Zinc is packaged in, and released from, a subset of glutamatergic synapses in the mammalian telencephalon where it has been shown to act as a potent neuromodulator. In order to establish the functional role for zincergic neurons in visual cortical function and plasticity we have compared the topographic distribution of zincergic terminals in the primary visual cortex (V1) of normal adult vervet monkeys (Cercopithicus aethiops) to that in monkeys monocularly deprived of visual input for short (24 h) or long (3 months) survival times. In normal animals, staining levels for zinc were highest in layers 1–3, 4b, 5 and 6 and lowest in layers 4a and 4c. The laminar and tangential patterns of zinc staining were complementary to staining patterns demonstrated using cytochrome oxidase (CO) histochemistry. Following 3 months of monocular deprivation by enucleation, levels of zinc staining in layers 3, 4cα and 6a were heterogeneously reduced, clearly revealing the ocular dominance pattern in V1. When compared with the pattern of CO staining, levels of both CO and zinc were reduced in cortical territory innervated by the enucleated eye. Zinc histochemistry also revealed the ocular dominance pattern after only 24 h of monocular impulse blockade induced by enucleation or intravitreal tetrodotoxin infusion. However, by either means of deprivation for 24 h, levels of zinc were increased in deprived-eye stripes relative to nondeprived-eye stripes. These results indicate that zincergic terminals demarcate distinct compartments in the primate visual cortex. Furthermore, levels of synaptic zinc are rapidly and dynamically regulated, suggesting that zinc and/or zincergic neurons participate in mediating activity-dependent changes in the organization of the adult neocortex.
The circuitry provided by excitatory neurons plays a central role in the transfer and processing of synaptic information in the mammalian cerebral cortex. In the visual cortex, the major excitatory inputs originate predominantly from thalamic neurons in the lateral geniculate nucleus (Montero, 1990; Kaneko and Mizuno, 1992) or from local and distant intra-cortically projecting neurons (Tsumoto et al., 1986; Tsumoto, 1990). Anatomical and physiological studies have revealed that visual cortical afferents terminate in rather discrete laminar and columnar compartments (Carder and Hendry, 1994) that may correspond to functional compartments (Livingstone and Hubel, 1987, 1988). Thus, the ability to characterize compartmentally distinct populations of excitatory inputs would provide important information regarding the functional connectivity of visual cortex.
Subpopulations of glutamatergic neurons in the mammalian CNS sequester zinc into synaptic vesicles of their terminal boutons (Martinez-Guijarro et al., 1991; Beaulieu et al., 1992; Sensi et al., 1997). These zinc-containing axon terminals can be anatomically distinguished using a simple histochemical staining method (Danscher, 1982). Using this method, the highest levels of staining in the adult mammalian brain are found in the telencephalon, particularly in the cerebral cortex and hippocampus (Haug, 1973; Frederickson, 1989; Frederickson and Danscher, 1990). The pattern of zinc staining in the cerebral cortex is region- and lamina-specific, with the highest levels distributed in supragranular and infragranular layers and lowest levels in layer 4 (Garrett et al., 1991; Dyck et al., 1993a; Garrett et al., 1994). This laminar differentiation is most distinct in primary sensory areas, where low levels of zinc staining in thalamocortical recipient layers define their boundaries with adjacent cortical areas (Dyck et al., 1993a). Within layer 4 of the cat visual cortex and in layers 2/3 of primate visual cortex, zinc-containing terminals are enriched in, and distinguish, novel columnar compartments (Dyck et al., 1993b), whose functional significance has not yet been established.
Increased attention has recently been focused on determining the functional significance for synaptic zinc in the CNS. There is substantive evidence showing that zinc is actively released from zinc-containing terminals in a calcium-dependent manner (Assaf and Chung, 1984; Howell et al., 1984; Budde et al., 1997). Zinc acts as a modulator of both ligand-gated (Smart et al., 1994), and voltage-gated (Harrison and Gibbons, 1994) ion channels in the CNS. In particular, zinc has been shown to be a potent modulator of excitatory neurotransmission by acting at zinc-specific binding sites on the N-methyl-d-aspartate (NMDA) subtype of glutamate receptors (Westbrook and Mayer, 1987; Christine and Choi, 1990; Legendre and Westbrook, 1990; Hollmann et al., 1993). As NMDAergic mechanisms have been implicated in forms of synaptic plasticity such as those involved in long-term potentiation or long-term depression (Kirkwood et al., 1996), it is possible that zincergic neurons participate in activity-dependent mechanisms of cortical plasticity in the visual cortex by modulating glutamatergic neurotransmission.
Here we sought to describe activity-dependent modifications in the levels of zinc staining in primate visual cortex induced by monocular visual deprivation. We show that levels of synaptic zinc undergo rapid, dynamic modifications in eye-specific cortical compartments suggesting that zincergic neurons might be key mediators of synaptic plasticity in the adult cerebral cortex.
Materials and Methods
We have used adult male vervet monkeys (Cercopithicus aethiops) that were obtained from a feral colony in St Kitts, West Indies for this study. A total of 10 animals were used. These included three normal animals, two animals that received unilateral visual deprivation for 24 h by way of intravitreal tetrodotoxin injection (TTX; 1 μg/ml; 20 μl), two animal that were monocularly enucleated for 24 h, and three animals that were monocularly enucleated, followed by 3 months survival.
At the conclusion of the exposure period, all animals were sedated with ketamine (10 mg/kg, i.m.; Sigma, Oakville, Canada) and removed from their cages. While under ketamine sedation, they were given an intravenous injection of sodium selenite (10 mg/kg, i.v.; Sigma) followed by 500 units of heparin (Hepalean; 1000 USP units/ml; Sigma). After a period of 20 min, the monkeys were killed with an overdose of sodium pentobarbital.
The chest was then opened with a bilateral incision through the ribcage and the heart was exposed. A small incision was made on the dorsal tip of the right atrium and blood was drained for a few seconds to relieve pressure. The animals were then perfused with Sorenson’s buffer through the left ventricle, until completely exsanguinated. The descending aorta was clamped shortly after onset of perfusion
The brain was removed, blocked along the midline, and followed by coronal block of both hemispheres at the lunate sulcus. The operculum from one hemisphere was dissected and manually flattened between glass slides to allow tangential sectioning. The operculum from the other hemisphere was left intact to allow coronal sectioning. All tissue blocks were rapidly frozen by immersion in 2-methylbutane that was maintained at −50°C in a dry-ice bath and then stored at −70°C. Prior to sectioning, pin holes were placed in the flattened hemisphere to provide fiducial landmarks to facilitate section alignment. Serial sections were cut on a cryostat (−20°C) at 25 mm thickness and thaw-mounted on gelatin-coated glass slides.
The procedures we used to localize synaptic zinc in the mammalian visual cortex have been previously detailed (Dyck et al., 1993b) and are based on a variation of the Timm stain (Danscher, 1982). Briefly, the slide-mounted sections were thawed and allowed to dry at room temperature, fixed in a descending series of ethanol washes (95%, 15 min; 70%, 2 min; 50%, 2 min), hydrated, and then dipped in 0.5% gelatin to prevent autocatalytic staining.
Zinc-selenide precipitate was visualized on slide-mounted sections in freshly prepared developer. Each batch of 18 slides was immersed in 250 ml of developer (50% Gum arabic, 125 ml; 2.0 M sodium citrate buffer, 25 ml; 0.5 M hydroquinone, 30 ml; 37 mM silver lactate, 30 ml; deionized H2O, 40 ml) in complete darkness for 1.5–2 h at 26°C. The slides were washed for 10 min in running tap water, rinsed in two changes of distilled water, and stabilized in a 5% sodium thiosulfate solution for 12 min. Slides were post-fixed in 70% ethanol for at least 30 min, dehydrated to 100% ethanol, cleared in xylene and coverslipped using Permount.
Cytochrome Oxidase Histochemistry
The topographic distribution of zinc was compared, in adjacent sections, to that of cytochrome oxidase (CO) which was visualized histochemically using a modification of the method described by Silverman and Tootell (Silverman and Tootell, 1987). Prior to staining, the sections were fixed for 5 min in 4% paraformaldehyde in 50 mM phosphate buffer (PB, pH 7.4). The slides were rinsed twice for 5 min in PB and then incubated at 37°C, for 45–60 min, in a solution consisting of 50 mg nickel ammonium sulfate, 250 μl of 1 M imidazole, 1 g sucrose, 25 mg 3, 3′-diaminobenzidine tetrahydrochloride (Sigma), 15 mg cytochrome C (Type III, Sigma) and 10 mg catalase (Sigma) in 100 ml PB. After incubation, the slides were rinsed in buffer, dehydrated in an ascending series of alcohols, cleared in xylene and coverslipped using Permount. Our modification, which reduces the osmolarity of the buffer (100 mM to 50 mM) and replaces cobalt with nickel, required only one incubation step and resulted in a much more sensitive and reliable method for staining CO in unfixed, fresh-frozen tissues.
Cytochrome oxidase histochemistry indicates eye-specific patterns of staining with longer periods of monocular activity blockade but not reliably with a survival period of 24 h or less, therefore, the topographic distribution of zinc was also compared to that of the transcription factor Zif268 using immunocytochemical methods. A down-regulation of Zif268 staining in deprived-eye stripes within V1 can be resolved within 5 h of the onset of monocular visual deprivation (Chaudhuri and Cynader, 1993). The reliability with which Zif268 immunohistochemistry demarcates deprived-eye columns with very short post-enucleation survival periods (3–5.5 h) has been questioned in a recent paper (Horton et al., 2000). The authors show that labeling was sometimes greater in the intact eye’s columns and sometimes in the enucleated eye’s columns. However, we did not see any irregularities in the Zif268 immunostaining in any of our short survival period animals. The consistency of our results is likely due to two factors, the first is that our short survival periods were longer (24 h) and, secondly, the results presented by Horton et al. (Horton et al., 2000) were derived only from animals monocularly deprived by enucleation whereas we deprived input using TTX and enucleation. It is possible that the activity and transduction mechanisms of ascending visual pathways is variable when deprivation is induced by damage rather than by action potential blockade.
Slide-mounted sections were incubated, at 4°C for 48 h, in PB containing a rabbit polyclonal antibody recognizing Zif268 (1:10 000; provided by R. Bravo, Bristol-Myers Squibb Pharmaceutical, Princeton, NJ). The sections were washed three times 10 min in PB following this, and each subsequent incubation step. The sections were incubated in PB containing biotinylated anti-rabbit IgG raised in goat (1:1000; Vector, Burlingame, CA) for 3 h at room temperature. After washing, the sections were transferred to avidin-HRP (1:1000; Vectastain Elite, Vector) where they were incubated for 1 h. Peroxidase labeling was visualized by incubation in 50 mM Tris-HCl buffer (pH 7.6) containing 1% nickel ammonium sulfate, 50 mM imidazole, 0.01% 3,3′-diaminobenzidine tetrahydrochloride and 0.004% H2O2. Some sections were processed in parallel but omitted either the primary antibody or secondary antibody incubation steps. Under either of these conditions, no specific immuno-reactivity was observed.
Normal Distribution of Zinc
At low magnification the distribution of zinc in the visual cortex of normal adult vervet monkeys (Fig. 1A) appeared to be similar to that previously described for mammalian primary sensory cortices [adult cats (Dyck et al., 1993a), wallabies (Garrett and Geneser, 1993), rodents (Haug, 1973; Zilles et al., 1990; Garrett et al., 1991; Brown et al., 2003)]. Middle cortical layers are distinguished from superficial and deep layers by staining less intensely for zinc. In the monkey, this laminar differentiation was most distinct in primary visual cortex (V1), where low levels of staining in a middle band clearly demarcated its boundaries with adjacent visual cortical area, V2 (Fig. 1A, arrow demarcates V1/V2 boundary).
Lamina-specific Staining for Zinc
The laminar specificity of zinc staining in V1 was determined by comparison to adjacent sections stained for CO [Fig. 1B (Horton and Hubel, 1981)] and Nissl-substance [not shown (Lund, 1973)]. In frontal sections, most cortical layers could be clearly differentiated, based on variations in the density and pattern of zinc staining (Fig. 1A,C). Laminar boundaries were also conspicuous when compared to sections stained for CO (Fig. 1B). With respect to zinc and CO, the relative distributions of these two markers appeared to be surprisingly complementary — cortical layers with highest levels of staining for CO exhibited the lowest levels for zinc staining and vice versa. As is apparent at higher magnification in frontal sections (Fig. 1C), the major laminar differences in zinc staining in V1 are attributed to variations in staining levels and patterns in layer 4. Layer 4a is characterized by alternating columns of lightly and densely stained columnar compartments while layer 4b is homogeneously stained, at high levels. Layer 4c is the lightest staining of all cortical layers with α and β sub-laminae further differentiable by variations in zinc staining intensity and patterning. Specifically, layer 4cβ stains lighter than 4cα with the highest levels of staining found along radially oriented fibers that originate in layer 5 and traverse layer 4cβ.
Laminar and sublaminar boundaries could be even more easily distinguished, on the basis of characteristic patterns and intensities of zinc staining, in sections cut parallel to the cortical surface (Fig. 2). From this perspective it was, again, apparent that the zincergic innervation was most dense in layers 1, 2, 3, 4b, 5 and 6. Zinc staining was least dense in layer 4c and the subcortical white matter. Within layer 4c, staining levels distinguished sublayers 4cα and 4cβ. Layer 4cα was stained at moderate levels while 4cb was stained at lower levels. At high magnification, zinc staining in most layers consisted of a reticulated network of highly stained profiles embedded within a moderately stained neuropil consisting of fine, granular puncta and unstained foci (Fig. 2, plates at right; see layers 3, 4b and 6a for example). In adjacent sections counterstained for Nissl substance (data not shown), the majority of unstained foci were found to correspond to profiles of cell somata, the remainder to cross-sectional profiles of cortical blood vessels.
At lower magnification, the subcortical white matter appeared unstained (Fig. 1, wm), but when viewed at higher magnification, in tangential sections, more superficial strata of the white matter (Fig. 2, wms) exhibited a uniform distribution of zinc-positive puncta, while at deeper levels (Fig. 2, wmd) zinc-positive puncta were aligned along fibers which coursed parallel to the plane of section.
Patterned Staining for Zinc
Although individual layers in V1 could be characterized on the basis of differential staining intensities, patterned modulations in the levels of zinc staining were conspicuous in sub-laminae within layers 3, 4 and 6.
In layer 3, a periodic modulation of high to moderate levels of zinc staining was not readily apparent in the frontal plane (Fig. 1A,C). However, in sections cut parallel to the cortical surface (Fig. 2) a periodic distribution was manifested by reduced levels of zinc staining in irregularly shaped, blob-like regions, embedded within an intensely zinc-stained lattice. In a direct comparison to adjacent sections stained for CO it was clear that the areas of reduced zinc staining corresponded to blobs while the inter-blob regions, which are CO-poor, are zinc-rich (Fig. 3). Thus, the zincergic innervation of layer 3 respects functional compartments complementary to those demarcated by CO.
Layer 4a featured a periodic distribution of intensely and lightly stained columns of zinc viewed in the frontal plane (Fig. 1C). The zinc-rich columns often appeared cone-shaped, broader in deepest portions of layer 4a. In cross-section (Fig. 2; 4a), these zinc-rich columns in layer 4a appeared as circular lacunae, measuring between 30 and 100 μm in diameter, that were surrounded by regions exhibiting lower staining levels. CO histochemistry intensely stains a lattice in layer 4a that surrounds weakly stained lacunae (Horton and Hubel, 1981). A comparison with adjacent sections stained for CO indicated that, as in lamina 3, the patterns of zinc and CO staining were complementary in layer 4a (not shown).
Although layer 4cβ exhibited the lowest level of staining of any layer, higher staining levels were organized along vertical fibers that emerged from layer 5, coursed through layer 4cβ and disappeared into layer 4cα (Fig. 1C). In tangential sections through layer 4cα, these fibrils appeared as circular patches ~10 μm across and spaced at ~30 μm intervals (Fig. 2). This staining pattern suggests that the apical dendrites of layer 5 neurons, which extend through layer 4c and into supragranular layers, are selectively innervated by zincergic axon terminals in layer 4cβ.
Effects of Monocular Deprivation
Monocular Enucleation with 3 month survival
The visual cortex of all TTX-injected and enucleated monkeys showed striking ocular dominance-specific (OD) modifications in the level of staining for synaptic zinc. Three months after monocular enucleation, reductions in zinc staining were apparent in layers 3, 4cα and 6a (Fig. 4A,C,E,G,I). Adjacent sections stained for CO revealed marked reductions of CO staining in deprived eye-stripes within all cortical layers (Fig. 4B,D,F,H,J). When aligned using fiducial landmarks, stripes exhibiting reduced levels of zinc staining were found to be coincident with those showing reduced levels of CO staining (arrows indicate corresponding points in each pair of sections).
The level of staining for zinc and CO along enucleated-eye stripes in layer 3 was not homogeneously reduced (Fig. 4A,B). With CO, the blobs that were aligned along deprived eye stripes were reduced in size, while those corresponding to the intact eye were larger and appeared fused as a result of increased staining within inter-blob zones (Fig. 4B,D). The reduction of zinc staining in stripes corresponding to the enucleated eye appeared to be predominantly in inter-blob zones, thus forming stripes within this layer (Fig. 4A,C). Zinc staining in deprived-eye stripes within 4cα (Fig. 4E,G) and 6a (Fig. 4I) was similarly reduced. In all affected layers, with both stains, OD stripes corresponding to the enucleated eye were narrower than intact eye stripes, but this difference was much more pronounced with zinc staining.
Effects of 24 h of Monocular Deprivation
Levels of staining for synaptic zinc were markedly modified in an eye-specific manner in layers 3, 4a, 4cα and 4cβ, only 24 h after unilateral retinal ganglion cell inactivation by intravitreal infusion of tetrodotoxin (TTX; Fig. 5) or by enucleation (not shown). The modification of zinc staining levels in layer 3 was more pronounced following 24 h deprivation (Fig. 5A) than it was for 3 months deprivation (Fig. 4A,C). This was true of layer 4a as well (Fig. 5A,B), which appeared relatively unaffected after 3 months of deprivation (Fig. 4A).
In order to determine the eye-specificity of the altered zinc staining levels, we compared the patterned distribution of zinc staining in serially adjacent sections stained for CO and Zif268 (Fig. 6). The appearance of blobs, stained by CO in layer 3, was not sufficiently altered by monocular activity blockade for 24 h. However, as has been previously documented using Zif268 immunohistochemistry (Chaudhuri and Cynader, 1993; Chaudhuri et al., 1995) the ocular dominance distribution was clearly visible in the pattern of Zif268 staining, with reduced levels clearly demarcating injected-eye stripes (Fig. 6, Zif, solid white arrows). Surprisingly, when aligned using fiducial land-marks (open arrows) and compared, the stripes revealed by higher levels of staining for synaptic zinc corresponded to the TTX-injected eye while those corresponding to the non-deprived eye stained relatively lighter (Fig. 6, solid arrows indicate corresponding injected-eye stripes).
Although CO staining failed to reveal the ocular dominance pattern in layer 3, 24 h after monocular deprivation by enucleation or TTX infusion, reductions of CO staining in layer 4c were apparent in all animals following only 24 h of activity blockade, and provided unambiguous confirmation of the results provided using Zif268, that zinc staining was higher in deprived-versus nondeprived-eye domains (Fig. 7, arrows).
The effect of enucleation on zinc staining, with a 24 h premortal survival time, appeared less pronounced than that induced by TTX. Unlike that seen following TTX infusion (Figs 5 and 6), the eye-specific reduction of zinc staining that was apparent 24 h following enucleation was negligible in the upper portions of layer 3 (Fig. 8A compared with Figs 5A and 6) but clearly apparent in sections obtained near the layer 3/4a border (Fig. 8C).
In order to determine whether the modified zinc staining levels could be attributed to changes within specific cytological compartments with long- or short-term deprivation, we compared zinc staining within deprived and nondeprived compartments in all layers. Regardless of deprivation method and duration we could not discern the cytological correlate of altered zinc staining by light microscopy. Differences in zinc staining within deprived and nondeprived compartments of layer 3, at 3 months after enucleation, are shown in Figure 9, while similar compartments in layer 4cα from a 24 h deprived animal are illustrated in Figure 10) at higher magnification. Modulations of zinc staining levels in all cases were reflected in relative changes in the density of zinc-stained puncta both in layer 3 (Fig. 9C vs 9B) and layer 4cα (Fig. 10B vs 10C). Further characterization of the specific cytological locus of change would require ultra-structural analysis by electron microscopy.
Histochemical methods were used to examine the laminar and tangential distribution of zinc-containing axon terminals in the adult primate visual cortex and to determine the extent to which the levels of zinc were regulated by visual input. Each layer in visual cortex was found to exhibit characteristic levels and patterns of staining for zinc. In most layers, the pattern and density of staining appeared complementary to that of CO. The pattern of zinc staining in several thalamocortical recipient layers (3 and 4a) was particularly striking. In these layers, the levels of synaptic zinc were found to vary periodically, in the tangential plane, in a manner that is precisely complementary to compartments stained for CO. When adult monkeys were monocularly deprived by enucleation, and allowed to survive for 3 months, a significant reduction of zinc staining within deprived-eye stripes revealed the ocular dominance pattern. However, when the cortex was stained after only 24 h of deprivation either by monocular tetrodotoxin infusion or by enucleation, the pattern of zinc staining was opposite. Here, levels of zinc staining were increased in the deprived-eye, relative to non-deprived eye stripes.
Activity and experience-dependent plasticity in the adult cerebral cortex is thought to be mediated, principally, by horizontal connections that provide feedback and feed-forward interactions between and within higher and lower-order cortical areas (Gilbert, 1998). These forms of plasticity are thought to be determined by alterations of cortical functional architecture that, in the short term, are mediated by modifications of synaptic strength (Calford and Tweedale, 1988, 1991; Gilbert and Wiesel, 1992) and in the longer term by remodeling of intrinsic cortical connectivity (Darian-Smith and Gilbert, 1994). We contend, based on physiological and anatomical evidence, that zincergic neurons actively participate in both of these plastic processes.
Electrophysiological studies have demonstrated that the functional organization of the adult cerebral cortex is rapidly modified by changes in sensory experience. For example, the receptive field properties of neurons in the primary visual cortex (Gilbert and Wiesel, 1992) and the somatosensory cortex of primates (Calford and Tweedale, 1991) can be modulated by only minutes of altered sensory experience. In an effort to understand the underlying molecular correlates of these physiological changes, several labs have examined the effects of altered sensory experience on the expression of neuroactive signaling molecules. In the primary visual cortex of cats and monkeys, monocular deprivation results in decreased expression of glutamate (Carder and Hendry, 1994), GAD (Hendry and Jones, 1988), GABA (Hendry and Jones, 1988), GABAA receptors (Hendry et al., 1990) and NMDA receptor subunits (Catalano et al., 1997), while increasing the expression of calmodulin-dependent protein kinase (Hendry and Kennedy, 1986) and neurotrophic factors (Obata et al., 1999) in cortical domains corresponding to the deprived input. However, in each of these studies deprivation periods of several days to several weeks were required to affect the expression of these molecules.
The search for molecules whose expression patterns correlate temporally and spatially with rapid electrophysiological changes in the cortex has been more elusive. Until now, only CO (Horton et al., 2000), NMDA receptor subunits (Quinlan et al., 1999), immediate early genes (Rosen et al., 1992; Beaver et al., 1993) and transcription factors such as Zif268 (Chaudhuri and Cynader, 1993) and cAMP response element binding protein (Barth et al., 2000) have been shown to exhibit altered levels of expression after only a few hours of altered sensory experience.
Here, we have shown that synaptic zinc levels are dynamically modulated in primate V1 with differential immediate and long-term effects. The relative reduction in the level of synaptic zinc in deprived-eye stripes with long-term monocular deprivation has not been described before, in any model system. In the only related study, thus far reported, permanent vibrissectomy was found to be without long-term effect on zinc staining levels in barrel cortex (Czupryn and Skangiel-Kramska, 2001b). On the other hand, the rapid increases in zinc staining following monocular deprivation are consistent with recent results obtained from studies using the adult mouse somatosensory cortex. Here, synaptic zinc has been found to increase rapidly in barrel compartments corresponding to whiskers that are removed by trimming or plucking (Czupryn and Skangiel-Kramska, 2001a; Brown and Dyck, 2002, 2003b). We have found that levels of synaptic zinc are significantly elevated within 3 h of whisker removal, became maximal by 24 h (>20% increase), and remained high until 7 days post-deprivation (Brown and Dyck, 2002). Thereafter, zinc levels became reduced to normal levels as sensory input was recovered, and was directly correlated with the length of whiskers as they re-grew (Brown and Dyck, 2002, 2003b). Based on these findings, it is clear that levels of synaptic zinc are rapidly increased in cortical compartments that are deprived of ascending sensory input. There is no reason to believe that the activity-dependent modulation of zinc would be different in the primate visual cortex with short periods of input blockade, however, further investigations are required to unambiguously establish a similar temporo-spatial response.
Two other important, functionally relevant, issues arise from these findings. The first relates to the mechanisms responsible for modulating levels of zinc in axon terminals in an experience- or activity-dependent manner and whether these changes have functional consequences for synaptic transmission in the cerebral cortex. The second is concerned with the rapid, compartment-specific increase of zinc staining levels in deprived-eye domains of the visual cortex that are reversed with longer survival times. Both suggest that there are two distinct, temporally specific physiological responses of zincergic neurons to alterations of sensory input and neuronal activity.
To address the first issue, one could argue that increased levels of zinc staining in deprived-eye stripes reflect activity-dependent changes in the uptake of zinc into, and its release from, axon terminals. Previous work that has assessed the kinetics of zinc turnover shows that zinc is released in an activity-dependent manner (Assaf and Chung, 1984; Howell et al., 1984). Once released, extracellular concentrations of zinc are regulated by transporters that facilitate the re-uptake of zinc into the presynaptic terminal and into synaptic vesicles (Palmiter et al., 1996; Wenzel et al., 1997). However, if the release of zinc is activity-dependent but the uptake is not, then it would seem plausible that, in situations that result in decreased afferent neuronal activity in the cortex, such as that caused by monocular visual deprivation or by whisker removal (Durham and Woolsey, 1978; Kelly et al., 1999), one disrupts zinc homeostasis such that more zinc is taken up into zincergic axon terminals than is released.
Alternatively, modulations of zinc levels in presynaptic terminals might reflect, or contribute to, experience-dependent changes of the synaptic organization of the cerebral cortex. Current hypotheses for mechanisms supporting experience-dependent plasticity in the cortex suggest that NMDA receptor-dependent forms of long-term potentiation (LTP) and long-term depression (LTD) might play an important role in this phenomenon (Artola and Singer, 1987; Kirkwood et al., 1996; Feldman, 2000). The appeal of NMDA-dependent LTP and LTD as processes that mediate experience-dependent plasticity is in part due to the fact that (a) NMDA-dependent LTP and LTD can be readily induced in primary sensory regions of the cortex (Bear and Kirkwood, 1993; Donoghue, 1995), (b) manipulations of sensory experience can produce LTP and LTD-like changes in the response properties of cortical neurons (Diamond et al., 1993, 1994; Wallace and Fox, 1999) and (c) pharmacological blockade of NMDA receptors disrupts experience-dependent reorganizations of the synaptic organization of the cortex (Jablonska et al., 1995; Garraghty and Muja, 1996; Rema et al., 1998). With this in mind, it is possible that activity-dependent changes in the levels of zinc release may provide a substrate for these processes to occur (Weiss et al., 1989). Although the precise functional role of synaptic zinc is unknown in this regard, studies have shown that zinc is capable of modulating NMDA-dependent forms of LTP and LTD in the hippocampus (Xie and Smart, 1994; Lu et al., 2000; Li et al., 2001). Synaptic zinc, thus, seems well positioned to function as a mediator of rapid changes in synaptic efficacy by virtue of its potent neuro-modulatory effects on NMDA and non-NMDA receptor-mediated glutamatergic neurotransmission (Westbrook and Mayer, 1987; Christine and Choi, 1990; Smart et al., 1994; Vogt et al., 2000). Increased levels of vesicular zinc in axon terminals could, therefore, translate into enhanced modulation of postsynaptic glutamate receptors and the facilitation of synaptic plasticity in eye-specific compartments. Our current studies are directed at assessing the possibility that increased levels of synaptic zinc inhibit NMDA receptors, thus facilitating LTD-like processes and allowing intact-eye circuits to subsume deprived-eye domains.
Based on ultrastructural (Danscher, 1982; Pérez-Clausell and Danscher, 1985) and neurochemical (Beaulieu et al., 1992) criteria axon terminals in the neocortex that are enriched with zinc are also glutamatergic. Although a large proportion of both thalamocortical and cortico-cortical projection neurons are considered glutamatergic, it is not likely that the zincergic innervation of primate visual cortex arises from the thalamus but, rather, from intracortical projections. First, the distribution of synaptic zinc is compartmented, displaying highest levels of organization in laminar and columnar domains that are complementary to those labeled by CO. Thalamocortical inputs to visual cortex largely target CO-rich layers (Hubel and Wiesel, 1972; Blasdel and Lund, 1983; Hendry and Yoshioka, 1994), which are relatively zinc-poor. Furthermore, CO staining is selectively enriched in blob and lattice compartments in layers 3 and 4a, respectively (Horton and Hubel, 1981), while zincergic terminals are enriched in complementary interblob and lacunar domains. Secondly, retrograde tracing techniques specific for zincergic projections label large numbers of neurons in layers 2, 3 and 6, with very few originating in layer 5, and none from layer 4 (Garrett et al., 1992; Casanovas-Aguilar et al., 2002). These neurons have been shown to provide local and long distance horizontal projections that arise from cortical regions within the ipsilateral hemisphere, as well as from supra- and infra-granular neurons of the contralateral cortical hemisphere, but no neurons are labeled in the thalamus (Garrett et al., 1992; Casanovas-Aguilar et al., 1995, 2002) (Brown and Dyck, 2003a).
Many of the excitatory pyramidal neurons in supragranular layers of primate V1 send long-range, horizontal connections forming patchy terminals (Rocklund and Lund, 1983; McGuire et al., 1991) that appear to align with common functional, columnar compartments (Malach et al., 1993; Yabuta and Callaway, 1998). As indicated above, zincergic neurons in the mammalian cortex contribute to a subset of excitatory connections that contribute to the vast associational network of cortico-cortical projections. Support for zincergic neurons being the ones responsible for providing this periodic array of tangential inputs comes from recent anatomical studies in rat visual cortex. Here, groups or clusters of zincergic neurons in layers 2/3 of rat visual cortex have been shown to innervate neurons in supragranular and infragranular layers of primary and secondary visual areas in a patchy manner (Casanovas-Aguilar et al., 2002). Thus, alterations in neuronal activity levels caused by blockade of ascending monocular inputs could be manifested by column-specific modulations in the levels of zinc staining in these horizontal pathways, as is described here in primate V1.
In addition to long-range horizontal projections that connect functional units within visual areas or between them, pyramidal neurons in the visual cortex are also key participants in the circuitry of local horizontal as well as vertical, intra-column projections. In the rat visual cortex, pyramidal neurons having a zincergic phenotype have been shown to be also involved in these intra-areal circuits, densely innervating all zinc-rich neocortical layers [layers 1, 2/3, 5 and 6 (Casanovas-Aguilar et al., 2002)]. Infragranular zincergic neurons were also shown to provide local connections in the visual cortex, but only within infragranular layers, not to overlying supragranular layers. Based on these results, it is unlikely that infragranular zincergic neurons are the ones that modulate levels of zinc in their synaptic boutons but, rather, it is more likely to be the population of layer 2/3 neurons that do so. Anatomical tract tracing studies of the type that have been performed in the rat will be necessary to confirm this hypothesis.
Taken together, the evidence provided in this paper suggests that a subpopulation of short- and long-range intracortically projecting neurons dynamically modulate zinc levels in their synaptic terminals in an activity- and/or experience-dependent manner. It remains to be seen how these zinc-specific responses are translated into changes in the efficacy of synaptic signaling and, specifically, how zinc and zincergic neurons participate in short and longer-term forms of plasticity in the adult cerebral cortex.
We are grateful to Rodrigo Bravo for providing the Zif268 antibody and to Jirin Tan for his expert technical assistance. This work was supported by an NSERC fellowship and operating grant to RHD and grants from the Canadian Institutes for Health Research to MSC and RHD.
Address correspondence to Richard Dyck, Department of Psychology, The University of Calgary, Alberta, Canada, T2N 1N4. Email: firstname.lastname@example.org.