Anti-inflammatory and anti-excitoxic effects of diethyl oxopropanamide, an ethyl pyruvate bioisoster, exert robust neuroprotective effects in the postischemic brain

Ethyl pyruvate (EP) is a simple aliphatic ester of pyruvic acid and has been shown to have robust neuroprotective effects via its anti-inflammatory, anti-oxidative, and anti-apoptotic functions. In an effort to develop novel EP derivatives with greater protective potencies than EP, we generated four EP isosteres, among them the neuroprotective potency of N,N-diethyl-2-oxopropanamide (DEOPA), in which the ethoxy group of EP was replaced with diethylamine, was far greater than that of EP. When DEOPA was administered intravenously (5 mg/kg) to rat middle cerebral artery occlusion (MCAO) model at 6 hrs post-surgery, it suppressed infarct formation, ameliorated neurological and sensory/motor deficits, and inhibited microglial activation and neutrophil infiltrations in the postischemic brain more effectively than EP. In particular, DEOPA markedly suppressed LPS-induced nitrite production and cytokine/chemokine inductions in microglia, neutrophils, and endothelial cells and these effects are attributable to inhibition of the activity of NF-κB by suppressing IκB-α degradation and p65 to DNA binding. In addition, DEOPA suppressed NMDA-induced neuronal cell death in primary cortical neuron cultures by NAD replenishment and suppression of NF-κB activity. Together, these results indicate DEOPA has multi-modal protective effects against ischemic brain damage targeting numerous cell types in the brain and also against other inflammation-related diseases.

Scientific RepoRts | 7:42891 | DOI: 10.1038/srep42891 of EP against Zn 2+ toxicity might be due to two effects, namely, NAD replenishment and direct Zn 2+ chelation 15 . In addition, we also found direct Ca 2+ chelation by EP suppresses the phosphorylation and secretion of HMGB1 in microglia 8 . Therefore, EP appears to have a variety of protective effects that might be conveyed directly by EP or by pyruvate produced by the metabolization of EP.
Protective effects of EP have been documented in various diseases that affect the central nervous system (CNS). For example, EP suppressed microglia activation and inflammatory marker inductions, significantly reduced infarct volumes, and mitigated neurological deficits in rat model of middle cerebral artery occlusion (MCAO) 4 and attenuated kainic acid-induced neuronal cell death in the CA1 and CA3 regions of the mouse hippocampus 16 . EP also suppressed the death of nigrostriatal dopaminergic (DA) neurons in the mouse model of Parkinson's disease 17 and improved motor function scores in models of spinal cord ischemia and traumatic brain injury 18,19 . The anti-inflammatory and anti-oxidative effects of EP exhibited robust neuroprotective effects in above-mentioned diseases of the CNS.
In an effort to develop more effective therapeutics based on EP, many researchers have screened or generated derivatives of EP or of pyruvate. Sappington et al. 20 screened 15 commercially available EP-related compounds for cytoprotective and anti-inflammatory effects and reported the anti-inflammatory effects of diethyl oxaloproprionate (DEOP) and 2-acetamidoacrylate (2AA) were comparable to those of EP, and that methyl 2-acetamidoacrylate (Me-2AA; a methyl derivative of 2AA) had greater anti-inflammatory effects than EP or 2AA. However, when two α -keto-carboxylic acid derivatives of EP were examined, their protective potencies were less than those of EP in a rat model of multivisceral ischemia-reperfusion 21 . Because, many beneficial effects of EP are common to pyruvate, for example, NAD replenishment and hydroxyl radical scavenging 15,22 , in the present study, we generated four EP isosteres by saving the pyruvoyl group and replacing the ethoxy group of EP with ethylamine, diethylamine, thioester, or n-propyl to form the corresponding amides. We subsequently evaluated the neuroprotective effects of these isosteres in our rat model of MCAO and sought to identify the molecular mechanisms responsible for their effects in this model and in cell lines for microglia, neutrophil, and endothelial cells and primary cortical cultures.

Suppressions of neurological deficits and motor impairments by DEOPA.
To determine whether DEOPA could improve neurological deficits and motor impairment, 5 mg/kg of DEOPA was administered 6 hrs post-MCAO and modified neurological severity scores (mNSSs) were measured from 2 to 14 days post-MCAO. In the MCAO + DEOPA group, mean modified neurological severity score (mNSS) was 6.2 ± 0.9 (n = 8, p < 0.01) at 2 days post-MCAO, which was significantly lower than that of PBS-treated MCAO controls (11.5 ± 0.5, n = 8) (Fig. 3A). Furthermore, these improvements in the MCAO + DEOPA group were maintained until 14 days post-MCAO (Fig. 3A). When motor activities were assessed by rota-rod testing at 5 rpm, mean latency (time spent on the rod) in the MCAO + DEOPA group was significantly greater than that of PBS-treated MCAO controls (MCAO + PBS) until 14 days (Fig. 3B). Mean latency in the MCAO + DEOPA group was significantly greater than that of MCAO + EP group until 5 days, although this difference disappeared at 7 days (Fig. 3B). Interestingly, at 15 rpm, mean latencies in the MCAO + DEOPA was significantly greater than that of PBS-treated MCAO controls throughout the experimental period until 14 days or of MCAO + EP group, in particular at 2, 5, and 14 days post-MCAO (Fig. 3B). These results suggest that the marked infarct suppression observed for DEOPA was accompanied by better neurological and motor outcomes, and that these effects were evident by a task with a high degree of difficulty and at delayed time points. Furthermore, pH, PaO 2 , PaCO 2 , and blood glucose levels were similar in the MCAO + DEOPA and the PBS-treated MCAO group, indicating that physiological parameters were not influenced by DEOPA (Supplementary Table S1).
Scientific RepoRts | 7:42891 | DOI: 10.1038/srep42891 Anti-inflammatory effects of DEOPA in the postischemic brain. Since EP is known to have strong anti-inflammatory effects, we examined anti-inflammatory potency of DEOPA in the postischemic brain. DEOPA (5 mg/kg, i.v.) was administered at 6 hrs post-MCAO and brain sections obtained at 2 days post-MCAO were stained with antibody against Iba-1 (a marker of cells of myeloid origin), Mac2 (a marker of activated resident microglia), or MPO-1 (myeloperoxidase-1, neutrophil marker). In sham controls, Iba-1 + cells exhibited a ramified morphology and were detected throughout brains (Fig. 4A). However, at 2 days post-MCAO, Iba-1 + cells in the cortices of ipsilateral hemispheres of PBS-treated MCAO controls were amoeboid in shape, indicating an activated state (Fig. 4B). However, in the MCAO + DEOPA group, most Iba-1 + cells retained a ramified morphology (Fig. 4C). In contrast to Iba-1 + cells, Mac2 + and MPO-1 + cells were barely detected in sham controls ( Fig. 4E and I), whereas in treatment naïve-MCAO controls, numbers of these cells were markedly elevated at 2 days post-MCAO (Fig. 4F, J, M and N). However, they were hardly detected in the MCAO + DEOPA group (Fig. 4G, K, M and N). Notably, the suppressive effects of DEOPA were greater than those of EP in both staining ( Fig. 4H and L-N). In addition, the inductions of proinflammatory markers (iNOS, TNF-α , Cox-2, IL-1β ) observed at 2 days post-MCAO in the PBS-treated MCAO group were significantly suppressed in the MCAO + DEOPA group to an extent greater than that was observed in the MCAO + EP group ( Fig. 4O and P). These results indicated the anti-inflammatory potency of DEOPA was greater than EP.

Suppressions of NO and of proinflammatory marker inductions in activated microglia by DEOPA.
To confirm the anti-inflammatory effects of DEOPA, BV2 cells (a microglia cell line) were treated with LPS (200 ng/ml) for 24 hrs and nitrite levels were measured in the presence or absence of DEOPA. NO production was dose-dependently suppressed by co-treating DEOPA (1, 5, or 10 mM) and reduced to the basal level at 10 mM (Fig. 5A). Importantly, the efficacy of DEOPA was greater than that of EP at all doses tested (Fig. 5A). Similarly, iNOS induction was markedly suppressed by DEOPA (5 or 10 mM) and these suppressive effects were also greater than those achieved by the same doses of EP (Fig. 5B). Furthermore, proinflammatory marker (TNF-α , Cox-2, and IL-1β ) levels induced by LPS were also markedly suppressed by DEOPA, and once again the suppressive effects of DEOPA were greater than those of EP at all doses tested (Fig. 5B). Suppression of NF-κB activity by DEOPA via the inhibition of IκB-α degradation in cytosol and of p65 to DNA binding in nucleus. The marked suppression of nitrite production and proinflammatory marker induction by DEOPA (Fig. 5) prompted us to examine the effects of DEOPA on the NF-κ B signaling pathway. The amount of Iκ B-α in the cytoplasm of LPS-treated BV2 cells was significantly decreased after treatment with LPS (200 ng/ml) for 15 min and this decrease was suppressed by pre-treating DEOPA (10 mM) but not by pre-treating EP (10 mM) (Fig. 6A). Consistent with these observations, the nuclear translocation of p65 was markedly suppressed by DEOPA but not by EP (Fig. 6B), indicating DEOPA suppressed the nuclear translocation of p65 probably by inhibiting Iκ B-α degradation in cytosol. Under these conditions, levels of α -tubulin and lamin B were unchanged in cytoplasm and nucleus, respectively ( Fig. 6A and B). In addition, when the DNA binding activity of p65 was examined using the TransAM NF-κ b p65 assay kit and recombinant p65, LPS-induced p65 to DNA binding activity was found to be suppressed by DEOPA and by EP, and the suppressive effect of DEOPA was comparable to EP (Fig. 6C). Furthermore, when we measured LPS-induced NF-κ B activity after transfecting BV2 cells with NF-kB-Luc reporter plasmid containing five copies of the NF-κ B consensus sequence (GGGAATTTCC), NF-κ B activity was suppressed by pre-treating DEOPA (10 mM) to 45.2 ± 4.0% (n = 4) of that observed for LPS-induced cells, and this suppression was significantly greater than that obtained by pre-treating the same amount of EP (68.1 ± 4.7%, n = 4; Fig. 6D). Importantly, in the postischemic brain, the decrease of the cytoplasmic Iκ B-α amounts and the induction of nuclear translocation of p65 were observed in the cortical penumbra of the ischemic hemisphere at 4 or 12 hrs post-MCAO, respectively ( Fig. 6E-G). They were significantly suppressed by DEOPA (5 mg/kg) administered at 1 hr post-MCAO ( Fig. 6E-G), further confirming these modulations also occurred in vivo. These results indicate that DEOPA exerted a robust anti-inflammatory effect by suppressing NF-κ B activity at two points, by inhibition of Iκ B-α degradation in cytoplasm and p65 to DNA binding in nuclei.
Blockades of neutrophil-endothelial adhesion and of transendothelial neutrophil migration by DEOPA. The almost complete absence of neutrophils in cortical penumbra of MCAO + DEOPA animals ( Fig. 4K and N) prompted us to investigate whether DEOPA-mediated suppression of NF-κ B activity is also responsible for blocking neutrophil infiltration. Differentiated human promyelocytic leukemia cells (dHL-60 cells) and human umbilical vein endothelial cells (HUVECs) were activated by treating them with TNF-α (10 U/ml) for 12 hrs (Fig. 7A). dHL-60 cells displayed multi-lobular nuclear morphology, indicative of the differentiated state, which was confirmed by staining with anti-CD11 antibody (Fig. 7A). After 30 min of co-culture of dHL-60 cells and HUVECs, the number of dHL60 cells that attached to HUVECs was increased almost 15 fold as compared with TNF-α -untreated cells ( Fig. 7B and C). However, pre-treatment of dHL-60 cells and HUVECs with DEOPA (5 or 10 mM) before TNF-α treatment significantly reduced attached cell numbers ( Fig. 7B and C). Pre-treatment of EP (10 or 20 mM) also reduced attached cell numbers but to a lesser extent ( Fig. 7B and C). Similarly, the numbers of dHL-60 cells that migrated through HUVECs were markedly reduced by pre-treating both cells with DEOPA (10 mM) and these suppressive effects were greater than that of the same amount of EP

Suppression of cell adhesion molecule inductions in endothelial cells and neutrophils by
DEOPA. ICAM-1 and P-selection are known to be induced in endothelial cells during neutrophil-endothelial interactions 23 and modulated by NF-kB 24,25 . Expressions of ICAM-1 and P-selection were significantly increased in TNF-α -treated HUVECs and these inductions were effectively suppressed by pre-treating DEOPA (5, 10 mM) for 1 hr before TNF-α treatment ( Fig. 7F and G). Similarly, inductions of PSGL-1 and Mac-1 in dHL-60 cells, that binds to endothelial P-selectin 26 and ICAM-1 27 , respectively, were significantly suppressed by pre-treating DEOPA (5, 10 mM) ( Fig. 7H and I). Interestingly, EP also suppressed the upregulations of those genes but the suppressive potencies of DEOPA were greater than EP in all cases ( Fig. 7G and I). Together these results indicate that the DEOPA-mediated suppression of NF-κ B activity and subsequent inhibitions of cell adhesion molecule expressions might be responsible for blocking neutrophil infiltration into the postischemic brain.
Anti-excitotoxic effects of DEOPA in primary cortical cultures. We next investigated whether DEOPA protects neurons from acute damaging processes after ischemic/reperfusion, such as, excitotoxicity and Zn 2+ -toxicity. In primary cortical cultures, NMDA (30 μ M, 10 min)-induced LDH release was suppressed by EP and DEOPA (5, 10, or 20 mM) (Fig. 8A). Although, at 5 mM, the anti-excitotoxic effect of EP was greater than that of DEOPA, at 10 or 20 mM, anti-excitotoxic efficacies of DEOPA were comparable to those of EP (Fig. 8A). Interestingly, DEOPA did not protect neurons from Zn 2+ -induced toxicity in acute (300 μ M, 30 min)-or chronic (40 μ M, 24 hrs) Zn 2+ -treated condition, whereas EP did (Supplementary Figure S1) 15 . Since NMDA receptors regulate neuronal PARP-1 expression and activity 28 and PARP-1 causes cell death by depleting NAD and ATP 29 , we examined NAD levels in NMDA-treated primary cortical cultures in the presence or absence of DEOPA. DEOPA (5 mM) enhanced basal NAD levels in normal primary cortical cultures, although the effect was weaker than those of EP (5 mM) or pyruvate (5 mM) (Fig. 8B). In addition, NMDA (30 μ M, 10 min)-induced NAD depletion was significantly suppressed by co-treating DEOPA (5 mM) with NMDA for 10 min, and DEOPA was as effective as EP (5 mM) (Fig. 8C). Interestingly, NAD level was decreased in ischemic hemispheres to 49.5% of that of contralateral hemisphere at 24 hrs after MCAO and it was also restored by DEOPA (5 mg/kg i.v.) to 88.9% of the control (Fig. 8D), further confirmed that the suppression of NAD depletion by DEOPA contributes to a robust neuroprotective effect of DEOPA in the postischemic brain. Moreover, DEOPA also inhibited Iκ B-α degradation    in cytoplasm and suppressed NMDA-induced NF-κ B activation in primary cortical cultures (Fig. 8E and F). As it was found in LPS-treated BV2 cells, EP failed to inhibit Iκ B-α degradation in cytoplasm (Fig. 8E). These results indicate DEOPA has an anti-excitotoxic effect, and suggest this also contributes to its protective effects in the postischemic rat brain.

Discussion
EP has been shown to have anti-inflammatory, anti-oxidative, and anti-apoptotic effects, and to chelate several metal ions. During our continues efforts to develop novel therapeutics with improved efficacies based on EP, we synthesized four isosteres of EP and found that one of these isosteres, DEOPA, in which the ethoxy group of EP was replaced with diethylamine, was more neuroprotective than EP, more specifically, it suppressed infarct formation and mitigated neurological deficits more effectively than EP in the postischemic brain. Intriguingly, the anti-inflammatory activity of DEOPA was superior to EP and it was attributed to its inhibition of NF-κ B activity, which was considerably greater than EP. Interestingly, marked suppression of NF-κ B activity was responsible for suppressing activation of microglia and for anti-excitotoxic effect in neurons and together these effects contribute to a robust neuroprotective effect in the postischemic brain (summarized in Supplementary Figure S4). Although efforts have been made to develop new therapeutics based on EP, this study represents the first systematic approach to generate and evaluate EP isosteres and present a potential candidate molecule. Regarding the mechanism underlying its anti-inflammatory effect, our findings indicate DEOPA inhibits both Iκ B-α degradation in cytosol and p65 to DNA binding in the nuclei of microglial cells (Fig. 6), which stands in contrast to the mechanism of EP, that involves the blockage of p65 to DNA binding but not Iκ B-α degradation 10 . Accordingly, these two mechanisms probably explain the robust anti-inflammatory effect observed for DEOPA in vivo and in vitro, which was also considerably greater than EP (Figs 4 and 5). Regarding the molecular mechanism underlying inhibition of p65 to DNA binding by DEOPA, we observed direct inhibition of this binding under cell free conditions (Fig. 6C). However, it considered probable that additional mechanisms might also be involved, such as, covalent modification of p65 at Cys 38 10 , GSH-depletion-mediated redox state change of p65 9 , or the inhibition of ROS-dependent STAT signaling 31 , as have been reported for EP. We speculate that the mechanisms responsible for the robust anti-inflammatory effect of DEOPA are likely to be multifactorial and regardless of the nature of the mechanisms involved, DEOPA might be more potent than EP. Interestingly, DEOPA failed to reduce HMGB1 release in LPS-treated BV2 cells (Supplementary Figure S2), whereas EP has been used to inhibit HMGB1 release and this inhibition contributes to its anti-inflammatory effect 8,12,32,33 . We suggest structure-functional studies are needed to be undertaken to explore the effects of the functional groups of DEOPA on the inhibitions of NF-κ B signaling or HMGB1 release.
In the present study, we found that suppression of NF-κ B activity by DEOPA not only inhibits microglia activations but suppresses neutrophil infiltration and excitotoxicity in neurons (Figs 4, 7 and 8). Although microglia are the major resident immune cells and play important roles in postischemic inflammation, circulating immune cells that infiltrate brain parenchyma, especially after blood brain barrier (BBB) disruption, also play pivotal roles. Stroke is well-known to be associated with acute and massive influxes of neutrophils, which are the first leukocytes to accumulate around injured brain tissues and aggravate tissue injury 34 . In the present study, we suggested inhibition of the NF-κ B activation, and thus the expressions of cell adhesion molecules, by DEOPA underlies effective blocking of neutrophil infiltration into the cortical penumbra of the postischemic brain (Fig. 4). Accumulating evidence indicates an association exists between the proinflammatory activation of neutrophils and infarct size, BBB disruption, hemorrhagic transformation, and poorer neurologic outcomes in the postischemic brain [35][36][37] . Moreover, activation of NF-κ B has known to be involved in NMDA-induced neuronal apoptosis in rat striatum and in primary cortical cultures 30,38 . Accordingly, it appears the neuroprotective effect of DEOPA is attributable to its anti-inflammatory effects on microglia and neutrophils and anti-excitotoxic effect on neurons and suppression of NF-κ B is critical for both effects.
Regarding the structure-function relationships of EP, some functions appear to be specifically associated with EP and not pyruvate, whereas others are common to both, which suggests that these other functions are due to the pyruvate generated by the de-esterification of EP or to the pyruvoyl moiety of EP. Direct ion chelation, such as, of Ca 2+ or Zn 2+ , and Nrf2 activation are examples of EP-specific functions 8,14,15 , whereas NAD replenishment and scavenging hydroxyl radical are examples of the latter 15,39 . However, such a classification is incomplete for some functions. For DEOPA, since it cannot be metabolized to pyruvate the observed effects of DEOPA might not be derived from pyruvate itself. In this regard, it is interesting to note that in contrast to EP, DEOPA had no protective effect in primary neuronal cultures exposed to Zn 2+ (Supplementary Figure S1) and failed to decompose H 2 O 2 under cell free conditions (Supplementary Figure S3). These findings demonstrate the pharmacologic properties of DEOPA and EP intrinsically differ, and suggest the robust protective effects of DEOPA are due to quite different underlying mechanisms.
In addition to reports issued on commercially available EP-related compounds 20,21 , recently, Min et al. 40 reported on the anti-inflammatory effect of EOP (S-ethyl 2-oxopropanethioate) in LPS-treated BV2 cells. However, in the present study, we found that the neuroprotective effect of EOP was unimpressive in the postischemic brain (Fig. 1B). Although more remains to be done to develop improved therapeutics, one important point raised by the above-mentioned studies is that lipophilicity may critically affect therapeutic efficacy 20,21 . In terms of the lipophilicities of EP isosteres, the ClogP value of DEOPA is 0.126, which is comparable to that of EP, 0.161, and thus, the comparatively high lipophilicity of DEOPA might also contribute to its neuroprotective effects.
In view of our finding of the substantially greater protective effect of DEOPA than EP in the postischemic brain, we believe DEOPA has unknown functions that contribute to its effects. We suggest further studies be conducted to identify these functions and to elucidate the mechanisms involved.

Materials and Methods
General procedure for synthesis of bioisosteres of EP. 2-Oxopropanoyl chloride (2), which was prepared from pyruvic acid (1) and α ,α -dichloromethyl methyl ether, was added to a solution of ethylamine (1a), diethylamine (1b), or ethanethiol (1c) in dichloromethane (40 mL) at 0 °C. Mixtures were stirred at room temperature for 4 hrs and then quenched with 1 N HCl. Solutions were extracted with ethyl acetate and organic layers were washed with water, brine, dried over Na 2 SO 4 , and evaporated in vacuum. Crude residues were purified by column chromatography to give the title compounds.  N-Ethyl-2-oxopropanamide (1a). The title compound was obtained as a yellow oil (620 mg, 19%).

N,N-Diethyl
Spectral data were consistent with previously reported data data 41 . 2-oxopropanethioate (1c). The title compound was obtained as a yellow oil (1.11 g, 45%).

S-Ethyl
Spectral data were consistent with the previously reported data data 41 . hexane-2,3-dione. The compound was purchased from Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO).
Infarct volume assessments. Rats were decapitated at 2 days post-MCAO, and whole brains were dissected coronally into 2-mm slices using a metallic brain matrix (RBM-40000, ASI, Springville, UT). Slices were immediately stained by immersing them in 2% 2,3,5-triphenyl tetrazolium chloride (TTC) at 37 °C for 15 min and then fixed in 4% paraformaldehyde. Infarcted tissue areas were quantified using the Scion Image program (Frederick, MD). To account for edema and shrinkage, areas of ischemic lesions were calculated using (contralateral hemisphere volume -(ipsilateral hemisphere volume -measured injury volume)). Infarct volumes were quantified (in mm 3 ) by multiplying summed section infarct areas by section thickness.
Rota-rod test. Twenty-four hours before surgery, rats were conditioned on a rota-rod unit at a constant speed (3 rpm) until they were able to remain on the rotating spindle for 180 s. At 2, 5, 7, 10, or 14 days post-MCAO, rats were subjected to rota-rod testing at spindle speeds of 5 and 15 rpm and residence times on the spindle were recorded with a 1 hr rest period after each test.
Cell cultures. BV2 cell (a microglial cell line), human umbilical vein endothelial cells (HUVECs), and HL-60 cells (a human myelocytic leukemia cell line) were obtained from the American Type Culture Collection (ATCC, Rockville, MD). BV2 cells were cultured as previously described 14 . HUVECs were cultivated in Endothelial Cell Medium (ScienCell, Corte Del Cedro Carlsbad, CA). HL-60 cells were cultivated in DMEM supplemented with 20% FBS and 100 U/ml penicillin/streptomycin. HL-60 differentiation was induced by treating 1 μ M all-trans retinoic acid (ATRA) (Sigma-Aldrich, St. Louis, MO) for 3 days. All these cell types were incubated at 37 °C in a 95% air/5% CO 2 humidified atmosphere.