Gabriel Olmos

Tumor Necrosis Factor Alpha: A Link between Neuroinflammation and Excitotoxicity

Tumor necrosis factor alpha (TNF-α) is a proinflammatory cytokine that exerts both homeostatic and pathophysiological roles in the central nervous system. In pathological conditions, microglia release large amounts of TNF-α; this de novo production of TNF-α is an important component of the so-called neuroinflammatory response that is associated with several neurological disorders. In addition, TNF-α can potentiate glutamate-mediated cytotoxicity by two complementary mechanisms: indirectly, by inhibiting glutamate transport on astrocytes, and directly, by rapidly triggering the surface expression of Ca+2 permeable-AMPA receptors and NMDA receptors, while decreasing inhibitory GABAA receptors on neurons. Thus, the net effect of TNF-α is to alter the balance of excitation and inhibition resulting in a higher synaptic excitatory/inhibitory ratio. This review summarizes the current knowledge of the cellular and molecular mechanisms by which TNF-α links the neuroinflammatory and excitotoxic processes that occur in several neurodegenerative diseases, but with a special emphasis on amyotrophic lateral sclerosis (ALS). As microglial activation and upregulation of TNF-α expression is a common feature of several CNS diseases, as well as chronic opioid exposure and neuropathic pain, modulating TNF-α signaling may represent a valuable target for intervention.

Protection by imidazol(ine) drugs and agmatine of glutamate-induced neurotoxicity in cultured cerebellar granule cells through blockade of NMDA receptor.

This study was designed to assess the potential neuroprotective effect of several imidazol(ine) drugs and agmatine on glutamate‐induced necrosis and on apoptosis induced by low extracellular K+ in cultured cerebellar granule cells. Exposure (30 min) of energy deprived cells to L‐glutamate (1–100 μM) caused a concentration‐dependent neurotoxicity, as determined 24 h later by a decrease in the ability of the cells to metabolize 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazoliumbromide (MTT) into a reduced formazan product. L‐glutamate‐induced neurotoxicity (EC50=5 μM) was blocked by the specific NMDA receptor antagonist MK‐801 (dizocilpine). Imidazol(ine) drugs and agmatine fully prevented neurotoxicity induced by 20 μM (EC100) L‐glutamate with the rank order (EC50 in μM): antazoline (13)>cirazoline (44)>LSL 61122 [2‐styryl‐2‐imidazoline] (54)>LSL 60101 [2‐(2‐benzofuranyl) imidazole] (75)>idazoxan (90)>LSL 60129 [2‐(1,4‐benzodioxan‐6‐yl)‐4,5‐dihydroimidazole] (101)>RX821002 (2‐methoxy idazoxan) (106)>agmatine (196). No neuroprotective effect of these drugs was observed in a model of apoptotic neuronal cell death (reduction of extracellular K+) which does not involve stimulation of NMDA receptors. Imidazol(ine) drugs and agmatine fully inhibited [3H]‐(+)‐MK‐801 binding to the phencyclidine site of NMDA receptors in rat brain. The profile of drug potency protecting against L‐glutamate neurotoxicity correlated well (r=0.90) with the potency of the same compounds competing against [3H]‐(+)‐MK‐801 binding. In HEK‐293 cells transfected to express the NR1‐1a and NR2C subunits of the NMDA receptor, antazoline and agmatine produced a voltage‐ and concentration‐dependent block of glutamate‐induced currents. Analysis of the voltage dependence of the block was consistent with the presence of a binding site for antazoline located within the NMDA channel pore with an IC50 of 10–12 μM at 0 mV. It is concluded that imidazol(ine) drugs and agmatine are neuroprotective against glutamate‐induced necrotic neuronal cell death in vitro and that this effect is mediated through NMDA receptor blockade by interacting with a site located within the NMDA channel pore.

Vascular endothelial growth factor protects spinal cord motoneurons against glutamate‐induced excitotoxicity via phosphatidylinositol 3‐kinase

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterized by the selective death of motoneurons. Recently, vascular endothelial growth factor (VEGF) has been identified as a neurotrophic factor and has been implicated in the mechanisms of pathogenesis of ALS and other neurological diseases. The potential neuroprotective effects of VEGF in a rat spinal cord organotypic culture were studied in a model of chronic glutamate excitotoxicity in which glutamate transporters are inhibited by threohydroxyaspartate (THA). Particularly, we focused on the effects of VEGF in the survival and vulnerability to excitotoxicity of spinal cord motoneurons. VEGF receptor‐2 was present on spinal cord neurons, including motoneurons. Chronic (3 weeks) treatment with THA induced a significant loss of motoneurons that was inhibited by co‐exposure to VEGF (50 ng/mL). VEGF activated the phosphatidylinositol 3‐kinase/Akt (PI3‐K/Akt) signal transduction pathway in the spinal cord cultures, and the effect on motoneuron survival was fully reversed by the specific PI3‐K inhibitor, LY294002. VEGF also prevented the down‐regulation of Bcl‐2 and survivin, two proteins implicated in anti‐apoptotic and/or anti‐excitotoxic effects, after THA exposure. Together, these findings indicate that VEGF has neuroprotective effects in rat spinal cord against chronic glutamate excitotoxicity by activating the PI3‐K/Akt signal transduction pathway and also reinforce the hypothesis of the potential therapeutic effects of VEGF in the prevention of motoneuron degeneration in human ALS.

Activation of I2‐imidazoline receptors enhances supraspinal morphine analgesia in mice: a model to detect agonist and antagonist activities at these receptors

This work investigates the receptor acted upon by imidazoline compounds in the modulation of morphine analgesia. The effects of highly selective imidazoline ligands on the supraspinal antinociception induced by morphine in mice were determined. Intracerebroventricular (i.c.v.) or subcutaneous (s.c.) administration of ligands selective for the I2‐imidazoline receptor, 2‐BFI, LSL 60101, LSL 61122 and aganodine, and the non selective ligand agmatine, increased morphine antinociception in a dose‐dependent manner. Neither moxonidine, a mixed I1‐imidazoline and α2‐adrenoceptor agonist, RX821002, a potent α2‐adrenoceptor antagonist that displays low affinity at I2‐imidazoline receptors, nor the selective non‐imidazoline α2‐adrenoceptor antagonist RS‐15385‐197, modified the analgesic responses to morphine. Administration of pertussis toxin (0.25 μg per mouse, i.c.v.) 6 days before the analgesic test blocked the ability of the I2‐imidazoline ligands to potentiate morphine antinociception. The increased effect of morphine induced by I2‐imidazoline ligands (agonists) was completely reversed by idazoxan and BU 224. Identical results were obtained with IBI, which alkylates I2‐imidazoline binding sites. Thus, both agonist and antagonist properties of imidazoline ligands at the I2‐imidazoline receptors were observed. Pre‐treatment (30 min) with deprenyl, an irreversible inhibitor of monoamine oxidase B (IMAO‐B), produced an increase of morphine antinociception. Clorgyline, an irreversible IMAO‐A, given 30 min before morphine did not alter the effect of the opioid. At longer intervals (24 h) a single dose of either clorgyline or deprenyl reduced the density of I2‐imidazoline receptors and prevented the I2‐mediated potentiation of morphine analgesia. These results demonstrate functional interaction between I2‐imidazoline and opioid receptors. The involvement of Gi‐Go transducer proteins in this modulatory effect is also suggested.

Chronic treatment with the monoamine oxidase inhibitors clorgyline and pargyline down‐regulates non‐adrenoceptor [3H]‐idazoxan binding sites in the rat brain

1 The binding of [3H]‐idazoxan in the presence of 10−6m (−)‐adrenaline was used to quantitate non‐adrenoceptor idazoxan binding sites (NAIBS) in the rat brain after treatment with various psychotropic drugs. 2 Chronic treatment (14 days) with the monoamine oxidase (MAO) inhibitors clorgyline (0.3–10 mg kg−1, i.p.) and pargyline (10 mg kg−1, i.p.), but not with Ro 41–1049 (1 mg kg−1, i.p.), markedly decreased (30–50%) the density of NAIBS in the cerebral cortex without any apparent change in the affinity of the radioligand. 3 Acute (1 day) and/or chronic treatments (14 days) with other psychotropic drugs such as desipramine (3 mg kg−1, i.p.), cocaine (10 mg kg−1, i.p.), reserpine (0.12 mg kg−1, s.c.), haloperidol (1 mg kg−1, i.p.) and diazepam (10 mg kg−1, i.p.) did not alter the density of NAIBS in the cerebral cortex. 4 In vitro, the propargylamines clorgyline, pargyline and deprenyl displaced the binding of [3H]‐idazoxan to NAIBS from two distinct sites, but only clorgyline displayed an apparent very high affinity for a relevant population of NAIBS (KiH = 40 pm; KiL = 10.6 μm). The structurally diverse MAO inhibitors Ro 16–6491 (selective for MAO‐B) and Ro 41–1049 (selective for MAO‐A), as well as the other psychotropic drugs (desipramine, cocaine, reserpine and haloperidol) displaced the binding of [3H]‐idazoxan to NAIBS monophasically and with very low potencies. As expected, the MAO inhibitors clorgyline and Ro 41–1049 displaced the binding of [3H]‐Ro 41–1049 to MAO‐A monophasically and with high potencies (Ki values: 0.18 nm and 22 nm, respectively). In contrast, idazoxan displayed very low affinity (Ki = 40 μm) against the binding of [3H]‐Ro 41–1049 to MAO‐A. These results disprove a direct interaction between [3H]‐idazoxan and the enzyme MAO. 5 Preincubation of cortical membranes with clorgyline (10−9 m or 10−6 m for 30 min) or pargyline (10−6 m or 10−5 m for 30 min), reduced by 30–50% and by 17–30%, respectively, the total density of NAIBS without any apparent change in the affinity of the radioligand. Preincubation with 10−6 m clorgyline did not alter the affinity of cirazoline for the two populations of NAIBS, but reduced by 60% the binding of [3H]‐idazoxan to the high affinity site without affecting the binding of the radioligand to the low affinity site. These results indicate that the two MAO inhibitors irreversibly block the binding of [3H]‐idazoxan to NAIBS. 6 In vivo, however, various acute treatments with clorgyline (1–20 mg kg−1, i.p.) for different time intervals (6–48 h) did not alter the density of NAIBS. In vivo, only very high doses of clorgyline (40 and 80 mg kg−1, i.p.) induced modest decreases (21–28%) in the density of NAIBS in the cerebral cortex. 7 Together the results indicate that the irreversible binding of clorgyline and pargyline to NAIBS found in vitro does not fully explain the marked decreases in the density of NAIBS found in vivo after the chronic treatments. It is suggested that the down‐regulation of NAIBS induced in vivo by clorgyline and pargyline, through a direct or indirect mechanism, may have functional implications.

Attenuation of tolerance to opioid-induced antinociception and protection against morphine-induced decrease of neurofilament proteins by idazoxan and other I2-imidazoline ligands

1 Agmatine, the proposed endogenous ligand for imidazoline receptors, has been shown to attenuate tolerance to morphine‐induced antinociception ( Kolesnikov et al., 1996 ). The main aim of this study was to assess if idazoxan, an α2‐adrenoceptor antagonist that also interacts with imidazoline receptors, could also modulate opioid tolerance in rats and to establish which type of imidazoline receptors (or other receptors) are involved. 2 Antinociceptive responses to opioid drugs were determined by the tail‐flick test. The acute administration of morphine (10 mg kg−1, i.p., 30 min) or pentazocine (10 mg kg−1, i.p., 30 min) resulted in marked increases in tail‐flick latencies (TFLs). As expected, the initial antinociceptive response to the opiates was lost after chronic (13 days) treatment (tolerance). When idazoxan (10 mg kg−1, i.p.) was given chronically 30 min before the opiates it completely prevented morphine tolerance and markedly attenuated tolerance to pentazocine (TFLs increased by 71–143% at day 13). Idazoxan alone did not modify TFLs. 3 The concurrent chronic administration (10 mg kg−1, i.p., 13 days) of 2‐BFI, LSL 60101, and LSL 61122 (valldemossine), selective and potent I2‐imidazoline receptor ligands, and morphine (10 mg kg−1, i.p.), also prevented or attenuated morphine tolerance (TFLs increased by 64–172% at day 13). This attenuation of morphine tolerance was still apparent six days after discontinuation of the chronic treatment with LSL 60101‐morphine. The acute treatment with these drugs did not potentiate morphine‐induced antinociception. These drugs alone did not modify TFLs. Together, these results indicated the specific involvement of I2‐imidazoline receptors in the modulation of opioid tolerance. 4 The concurrent chronic (13 days) administration of RX821002 (10 mg kg−1, i.p.) and RS‐15385‐197 (1 mg kg−1, i.p.), selective α2‐adrenoceptor antagonists, and morphine (10 mg kg−1, i.p.), did not attenuate morphine tolerance. Similarly, the concurrent chronic treatment of moxonidine (1 mg kg−1, i.p.), a mixed I1‐imidazoline receptor and α2‐adrenoceptor agonist, and morphine (10 mg kg−1, i.p.), did not alter the development of tolerance to the opiate. These results discounted the involvement of α2‐adrenoceptors and I1‐imidazoline receptors in the modulatory effect of idazoxan on opioid tolerance. 5 Idazoxan and other imidazol(ine) drugs fully inhibited [3H]‐(+)‐MK‐801 binding to N‐methyl‐D‐aspartate (NMDA) receptors in the rat cerebral cortex with low potencies (Ki: 37–190 μM). The potencies of the imidazolines idazoxan, RX821002 and moxonidine were similar, indicating a lack of relationship between potency on NMDA receptors and ability to attenuate opioid tolerance. These results suggested that modulation of opioid tolerance by idazoxan is not related to NMDA receptors blockade. 6 Chronic treatment (13 days) with morphine (10 mg kg−1, i.p.) was associated with a marked decrease (49%) in immunolabelled neurofilament proteins (NF‐L) in the frontal cortex of morphine‐tolerant rats, suggesting the induction of neuronal damage. Chronic treatment (13 days) with idazoxan (10 mg kg−1) and LSL 60101 (10 mg kg−1) did not modify the levels of NF‐L proteins in brain. Interestingly, the concurrent chronic treatment (13 days) of idazoxan or LSL 60101 and morphine, completely reversed the morphine‐induced decrease in NF‐L immunoreactivity, suggesting a neuroprotective role for these drugs. 7 Together, the results indicate that chronic treatment with I2‐imidazoline ligands attenuates the development of tolerance to opiate drugs and may induce neuroprotective effects on chronic opiate treatment. Moreover, these findings offer the I2‐imidazoline ligands as promising therapeutic co‐adjuvants in the management of chronic pain with opiate drugs.

Inhibition of monoamine oxidase A and B activities by imidazol(ine)/guanidine drugs, nature of the interaction and distinction from I2-imidazoline receptors in rat liver

I2‐Imidazoline sites ([3H]‐idazoxan binding) have been identified on monoamine oxidase (MAO) and proposed to modulate the activity of the enzyme through an allosteric inhibitory mechanism ( Tesson et al., 1995 ). The main aim of this study was to assess the inhibitory effects and nature of the inhibition of imidazol(ine)/guanidine drugs on rat liver MAO‐A and MAO‐B isoforms and to compare their inhibitory potencies with their affinities for the sites labelled by [3H]‐clonidine in the same tissue. Competition for [3H]‐clonidine binding in rat liver mitochondrial fractions by imidazol(ine)/guanidine compounds revealed that the pharmacological profile of the interaction (2 ‐ styryl ‐ 2 ‐ imidazoline, LSL 61112>idazoxan>2 ‐ benzofuranyl ‐ 2 ‐ imidazoline, 2‐BFI=cirazoline>guanabenz>oxymetazoline>>clonidine) was typical of that for I2‐sites. Clonidine inhibited rat liver MAO‐A and MAO‐B activities with very low potency (IC50s: 700 μM and 6 mM, respectively) and displayed the typical pattern of competitive enzyme inhibition (Lineweaver‐Burk plots: increased Km and unchanged Vmax values). Other imidazol(ine)/guanidine drugs also were weak MAO inhibitors with the exception of guanabenz, 2‐BFI and cirazoline on MAO‐A (IC50s: 4–11 μM) and 2‐benzofuranyl‐2‐imidazol (LSL 60101) on MAO‐B (IC50: 16 μM). Idazoxan was a full inhibitor, although with rather low potency, on both MAO‐A and MAO‐B isoenzymes (IC50s: 280 μM and 624 μM, respectively). Kinetic analyses of MAO‐A inhibition by these drugs revealed that the interactions were competitive. For the same drugs acting on MAO‐B the interactions were of the mixed type inhibition (increased Km and decreased Vmax values), although the greater inhibitory effects on the apparent value of Vmax/Km than on the Vmax value indicated that the competitive element of the MAO‐B inhibition predominated. Competition for [3H]‐Ro 41‐1049 binding to MAO‐A or [3H]‐Ro 19‐6327 binding to MAO‐B in rat liver mitochondrial fractions by imidazol(ine)/guanidine compounds revealed that the drug inhibition constants (Ki values) were similar to the IC50 values displayed for the inhibition of MAO‐A or MAO‐B activities. In fact, very good correlations were obtained when the affinities of drugs at MAO‐A or MAO‐B catalytic sites were correlated with their potencies in inhibiting MAO‐A (r=0.92) or MAO‐B (r=0.99) activity. This further suggested a direct drug interaction with the catalytic sites of MAO‐A and MAO‐B isoforms. No significant correlations were found when the potencies of imidazol(ine)/guanidine drugs at the high affinity site (pKiH, nanomolar range) or the low‐affinity site (pKiL, micromolar range) of I2‐imidazoline receptors labelled with [3H]‐clonidine were correlated with the pIC50 values of the same drugs for inhibition of MAO‐A or MAO‐B activity. These discrepancies indicated that I2‐imidazoline receptors are not directly related to the site of action of these drugs on MAO activity in rat liver mitochondrial fractions. Although these studies cannot exclude the presence of additional binding sites on MAO that do not affect the activity of the enzyme, they would suggest that I2‐imidazoline receptors represent molecular species that are distinct from MAO.

TNF-α potentiates glutamate-induced spinal cord motoneuron death via NF-κB.

Besides glutamate excitotoxicity, the neuroinflammatory response is emerging as a relevant contributor to motoneuron loss in amyotrophic lateral sclerosis (ALS). In this regard, high levels of circulating proinflammatory cytokines such as tumor necrosis factor-alpha (TNF-α) have been shown both in human patients and in animal models of ALS. The aim of this work was to study the effects of TNF-α on glutamate-induced excitotoxicity in spinal cord motoneurons. In rat spinal cord organotypic cultures chronic glutamate excitotoxicity, induced by the glutamate-uptake inhibitor threohydroxyaspartate (THA), resulted in motoneuron loss that was associated with a neuroinflammatory response. In the presence of TNF-α, THA-induced excitotoxic motoneuron death was potentiated. Co-exposure to TNF-α and THA also resulted in down-regulation of the astroglial glutamate transporter 1 (GLT-1) and in increased extracellular glutamate levels, which were prevented by nuclear factor-kappaB (NF-κB) inhibition. Furthermore, TNF-α and THA also cooperated in the induction of oxidative stress in a mechanism involving the NF-κB signalling pathway as well. The inhibition of this pathway abrogated the exacerbation of glutamate-mediated motoneuron death induced by TNF-α. These data link two important pathogenic mechanisms, excitotoxicity and neuroinflammation, suggested to play a role in ALS and, to our knowledge, this is the first time that TNF-α-induced NF-κB activation has been reported to potentiate glutamate excitotoxicity on motononeurons.

The effects of phenelzine and other monoamine oxidase inhibitor antidepressants on brain and liver I2 imidazoline‐preferring receptors

1 The binding of [3H]‐idazoxan in the presence of 10−6 m (−)‐adrenaline was used to quantitate I2 imidazoline‐preferring receptors in the rat brain and liver after chronic treatment with various irreversible and reversible monoamine oxidase (MAO) inhibitors. 2 Chronic treatment (7–14 days) with the irreversible MAO inhibitors, phenelzine (1–20 mg kg−1, i.p.), isocarboxazid (10 mg kg−1, i.p.), clorgyline (3 mg kg−1, i.p.) and tranylcypromine (10 mg kg−1, i.p.) markedly decreased (21–71%) the density of I2 imidazoline‐preferring receptors in the rat brain and liver. In contrast, chronic treatment (7 days) with the reversible MAO‐A inhibitors, moclobemide (1 and 10 mg kg−1, i.p.) or chlordimeform (10 mg kg−1, i.p.) or with the reversible MAO‐B inhibitor Ro 16–6491 (1 and 10 mg kg−1, i.p.) did not alter the density of I2 imidazoline‐preferring receptors in the rat brain and liver; except for the higher dose of Ro 16–6491 which only decreased the density of these putative receptors in the liver (38%). 3 In vitro, phenelzine, clorgyline, 3‐phenylpropargylamine, tranylcypromine and chlordimeform displaced the binding of [3H]‐idazoxan to brain and liver I2 imidazoline‐preferring receptors from two distinct binding sites. Phenelzine, 3‐phenylpropargylamine and tranylcypromine displayed moderate affinity (KiH = 0.3–6 μm) for brain and liver I2 imidazoline‐preferring receptors; whereas chlordimeform displayed high affinity (KiH = 6 nm) for these receptors in the two tissues studied, Clorgyline displayed very high affinity for rat brain (KiH = 40 pm) but not for rat liver I2 imidazoline‐preferring receptors (KiH = 169 nm). 4 Preincubation of cortical or liver membranes with phenelzine (10−4 m for 30 min) did not alter the total density of I2 imidazoline‐preferring receptors, indicating that this irreversible MAO inhibitor does not irreversibly bind to I2 imidazoline‐preferring receptors. In contrast, preincubation with 10−6 m clorgyline reduced by 40% the Bmax of [3H]‐idazoxan to brain and liver I2 imidazoline‐preferring receptors. 5 Chronic treatment (7 days) with the inducers of cytochrome P‐450 enzymes phenobarbitone (40 or 80 mg kg−1, i.p.), 3‐methylcholanthrene (20 mg kg−1, i.p.) or 2‐methylimidazole (40 mg kg−1, i.p.) did not alter the binding parameters of [3H]‐idazoxan to brain and liver I2 imidazoline‐preferring receptors. The compound SKF 525A, a potent inhibitor of cytochrome P‐450 enzymes which forms a tight but reversible complex with the haemoprotein, completely displaced with moderate affinity (KiH = 2–10 μm) the specific binding of [3H]‐idazoxan to brain and liver I2 imidazoline‐preferring receptors. Preincubation of total liver homogenates with 3 times 10−4 m phenelzine in the presence of 10−3 m NADH, a treatment that irreversibly inactivates the haeme group of cytochrome P‐450, did not reduce the density of liver I2 imidazoline‐preferring receptors. These results discounted a possible interaction of [3H]‐idazoxan with the haeme group of cytochrome P‐450 enzymes. 6 Together the results indicate that the down‐regulation of I2 imidazoline‐preferring receptors is associated with an irreversible inactivation of MAO (at least in the brain) that is not related either to the affinity of the MAO inhibitors for I2 imidazoline‐preferring receptors or to an irreversible binding to these putative receptors. These findings indicate a novel effect of irreversible MAO inhibitors in the brain and suggest a new target for these compounds that could be of relevance in the treatment of depression, a disease in which an increased density of brain I2 imidazoline‐preferring receptors has been reported.

The effects of chronic imidazoline drug treatment on glial fibrillary acidic protein concentrations in rat brain.

1 The concentration of the astrocytic marker, glial fibrillary acidic protein (GFAP) was quantitated by immunoblotting (western blotting) in the rat brain after treatment with various imidazoline drugs and other agents. 2 Chronic (7 days) but not acute (1 day) treatment with the imidazoline drugs, cirazoline (1 mg kg−1, i.p.) and idazoxan (10 mg kg−1, i.p.), but not with the structurally related α2‐adrenoceptor antagonists, RX821002 (2‐methoxy idazoxan) (10 mg kg−1, i.p.) and efaroxan (10 mg kg−1, i.p.), markedly increased (45%) GFAP immunoreactivity in the rat cerebral cortex. Chronic treatment (7 days) with yohimbine (10 mg kg−1, i.p.), a non‐imidazoline α2‐adrenoceptor antagonist, did not significantly modify GFAP immunoreactivity in the cerebral cortex. 3 Chronic treatment (7 days) with cirazoline and idazoxan did not alter the density of brain monoamine oxidase (MAO)‐B sites labelled by [3H]‐Ro 19–6327 (lazabemide), another relevant astroglial marker. Moreover, these imidazoline drug treatments did not modify the levels of α‐tubulin in the cerebral cortex. These negative results reinforced the specificity of the effects of imidazoline drugs on GFAP. 4 Irreversible inactivation of brain α2‐adrenoceptors (and other neurotransmitters receptors) after treatment with an optimal dose of the peptide‐coupling agent EEDQ (1.6 mg kg−1, i.p., for 6–24 h) did not alter GFAP immunoreactivity in the cerebral cortex. These results further disproved the involvement of these receptors on astroglial cells in the tonic control of GFAP levels. 5 The binding of [3H]‐idazoxan in the presence of 10−6 m (–)‐adrenaline was used to quantitate in parallel I2‐imidazoline preferring sites in the rat brain after the same treatments. Chronic treatment (7 days) with cirazoline and idazoxan, but not with RX821002, efaroxan or yohimbine, significantly increased (25%) the density of I2‐sites in the cerebral cortex. The up‐regulation of I2‐sites induced by cirazoline was not observed in the liver, a tissue that also expresses I2‐sites but lacks glial cells. 6 A strong correlation (r = 0.97) was found when the mean percentage changes in GFAP immunoreactivity were related to the mean percentage changes in I2 imidazoline sites after the various drug treatments. 7 Together the results suggest a direct physiological function of glial I2‐imidazoline preferring sites in the regulation of GFAP levels.