Astrocytic gap junction blockade markedly increases extracellular potassium without causing seizures in the mouse neocortex

Accepted Manuscript Astrocytic gap junction blockade markedly increases extracellular potassium without causing seizures in the mouse neocortex

Paolo Bazzigaluppi, Iliya Weisspapir, Bojana Stefanovic, Luc Leybaert, Peter L. Carlen PII: DOI: Reference:

S0969-9961(16)30298-4 doi: 10.1016/j.nbd.2016.12.017 YNBDI 3884

To appear in:

Neurobiology of Disease

Received date: Revised date: Accepted date:

6 October 2016 24 November 2016 18 December 2016

Please cite this article as: Paolo Bazzigaluppi, Iliya Weisspapir, Bojana Stefanovic, Luc Leybaert, Peter L. Carlen , Astrocytic gap junction blockade markedly increases extracellular potassium without causing seizures in the mouse neocortex. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Ynbdi(2016), doi: 10.1016/j.nbd.2016.12.017

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Astrocytic Gap Junction blockade markedly increases extracellular potassium without causing seizures in the mouse neocortex Paolo Bazzigaluppi1,2, Iliya Weisspapir1, Bojana Stefanovic2,3, Luc Leybaert4 and Peter L. Carlen1

PT

1: Fundamental Neurobiology, Krembil Research Institute, University Health Network, M5T 2S8 Toronto, Ontario, Canada; 2:

SC

RI

Physical Sciences, Sunnybrook Research Institute, M4N 3M5 Toronto, Ontario, Canada; 3: Department of Medical Biophysics, University of Toronto, Ontario, Canada; 4: Department of Basic Medical Sciences, University of Ghent, 9000 Ghent, Belgium.

MA

AC

CE

PT E

D

Corresponding author: Dr. Paolo Bazzigaluppi Krembil Research Institute University Health Network 7KDT 430, 60 Leonard Av. M5T 2S8 Toronto, ON Canada. E-mail: [email protected]

NU

Keywords: extracellular potassium, GAP27, TAT-GAP19, 4-AP, in vivo, seizures

1

ACCEPTED MANUSCRIPT Abstract Extracellular potassium concentration, [K+]o, is a major determinant of neuronal excitability. In the healthy brain, [K+]o levels are tightly controlled. During seizures, [K+]o increases up to 15mM and is thought to cause seizures due to its depolarizing effect. Although astrocytes have been suggested

PT

to play a key role in the redistribution (or spatial buffering) of excess K+ through Connexin-43 (Cx43)-based Gap Junctions (GJs), the relation between this dynamic regulatory process and

RI

seizure generation remains unknown. Here we contrasted the role of astrocytic GJs and

SC

hemichannels by studying the effect of GJ and hemichannel blockers on [K+]o regulation in vivo.

NU

[K+]o was measured by K+-sensitive microelectrodes. Neuronal excitability was estimated by local field potential (LFP) responses to forepaw stimulation and changes in the power of resting state

MA

activity. Starting at the baseline [K+]o level of 1.61 ± 0.3 mM, cortical microinjection of CBX, a broad spectrum connexin channel blocker, increased [K+]o to 11 ± 3 mM, Cx43 GJ/hemichannel

D

blocker Gap27 increased it from 1.9 ± 0.7 to 9 ± 1 mM. At these [K+]o levels, no seizures were

PT E

observed. Cx43 hemichannel blockade with TAT-Gap19 increased [K+]o by only ~1 mM. Microinjection of 4-aminopyridine, a known convulsant, increased [K+]o to ~10 mM and induced

CE

spontaneously recurring seizures, whereas direct application of K+ did not trigger seizure activity. These findings are the first in vivo demonstration that astrocytic GJs are major determinants for

neocortex.

AC

the spatial buffering of [K+]o and that an increase in [K+]o alone does not trigger seizures in the

2

ACCEPTED MANUSCRIPT Introduction A long standing question in epilepsy research is the mechanism of neuronal (hyper)excitability underlying seizure generation. The neuronal excitability is dictated by extracellular ionic concentrations and synaptic activity (Florence et al., 2009). Non-synaptic mechanisms of seizure

PT

generation have been observed (Jefferys and Haas, 1982; Taylor and Dudek, 1982; Haas and Jefferys, 1984) and consist preponderantly of ionic fluctuations in the extracellular space.

RI

Experimental observation and computer models suggest neuronal excitability is primarily dictated

SC

by the intra- vs. extra-cellular potassium gradient (for review see, Florence et al., 2009). In particular, the observation that extracellular potassium ([K+]o) rises during seizures (Moody et al.,

NU

1974; Pedley et al., 1974; Heinemann et al 1977, Raimondo et al., 2015) led to the “potassium

MA

accumulation hypothesis”, which posits that an initial supra-threshold increase in [K+]o without any other experimental manipulations (such as low calcium or the presence of a convulsant) triggers

D

seizures via a positive feedback mechanism (Traynelis and Dingledine, 1988; Frohlich et al., 2008).

PT E

Although it is widely recognized that understanding the underlying relationship between [K+]o and seizure generation is paramount, this hypothesis has hitherto not been tested in vivo in the

CE

neocortex. It is known that astrocytes are involved in the regulation of [K+]o via the uptake of K+ through ion transporters and its dissipation via the so-called K+ spatial buffering (Orkand et al.,

AC

1966; Somjen, 2001; Kofuji and Newman, 2004). This process relies on inter-astrocytic communication via Gap Junctions (GJs), which are prevalently formed by Connexin43 (Cx43, for review (Scemes and Spray, 2012)) and Cx30 (Kunzelmann et al., 1999). While potassium spatial buffering is well established for the Müller cells of the retina (Newman, 1984, 1993), the importance of Cx43 GJs on K+ buffering via astrocytes in the Central Nervous System has been confirmed only in vitro (Wallraff et al., 2006). This mechanism has been presumed to regulate [K+]o globally, determine neuronal excitability, and, furthermore, play a key role in seizure generation.

3

ACCEPTED MANUSCRIPT To recapitulate the link between [K+]o dynamics and seizures, 4-Amino Pyridine (4-AP), a known Kv channel blocker and a convulsive agent (Reddy and Kuruba, 2013) without a direct action on GJs, is widely employed in preclinical research to generate seizures with a concomitant increase in [K+]o (Avoli et al 1996a,b, for review Avoli and de Curtis, 2011) In the present work, we examined the role of astrocytic Cx43-Gjs and hemi-channels in [K+]o regulation and seizure generation in the

PT

mouse neocortex in vivo, contrasting the effects on [K+]o of the broad-spectrum connexin channel

RI

blocker Carbenoxolone (CBX), of the Gap27 peptide (blocking Cx43 hemichannels and gap

SC

junctions, Evans and Boitano, 2001; Evans and Leybaert, 2007; Wang et al., 2012; Wang et al., 2013a), and of the TAT-GAP19 peptide (specifically blocking Cx43 hemichannels without inhibiting

NU

gap junctions (Wang et al., 2013b)). We observed a marked rise in [K+]o following blockade of GJs

MA

and hemichannels with CBX and Gap27. However, specific block of Cx43 hemichannels with TATGap19 did not influence [K+]o, indicating that mainly inter-astrocytic GJs were responsible for [K+]o

D

alterations, in line with the primary role of GJs in spatial K+ buffering. Although the ensuing

PT E

increase in [K+]o from blocking interastrocytic GJs was comparable to that observed after the 4-AP application, and local field potentials evoked by peripheral stimulation were depressed, no

CE

seizures were observed. These data suggest that 4-AP convulsive action is probably due to its blocking effect on Kv channels. To rule out definitively the hypothesis that high [K +]o alone causes

AC

cortical seizures in vivo, we progressively increased [K+]o by applying solutions with increasing [K+] onto the exposed cortex. We observed a bimodal effect of exogenous K+ solutions on neuronal excitability, in line with the endogenous K+ increases: increased somatosensory evoked responses and neuronal power for concentrations between 4 and 10 mM followed by depressed neuronal transmission for concentrations higher than 12 mM. We did not observe seizures at any of the concentrations tested. The present findings are the first in vivo data on the effects of local

4

ACCEPTED MANUSCRIPT astrocytic GJ blockade on [K+]o homeostasis, neuronal excitability, and responsivity to peripheral stimulation.

Material and Methods

PT

Animal preparation: Experiments were conducted on young adult CD-1 mice (4-5 weeks of age) in accordance with the guidelines of the animal welfare committee of the University Health Network.

RI

Mice were anesthetized by intraperitoneally injected Ketamine-Xylazine (respectively 95 and 5

SC

mg/kg b.w.) and placed into a stereotaxic frame. The animal body temperature was maintained at 37.5 °C using a heating pad (Physitemp, TCAT-2DF). Hind limb withdrawal reflexes and breathing

NU

rates were observed at regular intervals throughout the experiment to ensure that the animal

MA

remained at a surgical plane of anesthesia. Forepaw stimulation (train of 3 pulses of 100µs at 10Hz, 0.5-0.8mA) was delivered via needles implanted into the left forepaw. A local anesthetic

D

(Sensoricaine, AstraZeneca Canada Inc.) was injected subcutaneously into the scalp region to be

PT E

incised. A small craniotomy (diameter of ~2 mm) was performed over the right somatosensory cortex of the mouse, leaving the dura mater intact and a well was built around the opening to

CE

preserve constant level of CSF over the brain. Phosphate Buffered Saline (PBS, pH 7.4, Sigma) was applied over the exposed cortex to prevent tissue damage and dehydration. Five groups were

AC

studied, each undergoing a different cortical microinjection (see below): Control (PBS, n=4), TATGAP19 (n=3), Carbenoxolone (CBX, n=7), GAP-27 (n=7), and 4-AP (n=5). In the exogenous K+ experiments we sequentially applied onto the cortex buffered saline (pH 7.4, 0.4-0.5 ml) solutions with increasing [K+] (4, 6, 8, 12 and 20mM). Electrophysiological recordings: The procedure used to manufacture the K+-sensitive electrodes was similar to the one described in previous studies (Dufour et al., 2011; Bazzigaluppi et al., 2015; Wang et al., 2016). Briefly, K+-sensitive electrodes were made from pulled borosilicate capillaries

5

ACCEPTED MANUSCRIPT (tip diameter ~1 m, World Precision Instruments, Sarasota, FL). The interior wall of the capillary was silanized with dimethyldichlorosilane vapor and dried at 120°C for 2 h. The tip was then filled with the potassium Ionophore I-cocktail B (Sigma–Aldrich Canada Ltd, Oakville). The rest of the barrel was backfilled with a 0.2 M KCl. The signal at the reference barrel was subtracted from the

PT

signal at the K+-selective barrel to obtain a signal proportional to [K+]o. Local field potentials (LFP) were recorded with a pulled borosilicate capillary filled with PBS (in the control experiments), or a

RI

mixture of PBS and Carbenoxolone (CBX, 1 mM, Sigma), or the GAP27 peptide (500 µM,

SC

(Samoilova et al., 2008), Severn Biotech), or the TAT-GAP19 peptide (500 µM, (Wang et al., 2013b)) or 4-AP (5 mM, Sigma). The focal application of the pharmacological agents used (or

NU

vehicle) was performed by ten repetitions of 3 to 5-ms microinjections from the LFP pipette

MA

(PicoSpritzer III, Parker), applied over a time period of 5 mins; this resulted in a total injection volume of ~1 µl. The signal recorded with the LFP electrode was subtracted from that recorded

D

from the K+-sensitive electrode. Data was acquired with Axopatch 200B amplifiers and sampled at

PT E

10 kHz. Under an Olympus BX-61W1 microscope with 4X PlanN objectives, both electrodes were placed into the cortical layers 2–3 of the forelimb region of the mouse somatosensory cortex (-0.3

CE

to -0.5 mm from bregma, 1 to 1.3 mm from midline, 150 to 250 µm depth), so that their tips were within 10 µm of each other. LFP and extracellular K+ signals were low pass filtered (5 kHz) and

AC

digitized (Digidata 1440, Axon instruments). K+-sensitive electrodes were calibrated using solutions containing 2.5, 4.5, 6.5 and 22.5 mM KCl. The relationship between the measured voltage and the K+ concentration of the respective solution was derived using the Nicolsky-Eisenmann equation (Ammann, 1986). Data Analysis and Statistics: Given the relatively slow dynamics of [K+]o increase following drug application, resting state [K+]o was estimated at: before injection (i.e. baseline) and at 10 mins, 30 mins, 60 mins and 90 mins after injection, and expressed as a 1 minute average preceding each

6

ACCEPTED MANUSCRIPT somatosensory stimulation. In the same epochs, the Fast Fourier Transform of the cortical EEG was computed to estimate the power of the neuronal activity. In the experiments where exogenous K+ was applied onto the cortex, we waited up to 10 minutes for [K +]o to stabilize before recording spontaneous and evoked activity. Raw power at different time points was divided into

PT

the following frequency bands: Theta (4-8 Hz), Alpha (9-14 Hz), Beta (15-30 Hz). The raw power in each band was averaged across animals and then normalized to the baseline value (i.e. before any

RI

somatosensory stimulation or pharmacological manipulation) in that band. Twenty trains of

SC

somatosensory stimuli were delivered at each time point (or at different [K +]os) and the evoked field (LFP) responses were estimated as the ten-repetitions-average. We assessed normality of

NU

data distribution within groups with normal probability plots and Kolmogorov-Smirnov test, then

MA

tested for significance with one way-ANOVA (for normally distributed data) or with Kruskal-Wallis ANOVA (for non-normally distributed data). Within each group, statistical difference between the

D

different time points was assessed via one-way-ANOVA or Kruskal-Wallis one-way-ANOVA (see

PT E

Statistic Table – supplementary table 1) followed by Least Significant Difference post-hoc test for multiple comparisons. The p-values represent the result of one-way-ANOVA. Analysis was

AC

Results

CE

performed in Matlab. All the values represent mean ± SEM.

Connexin channel blockade increases [K+]o We performed measurements of intracortical (parenchymal) [K+]o and LFPs making use of two electrodes placed in the forelimb region of mouse somatosensory cortex positioned at the level of cortical layers 2-3. Given the slow changes in the [K+]o (see examples in Figure 1A), we present the results of the analyses performed at different time points (10, 30, 60 and 90 minutes). To unravel the relationship between [K+]o and seizure generation, we microinjected 4-AP via the

7

ACCEPTED MANUSCRIPT pipette electrode used for LFP recordings. This induced a rapid and significant (n=5, p=2*10 -9) increase in [K+]oa (Figure 1A,B brown), from 1.34 ± 0.5 mM before injection to 3.66 ± 0.5 mM after 10 mins; 5.41 ± 0.5 mM after 30 mins; 10.67 ± 0.5 mM after 60 mins; and 8.10 ± 0.6 mM after 90 mins. In all 4-AP experiments, 30-40 minutes after injection, we observed spontaneous recurring seizures characterized by significant increases in neuronal power in the theta b1, alphab2 and betab3

PT

bands (Figure 3 brown bars, p=0.007, p=0.0018, p=0.0055, respectively). At these [K+]o values, the

RI

neuronal responsivity to somatosensory forepaw stimulation was compromised: LFP response

SC

amplitudesc were significantly reduced (p=1*10-7), from 0.47 ± 0.03 mV before injection to 0.12 ± 0.03 mV after 60 mins, and to 0.09 ± 0.03 mV after 90 mins (Figure 2C, brown). We controlled for

NU

the effects of our experimental approach on [K+]o and LFP responses by injection of vehicle (PBS) in

MA

a separate cohort of animals (n=3). In these experiments, we did not find any changes in [K+]oo (p = 0.13), nor in resting statep (all bands p>0.2) or somatosensory evokedp activity (p=0.33, Figures 1

D

A,B, Figure 2 Ai, C blue and Figure 3 blue bars). Before injection of the vehicle, the control group

PT E

showed a baseline [K+]o of 1.5 ± 0.2 mM, not different from the baseline values measured in the other experimental cohorts in this work and similar to the 1.5 ± 0.2 mM observed in the mouse by

CE

Tong (Tong et al., 2014), although lower than 3 mM measured in other animal models (for review

AC

see (Frohlich et al., 2008, Raimondo et al., 2015). To examine the effect of astrocytic K+ spatial buffering on [K+]o, we first tested the effects of the broad-spectrum connexin channel blocker, Carbenoxolone (CBX, 1mM), which acts on multiple connexins including neuronal Cx36 (Sohl et al., 2005) as well as on astrocytic Cx30 (Pannasch et al., 2014) and Cx43 (Sagar and Larson, 2006). Parenchymal microinjections of CBX resulted in a significant (P=8*10-9) increase in [K+]od (Figure 1A, black trace), from 1.6 ± 0.3 mM during baseline, to 5.0 ± 1.5 mM after 10 mins, 7.5 ± 3.1 mM after 30 mins, 11.5 ± 1.9 mM after 60 mins and 11 ± 3 mM after 90mins (Figure 1B, black trace). During such supra-physiological [K+]o,

8

ACCEPTED MANUSCRIPT we examined spontaneous neuronal activity (Figure 3 black bars) and did not find a significant change in neuronal activity power in any bande (p=0.81). On the other hand, somatosensory evoked LFP responses showed decreased amplitudesf, from 0.52 ± 0.11 mV before injection to 0.15 ± 0.11 mV after 60 mins and 0.10 ± 0.06 after 90 mins (p=0.018, Figure 2C black). These

PT

results show that CBX increases [K+]o and alters neuronal activity, without generating seizures. However, CBX affects multiple targets in addition to connexins, including neuronal pannexin

RI

channels (Thompson et al., 2008), voltage-gated calcium channels (Vessey et al., 2004), intrinsic

SC

neuronal properties (Rouach et al., 2003), neurotransmitter release (Connors, 2012), and acts as an anticonvulsant (Jahromi et al., 2002). On the other hand, potassium buffering depends largely

NU

on the expression of astrocytic Cx43 based GJs: indeed, decreased Cx43 expression, induced by

MA

astrocyte-specific TSC1 gene KO, has been shown to be associated with impaired K+ buffering and the appearance of seizures in a Tuberous Sclerosis mouse model (Xu et al., 2009).

D

In order to obtain connexin channel inhibition in a more specific manner, we made use of

PT E

the peptide Gap27 that is identical to a sequence present on the second extracellular loop of Cx43 containing the conserved SRPTEK motif (Giaume et al, 2013, Evans and Boitano, 2001). Gap27

CE

rapidly inhibits Cx43 hemichannels (Wang et al., 2012) and more slowly affects GJs (Boitano and

AC

Evans, 2000; Decrock et al., 2009; De Bock et al., 2012; Samoilova et al., 2008). We performed a

series of experiments (n=7) with parenchymal injections of Gap27 (500 µM). This provoked a [K +]og rise with time (P=8*10-6): from 1.9 ± 0.7 at baseline to 6.0 ± 0.7 after 30 mins, 8.0 ± 0.7 after 60 mins and 8.9 ± 1.0 after 90 mins (Figure 1A,B green). No seizures were observed and there was no significant change in the power of any of the bands examined h (all bands p>0.13; Figure 3 green bars). However, somatosensory evoked LFP amplitudei significantly decreased (P=0.01), from 0.43 ± 0.07 mV to 0.20 ± 0.06 after 60 mins. The extracellular K+ buildup in this experiment thus showed a different dynamic than those observed following CBX and 4-AP microinjections.

9

ACCEPTED MANUSCRIPT Consistently with the observations of Boitano (Boitano and Evans, 2000) describing a timedependent effect of GAP27 peaking after 60min from application, we observed slow build up in [K+]o (Fig 1 A,B green traces). The [K+]o reached 90-mins following GAP27 application was comparable to the levels achieved by microinjections of CBX (9 ± 1 with GAP27 vs. 11 ± 3 mM following CBX, p=0.227, unpaired t-test) indicating that more specific ways of interfering with

PT

connexin channel function with a Cx43 mimetic peptide is still able to induce [K+]o elevation.

RI

To understand the role of Cx43 in K+ buffering, we next targeted Cx43 hemichannels by injecting

SC

the specific Cx43 hemichannel blocking peptide TAT-GAP19 (Wang et al., 2013b; Abudara et al.,

NU

2014). We did not observe a rise in [K+]ol (p=0.42, Figure 1A,B cyan) nor any difference in resting statem (p=0.144) or somatosensory evoked neuronal activityn (p=0.9, Figure 2 and 3 cyan).

MA

Increase in [K+]o is not sufficient to generate seizures

Our experiments show that the homeostasis of K+ is susceptible to GJ blockade and suggest that a

PT E

D

rise in [K+]o induced by GJ blockade is not sufficient to trigger seizures. To elucidate the role of [K+]o on seizure generation without the confounding effect of GJ blockade, we injected 50 mM [K+] solution into the parenchyma. The local increase in [K+]o (12.1 ± 2.3 mM, n=2) was transient and failed to

CE

elicit seizures. Figure 4A shows the effect of five 5ms-long spritzes: the [K+]o rapidly rose on average

AC

up to 13.1 mM and then returned to baseline within the following 10 seconds, showing the great power of local buffering mechanisms for [K+]o. We then opted for a more global approach and applied solutions of increasing [K+] onto the exposed dura to generate steady levels of raised [K +]o during which we could measure somatosensory evoked responses and neuronal power (Figure 4B and 4C). At baseline ([K+]o =1.5 ± 0.1 mM), forepaw stimulations elicited responses of 0.71 ± 0.1 mV, which significantly (p=0.0007) increasedr to 1.41 ± 0.2 mV at [K+]o = 7.9 ± 0.2 mM and then decreased to 0.31 ± 0.1 mV at [K+]o =18.2 ± 0.3 mM (n=8, Figure 4D). Analysis of neuronal power

10

ACCEPTED MANUSCRIPT however did not reveal significant increases in any of the bands with the exogenously raised [K+]o` (figure 4E, all bands p>0.46).

Discussion Connexin43 is a major regulator of [K+]o

PT

In the present work we explored the role of astrocytic connexin channels in regulating

RI

extracellular potassium concentration and the consequences of their blockade on seizure

SC

generation. Our data show that [K+]o is regulated mainly by astrocytic Cx43 GJs whose blockade increases [K+]o, compromising forepaw stimulated evoked local field responses. In contrast to GJ

NU

blockade, 4-AP administration generates a comparable rise in [K+]o as well as seizures (that were

MA

not preceeded by interictal events in the buildup phase of [K+]o). Elevated [K+]o alone is thus not sufficient to trigger seizures. The [K+]o peaks measured in vivo in our experiments are in line with

D

other observations of [K+]o elevations of 9-12 mM, noted in different pharmacological models of

PT E

seizures (for review see (Raimondo et al., 2015)). We found that the broad spectrum GJ blocker, Carbenoxolone, compromises K+-homeostasis, causing a rapid increase in [K+]o. However, the

CE

mechanism of this effect is difficult to identify because of the multiple sites of action of this agent (Thompson et al., 2008; Connors, 2012). Cx43 is the most common astrocytic connexin (Nagy and

AC

Rash, 2000; Brand-Schieber et al., 2005; Blomstrand and Giaume, 2006); moreover conditional KO of the Tuberous Sclerosis Complex-1 (TSC1) gene in astrocytes decreases Cx43 expression and determines the epileptic phenotype (Xu et al., 2009). Accordingly, we used the Gap27 peptide that is composed of a sequence on the second extracellular loop of Cx43. However, because of the conserved nature of the Gap27 sequence, it may also inhibit Cx30 channels. Given the slow dynamics of the [K+]o drift (examples in Figure 1A), we analyzed the acquired recordings at discrete delays following agent administration. The effect of Gap27 on [K+]o is likely to be the consequence of effects on inter-astrocytic GJs. Indeed, following GAP27 (but not CBX or 4-AP) administration,

11

ACCEPTED MANUSCRIPT we observed that at least 30 minutes were required to detect a significant increase in [K +]o (Figure 1 A,B), in agreement with the in vitro observations by Boitano (Boitano and Evans, 2000), who showed a peak of GAP27 activity being reached after 30-60 minutes. Gap27 also blocks Cx37 GJs in cultured vascular cells (Martin et al., 2005), making it possible that some of the effects of the peptide on [K+]o are mediated via inhibition of vascular clearance of K+. Ninety minutes following

PT

GAP27 administration, [K+]o rose to 8.90 ± 1.0 mM, while the [K+]o level reached at the same delay

RI

following CBX injection was 11 ± 3 mM, which is not significantly different from the Gap27 result.

SC

CBX is known to block Cx30 (Pannasch et al., 2014), whose expression levels are lower than those of Cx43 (Nagy and Rash, 2000; Blomstrand and Giaume, 2006). We presently perturbed GJs and

NU

hemichannels via the mimetic peptide, GAP27. One of the proposed mechanisms of action of this

MA

drug is its rapid (within a minute) blockade of plasma membrane hemichannels (Wang et al., 2012) which hinders the subsequent formation of GJs during the physiological turnover of these

D

proteins (Leybaert et al., 2003). In terms of a contribution of hemichannels, the observation that

PT E

Cx43 hemichannel blockade with TAT-Gap19 had no effect on [K+]o indicates that Cx43 hemichannels are not involved in [K+]o regulation. Indeed, the blockade of Cx43 hemichannels with

CE

TAT-GAP19 failed to show a comparable effect on [K+]o. This is probably because potassium redistribution relies on the inter-astrocytic syncytium which is based on GJs and that possibly does

AC

not require Cx43 hemichannels. However, this does not exclude that astrocytic Cx43 hemichannels may influence neuronal excitability (i.e. by facilitating glutamate release), but this requires further investigation in future studies. Although the use of astrocyte-specific knockout of Cx30 and Cx43 has been used for in vitro experiments (Wallraff et al. 2006), we chose not to do so because of potential developmental changes. Also our in vivo neocortical data show very robust responses to blockade of astrocytic gap junctional communication, suggesting that the dynamics of these

12

ACCEPTED MANUSCRIPT responses are greatly increased in vivo. Our findings thus provide the first in vivo evidence that astrocytic GJs play a key role in [K+]o regulation.

Effects of astrocitic GJs blockade on neuronal activity Regardless of the injected agent, the effects of the increased [K +]o were deleterious on synaptic

PT

transmission. At [K+]o above 8 mM, the amplitude of LFP responses to somatosensory stimulation

RI

decreased, as previously observed in vitro by Rausche (Rausche et al., 1990). This effect is probably

SC

due to the transition of Na+ channels to the inactivated state and their consequent inability to fire action potentials (i.e. depolarizing block, as reviewed in (Raimondo et al., 2015)). This conclusion is

NU

strengthened by the experiments with exogenous potassium: we indeed observe the peak in

MA

evoked responses for [K+]o between 6 and 8 mM, followed by a decrease for [K +]o larger than 10 mM. These observations constitute the first in vivo demonstration of K+ depolarizing block of

PT E

D

somatosensory transmission.

Rise in [K+]o and seizure generation

CE

In contrast to GJ blockade, 4-AP administration generates a comparable rise in [K+]o and seizures. 4-AP is a known broad-spectrum voltage-activated Potassium channel (Kv) blocker

AC

(Mathie et al., 1998; Judge and Bever, 2006). Local microinjection of 5 mM 4-AP in our experiments increased [K+]o and induced recurrent seizures. On the other hand, CBX and GAP27 had comparable effects on [K+]o without kindling seizures. The latter observation contrasts with the so called “potassium accumulation hypothesis” (Fertziger and Ranck, 1970) (Green, 1964), which posits that an increase in [K+]o above a certain threshold triggers a positive feedback loop that depolarizes neurons, enhancing their excitability and promoting further increases in [K +]o, ultimately generating seizures (for review see (Frohlich et al., 2008)). In line with our conclusion,

13

ACCEPTED MANUSCRIPT blockade of GJs in a 4-AP induced regimen of elevated [K+]o reduces ictal events and interictal synchronization, showing how electrical synapses are fundamental for seizure generation (Gigout et al 2006). Our data suggest that the [K+]o increase observed in pharmacological models of seizure (for review (Raimondo et al., 2015) is a result rather than a cause of cortical seizures, as has been suggested in vitro for the pilocarpine model (Kohling et al., 1995). To the best of our knowledge,

PT

our work constitutes the first in vivo demonstration that either a local or global increased [K+]o, of

RI

9 to 12 mM, can be reached without kindling neocortical seizures, indicating elevated [K+]o alone

SC

does not trigger seizures in the neocortex. This does however not exclude that increased [K+]o may still prolong the seizures, although in other experiments, we show that exogenously raised [K +]o

NU

above 9mM suppressed 4-AP induced neocortical seizures also in vivo in mice (Wang et al., 2016).

MA

It is however noteworthy that K is a major determinant of neuronal excitability and therefore even if increased [K+]o does not generate seizures (or interictal events) in the neocortex, it is still likely

D

to disrupt network activity. This stands in contrast to the acute hippocampal slice preparations

PT E

where [K+]o higher than 6.5 mM caused spontaneous recurrent seizures (Rutecki et al., 1985; Traynelis and Dingledine, 1988). Our observation is key for preclinical epilepsy studies employing

CE

4-AP for seizure induction, as the increased [K+]o can have a confounding effect. Alternative interpretations of convulsive effect of the 4-AP rely on its neurotransmitter release enhancing

AC

effect (Tapia and Sitges, 1982; Thomsen and Wilson, 1983) or in the synchronizing effect of its blockage on voltage-gated K+ channels. Indeed, it has been shown that Kv blockade increases the frequency of Ca2+ oscillations at the single neuronal (Alonso et al., 2010) and the population levels (Cao et al., 2014). In summary, blockade of neocortical inter-astrocytic Cx43-based gap junctional communication (i.e. [K+]o spatial buffering) causes a remarkable rise in [K+]o which depresses somatosensory evoked LFPs but does not cause seizures.

14

ACCEPTED MANUSCRIPT Acknowledgements

AC

CE

PT E

D

MA

NU

SC

RI

PT

The present work has been supported by CIHR and Brain Canada.

15

ACCEPTED MANUSCRIPT References

AC

CE

PT E

D

MA

NU

SC

RI

PT

Abudara V, Bechberger J, Freitas-Andrade M, De Bock M, Wang N, Bultynck G, Naus CC, Leybaert L, Giaume C (2014) The connexin43 mimetic peptide Gap19 inhibits hemichannels without altering gap junctional communication in astrocytes. Frontiers in cellular neuroscience 8:306. Alonso E, Vale C, Sasaki M, Fuwa H, Konno Y, Perez S, Vieytes MR, Botana LM (2010) Calcium oscillations induced by gambierol in cerebellar granule cells. Journal of cellular biochemistry 110:497-508. Avoli M, Barbarosie M, Lucke A, Nagao T, Lopantsev V, Kohling R (1996a), Synchronous GABA-Mediated Potentials and epileptiform discharges in the rat limbic system in vitro. The Journal of Neuroscience, 16:3912-3924. Avoli M, Louvel J, Kurcewicz, Pumain R, Barbarosie M (1996b) Extracellular free potassium and calcium during synchronous activity induced by 4-aminopyridine in the juvenile rat hippocampus. Journal of physiology 493,707-717. Avoli M, de Curtis M (2011) Gabaergic synchronization in the limbic system and its role in the generation of epileptiform activity. Progress in Neurobiology 95:104-132. Bazzigaluppi P, Dufour S, Carlen PL (2015) Wide field fluorescent imaging of extracellular spatiotemporal potassium dynamics in vivo. NeuroImage 104:110-116. Blomstrand F, Giaume C (2006) Kinetics of endothelin-induced inhibition and glucose permeability of astrocyte gap junctions. Journal of neuroscience research 83:996-1003. Bohmer C, Kirschner U, Wehner F (2001) 18-beta-Glycyrrhetinic acid (BGA) as an electrical uncoupler for intracellular recordings in confluent monolayer cultures. Pflugers Archiv : European journal of physiology 442:688-692. Boitano S, Evans WH (2000) Connexin mimetic peptides reversibly inhibit Ca(2+) signaling through gap junctions in airway cells. American journal of physiology Lung cellular and molecular physiology 279:L623-630. Brand-Schieber E, Werner P, Iacobas DA, Iacobas S, Beelitz M, Lowery SL, Spray DC, Scemes E (2005) Connexin43, the major gap junction protein of astrocytes, is down-regulated in inflamed white matter in an animal model of multiple sclerosis. Journal of neuroscience research 80:798-808. Cao Z, Cui Y, Busse E, Mehrotra S, Rainier JD, Murray TF (2014) Gambierol inhibition of voltage-gated potassium channels augments spontaneous Ca2+ oscillations in cerebrocortical neurons. The Journal of pharmacology and experimental therapeutics 350:615-623. Connors BW (2012) Tales of a dirty drug: carbenoxolone, gap junctions, and seizures. Epilepsy currents / American Epilepsy Society 12:66-68. De Bock M, Wang N, Bol M, Decrock E, Ponsaerts R, Bultynck G, Dupont G, Leybaert L (2012) Connexin 43 hemichannels contribute to cytoplasmic Ca2+ oscillations by providing a bimodal Ca2+-dependent Ca2+ entry pathway. The Journal of biological chemistry 287:12250-12266. Decrock E, De Vuyst E, Vinken M, Van Moorhem M, Vranckx K, Wang N, Van Laeken L, De Bock M, D’Herde K, Lai CP, Rogiers V, Evans WH, Naus CC, Leybaert L (2009) Connexin 43 hemichannels contribute to the propagation of apoptotic cell death in a rat C6 glioma cell model. Cell death and differentiation 16:151-163. Dufour S, Dufour P, Chever O, Vallee R, Amzica F (2011) In vivo simultaneous intra- and extracellular potassium recordings using a micro-optrode. Journal of neuroscience methods 194:206-217. Evans WH, Boitano S (2001) Connexin mimetic peptides: specific inhibitors of gap-junctional intercellular communication. Biochemical Society transactions 29:606-612. Evans WH, Leybaert L (2007) Mimetic peptides as blockers of connexin channel-facilitated intercellular communication. Cell communication & adhesion 14:265-273. Fertziger AP, Ranck JB, Jr. (1970) Potassium accumulation in interstitial space during epileptiform seizures. Experimental neurology 26:571-585. Florence G, Dahlem MA, Almeida AC, Bassani JW, Kurths J (2009) The role of extracellular potassium dynamics in the different stages of ictal bursting and spreading depression: a computational study. Journal of theoretical biology 258:219-228.

16

ACCEPTED MANUSCRIPT Frohlich F, Bazhenov M, Iragui-Madoz V, Sejnowski TJ (2008) Potassium dynamics in the epileptic cortex: new insights on an old topic. The Neuroscientist : a review journal bringing neurobiology, neurology and psychiatry 14:422-433. Giaume C, Leybaert L, Naus CC, Saez JC (2013) Connexin and pannexin hemichannels in brain glial cells: properties, pharmacology, and roles. Front Pharmacology 17; 4:88. Gigout S, Louvel J, Kawasaki H, D’Antuono M, Armand V, Kurcewicz I, Olivier A, Laschet J, Turak B, Devaux B, Pumain R, Avoli M (2006) Effects of gap junction blockers on human neocortical synchronization. Neurobiology of Disease 22:496-508

AC

CE

PT E

D

MA

NU

SC

RI

PT

Haas HL, Jefferys JG (1984) Low-calcium field burst discharges of CA1 pyramidal increas in rat hippocampal slices. The Journal of physiology 354:185-201. Heinemann U, Lux HD, Gutnick MJ (1977) Extracellular free calcium and potassium during increasing activity in the cerebral cortex of the cat. Experimental brain research 27:237-243. Jahromi SS, Wentlandt K, Piran S, and Carlen PL. (2002) Anticonvulsant actions of gap junctional blockers in an in vitro seizure model. J Neurophysiol 88(4):1893-1902 Jefferys JG, Haas HL (1982) Synchronized bursting of CA1 hippocampal pyramidal cells in the absence of synaptic transmission. Nature 300:448-450. Judge SI, Bever CT, Jr. (2006) Potassium channel blockers in multiple sclerosis: neuronal Kv channels and effects of symptomatic treatment. Pharmacology & therapeutics 111:224-259. Kofuji P, Newman EA (2004) Potassium buffering in the central nervous system. Neuroscience 129:10451056. Kohling R, Lucke A, Nagao T, Speckmann EJ, Avoli M (1995) Extracellular potassium elevations in the hippocampus of rats with long-term pilocarpine seizures. Neuroscience letters 201:87-91. Kunzelmann P., Schroder W., Traub O., Steinhauser C., Dermietzel R., and Willecke K. (1999) Late onset and 17ncreasing expression of the Gap Junction protein Connexin30 in adult murine brain and long-term cultured astrocytes. Glia 25:111-119. Leybaert L, Braet K, Vandamme W, Cabooter L, Martin PE, Evans WH (2003) Connexin channels, connexin mimetic peptides and ATP release. Cell communication & adhesion 10:251-257. Martin PE, Wall C, Griffith TM (2005) Effects of connexin-mimetic peptides on gap junction functionality and connexin expression in cultured vascular cells. British journal of pharmacology 144:617-627. Mathie A, Wooltorton JR, Watkins CS (1998) Voltage-activated potassium channels in mammalian neurons and their block by novel pharmacological agents. General pharmacology 30:13-24. Moody WJ, Futamachi KJ, Prince DA (1974) Extracellular potassium activity during epileptogenesis. Experimental neurology 42:248-263. Nagy JI, Rash JE (2000) Connexins and gap junctions of astrocytes and oligodendrocytes in the CNS. Brain research Brain research reviews 32:29-44. Newman EA (1984) Regional specialization of retinal glial cell membrane. Nature 309:155-157. Newman EA (1993) Inward-rectifying potassium channels in retinal glial (Muller) cells. The Journal of neuroscience : the official journal of the Society for Neuroscience 13:3333-3345. Orkand RK, Nicholls JG, Kuffler SW (1966) Effect of nerve impulses on the membrane potential of glial cells in the central nervous system of amphibia. Journal of neurophysiology 29:788-806. Pannasch U, Freche D, Dallerac G, Ghezali G, Escartin C, Ezan P, Cohen-Salmon M, Benchenane K, Abudara V, Dufour A, Lubke JH, Deglon N, Knott G, Holcman D, Rouach N (2014) Connexin 30 sets synaptic strength by controlling astroglial synapse invasion. Nature neuroscience 17:549-558. Pedley TA, Fisher RS, Moody WJ, Futamachi KJ, Prince DA (1974) Extracellular potassium activity during epileptogenesis: a comparison between neocortex and hippocampus. Transactions of the American Neurological Association 99:41-45. Raimondo JV, Burman RJ, Katz AA, Akerman CJ (2015) Ion dynamics during seizures. Frontiers in cellular neuroscience 9:419.

17

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

Rausche G, Igelmund P, Heinemann U (1990) Effects of changes in extracellular potassium, magnesium and calcium concentration on synaptic transmission in area CA1 and the dentate gyrus of rat hippocampal slices. Pflugers Archiv : European journal of physiology 415:588-593. Reddy DS, Kuruba R (2013) Experimental models of status epilepticus and neuronal injury for evaluation of therapeutic interventions. International journal of molecular sciences 14:18284-18318. Rouach N, Segal M, Koulakoff A, Giaume C, Avignone E (2003) Carbenoxolone blockade of neuronal network activity in culture is not mediated by an action on gap junctions. The Journal of physiology 553:729-745. Rutecki PA, Lebeda FJ, Johnston D (1985) Epileptiform activity induced by changes in extracellular potassium in hippocampus. Journal of neurophysiology 54:1363-1374. Sagar GD, Larson DM (2006) Carbenoxolone inhibits junctional transfer and upregulates Connexin43 expression by a protein kinase A-dependent pathway. Journal of cellular biochemistry 98:15431551. Samoilova M, Wentlandt K, Adamchik Y, Velumian AA, Carlen PL (2008) Connexin 43 mimetic peptides inhibit spontaneous epileptiform activity in organotypic hippocampal slice cultures. Experimental neurology 210:762-775. Scemes E, Spray DC (2012) Extracellular K(+) and astrocyte signaling via connexin and pannexin channels. Neurochemical research 37:2310-2316. Sohl G, Maxeiner S, Willecke K (2005) Expression and functions of neuronal gap junctions. Nature reviews Neuroscience 6:191-200. Somjen GG (2001) Mechanisms of spreading depression and hypoxic spreading depression-like depolarization. Physiol Rev 81:1065-1096. Tapia R, Sitges M (1982) Effect of 4-aminopyridine on transmitter release in synaptosomes. Brain research 250:291-299. Taylor CP, Dudek FE (1982) Synchronous neural afterdischarges in rat hippocampal slices without active chemical synapses. Science 218:810-812. Thompson RJ, Jackson MF, Olah ME, Rungta RL, Hines DJ, Beazely MA, MacDonald JF, MacVicar BA (2008) Activation of pannexin-1 hemichannels augments aberrant bursting in the hippocampus. Science 322:1555-1559. Thomsen RH, Wilson DF (1983) Effects of 4-aminopyridine and 3,4-diaminopyridine on transmitter release at the neuromuscular junction. The Journal of pharmacology and experimental therapeutics 227:260-265. Tong X, Ao Y, Faas GC, Nwaobi SE, Xu J, Haustein MD, Anderson MA, Mody I, Olsen ML, Sofroniew MV, Khakh BS (2014) Astrocyte Kir4.1 ion channel deficits contribute to neuronal dysfunction in Huntington’s disease model mice. Nature neuroscience 17:694-703. Traynelis SF, Dingledine R (1988) Potassium-induced spontaneous electrographic seizures in the rat hippocampal slice. Journal of neurophysiology 59:259-276. Vessey JP, Lalonde MR, Mizan HA, Welch NC, Kelly ME, Barnes S (2004) Carbenoxolone inhibition of voltage-gated Ca channels and synaptic transmission in the retina. Journal of neurophysiology 92:1252-1256. Wallraff A, Kohling R, Heinemann U, Theis M, Willecke K, Steinhauser C (2006) The impact of astrocytic gap junctional coupling on potassium buffering in the hippocampus. The Journal of neuroscience : the official journal of the Society for Neuroscience 26:5438-5447. Wang L., Dufour S, Valiante TA, CarlenPL. (2016) Extracellular Potassium and seizures: excitation, inhibition and the role of Ih. International Journal of Neural Systems. In press Wang N, De Bock M, Decrock E, Bol M, Gadicherla A, Bultynck G, Leybaert L (2013a) Connexin targeting peptides as inhibitors of voltage- and intracellular Ca2+-triggered Cx43 hemichannel opening. Neuropharmacology 75:506-516. Wang N, De Bock M, Antoons G, Gadicherla AK, Bol M, Decrock E, Evans WH, Sipido KR, Bukauskas FF, Leybaert L (2012) Connexin mimetic peptides inhibit Cx43 hemichannel opening triggered by voltage and intracellular Ca2+ elevation. Basic research in cardiology 107:304. Wang N et al. (2013b) Selective inhibition of Cx43 hemichannels by Gap19 and its impact on myocardial ischemia/reperfusion injury. Basic research in cardiology 108:309.

18

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

Xu L, Zeng LH, Wong M (2009) Impaired astrocytic gap junction coupling and potassium buffering in a mouse model of tuberous sclerosis complex. Neurobiology of disease 34:291-299.

19

ACCEPTED MANUSCRIPT Legends Figure 1. Effects of different blockers on extracellular K+. A) Example traces showing the time

RI

PT

course of [K+]o following injection (black arrowhead) of each of the screened drugs. Injection of vehicle (PBS, blue trace) or the Cx43 hemichannel blocker (TAT-GAP19, cyan trace) did not affect [K+]o. Conversely, the general GJ blocker (CBX, black trace), the Cx43 GJ blocker (GAP27, green trace), and 4-AP (brown trace) progressively raised [K+]o. Following 4-AP administration, spontaneous recurring seizures were observed with transient “spikes” in [K +]o. B) Population data from the same experiment in A, with points representing the group average ± SEM at discrete delays following an agent’s administration. Legend is the same as in A. Thirty minutes following injection of CBX, GAP27, or 4-AP, [K+]o showed a significant increase (P<0.01) relative to the respective pre-drug administration baseline.

SC

Figure 2. Effects of extracellular K+ on neuronal activity. A) Example LFP recordings before

MA

NU

(‘Baseline’) and 90 minutes after injection (‘90 min’) of the screened drug: Ai: vehicle (PBS), Aii: GAP27, Aiii: TAT-GAP19, Aiv: CBX, Av: 4-AP. B) Examples of power spectra at baseline (blue) and after 90 minutes (red) of the traces in Aiv (CBX) and Av (4-AP). C) Population graph showing the effects of increasing [K+]o levels measured at different time points, on the somatosensory evoked LFP. Injection of CBX, GAP27 and 4-AP resulted in decreased amplitude of the LFP for [K+]o above 8mM.

Figure 3. Frequency bands population histogram. Power is normalized to baseline. Only 4-AP

D

produced a significant (P<0.01) increase in power at 90 minutes after injection relative to baseline.

PT E

Figure 4. Extracellular K+ alone does not cause seizures. A) Representative experiment of intra-

AC

CE

cortical K+ spritz, five repetitions of 5ms spritz are shown in black and their average in cyan: [K +]o is rapidly cleared from the extracellular space after a short increase; Example experiments showing the modulation of global [K+]o increases on somatosensory evoked responses (B, ten repetitions of somatosensory evoked responses in gray and their average in red) and neuronal power (C). Population data (n=8) shows increased somatosensory evoked responses for [K +]o up to 8 mM followed by decreased responses amplitudes for [K+]o greater than 12 mM (p=0.0007). Power analysis of neuronal activity (E) shows the absence of seizures.

20

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

21

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

22

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

23

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

FIGURE 4

24

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

Highlights: astrocytic Connexin43 play a key role in extracellular K homeostasis in vivo; + connexin43 specific blockade raises extracellular K concentration ([K ]e ); + rise in [K ]e alone is not sufficient to trigger seizures in the neocortex; + exogenous increase in [K ]e does not generate seizures.

25