Search

Purpose: Retinal ischemia leads to neuronal death. The effects of riluzole, a drug that protects against the deleterious effect of cerebral ischemia by acting on several types of ion channels and blocking glutamatergic neurotransmission, were investigated in a rat model of retinal ischemic injury., Methods: Retinal ischemia was induced by increasing intraocular pressure above systolic blood pressure for 30 minutes. Electroretinograms were recorded before ischemia and at different periods of reperfusion. Riluzole was injected or topically applied to the eye before or after ischemia and twice daily during the reperfusion period. Retinas were harvested for histopathology (toluidine blue and silver-impregnation stainings, Tdt-dUTP terminal nick-end labeling [TUNEL] method) and immunohistochemistry for cytoskeletal glial fibrillary acid protein and c-jun NH2-terminal kinase (p-JNK)., Results: Ischemia for 30 minutes caused a reduction of a- and b-waves of the electroretinogram. Systemic and topical treatments with riluzole significantly enhanced the recovery of the reduced a- and b-waves after defined reperfusion times. Riluzole also prevented or attenuated ischemia-induced retinal cell death (necrosis and apoptosis) and reduced the activation of p-JNK, c-jun phosphorylation, and the increase of cytoskeletal proteins induced by ischemic injury., Conclusions: Riluzole acted in vivo as a potent neuroprotective agent against pressure-induced ischemia. Therefore, riluzole may be a major drug for use in protection against retinal injury.

TWIK-1, TREK-1 and TASK K+ channels comprise a class of pore-forming subunits with four membrane-spanning segments and two P domains. Here we report the cloning of TRAAK, a 398 amino acid protein which is a new member of this mammalian class of K+ channels. Unlike TWIK-1, TREK-1 and TASK which are widely distributed in many different mouse tissues, TRAAK is present exclusively in brain, spinal cord and retina. Expression of TRAAK in Xenopus oocytes and COS cells induces instantaneous and non-inactivating currents that are not gated by voltage. These currents are only partially inhibited by Ba2+ at high concentrations and are insensitive to the other classical K+ channel blockers tetraethylammonium, 4-aminopyridine and Cs+. A particularly salient feature of TRAAK is that they can be stimulated by arachidonic acid (AA) and other unsaturated fatty acids but not by saturated fatty acids. These channels probably correspond to the functional class of fatty acid-stimulated K+ currents that recently were identified in native neuronal cells but have not yet been cloned. These TRAAK channels might be essential in normal physiological processes in which AA is known to play an important role, such as synaptic transmission, and also in pathophysiological processes such as brain ischemia. TRAAK channels are stimulated by the neuroprotective drug riluzole.

TASK is a new member of the recently recognized TWIK K+ channel family. This 395 amino acid polypeptide has four transmembrane segments and two P domains. In adult human, TASK transcripts are found in pancreas

Acid-sensing is associated with both nociception and taste transduction. Stimulation of sensory neurons by acid is of particular interest, because acidosis accompanies many painful inflammatory and ischaemic conditions. The pain caused by acids is thought to be mediated by H+-gated cation channels present in sensory neurons. We have now cloned a H+-gated channel (ASIC, for acid-sensing ionic channel) that belongs to the amiloride-sensitive Na+ channel/degenerin family of ion channels. Heterologous expression of ASIC induces an amiloride-sensitive cation (Na+ > Ca2+ > K+) channel which is transiently activated by rapid extracellular acidification. The biophysical and pharmacological properties of the ASIC channel closely match the H+-gated cation channel described in sensory neurons. ASIC is expressed in dorsal root ganglia and is also distributed widely throughout the brain. ASIC appears to be the simplest of ligand-gated channels.

The two P domain K+ channel mTWIK-1 has been cloned from mouse brain. In Xenopus oocytes, mTWIK-1 currents are K+-selective, instantaneous, and weakly inward rectifying. These currents are blocked by Ba2+ and quinine, decreased by protein kinase C and increased by internal acidification. The apparent molecular weight of mTWIK-1 in brain is 81 kDa. A 40 kDa form is revealed after treatment with a reducing agent, strongly suggesting that native mTWIK-1 subunits dimerize via a disulfide bridge. TWIK-1 mRNA is expressed abundantly in brain and at lower levels in lung, kidney, and skeletal muscle. In situ hybridization shows that mTWIK-1 expression is restricted to a few brain regions, with the highest levels in cerebellar granule cells, brainstem, hippocampus and cerebral cortex.

Chimeras have been constructed using three different fragments (N-terminal, central and C-terminal) of IRK3, a constitutive inward rectifier K+ channel subunit, and GIRK2, a G-protein activated inward rectifier K+ channel subunit and have been coinjected into Xenopus oocytes together with IRK3 or IRK1 (another constitutive inward rectifier) cRNA. Both IRK1 and IRK3 expression was inhibited by coinjection with chimeras containing a N-terminal fragment of IRK3 suggesting that subunits of K+ channels in the IRK family form a functional multimeric assembly where the N-terminal end has an important role. In situ hybridization shows that IRK1 and IRK3 are coexpressed in the same areas of the brain and probably in the same cells. Taken together both the localization and the oocyte expression results suggest that not only homomultimeric IRK1 or homomultimeric IRK3 assemblies take place but that heteromultimeric IRK1/IRK3 assemblies are also formed.

Severe forebrain ischemia induces a large increase in expression of NMDA receptor subunits in rat brain. One week after ischemia, levels of NMDA-R1 mRNAs in the CA1 pyramidal cells of hippocampus are 7 times higher than those observed in control rats. At 7 days postischemia, an enhanced immunostaining of the NMDA-R1 subunit was observed in all hippocampal structures indicating that changes in mRNA levels are accompanied by changes in receptor protein level. Riluzole, a potent inhibitor of glutamate release and CNQX, a selective AMPA/kainate antagonist, drastically reduced the ischemia-induced expression of mRNAs for the three NMDA receptor subunits while D-AP5, a selective NMDA antagonist, had essentially no effect. Therefore ischemia-induced expression of NMDA receptor subunits is associated with glutamate release and proceeds via an AMPA/kainate pathway. These results together with those of other groups concerning ischemia effects on AMPA and GABAA receptor levels, suggest an important role of the induced expression of NMDA receptor subunits in the deleterious effects of ischemia.

A cDNA encoding a N-type Ca2+ channel has been cloned from the murine neuroblastoma cell line N1A103. The open reading frame encodes a protein of 2,289 amino acids (257 kDa). Analysis of different clones provided evidence for the existence of distinct isoforms of N-type channels. High levels of mRNA were found in the pyramidal cell layers CA1, CA2 and CA3 of the hippocampus, in the dentate gyrus, in the cortex layers 2 and 4, in the subiculum and the habenula. The N-type Ca2+ channel gene has been localized on the chromosome 2, band A.

The present study analyses the effects of learning on the spatial pattern and the time-course of changes of immediate early gene messenger RNA's (c-fos and c-jun) in mouse brain produced by training in an appetitive bar-pressing task. Activation of c-fos and c-jun after training is strictly located in the hippocampal formation and is learning-dependent. Levels of both proto-oncogene mRNAs in the trained group were 4 to 5 times higher than in the sham-conditioned group. Injections of apamin, a bee venom neurotoxin that selectively blocks a class of Ca(2+)-activated K+ channels and improves learning and memory retention, produced as compared to untrained animals a 3- to 5-fold increase of expression of c-fos and c-jun with the same pattern as that observed in the trained animals. Post-training injection of 0.2 mg/kg apamin enhanced 1.4-fold the expression of both immediate early genes in CA1, CA3 and dentate gyrus as compared to trained saline-injected mice. All these results suggest that apamin-induced increase of immediate early gene expression might be related to the apamin-induced facilitation of learning.

The effects of the potent ATP-sensitive K+ channel blocker glipizide were measured on the locomotor activity of rats after bilateral intracerebral administration into the dorsal pallidum. Glipizide (10 pmol) was found to reduce spontaneous locomotor activity measured during the night cycle of the rats, whereas the ATP-sensitive K+ channel activator (-)-cromakalim (5 fmol) enhanced spontaneous locomotor activity. Glipizide (0.5, 2.5 and 10 pmol) was also found to depress noticeably d-amphetamine-induced locomotor activity (1 mg/kg s.c.). These results are in agreement with the idea that ATP-dependent potassium channels within the dorsal pallidum are involved in controlling motor activity in the rat.

Mast cell degranulating (MCD) peptide and Dendrotoxin I (DTXI), two potent hyperexcitability-inducing toxins acting on voltage-dependent potassium channels, induce the expression of both c-fos and c-jun mRNA in i.c.v. treated rats. The distribution of c-fos and c-jun expression has been analyzed by in situ hybridization. The expression is particularly high in the cerebral cortex and hippocampus for both toxins. However differences of expression between MCD and DTXI-treated animals have been observed in hypothalamus and thalamic and amygdaloid nuclei. Moreover, brain areas such as cerebellum which have high amounts of binding sites for both MCD and DTXI do not show any induction of c-fos and c-jun. Lemakalim, a K+ channel opener, prevents the MCD-induced activation of both 'immediate-early genes' in all brain areas but is unable to inhibit the induction of c-fos and c-jun induced by DTXI. These two toxins which are generally believed--from molecular approaches--to act on the same voltage-dependent K+ channel, clearly act in vivo on two distinct classes of channels.

Using the stable isotopes of lithium 6Li and 7Li, and the nuclear reaction 6Li(n,alpha)3H for detection, we have studied the isotopic exchange of lithium in various areas of the mouse brain and in the mouse plasma, under conditions of constant concentration of total lithium. The neutron irradiations were performed using 'cold' neutrons, at the European Institute Von Laue-Langevin. The nuclear reaction track densities were determined using an automatic image analyser. In the plasma, the isotopic ratios, 6Li/7Li, were measured using 'Secondary Ion Mass Spectrometry'. The concentration of total lithium in the plasma was kept close to 0.28 mM. The brain concentration of total lithium (referred to the tissue water content) ranged from more than 2 mM in the thalamus to less than 0.65 mM in the white matter of the cerebellum. The Nernst potential of lithium thus ranged from approx. -50 to approx. -20 mV, which means that lithium is probably not far from electrochemical equilibrium between brain cells and plasma. At any moment, the isotopic abundance of 6Li (ratio of 6Li to total lithium) in the different brain areas, were not significantly different from one another. The time-course of the isotopic abundance of 6Li in the brain was fitted by the composition of two exponential terms. The time-course of the isotopic abundance of 6Li in the plasma was also fitted by the composition of two exponential terms. These analytic curves (for the brain and for the plasma) were not significantly different from each other, at the precision of the measurements. This means that the isotopic equilibration of lithium between brain and plasma is almost instantaneous (i.e. accomplished in a few min at the most).

The compartmental analysis of lithium in the mouse plasma has been performed using the stable isotopes, 6Li and 7Li, as tracers. The animals were kept under stationary conditions (concentration of total lithium in plasma maintained equal to 0.28 mM) during the experiment. The isotopic exchange, 6Li/7Li, was described to a good approximation by combining two first-order processes (characteristic parameters: 89 microM and 2.071 X 10(-3) X min-1 for the first one, and 193 microM and 2.215 X 10(-4) X min-1 for the second one). From these kinetic data, it was estimated that the lithium capacities of the plasma and of the cells were 282 and 454 nmol X ml-1, and that the unidirectional fluxes of lithium between plasma and cells and through the kidneys were 0.259 and 0.227 nmol X ml-1 X mn-1. Since the method makes use only of stable, hence harmless, isotopes, one might think of extending it to direct experimentation in man.

Using nuclear reaction 6Li(n, alpha)3H and dielectric detectors, we have studied the distribution of Li in the brain of adult mice, following Li treatment of the animals. Two strains of animals were used in parallel: "quaking" dysmyelinating mutants and normally myelinated controls. The distribution appeared to be sharply regionalized in the brain of the normal mice (higher Li concentration in the gray rather than in the white matter, with the area postrema being particularly Li rich). In contrast, the Li distribution was practically homogeneous in the brain of the quaking dysmyelinating mutants, with a mean Li concentration comparable to that in the gray matter of the controls. The present method of Li detection has made it possible to estimate the Li equilibrium potentials (nerve cells with regard to plasma) in the different brain substructures. The results are consistent with (a) Li being actively extruded from nerve cells in all the cases and (b) myelination decreasing the relative importance of the passive component of Li transport in the nerve cells, as compared with the active component.

Lithium is detected in the mouse brain with the aid of a (n,alpha) nuclear reaction. The resolving power is of a few micron. The detection efficiency is estimated in a first series of experiments with brain sections of progressively increasing thickness. This allows us to make the microlocation of lithium in the brain sections quantitative. Lithium accumulation in the different brain substructures is then measured for increasing lithium doses given to the animals (1--15 meq/kg animal). For measurements made 6 h after the last injection of lithium, it is corroborated that the accumulation of Li is much larger in the areas with cellular bodies of neurons than in those with nerve trunks only, especially for the smaller doses of lithium given as treatment. Whereas the accumulation curve as a function of the Li concentration of the blood is linear for areas with nerve trunks only, it is negatively bent for the areas with cellular bodies. This and previous data allow us to attempt an estimate of the Li content of the cellular bodies only. A saturation effect seems to appear as soon as the Li concentration in the blood is greater than 1 mM.

The time-constant of lithium (Li)-clearance from the plasma was of the order of 2 hr, while the mean time-constant of lithium-clearance from the bulk of the body cells into the plasma was about 10 times this value. The mean time-constant of lithium ion (Li+)-clearance from various brain substructures (corpus callosum, striatum, hippocampus, neocortex) was comparable to the latter (15-25 hr). Comparison of samples where either a slight diffusion or no diffusion at all had occurred, suggest that between the nerve bodies and the trunks, lithium ions are not uniformly distributed.