ca2+ imaging of mouse neocortical interneurone -
TRANSCRIPT
Jou
rnal
of P
hysi
olog
y
Neurones in the mammalian CNS are characterized by an
exuberant diversity of dendritic morphologies (Ramón y
Cajal, 1904). Dendrites were thought for decades to be
passive cables, yet it has become clear than many
mammalian neurones have dendrites with active
conductances and rich intrinsic electrophysiological
properties (Johnston et al. 1996; Yuste and Tank, 1996;
Llinás, 1988; Stuart & Sakmann, 1994). In particular, in
pyramidal cells, electrophysiological and imaging studies
have demonstrated the existence of backpropagating
sodium-based action potentials (APs), which can quickly
propagate through large territories of the dendritic tree
and trigger essentially instantaneous calcium accumul-
ations in spines and dendritic shafts (Stuart & Sakmann,
1994; Yuste & Denk, 1995). In addition, local dendritic
spikes, mediated by sodium or calcium channels or by
regenerative activation of NMDA receptors (NMDARs),
can activate restricted regions of the dendritic tree and
trigger more localized calcium accumulations (Pockberger,
1991; Amitai et al. 1993; Yuste et al. 1994; Schiller et al.1997; Schiller et al. 2000). These different types of dendritic
spiking have been implicated in the implementation of
synaptic learning rules (Magee & Johnston, 1997;
Markram et al. 1997) and in the temporal firing patterns of
the cell (Larkum et al. 2001).
GABAergic cells are thought to play an essential role in
controlling the excitability and spike timing in cortical
networks (Somogyi et al. 1998; Pouille & Scanziani, 2001).
Although they have prominent dendritic trees with a large
diversity of morphologies, their dendritic physiology is
relatively unexplored. An indication that the dendrites of
GABAergic cells are endowed with spiking properties
came from modelling studies to explain the paradoxical
activation of interneurones by single release site EPSPs
(Gulyas et al. 1993; Traub & Miles, 1995). Two recent
studies have demonstrated that one class of hippocampal
interneurone and a potentially homologue neocortical cell
type also have active dendrites, although it is still unclear if
other classes of interneurone behave similarly. Specifically,
dendritic recordings from oriens-alveus interneurones in
the hippocampus have established that these cells exhibit
dendritic APs that are mediated by sodium channels and
can backpropagate to the dendritic tree (Martina et al.2000). In addition, bitufted, somatostatin-positive inter-
neurones in layer 2/3 from the rat neocortex also have
backpropagating dendritic APs, which cause EPSP
depression via dendritic calcium accumulations (Zilberter,
2000; Kaiser et al. 2001). These calcium accumulations
were reported to be smaller than those measured in
pyramidal neurones, perhaps due to the larger calcium-
Ca2+ imaging of mouse neocortical interneurone dendrites:Ia-type K+ channels control action potential backpropagationJesse H. Goldberg, Gabor Tamas* and Rafael Yuste
Department of Biological Sciences, Columbia University, New York, NY 10027, USA and *Department of Comparative Physiology, University ofSzeged, Szeged, Hungary H-6726
GABAergic interneurones are essential in cortical processing, yet the functional properties of their
dendrites are still poorly understood. In this first study, we combined two-photon calcium imaging
with whole-cell recording and anatomical reconstructions to examine the calcium dynamics during
action potential (AP) backpropagation in three types of V1 supragranular interneurones:
parvalbumin-positive fast spikers (FS), calretinin-positive irregular spikers (IS), and adapting cells
(AD). Somatically generated APs actively backpropagated into the dendritic tree and evoked
instantaneous calcium accumulations. Although voltage-gated calcium channels were expressed
throughout the dendritic arbor, calcium signals during backpropagation of both single APs and AP
trains were restricted to proximal dendrites. This spatial control of AP backpropagation was
mediated by Ia-type potassium currents and could be mitigated by by previous synaptic activity.
Further, we observed supralinear summation of calcium signals in synaptically activated dendritic
compartments. Together, these findings indicate that in interneurons, dendritic AP propagation is
synaptically regulated. We propose that interneurones have a perisomatic and a distal dendritic
functional compartment, with different integrative functions.
(Received 7 March 2003; accepted after revision 8 May 2003; first published online 4 July 2003)
Corresponding author J. H. Goldberg: Department of Biological Sciences, Columbia University, 1212 Amsterdam Avenue, Box2435, New York, NY 10027, USA. Email: [email protected]
J Physiol (2003), 551.1, pp. 49–65 DOI: 10.1113/jphysiol.2003.042580
© The Physiological Society 2003 www.jphysiol.org
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
buffering capacity of interneurones (Lee et al. 2000a;
Kaiser et al. 2001).
We have used two-photon calcium imaging to
systematically explore the phenomenology and mechanisms
underlying calcium accumulations in different types of
supragranular V1 neocortical interneurones. We focused
on two groups (multipolar, parvalbumin-positive fast
spikers (FS) and bipolar, calretinin-positive irregular spikers
(IS)) based on their morphology, intrinsic electro-
physiology, and immunocytochemistry. In addition, we
include data from a third, heterogeneous, group of
interneurones which we term adapting (AD), due to their
spike frequency adaptation during depolarizing current
injections (Dantzker & Callaway, 2000).
We found that APs required sodium channels to
backpropagate and produced calcium accumulations
mediated by voltage-gated calcium channels (VGCCs).
We observed that VGCCs were expressed throughout the
dendritic tree, and that calcium signals during back-
propagating APs were proximally restricted by potassium
currents. In addition, we found that calcium influx due to
dendritic AP invasion was enhanced specifically in
synaptically activated dendritic compartments.
METHODS Slice preparation and electrophysiologyExperiments were carried out in accordance with the NIH Guidefor the Care and Use of Laboratory Animals (NIH publication no.86–23, revised 1987) and with the Society for Neuroscience 1995Statement (http://www.jneurosci.org/misc/itoa.shtml). Coronalslices of primary visual cortex were made from P13–17 C57BL/6mice. Animals were anaesthetized with ketamine–xylazine (50and 10 mg kg_1). After decapitation, brains were rapidly removedand transferred into ice-cold cutting solution containing (mM):222 sucrose, 27 NaHCO3, 2.5 KCl, 1.5 NaH2PO4, bubbled with95 % O2–5 % CO2 to pH 7.4. Brains were cooled for at least 2 minand 300-mm-thick slices were prepared with a Vibratome(VT1000, Leitz, Germany). Slices were then transferred to aheated solution (35 °C) containing (mM): 126 NaCl, 3 KCl, 1.1NaH2PO4, 26 NaHCO3, 1 CaCl2, 3 MgSO4, bubbled with 95 %O2_5 % CO2 to pH 7.4, which cooled down in the next 30 min toroom temperature. Slices were transferred to the imagingchamber 1–7 h after cutting. Artificial cerebral spinal fluid(ACSF) during experiments contained (mM): 126 NaCl, 3 KCl, 1.1NaH2PO4, 26 NaHCO3, 3 CaCl2, 1 Mg2SO4, bubbled with 95 %O2–5 % CO2 to pH 7.4. All experiments were performed at 37 °C.Whole-cell recordings from non-pyramidal cells in layer 2/3 wereobtained with a patch-clamp amplifier (Axoclamp 2B, AxonInstruments, Foster City, CA, USA, or BVC-700, Dagan Corp.,Minneapolis, MN, USA). Mechanisms of backpropagation wereexplored with several drugs (Sigma), including CPA (50 mM),NiCl2 (1 mM), TTX (1 mM), TEA (24 mM), 4-AP (1 mM), andDl-APV (100–200mM). 6-Cyano-7-nitroquinoxaline-2,3-dione(CNQX) (100 mM) was washed in during some 4-AP experimentsto prevent background synaptic activity, and Trolox (100 mM,Aldrich) was used in some Fluo-4 experiments to reducephototoxicity. Neurones were stimulated synaptically using anextracellular pipette filled with 200 mM Alexa-488 dextran
(Molecular Probes, Eugene, OR, USA) in ACSF. Tips ofstimulation pipettes were bent by about 70 deg with a microforge(Narishige, Japan). This allowed positioning the stimulationpipette perpendicular to the slice surface. In order to achieve localsubthreshold stimulation it was necessary to place glass electrodesin the immediate vicinity (< 15 mm) of the dendrite of interest,and use small amplitude (5–20 mA or 1 V), and short duration(100 ms) single shocks.
Two-photon imagingCells were filled via patch pipette with 200 mM CaGreen-1 or400 mM Fluo-4 (Molecular Probes). Pipette solution contained(mM): 130 KMeO4 , 5 KCl, 5 NaCl, 10 Hepes, 2.5 Mg-ATP, 0.3GTP, 0.2 CaGreen-1 (or 0.4 Fluo-4), and 0.03 % biocytin and wastitrated to pH 7.3. Following break-in, we waited for 30 minbefore imaging to ensure that dendrites filled with indicator.Imaging was done using a custom-made two-photon laser scanningmicroscope, consisting of a modified Fluoview (Olympus,Melville, NY, USA) confocal microscope with a Ti:sapphire laserproviding 130 fs pulses at 75 MHz (Mira, Coherent, Santa Clara,CA, USA), and pumped by a solid-state source (Verdi, Coherent).A 60 w, 0.9 NA water immersion objective (IR1, Olympus) wasused. Fluorescence was detected with photo-multiplier tubes(HC125-02, Hamamatsu, Hamamatsu City, Japan) in externalwhole-area detection mode, and images were acquired andanalysed with Fluoview (Olympus) software. Images of dendriteswere acquired at 10 w digital zoom, resulting in a nominal spatialresolution of 30 pixels mm_1 and at a time resolution of 12.64 msper point (79 Hz) in line-scan mode.
AnalysisFluorescence levels of calcium measurements were analysed usingFluoview (Olympus) and ImageJ (NIH, Bethesda, MD, USA).Time courses were analysed using Igor (Wavemetrics, LakeOswego, OR, USA). Calcium signals during AP generation weredetected in line-scan mode and were corrected for backgroundfluorescence by measuring a non-fluorescent area close to thedendrite. The relative change of fluorescence of baseline (from400 ms prior to AP generation) (DF/F) was used as an indicatorfor the change in calcium. Between 5 and 15 line scans weretypically averaged to generate DF/F transients during APs. Decaykinetics were fitted using single exponential fitting algorithms ofIgor. Unless mentioned, two-sided Student t tests were used, anddata are presented as mean ± standard deviation (S.D.). Distancesfrom the soma were measured from the site of dendritic imagingto the location where the parent dendrite emerged from the soma.AP repolarization in 1 mM TEA experiments was measured as thetime from initial resting potential to return to resting potentialafter a single AP. Calcium transients in Fig. 4 were filtered with asliding Hanning kernel.
HistologyVisualization of biocytin was performed as described (Buhl et al.1994; Tamas et al. 1997). Three-dimensional light microscopicreconstructions were carried out using Neurolucida and NeuroExplorer (MicroBrightfield, Colchester, VT, USA) with a 100 w oilobjective. Monoclonal antibodies to parvalbumin (Swant,Bellinzona, Switzerland, diluted 1:2000) and calretinin (Swant,1:1000) were applied to characterize interneurones. Dualfluorescence labelling of cortical slices was carried out as described(Reyes et al. 1998; Tamas et al. 2000), using Alexa488-conjugatedstreptavidin (Molecular Probes) revealing biocytin and CY3-conjugated anti-mouse IgG (Jackson Labs, West Grove, PA, USA)for parvalbumin and calretinin.
J. H. Goldberg, G. Tamas and R. Yuste50 J Physiol 551.1
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
RESULTSDifferent classes of interneurones in layer 2/3 frommouse primary visual cortexTo characterize the intrinsic calcium dynamics of dendrites
from neocortical interneurones we performed combined
imaging–electrophysiological–anatomical experiments with
100 layer 2/3 non-pyramidal neurones in slices from
mouse primary visual cortex. Somata without detectable
apical dendrites were targeted using differential
interference contrast (DIC) for whole-cell recordings,
filled with the high-affinity calcium indicator calcium
green (200 mM) or Fluo-4 (400 mM), imaged with a
custom-made laser scanning two-photon microscope and,
in some cases, also reconstructed anatomically.
We characterized two groups of interneurones (Fig. 1): fast
spiking cells (FS, n = 41) and irregularly spiking cells (IS,
n = 39). FS cells (Connors & Gutnick, 1990) were
characterized by high-frequency non-adapting spike
trains during sustained current injection, multipolar
dendritic morphologies with dense axonal arborizations
generally restricted to layer 2/3, and parvalbumin
immunoreactivity (n = 12/13) (Fig. 1A). IS cells (Cauli etal. 1997) fired an initial burst of APs often followed
by single spikes at an irregular frequency, had
characteristically bipolar dendritic morphologies with a
narrow columnar axonal arbor that could reach layer 6,
and were preferentially immunoreactive for calretinin
(n = 7/10) (Fig. 1B). Finally, we encountered a third,
heterogeneous group of interneurones with adapting
firing patterns, characteristically different from the FS
and IS neurones (AD, n = 20) (Fig. 1C). Some AD cells
had bitufted dendritic arbors (n = 8) and expressed
somatostatin (n = 1/2) (data not shown) (Reyes et al. 1998;
Kaiser et al. 2001), whereas other AD cells, of variable
morphology, had regular spiking electrophysiological
characteristics (n = 9) (Szabadics et al. 2001).
Action potentials triggered dendritic calciumaccumulations in interneuronesTo explore the dendritic expression of voltage-gated
channels and to characterize the intrinsic calcium
dynamics we used APs to trigger stereotyped and
reproducible calcium accumulations (Majewska et al.2000). In these experiments we stimulated neurones with
single APs or trains of 10 APs (40 Hz), evoked with
depolarizing current steps injected in the soma while we
imaged different regions of the dendritic tree using line
scans with 79 Hz resolution (Fig. 2).
We found similar dendritic calcium accumulations in all
three types of interneurones, which were significantly
different from those found in neighbouring pyramidal
neurones (Fig. 2D–F). With 200 mM calcium green as the
indicator, in all FS, IS and AD cells, single APs caused
barely detectable calcium accumulations even in proximal
dendrites, at distances of < 50 mm from the soma
(Fig. 2C–F, left panels). However, trains of APs at
20–100 Hz reliably caused a time-locked calcium
accumulation which depended on the number of APs
fired. To investigate calcium accumulations during back-
propagation systematically, we chose a physiologically
relevant stimulation protocol of 10 APs at 40 Hz (Csicsvari
et al. 1998; Swadlow et al. 1998). We compared peak
calcium signals during 10 APs (40 Hz) in proximal
dendrites (< 50 mm), and did not observe significant class-
specific differences (Fig. 2D–F, right panels; for FS:
48 ± 35 % DF/F, n = 31; for IS: 47 ± 39 % DF/F, n = 26;
for AD: 78 ± 43 % DF/F, n = 15; mean ± S.D. for all
measurements). However, we observed that across classes
interneurones with < 50 % DF/F peak amplitude had
significantly longer decay time constant (t) values than
those with > 100 % DF/F peak amplitude (1500 ± 850 and
860 ± 370 ms, respectively P < 0.01). We assume that
indicator completely washed into the cell in the 30 min we
waited prior to imaging; thus, these results were consistent
with a high interneurone-group non-specific variability in
endogenous buffer capacities (Lee et al. 2000a).
Calcium signals during AP backpropagation in inter-
neurones were significantly smaller and slower than in
pyramidal cells. Peak amplitudes of calcium signals
(%DF/F) in proximal ( < 50 mm) dendrites of inter-
neurones during 10 APs were comparable to those found
in proximal apical dendrites from pyramidal cells during a
single AP (Fig. 2C, 48 ± 27 % DF/F, n = 45 (K. Holthoff &
R. Yuste, unpublished observations). In addition, decay
time constants of the calcium accumulations in pyramidal
cells (430 ± 240 ms; n = 42) were faster than those of FS
(1050 ± 650 ms; n = 31, P < 0.001), IS (1490 ± 810 ms;
n = 26, P < 0.001) and AD cells (1170 ± 550 ms, n = 15,
P < 0.001). Since higher buffer capacities decrease the
amplitude and prolong the decays of fluorescent calcium
transients (Helmchen, 1999), these results were consistent
with studies demonstrating that interneurones have
higher endogenous buffer capacities than pyramidal
neurones (Lee et al. 2000a; Kaiser et al. 2001).
Distinct populations of cortical interneurones express a
diversity of calcium-binding proteins in a cell-type-
specific fashion (DeFelipe, 1993). FS cells expressed
parvalbumin (Fig. 1A) (Kawaguchi & Kubota, 1993),
bipolar IS cells expressed calretinin (Fig. 1B) (Cauli et al.1997), and AD cells were a heterogeneous class with
potentially different calcium buffers. Because kinetically
distinct buffers are predicted to differentially shape dendritic
calcium dynamics during AP backpropagation (Markram
et al. 1998), we wondered whether there were kinetic
differences in the FS, IS or AD calcium transients. At all
distances from the soma, FS cells had faster decay kinetics
than IS cells, although this trend was only significant at
intermediate distances (P = 0.007, Mann–Whitney U-test;
Backpropagation in interneuronesJ Physiol 551.1 51
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
yJ. H. Goldberg, G. Tamas and R. Yuste52 J Physiol 551.1
Figure 1. Morphology and intrinsic electrophysiology of different types of interneuronesAa, fast spiking (FS) firing pattern in response to 800 ms depolarizing (above) and hyperpolarizing (below)current injections. Ab, representative FS cell morphology with multipolar dendritic arbor (blue) and localaxonal collaterals (red). Ac, parvalbumin (PV) immunopositivity of an FS cell with firing pattern andmorphology as shown above. The red PV immunostained cell in the left panel and the green biocytin-filledcell in the right panel indicated by arrows are the same cell. Ba, IS firing pattern, same regime as in Aa.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
Fig. 3D). This may have been due to the presence of
parvalbumin in these cells, which has been shown to
accelerate the initial component of the decay phase due to
its high affinity for but slow binding to calcium (Chard etal. 1993; Lee et al. 2000b). It is important to note that the
intracellular environment during whole-cell recording is
highly dialysed, suggesting that under more physiological
conditions, the impact of mobile buffers such as
parvalbumin and calretinin on calcium kinetics could be
more profound.
Calcium accumulations induced by AP trains wererestricted from the distal dendritic treeThe spatial extent of calcium accumulations induced by a
train of APs was not uniform along the dendritic tree. In all
cell types, (FS, n = 30; IS, n = 23; AD, n = 13), the peak
amplitude of the accumulations was reduced at distal
dendritic sites (Fig. 3A). Across cell types, the average
amplitude (DF/F %) at < 50 mm from the soma was
61 ± 37 (n = 49), at 50–100 mm from the soma was
55 ± 40 (n = 30, P = 0.426; Fig. 3B), whereas at > 100 mm
from the soma it was 19 ± 21 (n = 20 all cells, P < 0.001
compared to proximal measurement; Fig. 3B). There were
no significant differences between FS and IS cells at either
proximal, intermediate or distal dendritic positions
(proximal: 56 ± 31, n = 25 FS; 60 ± 42, n = 16 IS, P = 0.71;
intermediate: 56 ± 41, n = 15 FS; 42 ± 36, n = 12 IS,
P = 0.27; distal: 19 ± 24, n = 10 FS; 19 ± 19, n = 8 IS,
P = 0.94), but at intermediate positions, AD cells had
higher calcium accumulations than the other two cell groups
(proximal: 85 ± 41, n = 7 P = 0.425 vs. FS, P = 0.224 vs. IS;
intermediate: 100 ± 20, n = 4 P = 0.044 vs. FS, P = 0.010
vs. IS; distal: 21 ± 20 n = 4, P = 0.89 vs. FS, P = 0.84 vs. IS,
n = 4).
Mechanisms of calcium influx and efflux duringaction potential backpropagationWhy was there a limited spatial spread of the AP-induced
calcium accumulations in interneurone dendrites?
Although multicompartamental models suggest that the
passive cable properties of interneurones are well suited
for efficient AP backpropagation (Vetter et al. 2001b), a
non-uniform distribution of dendritic conductances or
buffer capacity could greatly influence the extent of AP
backpropagation or subsequent AP-triggered calcium
accumulations, respectively. We therefore considered the
following hypotheses: (1) distal dendrites had a higher
endogenous buffer capacity, (2) the AP train did not
invade distal dendrites, or (3) the AP train faithfully
invaded the distal dendritic tree but no calcium
Backpropagation in interneuronesJ Physiol 551.1 53
Bb, representative IS morphology, with bipolar dendritic organization (blue). Basal dendrites tended to bemore branched than apical, especially in lower layers, and axonal collaterals (red) were vertically distributed.Bc, calretinin immunopositivity of an IS cell with firing pattern and morphology as shown above. Filled(right) and labelled (left) cell indicated by arrows. Ca, firing pattern and Cb, light microscopic reconstructionof an adapting (AD) cell.
Figure 2. AP-induced calcium accumulations in interneurone dendrites were slower andsmaller than in pyramidal cellsA, projected two-photon z-scan of the basal dendritic tree of an FS cell, pia top. B, protocol used. Left: singleAP; right: train of ten APs (40 Hz). C, pyramidal cell dendritic calcium accumulations during both protocols.D–F, FS, IS and AD responses. Time constants (t) are from mono-exponential fits to decays.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
accumulations were produced, due to a lack of VGCCs
distally. For the rest of the study we focused exclusively on
interneurones of the FS and IS types, because they could be
immunohistochemically defined, and therefore represented
a more homogenous group, and because they could be
reliably targeted under DIC.
If distal dendritic domains were targeted with higher
endogenous buffer capacity, we would expect to see at
distal sites a prolonged decay as well as a reduction of peak
amplitude of the calcium transient. Decays of calcium
transients did not change significantly over distance from
soma (Fig. 3C), suggesting that the reduction of peak
signal was either due to absence of voltage-gated calcium
channels distally, or to failure of the AP train to
regeneratively propagate to distal sites.
One explanation for the poor calcium signals into the
distal dendrites could be that APs were passively
propagating along dendrites devoid of sodium channels.
We measured calcium accumulations in the presence of
the voltage-gated sodium channel blocker, TTX (1 mM)
using a train of brief (3 ms), large-amplitude (100 mV)
depolarizing currents to simulate APs (Fig. 4B). To avoid
underestimating the extent of passive AP propagation, we
simulated APs 3–6 times wider than the normal APs in
these cells. Still, calcium influx was reduced even in the
proximal 50 mm of the dendritic tree (27 ± 9 % from
control, n = 3 FS, P < 0.05; 41 ± 15 % from control, n = 3
IS, P = 0.11 Fig. 4B), and failed at distances greater than
50 mm (4 ± 5 % from control, n = 3 FS, P < 0.05; 13 ± 19 %
from control, n = 3 IS, P < 0.05). We concluded that, since
passive propagation alone could initiate calcium influx
only very proximally (< 50 mm) and at reduced amplitudes,
sodium channels were expressed on the dendrites of both
FS and IS cells.
To confirm that the AP-induced dendritic calcium influx
was due to the opening of voltage-gated calcium channels,
we applied nickel at a high concentration (1 mM) to block
both high- and low-voltage-activated calcium channels. In
both FS and IS cells, practically all calcium accumulations
were blocked by Ni2+ (Fig. 4C; 12 ± 2 %, n = 2 FS;
15 ± 4 %, n = 2 IS; P < 0.001, all cells) without any
significant effect on AP physiology (not shown). Calcium
influx through VGCCs can initiate further calcium release
from internal stores (Nakamura et al. 1999), and we tested
this possibility by depleting internal calcium stores with
the SERCA (smooth endoplasmic reticulum calcium
ATPase)-pump antagonist cyclopiazonic acid (CPA)
(Kovalchuk et al. 2000). CPA (30–50 mM) did not change
the amplitude of the calcium transients significantly
(Fig. 4D; 79 ± 21 %, n = 4 FS; 70 ± 14 %, n = 6 IS of
control), but prolonged the decay time constants of
calcium transients (172 ± 50 % n = 4 FS; 240 ± 140 % in
CPA, n = 6 IS), confirming wash-in of drug, and
indicating that SERCA pumps were involved in calcium
clearance.
We conclude that sodium-based APs actively back-
propagated into the dendritic tree and caused calcium
influxes via activation of voltage-gated calcium channels.
These calcium accumulations were then cleared in part by
SERCA pumps, into intracellular calcium stores.
Existence of VGCCs throughout the interneuronedendritic treeSince the calcium influxes we measured during AP
backpropagation were due to opening of VGCCs, it
remained possible that AP trains successfully invaded
distal dendrites but failed to elicit calcium accumulations
due to an absence of calcium channels distally. We thus
tested if VGCCs were expressed on distal dendrites by
J. H. Goldberg, G. Tamas and R. Yuste54 J Physiol 551.1
Figure 3. Calcium influx during backpropagation of AP trains was proximally restrictedA, calcium transients during 10 APs (40 Hz) from IS (1), FS (•) and AD (8) cell types, plotted againstdistance from the soma. Each line represents signals from a single cell imaged at different distances from thesoma. B, data were pooled into three compartments: proximal, intermediate, and distal, and comparedbetween FS (4), IS (5) and AD (Æ). * P < 0.05 on two-tailed Student’s t test, distal signals versus proximalfor each cell type. C, time constants (t) of mono-exponential fits of calcium decays, plotted versus distancefrom soma as in B.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
somatically injecting sustained (250 ms) depolarizing
currents in the presence of high concentrations of the
potassium channel blocker TEA (24 mM) and the sodium
channel antagonist TTX (1 mM). Because calcium
diffusion is relatively slow (Allbritton et al. 1992; Neher &
Augustine, 1992), imaging calcium accumulations in
dendrites under TEA/TTX in response to somatic
depolarizations can reveal regions of the dendritic tree that
have functional VGCCs (Yuste et al. 1994).
In all tested cells (n = 4 FS and 6 IS), sustained somatic
current depolarizations in TEA/TTX gave rise to
immediate calcium transients throughout the dendritic
tree which were significantly larger than those produced
by trains of backpropagating APs (Fig. 5). Importantly,
sustained depolarizations caused large transients even at
distal dendrites where AP trains had previously failed to
produce detectable signals (Fig. 5). Although there were no
systematic dendritic distance-dependent trends in peak
signal amplitude during sustained depolarizations, peak
signals in the soma were significantly smaller. We attribute
this to the low surface-to-volume ratio in the soma.
Interestingly, plateau potentials were triggered by these
depolarizations and closely resembled those previously
observed in pyramidal neurones under TTX/TEA (arrow,
Fig. 5C; Reuveni et al. 1993; Yuste et al. 1994).
We conclude that our inability to detect calcium signals in
distal dendrites during AP trains was not due to an absence
of VGCCs in distal dendrites. Rather, these data suggested
that AP trains did not invade distal dendrites.
Potassium channels controlled calcium influxduring backpropagating AP trainsThe failure of AP trains to invade distal dendritic
compartments could be explained by several different
scenarios: (1) a lack of functional sodium channels in
distal dendrites, (2) a high density of potassium currents in
distal dendrites, or (3) slow sodium channel inactivation
developing during the train and disproportionately
affecting the distal compartment (Mickus et al. 1999).
To test if potassium currents controlled AP propagation in
interneurone dendrites, we measured AP-induced
calcium accumulations in FS and IS cells in the presence of
1 mM 4-AP, a concentration that is relatively specific for Ia
and Kv3-type potassium channels (Kirsch & Drewe, 1993).
4-AP had a powerful effect on calcium accumulations
during backpropagating AP trains and preferentially
enhanced signals at distal sites (Fig. 6). In proximal
(< 50 mm) and intermediate (50–100 mm) dendritic
regions, no significant increases were observed. However,
at distal sites, addition of 4-AP endowed unresponsive
distal dendritic segments with prominent AP-initiated
Ca2+ events (Fig. 6; control/4-AP peak response ratios (%)
were: proximal, 97 ± 21, P = 0.76, n = 7; intermediate,
77 ± 36, P = 0.23, n = 6; distal, 18 ± 18, P < 0.0005, n = 5).
Backpropagation in interneuronesJ Physiol 551.1 55
Figure 4. Mechanism of backpropagation-initiatedcalcium transientsA, percentage of control (dashed line at 100 %) DF/F signal afteraddition of TTX (1 mM), nickel (1 mM) (55 mm from soma) or CPA(50 mM) (35 mm from soma). FS, filled bars; IS, open bars.* P < 0.05. B–D, effect of drug addition (light trace) on controlDF/F signal (dark trace) during 10 APs (40 Hz). Examples from FScells are shown on the left and IS cells on the right. B, light traces arecalcium response to ten 3-ms-wide simulated APs in the presenceof TTX. Left, FS cell at 25 mm from the soma, above, and 60 mmfrom the soma along the same dendrite, below. Right, IS cell 20 mmfrom the soma (upper) and 90 mm from the soma along the samedendrite, below. C, nickel (1 mM) blockade. E, CPA failed to blockthe signal, but prolonged decay kinetics.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
yJ. H. Goldberg, G. Tamas and R. Yuste56 J Physiol 551.1
Figure 5. Voltage-gated calcium channels wereexpressed throughout the dendritic treeA, basal dendritic arbor of a bipolar IS cell. The top halfof the soma was clipped during imaging to sample distalbasal dendrites. Lines transecting the dendrites indicatesites that were selected for line-scan imaging at anadditional 10 w digital zoom (not shown). B, a train of10 APs at 40 Hz was generated by 10 separate 3 mscurrent injections in the soma (shown at top). Calciumtransients were imaged in line-scan mode at the soma,and at three positions along the basal dendrite. C, a250 ms somatic current injection in the presence of TEA(24 mM) and TTX (1 mM) caused a plateau potential(arrow). Calcium transients were imaged at identicalsites to those in B. D, data pooled from 6 IS and 4 FScells. In each experiment, peak signals were normalizedto the control AP train signal at the soma. * P < 0.05,** P < 0.01.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
Dendritic invasion of single action potentials wasalso controlled by K+ currentsAfter observing the prominent role of potassium channels
in limiting calcium influx during backpropagation of AP
trains, we wondered whether single APs were similarly
controlled. We were able to image the calcium influx
during a single backpropagating AP by switching to the
calcium indicator Fluo-4. Because Fluo-4 undergoes a
near 100-fold increase in fluorescence on binding calcium,
it is more responsive to small calcium influxes than
calcium green. As shown in Fig. 7A, calcium influx during
a single backpropagating AP was also reduced, and often
undetectable, at distal (> 100 mm) sites of FS and IS cells
(P << 0.001 n = 8 FS; P < 0.001, n = 12 IS). We again
observed a reduction in calcium accumulations with
increasing distance from the soma during AP trains (10
at 40 Hz) (P = 0.006, n = 8 FS; P = 0.006, n = 12 IS),
although calcium accumulations imaged with Fluo-4 were
often detectable even at terminal dendrites > 170 mm
from the soma (Fig. 7B). Thus, in a separate set of
experiments under different exogenous buffer conditions,
we confirmed that calcium signals due to AP back-
propagation were spatially restricted.
In addition, we again observed that application of 1 mM
4-AP preferentially increased the distal Fluo-4 signal of
both single APs (P = 0.021, n = 6 FS; P = 0.001, n = 7 IS)
and AP trains (P = 0.007, n = 6 FS; P = 0.049, n = 6 IS;
Fig. 8). In both cell types, 4-AP application also
significantly increased the DF/F signals at intermediate
(51–100 mm) dendritic segments during single APs
(P = 0.004, n = 6 FS; P = 0.001, n = 7 IS). Calcium signals
Backpropagation in interneuronesJ Physiol 551.1 57
Figure 6. AP trains were proximally restricted bypotassium currentsA, AP train (10 APs at 40 Hz) was generated at the soma andimaged 65 mm and 120 mm from the soma of an IS cell, undercontrol conditions (thick line) and in the presence of 4-AP (thinline). B, same as in A for an FS cell. Tested sites were 70 and 140 mmfrom the soma. C, pooled data from 6 cells showing the % DF/Famplitude before (filled bars) and after (open bars) 4-AP onproximal, intermediate, and distal dendritic sites. Note that theeffect of 4-AP was only significant at distal sites, *** P < 0.0005.
Figure 7. Single APs and AP trains imaged with Fluo-4were also proximally restrictedIn a separate set of experiments using Fluo-4 as the calciumindicator, calcium signals due to single APs could be resolved.A and B, left, data presented as in Fig. 3. Each line represents signalsfrom a single cell imaged at different distances from the somaduring a single backpropagating AP (A) and during 10 APs at40 Hz (B). Right , data from the graph at the left were pooled intoproximal, intermediate and distal groups. * P < 0.005,** P < 0.001.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
yJ. H. Goldberg, G. Tamas and R. Yuste58 J Physiol 551.1
Figure 8. For legend see facing page.) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
evoked by single APs and AP trains were greatly reduced at
the distal compartment, and 1 mM 4-AP eliminated this
distance-dependent reduction. These results demonstrated
that K+ currents altered the propagation of both single APs
and AP trains into the dendritic tree and preferentially
affected the excitability and calcium dynamics of the distal
compartment.
4-AP also produced an increase in somatic AP width
(193 ± 49 % of control, n = 13 for FS; 196 ± 67 % of
control, n = 7 for IS; Fig. 8C). It is thus possible that the
effect that 4-AP had on dendritic AP-induced calcium
accumulations was due to an enhanced backpropagation
of wider spikes. However, the impact of 4-AP on dendritic
excitability appeared essential since the distal compart-
ments were primarily affected by the drug.
Ia-type potassium currents controlled APbackpropagationThe two targets of the 1 mM 4-AP, Ia- and Kv3-type
potassium currents, are both expressed in interneurones
(Zhang & McBain, 1995; Rudy & McBain, 2001). To
determine which of these potassium channel subtypes
controlled calcium influx during AP backpropagation, we
repeated experiments in the presence 1 mM TEA. This
concentration of TEA targets Kv3 channels while leaving
Ia-type channels intact (Erisiret al. 1999; Lien et al. 2002).
We found that blockade of Kv3 channels alone did not
affect calcium influx during AP backpropagation (Fig. 9).
Together with the 4-AP results, these data indicate that Ia
channels controlled AP backpropagation.
Interestingly, 1 mM TEA did not significantly increase the
half-width in either IS or FS cells (TEA/control: 99 ± 6 %,
n = 3 IS; 110 ± 11 %, n = 3 FS); however, specifically in FS
cells, 1 mM TEA slowed AP repolarization (TEA/control:
86 ± 14 %, n = 3 IS; 151 ± 9 %, n = 3 FS).
EPSP–AP coupling caused supralinear calciuminflux adjacent to activated synapsesOur results in FS and IS interneurones were reminiscent of
Ia-type potassium channel control of AP propagation in
pyramidal neurones (Hoffman et al. 1997). An important
characteristic of Ia currents is that they inactivate during
subthreshold depolarizations, such as during synaptic
activity (Migliore et al. 1999a). We wondered if in
interneurones, previous synaptic activation could affect
the invasion of APs into activated dendritic compart-
ments, and if calcium influx at synaptic sites was
modulated by backpropagating APs. To address these
issues, we somatically generated a single AP 10 ms after
evoking an EPSP with a stimulation electrode placed in the
immediate vicinity (< 15 mm) of the dendrite of interest
(see Methods and accompanying paper, Goldberg et al.2003). Importantly, we exclusively sought out dendritic
segments with orientations parallel to our line scan, and
used small stimulation intensities to activate restricted
domains along that segment (Fig. 10A). This allowed us to
quantitatively examine the interplay between EPSPs and
APs at three sites: (1) synaptic sites (where synaptic
activation alone caused calcium signals), (2) immediately
adjacent to synaptic sites but where synaptic activation
alone did not cause calcium influx (9.2 ± 2.3 mm, n = 10
dendrites in 2 FS, 4 IS and 4 AD cells) and (3) on dendrites
where no synaptic calcium signal was detected.
We reliably observed supralinear calcium signals during
EPSP–AP pairing only at sites adjacent to activated
synapses. In the experiment illustrated in Fig. 10, the apical
branch of a bipolar IS cell was activated at a local site in a
dendritic compartment. In interleaved trials, synaptic
stimulation or APs were evoked in isolation or coupled
with EPSPs preceding the AP by 10 ms. At the site of the
synaptically evoked calcium entry, this coupling did not
significantly alter the signal. However, 6 mm distal along
the branch, the pairing of the AP and EPSP resulted in a
large supralinear calcium influx (arrow, Fig. 10D). This
effect was observed in all three cell types (Fig. 10E, n = 2/2
FS, 4/4 IS, 3/4 AD).
As shown in Fig. 10E, we found that the degree of
supralinearity at sites where synaptic activation alone
caused calcium influx was highly variable in all three cell
types. While some cells exhibited modest supralinearity at
synaptic sites, suggestive of AP-mediated NMDA receptor
(NMDAR) recruitment (Yuste & Denk, 1995; Magee &
Johnston, 1997), others showed sublinearity, suggesting
Backpropagation in interneuronesJ Physiol 551.1 59
Figure 8. Single action potentials were also proximally restricted by K+ currentsA, left, XYZ-projection of a basal dendritic branch of an FS cell. Lines transecting dendrites at 10, 50 and
120 mm from the soma, indicate regions of interest where line scans were conducted at an additional 10 wdigital zoom (not shown). B, traces were recorded while eliciting single APs (left, top) or trains of 10 APs at40 Hz (right, top). Note the different % DF/F scale bars for the two stimulation regimes. Relative to control(dark traces), addition of 4-AP (1 mM, light traces) specifically increased distal signals, eliminating thedistance-dependent reduction in calcium signal for both single APs (left) and AP trains (right). C, 4-APincreased AP half-width. Top, single AP in control (dark trace), and in the presence of 1 mM 4-AP (lighttrace), generated by a 5 ms current injection at the soma, bottom. D, pooled data from FS cell group (n = 6)demonstrate the effect of 4-AP (open bars) on single APs (black bars) and AP trains (grey bars) at proximal,intermediate and distal dendritic sites. Data are normalized to the control signal at the proximal site. E, datapresented as in D for IS cell group (n = 7). * P < 0.05, ** P < 0.01.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
yJ. H. Goldberg, G. Tamas and R. Yuste60 J Physiol 551.1
Figure 9. For legend see facing page.) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
that VGCCs activated synaptically were left unavailable to
the backpropagating AP. However most fell very close to
the linear range, as shown in Fig. 10D. Although addition
of APV (100 mM) to block NMDAR-mediated calcium
entry significantly reduced synaptic calcium signals (see
accompanying paper, Goldberg et al. 2003), it did not
significantly change the distribution of paired signals
about the computed linear sum, revealing that in at least
some cases, supralinear EPSP–AP interaction was not
NMDAR-dependent.
Our failure to observe supralinearity at synaptic sites could
be due to indicator saturation during the strong synaptic
calcium signal. However, the peak amplitude of the
synaptic signal in the sublinear cases (151 ± 105 %DF/F,
n = 13 dendritic branches) was not significantly different
from that in supralinear cases (128 ± 118 %DF/F, n = 11
branches). Further, we found that the saturating calcium
influx in many of these experiments was at a DF/F around
500 % (n = 15). Thus we think it was unlikely that our
inability to reliably observe supralinearity at synaptic sites
was due to indicator saturation. Moreover, in identical
recording and indicator conditions during stronger synaptic
stimulation of clustered synapses we routinely observed
peak influxes 2–3 times greater (298 ± 101 %DF/F, n = 7).
We did not consistently observe supralinearity on
dendritic segments where we did not observe synaptic
calcium signals (ratio observed/predicted linearity =
1.64 ± 1.4, P = 0.14, n = 12). Thus signals were reliably
enhanced specifically at synaptically activated branches.
This finding suggests that previous synaptic activity is
capable of controlling the spatial dynamics of dendritic AP
invasion, and is consistent with our finding that Ia-type
potassium channels controlled calcium influx during AP
backpropagation.
Backpropagation in interneuronesJ Physiol 551.1 61
Figure 9. Ia-type potassium channels controlled AP backpropagationData are laid out as in Figure 8. A, XYZ-projection of an FS cell, pia at right, medial is at bottom. Open barsindicate regions of interest examined at an additional 10 w zoom for traces in B. B, traces were recorded whileeliciting single APs (left, top) or trains of 10 APs at 40 Hz (right, top). Note the different % DF/F scale bars forthe two stimulation regimes. Relative to control (dark traces), addition of TEA (1 mM, light traces) did notsignificantly alter calcium signals at any distance from the soma during both single APs (left) or AP trains(right). C, 1 mM TEA did not significantly change the AP half-width, but slowed repolarization in FS cells.Top, single AP in control (dark trace), and in the presence of 1 mM TEA (light trace), generated by a 4 mscurrent injection at the soma, bottom. D, pooled data from FS cell group (n = 3) demonstrate the effect of1 mM TEA (5) on single APs (4) and AP trains (Æ) at proximal, intermediate and distal dendritic sites. Dataare normalized to control signal at the proximal site. E, data presented as in D for IS cell group (n = 3).
Figure 10. EPSP–AP coupling caused supralinear calcium influxes adjacent to activatedsynapsesA, the position of the line scan, arrowheads, is indicated on an apical dendrite of an IS cell 40 mm from thesoma. The stimulation electrode, S, was placed approximately 8 mm beneath the dendrite. B–D, line scansand calcium transients (red at synaptic site, black 6 mm distal) during three stimulation protocols: synapticstimulation alone, Syn (B), single AP alone, 1 AP (C), and paired, Syn + 1 AP (D). Blue traces represent thecalculated sum of 1 AP and Syn signals. Physiology traces for each experimental protocol are at the bottom.Each line-scan image is an average of four interleaved trials. Arrows at 400 ms into the line scan, also beneatheach physiology trace, indicate the time of stimulation. Note different time scales for physiology and calciumtraces. E, summed responses during Syn + 1 AP at synaptic sites (syn), adjacent (adj), and under NMDAreceptor blockade (apv + mk-801) were normalized to the computed sum (blue line at y = 1). Data are from11 IS, 10 FS and 4 AD cells, shown individually at the left, and pooled at the right. Multiple dendriticpositions were tested on individual cells.
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
DISCUSSION Using two-photon calcium imaging we have characterized
the AP-induced calcium dynamics in dendrites from
different classes of interneurones. Our goals were to
investigate their intrinsic calcium dynamics and explore
their electrical excitability using calcium accumulations to
monitor the activation of particular dendritic regions by
single APs and AP trains. The main findings of this study
were (1) dendrites of FS, IS and AD cells actively support
AP backpropagation due the expression of voltage-gated
sodium, potassium and calcium channels, (2) calcium
signals during backpropagation of both single APs and AP
trains were reduced at distal dendritic regions due to the
activation of Ia-type potassium currents, and (3) synaptic
activation was able to enhance AP-mediated calcium
signals specifically in activated compartments. Together,
these findings indicate that interneurone dendrites are
active structures which can dynamically control signal
propagation.
Mechanisms of AP-induced calcium accumulationsin neocortical interneuronesSomatically generated APs propagated actively to the
dendritic tree and by opening VGCCs produced calcium
influx in the proximal regions. The peak accumulations
were unaffected by blockers of internal release, although
SERCA pumps were involved in dendritic calcium
clearance. Therefore both calcium influx and efflux
pathways in interneurone dendrites were the same as has
been reported in pyramidal cells (Regehr & Tank, 1994;
Yuste et al. 1994; Markram et al. 1995; Yuste & Denk, 1995;
Helmchen, 1999).
We did not encounter any appreciable differences in these
mechanisms among different types of interneurones
examined. However, calcium signals during AP back-
propagation had smaller amplitudes and slower offset
kinetics compared with pyramidal cells (Fig. 2), consistent
with a larger interneurone calcium-buffering capacity (Lee
et al. 2000a; Kaiser et al. 2001; Rozov et al. 2001).
Ia-type K+ channel control of AP backpropagation inGABAergic cellsOur data reveal that the dendrites of multiple types of
neocortical interneurones actively express voltage-gated
sodium, potassium and calcium channels. We observed
that, although VGCCs were located throughout the
dendritic tree (Fig. 5), and passive cable properties appear
ideal for backpropagation in interneurones (Vetter et al.2001a), under normal conditions the AP-induced calcium
accumulations were reduced at distal (> 100 mm)
dendritic regions (Figs 3 and 7).
Based on the large effect that 4-AP had on the spatial
pattern of AP-induced calcium accumulations (Figs 6 and
8), we conclude that AP backpropagation was controlled
by potassium channels. In pyramidal neurones, Ia-type
potassium channels appear at high densities in apical
dendrites (Hoffman et al. 1997; Korngreen & Sakmann,
2000), where they play a prominent role in regulating
dendritic AP invasion (Hoffman et al. 1997), and in fast-
spiking interneurones Kv3-type potassium channels
facilitate high-frequency firing (Martina et al. 1998; Erisir
et al. 1999; Rudy & McBain, 2001). Both targets of 1 mM
4-AP, Ia and Kv3-type K+ currents, are reportedly
expressed in interneurones, and thus may have been
involved in controlling dendritic AP propagation (Zhang
& McBain, 1995; Rudy & McBain, 2001). In order to
distinguish between these two potassium channel subtypes
implicated in the 1 mM 4-AP experiments, we specifically
blocked Kv3 potassium channels in 1 mM TEA (Erisir et al.1999; Lien et al. 2002). Since blockade of Kv3 alone did not
significantly enhance calcium signals during AP
backpropagation (Fig. 9), we conclude that Ia-type
potassium channels controlled dendritic calcium
accumulations during AP backpropagation.
We found it particularly interesting that Ia potassium
currents were so important in controlling AP propagation,
in the light of two recent studies in interneurones which
have emphasized the importance of dendritic potassium
currents in regulating synaptic activation and spike
initiation (Fricker & Miles, 2000; Galarreta & Hestrin,
2001). High dendritic Ia-potassium channel expression on
the dendrites of interneurones could thus serve the dual
function of regulating both spike initiation and
propagation.
EPSP interaction with backpropagating actionpotentialsAlthough AP backpropagation has recently been
demonstrated in a variety of interneuronal classes
(Tombaugh, 1998; Martina et al. 2000; Kaiser et al. 2001),
there has, as yet, been no description of how bAPs interact
with EPSPs in interneurones. Because of the unique
subthreshold inactivation kinetics of Ia channels, we
wondered if synaptic activity could regulate AP
backpropagation. We found that when evoked EPSPs
preceded somatically generated APs by 10 ms, calcium
influx in the compartment of the activated synapse was
supralinearly increased in some cells (Fig. 9).
There are at least two possible mechanisms for the
enhancement of AP-mediated calcium signals by
preceding EPSPs. First, EPSPs could eliminate potassium
channel control of APs by inactivating Ia, as has been
shown in CA1 pyramidal neurones (Hoffman et al. 1997;
Migliore et al. 1999a). Alternatively, EPSPs could boost
dissipating backpropagating APs by providing the
necessary dendritic depolarization to reach sodium
channel threshold (Stuart & Hausser, 2001). Given the
importance of K+ channels in controlling AP propagation
under control conditions (Figs 6 and 8), we prefer the first
explanation. However, direct dendritic recordings of these
J. H. Goldberg, G. Tamas and R. Yuste62 J Physiol 551.1
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
narrow dendrites will be necessary to clarify the precise
mechanisms of EPSP-mediated enhancement of AP-
induced calcium influxes.
It is important to note that we observed no supralinearity
in dendritic compartments where we did not observe
synaptic calcium influx. Given that the propagation of
dendritic APs through gap junctions may facilitate the
synchronization of the interneuronal syncytium (Beierlein
et al. 2000), our finding that EPSPs can control AP
propagation specifically in activated compartments
suggests that the spatial component of synaptic activity
may affect synchronization of interneuronal ensembles.
Functional compartments in GABAergic dendritesOur experiments suggest that spatially limited calcium
signalling during active backpropagation of somatic APs
defines at least two functional compartments in the
somatodendritic domain of neocortical interneurones: a
perisomatic region of the dendritic tree, where back-
propagating APs can reach, in normal conditions, and a
distal region, which may only be affected by back-
propagating APs when dendritic potassium channels are
inactivated. While proximal dendrites may undergo
calcium-dependent processes that involve the timing of
the firing of the postsynaptic cell, synaptic mechanisms of
calcium influx may dominate distal compartments. Also,
selective invasion of APs to proximal or recently activated
dendrites favours the propagation of both electrical and
calcium signals through gap junctions, which are
expressed within the networks of FS and regular spiking
non-pyramidal cells (Galarreta & Hestrin, 1999; Gibson etal. 1999; Tamas et al. 2000; Szabadics et al. 2001). In
addition, Ia potassium currents are regulated by
neurotransmitters and second messenger pathways (Hille,
1992; Migliore et al. 1999b; Atzori et al. 2000) suggesting
that the multi-compartment picture of interneurone
dendritic physiology is a dynamic one.
Differential calcium dynamics in proximal and distal
dendritic regions of interneurones might selectively
interact with inputs targeting different somatodendritic
regions. The proximal dendritic domain of cortical
interneurones is selectively innervated by glutamatergic
afferents from the thalamus (Freund et al. 1985) and by
GABAergic inputs from subcortical sources (Freund &
Meskenaite, 1992) and from local basket cells (Tamas et al.1998). Distal dendritic branches of interneurones, on the
other hand, are innervated by other local GABAergic cell
classes (Tamas et al. 1998) and by predominantly cortical
glutamatergic afferents. These two dendritic regions
correspond nicely to those identified in our study as
having different backpropagation and calcium dynamics.
It is possible that functional pairing between AP back-
propagation, postsynaptic calcium dynamics and particular
afferent pathways could dramatically increase the
computational power of individual cortical interneurones.
REFERENCESAllbritton NL, Meyer T & Stryer L (1992). Range of messenger action
of calcium ion and inositol 1,4,5-trisphosphate. Science 258,
1812–1815.
Amitai Y, Friedman A, Connors BW & Gutnick MJ (1993).
Regenerative activity in apical dendrites of pyramidal cells in
neocortex. Cereb Cortex 3, 26–38.
Atzori M, Lau D, Tansey EP, Chow A, Ozaita A, Rudy B & McBain CJ
(2000). H2 histamine receptor-phosphorylation of Kv3.2
modulates interneuron fast spiking. Nat Neurosci 3, 791–798.
Beierlein M, Gibson JR & Connors BW (2000). A network of
electrically coupled interneurons drives synchronized inhibition in
neocortex. Nat Neurosci 3, 904–910.
Buhl EH, Halasy K & Somogyi P (1994). Diverse sources of
hippocampal unitary inhibitory postsynaptic potentials and the
number of synaptic release sites. Nature 368, 823–828.
Cauli B, Audinat E, Lambolez B, Angulo MC, Ropert N, Tsuzuki K,
Hestrin S & Rossier J (1997). Molecular and physiological diversity
of cortical nonpyramidal cells. J Neurosci 17, 3894–3906.
Chard PS, Bleakman D, Christakos S, Fullmer CS & Miller RJ (1993).
Calcium buffering properties of calbindin D28k and parvalbumin
in rat sensory neurones. J Physiol 472, 341–357.
Connors BW & Gutnick MJ (1990). Intrinsic firing patterns of
diverse neocortical neurons. Trends Neurosci 13, 99–104.
Csicsvari J, Hirase H, Czurko A & Buzsaki G (1998). Reliability and
state dependence of pyramidal cell-interneuron synapses in the
hippocampus: an ensemble approach in the behaving rat. Neuron21, 179–189.
Dantzker JL & Callaway EM (2000). Laminar sources of synaptic
input to cortical inhibitory interneurons and pyramidal neurons.
Nat Neurosci 3, 701–707.
Defelipe J (1993). Neocortical neuronal diversity: chemical
heterogeneity revealed by colocalization studies of classic
neurotransmitters, neuropeptides, calcium-binding proteins, and
cell surface molecules. Cereb Cortex 3, 273–289.
Erisir A, Lau D, Rudy B & Leonard CS (1999). Function of specific
K(+) channels in sustained high-frequency firing of fast-spiking
neocortical interneurons. J Neurophysiol 82, 2476–2489.
Freund TF, Martin KA, Somogyi P & Whitteridge D (1985).
Innervation of cat visual areas 17 and 18 by physiologically
identified X- and Y-type thalamic afferents. II. Identification of
postsynaptic targets by GABA immunocytochemistry and Golgi
impregnation. J Comp Neurol 242, 275–291.
Freund TF & Meskenaite V (1992). gamma-Aminobutyric acid-
containing basal forebrain neurons innervate inhibitory
interneurons in the neocortex. Proc Natl Acad Sci U S A 89,
738–742.
Fricker D & Miles R (2000). EPSP amplification and the precision of
spike timing in hippocampal neurons. Neuron 28, 559–569.
Galarreta M & Hestrin S (1999). A network of fast-spiking cells in the
neocortex connected by electrical synapses. Nature 402, 72–75.
Galarreta M & Hestrin S (2001). Spike transmission and synchrony
detection in networks of GABAergic interneurons. Science 292,
2295–2299.
Gibson JR, Beierlein M & Connors BW (1999). Two networks of
electrically coupled inhibitory neurons in neocortex. Nature 402,
75–79.
Goldberg JH, Yuste R & Tamas G (2003). Ca2+ imaging of mouse
neocortical interneurone dendrites: contribution of Ca2+-
permeable AMPA and NMDA receptors to subthreshold Ca2+
dynamics. J Physiol 551, 67–78.
Backpropagation in interneuronesJ Physiol 551.1 63
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
Gulyas AI, Miles R, Sik A, Toth K, Tamamaki N & Freund TF (1993).
Hippocampal pyramidal cells excite inhibitory neurons through a
single release site. Nature 366, 683–687.
Helmchen F (1999). Dendrites as biochemical compartments. In
Dendrites. ed. Stuart G, Spruston N & Hausser M, pp. 161–192.
Oxford University Press, Oxford.
Hille B (1992). Ionic Channels in Excitable Membranes. Sinauer,
Sunderland, MA, USA.
Hoffman DA, Magee JC, Colbert CM & Johnston D (1997).
Potassium channel regulation of signal propagation in dendrites of
hippocampal pyramidal neurons. Nature 387, 869–875.
Johnston D, Magee JC, Colbert CM & Cristie BR (1996). Active
properties of neuronal dendrites. Annu Rev Neurosci 19, 165–186.
Kaiser KM, Zilberter Y & Sakmann B (2001). Back-propagating
action potentials mediate calcium signalling in dendrites of
bitufted interneurons in layer 2/3 of rat somatosensory cortex.
J Physiol 535, 17–31.
Kawaguchi Y & Kubota Y (1993). Correlation of physiological
subgroupings of nonpyramidal cells with parvalbumin- and
calbindinD28k-immunoreactive neurons in layer V of rat frontal
cortex. J Neurophysiol 70, 387–396.
Kirsch GE & Drewe JA (1993). Gating-dependent mechanism of
4-aminopyridine block in two related potassium channels. J GenPhysiol 102, 797–816.
Korngreen A & Sakmann B (2000). Voltage-gated K+ channels in
layer 5 neocortical pyramidal neurones from young rats: subtypes
and gradients. J Physiol 525, 621–639.
Kovalchuk Y, Eilers J, Lisman J & Konnerth A (2000). NMDA
receptor-mediated subthreshold Ca(2+) signals in spines of
hippocampal neurons. J Neurosci 20, 1791–1799.
Larkum ME, Zhu JJ & Sakmann B (2001). Dendritic mechanisms
underlying the coupling of the dendritic with the axonal action
potential initiation zone of adult rat layer 5 pyramidal neurons.
J Physiol 533, 447–466.
Lee SH, Rosenmund C, Schwaller B & Neher E (2000a). Differences
in Ca2+ buffering properties between excitatory and inhibitory
hippocampal neurons from the rat. J Physiol 525, 405–418.
Lee SH, Schwaller B & Neher E (2000b). Kinetics of Ca2+ binding to
parvalbumin in bovine chromaffin cells: implications for [Ca2+]
transients of neuronal dendrites. J Physiol 525, 419–432.
Lien CC, Martina M, Schultz JH, Ehmke H & Jonas P (2002). Gating,
modulation and subunit composition of voltage-gated K(+)
channels in dendritic inhibitory interneurones of rat
hippocampus. J Physiol 538, 405–419.
Llinas R (1998). The intrinsic electrophysiological properties of
mammalian neurons: insights into central nervous system
function. Science 242, 1654–1664.
Magee JC & Johnston D (1997). A synaptically controlled, associative
signal for hebbian plasticity in hippocampal neurons. Science 275,
209–212.
Majewska A, Brown E, Ross J & Yuste R (2000). Mechanisms of
calcium decay kinetics in hippocampal spines: role of spine
calcium pumps and calcium diffusion through the spine neck in
biochemical compartmentalization. J Neurosci 20, 1722–1734.
Markram H, Helm PJ & Sakmann B (1995). Dendritic calcium
transients evoked by single back-propagating action potentials in
rat neocortical pyradmidal neurons. J Physiol 485, 1–20.
Markram H, Luebke J, Frotscher M & Sakmann B (1997). Regulation
of synaptic efficacy by coincidence of postsynaptic APs and EPSPs.
Science 275, 213–215.
Markram H, Roth A & Helmchen F (1998). Competitive calcium
binding: Implications for dendritic calcium signaling. J ComputNeurosci 5, 331–348.
Martina M, Schultz JH, Ehmke H, Monyer H & Jonas P (1998).
Functional and molecular differences between voltage-gated K+
channels of fast-spiking interneurons and pyramidal neurons of
rat hippocampus. J Neurosci 18, 8111–8125.
Martina M, Vida I & Jonas P (2000). Distal initiation and active
propagation of action potentials in interneuron dendrites. Science287, 295–300.
Mickus T, Jung H & Spruston N (1999). Properties of slow,
cumulative sodium channel inactivation in rat hippocampal CA1
pyramidal neurons. Biophys J 76, 846–860.
Migliore M, Hoffman DA, Magee JC & Johnston D (1999a). Role of
an A-type K+ conductance in the back-propagation of action
potentials in the dendrites of hippocampal pyramidal neurons.
J Comput Neurosci 7, 5–15.
Migliore M, Hoffman DA, Magee JC & Johnston D (1999b). Role of
an A-type K+ conductance in the back-propagation of action
potentials in the dendrites of hippocampal pyramidal neurons.
J Comput Neurosci 7, 5–15.
Nakamura T, Barbara JG, Nakamura K & Ross WN (1999).
Synergistic release of Ca2+ from IP3-sensitive stores evoked by
synaptic activation of mGluRs paired with backpropagating action
potentials. Neuron 24, 727–737.
Neher E & Augustine GJ (1992). Calcium gradients and buffers in
bovine chromaffin cells. J. Physiol 450, 273–301.
Pockberger H (1991). Electrophysiological and morphological
properties of rat motor cortex neurons in vivo. Brain Res. 539,
181–190.
Pouille F & Scanziani M (2001). Enforcement of temporal fidelity in
pyramidal cells by somatic feed-forward inhibition. Science 293,
1159–1163.
Ramón y Cajal S (1904). La Textura del Sistema Nerviosa del Hombrey los Vertebrados, vol. 2. Moya, Madrid.
Regehr W & Tank D (1994). Dendritic calcium dynamics. Curr OpinNeurobiol 4, 373–382.
Reuveni I, Friedman A, Amitai Y & Gutnick MJ (1993). Stepwise
repolarization from Ca2+ plateaus in neocortical pyramidal cells:
evidence for nonhomogeneous distribution of HVA Ca2+ channels
in dendrites. J Neurosci 13, 4609–4621.
Reyes A, Lujan R, Rozov A, Burnashev N, Somogyi P & Sakmann B
(1998). Target-cell-specific facilitation and depression in
neocortical circuits. Nat Neurosci 1, 279–285.
Rozov A, Burnashev N, Sakmann B & Neher E (2001). Transmitter
release modulation by intracellular Ca2+ buffers in facilitating and
depressing nerve terminals of pyramidal cells in layer 2/3 of the rat
neocortex indicates a target cell-specific difference in presynaptic
calcium dynamics. J Physiol 531, 807–826.
Rudy B & McBain CJ (2001). Kv3 channels: voltage-gated K+
channels designed for high-frequency repetitive firing. TrendsNeurosci 24, 517–526.
Schiller J, Major G, Koester HJ & Schiller Y (2000). NMDA spikes in
basal dendrites of cortical pyramidal neurons. Nature 404,
285–289.
Schiller J, Schiller Y, Stuart G & Sakmann B (1997). Calcium action
potentials restricted to distal apical dendrites of rat neocortical
pyramidal neurons. J Physiol 505, 605–616.
Somogyi P, Tamas G, Lujan R & Buhl E (1998). Salient features of
synaptic organisation in the cerebral cortex. Brain Res Brain ResRev 26, 113–135.
Stuart GJ & Hausser M (2001). Dendritic coincidence detection of
EPSPs and action potentials. Nat Neurosci 4, 63–71.
Stuart GJ & Sakmann B (1994). Active propagation of somatic action
potentials into neocortical pyramidal cell dendrites. Nature 367,
69–72.
J. H. Goldberg, G. Tamas and R. Yuste64 J Physiol 551.1
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (
Jou
rnal
of P
hysi
olog
y
Swadlow HA, Beloozerova IN & Sirota MG (1998). Sharp, local
synchrony among putative feed-forward inhibitory interneurons
of rabbit somatosensory cortex. J Neurophysiol 79, 567–582.
Szabadics J, Lorincz A & Tamas G (2001). Beta and gamma frequency
synchronization by dendritic gabaergic synapses and gap junctions
in a network of cortical interneurons. J Neurosci 21, 5824–5831.
Tamas G, Buhl EH, Lorincz A & Somogyi P (2000). Proximally
targeted GABAergic synapses and gap junctions synchronize
cortical interneurons. Nat Neurosci 3, 366–371.
Tamas G, Buhl EH & Somogyi P (1997). Massive autaptic self-
innervation of GABAergic neurons in cat visual cortex. J Neurosci17, 6352–6364.
Tamas G, Somogyi P & Buhl EH (1998). Differentially
interconnected networks of GABAergic interneurons in the visual
cortex of the cat. J Neurosci 18, 4255–4270.
Tombaugh GC (1998). Intracellular pH buffering shapes activity-
dependent Ca2+ dynamics in dendrites of CA1 interneurons.
J Neurophysiol 80, 1702–1712.
Traub RD & Miles R (1995). Pyramidal cell-to-inhibitory cell spike
transduction explicable by active dendritic conductances in
inhibitory cell. J Comput Neurosci 2, 291–298.
Vetter P, Roth A & Hausser M (2001a). Propagation of action
potentials in dendrites depends on dendritic morphology.
J Neurophysiol 85, 926–937.
Vetter P, Roth A & Hausser M (2001b). Propagation of action
potentials in dendrites depends on dendritic morphology.
J Neurophysiol 85, 926–937.
Yuste R & Denk W (1995). Dendritic spines as basic units of synaptic
integration. Nature 375, 682–684.
Yuste R, Gutnick MJ, Saar D, Delaney KD & Tank DW (1994).
Calcium accumulations in dendrites from neocortical neurons: an
apical band and evidence for functional compartments. Neuron13, 23–43.
Yuste R & Tank DW (1996). Dendritic integration in mammalian
neurons, a century after Cajal. Neuron 16, 701–716.
Zhang L & McBain CJ (1995). Voltage-gated potassium currents in
stratum oriens-alveus inhibitory neurones of the rat CA1
hippocampus. J Physiol 488, 647–660.
Zilberter Y (2000). Dendritic release of glutamate suppresses
synaptic inhibition of pyramidal neurons in rat neocortex.
J Physiol 528, 489–496.
Acknowledgements We thank Misha Beierlein, Josh Brumberg and Jason MacLean forcomments. This study was funded by the NEI (EY11787 ), NINDS(NS40726), the New York STAR Center for High ResolutionImaging of Functional Neural Circuits and the John Merck Fund.
Backpropagation in interneuronesJ Physiol 551.1 65
) at MASS INST OF TECHNOLOGY on October 4, 2011jp.physoc.orgDownloaded from J Physiol (