Systematic regulation of spine sizes and densities in pyramidal neurons

Dendritic spines receive most excitatory inputs in the CNS. Recent evidence has demonstrated that the spine head volume is linearly correlated with the readily releasable pool of neurotransmitter and the PSD size. These correlations can be used to functionally interpret spine morphology. Using Golgi impregnations and light microscopy, we reconstructed 23000 spines from pyramidal neurons in layers 2/3, 4, 5 and 6 of mouse primary visual cortex and CA1 hippocampal region and measured their spine head diameters and densities. Spine head diameters and densities are variable within and across cells, although they are similar between apical and basal dendrites. When compared to other regions, layer 5 neurons have larger spine heads and CA1 neurons higher spine densities. Interestingly, we detect a correlation between spine head diameter and interspine distance within and across cells, whereby larger spines are spaced further away from each other than smaller spines. Finally, in CA1 neurons, spine head diameters are larger, and spine density lower, in distal apical dendrites (>200 microm from soma) compared to proximal regions. These results reveal that spine morphologies and densities, and therefore synaptic properties, are jointly modulated with respect to cortical region, laminar position, and, in some cases, even the position of the spine along the dendritic tree. Individual neurons also appear to regulate their apical and basal spine densities and morphologies in concert. Our data provide evidence for a homeostatic control of excitatory synaptic strength.     

Cortical integration in the visual system of the macaque monkey: large-scale morphological differences in the pyramidal neurons in the occipital, parietal and temporal lobes.

Layer III pyramidal neurons were injected with Lucifer yellow in tangential cortical slices taken from the inferior temporal cortex (area TE) and the superior temporal polysensory (STP) area of the macaque monkey. Basal dendritic field areas of layer III pyramidal neurons in area STP are significantly larger, and their dendritic arborizations more complex, than those of cells in area TE. Moreover, the dendritic fields of layer III pyramidal neurons in both STP and TE are many times larger and more complex than those in areas forming 'lower' stages in cortical visual processing, such as the first (V1), second (V2), fourth (V4) and middle temporal (MT) visual areas. By combining data on spine density with those of Sholl analyses, we were able to estimate the average number of spines in the basal dendritic field of layer III pyramidal neurons in each area. These calculations revealed a 13-fold difference in the number of spines in the basal dendritic field between areas STP and V1 in animals of similar age. The large differences in complexity of the same kind of neuron in different visual areas go against arguments for isopotentiality of different cortical regions and provide a basis that allows pyramidal neurons in temporal areas TE and STP to integrate more inputs than neurons in more caudal visual areas.     

Alterations to Dendritic Spine Morphology,but Not Dendrite Patterning,of Cortical Projection Neurons in Tc1 and Ts1Rhr Mouse Models of Down Syndrome

Down Syndrome (DS) is a highly prevalent developmental disorder, affecting 1/700 births. Intellectual disability, which affects learning and memory, is present in all cases and is reflected by below average IQ. We sought to determine whether defective morphology and connectivity in neurons of the cerebral cortex may underlie the cognitive deficits that have been described in two mouse models of DS, the Tc1 and Ts1Rhr mouse lines. We utilised in utero electroporation to label a cohort of future upper layer projection neurons in the cerebral cortex of developing mouse embryos with GFP, and then examined neuronal positioning and morphology in early adulthood, which revealed no alterations in cortical layer position or morphology in either Tc1 or Ts1Rhr mouse cortex. The number of dendrites, as well as dendrite length and branching was normal in both DS models, compared with wildtype controls. The sites of projection neuron synaptic inputs, dendritic spines, were analysed in Tc1 and Ts1Rhr cortex at three weeks and three months after birth, and significant changes in spine morphology were observed in both mouse lines. Ts1Rhr mice had significantly fewer thin spines at three weeks of age. At three months of age Tc1 mice had significantly fewer mushroom spines - the morphology associated with established synaptic inputs and learning and memory. The decrease in mushroom spines was accompanied by a significant increase in the number of stubby spines. This data suggests that dendritic spine abnormalities may be a more important contributor to cognitive deficits in DS models, rather than overall neuronal architecture defects.     

Structural dynamics of dendritic spines: Molecular composition,geometry and functional regulation

The development of dendritic spines with specific geometry and membrane composition is critical for proper synaptic function. Specific spine membrane architecture, sub-spine microdomains and spine head and neck geometry allow for well-coordinated and compartmentalized signaling, disruption of which could lead to various neurological diseases. Research from neuronal cell culture, brain slices and direct in vivo imaging indicates that dendritic spine development is a dynamic process which includes transition from small dendritic filopodia through a series of structural refinements to elaborate spines of various morphologies. Despite intensive research, the precise coordination of this morphological transition, the changes in molecular composition, and the relation of spines of various morphologies to function remain a central enigma in the development of functional neuronal circuits. Here, we review research so far and aim to provide insight into the key events that drive structural change during transition from immature filopodia to fully functional spines and the relevance of spine geometry to function.     

Electrical advantages of dendritic spines

Many neurons receive excitatory glutamatergic input almost exclusively onto dendritic spines. In the absence of spines, the amplitudes and kinetics of excitatory postsynaptic potentials (EPSPs) at the site of synaptic input are highly variable and depend on dendritic location. We hypothesized that dendritic spines standardize the local geometry at the site of synaptic input, thereby reducing location-dependent variability of local EPSP properties. We tested this hypothesis using computational models of simplified and morphologically realistic spiny neurons that allow direct comparison of EPSPs generated on spine heads with EPSPs generated on dendritic shafts at the same dendritic locations. In all morphologies tested, spines greatly reduced location-dependent variability of local EPSP amplitude and kinetics, while having minimal impact on EPSPs measured at the soma. Spine-dependent standardization of local EPSP properties persisted across a range of physiologically relevant spine neck resistances, and in models with variable neck resistances. By reducing the variability of local EPSPs, spines standardized synaptic activation of NMDA receptors and voltage-gated calcium channels. Furthermore, spines enhanced activation of NMDA receptors and facilitated the generation of NMDA spikes and axonal action potentials in response to synaptic input. Finally, we show that dynamic regulation of spine neck geometry can preserve local EPSP properties following plasticity-driven changes in synaptic strength, but is inefficient in modifying the amplitude of EPSPs in other cellular compartments. These observations suggest that one function of dendritic spines is to standardize local EPSP properties throughout the dendritic tree, thereby allowing neurons to use similar voltage-sensitive postsynaptic mechanisms at all dendritic locations.     

A Septin-Dependent Diffusion Barrier at Dendritic Spine Necks

Excitatory glutamatergic synapses at dendritic spines exchange and modulate their receptor content via lateral membrane diffusion. Several studies have shown that the thin spine neck impedes the access of membrane and solute molecules to the spine head. However, it is unclear whether the spine neck geometry alone restricts access to dendritic spines or if a physical barrier to the diffusion of molecules exists. Here, we investigated whether a complex of septin cytoskeletal GTPases localized at the base of the spine neck regulates diffusion across the spine neck. We found that, during development, a marker of the septin complex, Septin7 (Sept7), becomes localized to the spine neck where it forms a stable structure underneath the plasma membrane. We show that diffusion of receptors and bulk membrane, but not cytoplasmic proteins, is slower in spines bearing Sept7 at their neck. Finally, when Sept7 expression was suppressed by RNA interference, membrane molecules explored larger membrane areas. Our findings indicate that Sept7 regulates membrane protein access to spines.     

Biphasic Synaptic Ca Influx Arising from Compartmentalized Electrical Signals in Dendritic Spines

Excitatory synapses on mammalian principal neurons are typically formed onto dendritic spines, which consist of a bulbous head separated from the parent dendrite by a thin neck. Although activation of voltage-gated channels in the spine and stimulus-evoked constriction of the spine neck can influence synaptic signals, the contribution of electrical filtering by the spine neck to basal synaptic transmission is largely unknown. Here we use spine and dendrite calcium (Ca) imaging combined with 2-photon laser photolysis of caged glutamate to assess the impact of electrical filtering imposed by the spine morphology on synaptic Ca transients. We find that in apical spines of CA1 hippocampal neurons, the spine neck creates a barrier to the propagation of current, which causes a voltage drop and results in spatially inhomogeneous activation of voltage-gated Ca channels (VGCCs) on a micron length scale. Furthermore, AMPA and NMDA-type glutamate receptors (AMPARs and NMDARs, respectively) that are colocalized on individual spine heads interact to produce two kinetically and mechanistically distinct phases of synaptically evoked Ca influx. Rapid depolarization of the spine triggers a brief and large Ca current whose amplitude is regulated in a graded manner by the number of open AMPARs and whose duration is terminated by the opening of small conductance Ca-activated potassium (SK) channels. A slower phase of Ca influx is independent of AMPAR opening and is determined by the number of open NMDARs and the post-stimulus potential in the spine. Biphasic synaptic Ca influx only occurs when AMPARs and NMDARs are coactive within an individual spine. These results demonstrate that the morphology of dendritic spines endows associated synapses with specialized modes of signaling and permits the graded and independent control of multiple phases of synaptic Ca influx.     

Spine-neck geometry determines NMDA receptor-dependent Ca2+ signaling in dendrites

Increases in cytosolic Ca2+ concentration ([Ca2+]i) mediated by NMDA-sensitive glutamate receptors (NMDARs) are important for synaptic plasticity. We studied a wide variety of dendritic spines on rat CA1 pyramidal neurons in acute hippocampal slices. Two-photon uncaging and Ca2+ imaging revealed that NMDAR-mediated currents increased with spine-head volume and that even the smallest spines contained a significant number of NMDARs. The fate of Ca2+ that entered spine heads through NMDARs was governed by the shape (length and radius) of the spine neck. Larger spines had necks that permitted greater efflux of Ca2+ into the dendritic shaft, whereas smaller spines manifested a larger increase in [Ca2+]i within the spine compartment as a result of a smaller Ca2+ flux through the neck. Spine-neck geometry is thus an important determinant of spine Ca2+ signaling, allowing small spines to be the preferential sites for isolated induction of long-term potentiation.     

The impacts of geometry and binding on CaMKII diffusion and retention in dendritic spines

We used a particle-based Monte Carlo simulation to dissect the regulatory mechanism of molecular translocation of CaMKII, a key regulator of neuronal synaptic function. Geometry was based upon measurements from EM reconstructions of dendrites in CA1 hippocampal pyramidal neurons. Three types of simulations were performed to investigate the effects of geometry and other mechanisms that control CaMKII translocation in and out of dendritic spines. First, the diffusional escape rate of CaMKII from model spines of varied morphologies was examined. Second, a postsynaptic density (PSD) was added to study the impact of binding sites on this escape rate. Third, translocation of CaMKII from dendrites and trapping in spines was investigated using a simulated dendrite. Based on diffusion alone, a spine of average dimensions had the ability to retain CaMKII for duration of ~4 s. However, binding sites mimicking those in the PSD controlled the residence time of CaMKII in a highly nonlinear manner. In addition, we observed that F-actin at the spine head/neck junction had a significant impact on CaMKII trapping in dendritic spines. We discuss these results in the context of possible mechanisms that may explain the experimental results that have shown extended accumulation of CaMKII in dendritic spines during synaptic plasticity and LTP induction.     

The actin binding domain of βI-spectrin regulates the morphological and functional dynamics of dendritic spines

Actin microfilaments regulate the size, shape and mobility of dendritic spines and are in turn regulated by actin binding proteins and small GTPases. The βI isoform of spectrin, a protein that links the actin cytoskeleton to membrane proteins, is present in spines. To understand its function, we expressed its actin-binding domain (ABD) in CA1 pyramidal neurons in hippocampal slice cultures. The ABD of βI-spectrin bundled actin in principal dendrites and was concentrated in dendritic spines, where it significantly increased the size of the spine head. These effects were not observed after expression of homologous ABDs of utrophin, dystrophin, and α-actinin. Treatment of slice cultures with latrunculin-B significantly decreased spine head size and decreased actin-GFP fluorescence in cells expressing the ABD of α-actinin, but not the ABD of βI-spectrin, suggesting that its presence inhibits actin depolymerization. We also observed an increase in the area of GFP-tagged PSD-95 in the spine head and an increase in the amplitude of mEPSCs at spines expressing the ABD of βI-spectrin. The effects of the βI-spectrin ABD on spine size and mEPSC amplitude were mimicked by expressing wild-type Rac3, a small GTPase that co-immunoprecipitates specifically with βI-spectrin in extracts of cultured cortical neurons. Spine size was normal in cells co-expressing a dominant negative Rac3 construct with the βI-spectrin ABD. We suggest that βI-spectrin is a synaptic protein that can modulate both the morphological and functional dynamics of dendritic spines, perhaps via interaction with actin and Rac3.     

Subtle re-location of dendritic spine branches containing membrane with fast spiking mechanisms can alter synaptic efficacy

The potential physiological impact of morphological changes in the active dendritic spines, which are believed to be associated with altered synaptic efficacy, was investigated in a computer simulation study using the NEURON package [1]. A compartmental model of a simplified neuron was built, which included 30 complex spines (neck, head, and active zone) and accommodating AMPA-type synaptic inputs with alpha-function conductances. Hodgkin-Huxley type excitable membranes were inserted into the spine heads. It was shown that arranging spines in dense clusters, as opposed to a uniformly random spine distribution, has a negligible effect on the synaptic signal transfer (other model conditions, including synaptic input and spine density, remained unchanged). However, if a proportion (e.g., 3–20%) of the spines partly fuse with their neighbors forming branched spines, this could increase dramatically the cell response to the unchanged synaptic input. Results of this pilot study provide the basis for a more detailed investigation of the relationship between the spine arrangement and synaptic function, considering dual-component synaptic currents and mechanisms controlling ion fluxes in the dendritic compartments.     

Dynamic microtubules promote synaptic NMDA receptor-dependent spine enlargement

Most excitatory synaptic terminals in the brain impinge on dendritic spines. We and others have recently shown that dynamic microtubules (MTs) enter spines from the dendritic shaft. However, a direct role for MTs in long-lasting spine plasticity has yet to be demonstrated and it remains unclear whether MT-spine invasions are directly influenced by synaptic activity. Lasting changes in spine morphology and synaptic strength can be triggered by activation of synaptic NMDA receptors (NMDARs) and are associated with learning and memory processes. To determine whether MTs are involved in NMDAR-dependent spine plasticity, we imaged MT dynamics and spine morphology in live mouse hippocampal pyramidal neurons before and after acute activation of synaptic NMDARs. Synaptic NMDAR activation promoted MT-spine invasions and lasting increases in spine size, with invaded spines exhibiting significantly faster and more growth than non-invaded spines. Even individual MT invasions triggered rapid increases in spine size that persisted longer following NMDAR activation. Inhibition of either NMDARs or dynamic MTs blocked NMDAR-dependent spine growth. Together these results demonstrate for the first time that MT-spine invasions are positively regulated by signaling through synaptic NMDARs, and contribute to long-lasting structural changes in targeted spines.     

Cortical heterogeneity: Implications for visual processing and polysensory integration

Recent studies have revealed substantial variation in pyramidal cell structure in different cortical areas. Moreover, cell morphology has been shown to vary in a systematic fashion such that cells in visual association areas are larger and more spinous than those in the primary visual area. Various aspects of these structural differences appear to be important in influencing neuronal function. At the cellular level, differences in the branching patterns in the dendritic arbour may allow for varying degrees of non-linear compartmentalisation. Differences in total dendritic length and spine number may determine the number of inputs integrated by individual cells. Variations in spine density and geometry may affect cooperativity of inputs and shunting inhibition, and the tangential dimension of the dendritic arbours may determine sampling strategies within cortex. At the systems level, regional variation in pyramidal cell structure may determine thedegree of recurrent excitation through reentrant circuits influencing the discharge properties of individual neurones and the functional signature of the circuits they compose. The ability of pyramidal neurones in visual areas of the parietal and temporal lobes to integrate large numbers of excitatory inputs may also facilitate cortical binding. Here I summarise what I consider to be among the most salient, and testable, aspects of an inter-relationship between morphological and functional heterogeneity in visual cortex.     

Geometry and structural plasticity of synaptic connectivity

Changes in synaptic connectivity patterns through the formation and elimination of dendritic spines may contribute to structural plasticity in the brain. We characterize this contribution quantitatively by estimating the number of different synaptic connectivity patterns attainable without major arbor remodeling. This number depends on the ratio of the synapses on a dendrite to the axons that pass within a spine length of that dendrite. We call this ratio the filling fraction and calculate it from geometrical analysis and anatomical data. The filling fraction is 0.26 in mouse neocortex, 0.22-0.34 in rat hippocampus. In the macaque visual cortex, the filling fraction increases by a factor of 1.6-1.8 from area V1 to areas V2, V4, and 7a. Since the filling fraction is much smaller than 1, spine remodeling can make a large contribution to structural plasticity.     

The life cycle of Ca(2+) ions in dendritic spines

Spine Ca(2+) is critical for the induction of synaptic plasticity, but the factors that control Ca(2+) handling in dendritic spines under physiological conditions are largely unknown. We studied [Ca(2+)] signaling in dendritic spines of CA1 pyramidal neurons and find that spines are specialized structures with low endogenous Ca(2+) buffer capacity that allows large and extremely rapid [Ca(2+)] changes. Under physiological conditions, Ca(2+) diffusion across the spine neck is negligible, and the spine head functions as a separate compartment on long time scales, allowing localized Ca(2+) buildup during trains of synaptic stimuli. Furthermore, the kinetics of Ca(2+) sources governs the time course of [Ca(2+)] signals and may explain the selective activation of long-term synaptic potentiation (LTP) and long-term depression (LTD) by NMDA-R-mediated synaptic Ca(2+).     

Activity-dependent dendritic spine structural plasticity is regulated by small GTPase Rap1 and its target AF-6

Activity-dependent remodeling of dendritic spines is essential for neural circuit development and synaptic plasticity, but the mechanisms that coordinate synaptic structural and functional plasticity are not well understood. Here we investigate the signaling pathways that enable excitatory synapses to undergo activity-dependent structural modifications. We report that activation of NMDA receptors in cultured cortical neurons induces spine morphogenesis and activation of the small GTPase Rap1. Rap1 bimodally regulates spine morphology: activated Rap1 recruits the PDZ domain-containing protein AF-6 to the plasma membrane and induces spine neck elongation, while inactive Rap1 dissociates AF-6 from the membrane and induces spine enlargement. Rap1 also regulates spine content of AMPA receptors: thin spines induced by Rap1 activation have reduced GluR1-containing AMPA receptor content, while large spines induced by Rap1 inactivation are rich in AMPA receptors. These results identify a signaling pathway that regulates activity-dependent synaptic structural plasticity and coordinates it with functional plasticity.     

Morphological constraints on calcium dependent glutamate receptor trafficking into individual dendritic spine

Glutamate receptor trafficking into dendritic spines is a pivotal step in synaptic plasticity, yet the relevance of plasticity-producing rise of [Ca2+]i and of spine morphology to subsequent delivery of glutamate receptors into dendritic spine heads are still not well understood. Following chemical induction of LTP, an increase in eGFP-GluR1 fluorescence in short but not long dendritic spines of cultured hippocampal neurons was found. Repeated flash photolysis of caged calcium, which produced a transient rise of [Ca2+]i inside spine heads caused a selective, actin and protein synthesis dependent increase of eGFP-GluR1 in these spines. Strikingly, GluR1 increase was correlated with the ability of a calcium transient generated in the spine head to diffuse into the parent dendrite, and inversely correlated with the length of the spine: short spines were more likely to raise GluR1 than long ones. These observations link, for the first time, calcium transients in dendritic spines with spine morphology and its ability to undergo synaptic plasticity.     

Reelin in the extracellular matrix and dendritic spines of the cortex and hippocampus: a comparison between wild type and heterozygous reeler mice by immunoelectron microscopy

Reelin is a glycoprotein (～400 kDa) secreted by GABAergic neurons into the extracellular matrix of the neocortex and hippocampus as well as other areas of adult rodent and nonhuman primate brains. Recent findings indicate that the heterozygote reeler mouse (haploinsufficient for the reeler gene) shares several neurochemical and behavioral abnormalities with schizophrenia and bipolar disorder with mania. These include (1) a downregulation of both reelin mRNA and the translated proteins, (2) a decrease in the number of dendritic spines in cortical and hippocampal neurons, (3) a concomitant increase in the packing density of cortical pyramidal neurons, and (4) an age-dependent decrease in prepulse inhibition of startle. Interestingly, the heterozygous reeler mouse does not exhibit the unstable gait or the neuroanatomy characteristic of the null mutant reeler mouse. Immunocytochemical studies of the expression of reelin in mice have been primarily limited to light microscopy. In this study we present new immunoelectron microscopy data that delineates the subcellular localization of reelin in the cortex and hippocampus of the wild-type mouse, and compares these results to reelin expression in the heterozygous reeler mouse. In discontinuous areas of cortical layers I and II and the inner blade area of the dentate gyrus of the wild type mouse, extracellular reelin is associated with dendrites and dendritic spine postsynaptic specializations. Similar associations have been detected in the CA1 stratum oriens and other areas of the hippocampus. In the hippocampus, reelin expression is more expansive and more widespread than in cortical layers I and II. In contrast, extracellular reelin immunoreactivity is greatly diminished in all areas examined in the heterozygous reeler mouse. However, some cell bodies of GABAergic neurons in the cortex and hippocampus demonstrate an increased accumulation of reelin in the Golgi and endoplasmic reticulum. We suggest that in the heterozygous reeler mouse a downregulation of reelin biosynthesis results in a decreased rate of secretion into the extracellular space. This inhibits dendritic spine maturation and plasticity and leads to dissociation of dendritic postsynaptic density integrity and atrophy of spines. We speculate that the haploinsufficient reeler mouse may provide a model for future studies of the role of reelin, as it may be related to psychosis vulnerability.     

LOTOS-based two-photon calcium imaging of dendritic spines in vivo

Neurons in the mammalian brain receive thousands of synaptic inputs on their dendrites. In many types of neurons, such as cortical pyramidal neurons, excitatory synapses are formed on fine dendritic protrusions called spines. Usually, an individual spine forms a single synaptic contact with an afferent axon. In this protocol, we describe a recently established experimental procedure for measuring intracellular calcium signals from dendritic spines in cortical neurons in vivo by using a combination of two-photon microscopy and whole-cell patch-clamp recordings. We have used mice as an experimental model system, but the protocol may be readily adapted to other species. This method involves data acquisition at high frame rates and low-excitation laser power, and is termed low-power temporal oversampling (LOTOS). Because of its high sensitivity of fluorescence detection and reduced phototoxicity, LOTOS allows for prolonged and stable calcium imaging in vivo. Key aspects of the protocol, which can be completed in 5-6 h, include the use of a variant of high-speed two-photon imaging, refined surgery procedures and optimized tissue stabilization.     

Changes in the ultrastructure and morphometric parameters of the cortical axospinal synapses as affected by a calcium-free medium

Slice preparations of the white mouse sensorimotor cortex were examined for the morphometry of axospinal synapses in the material obtained under a standard experimental procedure and in sections perfused in calcium-free medium. The calcium-free medium in slice preparations was shown to reduce considerably the spine length and to increase the amount of spines with a negative curvature of the spine head in the contact area. Besides, it is noticed that the postsynaptic thickening grows, the active zone length decreases, the number of perforated contacts drops and the spine apparatus moves to the dendritic zone. It is supposed that the spine apparatus transfer is probably due to the reorganization of spine cytoskeleton, i.e., it is determined by both the motor activity of dendritic microtubules and the polymerization level of spine microfilaments.