Modulation of sensory input to the spinal cord by presynaptic ionotropic glutamate receptors.

Sensory input from peripheral nerves to the dorsal horn of the spinal cord is mediated by a variety of agents released by the central terminals of dorsal root ganglion (DRG) neurons. These include, but are not limited to, amino acids, especially glutamate, peptides and purines. The unraveling of the mechanisms of synaptic transmission by central terminals of DRG neurons has to take into account various ways in which the message from the periphery can be modulated at the level of the first central synapse. These include postsynaptic and presynaptic mechanisms. Homomeric and heteromeric complexes of receptor subunits for the different transmitters released by DRG neurons and interneurons, clustered at the postsynaptic site of central synapses, can be expressed in different combinations and their rate of insertion into the postsynaptic membrane is activity-regulated. Inhibitory mechanisms are an important part of central modulation, especially via presynaptic inhibition, currently believed to involve GABA released by inhibitory intrinsic neurons. Recent work has established the occurrence of another way by which sensory input can be modulated, i.e. the expression of presynaptic ionotropic and metabotropic receptors in central terminals of DRG neurons. Microscopic evidence for the expression, in these terminals, of various subunits of ionotropic glutamate receptors documents the selective expression of glutamate receptors in functionally different DRG afferents. Electrophysiological and pharmacological data suggest that activation of presynaptic ionotropic glutamate receptors in central terminals of DRG neurons may result in inhibition of release of glutamate by the same terminals. Glutamate activating presynaptic receptors may spill over from the same or adjacent synapses, or may be released by processes of astroglial cells surrounding synaptic terminals. The wide expression of presynaptic ionotropic glutamate receptors, especially in superficial laminae of the dorsal horn, where Adelta- and C fibers terminate, provides an additional or alternative mechanism, besides GABA-mediated presynaptic inhibition, for the modulation of glutamate release by these fibers. Since, however, presynaptic ionotropic glutamate receptors are also expressed in terminals of GABAergic intrinsic interneurons, a decrease of GABA release resulting from activation of these receptors in the same laminae, may also play a role in central sensitization and hyperalgesia.

Kainate receptors in primary afferents to the rat gracile nucleus.

In a previous work, we demonstrated that under weak paraformaldehyde fixation, kainate receptors (KR) (GluR5/6/7) are expressed in primary afferent terminals in superficial dorsal horn. We extended our study to primary afferents to the gracile nucleus; immunostaining for GluR5/6/7 in weakly fixed sections was in puncta of variable size. In double-stained sections, the majority of immunostained puncta colocalized with synaptophysin. Because of their large size and relations with smaller puncta, single-stained for synaptophysin, these terminals were presumed to be of dorsal column primary afferents. This was confirmed by anterograde labeling with cholera toxin B and with electron microscopy, which showed that GluR5/6/7 was present in terminals with morphology of primary afferents. These observations demonstrate that expression of presynaptic KR is a general feature of primary afferents with different functional properties.

Palladin is expressed preferentially in excitatory terminals in the rat central nervous system.

Palladin is a recently described intracellular protein associated with the actin cytoskeleton and cell adhesion in fibroblasts. In Western and Northern blot analyses, palladin expression is ubiquitous in embryonic mice, but it is down-regulated dramatically in most adult tissues. Significant amounts of palladin persist in the brain of adult rodents, as assessed by Western blot analysis. With this work, we extend preliminary observations and determine the overall distribution and subcellular location of palladin throughout the rat brain. In sagittal and coronal sections of the central nervous system, immunostain for palladin is present throughout the brain and spinal cord, but not uniformly. The densest regions of immunostain include the olfactory bulb, cerebral and cerebellar cortex, hippocampus, amygdala, superior colliculus, and superficial laminae of the spinal dorsal horn. Because immunostain characteristically is punctate, we performed double staining for palladin and the presynaptic marker synaptophysin. Confocal microscopy showed that palladin-immunopositive puncta are also immunopositive for synaptophysin; the proportion of synaptophysin-immunopositive puncta that also stained for palladin ranged from 100% of mossy fiber terminals in field CA3 of the hippocampus and in the cerebellar cortex to 60--70% of terminals in the cerebral cortex, striatum, and spinal dorsal horn. The presence of palladin in synaptic terminals was confirmed by electron microscopy. Because immunostained terminals commonly establish asymmetric synapses, the selectivity of palladin expression in synaptic terminals was tested by double staining for palladin and gamma-aminobutyric acid. The modest level of colocalization in this material at both the light microscopic and electron microscopic levels suggests a selectivity of palladin for terminals that release excitatory neurotransmitters. As concomitant work in cell cultures has shown that palladin participates in axonal development and migration, the present results suggest that palladin persists at excitatory synapses of the adult nervous system.

Vanilloid receptor VR1 is both presynaptic and postsynaptic in the superficial laminae of the rat dorsal horn.

Terminals in the rat spinal cord that express the vanilloid receptor VR1 are from small and medium dorsal root ganglion (DRG) neurons and appear prominent in lamina I and inner lamina II. Because primary afferents from these neurons can be myelinated or unmyelinated and their terminals in these laminae can be of various morphological and functional types, we undertook this study to identify the type(s) of VR1-positive afferent fibers and terminals. In the DRG, many small and medium-sized neurons are immunopositive. Under electron microscopy, dorsal root afferents that are immunopositive for VR1 are predominantly unmyelinated. Large numbers of VR1-positive terminals in lamina I are of the nonglomerular type and may contain dense core vesicles. VR1 immunoreactivity in terminals in lamina I is in good agreement with data on noxious, heat-sensitive neurons in the dorsal horn. Two types of glomerular afferent terminals in lamina II also are immunopositive for VR1. In both laminae, most VR1-positive terminals are distinct from substance P-positive terminals. However, the immunoreactivity in lamina II also is prominent in dendrites that are contacted by primary afferent endings. Because we also observed patchy immunostaining in cell bodies in lamina II, this unexpected result may reflect synthesis of VR1 by neurons in this lamina. However, because dorsal rhizotomy abolishes VR1 staining in both laminae I and II, it is suggested that the expression and intracellular dynamics of VR1 in lamina II neurons are controlled by presynaptic input.

Presynaptic kainate receptors in primary afferents to the superficial laminae of the rat spinal cord.

Subunits of glutamate receptors participate in the regulation of sensory transmission at primary afferent synapses in the superficial laminae of dorsal horn (DH). We report here on the distribution of kainate receptors (GluR5/6/7) in these laminae by using light microscope (LM) and electron microscope (EM) immunocytochemistry. Standard (4%) paraformaldehyde fixation resulted in immunostaining for GluR5/6/7 in perikarya and fine processes in lamina II, especially its inner part (IIi). Preembedding EM revealed immunostaining of dendrites, perikarya, and occasional terminals, presumed to be from primary afferent fibers, at the center of glomerular arrangements. In rats perfused with 0.5% paraformaldehyde, LM showed a more punctate staining, mainly in the ventral part of lamina IIi and lamina III, than in material fixed with 4% paraformaldehyde. Approximately two-thirds of GluR5/6/7 puncta were also immunostained with synaptophysin, suggesting that in material fixed with 0.5% paraformaldehyde, a large fraction of these are synaptic terminals. Double immunostained puncta disappear 4 days after dorsal rhizotomy, suggesting that most of GluR5/6/7-immunopositive terminals are from primary afferent fibers. EM material fixed with 0.5% paraformaldehyde confirmed the expression of GluR5/6/7 in numerous synaptic endings with morphology of primary afferents. To determine the type of primary afferent terminals that express GluR5/6/7, two neuroanatomic tracers were injected in the sciatic nerves. The lectin from Bandeiraea simplicifolia (IB4) is selectively taken up by unmyelinated primary afferent fibers that terminate in the outer part of lamina II (IIo) and dorsal part of lamina IIi, whereas the B subunit of the cholera toxin (CTB) is selectively taken up by a broader class of primary afferents which, in superficial DH, terminate mainly in laminae I, ventral part of IIi, and III. Approximately 20% of GluR5/6/7-immunoreactive puncta colocalized with IB4, whereas approximately 40% of GluR5/6/7-immunoreactive puncta colocalized with CTB. The present study shows that (1) GluR5/6/7 does not have a clear and consistent spatial relation with postsynaptic sites, (2) a large number of primary afferents express GluR5/6/7, and (3) these are not limited to one functional class. Thus, modulation by glutamate of primary afferent terminals by means of kainate receptors in the superficial laminae of DH may predominantly involve presynaptic mechanisms.

Metabotropic glutamate receptors in superficial laminae of the rat dorsal horn.

Light and electron microscopic immunocytochemistry were employed here to show the distribution of metabotropic glutamate receptors (mGluRs) mGluR2/3 and mGluR5 in laminae I and II of the dorsal horn, to identify their pre- and postsynaptic location, and to test colocalization with gamma-aminobutyric acid (GABA). mGluR2/3 was mainly in the inner part of lamina II; mGluR5 was mainly in laminae I and II. Electron microscopy showed that both mGluR2/3 and mGluR5 were in perikarya, dendrites, and vesicle-containing profiles. Two main morphological types of primary afferent terminals can be distinguished in the superficial laminae: C1, likely to be endings of unmyelinated fibers, and C2, of small myelinated fibers. Quantitative data show that only a small fraction of C2s stained for either receptor; more common were immunopositive dendrites postsynaptic to these terminals, and most common were appositions between C2s and mGluR5 immunopositive dendrites. Vesicle-containing profiles were characteristically apposed to primary afferent terminals, mainly C2s. Immunopositivity for mGluRs, especially mGluR2/3, was present in vesicle-containing profiles apposed to C2, none to C1, and about half of the profiles immunostained for either receptor were also stained for GABA. The presence of presynaptic and postsynaptic mGluRs in both inhibitory and excitatory interneurons may contribute to complex processing of fast and slow responses to peripheral input in superficial laminae. As selective agonists of mGluRs may modulate GABA release, the present demonstration of mGluRs in GABAergic terminals of presumed interneurons suggests that facilitatory effects may involve a mechanism of disinhibition.

Expression of three forms of nitric oxide synthase in peripheral nerve regeneration.

Nitric oxide (NO) is a short-lived molecule with messenger and cytotoxic functions in nervous, cardiovascular, and immune systems. Nitric oxide synthase (NOS), the enzyme responsible for NO synthesis, exists in three different forms: the neuronal (nNOS), present in discrete neuronal populations; the endothelial (eNOS), present in vascular endotheliun, and the inducible isoform (iNOS), expressed in various cell types when activated, including macrophages and glial cells. In this study, we have investigated the possible involvement of NO in Wallerian degeneration and the subsequent regeneration occurring after sciatic nerve ligature, using histochemistry and immunocytochemistry for the three NOS isoforms, at different postinjury periods. Two days after lesion, the three NOS isoforms are overexpressed, reaching their greatest expression during the second week. nNOS is upregulated in dorsal root ganglion neurons, centrifugally transported and accumulated in growing axons. eNOS is overexpressed in vasa nervorum of the distal stump and around ligature, and iNOS is induced in recruited macrophages. These findings indicate that different cellular sources contribute to maintain high levels of NO at the lesion site. The parallelism between NOS inductions and well-known repair phenomena suggests that NO, acting in different ways, may exert a beneficial effect on nerve regeneration.

Nitric oxide synthase and growth-associated protein are coexpressed in primary sensory neurons after peripheral injury.

A possible role for nitric oxide in growth and regeneration of dorsal root ganglion (DRG) afferents has been explored in lesion experiments by comparing immunocytochemistry for nitric oxide synthase (NOS) with that for the growth-associated phosphoprotein 43 (GAP-43). Sciatic nerve ligature induced a progressive increase in the number of small DRG cell profiles immunopositive for NOS between 2 days and 4 weeks of survival. In the proximal stump of the ligature, NOS-immunopositive fibers began to appear 2 days after injury and their growth cones were especially evident after 7 days. NOS-immunopositive fibers appeared past (i.e., distal to) the ligature at 14 days of survival and extended for at least 6 mm in either direction 4 weeks after the lesion. Dorsal root ligature alone at L4-L5 did not result in expression of NOS in DRG neurons or in the appearance of NOS-immunopositive fibers. In rats with dorsal root ligature and nerve ligature, the results were similar to those with nerve ligature only. DRG cell profiles immunopositive for GAP-43 kept increasing from 2 days to 4 weeks after sciatic nerve ligature and included small neurons initially and large neurons subsequently. Numerous axons became GAP-43 immunopositive on both sides of the ligature from 2 days after injury. In double-labeled material, about 80% of DRG cell profiles immunopositive for NOS were also immunopositive for GAP-43. The two antigens co-occurred in peripheral nerve axons proximal to the ligature starting at about 7 days and distal to it at about 2 weeks after ligature. Thus, in response to nerve lesion, nitric oxide may not only provide an injury signal to the central nervous system but may also contribute to the growth and regeneration of injured axons.

AMPA receptors at primary afferent synapses in substantia gelatinosa after sciatic nerve section.

Increased excitability of superficial laminae of the spinal cord may contribute to the pathological pain consequent to peripheral nerve injury. Among several mechanisms that may be responsible for this occurrence is upregulation of receptors for glutamate in the spinal cord. To explore this possibility, we investigated changes in AMPA receptors in substantia gelatinosa of rats after section of the sciatic nerve. Immunofluorescence was performed on sections from the fourth lumbar segment. Quantitative analysis of digitally captured images suggested that staining for an antibody to a sequence shared by GluR2 and GluR3 (GluR2/3) was increased on the side ipsilateral to the lesion. To determine whether antigen accumulation was at synaptic sites and to probe whether it was selective for primary afferent terminals, we performed electron microscopy on immunogold-labelled material. Gold particles coding for GluR2/3 subunits were counted from synaptic active zones of glomerular terminals in substantia gelatinosa that originate from small calibre afferent fibres, and from active zones of terminals of probable intrinsic origin. Counts were significantly increased on the side ipsilateral to the lesion only at synapses of primary afferent terminals. These results document selective upregulation of receptor protein at the synapse. This upregulation may contribute to the increased sensitivity of dorsal horn neurons following peripheral nerve injury.

NMDAR1 and primary afferent terminals in the superficial spinal cord.

We previously demonstrated a predominance of contacts by one type of glomerular ending (presumably from unmyelinated fibers) upon postsynaptic sites expressing the GluR1 subunit of the AMPA receptor, and of contacts by another type of glomerular ending (presumably from small myelinated fibers) upon postsynaptic sites expressing GLuR2. We here investigate whether any one of three types of primary afferent terminals, two glomerular and one non-glomerular, have direct contacts with postsynaptic sites containing the NMDAR1 subunit. Counts of gold particles revealed that contacts by primary afferents with NMDAR1-positive sites were less frequent than in material processed for AMPA receptor subunits, but that all three types of terminals contact NMDAR1-immunopositive postsynaptic sites in about equal proportions.

Heterogeneity of AMPA receptors in the dorsal column nuclei of the rat.

We have combined immunocytochemistry with retrograde tracing to demonstrate that projecting neurons in the gracile and cuneate nuclei express predominantly the GluR3 subunit of the AMPA receptor while interneurons in these nuclei express predominantly the GluR1 subunit. Interneurons expressing the GluR2 subunit are also present. It is speculated that the two classes of interneurons may release different inhibitory transmitters.

Arginine and NADPH diaphorase in the rat ventroposterior thalamic nucleus.

Simultaneous immunocytochemical staining for arginine (Arg) and histochemical staining for reduced nicotinamide adenine dinucleotide phosphate diaphorase (NADPHd, a marker for nitric oxide synthase) reveals that neuropil in the ventroposterior nucleus of the thalamus is enriched with both Arg-positive glial profiles and NADPHd-positive fibers. NADPHd-positive fibers are often apposed to Arg-positive astrocytes and oligodendrocytes. NADPHd-positive endothelial cells are often adjacent to Arg-positive astrocytes. The results suggest that Arg may be stored in supporting cells, whence it could be supplied to nearby nerve fibers or endothelial cells as substrate for nitric oxide synthase.

Glutamate receptors in spinal motoneurons after sciatic nerve transection.

Severing the axon of a neuron triggers profound changes in its soma, beginning within a few days and becoming maximal within a few weeks. Unravelling these changes bears directly on our understanding of degeneration and regeneration after injury. Classically described chromatolysis arises from reorganization of rough endoplasmic reticulum, associated with biosynthetic changes in response to injury. Since motoneurons, in contrast with other central neurons, are able to regenerate their axons, their response to axotomy is of special interest. For successful regeneration, a neuron must shift its cellular machinery from "operational" (e.g., integration of synaptic currents, conduction of action potentials, release of transmitter) to "regenerative" (e.g., repair of membrane and axoplasm, remyelination, growth cone guidance). Motoneurons become unresponsive to synaptic input after axotomy, and the conduction velocity of the proximal stump is reduced. The loss of synaptic contacts on to axotomized neurons has been suggested to underlie this lost responsiveness. Here, we demonstrate rapid, selective and dramatic changes in immunostaining for ionotropic glutamate receptors in axotomized motoneurons and in supporting cells, suggesting that altered expression of glutamate receptors underlies the changed reflex responsivity.

AMPA receptor subunits underlying terminals of fine-caliber primary afferent fibers.

Postembedding immunogold electron microscopy was used to determine the relation of primary afferent terminals in superficial laminae of the spinal dorsal horn with AMPA receptor subunits. Immunogold particles coding for GluR1 and GluR2/3 were concentrated at synaptic sites, between 30 nm outside and 40 nm inside the postsynaptic membrane. Immunopositive synapses displayed round vesicles and asymmetric specializations, characteristic of terminals releasing excitatory neurotransmitters; symmetric synapses, characteristic of terminals releasing inhibitory amino acids, were immunonegative. In superficial laminae, large terminals of two main types at the center of a synaptic glomerulus originate from primary afferents: C1 terminals are mainly endings of unmyelinated afferent fibers; C2 terminals are mainly endings of thinly myelinated afferent fibers. Terminals of both types were presynaptic to AMPA subunits, but in different proportions: C1 terminals were related more to GluR1 than to GluR2/3, whereas the reverse was true for C2 terminals. These results suggest that functional properties of peripheral afferents to the spinal cord may be specified by the density and combination of receptor subunits in the postsynaptic membrane, and raise the possibility that calcium-permeable AMPA channels may play a special role in the mediation of sensory input by unmyelinated fibers.

EM colocalization of AMPA and NMDA receptor subunits at synapses in rat cerebral cortex.

Electrophysiology and light microscopy suggest that a single excitatory synapse may use both amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) and N-methyl-D-aspartate (NMDA) receptors. Using immunogold electron microscopy, we here provide direct evidence for colocalization at individual synapses in sensorimotor cortex of adult rats. Colocalization was most commonly observed on dendritic spines; subunits of the two classes of receptors seemed to be independently distributed within the synaptic active zone.

Colocalization of GABA and glycine in the rat dorsal column nuclei.

About half the neurons in the rat dorsal column nuclei were immunopositive for glycine or for GABA; these were smaller than immunonegative neurons. In double-stained material, 29% of stained neurons were immunopositive for GABA only, and 42% for both antigens. The results resemble those reported for spinal cord laminae that receive fast-conducting primary afferents, and suggest that glycine is an inhibitory neurotransmitter in the dorsal column nuclei.

Colocalization of neuronal nitric oxide synthase with GABA in rat cuneate nucleus.

Pre- and post-embedding immunocytochemistry were employed in this electron microscopic investigation of cuneate neurons that are enriched in GABA and in nitric oxide synthase, the enzyme responsible for the synthesis of nitric oxide. GABAergic neurons are local circuit interneurons; 10-20% of them also contain nitric oxide synthase. These are among the smallest GABA-positive perikarya. We describe a network of processes in the rat cuneate nucleus that are immunopositive for nitric oxide synthase. Axon terminals positive for nitric oxide synthase are small and make synapses mainly onto dendrites; they make only occasional axo-axonic contacts. Double-labelling immunocytochemistry verified that the large majority of terminals positive for nitric oxide synthase also contained GABA. However, most GABA-positive profiles were negative for nitric oxide synthase and GABA-positive terminals that are negative for nitric oxide synthase frequently made axo-axonic contacts. These results suggest that nitric oxide synthase is within a specialized subpopulation of interneurons in the cuneate nucleus.

An osmium-free method of epon embedment that preserves both ultrastructure and antigenicity for post-embedding immunocytochemistry.

Immunocytochemistry for amino acids with post-embedding gold is compatible with glutaraldehyde fixation, osmication, and embedding in epoxy-based plastics, but immunogold detection of larger molecules in the central nervous system commonly requires special procedures, e.g. minimizing exposure to glutaraldehyde, eliminating osmium, cryosectioning, and/or embedding in acrylic plastics. These make samples more difficult to prepare and view and may compromise structural preservation. We report a new technique, fixing with high levels of glutaraldehyde, replacing osmium with tannic acid followed by other heavy metals and p-phenylenediamine, and embedding in Epon. This method optimizes antigenicity while retaining the structural preservation and convenient handling of standard embedding techniques. Compared to standard Epon embedment, labeling for neuropeptides in brain and spinal cord is improved. Moreover, the present method yields excellent labeling of glutamate receptors (difficult to identify with traditional post-embedding techniques) and enables simultaneous visualization of associated neurotransmitters.

Central release of tracer after noxious stimulation of the skin suggests non-synaptic signaling by unmyelinated fibers.

Injury to a peripheral nerve causes central changes of various nature and complexity reflecting activation of multiple signaling mechanisms. In a previous study we reported that nerve lesion triggers central release of a tracer, wheatgerm-agglutinin conjugated to horse-radish peroxidase, by unmyelinated fibers in the spinal cord. The released tracer occupies the space between nerve terminals and dendrites without extending into the synaptic cleft. We interpreted this to suggest release of unidentified endogenous factor(s) at nonsynaptic sites, which may contribute to the signaling of peripheral injury to the central nervous system. For such signaling to occur, a message must first be communicated along the axon. This message may depend on axonal transport and/or altered electrical activity. In pilot experiments we observed that application of tetrodotoxin (to block impulse conduction) to the intact nerve did not result in tracer release. We hypothesized that the message might be the sustained discharge of C fibers that occurs after injury. We show here that selective activation of C fibers (by applying mustard oil to the hindlimb of anesthetized rats) causes central release of tracer previously transported from the sciatic nerve to superficial laminae of the dorsal horn.