Jump to
Abstract
Objective—Nitro-oxidative stress exerts a significant role in the genesis of hypoxic-ischemic (HI) brain injury. We previously reported that the ω-6 long chain fatty acids, transarachidonic acids (TAAs), which are nitrative stress-induced nonenzymatically generated arachidonic acid derivatives, trigger selective microvascular endothelial cell death in neonatal neural tissue. The primary molecular target of TAAs remains unidentified. GPR40 is a G protein–coupled receptor activated by long chain fatty acids, including ω-6; it is highly expressed in brain, but its functions in this tissue are largely unknown. We hypothesized that TAAs play a significant role in neonatal HI-induced cerebral microvascular degeneration through GPR40 activation.
Approach and Results—Within 24 hours of a HI insult to postnatal day 7 rat pups, a cerebral infarct and a 40% decrease in cerebrovascular density was observed. These effects were associated with an increase in nitrative stress markers (3-nitrotyrosine immunoreactivity and TAA levels) and were reduced by treatment with nitric oxide synthase inhibitor. GPR40 was expressed in rat pup brain microvasculature. In vitro, in GPR40-expressing human embryonic kidney (HEK)-293 cells, [14C]-14E-AA (radiolabeled TAA) bound specifically, and TAA induced calcium transients, extracellular signal–regulated kinase 1/2 phosphorylation, and proapoptotic thrombospondin-1 expression. In vivo, intracerebroventricular injection of TAAs triggered thrombospondin-1 expression and cerebral microvascular degeneration in wild-type mice, but not in GPR40-null congeners. Additionally, HI-induced neurovascular degeneration and cerebral infarct were decreased in GPR40-null mice.
Conclusions—GPR40 emerges as the first identified G protein–coupled receptor conveying actions of nonenzymatically generated nitro-oxidative products, specifically TAAs, and is involved in (neonatal) HI encephalopathy.
Introduction
Abundant evidence points to a role for nitro-oxidative stress in hypoxic-ischemic (HI) encephalopathies, including in the newborn devoid of antioxidant systems.1–4 In HI insult, the neurovascular endothelium is particularly susceptible to ensue nitro-oxidative stress,5–8 resulting in marked microvascular degeneration, which contributes to the pathogenesis of HI brain injury.9,10 Antioxidants and free radical scavengers have been found to be effective in reducing lesions associated with experimental neonatal HI.11 Cytotoxicity in response to nitro-oxidative stress involves modifications of proteins, DNA, and lipids.5,6 In this context, numerous mediators of peroxidation and nitration are generated and found to reproduce cytotoxicity associated with HI brain injury.12–15 We described a lipid nitrative process mediated by the nitrogen dioxide radical (NO2● ) resulting in a cis- to transisomerization of arachidonic acid, an ω-6 C20 polyunsaturated fatty acid (FA).16 This nonenzymatic reaction produces stable free transisomeric FAs named transarachidonic acids (TAAs).17 We have shown that TAA levels increase in a model of ischemic retinopathy of prematurity and seem to contribute to the microvascular degeneration observed in the retina in vivo18; analogous observations were made in a model of ischemic encephalopathy of the newborn.7 However, as is also the case for peroxidation products, although certain cell signaling mechanisms associated with microvascular degeneration induced by TAAs have been unveiled,18 the primary molecular target of TAAs has yet to be identified; this aspect is particularly relevant in the context of the relative role of TAAs in the pathology (herein HI) associated with their generation.
GPR40 has been described as a receptor for long chain FAs, including ω-6,19–21 whereas GPR120, the other long chain FA receptor, is mostly activated by ω-3 long chain FAs.21 In humans, GPR40 is principally expressed in pancreatic islets and brain tissue.19 Numerous studies examined the role of GPR40 in the FA-induced potentiation of insulin secretion and in β-cell dysfunctions associated with type 2 diabetes mellitus, and evaluated the use of GPR40 agonists as potential therapeutic options for the treatment of type 2 diabetes mellitus.20,22,23 However, the role of GPR40 in brain specifically in HI insult is elusive. Moreover, whether the long chain ω-6 FA TAA acts via GPR40 to elicit microvascular injury as seen in HI is also unknown. We therefore proceeded to address these 2 interrelated challenges.
We hereby report involvement of nitrative stress in the production of TAAs in a rodent model of neonatal HI and present data suggesting that TAAs mediate HI-induced brain microvascular degeneration and ensue brain infarct via binding and activation of the GPR40 receptor; these findings unveil for the first time a primary target for a nonenzymatically generated nitro-peroxidation products.
Materials and Methods
Materials and Methods are available in the online-only Supplement.
Results
Neonatal HI Induces a Nitrative Stress-Dependent Brain Infarct Associated With Diffuse Vascular Degeneration
Twenty-four hours after exposure of P7 rat pups to HI insult using the Rice-Vannucci model (see details in Methods in the online-only Data Supplement), we observed an infarct on the brain hemisphere (mostly cortical) ipsilateral to the carotid ligation, as well as a diffuse 40% decrease in vascular density on the same side in peri-infarct regions (Figure 1A–1C). These changes were associated with a rapid (within 18 hours of insult) increase in protein and lipid nitration in ipsilateral brain parenchyma as attested by a rise in 3-nitrotyrosine adducts primarily localized on brain microvasculature (Figure 1D) and a concomitant increase in TAA (Figure 1E) to values corresponding to 2 to 5 μmol/L24; this increase in TAA was rapidly cleared as previously reported.7 As illustrated, sham, hypoxic alone, or ischemic alone control brain sections did not reveal vascular degeneration, infarct, or increases in nitration products. Inhibition of nitric oxide synthase (NOS) with intracerebroventricular L-NAME–abrogated TAA increases and markedly attenuated brain infarct size and vasoobliteration (Figure 1F–1J).
Hypoxia-ischemia (HI) induces microvascular degeneration and brain infarct via nitrative stress and ensuing production of transarachidonic acids (TAAs). A, Representative brain sections illustrating the infarcted zone (white) evaluated by TTC staining in sham and HI rat pups. B, Representative brain sections stained with lectin Griffonia simplicifolia (red) illustrating microvascular degeneration evaluated by decreased vascular density. Scale bar=100 µm. C, Time course quantification of microvascular density (n=3–10). D, Representative brain sections illustrating the nitrative stress within microvessels as evidenced by costaining of 3-nitrotyrosine (green) and lectin (red), 18 hours post-HI. Sham, hypoxia, and ischemia serve as controls for the model. Scale bar=100 µm. E, Time course quantification of total brain TAA levels (free + bound [phospholipid-esterified]) in HI-induced rat pups (n=5–11). F, Total brain TAA levels in HI rat pups treated with artificial cerebrospinal fluid (aCSF) or L-NAME (2 mg/kg) 18 hours postinjury (n=3–4). Values in histograms are mean±SEM. *P≤0.05 compared with their respective control. G, Representative brain sections illustrating the effect of aCSF and L-NAME (2 mg/kg) infusion on brain infarct in rat pups that underwent HI. H, Quantification of the infarcted volume in animals treated as indicated in H (n=3–4). I, Representative brain sections illustrating the effect of aCSF and L-NAME (2 mg/kg) infusion on microvascular density in rat pups that underwent HI. Scale bar=100 µm. J, Quantification of microvascular density in animals treated as indicated in F (n=4–10).
TAAs Bind and Activate the GPR40 Receptor
To elucidate the specific role of TAA in HI injury in the newborn, we proceeded to identify the primary target of TAA. Because GPR40 receptor is a putative receptor for long chain FAs19,20 and cis-AA and TAAs share very high structural homology,16,25 we determined whether TAA can bind and activate GPR40. GPR40 specifically localized on cerebral microvessels (Figure 2A) and was hardly detected on neurons and glia; concordantly, GPR40 mRNA expression was far greater in isolated microvessels compared with whole brain extracts (Figure 2B), despite the fact that cerebral microvasculature only corresponds to ≈2% of total brain mass.26 We next stably transfected HEK293 cells with the hGPR40 receptor (pIRESpuro-hGPR40) to assess binding properties of TAAs on GPR40. Because all TAA are equivalently cytotoxic on endothelial cells,18 we used 14E-AA for binding, efficacy, and mode of action. Specific binding of [14C]14E-AA was detected on pIRESpuro-hGPR40-expressing cells but hardly found in control cells transfected with the plasmid devoid of GPR40 gene; as expected, 14E-AA (as well as cis-AA) displaced bound [14C]14E-AA with an IC50 of ≈1 μmol/L (Figure 2C). Coherent with these observations, 14E-AA (5 μmol/L) specifically accrued intracellular calcium transients, extracellular signal–regulated kinase 1/2 phosphorylation, and the antiangiogenic (endothelial) proapoptotic factor thrombospondin-1 (TSP-1) in GPR40-containing cells, but not in GPR40-devoid cells (Figure 2D and 2E); the latter are consistent with extracellular signal–regulated kinase 1/2 and TSP-1–dependent TAA-induced endothelial cell death.18 Collectively, these data suggest that TAA bind and activate GPR40.
Transarachidonic acids bind and activate the GPR40 receptor. A, Representative brain sections illustrating the expression of GPR40 on microvessels as evidenced by costaining of lectin (red) and GPR40 antibody (green) in P7 rat pups. Scale bar=100 µm. B, GPR40 mRNA expression in whole rat brain extract and isolated microvessels. (n=3–4). C, Concentration displacement curve of [3C]-cis-AA by AA and 14E-AA in human embryonic kidney (HEK)293 cells stably transfected with the human GPR40. (n=3–4). D, Effect of 14E-AA (5×10−6 mol/L) on intracellular calcium mobilization in HEK293 cells stably transfected with the human GPR40 (pIRESpuro-hGPR40) or the empty vector (pIRESpuro). E, Representative Western blots of extracellular signal–regulated kinase (ERK)1/2 phosphorylation in HEK293 cells stably transfected with the human GPR40 (pIRESpuro-hGPR40) or the empty vector (pIRESpuro) and exposed to 14E-AA (5×10−6 mol/L). Immunoblots are representative of 3 independent experiments. Representative Western blots of thrombospondin-1 expression in HEK293 cells stably transfected with the human GPR40 (pIRESpuro-hGPR40) or the empty vector (pIRESpuro) and exposed to 14E-AA (5×10−6 mol/L) for 21 hours. Immunoblots are representative of 3 independent experiments. Values in histograms are mean±SEM. *P≤0.05 compared with their respective control.
TAA Induce Endothelial Cell Death Via GPR40
We next addressed whether TAA-induced endothelial cell death is GPR40 dependent. 14E-AA, but not cis-AA or cis-linoleic acid, decreased the viability of brain cerebral microvascular endothelial cells (human; Figure 3A); effects of 14E-AA were abrogated by knockdown of GPR40 (using siRNA-GPR40; Figure IA in the online-only Data Supplement). Consistent with this observation, 14E-AA interfered with microvascular sprouting of explants (in Matrigel) of aorta from wild-type (WT) mice, but not from congener GPR40 knockout (KO) mice (Figure 3B and 3C).
Transarachidonic acids induce endothelial cell death via the GPR40 receptor. A, Effects of 14E-AA, cis-arachidonic acid, and cis-linoleic acid (10−6 mol/L) on human brain microvascular endothelial cells 72 hours after transfection or not with a scrambled negative control siRNA or a siRNA targeting GPR40. B, Representative microvascular sprouting from Matrigel-embedded aortic rings harvested from wild-type (WT) and GPR40−/− knockout (KO) mice exposed to control solution or 14E-AA (5×10−6 mol/L). Scale bar=500 µm. C, Quantification of aortic ring vessel outgrowth as indicated in B (n=3–4). D, Representative brain sections illustrating the effects of intracerebroventricular injections of artificial cerebrospinal fluid (aCSF) and 14E-AA (5×10−6 mol/L) on microvascular density visualized with lectin (green) in WT and GPR40 KO mice. Scale bar=100 µm. E, Quantification of microvascular density as indicated in D (n=3).
To further confirm the role of GPR40 in TAA-induced brain microvascular degeneration, we conducted in vivo experiments by injecting intracerebroventricularly 14E-AA (estimated final brain concentration ≈5 μmol/L) in mice. Twenty-four hours after injection of 14E-AA, we observed a significant decrease in periventricular cerebral microvascular density in WT, but not in GPR40 KO mice (Figure 3D and 3E).
To ascertain that TAA-induced GPR40-dependent microvascular degeneration can occur throughout the extended central nervous system, including the retina, we tested this premise in retinas knocked down or not of GPR40 using a lentivirus encoded with shRNA-GPR40 (Figure IB in the online-only Data Supplement), as we previously used for other targets.27,28 14E-AA injected intravitreously at P5 interfered with normal retinal vascular development (detected at P8), as attested by decreased vascular area and density; the effects of 14E-AA were abrogated by GPR40 knockdown (Figure IIA–IID in the online-only Data Supplement). Hence TAA-induced neuromicrovascular degeneration is GPR40 dependent.
We previously reported that TAA exert neurovascular endothelial cytotoxicity by inducing expression of the proapoptotic antiproliferative TSP-1,18 which in turn acts by binding to its transmembrane receptor, CD3629,30; concordantly, we hereby show that 14E-AA induces TSP-1 in HEK293 cells, but this requires expression of GPR40 (Figure 2E). To ascertain that TSP-1 was involved in 14E-AA–induced cerebral microvascular degeneration, we tested its effects in brain explants. 14E-AA–induced cerebral microvascular degeneration was abrogated with an antibody that targets the binding site of TSP-1 on the CD36 receptor (Figure 4C and 4D), whereas, an antibody that targets the ox-LDL site of CD36 was ineffective. Because TAA primarily and specifically binds to GPR40 (Figure 2B and 2C), we proceeded to show its role in TAA-induced capillary drop out using GPR40 KO mice. Intracerebroventricularly injected 14E-AA–elicited TSP-1 expression (mostly localized on microvasculature [Figure 4A]) and microvascular toxicity observed in vivo in WT mice were not detected in GPR40 KO mice (Figure 4B). These data, along with our previous report showing markedly reduced vasoobliteration in an ischemic retinopathy model in CD36 null mice,18 suggest that TAAs induce neurovascular endothelial cell death via activation of GPR40 and the subsequent generation of TSP-1 (acting via its CD36 receptor).
In vivo, 14E-AA induces the expression of thrombospondin (TSP)-1 in a GPR40-dependent manner. A, Representative brain sections illustrating the microvascular (lectin positive; green) colocalization of TSP-1 (red) in rat pups injected with the control solution or 14E-AA (5×10−5 mol/L), 24 hours postintracerebroventricular injections. Scale bar=100 µm. B, Representative brain sections illustrating the microvascular (lectin positive; red) colocalization of TSP-1 (green) in wild-type (WT) and GPR40 KO mice injected with 14E-AA (5×10−5 mol/L), 24 hours postintracerebroventricular injections. Scale bar=100 µm. Values in histograms are mean±SEM. *P≤0.05 compared with their respective control. C, Representative brain explants illustrating the ex vivo effects of 14E-AA (5×10−6 mol/L) on vascular density in presence or absence of a CD36 monoclonal antibody specific for the TSP-1 binding site (200 µg/mL) or the ox-LDL binding site (100 µg/mL). Scale bar=100 µm. D, Quantification of microvascular density as indicated in C (n=4–10).
HI-Induced Brain Damage Involves TAA Production and GPR40-Dependent Mechanisms
Because neonatal HI elicits a nitrative stress-dependent increase of TAA levels (Figure 1D and 1E), which subsequently induces endothelial cell death via activation of the GPR40 receptor (Figure 3A–3E), we determined the specific contribution of GPR40 in the cerebromicrovascular degeneration and infarct size triggered by HI. In contrast to the ligand TAA, expression of GPR40 was not affected by HI (Figure IV in the online-only Data Supplement). As illustrated in Figure 5, neonatal HI insult induced an ipsilateral infarct associated with decreased brain vascular density on the ipsilateral side of peri-infarct regions of WT mice. In contrast, in GPR40 KO mice exposed to the same insult, brain infarct volume was markedly reduced and microvascular density was preserved (Figure 5A–5D). Likewise, in an established ischemic retinopathy model associated with microvascular degeneration (induced by exposure of rodent pups to hyperoxia) and high TAA levels,18 GPR40 knockdown (by injecting intravitreally a lentivirus encoded with shRNA-GPR40) prevented microvascular obliteration (Figure IIIA and IIIB in the online-only Data Supplement).
Hypoxia-ischemia (HI)–induced brain damages involve GPR40-dependent mechanisms. A, Representative brain sections illustrating the infarcted zone (white) evaluated by TTC staining in wild-type (WT) and GPR40 knockout (KO) mice subjected to HI, 24 hours postinjury. B, Quantification of the infarcted volume in mice as indicated in A (n=5–6). Values in histograms are mean±SEM. *P≤0.05 compared with their respective control. C, Representative brain sections stained with lectin (red) illustrating microvascular degeneration in WT and GPR40 KO mice subjected to HI, 24 hours postinjury. Scale bar=100 µm. D, Quantification of microvascular density in mice as indicated in C (n=9–10).
Discussion
Nitro-oxidative stress has been amply shown to be involved in the pathogenesis of HI brain injury of the neonate (and adult).1–4 Yet, antioxidants to date have not been successful therapeutic candidates for several reasons, which include the following: (1) compartmentalization of oxidation/nitration, (2) propensity of agent to scavenge preferentially reactive oxygen versus nitrogen species, (3) accessibility of antioxidants to appropriate site,31 (4) possible benefits of avoiding excess antioxidants to maintain a balanced redox potential,32,33 and (5) importantly, the induced toxic effects of stable products of oxidation/nitration (such as isoprostanes, isofurans, TAA [which are themselves not reactive]) after the oxidant stress subsides,25,34 thus acting as mediators of reactive oxygen and nitrogen species in tissue injury. Accordingly, identification of the primary site of action of stable products of oxidation/nitration would be of interest physiologically and therapeutically. Because HI brain (and retinal) injury of the newborn is significantly contributed to by nitrative stress,5–8 we proceeded to determine the primary site of action of stable products of nitro-oxidative stress, specifically TAA, because to date no specific receptors have been identified for nonenzymatic products derived from arachidonic acid. We found that TAA interact with GPR40, which in turn mediates its cytotoxic effects on neural microvascular endothelium by inducing the endothelial proapoptotic factor TSP-1; correspondingly, GPR40 participates significantly to HI brain (and retinal) injury of the newborn and represents the first receptor identified to be involved in cytotoxic effects of nonenzymatically generated nitro-oxidative stress metabolites of arachidonic acid.
Nitro-oxidative stress has marked effects on lipids.35,36 These nonenzymatic reactions convert cis-AA–derived into different classes of nitro-peroxidation products, including isoprostanes, isofurans, and TAA, may also occur.17,25,37 However, these products are not only markers but also mediators of cytotoxicity induced by oxidant stress15,38,39; this claim applies to oxidative and nitrative products. As observed herein (Figure 1), HI encephalopathies and retinopathies exhibit increases in the nitro-peroxidation products, TAA, and corresponding inhibition of NOS, essential to generate TAA,16,18 prevents their concentrations from increasing and limit the vascular degeneration and the neural injury.5–8 It is noteworthy that iNOS expression does not change in neonatal HI injury,40 and although nNOS activity does,41,42 it is the perivascular eNOS that is involved in free radical production in neonatal HI43 and ensued nitration specifically localized initially on microvasculature (Figure 1D). Additional support of this inference is that eNOS KO mice do not express augmented levels of TAAs when subjected to experimental conditions susceptible to induce nitrative stress and ensued TAAs biosynthesis.18 This suggests that eNOS is the main isoform involved in TAA production in neonatal HI brain injury.
A major finding of this study is the role of GPR40, which interacts with TAA and mediates its cytotoxic effects on endothelial cells, thus explaining its involvement in HI brain (and retinal) injury (Figure 5 and Figure II in the online-only Data Supplement). We focused on GPR40 because it is a long chain ω-6 FA receptor, which interacts with lipids, such as linoleic and arachidonic acids, whereas the related GPR120 receptor interacts mostly with ω-3 FAs.21 We confirmed our presumption that TAA interact with GPR40 and the latter mediates its cytotoxic effects on neural microvasculature. GPR40 has been particularly studied for its ability to upregulate the glucose-induced insulin secretion on FA stimulation44; reported GPR40-coupled calcium mobilization and kinase phosphorylation20 partly coincide with our observations. GPR40 activated by long chain cis-FA seems to be a validated target for type 2 diabetes mellitus by enhancing glycemic control20,45; yet, harmful effects mediated by GPR40 activation are still debated.46,47 Along these lines, activation of GPR40 by thiazolidinediones48 but, generally, not by long chain cis-FA, is cytotoxic49; in fact, cytoprotective effects of cis-AA is described.50,51 On the contrary, TAA have been consistently found to be toxic,16–18,25 and we now found that their effects are mediated via GPR40 (Figures 2–4). This apparent discrepancy between effects of different FAs is most likely attributable to activation of distinct allosteric sites on the receptor by different molecular structures.52
The microvasculature of newborn neural tissue is particularly susceptible to nitro-peroxidant stress as observed in ischemic encephalopathies and retinopathies.5–8,53 In fact, injury to the microvasculature precedes that to the tissue parenchyma.5–7 Preservation of the microvasculature salvages (part of) the penumbral region and accordingly diminishes the size of the infarct. Our findings are consistent with these claims because inhibition of NOS which diminished formation of nitrative stress product, TAA, as well as silencing of GPR40, tended to preserve (peri-infarct) microvascular density and decreased infarct size (Figures 1 and 5).
Altogether, our findings indicate that TAAs interact with the GPR40 receptor and mediate the microvascular degeneration induced by nitrative stress resulting from HI neural insults, providing strong evidence for an unprecedented link between GPR40 and neuromicrovascular injury in immature subjects. Accordingly, GPR40 emerges as the first receptor conveying actions of nonenzymatically generated nitro-oxidative products, namely TAA in the present case, which are importantly involved in ischemic encephalopathies (and retinopathies).
Acknowledgments
We acknowledge Dr M. Prentki (Center de Recherche du CHUM) and Drs D. Lin and H. Baribault (Amgen Inc) for providing us with the human GPR40 plasmid and the GPR40 KO breeders, respectively. We also thank G. Fergusson and M. Ethier for their invaluable technical skills and help.
Sources of Funding
This study was supported in part by grants from the Canadian Institutes of Health Research and the Heart and Stroke Foundation of Quebec. Drs Chemtob and Poitout hold Canada Research Chairs, respectively, in Vision Science and in Diabetes and Pancreatic β-cell function; Chemtob also holds the Leopoldine Wolfe Chair in translational research in macular degeneration. A. Kooli was recipient of a doctoral fellowship from the Fonds de Recherche en Santé du Québec. J.-C. Honoré and T. Alquier were recipients of Focus on Stroke and Canadian Diabetes Association postdoctoral fellowships, respectively.
Disclosures
None.
Footnotes
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.112.300943/-/DC1.
- Received December 4, 2012.
- Accepted March 4, 2013.
- © 2013 American Heart Association, Inc.
References
Significance
GPR40 is the first identified G protein–coupled receptor conveying actions of nonenzymatically generated nitro-oxidative products, specifically transarachidonic acids, and is involved in (neonatal) hypoxic-ischemic encephalopathy. We demonstrated that transarachidonic acids interact with GPR40 and mediate the microvascular degeneration induced by nitrative stress resulting from hypoxic-ischemic insults, providing evidence for an unprecedented link between GPR40 and neuromicrovascular injury in immature subjects. Accordingly, GPR40 emerges as the first receptor conveying actions of nonenzymatically generated nitro-oxidative products, namely transarachidonic acids in the present case, which are importantly involved in ischemic encephalopathies.
This Issue
Jump to
Article Tools
Share this Article
- Fatty Acid Receptor Gpr40 Mediates Neuromicrovascular Degeneration Induced by Transarachidonic Acids in RodentsSignificance
