Rapid Dendritic Morphogenesis in CA1 Hippocampal Dendrites Induced by Synaptic Activity

See allHide authors and affiliations

Science  19 Mar 1999:
Vol. 283, Issue 5409, pp. 1923-1927
DOI: 10.1126/science.283.5409.1923


Activity shapes the structure of neurons and their circuits. Two-photon imaging of CA1 neurons expressing enhanced green fluorescent protein in developing hippocampal slices from rat brains was used to characterize dendritic morphogenesis in response to synaptic activity. High-frequency focal synaptic stimulation induced a period (longer than 30 minutes) of enhanced growth of small filopodia-like protrusions (typically less than 5 micrometers long). Synaptically evoked growth was long-lasting and localized to dendritic regions close (less than 50 micrometers) to the stimulating electrode and was prevented by blockade of N-methyl-d-aspartate receptors. Thus, synaptic activation can produce rapid input-specific changes in dendritic structure. Such persistent structural changes could contribute to the development of neural circuitry.

Coordinated patterns of activity help to organize neural circuits throughout the brain (1). In particular, activity shapes the structure of sensory maps (2) and individual neurons (3) throughN-methyl-d-aspartate (NMDA) receptor–dependent processes, which suggests that synapse-specific associative changes are involved. Relatively little is known about the role of activity in the development of dendritic morphology. A number of studies have addressed whether long-term potentiation (LTP) produces postsynaptic structural changes. Using electron microscopy (EM) analysis of fixed tissue (4) and optical imaging of living tissue (5), these studies have produced inconsistent results. In the mammalian brain, dendrites develop in a stereotypical sequence: Soon after birth the relatively smooth dendrites of neonates sprout numerous thin filopodia-like protrusions, which are later replaced by dendritic spines as the brain matures (6). Studies in developing cultured rat brain slices (7) and disassociated cultures (8–10) from hippocampal area CA1 show that dendritic protrusions, including filopodia and spines, are structurally dynamic, perhaps powered by actin motility (10). Filopodia make synapses that coexist with axodendritic synapses in cultures (8, 9) and in vivo (11). These observations have led to the hypothesis that dendritic filopodia are crucial in establishing synaptic connections during development, paving the way for the development of mature spines (7,9, 11). Activity-dependent dendritic morphogenesis could therefore underlie Hebbian plasticity during development. We used vital imaging of dendrites in living rat brain slices to directly observe dendritic morphogenesis evoked by synaptic activity.

A small number of CA1 pyramidal neurons in organotypic cultured hippocampal slices (12, 13) were infected with a neurotropic recombinant Sindbis virus (14) expressing enhanced green fluorescent protein (EGFP) (15) (Fig. 1A). When visualized with a custom-built two-photon laser scanning microscope (TPLSM) (16,17), infected neurons showed bright homogeneous fluorescence throughout their dendritic and axonal arbors, revealing detailed morphology (Fig. 1, A through C) with numerous dendritic protrusions (the density of the protrusions was 0.52 ± 0.2 μm–1) (18). Time-lapse imaging of dendritic segments showed that dendritic morphology is dynamic on time scales of minutes to hours. Protrusions appeared and disappeared or changed shape (Fig. 1D) (7).

Figure 1

TPLSM imaging of CA1 pyramidal cell dendrites labeled with EGFP (in a slice 9 days in culture). Images shown are maximum value projections of several sections acquired 0.5 μm apart. (A) CA1 region of a hippocampal slice showing several infected neurons, including dendritic arbors and axons. (B) Apical dendrite and secondary branches from the boxed region in (A). (C) Higher magnification image from the boxed region in (B) showing dendritic protrusions. (D) Time-lapse image sequence showing sprouting (solid arrowhead) and retracting (open arrowhead) filopodia (in a slice 9 days in culture; eight sections per image). Time stamps (in minutes) are indicated in the lower left corners.

To examine the effects of synaptic activity on dendritic morphology, a glass electrode (0.4 to 1 megohm) was placed close (∼3 to 10 μm) to a target dendrite (Fig. 2A, left). The proximity of the stimulating electrode was absolutely necessary to ensure that the dendrite would receive localized synaptic stimulation (17). After the delivery of a tetanic stimulus (two trains of 100 pulses at 100 Hz) in a majority of cases (12 out of 14 experiments) new protrusions appeared (Fig. 2), and the total density of protrusions increased by 19 ± 4% (40 min after tetanus; mean ± SEM, n = 14). Comparison of changes in protrusion number after tetanus with changes in the absence of tetanus revealed that tetanus significantly increased the sprouting of new protrusions [P < 0.005, Wilcoxon matched-pairs signed-ranks (WMPSR) test] (19). Growth began only after a brief delay after tetanus, with no detectable change in dendritic structure at the first time point (3 min; Fig. 2, B and C). Growth was significant 20 min after tetanus and continued for at least another 40 min. The images revealed that the most dramatic feature of tetanus-evoked morphogenesis was an increase in the number of long (>2.75 μm) filopodia-like protrusions (Fig. 2A; length = 4.0 ± 1.5 μm). Before tetanus, filopodia were relatively rare (density = 0.053 ± 0.038 μm–1). Forty minutes after tetanus, their number had increased by 145 ± 65% (mean ± SEM, n = 14) (Fig. 2C, open circles). This included de novo genesis of new filopodia >2.75 μm (43%) as well as growth of existing protrusions to lengths exceeding 2.75 μm (57%) (Fig. 2). Because of their pronounced plasticity and their potential role as spine precursors, we focused further analysis on filopodia.

Figure 2

Growth of dendritic protrusions in response to tetanic stimulation. (A) Left, schematic showing the experimental geometry. A dendrite of a slice 9 days in culture was electrically stimulated with a glass pipette. Right, the stimulus produced growth mainly of filopodia (solid white arrows). Other structures disappeared (open arrows). Time stamps are in minutes. (B) Example of the time course of the number of protrusions in response to tetanic stimulation (gray arrow) at the test site (open circles, within ∼30 μm of the stimulating electrode) and a control site (solid circles, more than 100 μm from the stimulating electrode). Right, experimental schematic. (C) Average change in the density of filopodia (>2.75 μm) in response to tetanic stimulation [gray arrowhead, time (t) = 0] at test sites (open circles) and control sites (solid circles) (n = 14). To compute the time course, the density of filopodia just before tetanus was subtracted from individual experiments, and data falling in fixed time bins were averaged. To compute fractional changes, the time course was then normalized to the average density before tetanus. Horizontal error bars represent the standard deviation of the time points in each bin; vertical error bars represent the standard errors of the measurements.

An important property expected for cellular mechanisms underlying circuit formation during development, as well as information storage in the adult brain, is that they should be restricted to active synapses (20). In our experiments, this would correspond to synapses on dendritic regions close to the stimulating electrode. To test this type of synapse specificity we compared tetanus-induced growth at dendritic sites close to the stimulating electrode (test sites) with dendritic sites at least 100 μm away (control sites). Test sites showed significant growth (Fig. 2C, open circles,n = 14), whereas control sites remained unchanged (Fig. 2C, solid circles).

To determine whether growth is stimulated by synaptic transmission rather than by direct depolarization of the postsynaptic membrane by the current from the stimulating electrode, we usedd,l(-)-2-amino-5-phosphonopentanoic acid (APV), a specific antagonist for NMDA receptors. Slices were superfused with artificial cerebrospinal fluid (ACSF) containing 100 μM APV while baseline images were collected at test and control sites (n = 8). Simple measures of baseline morphological dynamics, such as the rate of formation and retraction of filopodia, were not affected by the presence of APV over the time scale of the experiment (Fig. 3A) (rates of sprouting and retraction were as follows: with APV, 2.5 ± 2.3% min–1; without APV, 2.3 ± 2.0% min–1). Delivery of a tetanus in the presence of APV showed that the drug abolished tetanus-evoked morphogenesis, with no significant changes at either the test (Fig. 3, B and C) or control (Fig. 3B) sites. APV was then washed out for >30 min and a second tetanus was delivered, which produced significant growth at the test site (Fig. 3, B and C). The density of filopodia increased significantly when compared to test sites in APV (P < 0.02, WMPSR test) and to control sites (P < 0.02). Synaptically evoked growth after APV washout was rapid, with significant changes in the number of filopodia occurring only 3 min after tetanus (Fig. 3C). This is in contrast to the experiments in Fig. 2, where a tetanus, without a prior tetanus in APV, produced delayed growth. The reason for this difference is unclear but may reflect some form of priming (21) produced by tetanus in APV.

Figure 3

Dendritic morphogenesis requires synaptic activation of NMDA receptors. (A) Time-lapse sequence showing sprouting (white solid arrowheads), retracting (open arrowheads), and changing (grey solid arrowheads) protrusions in the presence (top row) and after washout (bottom row) of APV. Time stamps in (A) and (B) are in minutes. (B) 1: Time course of the number of filopodia in response to tetanic stimulation (gray arrows) in the presence of APV (gray bar) and after APV washout (in a slice 8 days in culture). Tetanus in APV produced little change in the number of filopodia. After washout of APV, a subsequent tetanus produced new filopodia. 2: Image of a small side branch before and after tetanus after washout; note the appearance of new filopodia (white arrow). (C) Summary data for APV experiments (n = 8) expressed as changes in density of filopodia. Error bars, same as inFig. 2C. (D) Example of a time course of the number of filopodia after low-frequency stimulation (10 Hz, black arrow) and a subsequent high-frequency stimulus (100 Hz, gray arrow).

If strong synaptic NMDA receptor activation is required to trigger morphogenesis, then low-frequency stimulation, which leads to much smaller NMDA receptor–mediated postsynaptic currents (22), might not produce changes in dendritic morphology. Delivery of 2× 100 stimuli at 10 Hz failed to produce changes in filopodia density that were distinguishable from the baseline (P = 0.22,n = 5). Subsequent delivery of 100 stimuli at 100 Hz produced a robust increase in the number of filopodia (Fig. 3C).

We examined the morphological fates of filopodia induced by tetanic stimulation. In the absence of additional evoked activity, these structures persisted (Fig. 4, A and B). However, they did not continue to increase in length (Fig. 4C): The net growth decreased to baseline levels ∼45 min after tetanus. Furthermore, 27% (19 out of 71) of the new filopodia developed a bulbous head within 1 hour after the stimulus (Fig. 4B), which suggests that filopodia might mature to become spines.

Figure 4

Time course of the development of dendritic filopodia. (A) Example of a time course of the number of filopodia in response to tetanic stimulation at the test site (gray arrow; 100 Hz for 1 s, twice). (B) Images corresponding to the time course in (A). Time stamps are in minutes. Tetanic stimulation produced growth of several filopodia (compare 28 min and 56 min, solid white arrowhead) that later developed pronounced bulbous heads (101 min, gray arrowhead). (C) Time course of length changes per filopodium between consecutive observations separated by ∼10 min (Δlength/time). Growth (crosses, net change; solid circles, change in growing filopodia; open circles, change in retracting filopodia) was delayed by ∼5 min and peaked ∼15 min after tetanus (t = 0, gray arrow) and decayed to baseline within ∼40 min.

Our results show that synaptic activity can produce rapid growth in postsynaptic dendrites (23). Growth was input specific, occurring only close to activated parts of the dendrite, and required synaptic NMDA receptor activation. This activity-induced growth was most prominently expressed as an increase in long, thin, filopodia-like protrusions. Such structures often make synaptic contacts in vivo and in vitro and have been proposed as precursors of mature spines (7, 9, 11). Consistent with this hypothesis is our observation that new filopodia often develop bulbous heads, a morphological sign of mature spines (9,10, 24, 25) (Fig. 4C). Definite proof that these synaptic activity–induced filopodia make synapses will have to await ultrastructural studies using serial-section EM (24) and intracellular measurement of synaptic transmission. If these structures generate synapses, they will have greater likelihood of connecting with presynaptic axons that were active during the synaptic stimulus, providing a mechanism for synaptic plasticity satisfying Hebbian rules. Such a mechanism could play a role in the establishment of functional neural circuits during development and memory storage (1, 2, 20).


View Abstract

Stay Connected to Science

Navigate This Article