2024.02.02.578550v1.full-html.html

Chemogenetic silencing reveals presynaptic G

i/o

protein-mediated

inhibition of synchronized activity in the developing hippocampus

in vivo

Jürgen Graf

1,4

, Arash Samiee

2,4

, Tom Flossmann

1,3

, Knut Holthoff

1,5

and Knut Kirmse

2,5

∗

Department of Neurology, Jena University Hospital, 07747 Jena, Germany

Department of Neurophysiology, Institute of Physiology, University of Würzburg, 97070

Würzburg, Germany

Institute of Physiology I, Jena University Hospital, 07743 Jena, Germany

These authors contributed equally.

Senior authors

∗

Correspondence to:

knut.kirmse@uni-wuerzburg.de

Abbreviated title:

Presynaptic G

i/o

constrains synchronized activity

Number of Pages – 46

Number of Figures – 7

Number of Extended Data items – 1

Number of words in Abstract – 225

Number of words in Introduction – 518

Number of words in Discussion – 1,159

Conflict of interest statement:

The authors declare no competing financial interests.

Acknowledgments:

This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research

Foundation – Research Grants: KI 1816/6-1 #442107075, KI 1816/7-1 #448069679 to K.K.;

HO 2156/5-1 #442107075, HO 2156/6-1 #448069679 to K.H.; Research Unit 3004: KI

1816/5-1 #432559020, KI 1816/9-1 #415914819 to K.K.; Major Research Instrumentation:

INST 93/1103-1 FUGG #502670664 to K.K.) and the Thüringer Aufbaubank (2017 FGI 0020

to K.H.). We thank Ina Ingrisch and Maria Oppmann for excellent technical assistance.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Abstract

The development of neuronal circuits is an activity-dependent process. Research aimed at

deciphering the learning rules governing these developmental refinements is increasingly

utilizing a class of chemogenetic tools that employ G

i/o

protein signaling (G

-DREADDs) for

neuronal silencing. However, their mechanisms of action and inhibitory efficacy in immature

neurons with incompletely developed G

i/o

signaling are poorly understood. Here, we analyze

the impact of G

i/o

signaling on cellular and network excitability in the neonatal hippocampus by

expressing the G

-DREADD hM4Di in telencephalic glutamatergic neurons of neonatal mice of

both sexes. Using acousto-optic two-photon Ca

imaging, we report that activation of hM4Di

leads to a complete arrest of spontaneous synchronized activity in CA1

in vitro

Electrophysiological analyses demonstrate that hM4Di-mediated silencing is not accounted for

by changes in intrinsic excitability of CA1 pyramidal cells (PCs). Instead, activation of hM4Di

robustly restrains synaptic glutamate release by the first postnatal week, effectively reducing

recurrent excitation in CA1.

In vivo

, inhibition through hM4Di potently suppresses early sharp

waves (eSPWs) and discontinuous oscillatory network activity in CA1 of head-fixed mice before

eye opening. In summary, hM4Di dampens glutamatergic neurotransmission through a

presynaptic mechanism sufficient to terminate spontaneous synchronized activity in the

neonatal CA1. Our findings have implications for designing and interpreting experiments

utilizing G

-DREADDs in immature neurons and further point to a potential role of G

i/o

-dependent

endogenous neuromodulators in activity-dependent hippocampal development.

Keywords:

development, presynaptic inhibition, DREADD, chemogenetics,

in vivo

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Introduction

The immature rodent hippocampus displays spontaneous synchronized activity prior to the

developmental onset of spatial memory and environmental exploration. In CA1 during the first

and second postnatal week, activity is characterized by intermittent network bursts that

alternate with periods of low activity. Network bursts are driven by input from the sensory

periphery (Mohns and Blumberg, 2010; Valeeva et al., 2019; Chen et al., 2023; Gainutdinov et

al., 2023; Kostka and Hanganu-Opatz, 2023), but they can also be internally generated within

the entorhinal-hippocampal formation (Ben-Ari et al., 1989; Leinekugel et al., 2002; Graf et al.,

2021). By co-activating large numbers of neurons (Dard et al., 2022; Graf et al., 2022), network

bursts were proposed to inform nascent CA1 circuits about the fundamental statistical

properties of the external (environment) and internal (body) world, thereby preparing them for

later cognitive and sensory demands (for review see Cossart and Khazipov, 2022; Kirmse and

Zhang, 2022).

To elucidate the learning rules underlying activity-dependent refinements in brain development,

research is increasingly utilizing chemogenetic strategies, as they allow neuronal activity to be

manipulated in a temporally controlled and cell type-specific manner (Donato et al., 2017; Wong

et al., 2018; Murata and Colonnese, 2020; Leighton et al., 2021; Maldonado et al., 2021; Chen

et al., 2023; Kostka and Hanganu-Opatz, 2023; Leprince et al., 2023). The most widely used

inhibitory chemogenetic actuator is hM4Di, which is based on the human muscarinic M4

receptor employing G

i/o

signaling for neuronal silencing (Armbruster et al., 2007). In mature

neurons, endogenous neuromodulators activating G

i/o

signaling reduce excitability by multiple

mechanisms. In somatodendritic compartments, G

i/o

mediates hyperpolarization due to

activation of G-protein-coupled inwardly rectifying K

(GIRK/Kir3) channels (Reuveny et al.,

1994) and inhibition of Ca

1 channels involved in dendritic spike generation (Perez-Garci et al.,

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

2006; Perez-Garci et al., 2013). In addition, G

i/o

βγ

dimers inhibit evoked and spontaneous

vesicle release by interacting with SNARE proteins and Ca

2 channels in presynaptic terminals

(Herlitze et al., 1996; Sakaba and Neher, 2003; Zurawski et al., 2019; Alten et al., 2022). In a

manner analogous to that of endogenous G

i/o

-protein-coupled receptors, G

-DREADDs are

known to effectively silence neuronal activity via both hyperpolarization and suppression of

neurotransmitter release in the adult brain (Stachniak et al., 2014; Vardy et al., 2015). This,

however, might fundamentally differ from the situation in the neonatal neocortex and

hippocampus, when GIRK channel expression is low (Chen et al., 1997; Ehrengruber et al.,

1997; Bony et al., 2013) and the G

i/o

-dependent hyperpolarization (Luhmann and Prince, 1991;

Nurse and Lacaille, 1999; Verheugen et al., 1999) and inhibition of voltage-gated Ca

channels

(Gaiarsa et al., 1995) are not yet operative.

Here, we address this issue by analyzing the single-neuron mechanisms and circuit-level

efficacy of hM4Di-mediated G

i/o

-dependent silencing in the hippocampal CA1 region of

developing mice

in vitro

and

in vivo

. We focus on the first and second postnatal weeks, when

network activity comprises intermittent bursts that are generated in a relatively all-or-none

manner (Prida and Sanchez-Andres, 1999; Flossmann et al., 2019), rendering inhibition

particularly challenging. We demonstrate that the activation of hM4Di-induced G

i/o

signaling can

100

arrest spontaneous neuronal synchrony through the presynaptic silencing of glutamatergic

101

neurotransmission.

102

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Results

103

i/o

protein-mediated silencing of synchronized network activity in vitro

104

To examine the efficacy of G

i/o

protein-coupled chemogenetic silencing during development,

105

we conditionally expressed hM4Di in telencephalic glutamatergic neurons from embryonic

106

stages onward (

Emx1

IREScre

::hM4Di

LSL

hM4Di

Emx1

mice) (Gorski et al., 2002; Zhu et al., 2016;

107

Graf et al., 2021). Of note, both CA1 pyramidal cells (PCs) and the vast majority of

108

glutamatergic synapses in CA1 are derived from the

Emx1

lineage (Sando et al., 2017; Leprince

109

et al., 2023). We first used three-dimensional acousto-optic two-photon Ca

imaging in CA1

110

stratum pyramidale

of acute brain slices to quantify network activity at single-cell resolution at

111

postnatal days (P) 2–5 (

Fig. 1

A-B). Under control conditions, spontaneous activity in CA1 PCs

112

consisted of discrete network events, known as giant depolarizing potentials (GDPs) (Ben-Ari

113

et al., 1989), that alternated with extended periods of low activity (

Fig. 1

C). Activation of

114

hM4Di

Emx1

by the DREADD agonist C21 (1 µM) led to a profound reduction in the mean

115

frequency of somatic Ca

transients (CaTs) from 1.4

0.4 min

-1

to 0.0

0.0 min

-1

(n = 5 slices;

116

Fig. 1

B-E). To control for potential DREADD-independent effects of C21, we repeated these

117

experiments in wild-type (WT) mice lacking

hM4Di

Emx1

(n = 6 slices;

Fig. 1

D-E). Generalized

118

linear mixed-effects modeling (GLMM) revealed a significant interaction term, indicating that

119

C21 effects in slices obtained from

hM4Di

Emx1

mice were larger than in those from WT neonates

120

(genotype × treatment: P = 3.5

-6

; for detailed statistics, see Extended Data

Table 1-1

121

Post-hoc

testing confirmed that the suppression of neuronal activity by C21 was restricted to

122

hM4Di

Emx1

mice, whereas C21 was ineffective in WT pups (

hM4Di

Emx1

: P = 3.2

-3

, WT: P =

123

0.16, simple contrasts;

Fig. 1

E). Likewise, at the network level, C21-induced inhibition

124

abolished GDPs in slices from

hM4Di

Emx1

but not from WT neonates (

Fig. 1

F). Collectively,

125

these data demonstrate effective network silencing through hM4Di

Emx1

and identify G

i/o

126

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

signaling as a potential inhibitory actuator of synchronized network activity in the developing

127

hippocampus

in vitro

128

129

hM4Di

Emx1

-mediated silencing is not accounted for by changes in intrinsic excitability

130

i/o

protein-coupled chemogenetic actuators are commonly used tools for neuronal silencing in

131

the adult, as they can reduce spike rates through hyperpolarization and shunting, causing

132

subtractive and divisive inhibition (Roth, 2016). However, these mechanisms may be less

133

effective in the neonatal brain, as the coupling of somatodendritic GIRK channels to G

i/o

134

signaling is absent or weak during the perinatal period (Fukuda et al., 1993; Chen et al., 1997;

135

Nurse and Lacaille, 1999; Bony et al., 2013). We therefore set out to examine the mechanism(s)

136

of hM4Di

Emx1

-dependent inhibition in the neonatal hippocampus by performing somatic current-

137

clamp recordings. At P2–5, we found that bath-application of C21 (1 µM) did not significantly

138

affect the resting membrane potential (RMP) of CA1 PCs in either

hM4Di

Emx1

or WT mice

139

(

Fig. 2

A-B) (GLMM; for detailed statistics, see Extended Data

Table 1-1

). We noticed a minor

140

temporal drift in membrane resistance (R

) during these prolonged recordings. However, this

141

time-dependent change was independent of genotype (interaction: P = 0.30;

Fig. 2

A and 2C),

142

indicating that hM4Di

Emx1

activation left R

unaffected. We further explored potential effects of

143

hM4Di

Emx1

activation on intrinsic excitability during the second postnatal week (

Fig. 2

D), when

144

active and passive membrane properties are considerably more mature but burst-like network

145

activity is still prominent

in vivo

(Graf et al., 2022). We found that, at P8–12, both RMP and R

146

were unaltered by bath-application of C21 (

Fig. 2

E-F). In summary, activation of hM4Di

Emx1

did

147

not induce detectable hyperpolarization or shunting in CA1 PCs before eye opening.

148

We next sought to extend these findings by examining the input-output relationship and

149

characteristics of action-potential (AP) firing in CA1 PCs through current-clamp recordings in

150

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

slices from

hM4Di

Emx1

and WT mice at P2–5 (

Fig. 3

A-C). In each recorded PC, we applied three

151

series of depolarizing step-current injections in the absence and presence of C21 (1 µM).

152

Immediately prior to each experiment, the RMP was biased to

−

70 mV by somatic current

153

injection so as to exclude any contribution of differences in RMP on the analyzed properties.

154

We found that the maximum instantaneous firing rates of CA1 PCs were unaffected by C21 in

155

slices from both

hM4Di

Emx1

and WT mice (

Fig. 3

D; for detailed statistics, see Extended Data

156

Table 1-1

). During these experiments, maximum mean firing rates (per 500-ms-long current

157

step) showed a minor decline over time, which, however, did not significantly differ between

158

hM4Di

Emx1

and WT mice (interaction: P = 0.052;

Fig. 3

E). In either genotype, C21 lacked

159

significant effects on AP thresholds (

Fig. 3

F) or rheobase currents (

Fig. 3

G). We then repeated

160

these experiments during the second postnatal week (

Fig. 3

H-J). As compared to neurons of

161

the first postnatal week, CA1 PCs recorded at P8–12 exhibited substantially higher maximum

162

instantaneous (

Fig. 3

K) and mean (

Fig. 3

L) firing rates, more negative AP thresholds (

Fig. 3

163

and higher rheobase currents (

Fig. 3

N), implying that membrane properties had undergone a

164

considerable maturation within the period investigated here. In line with our data obtained at

165

P2–5, however, GLMM indicated that none of the analyzed quantities was significantly affected

166

by C21 (

Fig. 3

K-N and Extended Data

Table 1-1

167

Taken together, our data indicate that the observed hM4Di

Emx1

-mediated silencing of network

168

activity in the neonatal hippocampus (

Fig. 1

) is unlikely to result from changes in intrinsic

169

excitability of CA1 PCs and, thus, point to an alternative mechanism.

170

171

hM4Di

Emx1

activation potently suppresses synaptic glutamate release by the first

172

postnatal week

173

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Presynaptic G

i/o

signaling is functional already at early developmental stages (Fukuda et al.,

174

1993; Safiulina et al., 2005; Kirmse et al., 2008). Moreover, suppression of vesicular release

175

was previously identified as a major mechanism of G

i/o

-based chemogenetic actuators in the

176

adult brain (Stachniak et al., 2014; Vardy et al., 2015). We therefore investigated whether

177

inhibition of synaptic transmission underlies the hM4Di

Emx1

-mediated silencing at the network

178

level. To this end, we first recorded miniature excitatory postsynaptic currents (mEPSCs) in the

179

presence of blockers of voltage-gated Na

channels (tetrodotoxin, 0.5 µM) and GABA

180

receptors (gabazine, 10 µM) (

Fig. 4

A). Of note, most neurons providing glutamatergic input to

181

CA1 are derived from the

Emx1

lineage (Sando et al., 2017; Graf et al., 2021; Leprince et al.,

182

2023). We found that, in CA1 PCs recorded from

hM4Di

Emx1

mice at P2–5, C21 (1 µM) reduced

183

the frequency of mEPSCs from 0.13

0.03 min

−

to 0.05

0.01 min

−

(

Fig. 4

B). Conversely,

184

mEPSC frequency was unaffected in WT neurons, thereby excluding the possibility that hM4Di-

185

independent actions of C21 could underlie the effect (interaction: P = 3.5

−

;

hM4Di

Emx1

: P

186

= 1.1

−

, WT: P = 0.14, simple contrasts;

Fig. 4

B and Extended Data

Table 1-1

). In addition,

187

median mEPSC amplitudes, a measure of quantal size, remained unaltered after wash-in of

188

C21 in either genotype (

Fig. 4

C), pointing to a presynaptic site of action of hM4Di

Emx1

. Again,

189

we extended these investigations to the second postnatal week (

Fig. 4

D). At P8–12, akin to the

190

above observations, C21 reduced mEPSC frequencies by about 50% selectively in

hM4Di

Emx1

191

mice (

Fig. 4

E), without decreasing median mEPSC amplitudes (interaction: P = 0.056;

Fig. 4

F).

192

Collectively, our data suggest that activation of hM4Di

Emx1

leads to inhibition of vesicle release

193

at glutamatergic synapses impinging on CA1 PCs, which in turn can explain the observed

194

suppression of synchronized network activity.

195

To corroborate this conclusion, we next examined to what extent G

i/o

signaling controls

196

glutamate release at Schaffer collateral synapses, which provide a major excitatory drive to

197

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

CA1. To this end, EPSCs evoked by electrical stimulation of afferents in

stratum radiatum

198

(eEPSCs) were recorded from CA1 PCs of

hM4Di

Emx1

mice in the presence of gabazine (10 µM)

199

(

Fig. 5

A). At P2–5, bath-application of C21 (1 µM) was found to massively decrease mean

200

eEPSC amplitudes from 34.1

8.5 pA to 1.7

1.3 pA (

Fig. 5

B). This was associated with an

201

increase in eEPSC failure rate from 25

6% to 92

4% (

Fig. 5

C), supporting the conclusion

202

that the effect is presynaptic in origin. In these experiments, paired-pulse stimulation frequently

203

induced long-latency polysynaptic bursts of EPSCs, reflecting network disinhibition due to the

204

blockade of GABA

receptors (arrows in

Fig. 5

A). We therefore considered the possibility that

205

these EPSC bursts

per se

could change synaptic strength, i.e. independently of C21. To

206

address this point, we performed an additional set of experiments, in which vehicle was applied

207

instead of C21 (

Fig. 5

A). We found that both mean amplitudes (

Fig. 5

B) and failure rates

208

(

Fig. 5

C) of eEPSCs were stable over time. For each parameter analyzed, GLMM revealed a

209

significant interaction between treatment and dataset (eEPSC amplitude: P = 5.3

−

;

210

eEPSC failure rate: P = 2.9

−

), and

post-hoc

tests confirmed that the suppression of

211

eEPSCs was specific to C21 (

Fig. 5

B-C; for detailed statistics, see Extended Data

Table 1-1

212

Likewise, at P8–12, activation of hM4Di

Emx1

by bath-application of C21 profoundly reduced

213

mean eEPSC amplitudes and increased eEPSC failure rates, both of which were not observed

214

when vehicle was applied instead of C21 (

Fig. 5

D-F). Taken together, these data demonstrate

215

that activation of G

i/o

signaling through hM4Di

Emx1

potently suppresses synaptic glutamate

216

release at Schaffer collateral synapses on CA1 PCs by the first postnatal week.

217

218

Presynaptic G

i/o

signaling is an actuator of immature hippocampal network dynamics in

219

vivo

220

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Our data obtained so far indicate that the hM4Di

Emx1

-mediated inhibition of synaptic glutamate

221

release alone suffices to virtually abolish GDPs in acute slices and therefore point to

222

presynaptic G

i/o

signaling as an actuator of neuronal synchrony in the developing hippocampus.

223

However, the

in vivo

relevance of these findings remains unclear. We therefore performed local

224

field potential (LFP) recordings in

stratum radiatum

of CA1 in head-fixed mice. Experiments

225

were conducted under nitrous oxide anesthesia at P3–5. At this developmental stage, CA1 PCs

226

are incapable of sustaining persistent firing activity but instead generate intermittent population

227

bursts reminiscent of GDPs in slices (Leinekugel et al., 2002; Dard et al., 2022; Graf et al.,

228

2022). In accordance with previous studies, LFP activity at P3–5 was discontinuous and

229

dominated by early sharp waves (eSPWs;

Fig. 6

A-B) (Valeeva et al., 2019; Gainutdinov et al.,

230

2023). Less consistently, we also detected short-lasting LFP oscillations in the theta-to-beta

231

frequency range in several of the analyzed mice (

Fig. 6

A) (Graf et al., 2021). For quantification

232

of hM4Di

Emx1

effects, we computed the LFP bandpower (8–40 Hz) in non-overlapping 5-min-

233

long intervals and normalized it to the bandpower of the technical noise (see Methods for

234

details). In

hM4Di

Emx1

mice, a single dose of systemically applied C21 (3 mg/kg s.c.) reduced

235

the normalized bandpower to approximately noise levels, implying an almost complete

236

suppression of LFP activity (

Fig. 6

C-D). This effect set in within minutes, confirming the

237

favorable bioavailability of C21 when administered through subcutaneous injection in neonatal

238

mice (Thompson et al., 2018; Jendryka et al., 2019; Kostka and Hanganu-Opatz, 2023). The

239

suppression of LFP activity lasted for at least two hours (

Fig. 6

C) and was strictly dependent

240

on activation of hM4Di

Emx1

, since no decrease in normalized bandpower was observed upon

241

application of either vehicle in

hM4Di

Emx1

pups or C21 in WT mice at 30–60 min

post

injection

242

(interaction: P = 2.5

−

). Likewise, eSPWs were virtually abolished after injection of C21 in

243

hM4Di

Emx1

mice (control: 2.1

0.3 min

−

, C21: 0.0

0.0 min

−

;

Fig. 6

B and 6E; for detailed

244

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

statistics, see Extended Data

Table 1-1

). While the normalized bandpower in WT mice

245

remained unaffected, C21 reduced eSPW rates, although to a lesser extent than in

hM4Di

Emx1

246

pups.

247

We finally asked whether hM4Di

Emx1

-induced presynaptic inhibition of synaptic glutamate

248

release can effectively silence network activity at P10–12. By this time, discrete eSPW events

249

had been largely replaced by more complex, mostly discontinuous, oscillatory LFP activity

250

patterns in the theta-beta range (

Fig. 7

A), reflecting a profound developmental increase in

251

synaptic inputs to CA1. In

hM4Di

Emx1

mice, subcutaneous injection of C21 reduced the

252

normalized bandpower by approximately one order of magnitude, with little recovery during the

253

subsequent two-hour recording period (

Fig. 7

A and 7C). Maximum suppression was already

254

present 15 min after injection. Conversely, no apparent changes in normalized bandpower were

255

evident upon application of either vehicle in

hM4Di

Emx1

pups or C21 in WT mice (

Fig. 7

B-C).

256

Accordingly, GLMM revealed a significant interaction term (dataset

treatment: P = 2.1

−

257

and

post-hoc

testing confirmed that the suppression of neuronal activity by C21 was restricted

258

hM4Di

Emx1

mice (

Fig. 7

C; for detailed statistics, see Extended Data

Table 1-1

). Taken

259

together, our data demonstrate that hM4Di

Emx1

-induced synaptic suppression of glutamate

260

release can impose powerful inhibition on synchronized network activity in the developing

261

hippocampus before eye opening.

262

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Discussion

263

Generative mechanisms of network bursts in the developing CA1

264

Network bursts in the developing hippocampus are generated in a relatively all-or-none manner

265

(Prida and Sanchez-Andres, 1999; Gainutdinov et al., 2023). Theoretical work suggests that

266

this behavior arises from a transient, unstable network state evoked by supra-threshold

267

excitatory input, leading to a profound amplification of neuronal firing rates (Rahmati et al.,

268

2017). To further characterize the generative mechanisms underlying synchronized activity, we

269

here employed a transgenic approach (

hM4Di

Emx1

mice) (Gorski et al., 2002; Zhu et al., 2016)

270

to target the entire population of

Emx1

-lineage cells from embryonic stages onwards. We

271

examined the effect of hM4Di

Emx1

activation

in vitro

and found that it abolishes GDPs at P2–5

272

(

Fig. 1

). Likewise, a single systemic injection of the DREADD agonist C21 completely blocked

273

eSPWs as observed in

in vivo

LFP recordings from neonatal

hM4Di

Emx1

mice (

Fig. 6

). Thus,

274

glutamatergic neurons of the

Emx1

-lineage contribute most to the transient instability

275

underlying the supra-threshold amplification of firing rates during network bursts in CA1. This

276

further implies that neither GDPs

in vitro

nor eSPWs

in vivo

are sustained by depolarizing

277

GABAergic (

Emx1

–

) neurons alone, despite their partially excitatory action (Flossmann et al.,

278

2019; Murata and Colonnese, 2020; Graf et al., 2021). In addition, our data are fully compatible

279

with the previous conclusion that somatosensory feedback, in the form of eSPWs, is relayed to

280

CA1 via

Emx1

glutamatergic projections from the entorhinal cortex (Valeeva et al., 2019;

281

Leprince et al., 2023), whereas glutamatergic input from

Emx1

–

sources, such as from the

282

ventral midline thalamus (Leprince et al., 2023), modulates rather than sustains synchronized

283

activity in developing CA1. Similar considerations apply to the second postnatal week, when

284

short-lasting events (i.e., eSPWs) had been largely replaced by discontinuous oscillatory LFP

285

activity (

Fig. 7

), consistent with the single-cellular dynamics of network activity in developing

286

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

CA1

in vivo

(Graf et al., 2022). In methodological terms, our data demonstrate that transgenic

287

expression of hM4Di enables effective suppression of spontaneous synchronized activity

288

already shortly after birth, positioning it as a valuable and non-invasive alternative to viral

289

transfection or

in utero

electroporation for developmental studies (Zhu et al., 2016).

290

291

Cellular mechanism and efficacy of hM4Di

Emx1

in neonatal CA1 neurons

292

-DREADDs including hM4Di are normally localized to both neuronal cell bodies and axons

293

(Mahler et al., 2014). Accordingly, in the adult brain, G

-DREADD-based neuronal silencing

294

results from the combined effects of (i) hyperpolarization and increased membrane

295

conductance in somatodendritic domains and the (ii) suppression of vesicle release from

296

presynaptic terminals (Armbruster et al., 2007; Stachniak et al., 2014; Vardy et al., 2015). We

297

here examined the efficacy of hM4Di

Emx1

-mediated silencing in immature hippocampal neurons

298

during the neonatal period, when GIRK channel expression is low (Chen et al., 1997;

299

Ehrengruber et al., 1997) and endogenous G

i/o

-protein-coupled receptors are ineffective in

300

mediating hyperpolarization in CA1 neurons (Nurse and Lacaille, 1999; Verheugen et al., 1999).

301

Unlike what is observed in adult neurons, we discovered that activation of hM4Di

Emx1

did not

302

induce hyperpolarization or shunting (

Fig. 2

), thereby leaving evoked AP firing in somatic

303

current-clamp recordings unaffected (

Fig. 3

). Conversely, activation of hM4Di

Emx1

effectively

304

suppressed glutamatergic neurotransmission, resembling previous observations in adult

305

neurons. This synaptic silencing is likely of presynaptic origin, since hM4Di

Emx1

caused a

306

selective reduction in the frequency, but not amplitude, of mEPSCs (

Fig. 4

) as well as a

307

substantial increase in the failure rate of eEPSCs (

Fig. 5

). The observed dichotomy supposedly

308

reflects differences in the signaling cascades engaged by G

i/o

βγ dimers, which, in presynaptic

309

terminals, inhibit evoked and spontaneous (miniature) release chiefly by interacting with

310

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

SNARE proteins and P/Q- and N-type Ca

channels (Herlitze et al., 1996; Zurawski et al.,

311

2019; Alten et al., 2022), as opposed to activating GIRK channels in somata and dendrites. We

312

found that activating G

i/o

signaling via hM4Di

Emx1

caused an almost complete suppression of

313

glutamatergic synaptic responses evoked in

stratum radiatum

(i.e., putative Schaffer

314

collaterals) (

Fig. 5

B), indicating that presynaptic silencing is highly efficacious by P2–5. The

315

exclusively presynaptic effect is in line with previous studies showing that G

i/o

signaling matures

316

earlier at pre- compared to postsynaptic sites to inhibit neurotransmitter release by the time of

317

birth (Gaiarsa et al., 1995; Caillard et al., 1998; Nurse and Lacaille, 1999). It is important to

318

consider that over timescales longer than those addressed in this study, G

i/o

-dependent

319

actuators are anticipated to exert additional influence on canonical cAMP-protein kinase A

320

pathways. These pathways have previously been implicated in, for instance, the synaptic

321

incorporation of AMPA receptors (Esteban et al., 2003), postsynaptic long-term potentiation

322

(Yasuda et al., 2003), presynaptic long-term plasticity (Sakaba and Neher, 2003; Atwood et al.,

323

2014; Shahoha et al., 2022), synapse formation (Liang et al., 2021) as well as neuronal

324

migration and neurite growth (Bony et al., 2013). Taken together, we demonstrate that hM4Di

325

can serve as a valuable presynaptic silencer in mice shortly after birth. Our findings hold

326

potential significance for experiments utilizing hM4Di, other G

-DREADDs such as KORD

327

(Vardy et al., 2015) or likewise G

i/o

-dependent optogenetic actuators (Mahn et al., 2021) for

328

neuronal inhibition at early developmental stages.

329

330

Presynaptic G

i/o

signaling as an actuator of developing hippocampal network dynamics

331

Our analyses lead us to hypothesize that, prior to eye opening, hippocampal network activity is

332

under tight control of G

i/o

signaling acting on presynaptic terminals of glutamatergic excitatory

333

neurons. In the theoretical framework mentioned above, a reduction in quantal content of

334

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

evoked glutamate release raises the network’s amplification threshold, which effectively

335

renders network burst generation less likely (Flossmann et al., 2019). During the neonatal

336

period, GABA

R-mediated postsynaptic signaling is not yet functional (Nurse and Lacaille,

337

1999; Verheugen et al., 1999), and the chloride extrusion capacity of immature neurons is low

338

(Spoljaric et al., 2017). Therefore, G

i/o

-dependent presynaptic silencing could act as a powerful

339

inhibitory constraint on CA1 activity – at developmental stages when postsynaptic GABAergic

340

inhibition is weak (for review see Kirmse and Zhang, 2022). Currently, a major open question

341

relates to the identity and dynamics of endogenous transmitters mediating presynaptic silencing

342

through G

i/o

pathways. Earlier studies identified GABA and adenosine (acting via GABA

and

343

adenosine A1 receptors, respectively) as candidates released by local neurons and/or glial cells

344

(Gaiarsa et al., 1995; Nurse and Lacaille, 1999; Safiulina et al., 2005). Beyond that,

345

neuromodulatory projections from subcortical nuclei could target terminals for the presynaptic

346

silencing of excitation. While it is increasingly understood how G

i/o

-dependent neuromodulators,

347

such as acetylcholine, dopamine, noradrenaline, and serotonin, shape mnemonic functions

348

through the control of synaptic signaling and plasticity in the adult CA1 (Palacios-Filardo and

349

Mellor, 2019), their developmental emergence and functions are mostly unclear. A better

350

understanding of G

i/o

-dependent presynaptic silencing is clinically relevant, as many therapeutic

351

and recreational drugs modulating G

i/o

pathways can interfere with brain development during

352

pregnancy, thereby potentially causing long-lasting cognitive impairments in the offspring (Silva

353

et al., 2013; Fazeli et al., 2017). Exploring whether presynaptic G

i/o

-targeting neuromodulation

354

could serve as a brain state-dependent gating mechanism for neuronal dynamics in the

355

developing hippocampus therefore presents an interesting avenue for research.

356

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

References

362

Alten B, Guzikowski NJ, Zurawski Z, Hamm HE, Kavalali ET (2022) Presynaptic mechanisms

363

underlying GABA(B)-receptor-mediated inhibition of spontaneous neurotransmitter

364

release. Cell Rep 38:110255.

https://doi.org/10.1016/j.celrep.2021.110255

365

Armbruster BN, Li X, Pausch MH, Herlitze S, Roth BL (2007) Evolving the lock to fit the key to

366

create a family of G protein-coupled receptors potently activated by an inert ligand.

367

Proc Natl Acad Sci U S A 104:5163-5168.

https://doi.org/10.1073/pnas.0700293104

368

Atwood BK, Lovinger DM, Mathur BN (2014) Presynaptic long-term depression mediated by

369

Gi/o-coupled receptors. Trends Neurosci 37:663-673.

370

https://doi.org/10.1016/j.tins.2014.07.010

371

Ben-Ari Y, Cherubini E, Corradetti R, Gaiarsa JL (1989) Giant synaptic potentials in immature

372

rat CA3 hippocampal neurones. J Physiol 416:303-325.

373

https://doi.org/10.1113/jphysiol.1989.sp017762

374

Bony G, Szczurkowska J, Tamagno I, Shelly M, Contestabile A, Cancedda L (2013) Non-

375

hyperpolarizing GABAB receptor activation regulates neuronal migration and neurite

376

growth and specification by cAMP/LKB1. Nat Commun 4:1800.

377

https://doi.org/10.1038/ncomms2820

378

Caillard O, McLean HA, Ben-Ari Y, Gaiarsa JL (1998) Ontogenesis of presynaptic GABAB

379

receptor-mediated inhibition in the CA3 region of the rat hippocampus. J Neurophysiol

380

79:1341-1348.

https://doi.org/10.1152/jn.1998.79.3.1341

381

Chen SC, Ehrhard P, Goldowitz D, Smeyne RJ (1997) Developmental expression of the GIRK

382

family of inward rectifying potassium channels: implications for abnormalities in the

383

weaver mutant mouse. Brain Res 778:251-264.

https://doi.org/10.1016/s0006-

384

8993(97)00896-2

385

Chen YN, Kostka JK, Bitzenhofer SH, Hanganu-Opatz IL (2023) Olfactory bulb activity shapes

386

the development of entorhinal-hippocampal coupling and associated cognitive abilities.

387

Curr Biol 33:4353-4366 e4355.

https://doi.org/10.1016/j.cub.2023.08.072

388

Cossart R, Khazipov R (2022) How development sculpts hippocampal circuits and function.

389

Physiol Rev 102:343-378.

https://doi.org/10.1152/physrev.00044.2020

390

Dard RF, Leprince E, Denis J, Rao Balappa S, Suchkov D, Boyce R, Lopez C, Giorgi-Kurz M,

391

Szwagier T, Dumont T, Rouault H, Minlebaev M, Baude A, Cossart R, Picardo MA

392

(2022) The rapid developmental rise of somatic inhibition disengages hippocampal

393

dynamics from self-motion. Elife 11.

https://doi.org/10.7554/eLife.78116

394

Donato F, Jacobsen RI, Moser MB, Moser EI (2017) Stellate cells drive maturation of the

395

entorhinal-hippocampal circuit. Science 355.

https://doi.org/10.1126/science.aai8178

396

Ehrengruber MU, Doupnik CA, Xu Y, Garvey J, Jasek MC, Lester HA, Davidson N (1997)

397

Activation of heteromeric G protein-gated inward rectifier K+ channels overexpressed

398

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

by adenovirus gene transfer inhibits the excitability of hippocampal neurons. Proc Natl

399

Acad Sci U S A 94:7070-7075.

https://doi.org/10.1073/pnas.94.13.7070

400

Esteban JA, Shi SH, Wilson C, Nuriya M, Huganir RL, Malinow R (2003) PKA phosphorylation

401

of AMPA receptor subunits controls synaptic trafficking underlying plasticity. Nat

402

Neurosci 6:136-143.

https://doi.org/10.1038/nn997

403

Fazeli W, Zappettini S, Marguet SL, Grendel J, Esclapez M, Bernard C, Isbrandt D (2017)

404

Early-life exposure to caffeine affects the construction and activity of cortical networks

405

in mice. Exp Neurol 295:88-103.

https://doi.org/10.1016/j.expneurol.2017.05.013

406

Flossmann T, Kaas T, Rahmati V, Kiebel SJ, Witte OW, Holthoff K, Kirmse K (2019)

407

Somatostatin Interneurons Promote Neuronal Synchrony in the Neonatal

408

Hippocampus. Cell Rep 26:3173-3182.

https://doi.org/10.1016/j.celrep.2019.02.061

409

Fukuda A, Mody I, Prince DA (1993) Differential ontogenesis of presynaptic and postsynaptic

410

GABAB inhibition in rat somatosensory cortex. J Neurophysiol 70:448-452.

411

https://doi.org/10.1152/jn.1993.70.1.448

412

Gaiarsa JL, Tseeb V, Ben-Ari Y (1995) Postnatal development of pre- and postsynaptic

413

GABAB-mediated inhibitions in the CA3 hippocampal region of the rat. J Neurophysiol

414

73:246-255.

https://doi.org/10.1152/jn.1995.73.1.246

415

Gainutdinov A, Shipkov D, Sintsov M, Fabrizi L, Nasretdinov A, Khazipov R, Valeeva G

416

(2023) Somatosensory-Evoked Early Sharp Waves in the Neonatal Rat Hippocampus.

417

Int J Mol Sci 24.

https://doi.org/10.3390/ijms24108721

418

Gorski JA, Talley T, Qiu M, Puelles L, Rubenstein JL, Jones KR (2002) Cortical excitatory

419

neurons and glia, but not GABAergic neurons, are produced in the Emx1-expressing

420

lineage. J Neurosci 22:6309-6314.

https://doi.org/20026564

421

Graf J, Rahmati V, Majoros M, Witte OW, Geis C, Kiebel SJ, Holthoff K, Kirmse K (2022)

422

Network instability dynamics drive a transient bursting period in the developing

423

hippocampus in vivo. Elife 11.

https://doi.org/10.7554/eLife.82756

424

Graf J, Zhang C, Marguet SL, Herrmann T, Flossmann T, Hinsch R, Rahmati V, Guenther M,

425

Frahm C, Urbach A, Neves RM, Witte OW, Kiebel SJ, Isbrandt D, Hübner CA, Holthoff

426

K, Kirmse K (2021) A limited role of NKCC1 in telencephalic glutamatergic neurons for

427

developing hippocampal network dynamics and behavior. Proc Natl Acad Sci U S A

428

118.

https://doi.org/10.1073/pnas.2014784118

429

Herlitze S, Garcia DE, Mackie K, Hille B, Scheuer T, Catterall WA (1996) Modulation of Ca2+

430

channels by G-protein beta gamma subunits. Nature 380:258-262.

431

https://doi.org/10.1038/380258a0

432

Jendryka M, Palchaudhuri M, Ursu D, van der Veen B, Liss B, Katzel D, Nissen W, Pekcec A

433

(2019) Pharmacokinetic and pharmacodynamic actions of clozapine-N-oxide,

434

clozapine, and compound 21 in DREADD-based chemogenetics in mice. Sci Rep

435

9:4522.

https://doi.org/10.1038/s41598-019-41088-2

436

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Kirmse K, Zhang C (2022) Principles of GABAergic signaling in developing cortical network

437

dynamics. Cell Rep 38:110568.

https://doi.org/10.1016/j.celrep.2022.110568

438

Kirmse K, Dvorzhak A, Grantyn R, Kirischuk S (2008) Developmental downregulation of

439

excitatory GABAergic transmission in neocortical layer I via presynaptic adenosine A(1)

440

receptors. Cereb Cortex 18:424-432.

https://doi.org/10.1093/cercor/bhm077

441

Kostka JK, Hanganu-Opatz IL (2023) Olfactory-driven beta band entrainment of limbic

442

circuitry during neonatal development. J Physiol 601:3605-3630.

443

https://doi.org/10.1113/JP284401

444

Leighton AH, Cheyne JE, Houwen GJ, Maldonado PP, De Winter F, Levelt CN, Lohmann C

445

(2021) Somatostatin interneurons restrict cell recruitment to retinally driven

446

spontaneous activity in the developing cortex. Cell Rep 36:109316.

447

https://doi.org/10.1016/j.celrep.2021.109316

448

Leinekugel X, Khazipov R, Cannon R, Hirase H, Ben-Ari Y, Buzsaki G (2002) Correlated

449

bursts of activity in the neonatal hippocampus in vivo. Science 296:2049-2052.

450

https://doi.org/10.1126/science.1071111

451

Leprince E, Dard RF, Mortet S, Filippi C, Giorgi-Kurz M, Bourboulou R, Lenck-Santini PP,

452

Picardo MA, Bocchio M, Baude A, Cossart R (2023) Extrinsic control of the early

453

postnatal CA1 hippocampal circuits. Neuron 111:888-902 e888.

454

https://doi.org/10.1016/j.neuron.2022.12.013

455

Liang J, Kruijssen DLH, Verschuuren ACJ, Voesenek BJB, Benavides FFW, Saez Gonzalez

456

M, Ruiter M, Wierenga CJ (2021) Axonal CB1 Receptors Mediate Inhibitory Bouton

457

Formation via cAMP Increase and PKA. J Neurosci 41:8279-8296.

458

https://doi.org/10.1523/JNEUROSCI.0851-21.2021

459

Luhmann HJ, Prince DA (1991) Postnatal maturation of the GABAergic system in rat

460

neocortex. J Neurophysiol 65:247-263.

https://doi.org/10.1152/jn.1991.65.2.247

461

Mahler SV, Vazey EM, Beckley JT, Keistler CR, McGlinchey EM, Kaufling J, Wilson SP,

462

Deisseroth K, Woodward JJ, Aston-Jones G (2014) Designer receptors show role for

463

ventral pallidum input to ventral tegmental area in cocaine seeking. Nat Neurosci

464

17:577-585.

https://doi.org/10.1038/nn.3664

465

Mahn M, Saraf-Sinik I, Patil P, Pulin M, Bitton E, Karalis N, Bruentgens F, Palgi S, Gat A,

466

Dine J, Wietek J, Davidi I, Levy R, Litvin A, Zhou F, Sauter K, Soba P, Schmitz D, Luthi

467

A, Rost BR, Wiegert JS, Yizhar O (2021) Efficient optogenetic silencing of

468

neurotransmitter release with a mosquito rhodopsin. Neuron 109:1621-1635 e1628.

469

https://doi.org/10.1016/j.neuron.2021.03.013

470

Maldonado PP, Nuno-Perez A, Kirchner JH, Hammock E, Gjorgjieva J, Lohmann C (2021)

471

Oxytocin Shapes Spontaneous Activity Patterns in the Developing Visual Cortex by

472

Activating Somatostatin Interneurons. Curr Biol 31:322-333 e325.

473

https://doi.org/10.1016/j.cub.2020.10.028

474

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Mohns EJ, Blumberg MS (2010) Neocortical activation of the hippocampus during sleep in

475

infant rats. J Neurosci 30:3438-3449.

https://doi.org/10.1523/JNEUROSCI.4832-

476

09.2010

477

Murata Y, Colonnese MT (2020) GABAergic interneurons excite neonatal hippocampus in

478

vivo. Sci Adv 6:eaba1430.

https://doi.org/10.1126/sciadv.aba1430

479

Nurse S, Lacaille JC (1999) Late maturation of GABA(B) synaptic transmission in area CA1 of

480

the rat hippocampus. Neuropharmacology 38:1733-1742.

481

https://doi.org/10.1016/s0028-3908(99)00122-7

482

Palacios-Filardo J, Mellor JR (2019) Neuromodulation of hippocampal long-term synaptic

483

plasticity. Curr Opin Neurobiol 54:37-43.

https://doi.org/10.1016/j.conb.2018.08.009

484

Perez-Garci E, Larkum ME, Nevian T (2013) Inhibition of dendritic Ca2+ spikes by GABAB

485

receptors in cortical pyramidal neurons is mediated by a direct Gi/o-beta-subunit

486

interaction with Cav1 channels. J Physiol 591:1599-1612.

487

https://doi.org/10.1113/jphysiol.2012.245464

488

Perez-Garci E, Gassmann M, Bettler B, Larkum ME (2006) The GABAB1b isoform mediates

489

long-lasting inhibition of dendritic Ca2+ spikes in layer 5 somatosensory pyramidal

490

neurons. Neuron 50:603-616.

https://doi.org/10.1016/j.neuron.2006.04.019

491

Pnevmatikakis EA, Giovannucci A (2017) NoRMCorre: An online algorithm for piecewise rigid

492

motion correction of calcium imaging data. J Neurosci Methods 291:83-94.

493

https://doi.org/10.1016/j.jneumeth.2017.07.031

494

Prida LM, Sanchez-Andres JV (1999) Nonlinear frequency-dependent synchronization in the

495

developing hippocampus. J Neurophysiol 82:202-208.

496

https://doi.org/10.1152/jn.1999.82.1.202

497

Rahmati V, Kirmse K, Holthoff K, Kiebel SJ (2018) Ultra-fast accurate reconstruction of

498

spiking activity from calcium imaging data. J Neurophysiol 119:1863-1878.

499

https://doi.org/10.1152/jn.00934.2017

500

Rahmati V, Kirmse K, Holthoff K, Schwabe L, Kiebel SJ (2017) Developmental Emergence of

501

Sparse Coding: A Dynamic Systems Approach. Sci Rep 7:13015.

502

https://doi.org/10.1038/s41598-017-13468-z

503

Reuveny E, Slesinger PA, Inglese J, Morales JM, Iniguez-Lluhi JA, Lefkowitz RJ, Bourne HR,

504

Jan YN, Jan LY (1994) Activation of the cloned muscarinic potassium channel by G

505

protein beta gamma subunits. Nature 370:143-146.

https://doi.org/10.1038/370143a0

506

Roth BL (2016) DREADDs for Neuroscientists. Neuron 89:683-694.

507

https://doi.org/10.1016/j.neuron.2016.01.040

508

Safiulina VF, Kasyanov AM, Giniatullin R, Cherubini E (2005) Adenosine down-regulates giant

509

depolarizing potentials in the developing rat hippocampus by exerting a negative

510

control on glutamatergic inputs. J Neurophysiol 94:2797-2804.

511

https://doi.org/10.1152/jn.00445.2005

512

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Sakaba T, Neher E (2003) Direct modulation of synaptic vesicle priming by GABA(B) receptor

513

activation at a glutamatergic synapse. Nature 424:775-778.

514

https://doi.org/10.1038/nature01859

515

Sando R, Bushong E, Zhu Y, Huang M, Considine C, Phan S, Ju S, Uytiepo M, Ellisman M,

516

Maximov A (2017) Assembly of Excitatory Synapses in the Absence of Glutamatergic

517

Neurotransmission. Neuron 94:312-321 e313.

518

https://doi.org/10.1016/j.neuron.2017.03.047

519

Shahoha M, Cohen R, Ben-Simon Y, Ashery U (2022) cAMP-Dependent Synaptic Plasticity at

520

the Hippocampal Mossy Fiber Terminal. Front Synaptic Neurosci 14:861215.

521

https://doi.org/10.3389/fnsyn.2022.861215

522

Silva CG, Metin C, Fazeli W, Machado NJ, Darmopil S, Launay PS, Ghestem A, Nesa MP,

523

Bassot E, Szabo E, Baqi Y, Muller CE, Tome AR, Ivanov A, Isbrandt D, Zilberter Y,

524

Cunha RA, Esclapez M, Bernard C (2013) Adenosine receptor antagonists including

525

caffeine alter fetal brain development in mice. Sci Transl Med 5:197ra104.

526

https://doi.org/10.1126/scitranslmed.3006258

527

Spoljaric A, Seja P, Spoljaric I, Virtanen MA, Lindfors J, Uvarov P, Summanen M, Crow AK,

528

Hsueh B, Puskarjov M, Ruusuvuori E, Voipio J, Deisseroth K, Kaila K (2017)

529

Vasopressin excites interneurons to suppress hippocampal network activity across a

530

broad span of brain maturity at birth. Proc Natl Acad Sci U S A 114:E10819-E10828.

531

https://doi.org/10.1073/pnas.1717337114

532

Stachniak TJ, Ghosh A, Sternson SM (2014) Chemogenetic synaptic silencing of neural

533

circuits localizes a hypothalamus-->midbrain pathway for feeding behavior. Neuron

534

82:797-808.

https://doi.org/10.1016/j.neuron.2014.04.008

535

Thompson KJ, Khajehali E, Bradley SJ, Navarrete JS, Huang XP, Slocum S, Jin J, Liu J,

536

Xiong Y, Olsen RHJ, Diberto JF, Boyt KM, Pina MM, Pati D, Molloy C, Bundgaard C,

537

Sexton PM, Kash TL, Krashes MJ, Christopoulos A, Roth BL, Tobin AB (2018)

538

DREADD Agonist 21 Is an Effective Agonist for Muscarinic-Based DREADDs in Vitro

539

and in Vivo. ACS Pharmacol Transl Sci 1:61-72.

540

https://doi.org/10.1021/acsptsci.8b00012

541

Valeeva G, Janackova S, Nasretdinov A, Rychkova V, Makarov R, Holmes GL, Khazipov R,

542

Lenck-Santini PP (2019) Emergence of Coordinated Activity in the Developing

543

Entorhinal-Hippocampal Network. Cereb Cortex 29:906-920.

544

https://doi.org/10.1093/cercor/bhy309

545

Vardy E, Robinson JE, Li C, Olsen RHJ, DiBerto JF, Giguere PM, Sassano FM, Huang XP,

546

Zhu H, Urban DJ, White KL, Rittiner JE, Crowley NA, Pleil KE, Mazzone CM, Mosier

547

PD, Song J, Kash TL, Malanga CJ, Krashes MJ, Roth BL (2015) A New DREADD

548

Facilitates the Multiplexed Chemogenetic Interrogation of Behavior. Neuron 86:936-

549

946.

https://doi.org/10.1016/j.neuron.2015.03.065

550

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Figure 5.

hM4Di

Emx1

activation suppresses evoked glutamate release at Schaffer collateral

615

synapses.

A-C,

Data obtained at P2–5.

Sample EPSCs evoked by stimulation of Schaffer

616

collaterals in the absence (

top-left

) or presence (

top-right

) of C21 (1 µM). In a separate dataset

617

(

bottom

), vehicle (instead of C21) was applied to examine the temporal stability of evoked

618

responses. Recordings were performed in the continuous presence of gabazine (10 µM).

619

Stimulus artefacts were clipped for clarity. Arrows indicate long-latency polysynaptic bursts of

620

EPSCs. Gray traces correspond to ten successive trials, black traces represent their means.

621

B-C,

Mean amplitudes (

) and failure rates (

) of EPSCs evoked by the first pulse (eEPSC

622

D-F,

The same as in

A-C

, but at P8–12.

B-C

and

E-F,

Each open symbol represents a single

623

cell. Population data (closed symbols) are presented as mean

SEM.

∗∗∗

P < 0.001,

∗∗

P < 0.01,

624

ns – not significant (simple contrasts). For statistics, see Extended Data

Table 1-1

625

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Extended Data

654

Table 1-1.

Extended Data table supporting Figures 1–7: Descriptive and inductive statistics. GLMM, generalized linear mixed

655

model.

∗∗∗

P < 0.001,

∗∗

P < 0.01,

∗

P < 0.05, ns – not significant.

656

658

Parameter

Source

Sign. Genotype

Sign.

hM4Di

Cont

5 slices

1.4 ± 0.4

min

-1

Model:

75.9

8.7E-07

***

hM4Di

C21

0.0 ± 0.0

min

-1

Genotype:

29.7

4.1E-04

***

Cont vs. C21: 4.0

9.0

3.2E-03

Cont

6 slices

2.0 ± 0.4

min

-1

Condition:

143.4

7.8E-07

***

C21

1.4 ± 0.3

min

-1

Int.:

100.8

3.5E-06

***

Cont vs. C21: 1.5

9.0

1.6E-01

hM4Di

Cont

5 slices

2.2 ± 0.7

min

-1

Model:

14.3

1.8E-02

hM4Di

C21

0.0 ± 0.0

min

-1

Genotype:

5.2

5.9E-02

Cont vs. C21: 4.4

7.1

3.0E-03

Cont

6 slices

2.7 ± 0.5

min

-1

Condition:

16.4

4.8E-03

C21

2.1 ± 0.3

min

-1

Int.:

6.2

4.2E-02

Cont vs. C21: 1.2

7.1

2.8E-01

hM4Di

Cont

10 cells

-65.3 ± 1.8

Model:

3.1

5.4E-02

hM4Di

C21

-66.2 ± 1.7

Genotype:

8.0

1.3E-02

Cont

7 cells

-59.0 ± 1.6

Condition:

0.1

7.2E-01

C21

-58.6 ± 1.9

Int.:

1.0

3.4E-01

hM4Di

Cont

10 cells

649.8 ± 118.7

MΩ

Model:

4.5

1.0E-02

hM4Di

C21

539.9 ± 84.8

MΩ

Genotype:

0.2

6.5E-01

Cont

7 cells

602.2 ± 20.8

MΩ

Condition:

10.5

2.9E-03

C21

562.5 ± 38.1

MΩ

Int.:

1.1

3.0E-01

hM4Di

Cont

10 cells

-60.4 ± 2.1

Model:

0.031

9.9E-01

hM4Di

C21

-60.5 ± 2.5

Genotype:

0.0

8.3E-01

Cont

9 cells

-61.1 ± 1.3

Condition:

0.0

9.9E-01

C21

-61.0 ± 1.4

Int.:

0.0

8.8E-01

hM4Di

Cont

10 cells

276.4 ± 38.8

MΩ

Model:

0.51

6.8E-01

hM4Di

C21

268.0 ± 32.9

MΩ

Genotype:

0.1

7.5E-01

Cont

9 cells

249.0 ± 22.8

MΩ

Condition:

0.1

7.3E-01

C21

265.3 ± 37.1

MΩ

Int.:

1.2

2.9E-01

hM4Di

Cont

10 cells

43.0 ± 6.1

Model:

2.7

7.8E-02

hM4Di

C21

43.5 ± 5.0

Genotype:

2.5

1.3E-01

Cont

7 cells

59.4 ± 3.8

Condition:

2.8

1.1E-01

C21

51.9 ± 6.6

Int.:

3.6

7.6E-02

RMP

Max. inst.

frequency

GDP frequency

± SEM

Fixed effects (GLMM)

Simple contrasts

CaT frequency

# Figure

Genotype

(Dataset)

Condition

Biological

replicate (n)

Mean

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Table 1-1

(continued)

659

hM4Di

Cont

10 cells

14.0 ± 1.6

Model:

8.3

1.6E-03

hM4Di

C21

12.5 ± 1.3

Genotype:

0.0

9.8E-01

Cont

7 cells

15.2 ± 1.7

Condition:

22.9

2.8E-04

***

C21

11.4 ± 1.7

Int.:

4.5

5.2E-02

hM4Di

Cont

10 cells

-33.0 ± 0.6

Model:

1.1

4.6E-01

hM4Di

C21

-32.5 ± 0.8

Genotype:

2.4

1.4E-01

Cont

7 cells

-29.4 ± 2.4

Condition:

0.5

5.2E-01

C21

-29.0 ± 2.8

Int.:

0.0

9.0E-01

hM4Di

Cont

10 cells

51.0 ± 8.7

Model:

0.89

4.7E-01

hM4Di

C21

51.0 ± 7.2

Genotype:

0.4

5.4E-01

Cont

7 cells

42.9 ± 2.9

Condition:

1.5

2.5E-01

C21

47.1 ± 4.7

Int.:

1.5

2.5E-01

hM4Di

Cont

10 cells

116.2 ± 11.6

Model:

1.4

2.9E-01

hM4Di

C21

110.7 ± 9.6

Genotype:

2.4

1.4E-01

Cont

9 cells

135.2 ± 9.7

Condition:

0.3

5.7E-01

C21

136.7 ± 11.3

Int.:

1.0

3.2E-01

hM4Di

Cont

10 cells

31.5 ± 2.0

Model:

2.8

8.2E-02

hM4Di

C21

30.8 ± 2.6

Genotype:

5.1

3.9E-02

Cont

9 cells

39.6 ± 2.1

Condition:

1.2

3.0E-01

C21

37.6 ± 3.0

Int.:

0.2

6.3E-01

hM4Di

Cont

10 cells

-39.8 ± 2.1

Model:

2.3

9.9E-02

hM4Di

C21

-40.0 ± 2.1

Genotype:

1.1

3.1E-01

Cont

9 cells

-42.1 ± 1.3

Condition:

4.0

5.4E-02

C21

-43.1 ± 1.6

Int.:

2.0

1.7E-01

hM4Di

Cont

10 cells

94.0 ± 9.0

Model:

0.93

4.5E-01

hM4Di

C21

96.0 ± 8.3

Genotype:

1.8

2.0E-01

Cont

9 cells

80.0 ± 8.8

Condition:

0.0

9.7E-01

C21

77.8 ± 8.5

Int.:

0.6

4.4E-01

hM4Di

Cont

12 cells

0.13 ± 0.03

Model:

8.7

7.3E-04

***

hM4Di

C21

0.05 ± 0.01

Genotype:

2.8

1.1E-01

Cont vs. C21: 3.9

1.1E-03

Cont

8 cells

0.10 ± 0.01

Condition:

2.5

1.3E-01

C21

0.16 ± 0.06

Int.:

19.3

3.5E-04

***

Cont vs. C21: -1.6

1.4E-01

hM4Di

Cont

12 cells

21.8 ± 0.7

Model:

1.7

2.1E-01

hM4Di

C21

23.4 ± 1.8

Genotype:

0.2

6.7E-01

Cont

8 cells

26.1 ± 3.5

Condition:

1.2

2.9E-01

C21

21.2 ± 2.1

Int.:

4.4

5.0E-02

hM4Di

Cont

13 cells

0.31 ± 0.07

Model:

15.0

1.3E-05

***

hM4Di

C21

0.14 ± 0.02

Genotype:

1.1

3.0E-01

Cont vs. C21: 4.5

1.7E-04

***

Cont

10 cells

0.27 ± 0.05

Condition:

22.9

1.0E-04

***

C21

0.24 ± 0.03

Int.:

15.1

8.5E-04

***

Cont vs. C21: 0.59 21

5.6E-01

AP threshold

Rheobase

Max. mean

frequency

Max. inst.

frequency

Max. mean

frequency

AP threshold

mEPSC

frequency

mEPSC

amplitude

mEPSC

frequency

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Table 1-1

(continued)

661

hM4Di

Cont

13 cells

25.6 ± 1.5

Model:

3.6

3.0E-02

hM4Di

C21

25.7 ± 2.1

Genotype:

4.8

4.0E-02

Cont

10 cells

22.8 ± 1.6

Condition:

4.8

3.9E-02

C21

19.1 ± 1.2

Int.:

4.1

5.6E-02

hM4Di

Cont

8 cells

34.1 ± 8.5

Model:

70.0

1.4E-12

***

hM4Di

C21

1.7 ± 1.3

Dataset:

12.9

1.3E-03

Cont vs. C21: 3.3

3.1E-03

hM4Di

Cont

7 cells

36.9 ± 8.6

Condition:

111.7

6.6E-11

***

hM4Di

Veh.

23.7 ± 4.0

Int.:

72.6

5.3E-09

***

Cont vs. Veh.: 1.3

2.1E-01

hM4Di

Cont

8 cells

0.25 ± 0.06

Model:

82.3

4.9E-09

***

hM4Di

C21

0.92 ± 0.04

Dataset:

15.9

1.5E-03

Cont vs. C21: -15.1 13

1.3E-09

***

hM4Di

Cont

7 cells

0.25 ± 0.07

Condition:

121.9

5.6E-08

***

hM4Di

Veh.

0.29 ± 0.08

Int.:

91.9

2.9E-07

***

Cont vs. Veh.: -0.99 13

3.4E-01

hM4Di

Cont

9 cells

35.2 ± 7.0

Model:

42.4

6.7E-11

***

hM4Di

C21

2.5 ± 1.7

Dataset:

20.3

9.4E-05

***

Cont vs. C21: 4.0

3.5E-04

***

hM4Di

Cont

8 cells

38.2 ± 6.4

Condition:

57.1

2.0E-08

***

hM4Di

Veh.

31.6 ± 4.7

Int.:

43.4

2.7E-07

***

Cont vs. Veh.: 0.66 30

5.2E-01

hM4Di

Cont

9 cells

0.22 ± 0.06

Model:

77.6

9.0E-10

***

hM4Di

C21

0.91 ± 0.05

Dataset:

65.2

7.7E-07

***

Cont vs. C21: -12.2 15

3.4E-09

***

hM4Di

Cont

8 cells

0.14 ± 0.04

Condition:

71.5

4.3E-07

***

hM4Di

Veh.

0.15 ± 0.03

Int.:

68.9

5.4E-07

***

Cont vs. Veh.: -0.11 15

9.2E-01

hM4Di

Cont

5 animals

5.4 ± 0.9

Model:

27.6

3.2E-09

***

hM4Di

C21

0.9 ± 0.0

Dataset:

10.9

4.4E-04

***

Cont vs. C21: 3.2

3.6E-03

Cont

5 animals

8.9 ± 5.2

Condition:

15.3

6.6E-04

***

C21

9.8 ± 5.7

Int.:

50.6

2.5E-09

***

Cont vs. C21: -0.57 24

5.8E-01

hM4Di

Cont

5 animals

12.1 ± 2.5

hM4Di

Veh.

20.6 ± 6.0

Cont vs. Veh.: -2.3

3.4E-02

hM4Di

Cont

5 animals

2.1 ± 0.3

min

-1

Model:

14.5

1.4E-06

***

hM4Di

C21

0.0 ± 0.0

min

-1

Dataset:

8.5

1.6E-03

Cont vs. C21: 6.7

5.9E-07

***

Cont

5 animals

2.3 ± 0.4

min

-1

Condition:

21.3

1.1E-04

***

C21

1.4 ± 0.4

min

-1

Int.:

17.1

2.4E-05

***

Cont vs. C21: 2.8

1.0E-02

hM4Di

Cont

5 animals

2.1 ± 0.1

hM4Di

Veh.

2.5 ± 0.2

Cont vs. Veh.: -1.54 24

1.4E-01

hM4Di

Cont

5 animals

39.8 ± 7.8

Model:

60.9

7.6E-13

***

hM4Di

C21

2.5 ± 0.6

Dataset:

14.8

6.5E-05

***

Cont vs. C21: 2.6

1.4E-02

Cont

5 animals

57.6 ± 20.4

Condition:

73.4

9.1E-09

***

C21

61.9 ± 22.0

Int.:

100.7

2.1E-12

***

Cont vs. C21: -0.20 24

8.4E-01

hM4Di

Cont

5 animals

185.2 ± 69.1

hM4Di

Veh.

267.1 ± 116.9

Cont vs. Veh.: -1.3

2.1E-01

Normalized

bandpower (8-

40Hz)

eSPW

occurrence

eEPSC

failure

rate

mEPSC

amplitude

eEPSC

amplitude

eEPSC

amplitude

eEPSC

failure

rate

Normalized

bandpower (8-

40Hz)

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

Materials and Methods

663

Key Resources Table

664

Reagent type

(species) or

resource

Designation

Source or reference

Identifiers

Strain (

Emx1

IREScre

)

Emx1-IRES-cre

The Jackson Laboratory

RRID: IMSR_JAX:005628

Strain (

hM4Di

LSL

)

R26-LSL-hM4Di-

DREADD

The Jackson Laboratory

RRID: IMSR_JAX:026219

Strain (WT)

C57BL/6J

The Jackson Laboratory

RRID: IMSR_JAX:000664

Software

Matlab

Mathworks

RRID: SCR_001622

Software

Fiji

PMID: 22743772

RRID: SCR_002285

Software

pClamp (Clampfit)

Molecular Devices

RRID: SCR_011323

Software

Mini Analysis Program

Synaptosoft

RRID: SCR_002184

Software

OriginPro

OriginLab

RRID: SCR_014212

Software

IBM SPSS Statistics

IBM

RRID: SCR_016479

Software

Patchmaster

HEKA Elektronik

RRID: SCR_000034

665

Animals

666

All animal procedures were performed with approval of the local governments (Thüringer

667

Landesamt für Verbraucherschutz, reference no.: 02-012/16; Regierung von Unterfranken,

668

Arbeitsbereich Tierschutz) and complied with European Union norms (Directive 2010/63/EU).

669

Animals were housed in standard cages with

ad libitum

access to food and water.

Emx1

IREScre

670

(#005628) and

hM4Di

LSL

(#026219) mice were originally obtained from The Jackson Laboratory

671

and maintained on a C57BL/6J background. Double heterozygous offspring (referred to as

672

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

hM4Di

Emx1

mice

) was used for experiments. C57BL/6J mice served as WT controls. Mice of

673

either sex were used.

674

675

Preparation of acute brain slices

676

Animals were decapitated under deep isoflurane anesthesia. The brain was quickly removed

677

and transferred into ice-cold saline containing (in mM): 125 NaCl, 4 KCl, 10 glucose, 1.25

678

NaH

, 25 NaHCO

, 0.5 CaCl

, and 2.5 or 6 MgCl

, gassed with carbogen (5% CO

, 95%

679

; pH 7.4). Horizontal brain slices (350 µm) were cut on a vibratome and stored for at least

680

1 h before their use at room temperature in artificial cerebrospinal fluid (ACSF) containing (in

681

mM): 125 NaCl, 4 KCl, 10 glucose, 1.25 NaH

, 25 NaHCO

, 2 CaCl

, and 1 MgCl

, gassed

682

with carbogen (pH 7.4). For recordings, slices were transferred into a submerged-type recording

683

chamber on the microscope stage. All experiments were performed at ~32 °C.

684

685

Two-photon Ca

imaging in vitro

686

For single-cell Ca

imaging, cells were loaded with the membrane-permeable Ca

indicator

687

Oregon Green 488 BAPTA-1 AM (OGB1) using multi-cell bolus-loading in

stratum pyramidale

688

of hippocampal CA1. Imaging was performed using an acousto-optic deflection (AOD) two-

689

photon laser-scanning microscope operated by the software MES (Femto3D ATLAS,

690

Femtonics). Fluorescence excitation at 800 nm was provided by a tunable Ti:Sapphire laser

691

(Chameleon Ultra II, Coherent) using a 20×/1.0 NA water immersion objective (XLUMPLFLN

692

20XW, Olympus). Emission light was detected by photomultiplier tubes (16 bit, H11706P-40,

693

Hamamatsu) above and below the sample. In the upper detection pathway, emission light was

694

separated from excitation light using a primary dichroic mirror (700 nm), short-

pass ﬁltered with

695

an IR blocker (700 nm) and further band-pass filtered (520/60 nm). In the lower detection

696

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

pathway, an oil immersion condenser was used. Excitation light was blocked using an IR

697

blocker (700 nm). The emission light was then passed through a dichroic mirror (565 nm) and

698

band-pass filtered (520/60 nm). Signals from both photomultiplier tubes were summed up by a

699

signal combiner and digitized. To record the activity of neurons from various z-depths within the

700

brain slice, chessboard scanning was used, in which square scanning regions (50

50 pixels at

701

a resolution of 1 µm/pixel) were defined at various xyz-positions. Sampling rate depended on

702

the number of scanning regions and ranged from about 40 to 62 Hz. Spontaneous activity was

703

recorded for 8 min before and after wash-in of C21.

704

705

Patch-clamp recordings in vitro

706

Electrophysiological signals were acquired using a HEKA EPC 10 ampliﬁer, a built

-in 16-bit

707

AD/DA board and the software Patchmaster (HEKA Elektronik). Signals were low-

pass ﬁltered

708

at 2.9 kHz and sampled at 20 kHz. For whole-cell current-clamp recordings, glass pipettes (3–

709

5 MΩ) were filled with solution containing (in mM):

12 KCl, 123 K-gluconate, 5 NaCl, 0.2 EGTA,

710

10 HEPES, 1.8 Mg-ATP, 0.3 Na-GTP, 0.1 spermine, 10 phosphocreatine (pH 7.3). Ionotropic

711

glutamate and GABA receptor antagonists (10 µM DNQX, 50 µM APV, 10 µM gabazine) were

712

added to the ACSF to eliminate recurrent excitation and to minimize synaptic noise. Resting

713

membrane potential was measured for a total of 10 min at zero current. Once per 20 s, a test

714

pulse (duration: 500 ms; amplitude:

−

10 pA at P2–5,

−

20 pA at P8–12) was applied to estimate

715

membrane resistance. C21 (1 µM) was applied after 2 min of control recording. AP firing

716

characteristics were measured by applying a series of step currents (each 500 ms) of variable

717

amplitude (P2–5: from 0 to 120 pA in increments of 10 pA; P8–12: from 0 to 240 pA in

718

increments of 20 pA). Three series were recorded in each condition, both control and C21.

719

Immediately prior to the first series in each condition, the membrane potential was biased to

720

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

−

70 mV. For recording mEPSCs, whole-cell voltage-clamp measurements were performed at

721

a holding potential of

−

70 mV. Recording pipettes were filled with solution containing (in mM):

722

145 KCl, 5 NaCl, 0.2 EGTA, 10 HEPES, 1.8 Mg-ATP, 0.3 Na-GTP (pH 7.25). Tetrodotoxin (TTX,

723

0.5 µM) and gabazine (10 µM) were added to the ASCF to block voltage-gated Na

channels

724

and GABA

receptors, respectively. For recording eEPSCs, whole-cell voltage-clamp

725

measurements were performed at a holding potential of

−

70 mV. Schaffer collaterals were

726

electrically stimulated using an ACSF-filled glass pipette placed in

stratum radiatum

(resistance

727

about 1

MΩ

) and a Model 2100 isolated pulse stimulator (A-M Systems). After finding a

728

stimulation site, cells were allowed to recover for five minutes before the start of recording. Per

729

condition, each cell was stimulated 30 times

a paired-pulse manner (20 Hz) at an inter-trial

730

interval of 10 s. Recording pipettes were filled with solution containing (in mM): 12 KCl, 123 K-

731

gluconate, 5 NaCl, 0.2 EGTA, 10 HEPES, 1.8 Mg-ATP, 0.3 Na-GTP, 0.1 spermine, 10

732

phosphocreatine (pH 7.3). Gabazine (10 µM) was added to the ASCF to isolate eEPSCs. Only

733

recordings with an access resistance below 30 MΩ were accepted. Series resistance

734

compensation was not applied.

735

736

Surgical preparation, anesthesia and animal monitoring for in vivo LFP recordings

737

For analgesia, a subcutaneous injection of 200 mg/kg metamizol (Novacen) was administered

738

30 minutes prior to the start of the preparation. Animals were then placed onto a warm platform

739

and anesthetized with isoﬂurane (3.5% for

induction, 1–2% for maintenance) in pure oxygen

740

(flow rate: 1 l/min). The skin overlying the skull was disinfected and locally inﬁltrated with 2%

741

lidocaine (s.c.) for local analgesia. Scalp and periosteum were removed, and a custom-made

742

plastic chamber wit

h a central borehole (Ø 3 mm) was ﬁxed on the skull using cyanoacrylate

743

glue (UHU) (P3–5: 3.5 mm rostral from lambda and 1.5 mm lateral from midline; P10–12: 3.5

744

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

mm rostral from lambda and 2 mm lateral from midline). A subcutaneous catheter was placed

745

at the lateral abdomen to allow for an acute injection of C21 (3 mM dissolved in 0.9% NaCl)

746

and vehicle, respectively. To this end, a small tube (outer diameter: 0.61 mm, inner diameter:

747

0.28 mm, length: 70 mm), prefilled with 0.9% NaCl solution (volume: 4 µl) was inserted with a

748

guiding canula (19G) and fixed using cyanoacrylate glue. For the hippocampal window

749

preparation (Graf et al., 2021), the plastic chamber was tightly connected to a preparation stage

750

and subsequently perfused with warm artificial cerebrospinal fluid (ACSF) containing (in mM):

751

125 NaCl, 4 KCl, 25 NaHCO

, 1.25 NaH

, 2 CaCl

, 1 MgCl

and 10 glucose (pH 7.4, 35–

752

36°C). A circular hole was drilled into the skull using a tissue punch (outer diameter 2.7 mm).

753

The underlying cortical tissue and parts of corpus callosum were carefully removed by

754

aspiration using a vacuum supply and a blunt 27G or 30G needle. Care was taken not to

755

damage alveus fibers. As soon as bleeding stopped, the animal was transferred to the

756

microscope stage.

757

During

in vivo

recordings, body temperature was continuously monitored and maintained at

758

close to physiological values (36–37°C) by means of a heating pad and a temperature sensor

759

placed below the animal. Spontaneous respiration was monitored using a differential pressure

760

ampliﬁer (Spirometer Pod and PowerLab 4/35, ADInstruments). Isoﬂurane was discontinued

761

after completion of the surgical preparation and gradually substituted with the analgesic-

762

sedative nitrous oxide (up to the ﬁxed ﬁnal N

O/O

ratio of 3:1, flow rate: 1 l/min). Experiments

763

commenced

60 min after withdrawal of isoﬂurane. At the end of

the experiment, the animal was

764

decapitated under deep isoflurane anesthesia.

765

766

In vivo LFP recordings

767

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

The recording chamber was continuously perfused with warm ACSF (as above). A tungsten

768

microelectrode (Tunglass-

1, 0.8 MΩ impedance, Kation Scientific) was lowered just above the

769

hippocampal formation. ACSF was then removed and the hippocampal window was filled up

770

with agar (1.5%, in 0.9 mM NaCl) and covered with a custom-made cover glass that allowed

771

positioning of the electrode. As soon as the agar solidified, the chamber was reperfused with

772

ACSF. Before inserting the electrode into the brain, an LFP reference signal was recorded for

773

5 minutes to estimate the noise picked up by the electrode (technical noise). Next, the

774

microelectrode was slowly advanced into

stratum radiatum

(200–250 µm below the

775

hippocampal surface). The final electrode depth was determined when the recorded eSPWs

776

showed a polarity reversal. Electrophysiological signals were acquired and band-pass-

ﬁltered

777

at 3–3,000 Hz using an EXT-

02F/2 ampliﬁer (npi electronic), a 16

-bit AD/DA board (PowerLab

778

4/35, ADInstruments) and the software LabChart 8 (ADInstruments). Signals were sampled at

779

20 kHz. After recording spontaneous activity for 30 min, C21 was injected (s.c.) via the catheter

780

at a concentration of 3 mg/kg. NaCl (0.9%) served as the vehicle control. The LFP recording

781

continued for at least 2 hours after injection of C21.

782

783

Analysis and quantification

784

Two-photon Ca

imaging in vitro

785

Image stacks were registered using NoRMCorre (Pnevmatikakis and Giovannucci, 2017). Time

786

periods with residual z-drift were visually identified and considered as missing values in all

787

subsequent analyses. Regions of interest (ROIs) corresponding to the somata of putative CA1

788

PCs were manually selected (Fiji). Subsequent analyses were performed using custom scripts

789

in Matlab. For each cell, the time-course of fluorescence

𝐹𝐹

(

𝑡𝑡

)

was obtained by framewise

790

averaging across all pixels belonging to its ROI. Baseline fluorescence

𝐹𝐹

(

𝑡𝑡

)

was computed by

791

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

oversmoothing

𝐹𝐹

(

𝑡𝑡

)

using a second-order Savitzky-Golay algorithm (window length, 1,500

792

frames). Fluorescence signals were then expressed as relative changes from baseline levels,

793

i.e.

∆𝐹𝐹 𝐹𝐹

⁄

(

𝑡𝑡

)

. Ca

transients (CaTs) were detected from

∆𝐹𝐹 𝐹𝐹

⁄

(

𝑡𝑡

)

traces using UFARSA

794

(Rahmati et al., 2018). Network events were operationally defined as GDPs by following a

795

previous approach (Flossmann et al., 2019): (1) CaTs were classified as GDP-related if they

796

fell into a 500-ms-long time interval during which the fraction of active cells (i.e., cells

with ≥1

797

detected CaT)

was ≥20%. (2) Neighboring GDP

-related CaTs were assigned to the same GDP

798

if both shared a 500-ms-

long time interval during which the fraction of active cells was ≥20%,

799

otherwise to separate GDPs.

800

801

Electrophysiological recordings in vitro

802

Resting membrane potential was determined as the median of the recorded voltage

𝑉𝑉

𝑚𝑚

(

𝑡𝑡

)

(at

803

zero current) in consecutive 20-s intervals. Membrane resistance was estimated under current-

804

clamp conditions from the steady-state voltage change evoked by 500-ms-long hyperpolarizing

805

test pulses (

−

10 pA at P2–5,

−

20 pA at P8–12, applied at 20-s intervals). For AP detection, we

806

computed the first derivative of

𝑉𝑉

𝑚𝑚

(

𝑡𝑡

)

and smoothed it using a second-order Savitzky-Golay

807

algorithm (window: 5 samples), yielding

𝑉𝑉̇

𝑚𝑚

𝑠𝑠𝑚𝑚𝑠𝑠𝑠𝑠𝑠𝑠ℎ𝑒𝑒𝑒𝑒

(

𝑡𝑡

)

. Time points at which

𝑉𝑉̇

𝑚𝑚

𝑠𝑠𝑚𝑚𝑠𝑠𝑠𝑠𝑠𝑠ℎ𝑒𝑒𝑒𝑒

(

𝑡𝑡

)

808

crossed a threshold of 20 V/s were considered as AP onsets. The value of

𝑉𝑉

𝑚𝑚

(

𝑡𝑡

)

corresponding

809

to the onset of the first AP evoked at the lowest amplitude of injected current was defined as

810

AP threshold. The maximum mean AP frequency was computed as the maximum number of

811

evoked APs per current step divided by step duration. The maximum instantaneous AP

812

frequency was determined as the inverse of the minimum interval between two successive

813

evoked APs. Rheobase was estimated as the minimum amplitude of injected current that gave

814

rise to at least one AP. The quantities were computed separately for each of the three series of

815

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint

current injections per condition. For statistics, we computed the mean across the three series,

816

except for rheobase in which case the median was used. Analyses were performed using

817

custom scripts in Matlab. mEPSCs were detected using Minianalysis 6.0 (Synaptosoft). Events

818

were visually selected and semi-automatically detected based on an amplitude and area

819

criterion. eEPSCs were analyzed using Clampfit 11.2 (pClamp 11.2, Molecular Devices).

820

eEPSCs were visually classified as failures and successes, respectively. To correct for a minor

821

contamination of eEPSCs by the late decay of the stimulation artefact, eEPSC amplitudes were

822

computed by subtracting the mean amplitude of failures from raw eEPSC amplitudes.

823

824

In vivo LFP recordings

825

The LFP was analyzed using custom-written Matlab scripts. Artefacts and noise peaks were

826

removed semi-automatically. The LFP signal was downsampled to 500 Hz and baseline-

827

corrected using a median filter (window length: 1 s). As an unbiased approach to quantify the

828

input to CA1, we calculated the Welch’s power spectral density estimate (window length: 6 s,

829

window overlap: 20%) for a 30-min window during control and at 30–60 min after C21 injection.

830

The LFP reference signal was used for normalization. At P3–5, eSPWs were detected based

831

on amplitude (larger than

−

0.05 mV) and an increase in bandpower (15–30 Hz). Continuous

832

spectrograms were computed using the toolbox chronux (version 2.12,

http://chronux.org/

)

833

(window length: 1 s, number of tapers: 3, window overlap: 80%; frequency range: 3–100 Hz,

834

resulting in a datapoint every 0.2 s).

835

836

Statistical analysis

837

Statistical analyses were performed using Matlab (2021b), OriginPro (2020) and IBM SPSS

838

Statistics (28). Population data are reported as mean ± standard error of the mean (SEM),

839

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted February 2, 2024.

;

https://doi.org/10.1101/2024.02.02.578550

doi:

bioRxiv preprint