O-GlcNAc stimulation: A new metabolic approach to treat septic shock

Septic shock is a systemic inflammation associated with cell metabolism disorders and cardiovascular dysfunction. Increases in O-GlcNAcylation have shown beneficial cardiovascular effects in acute pathologies. We used two different rat models to evaluate the beneficial effects of O-GlcNAc stimulation at the early phase of septic shock. Rats received lipopolysaccharide (LPS) to induce endotoxemic shock or saline (control) and fluid resuscitation (R) with or without O-GlcNAc stimulation (NButGT–10 mg/kg) 1 hour after shock induction. For the second model, rats received cecal ligature and puncture (CLP) surgery and fluid therapy with or without NButGT. Cardiovascular function was evaluated and heart and blood samples were collected and analysed. NButGT treatment efficiently increased total O-GlcNAc without modification of HBP enzyme expression.Treatment improved circulating parameters and cardiovascular function in both models, and restored SERCA2a expression levels. NButGT treatment also reduced animal mortality. In this study, we demonstrate that in septic shock O-GlcNAc stimulation improves global animal and cardiovascular function outcomes associated with a restoration of SERCA2a levels. This pre-clinical study opens avenues for a potential therapy of early-stage septic shock.

to decrease (e.g.: DON and Azaserine) or increase (e.g.: glucosamine, PUGNAc, NButGT) the level of protein O-GlcNAcylation. Whilst protein O-GlcNAcylation modulation seems to cause different effects in acute and chronic pathologies 7 , there is little evidence to suggest that a global increase in this post-translational modification could become a future therapeutic avenue to treat septic shock. In trauma-haemorrhage situations, glucosamine or PUGNAc administered during reperfusion increase survival of animals through an improvement in inflammatory responses and organ function 8 . More recently, Hwang et al. demonstrated in a septic shock model that pre-treatment with glucosamine improves survival thanks to a decrease in the inflammatory state in mice 9 .
In this context, we propose that an increase in protein O-GlcNAcylation by NButGT post-treatment could improve the outcome of septic shock in an animal model., We demonstrate that despite no improvement in the inflammatory state, increased protein O-GlcNAcylation improves global outcome in our septic shock rats. In our opinion, this treatment resulted in amelioration of cardiac function.

Results
NButGT efficiently increases total O-GlcnAc in septic shock without any consequence on major enzymes involved in the hexosamine biosynthetic pathway. In septic shock rats (LPS and CLP) the level of O-GlcNAcylated proteins and HBP enzymes (OGT and OGA) were not modified when compared to control or Sham rat samples (Fig. 1), but an increase in GFAT, the limiting rate enzyme (p = 0.0210 vs Sham Fig. 2B) was observed in the CLP model. Except for OGA, which is slightly increased, neither OGlcNAc levels nor OGT and GFAT expression were modified by fluid resuscitation. As expected, NButGT treatment, an OGA blocker, efficiently increased total O-GlcNAcylation of proteins 2 fold in the LPS model and by 2.5 fold in the CLP model (Figs. 1A and 2A). In these two models, this increase was not associated with modification of the three main enzymes regulating O-GlcNAc levels.

O-GlcNAc stimulation did not impact inflammatory responses.
The evaluation of the inflammatory state of septic shock animals was measured on cardiac (RT-qPCR) and plasma (ELISA) samples (Fig. 3). LPS resulted in significant increases in pro-inflammatory IL-6 and TNFα cytokines in both analyses. However, fluid resuscitation alone or completed with NButGT treatment did not modify the inflammatory response of animals.
While fluid resuscitation alone was not associated with improvement of these parameters, in association with NButGT treatment it normalized circulating parameters such as HCO 3 − (p < 0.001 vs LPS + FR) and pCO 2 (p < 0.001 vs LPS + FR). Furthermore, NButGT treatment induced a normalization of lactate (3.02 ± 0.27 mmol/L, p < 0.001 vs LPS + R), creatinine (33.7 ± 4.7 µmol/L, p = 0.001 vs LPS + R) and troponin T (33.6 ± 11.6 ng/L, ns vs LPS + R) indicating improved organ function in NButGT animals. Finally, as shown in O-GlcnAc stimulation improves cardiovascular function in an endotoxemic rat model. To explore the putative beneficial effect of O-GlcNAc stimulation, our next step was to investigate cardiovascular function. As shown in Fig. 5A,B, endotoxemic shock induced by iv injection of LPS lead to hemodynamic alterations especially tachycardia. LPS-treated rats also presented systolic dysfunction (LVEF, 80.7 ± 2.1 in control vs 65.1 ± 2.9% in LPS, p < 0.001) and delayed relaxation (E/E′ ratio, 25.4 ± 2.9 in control vs 17.0 ± 1.34 in LPS, p < 0.05) evaluated by echocardiography (Fig. 5C,D). Fluid resuscitation efficiently improved mean arterial pressure (similar to the control group) without any impact on heart function.
NButGT supplementation in the fluid resuscitation to stimulate protein O-GlcNAcylation did not further impact the mean arterial pressure (similar to the control and LPS + R group). However, it slightly restored cardiac function with an improvement in systolic function (LVEF, 75.0 ± 1.7%, p < 0.05 vs LPS) and cardiac relaxation (E/E′, 23.0 ± 2.0, p = 0.090 vs LPS).
CLP rats also showed an increase in respiratory rate (54.1 ± 3.0 vs 44.9 ± 1.7 rpm in Sham, p < 0.05), in lactatemia  In this model, genes and proteins involved in autophagy were also evaluated in hearts, however these were not modified by NButGT treatment (Fig. 8).

Discussion
This study was designed to evaluate the therapeutic potential of O-GlcNAc stimulation as an early post-treatment for septic shock. Our results demonstrate that an increase in protein O-GlcNAcylation with NButGT, an inhibitor of OGA, improved global outcome of animals in two different models of septic shock. O-GlcNAc stimulation could be a valuable therapeutic strategy in clinical septic shock management. www.nature.com/scientificreports www.nature.com/scientificreports/ In this study, both LPS and CLP models produced a septic shock with organ dysfunction and metabolic perturbations seen at the plasma level (creatinine, troponin T and/or lactate), and cardiovascular dysfunction with pronounced tachycardia, low hypotension, and both systolic and diastolic dysfunction. Thus, according to the latest definition, our animals developed septic shock 2,10 .
Increase in protein O-GlcNAcylation with NButGT is not associated with a modification of HBP enzyme expression. Results from our study are based on two models of shock, an LPS-induced model, which develops a rapid and reproducible endotoxemic shock (3 h), and a CLP model which develops a longer term poly-microbial septic shock (24 h). In these two models, septic shock leads to an increase in GFAT expression but did not modify (i) the expression of OGT and OGA, the two direct enzymes involved in O-GlcNAc level regulation and (ii) the global protein O-GlcNAcylation in cardiac tissue. NButGT administration induced an increase in global protein O-GlcNAcylation in heart tissue but had no impact on the expression of OGA, OGT and GFAT. The relationship between O-GlcNAc levels and the pathological state is not clearly identified. Whilst our results are in accordance with previous studies demonstrating that intracellular O-GlcNAc levels are modulated by stress, regulation of GFAT, OGA and OGT may take a longer. For example, significant increases in intracellular glucose concentrations did not modify GFAT gene or protein expression 11,12 . In parallel, expression of OGT seems to be poorly sensitive to O-GlcNAc modulation 13,14 . However, GFAT and OGT activities are regulated by UDP-GlcNAc, GlcN-6-P and O-GlcNAc moieties 13,[15][16][17] . In vitro, a long term augmentation of O-GlcNAc levels with OGT adenovirus transfection or treatment with Thiamet G (10 µM) induced an increase in OGA gene and protein expression 14,15,18 . Collectively, these results suggest that the modulation of OGA requires a longer period of time.
Currently, iv fluid resuscitation in association with antibiotic therapy is the gold standard for the treatment of septic shock in patients 19 . As expected, in our study, fluid resuscitation with a bolus of colloid improved mean arterial pressure. However, the circulating markers of acidosis and organ dysfunction were not improved with fluid resuscitation only. Fluid resuscitation therapy has been recently discussed as it may have adverse effects on renal function 20 , with a potential detrimental effect in the long term for patients suffering from septic shock. Results from our study suggest a new interrogation of fluid resuscitation therapy is warranted. Interestingly, the addition of NButGT to the fluid resuscitation treatment in LPS-treated rats efficiently normalized pH and bicarbonates. This therapeutic approach also normalizes circulating markers of organ function such as lactate and troponin T, and improves survival time. Moreover, we observed a significant increase in plasma creatinine levels in LPS-treated rats suggesting acute kidney injury. Treatment with NButGT normalizes this plasma creatinine level in our two models suggesting a potentially beneficial effect on renal function. Additional physiological, biochemical and histological analyses would be necessary to validate this hypothesis. In this model, NButGT also improved www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ cardiac leading to an early improvement in the animal status. The beneficial effects of NButGT were also confirmed in our CLP-induced septic shock model, whereby cardiovascular function and respiration were improved. This observation is in accordance with other studies which also demonstrated the beneficial effects of O-GlcNAc stimulation in shock. In cardiac ischemia-reperfusion, after acute renal injury or during haemorrhagic shock, O-GlcNAc stimulation (by glucosamine, PUGNAc, siRNA against OGA) improved cardiac or renal function as well as increasing animal survival 8,[21][22][23][24][25] . Among the mechanisms underlying these effects, the authors concluded that there were improvements in the inflammatory response, autophagy regulation and/or calcium overload. Interestingly in our model, O-GlcNAc stimulation improves survival, circulating parameters and cardiac function without any impact on the inflammatory response. In our study, IL-6 and TNF-α were elevated 3 hours after shock induction, and treatment with fluid resuscitation with or without NButGT had no impact. Interestingly, acute increase in O-GlcNAc level at a later stage of the pathology in other models of heart failure was associated with a significant decrease in inflammatory markers 8 . In a recent study, Baudoin et al. defined the relationship between O-GlcNAcylation and inflammation as a 'vast territory to explore' 26 . In fact, the role of OGlcNAc on inflammation is controversial. In previous studies of septic shock 9 or inflammatory models such as in trauma-haemorrhage 27 , the authors demonstrated a reduction in plasma cytokines (IL-6 or TNF-α). Nevertheless, this elevated inflammatory response without improvement attributable to NButGT is concordant with the fact that the pro-inflammatory response begins 30 minutes after LPS injection and is maintained for at least 24 hours 28,29 . Other studies have evaluated the impact of increased protein O-GlcNAcylation at a later stage and early NButGT treatment might have an effect at a later stage of the pathology.
Autophagy is considered as a biological quality control mechanism, and it could result in massive self-degradation or accumulation of toxic material in septic shock. Reducing autophagy has been shown to improve cardiac contractility 30 . Based on this observation, we hypothesized that improvements observed in NButGT-treated rats could partially be explained by a modulation of autophagy. Indeed, the increase in protein O-GlcNAcylation has been shown to blunt the autophagic pathway through a decrease in Beclin-1 or LC3II protein expression 31,32 . However, in our study there was no difference in Beclin-1 or LC3 gene or protein expression in control and LPS-treated rats. According to Sun et al., the expression of these autophagic regulators depends on the time and dose of LPS administration 33,34 . Therefore, the effect of NButGT on the autophagic pathway cannot be linked to the beneficial effect of NButGT treatment under our experimental conditions. www.nature.com/scientificreports www.nature.com/scientificreports/ Cardiac improvements are also associated with improved calcium regulator expression. Considering the main role of cardiac dysfunction in the septic shock response, we investigated calcium homeostasis, a main controller of cardiac contractility, in the LPS rat model only. Whilst many studies describe an alteration of the calcium cycle at the late stage of septic shock, very few studies are available at the early phase 35 . These studies classically show a decreased or unchanged SERCA2 function during the later phase (6-20 h after endotoxemic shock in mouse and rat models), and Morse et al. showed an increase in SERCA2 activity at the recovery phase of septic shock 36 . In our study, SERCA2a was investigated at the early phase (3 h) of septic shock and SERCA2a protein levels appear to be increased. Interestingly, NButGT treatment re-established SERCA2a protein expression. We postulated that in our model, SERCA2a overexpression could be a compensatory mechanism aimed at reducing the potential cytoplasmic calcium overload observed in septic shock. In addition, analysis of calcium function of isolated cardiomyocytes was performed on a limited number of animals (n = 4), and are consistent with results from western www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ blot analysis showing a significant improvement in calcium handling in NButGT treated animal cells (data not shown). Whilst previous studies have demonstrated a potential relationship between modulation of O-GlcNAc levels and SERCA2a expression, the exact impact on SERCA2a expression remains unclear. Two studies demonstrated an absence of SERCA2a expression changes following modulation of O-GlcNAc levels 37,38 . However, an increase in O-GlcNAc was also associated with a decrease in SERCA2a mRNA and protein expression in the context of diabetes with a modification of cardiac function 39,40 . Moreover, while SERCA2a can be O-GlcNAcylated, the consequence on its activity remains unknown 41,42 . The beneficial cardiovascular effects of NButGT in early septic shock management could be partially due to improved calcium homeostasis protein levels.
In summary, we demonstrate that acute O-GlcNAc stimulation by NButGT post-treatment tends to improve cardiovascular function and global outcome in two different models of septic shock. In this study, we show that this improvement is associated with the normalization of proteins involved in the calcium cycle. This pre-clinical study opens new avenues for a potential therapy of acute pathologies and especially septic shock.

Methods
Reagents. All reagents and solutions were purchased from Thermo Fisher Scientific Inc. (Waltham, MA) or from Sigma-Aldrich (St. Louis, MO). NButGT was synthesized using MS methods 43 . An NButGT dose of 10 mg/ kg was determined from pilot studies as the optimal dose to achieve optimal cardiac protein O-GlcNAcylation.
Animal models. Rats were housed under standard conditions of temperature (21-24 °C), humidity (40-60%) and 12 h light/dark cycle with the light period starting at 07:00. Food and water were available ad libitum. Experiments were approved by the ethics committee in charge of animal experimentation, committee of the Pays de la Loire #6, (2687-2015102815445892) and performed in accordance with French law on animal welfare, EU Determination of the optimal dose of nButGt. Determination of the optimal dose of NButGT in healthy rats was based on the study of Macauley and Vocadlo 44 . Three doses were tested in the LPS model and explored at the cardiac level by O-GlcNAc measurements. For all subsequent protocols in this study, the dose of 10 mg.kg −1 was used. endotoxemic rat model. Endotoxemic shock was induced in ten-week-old male Sprague-Dawley rats (Janvier, Le Genest St, France) by intra-venous injection of lipopolysaccharide (LPS, 5 mg.kg-1, LPS from E. Coli O111:B4, Sigma, France), and was compared to control rats (iv injection of saline). One hour later, LPS-treated rats were resuscitated with Gelofusine (B. Braun) 15 mL/kg alone (LPS+ R) or with NButGT (10 mg/kg). Three hours after LPS injection, rats were investigated for in vivo measurements, sacrificed and the left ventricle was collected in liquid nitrogen. Heart powder was obtained by grinding with a mortar.
To evaluate mortality, rats were observed for 30 hours after endotoxemic shock induction. During this period, animals were prematurely euthanized with a lethal dose of pentobarbital (Dolethal ® , Vetoquinol, Paris, France) if they met specific criteria: incapacity to move, decubitus position, difficulty in breathing.
cecal ligation and puncture rat model. Cecal ligature and puncture (CLP) surgery was performed on ten-week-old male Wistar rats (Charles River Ecully, France) by a modified protocol published by Rittirsch et al. 45 . Rats were anaesthetized with a gas-mixture of 1% isoflurane (Forene ® , Abbott France, Rungis, France) in O 2 . Two ligations were performed. The first-one was just under the ileo-cecal valve (100% of cecum length) to avoid ligation of the ileo-cecal artery, and so as to in initiate a massive necrosis. The second one was located at the distal part of the cecum (20% of length) and included the artery in order to produce necrosis. Between these two ligations, a single puncture was performed using a 16-G needle. Exactly 0.1 mL of faeces was withdrawn and then distributed in the abdominal cavity. Sham rats were anaesthetized and underwent the laparotomy without cecal ligation and puncture. All animals received buprenorphine treatment (10 µg/kg −1 , subcutaneously, Buprecare ® , Med'Vet, France) before surgery and 8 h post-surgery. CLP rats also received 10 ml/kg −1 of saline subcutaneously for fluid resuscitation after surgery, and 8 and 22 h post-CLP, supplemented, or not, with NButGT treatment at 10 mg/kg −1 . Twenty-four hours after surgery, rats were investigated for in vivo pressure measurements, sacrificed, and the heart and blood collected for biochemical analysis.
In vivo cardiovascular evaluation. Three hours after LPS or saline injection, animals were anesthetized with 2% volume of isoflurane and 1 L.min −1 O 2 to limit hemodynamic repercussion. Transthoracic echocardiography was performed using an ultrasound system VIVID7 (GE Healthcare, Horton, Norway) equipped with a 10 MHz sectorial probe as previously described 39 . Measurements were made on five cardiac cycles and averaged for each data value. Pressure measurements were performed through the right carotid artery to evaluate the mean arterial pressure signals and heart rate. The right carotid artery was isolated, ligated at the proximal end and a 2 F microtip pressure catheter was insert (Millar instruments Inc, Houston, Texas). Pressure signal and heart rate were recorded using an A/D converter (EMKA Technologies, Paris, France) stored and displayed on a computer by the IOX1.5.7 Software System (EMKA Technologies). Thereafter, blood and left ventricles were collected for biological and biochemical analysis.
Western blotting experiments were performed on LV tissue samples as previously described 46 , with extraction in the presence of 10 µM NButGT in T-PER buffer (Thermo Scientific). Proteins were evaluated using the following antibodies: GFAT (Thermo Scientific,