Neuroscience 250 (2013) 208–221

ALTERATIONS IN MEMBRANE AND FIRING PROPERTIES OF LAYER2/3 PYRAMIDAL NEURONS FOLLOWING FOCAL LASER LESIONS INRAT VISUAL CORTEX

B. IMBROSCI * AND T. MITTMANN

Institute of Physiology and Pathophysiology, University

Medical Center of the Johannes-Gutenberg University Mainz,

D-55128 Mainz, Germany

Abstract—Focal cortical injuries are well known to cause

changes in function and excitability of the surviving cortical

areas but the cellular correlates of these physiological alter-

ations are not fully understood. In the present study we

employed a well established ex vivo–in vitro model of focal

laser lesions in the rat visual cortex and we studied mem-

brane and firing properties of the surviving layer 2/3 pyrami-

dal neurons. Patch-clamp recordings, performed in the first

week post-injury, revealed an increased input resistance, a

depolarized spike threshold as well as alterations in the fir-

ing pattern of neurons in the cortex ipsilateral to the lesion.

Notably, the reported lesion-induced alterations emerged or

became more evident when an exciting perfusing solution,

known as modified artificial cerebrospinal fluid, was used

to increase the ongoing synaptic activity in cortical slices.

Conversely, application of glutamatergic or GABAA receptor

blockers reduced the observed alterations and GABAB

receptor blockers abolished the differences completely. All

together the present findings suggest that changes in syn-

aptic receptors function, following focal cortical injuries,

can modulate membrane and firing properties of layer 2/3

pyramidal neurons. This previously unknown functional

interplay between synaptic and membrane properties may

constitute a novel cellular mechanism to explain alterations

in neuronal network function and excitability following focal

cortical injuries. � 2013 IBRO. Published by Elsevier Ltd. All

rights reserved.

Key words: cortical lesions, patch clamp recordings, input

resistance, spike threshold, mACSF, GABAB receptors.

0306-4522/13 $36.00 � 2013 IBRO. Published by Elsevier Ltd. All rights reservehttp://dx.doi.org/10.1016/j.neuroscience.2013.06.063

*Corresponding author. Address: Institute of Physiology and Patho-physiology, University Medical Center of the Johannes-GutenbergUniversity Mainz, Duesbergweg 6, D-55128 Mainz, Germany. Tel:+49-(0)6131-39-25715; fax: +49-(0)6131-39-25560.

E-mail address: imbrosci@uni-mainz.de (B. Imbrosci).Abbreviations: ADP, afterdepolarization; AHP, afterhyperpolarization;DNQX, 6,7-dinitroquinoxaline-2,3-dione; EGTA, ethylene glycoltetraacetic acid; GABAAR, GABAA receptor; GABABR, GABAB

receptor; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid;ISI, interspike interval; (m)ACSF, (modified) artificial cerebrospinalfluid; PTX, picrotoxin; resting Vm, resting membrane potential; Ri, inputresistance; sEPSCs, spontaneous EPSCs; sIPSCs, spontaneousIPSCs; SFA, spike frequency adaptation.

208

INTRODUCTION

Brain injuries are a leading cause of disability worldwide.

Traditionally, the loss of function following these

pathological events was attributed to the irreversible

damage of neuronal tissue. Accumulating lines of

evidence, however, suggest that functional alterations,

occurring in the structurally intact brain areas adjacent

to the lesion, could also contribute to the observed

neurological deficits and to processes of functional

recovery. These physiological disturbances, spreading

from the site of primary damage, include metabolic

changes (Witte et al., 2000; Alves et al., 2005) and

alterations in neuronal activity. In particular, several

electrophysiological in vivo studies revealed a period of

hyperexcitability strongly expressed in the first week

post-lesion. In a heat-lesion model in the cat visual

cortex, few days after the lesion induction, neurons

located between 1 to 2.5 mm from the border of the

injury exhibited an increased spontaneous and evoked

activity (Eysel and Schmidt-Kastner, 1991). Similarly, a

significant increase in the spontaneous firing frequency

was found in the tissue adjacent to a photochemically

induced infarction in the rat sensorimotor cortex

(Schiene et al., 1996). Furthermore, in a model of focal

cortical compression, recordings of multi-unit activity

from the rat barrel cortex showed that an increase in

neuronal responsiveness, upon whisker stimulation, was

already visible two hours after the injury onset (Ding

et al., 2011). At the cellular level an unbalance between

excitatory and inhibitory synaptic transmission has so

far been considered the most plausible explanation for

the hyperexcitability observed in vivo (Domann et al.,

1993; Mittmann et al., 1994; Neumann-Haefelin et al.,

1995; Buchkremer-Ratzmann et al., 1996; Schiene

et al., 1996; Reinecke et al., 1999; Neumann-Haefelin

and Witte, 2000). Indeed, a reduced inhibition, coupled

with an enhanced excitation, will inevitability increase

the synaptically driven depolarization and thereby the

firing activity of neurons. Although this scenario is a

quite realistic one, it may not tell the whole story.

Changes in ongoing synaptic activity are likely to

influence neuronal excitability in a more complex

fashion. For instance, several lines of evidence suggest

that background synaptic transmission can exert a

powerful influence on neuronal membrane and firing

properties (Destexhe and Pare, 1999; Chance et al.,

2002; Leger et al., 2005; Bar-Yehuda and Korngreen,

2007). Alterations in neuronal membrane properties are

d.

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 209

also expected following activation of extrasynaptic and

metabotropic receptors (Wang et al., 2010; Juuri et al.,

2010). Changes in background synaptic activity might

therefore produce subtle but at the same time robust

effects on what is normally referred as the ‘‘intrinsic’’

behavior of neurons. In the present study we employed

a well established ex vivo–in vitro infrared-laser lesion

model in the rat visual cortex and we investigated, in the

first week post-lesion, potential changes in passive and

active membrane properties of the surviving layer 2/3

pyramidal neurons. To better study the influence of

network activity on these neuronal properties, brain

slices were activated using an artificial cerebrospinal

fluid (ACSF) slightly different in ionic composition from a

standard bath solution (Bar-Yehuda and Korngreen,

2007). Similar ‘‘modified’’ ACSFs (mACSFs) have been

reported to enhance synaptic and neuronal activity in

cortical slices and have often been implemented to re-

create, at least to some extent, the high ongoing activity

found in vivo (Sanchez-Vives and McCormick, 2000;

Silberberg et al., 2004; Reig and Sanchez-Vives, 2007).

Interestingly, in the presence of this exciting medium we

observed robust lesion-induced alterations in both input

resistance (Ri) and neuronal firing properties. More

moderate alterations were present when slices were

perfused with a standard solution. Possibly, the reported

alterations might have been underestimated so far due

to the common use of ‘‘standard’’ ACSFs which dampen

neuronal activity in brain slices (Bar-Yehuda and

Korngreen, 2007). Synaptic transmission was found to

play an important role in the observed lesion-mediated

effects. Exciting cortical slices may have therefore

unmasked these changes by increasing the level of

ongoing synaptic activity and as a consequence the

contribution of synaptic receptors in shaping the

‘‘intrinsic’’ behavior of neurons. All together our findings

underscore that focal cortical lesions can cause

robust changes in neuronal membrane and firing

properties which may contribute, more than previously

thought, to the pathophysiological processes following

brain injuries.

EXPERIMENTAL PROCEDURES

Ethical statement

This study was carried out in strict accordance with the

German regulations for experimentation with vertebrate

animals. The protocol was approved by the Ethics

committee of the University of Mainz (G11-1-016). The

number of animals was kept to a minimum and all

efforts were made to minimize suffering.

Cortical lesion induction

Long–Evans rats (n= 59) at the age of 21 days were

anesthetized by an intraperitoneal injection of a mixture

of Ketamine (100 mg/kg) and Xylazine (8 mg/kg). The

animals were subsequently fixated in a stereotaxic

apparatus, the skull was exposed and cautiously drilled

above the right visual cortex parallel to the midline in a

rectangular area of 1-mm width beginning right anterior

to the lambda suture and extending 3 mm toward the

Bregma without penetrating the dura mater. Cortical

lesions were made under visual control with an 810-nm

infrared diode laser (OcuLight SLx, Iris Medical, USA)

attached to a binocular operating microscope. Multiple,

confluent lesions were induced about 2 to 2.5 mm

lateral from the midline in order to form an elongated

lesion of 1 mm mediolateral width and 3 mm

anteroposterior length starting anterior to the lambda

suture in the visual cortex (areas V1M, V2ML V2MM)

(Paxinos and Watson, 1998). Age-matched siblings

were used as sham-operated controls. They underwent

the same surgical procedure however after drilling the

skull no laser-lesions were induced.

Electrophysiology

Slice preparation. After a survival time of 2–5 days

the animals were deeply anesthetized with isoflurane

and decapitated. Coronal slices containing the visual

cortex (300 lm) were prepared from the lesioned

hemisphere by use of a vibratome (LEICA, VT-1000-S,

Germany). The tissue was kept at room temperature

and incubated for 1 h in a standard ACSF containing

(in mM): 125 NaCl, 25 NaHCO3, 2.5 KCl, 1.5 MgCl2, 2

CaCl2, 1.25 NaH2PO4, and 25 D-glucose (pH 7.4) and

bubbled with 95% O2 and 5% CO2. Single slices were

transferred into a submerged recording chamber.

Slices were superfused (perfusion rate: 3.5 ml/min) with

either the standard ACSF or two different ‘‘modified’’

ACSFs. The ‘‘modified’’ ACSF used for most of the

experiments will be referred as K+ 5 mM and

contained (in mM): 126 NaCl, 25 NaHCO3, 5 KCl, 1

MgCl2, 1 CaCl2, 1.25 NaH2PO4, and 25 D-glucose. A

second ‘‘modified’’ ACSF was used for a subset of

experiments. This solution will be referred as K+

3.5 mM since it had the same ionic composition as K+

5 mM but the KCl concentration was decreased from 5

to 3.5 mM. The perfusing solutions were always

bubbled with 95% O2 and 5% CO2. During the

experiments the ACSF temperature was kept at

31 ± 1 �C. The recording chamber was mounted on an

upright microscope (Olympus-BX50WI, Olympus,

Japan) equipped with 2.5� and 40� water immersion

type objectives.

Whole-cell patch-clamp recordings. Current clamp

whole-cell patch-clamp recordings were performed from

layers 2/3 pyramidal neurons under visual control using

DIC optics. Patch pipettes were pulled from borosilicate

glass capillaries (GB 150F-8P, Science Products,

Germany) and their resistance ranged from 4 to 6 MOwhen filled with intracellular solution. The intracellular

solution contained (in mM): 140 K-gluconate, 8 KCl, 2

MgCl2, 4 Na2-ATP, 0.3 Na2-GTP, 10 Na-

phosphocreatin and 10 HEPES. The pH was set to 7.3

with KOH. Bridge balance was applied throughout

current-clamp experiments. The membrane and firing

properties of pyramidal neurons were studied by

applying a series of 1-s-lasting square pulses of

hyperpolarizing and depolarizing currents through the

210 B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221

patch-clamp electrode (at 0.2 Hz). Typically we started by

applying �50 pA, then we incremented the magnitude of

the injected current by 25 pA for each step. To provide

a precise measurement of rheobase, we additionally

used a protocol in which neurons were kept at �70 mV

and the amplitude of the injected current was gradually

increased by relatively small steps (0.5 pA each step)

until the threshold membrane potential for spike

generation was reached. Firing frequency was

calculated by counting the number of action potentials

during the entire duration of each current pulse. To

investigate spike time reliability we injected consecutive

depolarizing sinusoidal current pulses at 2 or 5 Hz.

Since the Ri was increased post-lesion the magnitude of

current necessary to trigger one spike per sine wave

cycle in sham-operated rats produced more spikes per

cycle in lesioned animals (data not shown). Therefore,

we did not inject the same amount of current in all

neurons but adjusted the magnitude of the applied

current in each cell to induce at least one (and

preferentially only one) action potential per cycle. Initially

depolarizing currents, producing subthreshold

membrane oscillations, were applied. The magnitude of

the injected current was then increased in 2 pA steps

until one action potential was generated in each cycle.

For experiments conducted in voltage-clamp the

intracellular solution contained (in mM): 125 Cs-

gluconate, 5 CsCl, 10 EGTA, 2 MgCl2, 2 Na2-ATP, 0.4

Na2-GTP, 10 HEPES and 5 QX-314. The pH was set to

7.3 with CsOH. Access resistance was controlled before

and after each voltage clamp recording. Cells were not

further analyzed if this parameter was either higher than

20 MO or changed more than 20%. Spontaneous

excitatory and inhibitory postsynaptic currents

(spontaneous EPSCs (sEPSCs), spontaneous IPSCs

(sIPSCs)) were isolated by clamping neurons at

�60 mV and +10 mV, which are in very close proximity

to the reversal potential for GABAARs and AMPARs,

respectively.

Electrophysiological data acquisition and analysis. All

data performed in current-clamp mode were recorded

with an Axoclamp-2B amplifier (AXON Instrument,

USA). Axopatch 200B (Axon instrument, USA) was

used to collect data in voltage-clamp. Data were filtered

at 10 kHz and digitized at 20 kHz using a Digidata-1400

system with PClamp 10 software (Molecular Devices,

Sunnyvale, CA, USA). PClamp 10.1 software and

Matlab were used for off-line analysis. Somatic Ri was

calculated as the peak of voltage deflection in response

to hyperpolarizing (�50 pA) one-second-lasting current

pulses. To analyze spike threshold, spike amplitude and

spike half-width, we selected the first spike generated

with the lowest depolarizing current injected. To

measure spike threshold the maximum rate of change

of Vm (dVm/dt)MAX was computed during the upstroke

of each spike. The membrane potential threshold was

defined as the voltage at the onset of each spike at

which 3% of (dVm/dt)MAX was reached. This fraction

(3%) of (dVm/dt)MAX was chosen because it best

matched the threshold assigned by visual inspection of

the raw traces (Azouz and Gray, 2000). Phase plots

were constructed in each recorded cell by analyzing the

first action potential evoked by somatic current injection

of increasing amplitude. We selected a segment of the

voltage trace from 2 ms before to 5 ms after the peak of

the first evoked spike. Spike amplitude was calculated

as voltage difference between the threshold and the

peak of the spike. Spike half-width was measured as

spike duration at half spike amplitude. To examine the

degree of spike frequency adaptation (SFA) we

analyzed the firing pattern in response to 1-s-lasting

positive-current injection inducing a train of 8 to 12

spikes. The traces selected for this purpose did not

differ in the mean firing rate between the experimental

groups. The first interspike interval (ISI) was defined as

the time between the first and second action potential in

a train. The adaptation coefficient was then calculated

dividing the first ISI by the average of the last two ISIs

in the train. The difference between the peak of the

negative voltage deflection following a spike and the

spike threshold was used to measure

afterhyperpolarization (AHP) (in case the difference was

a negative value) or afterdepolarization (ADP) (in case

the difference was a positive value). To examine spike

timing in response to repetitive sinusoidal current

injections we considered a complete cycle having 180

degrees and then we measured the phase (expressed

in degrees) of the cycle at which the spike occurred (in

case two spikes were present in a cycle only the phase

of the first one was analyzed). To quantitatively

measure spike timing precision we calculated the

standard deviation of spike latencies (spike jitter) in

response to repeated sinusoidal current injections.

sEPSCs and sIPSCs were semi-automatically identified

by the software Clampfit 10.1 and were further validated

by careful visual inspection. Frequency and amplitude of

sESPCs and sIPSCs were calculated as the median of

500 events for each cell.

Drugs

The following blockers were used: D-(–)-2-amino-5-

phosphonopentanoic acid (D-AP5) (25 lM, Tocris

Biozol, Eching Germany), 6,7-dinitroquinoxaline-2,3-

dione (DNQX) (20 lM, Tocris Biozol, Eching Germany),

picrotoxin (PTX) (50 lM, Tocris Biozol, Eching

Germany), CGP55845 (1 lM, Biozol, Germany).

Statistics

The results are presented as mean ± SEM. The statistic

evaluation of the data was performed with an unpaired

Student-t test after verifying, with one-sample

Kolmogorov–Smirnov test, the normal distribution of the

data. In case more than two experimental groups were

compared statistical significance was assessed with

analysis of variance (ANOVA). Pearson’s correlation

coefficients were used when computing correlations

between two variables. Significantly different values

(⁄p < 0.05; ⁄⁄p < 0.01; ⁄⁄⁄p< 0.001) are indicated by

asterisks in the figures.

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 211

RESULTS

Cortical lesions were induced with an infrared laser in the

visual cortical areas V1M, V2MM, V2ML in the region

Bregma �5 mm to Bregma �8 mm (Paxinos and

Watson, 1998). As represented in the Nissl stained

coronal section in Fig. 1A the lesion consists in a focal

necrotic area measuring around 1 ± 0.2 mm in

mediolateral extent. Our lesion model therefore

represents a useful tool to investigate the

pathophysiological processes following a substantial, but

at the same time localized, damage of cortical tissue

(Roll et al., 2012). To study the spatial profile of lesion-

induced functional alterations, patch-clamp recordings

were obtained from layer 2/3 pyramidal neurons located

both in the cortical hemisphere ipsilateral and

contralateral to the lesion in coronal sections at

�6.5 ± 0.5-mm distance from Bregma. Neurons

selected for recordings were divided into four groups

depending on the location with respect to the border of

lesion: les 1 and les 2 included recordings of neurons

located in the cortex ipsilateral to the lesion at

respectively 1- and 2-mm distance from the border of

the lesion, contra and sham included neurons from the

undamaged hemisphere contralateral to the lesion and

from the cortex of sham-operated rats in a homotopic

area to the region les 1 (Fig. 1A). One representative

pyramidal neuron is shown in Fig. 1B. The first sets of

experiments were performed using the solution K+

5 mM as medium to perfuse cortical slices.

Increased Ri and evoked firing rate post-lesion

First, we measured somatic Ri by injecting small negative

current pulses (�50 pA) producing a hyperpolarization of

few mV from the resting membrane potential (resting Vm).

Ri was significantly increased in neurons located

ipsilateral to the lesion (sham: 42.21 ± 2.28 MO,n= 17 from 5 rats; les 1: 63.38 ± 2.97 MO, n= 21

from 5 rats, p< 0.001; les 2: 60.96 ± 4.45 MO, n= 17

from 4 rats, p< 0.005) while it remained unchanged in

the contralateral hemisphere (contra: 45.43 ± 2.74 MO,n= 17 from 5 rats, p> 0.05), (Fig. 2A, B). As

expected, the increased Ri reduced the amount of

current required to elicit action potentials causing a left

Fig. 1. Histology of the focal laser lesion. (A) Nissl-stained coronal section co

position of the recording electrodes located at around 1 and 2 mm from the

regions to les 1 were selected to record neurons from sham-operated rats. (B

pyramidal neuron.

shift in the firing frequency–current curve post-lesion (for

250 pA current injection, contra: 7.71 ± 1.29 Hz,

n= 17 from 5 rats, p> 0.05; sham: 4.35 ± 1.40 Hz,

n= 17 from 5 rats; les 1: 12.5 ± 1.26 Hz, n= 18 from

5 rats, p< 0.001; les 2: 11.41 ± 1.51 Hz, n= 16 from

4 rats, p< 0.005, Fig. 2A, C). The increase in Ri is

expected to reduce the amount of current required for

spike induction (rheobase). To provide a precise

measurement of rheobase we additionally recorded cells

from the sham and les 1 animals using a different

protocol in which the magnitude of the injected current

was incremented by relatively small steps (0.5 pA each

step). The rheobase was significantly reduced post-

lesion (sham: 213.08 ± 14.34 pA, n= 19, from 4 rats;

les 1: 161.10 ± 8.59 pA, n= 19, from 4 rats, p< 0.01).

Furthermore, a significant negative correlation was

found between Ri and rheobase (sham r= �0.69p< 0.001, les 1 r= �0.59, p< 0.01, Fig. 2D). These

results confirmed that the excitability of neurons from

lesion-treated animals was increased due to the higher Ri.

Post-lesional changes in spike threshold

Counteracting the increased neuronal excitability due to

the elevation in Ri, we observed a significant positive

shift in the membrane potential threshold for spike

induction (spike threshold) both at 1 and 2 mm lateral

from the lesion border (sham: �45.87 ± 0.54 mV,

n= 17 from 5 rats; les 1: �43.76 ± 0.39 mV, n= 21

from 5 rats, p< 0.05; les 2: �42.24 ± 1.01 mV, n= 15

from 4 rats, p< 0.001). The spike threshold remained

unaltered in the cortical hemisphere contralateral to the

lesion (contra: �46.19 ± 0.58 mV, n= 17 from 5 rats,

p> 0.05), (Fig. 3A1–2). Generally somatic action

potentials from neocortical pyramidal neurons display a

rapid onset, appearing as a ‘‘kink’’ at the base of the

spike (Naundorf et al., 2006; Yu et al., 2008). As a

consequence the spike threshold could be identified as

abrupt increase in the rate of change of the membrane

potential (dVm/dt). To examine the spike threshold in

more detail we therefore constructed a phase plot for

each experimental group where the derivative of the

membrane potential (dVm/dt) was plotted as a function

of the membrane potential. Each plot included one spike

for each recorded neuron. From these plots the shifted

ntaining the lesion in the right visual cortex. The symbols illustrate the

border of the lesion and in the contralateral hemisphere. Homotopic

) Confocal image of one representative lucifer yellow-labeled layer 2/3

Fig. 2. Increased input resistance post-lesion. (A) Voltage responses to a hyperpolarizing (�50 pA), a subthreshold depolarizing (50 pA) (top) and a

suprathreshold depolarizing (300 pA) (bottom) 1-s-lasting current step from a representative neuron in each experimental group. The current pulse

protocols are shown on the right side in gray. (B) Mean input resistance (Ri) measured by injecting a hyperpolarizing (�50 pA) current pulse. (C)

Firing frequency versus injected current curves. (D) Rheobase versus Ri in sham and les 1. In this plot rheobase was measured by increasing the

amplitude of somatic current injection in 0.5 pA steps.

Fig. 3. Depolarized spike threshold post-lesion. (A1) Peak-aligned averaged action potential waveform for each experimental group. The rectangle

area is magnified on the right side for clarity. Note the positive shift in spike threshold post-lesion. (A2) Summary plot of the mean spike threshold for

each experimental group. (B) Phase plots displaying the rate of change in membrane potential (dVm/dt) as a function of the membrane potential

(Vm) for each experimental group. Each curve in phase plot graphs represents the first action potential evoked by somatic current injection of

increasing amplitude in each recorded cell. The insets on the right side of each plot represent a magnification of the rectangle area to emphasize the

Vm at which dVm/dt suddenly increased.

212 B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221

spike threshold post-lesion is clearly evident as positive

shift in the membrane potential at which the rapid rise in

dVm/dt takes place (Fig. 3B). Spike half-width and

amplitude were also analyzed. The spike half-width was

not significantly altered (contra: 1.24 ± 0.05 ms, n= 17

from 5 rats, p> 0.05; sham: 1.18 ± 0.03 ms, n= 17

from 5 rats; les 1: 1.26 ± 0.05 ms, n= 21 from 5 rats,

p> 0.05; les 2: 1.29 ± 0.04 ms, n= 15 from 4 rats,

p> 0.05, Fig. 4A1–2). The spike amplitude was

significantly reduced in the hemisphere ipsilateral to the

lesion (contra: 95.32 ± 0.57 mV, n= 17 from 5 rats,

p> 0.05; sham: 94.51 ± 0.73 mV, n= 17 from 5 rats;

les 1: 90.82 ± 1.21 mV, n= 21 from 5 rats, p< 0.01;

les 2: 89.87 ± 1.23 mV, n= 15 from 4 rats, p< 0.005,

Fig. 4B1–2). To test whether the reduced amplitude of

action potentials was a direct consequence of the

positive shift in spike threshold, we additionally

measured the membrane potential at the peak of each

Fig. 4. Spike half-width and amplitude. (A1) Peak- and voltage-aligned averaged action potential waveform magnified in time to emphasize spike

width. (A2) Summary plot showing the mean spike half-width for each experimental group. (B1) Voltage-aligned averaged action potential waveform.

Summary plots showing (B2) the mean spike amplitude and (B3) the mean membrane potential at spike peak for each experimental group. Graphs

showing (B4) a high correlation between spike threshold and spike amplitude and (B5) no correlation between spike threshold and spike peak in all

recorded neurons.

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 213

spike (spike peak). This value was not significantly

altered post-lesion (contra: 51.14 ± 0.56 mV; sham:

50.64 ± 0.47 mV; les 1: 50.01 ± 0.75 mV; les 2:49.63 ± 0.59 mV, p> 0.05, Fig. 4B3). Furthermore, a

strong correlation was found between spike threshold

and amplitude (r= �0.06, p< 0.001, Fig. 4B4) but not

between spike threshold and spike peak (r= 0.01,

Fig. 4B5). All together these data suggest that the

reduced amplitude of action potentials was not an

independent phenomenon but was mainly resulting from

the depolarizing shift in spike threshold.

Previous experimental data have shown that the spike

threshold of cortical neurons is not static but can notably

vary depending on the level of depolarization preceding a

spike (Azouz and Gray, 2000). We therefore asked

whether the observed depolarized spike threshold was

caused by a lesion-induced positive shift in resting Vm.

Against this hypothesis we did not find any statistically

significant depolarization in resting Vm post-lesion

(contra: �71.83 ± 0.41 mV, n= 17 from 5 rats,

p> 0.05; sham: �72 ± 0.61 mV, n= 17 from 5 rats;

les 1: �70.7 ± 0.69 mV, n= 21 from 5 rats, p> 0.05;

les 2: �71.48 ± 0.56 mV, n= 15 from 4 rats,

p> 0.05). In addition, we observed a significant

correlation between spike threshold and resting Vm in

neurons from the cortical hemisphere contralateral to

the lesion (r= 0.65, p< 0.01, Fig. 5A) and from

sham-operated animals (r= 0.72, p= 0.001, Fig. 5B).

Surprisingly this correlation was lost in neurons located

at 1 mm ipsilateral to the lesion (r= 0.04, p> 0.05,

Fig. 5C) and was weaker at 2 mm distance (r= 0.46,

p> 0.05, Fig. 5D). This finding suggests that under

physiological conditions cortical neurons are able to

modulate their spike threshold in response to variations

in resting Vm (or vice versa). However, this mechanism

seems to be impaired following a focal cortical lesion.

Since we neither observed lesion-induced alterations

in the cortex contralateral to the lesion nor we observed

significant changes between les 1 and les 2, we decided

to proceed with recordings only from neurons located at

1-mm distance from the lesion (les 1) and from the

homotopic cortex in sham-operated animals (sham).

Spike train accomodation and spike time precision

The described changes in Ri and spike threshold post-

lesion will likely cause bidirectional perturbations in

neuronal excitability. These findings however do not

provide any information concerning the firing behavior of

neurons. We therefore further examined the neuronal

firing pattern by injecting a one-second-lasting

suprathreshold current step. All recorded cells presented

SFA typical of most pyramidal neurons (Connors and

Gutnick, 1990). The adaptation coefficient was however

Fig. 5. Correlations among resting membrane potential and spike threshold. Graphs showing the correlations between spike threshold and resting

membrane potential (resting Vm) for all experimental groups: (A) contra, (B) sham, (C) les 1, (D) les 2.

Fig. 6. Altered spike frequency adaptation post-lesion. (A) Representative firing responses to a suprathreshold depolarizing 1-s-lasting current

step. The traces on the right represent a magnification of the rectangle area to better visualize the difference in the first interspike interval (ISI). (B)

Bar plot showing the mean adaptation coefficient calculated dividing the first ISI by the average of the last two ISIs in the train. (C) Bar plot showing

the mean first ISI. (D) Diagram showing the afterhyperpolarization/afterdepolarization (AHP/ADP) following the first somatically induced action

potential for each recorded neuron and the mean value for the two experimental groups. Note the increased AHP post-lesion. (E) A strong

correlation between AHP/ADP and first ISI suggests that the level of repolarization following the first spike can predict well the time of occurrence of

the subsequent spike.

214 B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221

increased post-lesion, primarily due to a significant

prolongation in the first ISI (adap. coeff., sham:

0.26 ± 0.04, n= 14 from 5 rats; les 1: 0.42 ± 0.04,

n= 19 from 5 rats, p< 0.05; first ISI, sham:

30.93 ± 5.11, n= 14 from 5 rats; les 1: 48.13 ± 4.77,

n= 19 from 5 rats, p< 0.05, Fig. 6A–C). The prolonged

first ISI was accompanied by a significant enhancement

in the AHP following the first spike (sham:

�0.85 ± 0.71 mV, n= 14 from 5 rats; les 1:�4.68 ± 0.90 mV, n= 19 from 5 rats, p< 0.01,

Fig. 6D). The increased AHP was most likely responsible

for the prolonged first ISI post-lesion since a strong

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 215

correlation was found between the two variables

(r= �0.9, p> 0.001, Fig. 6E). There is evidence that

SFA may modulate spike timing of neurons and

ultimately contribute to the emergence of neuronal

network rhythms (Fuhrmann et al., 2002). We therefore

tested the effect of the lesion on spike fidelity in response

to a suprathreshold sinusoidal current injection. The

injection of sinusoidal current has been used to measure

spike timing precision since it induces fluctuations which

closely resemble membrane potential oscillations

occurring in vivo (Cea-del Rio et al., 2010). In a recent

study whole-cell in vivo recordings showed that in the rat

visual cortex layer 2/3 pyramidal neurons exhibit slow

spontaneous membrane potential oscillations at around

2 Hz (Sun and Dan, 2009). We therefore started by

applying 10 consecutive sine waves of current at 2 Hz.

As expected, neurons from both sham and lesion-treated

animals were able to respond with a relatively high

temporal precision at this input frequency (Fig. 7A).

Interestingly, while the first spike occurred at a very

similar phase of the first wave (sham: 92.94 ± 1.48�,n= 18, from 4 rats; les 1: 91.25 ± 1.50�, n= 19, from 4

rats, p> 0.05) spikes occurred at a significantly earlier

phase in the subsequent cycles post-lesion (10th cycle,

sham: 94.07 ± 1.04�; les 1: 87.12 ± 1.15�, p< 0.001,

Fig. 7A, C). This alteration was however not

accompanied by an improvement in spike fidelity. The

Fig. 7. Unchanged spike timing precision post-lesion. (A, B) Representativ

response to 2-Hz and 5-Hz suprathreshold repetitive sinusoidal current pulse

considering one sinusoidal cycle having 180�) at which spikes occurred in r

Mean spike jitter for 2-Hz and 5-Hz sinusoidal current injection.

jitter in spike timing, which was used as a measurement

of spike timing precision, remained unaltered post-lesion

(sham: 3.61 ± 0.35 ms; les 1: 3.21 ± 0.41 ms,

p> 0.05, Fig. 7E). To exclude that potential changes in

spike fidelity might not be detected due to the relative

weak protocol we additionally challenged cells with a

higher number of sine wave cycles (20) at a higher

frequency (5 Hz). Once again spike timing precision

remained unaffected (spike jitter, sham: 4.85 ± 0.33 ms;

les 1: 4.54 ± 0.31 ms, p> 0.05, Fig. 7E). The higher

spike jitter at 5 Hz compared with 2 Hz is most likely due

to the fact that at 5 Hz spikes occurred at a progressively

later phase of the sine wave cycles (Fig. 7B, D). These

findings suggest that neurons are capable to precisely

encode subsequent incoming inputs at low frequency but

their reliability declines if the input frequency increases.

The reduced capability to follow inputs at 5 Hz was not

affected by the lesion (Fig. 7D, E).

Modulation of synaptic and neuronal activity incortical slices alters the magnitude of the lesion-induced effects

In a previous study performed in the same lesionmodel we

reported a moderate but not statistically significant change

in neuronal Ri post-injury (Imbrosci et al., 2010). The herein

described robust alterations might have therefore become

e voltage changes (top) and raster plot for firing activity (bottom) in

s. (C, D) Graphs showing the mean phase (expressed in degrees and

esponse to each sinusoidal cycle at 2 Hz and 5 Hz, respectively. (E)

216 B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221

more prominent by the usage of an exciting medium (K+

5 mM) to perfuse cortical slices. To explore this

possibility we repeated the experiments while perfusing

cortical slices with two different ACSFs. The first

alternative bath solution (K+ 2.5 mM) corresponded to a

standard ACSF and contained a relatively low K+

concentration and a relatively high concentration of

divalent ions. Background synaptic activity and neuronal

activity have been reported to be low in slices perfused

with solutions having a similar ionic composition

(Silberberg et al., 2004; Bar-Yehuda and Korngreen,

2007). The second alternative ACSF was more similar to

K+ 5 mM but contained only 3.5 mM K+. Based on

previous studies this solution may enhance synaptic and

neuronal activity with respect to K+ 2.5 mM (Sanchez-

Vives and McCormick, 2000; Reig and Sanchez-Vives,

2007) but to a minor extent than K+ 5 mM (Bar-Yehuda

and Korngreen, 2007). In accordance with these studies,

recordings from layer 2/3 interneurons and layer 5

pyramidal cells in control animals revealed a higher

Fig. 8. Lesion-induced changes in firing frequency–current relationship and i

synaptic blockers. (A–C) Firing frequency–current relationships and mean

3.5 mM and K+ 5 mM, respectively. (D–F) Firing frequency–current relations

GABAA receptor blocker, PTX, glutamate (AMPA and NMDA) receptor blocke

respectively. (G) Percentage change in input resistance (Ri) and (H) percent

lesion in comparison to sham-operated animals in the six experimental cond

probability to observe spontaneous action potentials with

increasing concentration of extracellular K+ in the ACSF

(unpublished observations Imbrosci and Mittmann). In

line with this assumption the resting Vm was gradually

depolarized by the three solutions (K+ 2.5 mM,sham:

�79.28 ± 1.06 mV, n= 19 from 4 rats; les 1:�80.09 ± 0.52 mV, n= 14 from 3 rats; K+

3.5 mM,sham: �75.54 ± 0.94 mV, n= 18 from 4 rats;

les 1: �75.83 ± 0.46 mV, n= 18 from 3 rats; K+

5 mM,sham: �72.00 ± 0.61 mV, n= 17 from 5 rats; les1: �70.70 ± 0.69 mV, n= 21 from 5 rats). When slices

were perfused with K+ 2.5 mM or K+ 3.5 mM, Ri and

firing frequency induced by somatic current injections

were increased post-lesion but to a lesser extent in

comparison to the results obtained with K+ 5 mM (Ri, K+

2.5 mM, sham: 94.05 ± 6.24 MX, n= 19 from 4 rats;

les 1: 113.6 ± 5.19 MX, n= 14 from 3 rats, p< 0.05;

K+ 3.5 mM,sham: 72.71 ± 5.00 MX, n= 18 from 4 rats;

les 1: 88.13 ± 4.45 MX, n= 18 from 3 rats, p< 0.05,

Fig. 8A–C, G, H). In a previous work by use of the same

nput resistance (Ri) with different perfusing ACSFs and in presence of

Ri (insets) when cortical slices were perfused with K+ 2.5 mM, K+

hips and mean Ri (insets) with K+ 5 mM and in the presence of the

rs, (DNQX and D-AP5) and the GABAB receptor blocker CGP55845,

age change in firing rate in response to 250-pA current injection after

itions presented in A–F.

Fig. 9. Lesion-induced changes in spike threshold and afterhyperpolarization (AHP)/afterdepolarization (ADP) with different perfusing ACSFs and

in the presence of synaptic blockers. (A) Peak-aligned averaged action potential waveform for neurons from sham-operated (gray) and lesion-

treated animals (black) under different experimental conditions (spikes are truncated for clarity). Slices were perfused with K+ 2.5 mM, K+ 3.5 mM

or K+ 5 mM (top, from left to right) or with K+ 5 mM in the presence of PTX, DNQX and D-AP5 or CGP55845 (bottom, from left to right). (B) Peak-

and voltage-aligned averaged action potential waveform for neurons from sham-operated (gray) and lesion-treated animals (black) under the same

experimental conditions described in A. Summary plots of the mean (C) spike threshold and (D) AHP/ADP for sham-operated and lesion-treated

animals for each of the experimental conditions presented in A and B.

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 217

lesion model, Ri was not significantly altered post-lesion

when slices were perfused with a standard ACSF

(Imbrosci et al., 2010). The lesion-induced increase in Ri

observed in the present study with the same perfusing

solution (K+ 2.5 mM) is most likely due to the closer

recording position of 1 mm from the border of the lesion

as compared to the previous study (2–3 mm). Changes in

spike threshold emerged when using K+ 3.5 mM (sham:

�41.46 ± 0.45 mV, n= 18 from 4 rats; les 1:�39.53 ± 0.49 mV, n= 18 from 3 rats, p< 0.01) but

were absent when slices were perfused with K+ 2.5 mM

(sham: �38.36 ± 0.32 mV, n= 19 from 4 rats; les 1:�38.56 ± 0.98 mV, n= 14 from 3 rats, Fig. 9A, C) An

increased AHP post-lesion was observed with both

alternative solutions but again the lesion-induced

changes in AHP was modest in comparison with that

obtained using K+ 5 mM (K+ 2.5 mM, sham:

�5.62 ± 0.36 mV, n= 18 from 4 rats; les 1:�7.67 ± 0.66 mV, n= 17 from 3 rats, p< 0.05; K+

3.5 mM,sham: �4.34 ± 0.51 mV, n= 18 from 4 rats; les1: �5.85 ± 0.49 mV, n= 20 from 3 rats, p< 0.05)

(Fig. 9B, D).

The role of synaptic transmission in the lesion-mediated changes in neuronal membrane and firingneuronal properties

In the last part of the study we investigated the contribution

of synaptic activity in the observed lesion-induced

alterations by repeating our recordings (with K+ 5 mM) in

presence of glutamate (AMPA and NMDA) receptor

blockers (DNQX, 20 lM and D-AP5, 25 lM respectively),

GABAAR or GABABR blockers (PTX, 50 lM and

CGP55845, 1 lM respectively). Each blocker reduced

the differences in Ri and firing frequency between sham

and lesion. Interestingly, this was mainly due to an

increase in Ri in the sham group. Among synaptic

blockers, the GABAB receptor (GABABR) blocker,

CGP55845, exerted the most robust effect (PTX,sham:

58.56 ± 4.34 MX, n= 25 from 5 rats; les 1:67.54 ± 4.07 MX, n= 28 from 5 rats, p> 0.05; DNQX/D-AP5, sham: 55.22 ± 3.22 MX, n= 18 from 3 rats; les1: 63.48 ± 4.10 MX, n= 20 from 3 rats, p> 0.05; CGP,

sham: 70.38 ± 6.20 MX, n= 17 from 5 rats; les 1:64.32 ± 6.60 MX, n= 19 from 4 rats, p> 0.05)

(Fig. 8D–H). Consistently, a significant positive shift in

spike threshold post-lesion was still found in the

presence of PTX (sham: �43.84 ± 0.41 mV, n= 25

from 5 rats; les 1: �42.36 ± 0.28 mV, n= 28 from 5

rats, p< 0.01) and DNQX/D-AP5 (sham:

�43.86 ± 0.29 mV, n= 18 from 3 rats; les 1:�42.52 ± 0.56 mV, n= 19 from 3 rats, p< 0.05) but

not in the presence of CGP55845 (sham:

�43.58 ± 0.62 mV, n= 17 from 5 rats; les 1:�43.07 ± 0.55 mV, n= 19 from 4 rats) (Fig. 9A, C).

Finally, an increased AHP post-lesion was still present

when PTX was applied (sham: 1.78 ± 0.56 mV, n= 25

from 5 rats; les 1: �0.56 ± 0.46 mV, n= 28 from 5 rats,

Fig. 10. Spontaneous synaptic transmission remained unaltered

post-lesion. (A) Representative traces of sIPSCs and sEPSCs

recorded at the reversal potential for ionotropic glutamatergic and

GABAA receptors, respectively. The complete abolishment of the

signals following bath application of PTX (50 lM) and DNQX (20 lM)

confirmed that they were due to the activation of GABAA and AMPA

receptors, respectively. (B) Mean frequency and (C) mean amplitude

of sEPSCs and sIPSCs.

218 B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221

p< 0.01), was attenuated by DNQX/D-AP5 (sham:

0.95 ± 0.56 mV, n= 18 from 3 rats; les 1:�0.42 ± 0.53 mV, n= 19 from 3 rats, p> 0.05) and

disappeared with CGP55845 (sham: 0.61 ± 0.49 mV,

n= 17 from 5 rats; les 1: 0.22 ± 0.62 mV, n= 19 from

4 rats) (Fig. 9B, D). Although only the blockage of

GABABRs completely abolished the differences between

lesion-treated and control animals, the lesion-induced

changes were also attenuated through pharmacological

inhibition of glutamate or GABAA receptors. Therefore,

we aimed to directly test the effect of the lesion on

spontaneous synaptic activity by recording sEPSCs and

sIPSCs under voltage clamp conditions (Fig. 10A).

Excitatory and inhibitory spontaneous events were

isolated by clamping each cell at the reversal potential for

GABAA and glutamatergic receptors, �60 and +10 mV,

respectively (Fig. 10A). Surprisingly, we did not observe

any significant changes in frequency and amplitude of

both sEPSCs (frequency, sham: 10.76 ± 1.14 Hz,

n= 18 from 5 rats; les 1: 12.67 ± 0.98 Hz, n= 21 from

4 rats, p> 0.05; amplitude, sham: 14.6 ± 1.13 pA; les1: 16.77 ± 0.98 pA, p> 0.05) and sIPSCs post-lesion

(frequency, sham: 12.49 ± 0.86 Hz; les 1:14.84 ± 0.98 Hz, p> 0.05; amplitude, sham:

22.07 ± 1.32 pA; les 1: 26.79 ± 1.77 pA, p> 0.05,

Fig. 10B, C).

DISCUSSION

Aim of the study and chosen experimental condition

In the present study we explored the effect of laser-

induced lesions in the rat visual cortex on the

membrane and firing properties of the surviving layer 2/

3 pyramidal neurons. Notably, our aim was not to study

the physiology of single neurons isolated from the

network but rather the behavior of neurons functionally

integrated in active microcircuits. The chosen

experimental condition gave us the opportunity to

explore the potential influence of ongoing synaptic

activity on the membrane and firing properties of

neurons. For our in vitro recordings we used a bath

solution slightly different in ionic composition from a

standard ACSF with the purpose of ‘‘exciting’’ brain

slices. The reason of using a modified ASCF arose from

a series of studies which underscored that the

background synaptic inputs measured in acute brain

slices perfused with a standard ACSF (generally

containing, in mM, 2.5 K+, 2 Ca2+ and 2 Mg2+) is low

in comparison with the high ongoing activity found

in vivo. The data collected under this experimental

condition might therefore not reliably reproduce the

‘‘in vivo scenario’’ (Bar-Yehuda and Korngreen, 2007).

To increase the level of background activity the most

effective approach seems to be a modification in the

ionic composition of the perfusing medium. For

instance, replacing the standard ACSF (in mM: 2.5 K+,

2 Ca2+, 2 Mg2+) with a modified one, which more

closely resembles the brain interstitial fluid in situ (in

mM: 3.5 K+, 1.0 or 1.2 Ca2+ and 1 Mg2+) was shown

to be sufficient to generate in vivo-like spontaneous

rhythmic oscillations in otherwise silent ferret neocortical

slices (Sanchez-Vives and McCormick, 2000; Reig and

Sanchez-Vives, 2007). Likewise, in another in vitrostudy, layer 5 cortical pyramidal cells displayed high

frequency discharges when slices were excited with a

modified ACSF containing (in mM) 6.25 K+, 1.5 Ca2+

and 0.5 Mg2+ while no spike activity was detected when

a standard ACSF was used (Silberberg et al., 2004).

‘‘Exciting’’ acute cortical slices with a specific

extracellular bath solution might therefore constitute a

valid approach for in vitro investigations of the cellular

mechanisms of neuronal network function under

physiological and pathological conditions. Commonly,

the manipulation of a standard ACSF consists of an

increase in the K+ concentration, from 2.5 to 3.5, 5 or

even 6 mM, and a decrease in the concentration of

divalent ions (Ca2+, Mg2+), generally from 2 to 1.2,

1 mM (Bar-Yehuda and Korngreen, 2007). In the

present study, reducing Ca2+ and Mg2+ to similar

values and raising the concentration of K+ to 3.5 and

5 mM progressively depolarized the neuronal resting Vm

of layer 2/3 pyramidal cells nonetheless neither

spontaneous action potentials nor rhythmic membrane

potential fluctuations were observed. Nonetheless, the

same neurons were always able to produce action

potentials when artificially depolarized, ruling out the

possibility that this was the result of recording from

unhealthy cells.

On the contrary, a good portion of layer 2/3

interneurons as well as layer 5 pyramidal neurons

became spontaneously active when perfused with a

mACSF containing 5 mM K+ (unpublished observation,

Imbrosci and Mittmann). Most likely, even an exciting

medium containing 5 mM K+ was not sufficient to

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 219

generate a sufficient level of synchronous activity able to

spontaneously depolarize layer 2/3 pyramidal neurons to

the spike threshold. Consistently, even in vivo, layer 2/3pyramidal cells are characterized by a relative low firing

frequency (Greenberg et al., 2008; Sun and Dan, 2009).

Finally, it should be noted that the employed K+ 5 mM,

mACSF has an electrolyte composition very similar to

the serum (Seiffert et al., 2004). This can be important

under pathophysiological conditions such as following a

focal brain injury. Indeed, brain lesions often cause a

breakdown of the blood brain barrier with a subsequent

extravasation of serum into the brain extracellular space

(Ivens et al., 2007).

Lesion-mediated changes in Ri and firing propertiesof layer 2/3 pyramidal neurons

Using the so modified ACSF to perfuse cortical slices we

found robust alterations in the membrane and firing

properties of layer 2/3 pyramidal neurons in the

hemisphere ipsilateral to the lesion. The first

observation was a significant increase in Ri at 1 and

2 mm ipsilateral from the lesion with a consequent

reduction in rheobase (Fig. 2). Therefore, following

cortical lesions, the surrounding, surviving neurons are

likely to have an increased sensitivity to incoming

synaptic inputs and some of the previously subthreshold

inputs may reach the threshold for the generation of

action potentials. This may influence not only the

excitability but also the receptive field properties of

neurons. Based on these results there are reasons to

believe that not only changes in GABAAR-mediated

inhibition (Domann et al., 1993; Mittmann et al., 1994;

Neumann-Haefelin et al., 1995; Buchkremer-Ratzmann

et al., 1996; Schiene et al., 1996; Reinecke et al., 1999;

Neumann-Haefelin and Witte, 2000) but also changes in

Ri could substantially contribute to the increased

neuronal firing and to the enlargement of neuronal

receptive field size observed in vivo in different cortical

lesion models (Eysel and Schmidt-Kastner, 1991;

Schiene et al., 1996; Eysel and Schweigart, 1999; Ding

et al., 2011). While resting Vm was not altered post-

lesion the spike threshold of neurons in the cortex

ipsilateral to the lesion was significantly shifted to more

depolarized potentials (Fig. 3). Intuitively, a positive shift

in spike threshold should reduce neuronal excitability by

increasing the level of membrane depolarization needed

to generate an action potential. It is plausible that these

alterations are part of a homeostatic mechanism

attempting to compensate, at least to some extent, for

the increased Ri. Remarkably, in the cat visual cortex

the spike threshold was found to contribute to the

sharpness of orientation tuning, to increase cells

direction selectivity (Carandini and Ferster, 2000;

Volgushev et al., 2000) and to ameliorate the distinction

between simple and complex cells (Priebe et al., 2004).

The altered spike threshold post-lesion, by altering the

postsynaptic potential to spike transformation, could

therefore influence not only the excitability but also the

functional specificity of pyramidal neurons. It is

interesting to mention that in sham-operated animals the

cell to cell variability in spike threshold was found to

correlate with the resting Vm. However, this correlation

was lost post-lesion (Fig. 5). This suggests that the

shifted spike threshold did not arise from an overall

depolarization of the neurons. Alternatively, a decreased

availability of voltage gated Na+ channels or an

increase in K+ conductance might take place post-

injury. The laser-induced cortical lesion caused changes

in the firing behavior of the neurons as well. In

particular, we observed a reduced SFA due to a relative

reduction in the initial firing rate (Fig. 6A–C). The SFA is

commonly attributed to an action potential-dependent

K+ conductance which hyperpolarizes neurons slowing

the occurrence of subsequent spikes (Fuhrmann et al.,

2002). Indeed, a significantly increased AHP following

the first spike was evident post-lesion (Fig. 6D). This

suggests that changes in the activation of K+ or Ca2+-

activated K+ channels could be responsible for the

increased interval between first and second action

potential in the train. At the cellular level changes in

SFA are likely to have an impact on the neuronal input–

output relationship. For instance, SFA could cooperate

together with feedback inhibition to impose a certain

temporal constrain to the output of the neurons. Cortical

information processing might therefore be affected by

the observed changes in SFA. At the network level the

functional relevance of spike accommodation remains

unclear. However, it has been proposed that SFA could

influence frequency of neocortical oscillations by

modulating the spike timing reliability of neurons

(Fuhrmann et al., 2002). Spike timing precision,

measured as the ability of neurons to reliably respond to

repetitive suprathreshold sinusoidal current injection at 2

and 5 Hz, was however unaltered post-lesion (Fig. 7E).

Although the precision in spike timing remained

unchanged, the lesion caused action potentials to occur

at an earlier phase of the sinusoidal waves at 2 Hz

(Fig. 7C). A simple increase in Ri is unlikely to explain

these differences, because (1) the phase shift emerged

only with the second wave cycle and (2) the amplitude

of the injected current was carefully adjusted to be the

minimum required to induce one spike in each cycle.

Changes in active membrane properties following the

induction of the first spike are most likely responsible for

this phenomenon. One may speculate that this temporal

shift in spike occurrence could have an impact on

cortical representation of incoming stimuli as well as on

some forms of spike-timing dependent plasticity.

Contribution of synaptic receptors in the observedlesion-mediated effects

Interestingly, we found that the changes in Ri and firing

properties emerged or were enhanced by ‘‘exciting’’

perfusing solutions (K+ 3.5 mM, K+ 5 mM) while were

dampened or disappeared in the presence of a standard

ACSF (K+ 2.5 mM). This finding suggests that the

observed changes were unmasked by the usage of an

exciting medium which increased the activity of cortical

networks in vitro (Figs. 8 and 9). In the last part of the

study we tried to disclose if the alterations in neuronal

membrane and spike properties emerged from changes

at the level of single cells or, indirectly, from changes at

220 B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221

the level of the network. Indeed, recent studies have

shown that ‘‘exciting’’ brain slices with a mACSF can

affect not only intrinsic (Bar-Yehuda and Korngreen,

2007) but also synaptic (Reig and Sanchez-Vives, 2007)

properties of neurons. For example, alterations in the

extracellular K+ concentration, is likely to increase the

conductance of GIRK channels and to alter Vm at

presynaptic terminals with indirect effects on

neurotransmitter release. We therefore repeated our

recordings in the presence of either glutamate (AMPA

and NMDA) receptor, GABAAR or GABABR blockers.

The reported lesion-induced changes in Ri and spike

properties were completely abolished only in the

presence of the GABABR blocker, CGP55845 (Figs. 8

and 9). GABABRs, with their ability to modulate Ca2+

and K+ conductance (Misgeld et al., 1995), may

therefore contribute to the observed alterations. Yet,

application of either GABAA or AMPA and NMDA

receptor blockers (DNQX and D-AP5) did also reduce

the differences in Ri, spike frequency, spike threshold

and AHP (Figs. 8 and 9). Therefore, ionotropic

glutamatergic and GABAergic receptors may also, at

least in part, mediate the reported changes. In line with

this all synaptic receptor blockers had similar effects in

cells from control animals: they caused an increase in

Ri (Fig. 8), a depolarization of the spike threshold and a

robust shift from AHP to ADP (Fig. 9). However, we

cannot exclude that each pharmacological manipulation

may have indirect effects on the studied parameters by

altering neuronal network activity in cortical slices.

Finally, we evaluated the effect of the lesion on

spontaneous network activity by measuring,

spontaneous excitatory and inhibitory synaptic currents

in layer 2/3 pyramidal cells. Both frequency and

amplitude of sEPSCs and sIPSCs remained unaltered

post-lesion (Fig. 10). This result suggests that phasic

synaptic transmission is not likely to have a strong

impact on the observed changes in Ri and spike

threshold. Alternatively, tonic synaptic conductance may

play an important role here. The unaltered synaptic

transmission following lesion was somehow unexpected,

since a weaker inhibition has so far been considered the

most likely candidate to explain the post-lesion

hyperexcitability (Domann et al. 1993; Mittmann et al.

1994; Neumann-Haefelin et al. 1995; Buchkremer-

Ratzmann et al. 1996; Schiene et al. 1996; Reinecke

et al. 1999; Neumann-Haefelin and Witte, 2000). The

reason why we failed to observe any sign of reduction in

spontaneous GABAergic transmission could rely on the

modified perfusing ACSF solution which may have

increased the spontaneous activity of interneurons and

boosted neurotransmitter release. Our results raise

some interesting questions about the actual changes in

the strength of inhibition following brain injuries. To

address this issue, future studies should evaluate in

more detail, which, of the several and currently used

in vitro experimental conditions, can better represent the

‘‘in vivo scenario’’. Besides ACSFs with different ionic

composition, other parameters, such as higher

temperature and different perfusion rates, should be

tested.

To conclude, the data presented in this study highlight

that synaptic receptors-mediated changes in membrane

and firing properties are likely to significantly contribute

to the hyperexcitability and neuronal network

dysfunction often reported following cortical injuries.

Acknowledgments—We thank Simone Dahms-Praetorius for

excellent technical assistance. This work was supported by a

grant from the German Research Foundation (DFG) to T.M.

(MI 452/4-1).

REFERENCES

Alves OL, Bullock R, Clausen T, Reinert M, Reeves TM (2005)

Concurrent monitoring of cerebral electrophysiology and

metabolism after traumatic brain injury: an experimental and

clinical study. J Neurotrauma 22:733–749.

Azouz R, Gray CM (2000) Dynamic spike threshold reveals a

mechanism for synaptic coincidence detection in cortical

neurons in vivo. Proc Natl Acad Sci U S A 97:8110–8115.

Bar-Yehuda D, Korngreen A (2007) Cellular and network

contributions to excitability of layer 5 neocortical pyramidal

neurons in the rat. PLoS One 2:e1209.

Buchkremer-Ratzmann I, August M, Hagemann G, Witte OW (1996)

Electrophysiological transcortical diaschisis after cortical

photothrombosis in rat brain. Stroke 27:1105–1111.

Carandini M, Ferster D (2000) Membrane potential and firing rate in

cat primary visual cortex. J Neurosci 20:470–484.

Cea-del Rio CA, Lawrence JJ, Erdelyi F, Szabo G, McBain CJ (2010)

Cholinergic modulation amplifies the intrinsic oscillatory properties

of CA1 hippocampal cholecystokinin-positive interneurons. J

Physiol 589:609–627.

Chance FS, Abbott LF, Reyes AD (2002) Gain modulation from

background synaptic input. Neuron 35:773–782.

Connors BW, Gutnick MJ (1990) Intrinsic firing patterns of diverse

neocortical neurons. Trends Neurosci 13:99–104.

Destexhe A, Pare D (1999) Impact of network activity on the

integrative properties of neocortical pyramidal neurons in vivo. J

Neurophysiol 81:1531–1547.

Ding MC, Wang Q, Lo EH, Stanley GB (2011) Cortical excitation and

inhibition following focal traumatic brain injury. J Neurosci

31:14085–14094.

Domann R, Hagemann G, Kraemer M, Freund HJ, Witte OW (1993)

Electrophysiological changes in the surrounding brain tissue of

photochemically induced cortical infarcts in the rat. Neurosci Lett

155:69–72.

Eysel UT, Schmidt-Kastner R (1991) Neuronal dysfunction at the

border of focal lesions in cat visual cortex. Neurosci Lett

131:45–48.

Eysel UT, Schweigart G (1999) Increased receptive field size in the

surround of chronic lesions in the adult cat visual cortex. Cereb

Cortex 9:101–109.

Fuhrmann G, Markram H, Tsodyks M (2002) Spike frequency

adaptation and neocortical rhythms. J Neurophysiol 88:761–770.

Greenberg DS, Houweling AR, Kerr JN (2008) Population imaging of

ongoing neuronal activity in the visual cortex of awake rats. Nat

Neurosci 11:749–751.

Imbrosci B, Eysel UT, Mittmann T (2010) Metaplasticity of horizontal

connections in the vicinity of focal laser lesions in rat visual cortex.

J Physiol 588:4695–4703.

Ivens S, Kaufer D, Flores LP, Bechmann I, Zumsteg D, Tomkins O,

Seiffert E, Heinemann U, Friedman A (2007) TGF-beta receptor-

mediated albumin uptake into astrocytes is involved in neocortical

epileptogenesis. Brain 130:535–547.

Juuri J, Clarke VRJ, Lauri SE, Taira T (2010) Kainate receptor-

induced ectopic spiking of CA3 pyramidal neurons initiates

network bursts in neonatal hippocampus. J Neurophysiol

104:1696–1706.

B. Imbrosci, T. Mittmann /Neuroscience 250 (2013) 208–221 221

Leger JF, Stern EA, Aertsen A, Heck D (2005) Synaptic integration in

rat frontal cortex shaped by network activity. J Neurophysiol

93:281–293.

Misgeld U, Bijak M, Jarolimek W (1995) A physiological role for

GABAB receptors and the effects of baclofen in the mammalian

central nervous system. Prog Neurobiol 46:423–462.

Mittmann T, Luhmann HJ, Schmidt-Kastner R, Eysel UT, Weigel H,

Heinemann U (1994) Lesion-induced transient suppression of

inhibitory function in rat neocortex in vitro. Neuroscience

60:891–906.

Naundorf B, Wolf F, Volgushev M (2006) Unique features of action

potential initiation in cortical neurons. Nature 440:1060–1063.

Neumann-Haefelin T, Hagemann G, Witte OW (1995) Cellular

correlates of neuronal hyperexcitability in the vicinity of

photochemically induced cortical infarcts in rats in vitro.

Neurosci Lett 193:101–104.

Neumann-Haefelin T, Witte OW (2000) Periinfarct and remote

excitability changes after transient middle cerebral artery

occlusion. J Cereb Blood Flow Metab 20:45–52.

Paxinos G, Watson C (1998) The rat brain in stereotaxic coordinates.

4th ed. London: Academic press.

Priebe NJ, Mechler F, Carandini M, Ferster D (2004) The contribution

of spike threshold to the dichotomy of cortical simple and complex

cells. Nat Neurosci 7:1113–1122.

Reig R, Sanchez-Vives MV (2007) Synaptic transmission and

plasticity in an active cortical network. PLoS One 2:e670.

Reinecke S, Lutzenburg M, Hagemann G, Bruehl C, Neumann-

Haefelin T, Witte OW (1999) Electrophysiological transcortical

diaschisis after middle cerebral artery occlusion (MCAO) in rats.

Neurosci Lett 261:85–88.

Roll L, Mittmann T, Eysel UT, Faissner A (2012) The laser lesion of

the mouse visual cortex as a model to study neural extracellular

matrix remodeling during degeneration, regeneration and

plasticity of the CNS. Cell Tissue Res 349:133–145.

Sanchez-Vives MV, McCormick DA (2000) Cellular and network

mechanisms of rhythmic recurrent activity in neocortex. Nat

Neurosci 3:1027–1034.

Schiene K, Bruehl C, Zilles K, Que M, Hagemann G, Kraemer M,

Witte OW (1996) Neuronal hyperexcitability and reduction of

GABAA-receptor expression in the surround of cerebral

photothrombosis. J Cereb Blood Flow Metab 16:906–914.

Seiffert E, Dreier JP, Ivens S, Bechmann I, Tomkins O, Heinemann

U, Friedman A (2004) Lasting blood–brain barrier disruption

induces epileptic focus in the rat somatosensory cortex. J

Neurosci 24:7829–7836.

Silberberg G, Wu C, Markram H (2004) Synaptic dynamics control

the timing of neuronal excitation in the activated neocortical

microcircuit. J Physiol 556:19–27.

Sun W, Dan Y (2009) Layer-specific network oscillation and

spatiotemporal receptive field in the visual cortex. Proc Natl

Acad Sci U S A 106:17986–17991.

Volgushev M, Pernberg J, Eysel UT (2000) Comparison of the

selectivity of postsynaptic potentials and spike responses in cat

visual cortex. Eur J Neurosci 12:257–263.

Wang Y, Neubauer FB, Luscher H-R, Thurley K (2010) GABAB

receptor-dependent modulation of network activity in the rat

prefrontal cortex in vitro. Eur J Neurosci 31:1582–1594.

Witte OW, Bidmon HJ, Schiene K, Redecker C, Hagemann G (2000)

Functional differentiation of multiple perilesional zones after focal

cerebral ischemia. J Cereb Blood Flow Metab 20:1149–1165.

Yu Y, Shu Y, McCormick DA (2008) Cortical action potential

backpropagation explains spike threshold variability and rapid-

onset kinetics. J Neurosci 28:7260–7272.

(Accepted 27 June 2013)(Available online 9 July 2013)