A1 is predominantly regulated through changes in its expression (39).

While the bona fide NR4A1 ligand remains to be established, several small molecular inhibitors have been recently identified for the NR4A family of receptors (43). In the present study, we used a derivative of diindolymethane that prevents transactivation of the receptor (44). While the inhibition of NR4A1 by these compounds is dependent on the NR4A1 ligand binding domain, it is not clear if they directly bind in the ligand-binding cavity of NR4A1 or with a secondary molecule that interacts through the NR4A1 ligand-binding domain (45). Regardless of the mechanism of action, we found that chemical inhibition of NR4A1 transactivation was sufficient to attenuate Cer6-induced GFAT2 mRNA expression in retinal TR-MUL cells in culture. GFAT2 mRNA expression was enhanced in neonatal rat ventricular myocytes following adenoviral overexpression of NR4A1 (35). In agreement with the previous report, NR4A1 was sufficient to increase GFAT2 mRNA expression in retinal TR-MUL cell in culture. Further, NR4A1 increased O-GlcNAcylation in TR-MUL cells in a manner that was prevented by GFAT inhibition.

NR4A1 binds to DNA as a monomer, homodimer, and even as a heterodimer with other NR4A family members or retinoid x receptor (46). As a monomer, NR4A1 binds to the Nur77-binding response element (NBRE; AAAGGTCA) (38). As a homo- or hetero-dimer with other NR4A family members, NR4A1 binds to the Nur-responsive element [AAAT(G/A)(C/T)CA] (46). Previous characterization of the mouse GFAT2 gene promoter identified several putative regulatory elements including two Sp1 sites that were required for basal promoter activity (47). Intriguingly, a region from −243 to −253 of the mouse GFAT2 promoter (AAATTCCA) bears sequence homology to the Nur-response element. Although it is important to note that in the previous study by Lehmann et al (35), GFAT2 mRNA expression was also increased by a variant of NR4A1 wherein the DNA-binding domain was disrupted. This suggests that NR4A1 may act indirectly through a non-genomic mechanism to promote GFAT2 mRNA expression and increase HBP flux.

4.4. Conclusion

Current therapeutic options for DR largely address the neovascularization and microvascular dysfunction that characterizes later proliferative stages of disease progression. However, neuro-retinal deficits can clearly precede and even predict the visible signs of microvascular disease in diabetic patients [reviewed in (48)]. Importantly, clinical studies by multiple investigators have reported retinal defects (49,50) and impaired visual function in pre-diabetic humans with impaired glucose tolerance (51,52). Thus, there is an urgent need for interventions that target the initiating biochemical changes that contribute to the early non-proliferative stages of disease progression. While there is previous evidence that retinal O-GlcNAc signaling is impaired in both type 1 and type 2 diabetes (16), the present study suggests that increased O-GlcNAcylation in retina may precede the development of diabetes. Overall, the present study supports a model wherein consumption of a HFD increases retinal protein O-GlcNAcylation by promoting NR4A1-dependent GFAT2 expression and increased HBP flux. Thus, pharmacological targeting of NR4A1 or isoform specific inhibition of GFAT2 may represent novel therapeutic interventions to address increased O-GlcNAcylation in the retina of diabetic/pre-diabetic patients.

Highlights.

• Protein O-GlcNAcylation was increased in the retina of mice fed a high fat diet

• Consumption of a high fat diet promoted retinal GFAT2 and NR4A1 mRNA expression

• Palmitate and Cer6 promoted O-GlcNAcylation and increased GFAT2/NR4A1 expression

• NR4A1 promoted GFAT-dependent O-GlcNAcylation and increased GFAT2 expression

• NR4A1 inhibition attenuated Cer6-induced GFAT2 mRNA expression

Abbreviations:

Aza

azaserine

BSA

bovine serum albumin

Cer6

N-hexanoyl-D-sphingosine

DIM

diindoylmethane-C-pPhOH

DON

6-diazo-5-oxonorleucine

DR

diabetic retinopathy

eIF2

eukaryotic initiation factor 2

EMEM

Eagle’s Minimum Essential Medium

ERG

electroretinography

F-6-P

fructose-6-phosphate

FBS

fetal bovine serum

GFAT

glutamine-fructose-6-phosphate amidotransferase

GlcN-6-P

glucosamine-6-phosphate

HBP

hexosamine bioshythetic pathway

HFD

high fat diet

NR4A1

nuclear receptor subfamily 4 group A member 1

OGA

O-GlcNAcase

OGT

O-GlcNAc transferase

PAL

palmitate

PERK

protein kinase RNA-like endoplasmic reticulum kinase

PKA

cyclic AMP dependent protein kinase

TG

thapsigargin

UDP-GlcNAc

uridine diphosphate-N-Acetylglucosamine

XBP1

X box-binding protein 1