Introduction

Ischemic stroke, which results from cerebral arterial occlusion, is becoming a major cause of morbidity and mortality in today’s society and affects millions of people every year. Currently, the only approved treatment for the acute phase of stroke is the recombinant thrombolytic tissue-type plasminogen activator.¹ Identifying molecules that contribute to the ischemic damage may help to elucidate potential therapeutic targets.

Editorial see p 2002

Clinical Perspective on p 2051

The aryl hydrocarbon receptor (AhR) or dioxin receptor is a ligand-activated transcription factor that belongs to the basic helix-loop-helix/Per-Arnt-Sim of highly conserved proteins.² One of its main functions is to participate in the metabolism of xenobiotics from environmental pollutants such as halogenated and polycyclic aromatic hydrocarbons and polychlorinated biphenyls.³ On AhR activation by xenobiotics, AhR translocates into the nucleus and dimerizes with the AhR nuclear translocator (ARNT); the heterodimer then binds to dioxin-responsive elements located upstream of target genes such as Cyp1a1, Cyp1b1, and AhRR, leading to a wide variety of toxic responses, including severe thymic involution, wasting syndrome, chloracne, immune suppression, inflammation, reduced fertility, hepatotoxicity, tumor promotion, and death.^2,4 Furthermore, AhR is postulated to play important roles in normal cell physiology and function, as demonstrated by the phenotype observed in the AhR knockout mice⁵ and by different studies that link AhR activation by endogenous or naturally occurring ligands to vascular and cardiac homeostasis,⁶ immune system function⁷ and tumor development.⁸ Although AhR and its target genes are widely expressed in different brain regions, including cerebellum, ventrolateral medulla, hippocampus, and cortex,^9–11 little is known about its role in brain physiology and pathology. Of note, AhR inhibition has been shown to protect against N-methyl-d-aspartate excitotoxicity¹² and to decrease apoptosis in cortical neurons,¹³ suggesting its potential involvement in neurological insults such as stroke.

In this study, we investigated the role of AhR in acute ischemic brain injury by using in vivo and in vitro models of stroke induced either by middle cerebral artery occlusion (MCAO) in mouse or by exposure of cultured rat cortical neurons to oxygen-glucose deprivation (OGD), respectively. Our data show that AhR activation by its endogenous ligand L-kynurenine (L-Kyn) has a deleterious role in cerebral ischemia and identify the L-Kyn/AhR pathway as a potential therapeutic target for stroke.

Methods

Reagents

The AhR antagonists 6,2’,4’-trimethoxyflavone¹⁴ (TMF) and CH223191¹⁵ (CH), the putative AhR agonist L-Kyn,^8,16–18 and the specific tryptophan 2,3-dioxygenase (TDO) inhibitor 680C91 were from TOCRIS Bioscience. The AhR agonist benzo[a]pyrene and 1-methyl-d-tryptophan, a specific indoleamine-2,3-dioxygenase (IDO) inhibitor, were from Sigma. The AhR agonist 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) was from Accustandard.

Animals

Experiments were performed in male C57Bl/6 mice at 10 to 12 weeks of age obtained from The Jackson Laboratories and in wild-type (WT) control (C57Bl/6) and AhR^+/− heterozygous mice (C57Bl/6) described elsewhere.¹⁹ WT and AhR^−/− knockout mice (C57Bl/6) were obtained from Taconic. Animals were kept in a room with controlled temperature and a 12-hour dark/light cycle and fed with standard food and water ad libitum. All experimental protocols adhered to the regulations of the Animal Welfare Committee of the Universidad Complutense (following EU directives 86/609/CEE and 2003/65/CE).

In Vivo Experimental Groups

All experiments were performed and quantified in a randomized fashion by investigators blinded to specific treatments. Mice were subjected to a distal permanent or transient MCAO by ligation (online-only Data Supplement). To evaluate the contribution of AhR, several groups were used for the determination of infarct outcome in mice 24 to 48 hours after ischemia. AhR^+/− and AhR^−/− mice and their WT littermates were subjected to MCAO. Furthermore, C57Bl/6 mice received an intraperitoneal administration of either vehicle (dimethyl sulfoxide), the AhR antagonist TMF (5 mg/kg), or the AhR antagonist CH (10 mg/kg). Another group received either vehicle or the AhR agonist benzo[a]pyrene (40 mg/kg). In another set of experiments, C57BL/6 mice were treated with either vehicle or TMF±10 mg/kg L-Kyn. To study the AhR-dependent effects of TMF, benzo[a]pyrene, and L-Kyn, AhR^−/− mice were treated with these compounds. Finally, mice were treated with vehicle, a TDO inhibitor (680C91; 10 mg/kg), or an IDO inhibitor (1-methyl-d-tryptophan; 10 mg/kg). In several experiments, coadministration of the TDO inhibitor (10 mg/kg) plus L-Kyn (10 mg/kg) was also performed. All treatments were administered 10 minutes after MCAO except for the TDO and IDO inhibitors, which were administered 10 minutes and 8 hours after ischemia, after preliminary experiments in which we failed to find a significant effect when these inhibitors were administered only 10 minutes after ischemia.

Infarct Volume Measurement and Assessment of Functional Deficits

Twenty-four and 48 hours after MCAO, infarct outcome was assessed by 2,3,5-triphenyltetrazolium chloride (TTC) staining or magnetic resonance imaging. Functional deficits were graded by use of the grip test and the modified Neurological Severity Score (online-only Data Supplement).

In Vitro Experimental Groups

OGD in rat cortical neurons was performed as described in the online-only Data Supplement. In some experiments, TMF (3–10 μmol/L) in the absence or presence of L-Kyn (50 μmol/L) was included 24 hours before, during, and after OGD. In another set of experiments, TMF (10 μmol/L) with or without L-Kyn (50 μmol/L) was included only during and after OGD. In another group, neurons were treated with the TDO inhibitor 680C91 (10 μmol/L) after OGD. In several experiments, TCDD (20 nmol/L) was used as a positive control of AhR activation. Medium and cells were collected 24 hours after OGD for viability assessment. Cells were collected at 0.5, 1, 3, 5, 8, and 24 hours after OGD for mRNA and protein assays, nuclear/cytoplasmic extracts, and immunocytochemistry.

Measurement of Cell Viability After OGD

Lactate dehydrogenase release from damaged cells and morphological analysis were used to determine neuronal cell death after OGD (online-only Data Supplement).

Statistical Analysis

Results are expressed as mean±SEM for the indicated number of experiments. Data were tested for homoscedasticity with the Bartlett test. Statistical significance was determined by use of a nonparametric, 2-tailed Mann-Whitney t test; a nonparametric, 1-way Kruskal-Wallis ANOVA test followed by a Dunn post hoc testing; or a nonparametric, 2-way ANOVA followed by Bonferroni post hoc testing. Correlation analysis was performed by use of a nonparametric Spearman correlation, and a linear regression of the data is displayed. Values of P<0.05 were considered statistically significant. In each figure, the mean value of every group labeled with a specific symbol is significantly different (P<0.05) from the mean value of the reference group, which is indicated in each case. All statistical analyses were performed with Prism version 5.0 (GraphPad Software, Inc).

Results

AhR Expression Is Increased in the Postischemic Brain

Western blot analysis showed low levels of AhR expression in sham-operated mice (n=4; P<0.05; Figure 1A). After ischemia, AhR levels increased in the peri-infarct and core regions. In the peri-infarct, AhR upregulation was observed at 18 hours, peaked at 24 to 48 hours, and started to decrease at 72 hours. In the core, AhR expression reached a plateau at 5 hours and returned to baseline at 7 days. AhR immunostaining showed similar results (n=3–4; Figure 1B). AhR nuclear translocator, the heterodimer partner of AhR, was also increased in both peri-infarct and core regions (Figure IA and IB in the online-only Data Supplement).

• Download figure
• Download PowerPoint

To determine the cell type responsible for the increased AhR expression after ischemia, we used immunofluorescence staining and confocal analysis 24 hours after MCAO (n=3–4; Figure 1C). Sham animals showed some diffuse AhR immunoreactivity in neurons at the ipsilateral and contralateral sides. Exposure to ischemia increased AhR expression mainly in neurons (AhR and NeuN colocalization) in the peri-infarct area but also in the core at different time points after MCAO (II/III layers of neocortex; Figure 1C and Figure II in the online-only Data Supplement). AhR immunoreactivity was rarely observed in astrocytes (glial fibrillary acidic protein positive; Figure 1C) and was not detectable in microglial cells (Iba-1 positive; Figure 1C).

Exposure to MCAO Induces AhR Nuclear Translocation and Transcriptional Activity in Mouse Brain

Because the activity of AhR as a transcription factor is regulated by shuttling from the cytoplasm to the nucleus,²⁰ MCAO-induced changes in AhR subcellular location were assessed by protein analysis in nuclear/cytoplasmic fractions and immunofluorescence. Western blot analysis demonstrated an initial 2-fold increase in AhR nuclear/cytoplasmic ratio followed by a second 2.4-fold increase, at 1 and 18 hours after ischemia, respectively, in brains of MCAO-exposed animals compared with the sham group, returning to basal levels after 7 days (n=5; P<0.05; Figure 1D). AhR immunoreactivity showed a diffuse cytoplasmic and nuclear pattern in sham mice (Figure 1E); in contrast, 18 hours after MCAO, nuclear translocation of AhR in neurons in peri-infarct and core regions became evident (n=4; P<0.05; Figure 1E). At this time, AhR nuclear translocator was also expressed in neurons (Figure IB in the online-only Data Supplement), allowing AhR transcriptional activity. Indeed, analysis by reverse transcriptase–polymerase chain reaction confirmed the induction of the AhR target genes Cyp1a1 and AhRR 18 hours after MCAO compared with the sham group (n=5; P<0.05; Figure 1F). These data indicate that in vivo cerebral ischemia induces AhR nuclear translocation and transcriptional activity.

AhR Participates in Acute Ischemic Damage After MCAO in Mice

We next examined whether AhR participates in ischemic brain injury. First, we used a pharmacological approach to inhibit or to activate AhR. Mice were exposed to MCAO and, 10 minutes later, were treated with vehicle or the AhR antagonists TMF or CH (Figure 2A–2E). Administration of TMF (5 mg/kg) dramatically decreased the infarct size after permanent (MCAO; n=10–11; P<0.05; Figure 2B and Figure IIIA in the online-only Data Supplement) or transient ischemia (transient MCAO; Figure IIIB in the online-only Data Supplement), determined either by magnetic resonance imaging at 24 hours or by TTC staining at 48 hours. Animals treated with TMF also presented a lower modified Neurological Severity Score 48 hours after MCAO (Figure IIIA in the online-only Data Supplement). This improvement was also seen with another AhR antagonist, CH (10 mg/kg; n=7; P<0.05; Figure 2C). Finally, AhR activation by the prototypical AhR agonist benzo[a]pyrene increased infarct volume compared with the vehicle group (40 mg/kg; n=7; P<0.05; Figure 2D); therefore, AhR inhibition or activation results in neuroprotection or potentiation, respectively, of brain damage after stroke.

• Download figure
• Download PowerPoint

To confirm these results, WT, AhR^+/−, and AhR^−/− mice were used (Figure 2E and Figure IVA–IVC in the online-only Data Supplement). First, haploinsufficiency in AhR^+/−mice, in which the expression levels of AhR and the AhR target gene AhRR after MCAO were lower than in the WT group (n=5; P<0.05), was also neuroprotective, showing a reduction in infarct volume (n=7–9; P<0.05; Figure IVD in the online-only Data Supplement) and a decrease in neurological deficits 48 hours after ischemia (Figure IVE and IVF in the online-only Data Supplement) compared with their WT littermates. Complete loss of AhR in AhR^−/− mice also resulted in neuroprotection after MCAO (n=7–11; P<0.05; Figure 2E) compared with the AhR^+/+ group. Finally, AhR^−/− mice were treated with either TMF or benzo[a]pyrene after MCAO to demonstrate their AhR-dependent effects. Two-way ANOVA demonstrated a significant effect of AhR genotype (P<0.05), treatment (P<0.05), and AhR genotype–by–treatment interaction (P<0.05). Whereas treatment of AhR^+/+ animals with the AhR antagonist TMF or the AhR agonist benzo[a]pyrene caused a reduction or an increase in the infarct volume, respectively (n=7–11; P<0.05 versus AhR^+/+ vehicle; Figure 2E), these effects were lost in AhR^−/− mice (n=7–11; P>0.05 versus AhR^−/− vehicle; Figure 2E), demonstrating that the neuroprotective or harmful actions of TMF and benzo[a]pyrene, respectively, are entirely AhR dependent. Together, these results demonstrate a detrimental role of AhR in the acute phase of ischemic injury in vivo.

AhR Regulates cAMP Response Element-Binding Protein Survival/Death Signaling After Ischemia: Effect of AhR Pharmacological Modulation

AhR has been implicated in numerous physiological and pathological processes.^2,4 The absence of changes in the inflammatory milieu after AhR inhibition (Figure V in the online-only Data Supplement) and the specific neuronal expression of AhR in our in vivo model suggest that AhR exerts its main functions in neurons after cerebral ischemia. Indeed, our findings in rat cortical neurons exposed to OGD (Figure VI in the online-only Data Supplement) corroborate the results observed in vivo, inasmuch as AhR expression and activation were increased after OGD in cortical rat neurons and AhR inhibition protected neurons against ischemic damage.

cAMP response element (CRE)–binding protein (CREB) is one of the most important neuronal transcription factors known by its implication in the expression of survival and antiapoptotic genes such as brain-derived neurotrophic factor (BDNF) on binding to CRE.^21,22 Previous studies suggested that AhR might impair CREB-mediated neuronal BDNF gene expression.¹² Therefore, we explored BDNF expression after MCAO or AhR inhibition. The AhR antagonist TMF increased cortical BDNF levels in sham-operated mice (n=6; P<0.05; Figure 3A) 24 hours after surgery. After MCAO, BDNF levels increased, showing a positive correlation with lesion size (n=5; Spearman r=0.7062, P=0.0182; Figure 3B). As a result, BDNF values were normalized for infarct volumes, and thus, we found that animals treated with TMF had higher normalized BDNF levels compared with the vehicle group (n=5; P<0.05; Figure 3B), suggesting an inhibitory effect of AhR on CREB-mediated BDNF expression.

• Download figure
• Download PowerPoint

Because CREB signaling leads to inhibition of apoptotic cascades,²³ we reasoned that AhR could modulate proapoptotic and antiapoptotic proteins after MCAO. Indeed, the AhR antagonist TMF reduced proapoptotic proteins p53 and Puma and increased the antiapoptotic Bcl-X (n=5–6; P<0.05; Figure 3C–3E). No changes were observed in Bax, Bad, and Bix (data not shown). Furthermore, cleaved caspase 3 and terminal deoxynucleotidyl transferase dUTP nick-end labeling staining showed a decreased number of apoptotic cells in the peri-infarct region of MCAO animals treated with TMF compared with the vehicle group (n=3; Figure 3F and 3G).

Thus, our data suggest that AhR activation after cerebral ischemia inhibits CREB signaling and subsequent CREB-induced prosurvival pathways. Of note, CREB and phosphorylated CREB (ser133) levels were increased in vivo and in vitro after administration of the AhR antagonist TMF (n=5–6; P<0.05; Figure 3H and Figure VIIA in the online-only Data Supplement). Accordingly, nucleofection of cortical neurons with a CRE-responsive luciferase construct demonstrated that AhR inhibition by TMF increases CREB transcriptional activity in control and OGD neurons (n=12; P<0.05; Figure 3I). The increase in mRNA levels of Bdnf and Npas4, another CREB target gene, in TMF-treated cortical neurons exposed to OGD (Figure VIIB in the online-only Data Supplement) further supports that AhR inhibits CREB transcriptional activity.

Finally, AhR coimmunoprecipitation after in vivo MCAO (n=3–4; P<0.05; Figure 3J) or in vitro OGD (n=3–4; P<0.05; Figure VIIC in the online-only Data Supplement) showed an increased interaction of AhR with CREB-binding protein (CBP) and with CREB, suggesting that the AhR inhibitory effect on CREB signaling could be due, at least in part, to a direct interaction with the CREB/CBP complex. Consistently, AhR inhibition by TMF significantly reduced receptor association to both CBP and CREB.

L-Kyn Is an Endogenous AhR Ligand in Cerebral Ischemia

A variety of endogenous ligands have been demonstrated to activate AhR in different models.³ Our findings strongly suggest that AhR is activated by an endogenous agonist generated by cerebral ischemic damage. L-Kyn, a key metabolite of the kynurenine pathway (Figure 4A), has recently been identified as an in vivo AhR agonist.⁸ Because this pathway is activated after stroke,²⁴ we explored whether L-Kyn could account for endogenous AhR activation in cerebral ischemia.

• Download figure
• Download PowerPoint

First, levels of L-Kyn and its precursor, L-tryptophan (L-Trp), were determined in the ipsilateral cortex and in plasma (Figure 4B–4E). Two-way ANOVA showed a significant effect of surgery (P<0.05), time (P<0.05), and surgery-by-time interaction (P<0.05) in brain L-Kyn and L-Trp levels but not in plasma. Brain L-Kyn levels increased as early as 3 hours after MCAO and remained elevated 24 hours after (n=7–10; P<0.05; Figure 4B). This increase was associated with an initial increase and a subsequent reduction in L-Trp levels from 3 to 24 hours after MCAO (n=7–10; P<0.05; Figure 4C). After MCAO, no changes were observed in plasma L-Kyn and L-Trp levels (n=7–14; Figure 4D and 4E).

The specific increase in L-Kyn after MCAO in brain but not in plasma made us reason that local L-Kyn biosynthesis could be induced after ischemia. Consequently, we tested brain mRNA expression of Ido and Tdo, the 2 key enzymes in the initial step of the kynurenine pathway (Figure 4A).²⁵ Whereas no changes were detected in Ido1 and Ido2 mRNA levels (n=5; Figure 4F), ischemia increased the expression of Tdo compared with sham animals (n=5; P<0.05; Figure 4F). This result was confirmed by immunofluorescence (n=3; Figure 4G): Whereas sham animals presented low TDO levels in the ipsilateral cortex, MCAO increased TDO expression in the core and peri-infarct regions 5 and 24 hours after the occlusion. Colocalization experiments demonstrated that TDO is located mainly in neurons and by some astrocytes (n=3; Figure 4H). Ischemia-induced TDO upregulation indicates that, in addition to the augmented peripheral L-Kyn synthesis after stroke,²⁴ local L-Kyn production is increased and is expected to contribute to brain L-Kyn levels after MCAO.

Hence, L-Kyn could be the endogenous AhR agonist in this setting. Thus, we decided to examine the effects of L-Kyn in cultured cortical neurons. First, L-Kyn activated AhR in neurons 0.5 to 1 hour after the treatment as shown by an increased nuclear accumulation of AhR, comparable to that induced by TCDD (Figure VIIIA and VIIIB in the online-only Data Supplement). Moreover, L-Kyn increased the expression of the AhR target genes Cyp1a1 and Cyp1b1 mRNA in cortical neurons, an effect blocked by TMF (Figure VIIIC and VIIID in the online-only Data Supplement). Finally, L-Kyn decreased CRE-mediated transcription in neurons, demonstrated by a reduction in both Bdnf and Npas4 mRNA expression and in the CRE-luciferase reporter assay (Figure IXA and IXB in the online-only Data Supplement). Together, these data indicate that L-Kyn is a specific AhR agonist in neurons that could mediate AhR-dependent inhibition of CREB signaling after experimental stroke.

L-Kyn Plays a Deleterious Role in Cerebral Ischemia Through AhR Activation

The neuroprotection achieved by AhR inhibition in in vivo and in vitro ischemia implies a deleterious role of L-Kyn through AhR activation. Indeed, L-Kyn aggravated OGD-induced neuronal damage, assessed by morphological analysis and lactate dehydrogenase release determination (n=24; P<0.05; Figure 5A and 5B), an effect reversed by TMF, indicating that L-Kyn–induced deleterious effects after OGD depend on AhR activation. In vivo, intraperitoneal administration of L-Kyn (10 mg/kg) after MCAO increased brain and plasma L-Kyn levels (n=5–10; P<0.05; Figure 5C and 5D) concomitantly with an increase in infarct volume (n=7–10; P<0.05; Figure 5E and 5F) compared with the vehicle group, an effect not observed at a lower dose of L-Kyn (5 mg/kg; n=7; P>0.05; Figure 5E). Moreover, the L-Kyn–induced increase in infarct volume was AhR dependent; it was absent after coadministration of the AhR antagonist TMF or in AhR^−/− mice (n=7; P<0.05; Figure 5E and 5F).

• Download figure
• Download PowerPoint

Having demonstrated that exogenous L-Kyn has an AhR-dependent, detrimental role after ischemia, we finally explored whether inhibition of L-Kyn biosynthesis was neuroprotective in this setting. The TDO inhibitor 680C91 reduced infarct volume compared with the vehicle-treated group (n=10; P<0.05; Figure 6A). However, no differences were observed between vehicle and the IDO inhibitor 1-methyl-d-tryptophan (n=7; Figure 6B), supporting that brain TDO activity accounts for L-Kyn biosynthesis after stroke. Coadministration of L-Kyn abolished the neuroprotective actions of the TDO inhibitor (n=7; P<0.05; Figure 6C). In addition, TDO inhibition after MCAO abolished the postischemic brain L-Kyn augmentation while producing an accumulation of L-Trp (n=5–7; P<0.05; Figure 6D and 6E), reinforcing that neuroprotective actions of TDO inhibition are mediated by a reduction in L-Kyn synthesis after stroke.

• Download figure
• Download PowerPoint

Finally, we investigated whether TDO inhibition, by diminishing L-Kyn levels, might decrease AhR activation. Indeed, the TDO inhibitor decreased AhR activation after in vivo (n=7; P<0.05; Figure 6F) and in vitro (n=5; P<0.05; Figure 6G–6I) cerebral ischemia. Together, our results demonstrate that inhibition of TDO-mediated L-Kyn biosynthesis decreases AhR activation and its subsequent deleterious effects after experimental stroke.

Discussion

AhR or dioxin receptor is a transcription factor with important roles in xenobiotic-induced toxicity and carcinogenesis but also in some physiological functions such as reproduction, organ homeostasis, and adaptive immunity.^2,7,26 Now, the results presented here show that cerebral ischemia induces AhR overexpression and activation in neurons, which participate in ischemic brain damage concomitantly with the downregulation of CREB prosurvival pathways. We have also found that cerebral ischemia induces the synthesis of L-Kyn, an L-Trp metabolite with AhR agonistic properties that may account for the endogenous activation of AhR in this setting. These data are the first to demonstrate a pathophysiological role of this receptor in the context of stroke, a high-incidence cerebrovascular pathology with limited therapeutic treatment.

We have found that AhR is overexpressed in mouse brain after experimental stroke, mainly in neurons located at the peri-infarct but also at the ischemic core. Likewise, AhR expression is found in cultured rat cortical neurons, as previously reported,^12,13,27 and increases after in vitro experimental ischemia. To the best of our knowledge, this is the first evidence showing that AhR signaling pathways are induced in neurons after a deleterious stimulus such as cerebral ischemia. The early profiles of AhR activation and transcriptional activity strongly point to a damaging role of this receptor in stroke physiopathology.

Indeed, using pharmacological and genetic loss-of-function approaches, we have demonstrated the detrimental role of AhR in the ischemic brain. After in vivo experimental stroke by permanent MCAO, 2 different AhR antagonists, TMF¹⁴ and CH,¹⁵ were neuroprotective. The antagonism of AhR (with TMF) was also beneficial when ischemia was followed by recanalization (transient MCAO). Conversely, AhR activation by the prototypical AhR agonist benzo[a]pyrene aggravated stroke outcome. In addition, either AhR haploinsufficiency or total deletion decreased ischemic damage and improved neurological scores. The ligands used failed to affect stroke outcome in AhR^−/− mice, showing that their actions are AhR dependent. The deleterious effects of in vivo AhR activation were confirmed in cortical neurons exposed to OGD. All these data together demonstrate that AhR participates in acute cerebral damage after ischemic stroke and that specific antagonists of this receptor such as TMF may be useful for inhibiting the detrimental actions of this receptor after cerebral ischemia.

Several studies have demonstrated that AhR inhibition reduces neuronal apoptosis^27,28 and neurotoxicity concomitant with an increase in BDNF.^12,13 In this context, CREB is a critical transcription factor against ischemic injury because of its target genes that include the neuroprotective neurotrophin BDNF and the so-called activity-regulated inhibitor of death genes or by antagonizing apoptotic death signals.^29–31 Indeed, our results show that the AhR antagonist TMF increased normalized BDNF levels and decreased apoptosis by affecting apoptosis-related genes after MCAO. In vitro studies confirmed that AhR inhibits CREB transcriptional activity, as revealed by an upregulation of BDNF and NPAS4 and by an increase in CRE-responsive luciferase activity in cultured neurons exposed to OGD and AhR inhibition. These data strongly suggest that inhibition of CREB signaling participates in the detrimental role of AhR after stroke. Because in vivo and in vitro coimmunoprecipitation experiments of AhR demonstrate its direct interaction with CBP and CREB, our results support that cerebral ischemia increases AhR recruitment to CBP/CREB complexes playing an inhibitory role in CREB survival genes.

Supporting that AhR is an important modulator of physiological functions, several endogenous AhR ligands have been described. Our study demonstrates that one of them, L-Kyn,^8,16–18 an L-Trp derivative through the so-called kynurenine pathway, is a key AhR ligand in neurons. Interestingly, in both animals and humans, the kynurenine pathway is increased after stroke and is associated with stroke severity and poor prognosis.^24,32 Consistently, we found that brain L-Kyn levels increased as early as 3 hours and remained elevated 24 hours after MCAO, coinciding with the maximal nuclear AhR expression. In vitro, L-Kyn induced AhR nuclear translocation and upregulation of AhR target genes in a manner similar to the classic AhR agonist TCDD, effects that were inhibited by the AhR antagonist TMF or in AhR^−/− mice. All these results support that, in the ischemic setting, L-Kyn is the endogenous AhR ligand responsible for neuronal AhR activation.

In normal brain, 60% of L-Kyn is taken up from blood,³³ and 40% is generated locally from L-Trp, which is transported across the blood-brain barrier. The initial step of the brain kynurenine pathway is carried out mostly by 2 key enzymes, IDO and TDO, which metabolize L-Trp into different metabolites, including L-Kyn. TDO is expressed in mainly liver but is also present in brain,^34–36 suggesting a role in brain function. Supporting this, we demonstrate that TDO, but not IDO, is upregulated predominantly in neurons of the ipsilateral cortex of the ischemic brain. In addition, L-Trp, an amino acid that regulates TDO function,³⁷ is significantly increased in brain after ischemia. All these data point toward L-Kyn synthesis by TDO as the main pathway for the augmented L-Kyn concentration in the ischemic brain. Importantly, TDO inhibition abolished, at early times, both the L-Kyn increase and AhR nuclear translocation in the ischemic brain, as well as AhR activation and transcriptional activity after OGD in cultured neurons, strongly suggesting that, in vivo, AhR is activated by L-Kyn very rapidly after the ischemic occlusion and is thus causally involved in brain damage at those crucial times at which lesion is expanding after the arrest of the blood flow in the affected area. We cannot disregard that some cerebral L- Kyn may derive from the periphery because large amounts of L-Kyn are produced in the liver through L-Trp degradation by TDO.³⁸ In addition, local synthesis of L-Kyn by microglia, astrocytes, or macrophages in the ischemic brain could also occur in an inflammatory microenvironment.³⁹ Of note, both AhR and TDO are coexpressed by neurons placed in the infarct border (data not shown), suggesting both a paracrine and an autocrine model for AhR activation by L-Kyn.

Therefore, apart from AhR, we have identified TDO as an important therapeutic target because its inhibition drastically reduced brain ischemic damage concomitantly with a decrease in brain L-Kyn concentration and with an inhibition of AhR activation. From our data, we can conclude that a cascade of events, initiated by the early activation and overexpression of AhR and TDO-mediated L-Kyn biosynthesis followed by the subsequent downregulation of endogenous neuroprotective or antiapoptotic pathways, participates in brain damage (lesion size and neurological function) after ischemic stroke.

The kynurenine pathway is becoming recognized as a key player in the mechanisms of neuronal damage in several neurodegenerative disorders (for a review, see Vécsei et al²⁵). L-Kyn has a central role on this pathway because its degradation by different enzymes generates neuroactive metabolites such as quinolinic acid and kynurenic acid, which show agonistic and antagonistic N-methyl-d-aspartate receptor properties, respectively. Accordingly, studies on the effect of L-Kyn are controversial. L-Kyn, apparently by elevating brain kynurenic acid levels, results in neuroprotection when administered before hypoxia/ischemia and N-methyl-d-aspartate lesions^40–43; conversely, and in agreement with our data, postischemic L-Kyn administration exacerbated acute neuronal damage after MCAO.⁴⁴ Therefore, direct actions of L-Kyn, without the involvement of its downstream metabolites, remained obscure. In addition, although AhR-dependent, immunosuppressive, and anti-inflammatory properties of L-Kyn have been reported in several systems,^8,16–18 we show here that they do not play a role in our setting. To the best of our knowledge, our present data are the first to prove that the neurotoxic actions of L-Kyn are AhR dependent, thus providing a novel mechanistic insight for previously described deleterious actions of L-Kyn in the central nervous system.

Conclusions

Our study links both L-Kyn and AhR signaling routes in ischemic brain injury and thus recognizes the L-Kyn/AhR pathway as a potential therapeutic target in this setting. Our findings also demonstrate that this pathway may be therapeutically targeted at 2 different levels, L-Kyn biosynthesis by TDO and AhR activation, posing novel possibilities for the treatment of acute stroke. Furthermore, our results open new lines of investigation into the effects of the interference of this pathway at later time points or when the occluded blood vessel is recanalized by recombinant tissue-type plasminogen activator, as in clinical practice, or the consequences that polymorphisms of this receptor may have on stroke outcome. In addition, considering the well-described role of AhR in xenobiotic-induced toxicity and carcinogenesis, further studies are required to determine whether exposure to environmental contaminants such as dioxin could affect susceptibility to stroke damage in patients at risk or even the development of idiopathic neurodegenerative diseases.

Footnotes