Involvement of the trisynaptic hippocampal pathway in generating neural representations of object–place associations (an analytical review)

The possible mechanisms by which neural representations of object–place associations are generated in different parts of the network consisting of the hippocampus and the parahippocampal complex are analyzed. Spatial and non-spatial information arrives in the hippocampus via two streams from the parahippocampal complex, which consists of the perirhinal, postrhinal, and entorhinal areas of the cortex. It can be suggested that because there are no connections between the lateral and medial areas of the entorhinal cortex, these representations, as particular patterns of connected and discharging neurons, are generated mainly in the hippocampus, though they may also be generated in the entorhinal cortex because of the input from the postrhinal cortex. As both information streams converge on neurons in the dentate gyrus and field CA3, the trisynaptic pathway through the hippocampus may play a key role in generating these representations. As spatial information arrives in the neocortex and passes from there via the parahippocampal complex to the hippocampus about 20 msec earlier than non-spatial information, spatial information is processed first in the dentate gyrus and field CA3. Later, because of the return of excitation from field CA3c to the dentate gyrus, neural representations of object–place associations start to be generated in the dentate gyrus. Signals are transferred from the dentate gyrus to field CA3, where information arriving from the entorhinal cortex is superimposed on the neuronal patterns activated by these signals. As a result, more complex neural representations are generated in field CA3 and signals are sent to field CA1. In the dorsal (ventral) part of field CA1, non-spatial (spatial) information arriving from the lateral (medial) part of the entorhinal cortex is superimposed on the activated neuronal pattern. The result is that higher-order representations are generated in field CA1. In the parahippocampal cortex, the generation of neuronal representations of object–place associations can result from the transfer of activity from the dorsal part of hippocampal field CA1.     

Pentylenetetrazole-induced seizures affect binding site densities for GABA, glutamate and adenosine receptors in the rat brain

Pentylenetetrazole (PTZ) is a convulsant used to model epileptic seizures in rats. In the PTZ-model, altered heat shock protein 27 (HSP-27) expression highlights seizure-affected astrocytes, which play an important role in glutamate and GABA metabolism. This raises the question whether impaired neurotransmitter metabolism leads to an imbalance in neurotransmitter receptor expression. Consequently, we investigated the effects of seizures on the densities of seven different neurotransmitter receptors in rats which were repeatedly treated with PTZ (40 mg/kg) over a period of 14 days. Quantitative in vitro receptor autoradiography was used to measure the regional binding site densities of the glutamate α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), kainate and N-methyl-d-aspartate (NMDA) receptors, the adenosine receptor type 1 (A₁), which is part of the system controlling glutamate release, and the γ-aminobutyric acid (GABA) receptors GABA_A and GABA_B as well as the GABA_A-associated benzodiazepine (BZ) binding sites in each rat. Our results demonstrate altered receptor densities in brain regions of PTZ-treated animals, including the HSP-27 expressing foci (i.e. amygdala, piriform and entorhinal cortex, dentate gyrus). A general decrease of kainate receptor densities was observed together with an increase of NMDA binding sites in the hippocampus, the somatosensory, piriform and the entorhinal cortices. Furthermore, A₁ binding sites were decreased in the amygdala and hippocampal CA1 region (CA1), while BZ binding sites were increased in the dentate gyrus and CA1. Our data demonstrate the impact of PTZ induced seizures on the densities of kainate, NMDA, A₁ and BZ binding sites in epileptic brain. These changes are not restricted to regions showing glial impairment. Thus, an altered balance between different excitatory (NMDA) and modulatory receptors (A₁, BZ binding sites, kainate) shows a much wider regional distribution than that of glial HSP-27 expression, indicating that receptor changes are not following the glial stress responses, but may precede the HSP-27 expression.     

Repeated seizures cause a generalized increase in excitability in the hippocampus

Changes in excitability within the hippocampus were examined by recording in CA3 and in CA1 or the dentate gyrus and stimulating in the angular bundle (fiber bundle connecting the entorhinal cortex with the dentate gyrus) before and after 36 electrographic seizures in urethane anesthetized rats. The responses recorded in CA3 or CA1 to stimulation of the angular bundle every 30 s were unchanged after repeated seizures, while the responses in the dentate gyrus were augmented. When the angular bundle was stimulated every 200 ms, population spikes appeared in CA3 and CA1, with CA1 producing multiple population spikes before CA3. The results indicate that after repeated seizures there is a general enhancement of excitability throughout the hippocampus.     

Expression level of vascular endothelial growth factor in hippocampus is associated with cognitive impairment in patients with Alzheimer's disease

Although the enhanced expression of vascular endothelial growth factor (VEGF) in the brains of patients with Alzheimer's disease (AD) has been reported, the functional significance of VEGF level in the progression of AD is still unclear. We examined the VEGF expression in the hippocampus of patients with AD at different stages of progression by Western blot analysis, and found that the VEGF189 isoform (VEGF₁₈₉) was barely detectable in normal hippocampus, but significantly increased at the early stage of patients with AD. VEGF₁₈₉ was decreased with advancing stages of AD. Immunostaining shows that VEGF was significantly increased in the cells in the CA1, CA3, and dentate gyrus regions of hippocampus and layers III and V of entorhinal cortex of patients with AD, compared with normal brain. Confocal images show that VEGF was predominantly expressed in neurons and astrocytes in the hippocampus and entorhinal cortex of patients with AD. Our data suggest that VEGF level is associated with progressive loss of cognitive function in patients with AD.     

Colocalization of mineralocorticoid receptor and glucocorticoid receptor in the hippocampus and hypothalamus

We investigated the distribution and colocalization pattern of the two corticosteroid receptors, mineralocorticoid receptor (MR) and glucocorticoid receptor (GR), in the hippocampus and hypothalamus, the main target regions of corticosterone in the rat brain, using double immunofluorescence histochemistry in conjunction with specific polyclonal antibodies against MR and GR. In the hippocampus, MR- and GR-immunoreactivity (ir) were colocalized in CA1 and CA2 pyramidal neurons and granule cells of the dentate gyrus, while only MR-ir was seen in the CA3 pyramidal neurons. Colocalization of MR- and GR-ir was also observed in the parvocellular region, but not in the magnocellular region of the paraventricular nucleus (PVN). Subcellular distribution of MR-ir was more cytoplasmic in comparison with that of GR-ir, while the ratio of cytoplasmic to nuclear distribution of these receptors was different among the regions. After adrenalectomy (ADX), the distribution pattern of both receptors was changed to cytoplasmic, although the degree of the change of distribution was different among each region. Replacement of corticosterone after ADX recovered the distribution pattern to that of the sham-operated animals. These results suggest that the balance of MR and GR in the cell underlies the potential regulation of corticosteroid through the hippocampus and hypothalamus.     

Quantitative autoradiography of hippocampal GABAB and GABAA receptor changes in Alzheimer's disease

GABAB and GABAA receptors were examined by quantitative [3H]GABA autoradiography in postmortem human hippocampus from 6 histopathologically verified cases of dementia of the Alzheimer type (DAT) and 6 normal controls. Significant decrements in the Bmax for both types of GABA receptors were observed in DAT hippocampus as compared to normal controls. No significant differences in Kd values were revealed. As compared to controls, DAT hippocampus exhibited fewer GABAB receptors in stratum moleculare of the dentate gyrus, stratum lacunosum-moleculare and stratum pyramidale of CA1. Significant loss of GABAA receptors in DAT hippocampus was also observed in the CA1 pyramidal cell region. These changes could not be correlated with differences in age nor in postmortem delay between the two groups. These findings may reflect the neuronal pathologies in CA1 region, in dentate gyrus, and in projections from the entorhinal cortex which are associated with the memory impairment in DAT.     

Does a unique type of CA3 pyramidal cell in primates bypass the dentate gate?

The predominant excitatory synaptic input to the hippocampus arises from entorhinal cortical axons that synapse with dentate granule cells, which in turn synapse with CA3 pyramidal cells.Thus two highly excitable brain areas--the entorhinal cortex and the CA3 field--are separated by dentate granule cells, which have been proposed to function as a gate or filter. However, unlike rats, primates have "dentate" CA3 pyramidal cells with an apical dendrite that extends into the molecular layer of the dentate gyrus, where they could receive strong, monosynaptic, excitatory synaptic input from the entorhinal cortex. To test this possibility, the dentate gyrus molecular layer was stimulated while intracellular recordings were obtained from CA3 pyramidal cells in hippocampal slices from neurologically normal macaque monkeys. Stimulus intensity of the outer molecular layer of the dentate gyrus was standardized by the threshold intensity for evoking a dentate gyrus field potential population spike. Recorded proximal CA3 pyramidal cells were labeled with biocytin, processed with diaminobenzidine for visualization, and classified according to their dendritic morphology. In response to stimulation of the dentate gyrus molecular layer, action potential thresholds were similar in proximal CA3 pyramidal cells with different dendritic morphologies. These findings do not support the hypothesis that dentate CA3 pyramidal cells receive stronger synaptic input from the entorhinal cortex than do other proximal CA3 pyramidal cells.     

Autoradiographic distribution of [H]nicotine binding in human cortex: Relative abundance in subicular complex

Distinct patterns of [³H]nicotine (3 nm) binding were apparent in various regions of adult human neo- and archicortex. Receptor binding was greatest in the subicular complex—particularly presubiculum—and entorhinal cortex, where it was prominent in the characteristic parvo- and magnocellular islands of these regions and in middle layers of entorhinal cortex. In somatosensory cortex (Brodmann areas 3, 1 and 2) and occipital (area 17) cortex binding was highest in the upper and lower layers, and relatively sparse in the sensory input, layer IV. In primary motor (area 4) and temporal (area 21) cortex, binding in the outer half of the cortical ribbon was denser than that in the inner half and a distinct band was apparent in temporal and cingulate (area 32) in the lower portion of layer III. In prefrontal association cortex the pattern of binding was less distinct although slightly higher in the lower architectonic layers. There was generally little binding in the hippocampus (areas CA1–4) and dentate gyrus with the exception of the stratum lacunosum moleculare in CA2–3 and, to a lesser extent, supra- and subgranule zones of the dentate. These patterns of reactivity, which are distinct from that of the major cortical cholinergic innervation, suggest that the nicotinic receptor, detected using nanomolar concentrations of [³H]nicotine, may primarily be associated with intracortical circuitry in the neocortex. The relatively high density in entorhinal and subicular regions may be related to the extensive phylogenetic development of these regions which has occurred in conjunction with the development of multimodal association circuitry in the human cortex.     

Annexin 7-immunoreactive microglia in the hippocampus of control and adrenalectomized rats

Annexin 7 (ANX7), also termed synexin, is a member of the annexin family of calcium-binding proteins. In the present study, we examined the distribution and cellular localization of ANX7-immunoreactivity in the rat hippocampus and its response to adrenalectomy (ADX). ANX7 was co-localized with OX42 in microglia distributed throughout the hippocampus of both control and ADX animals. ANX7-immunoreactivity was not detected in GFAP-positive astrocytes or in hippocampal neurons. At 1-week and 4-weeks following ADX, we observed a population of large, ameboid, ANX7-immunopositive microglia ("reactive microglia") which were largely confined to the granule cell layer of the dentate gyrus throughout its rostrocaudal extent. No reactive microglia were present in the hippocampus of sham-ADX or ADX + corticosterone treated animals. In 4-weeks ADX animals but not 1-week ADX, ANX7-immunostaining was significantly increased in the mossy fiber layer of CA3, due to the presence of many small, dark-staining "activated microglia". Our results show that ANX7 is abundantly expressed in the rat hippocampus by different microglial forms (e.g., ramified, activated and reactive microglia), suggesting an important role for this calcium-binding protein in microglial Ca2+-dependent processes.     

Differences in lesion-induced hippocampal plasticity between mice and rats

We studied the differences between mice and rats in lesion-induced sprouting in the hippocampus. The entorhinal cortex was unilaterally lesioned with ibotenic acid in adult, female mice and rats. Four weeks later the subsequent axonal sprouting in the dentate gyrus was analysed, by measuring the density of the synaptophysin immunohistochemical and acetylcholinesterase histochemical staining in the termination area of the entorhinal cortex axons. The data demonstrate that both mice and rats display a significantly increased density of staining for synaptophysin and acetylcholinesterase in the molecular layer of the dentate gyrus, indicative of axonal sprouting. Both species also show an upregulation in the density of staining for acetylcholinesterase in the molecular layer of the dentate gyrus. Further, rats, but not mice, show a significant upregulation of synaptophysin staining in stratum lacunosum moleculare of CA1 following the lesions. However, whereas rats show significant shrinkage of the molecular layer of the dentate gyrus, mice do not show any shrinkage of that layer following entorhinal cortex lesions. Taken together, these data indicate that whereas the process of reinnervation in the hippocampus is similar between the mouse and the rat, the hippocampal response to denervation shows clear differences between these two species.     

Localization and glucocorticoid regulation of 11beta-hydroxysteroid dehydrogenase type 1 mRNA in the male mouse forebrain

The enzyme 11beta-hydroxysteroid dehydrogenase type 1 (11beta-HSD1) converts the inactive 11-dehydrocorticosterone into the active glucocorticoid corticosterone. There is accumulating evidence indicating widespread expression of 11beta-HSD1 in the brain. However, there is little information about regulation of 11beta-HSD1 expression in this tissue. Using in situ hybridization involving use of 35S-labeled cRNA probe, we have studied the distribution of cells expressing 11beta-HSD1 mRNA in the male mouse forebrain as well as the effects of adrenalectomy (ADX) and acute administration of corticosterone (3 and 24 h) on 11beta-HSD1 mRNA levels. Cells expressing 11beta-HSD1 mRNA were mostly detected in the cerebral cortex, hippocampus, amygdala and medial preoptic area, with the highest expression in the cerebral cortex (retrosplenial granular area) and hippocampus (CA3 and granular layer of the gyrus dentatus). Seven days following ADX, 11beta-HSD mRNA levels were increased by 50% in the gyrus dentatus, by 100% in the CA3 area, and 105% in the cerebral cortex. Administration of corticosterone to ADX mice induced a significant decrease in mRNA, in both the hippocampus and cerebral cortex so that, at the 24 h time interval, the levels were similar to those observed in intact mice. These results clearly indicate that circulating corticosterone is downregulating the expression of 11beta-HSD1 mRNA in the two forebrain areas studied. This downregulation might contribute to maintain low intracellular corticosterone levels in central regions and then prevent the deleterious effects induced by high glucocorticoid levels.     

Reverberation of Excitation in Living “Hippocampal Formation–Entorhinal Cortex” Slices from Rats. Optical Recording

A vital potential-dependent dye was used to conduct optical recording of the electrical activity of the hippocampal formation in living slices of the rat brain including the hippocampal formation and the entorhinal cortex. These studies showed that single electrical stimuli applied to the entorhinal cortex, subiculum, and dentate gyrus produced responses in which waves of excitation passed across the hippocampal formation sequentially from the dentate gyrus, through CA3, to the CA1 field of the hippocampus. When GABAergic inhibition was partially blocked with picrotoxin, the first wave of excitation was immediately followed by several further waves in all zones of the hippocampal formation, with a constant shift in latency, which increased from the dentate gyrus to CA3 and CA1. Reverberation of excitation in the hippocampal formation-entorhinal cortex structure is regarded as the most probable cause for the appearance of these sequences of waves.     

Excitation Waves Returning to the Hippocampus via the Entorhinal Cortex Can Reactivate Populations of “Trained” Field CA1 Neurons during Deep Sleep

It is suggested that information on new stimuli from the neocortex is transmitted to the hippocampus, where temporal traces persist in the form of mosaics of modified synapses. During sleep, populations of neurons storing these traces are reactivated and return the information required for consolidation of a permanent memory trace to the neocortex. A possible mechanism for the reactivation of “trained” hippocampal neurons during memory consolidation consists of the reverberation of excitation in the neuron circuits linking the hippocampus and entorhinal cortex. Our studies in rats included recording of responses in hippocampal field CA1 to stimulation of Schaffer collaterals with potentiated synapses during waking and sleep. During deep sleep, discharges of field CA1 neurons were followed by waves of excitation which passed through the entorhinal cortex and reached the hippocampus and dentate gyrus via fibers of the perforant path, evoking neuron discharges in the latter. Repeated neuron discharges in field CA1 occurred on interaction of the early excitation wave returning directly via perforant path fibers and the late wave returning via Schaffer collaterals, not via the trisynaptic path via the dentate gyrus and hippocampal field CA3 but probably via field CA2.     

Simultaneous induction of long-term potentiation in the hippocampus and the amygdala by entorhinal cortex activation: mechanistic and temporal profiles

The medial temporal lobe, including the entorhinal cortex, the amygdala and the hippocampus, has an important role in learning and memory, and its circuits exhibit synaptic plasticity (long-term potentiation [LTP]). The entorhinal cortex is positioned to exert a potent influence on the amygdala and the hippocampus given its extensive monosynaptic projections to both areas. We therefore studied the effects of activation of the entorhinal cortex with simultaneous recording of LTP in the hippocampus and amygdala in the anesthetized rat. theta Burst stimulation of the lateral entorhinal cortex induced LTP simultaneously in the basal amygdaloid nucleus and in the dentate gyrus. However, the mechanisms involved in the induction of LTP in the two areas differed. The N-methyl-D-aspartate receptor antagonist 3-[(+/-)-2-carboxypiperazine-4-yl)-propyl-1-phosphonic acid delivered 1 h before LTP induction (10 mg/kg, i.p.), blocked LTP in the dentate gyrus but not in the amygdala. In addition we found that the basal amygdala as well as the dentate gyrus sustained late-phase LTP (10 h) which may participate in memory encoding and/or modulation processes. Overall, the results suggest a coordinating role for the entorhinal cortex by simultaneously modulating activity and plasticity in these structures, albeit through different mechanisms. Interactive encoding of this sort is believed to endow memories with a different, more integrative, quality than when either pathway is activated alone.     

Regional and time dependent variations of low Mg2+ induced epileptiform activity in rat temporal cortex slices

Summary In order to study spatial interactions during low magnesium induced epileptiform activity, changes in extracellular potassium concentration ([K⁺]_o) and associated slow field potentials (f.p.'s) were recorded in thin rat temporal cortex slices (400 m) containing the neocortical temporal area 3 (Te3), the entorhinal cortex (EC) and the hippocampal formation with the dentate gyrus, area CA3 and CA1 and the subiculum (Sub). The epileptiform activity was characterized by short recurrent epileptiform discharges (40 to 80 ms, 20/min) in areas CA3 and CA1 and by interictal discharges and tonic and clonic seizure like events (SLE's) (13–88s) in the EC, Te3 and Sub. While interictal discharges occurred independent of each other in the different subfields, the three areas became synchronized during the course of a SLE. The EC, Te3 and Sub all could represent the focus for generation of the SLE's. This initiation site for SLE's sometimes changed from one area to another. The characteristics of the rises in [K⁺]_o and subsequent undershoots were comparable to previous observations in in vivo preparations. Interestingly, rises in [K⁺]_o could start before actual onset of seizure like activity in secondarily recruited areas. The epileptiform activity could change its characteristics to either a state of recurrent tonic discharge episodes or to a continuous clonic discharge state reminiscent of various forms of status epilepticus. We did not observe, in any of these states, active participation by area CA3 in the epileptiform activity of the EC in spite of clear projected activity to the dentate gyrus. Even after application of picrotoxin (20 M), area CA3 did not actively participate in the SLE's generated in the entorhinal cortex. When baclofen (2 M) was added to the picrotoxin containing medium, SLE's occurred both in the entorhinal cortex and in area CA3, suggesting that inhibition of inhibitory interneurons by baclofen could overcome the filtering of projected activity from the entorhinal cortex to the hippocampus.     

Gerbil Angiotensin II AT1 receptors are highly expressed in the hippocampus and cerebral cortex during postnatal development

Increasing evidence suggests that Angiotensin II, classically known from its many effects regulating salt and water homeostasis, is also involved in brain development and cognitive functions through activation of AT₁ Angiotensin II receptors. The recently cloned gerbil AT₁ receptor is expressed in brain areas controlling hydro-mineral homeostasis, and particularly highly expressed in limbic areas such as the hippocampal formation. We quantified the gerbil AT₁ receptor messenger RNA expression and receptor binding by quantitative in situ hybridization and receptor autoradiography, respectively, in the hippocampal formation and cerebral cortex of gerbils during postnatal development. The receptor messenger RNA and binding were present from birth and showed a gradual and sustained increase through postnatal maturation in the CA1 and CA2 regions of the hippocampus and in the dentate gyrus. Conversely, in the CA3 region, no binding was detected while receptor messenger RNA peaked at 15 days after birth and disappeared in the adult. The highest receptor messenger RNA expression and binding were found in the septomedial portions of the CA1 region and at septal levels of the CA2 region.We detected the highest receptor messenger RNA expression at postnatal day one in the frontolateral pole of the cerebral hemispheres. In these areas, and in the frontoparietal and insular cortex, receptor messenger RNA dramatically decreased during postnatal life. Similarly, we found receptor messenger RNA expression in the cingulate, retrosplenial, perirhinal and infralimbic cortex with higher values during the first two weeks of development and decreased expression in the adult. However, receptor binding in the cerebral cortex, did not decrease during postnatal life.The differential profile of receptor messenger RNA expression and binding in the gerbil cortex and hippocampus during postnatal maturation suggest a role for AT₁ receptors in the development and function of the corticohippocampal system.     

听源性惊厥易感大鼠惊厥和点燃后前脑结构内的c—fos表达

为探讨听源性惊厥点燃和前脑结构的关系，用免疫细胞化学方法结合体视学分析，研究Ｗｉｓｔａｒ种系的听源性惊厥易感大鼠惊厥和点燃后，前脑结构内ｃ－ｆｏｓ表达的差异。结果显示，１．正常Ｗｉｓｔａｒ大鼠接受一次强音刺激后，海马，齿状回，杏仁核，内嗅皮质，嗅周皮质和额－顶皮质内未见Ｆｏｘ阳性神经元；２．Ｐ７７ＰＭＣ大鼠一次惊厥后，除海马，齿状回外，上述被检各区内可见广泛的Ｆｏｓ阳性神经元，其分布具有区域差异．     

Adrenalectomy and corticosterone replacement differentially alter CA3 dendritic morphology and new cell survival in the adult rat hippocampus

Plastic changes in the adult mammal hippocampus can be altered by many factors and perhaps the most well-documented is stress. Stress and elevated corticosterone levels have been shown to decrease hippocampal neurogenesis and decrease the complexity of CA3 pyramidal neurons. However, the extent of these changes in relation to low and moderately elevated levels of corticosterone has yet to be fully investigated. Therefore, the aim of the present study was to determine how low to moderately elevated circulating corticosterone levels affect dendritic morphology of CA3 pyramidal cells and hippocampal neurogenesis in adult male rats. To do this, three groups of adult male Wistar rats were used: (1) Sham-operated, (2) Adrenalectomized (ADX), and (3) ADX + corticosterone replacement. Primary results show that adrenalectomy, but not moderately elevated levels of corticosterone replacement, resulted in significant atrophy of CA3 pyramidal neurons. Interestingly, moderate corticosterone replacement resulted in significantly more surviving new cells in the dentate gyrus when compared to sham controls. This work shows that circulating levels of corticosterone differentially affect plasticity in the CA3 region and the dentate gyrus.     

Short-term glucocorticoid manipulations affect neuronal morphology and survival in the adult dentate gyrus

In order to determine whether short-term glucocorticoid manipulations influence the morphology and survival of neurons in the adult mammalian hippocampal formation, we performed quantitative analyses of Golgi-impregnated and Nissl-stained tissue from the brains of sham operated male rats, adrenalectomized male rats and adrenalectomized male rats which received corticosterone replacement. Three days after adrenalectomy, massive cell death, as detected by a dramatic increase in number of pyknotic cells, was observed in the granule cell layer of the dentate gyrus. By seven days following adrenalectomy, the numbers of pyknotic cells were even greater. Moreover, significant decreases in cross-sectional cell body area and numbers of dendritic branch points of Golgi-impregnated dentate gyrus granule cells were detected at seven days after adrenalectomy. Replacement of corticosterone to adrenalectomized rats prevented the appearance of large numbers of pyknotic cells as well as the decrease in granule cell cross-sectional cell body area and the numbers of dendritic branch points. In contrast, no obvious signs of degeneration were detected in the pyramidal cell layers of the CA1 and CA3 regions of the hippocampus at either three or seven days following adrenalectomy. In addition, no significant changes in morphological characteristics were observed in CA1 or CA3 pyramidal cells with adrenalectomy. These results show that dentate gyrus granule cells require glucocorticoids for their survival and for the maintenance of normal morphology and suggest that granule cell morphology and/or survival may undergo constant fluctuation in response to diurnal rhythms or stress-induced changes in glucocorticoid levels.