2024.01.03.572456v1.full-html.html

Spatiotemporal Mapping and Molecular Basis of Whole-brain Circuit Maturation

Jian Xue

, Andrew T. Brawner

2, 3, 9

, Jacqueline R. Thompson

2, 3, 9

, Tushar D. Yelhekar

, Kyra T.

Newmaster

Qiang Qiu

5, 6

, Yonatan A. Cooper

, C. Ron Yu

5, 6

, Yasir H. Ahmed-Braima

, Yongsoo Kim

Yingxi Lin

1,10*

Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, 75390, USA

Department of Psychiatry and Behavioral Sciences, SUNY Upstate Medical University, Syracuse, NY

13210, USA

Neuroscience Graduate Program, SUNY Upstate Medical University, Syracuse, NY 13210, USA

Department of Neural and Behavioral Sciences, The Pennsylvania State University, Hershey, PA 17033,

USA

Stowers Institute for Medical Research, Kansas City, MO 64110, USA

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, MO 66160,

USA

Current address: Department of Human Genetics, David Geffen School of Medicine, University of

California, Los Angeles, CA 90095, USA.

Department of Biology, Syracuse University, Syracuse, NY 13244, USA

Equal contribution

Lead contact

*Correspondence: Yingxi.Lin@UTSouthwestern.edu

KEYWORDS

DevATLAS, circuit maturation, activity-dependent, whole-brain mapping, spatiotemporal map,

developmental brain templates, neurodevelopmental disorder, environmental enrichment, energy

metabolic pathway

SUMMARY

Brain development is highly dynamic and asynchronous, marked by the sequential maturation of

functional circuits across the brain. The timing and mechanisms driving circuit maturation remain elusive

due to an inability to identify and map maturing neuronal populations. Here we create DevATLAS

(Developmental Activation Timing-based Longitudinal Acquisition System) to overcome this obstacle. We

develop whole-brain mapping methods to construct the first longitudinal, spatiotemporal map of circuit

maturation in early postnatal mouse brains. Moreover, we uncover dramatic impairments within the deep

cortical layers in a neurodevelopmental disorders (NDDs) model, demonstrating the utility of this resource

to pinpoint when and where circuit maturation is disrupted. Using DevATLAS, we reveal that early

experiences accelerate the development of hippocampus-dependent learning by increasing the

synaptically mature granule cell population in the dentate gyrus. Finally, DevATLAS enables the

discovery of molecular mechanisms driving activity-dependent circuit maturation.

INTRODUCTION

The early postnatal period encompasses the remarkable transformation of functionally primitive

newborns into sophisticated decision-making individuals. This advancement in brain function is enabled

by the establishment of efficient neural circuits.

Neural circuits comprise interconnected neurons

spanning brain regions and serve as the fundamental units of neuronal processing.

The formation of

neural circuits requires preceding processes such as neurogenesis, migration, axon pathfinding, and

target innervation. However, the functional ability of neural circuits is conferred during the final stage of

neurodevelopment through the activation and remodeling of synapses, here referred to as circuit

maturation.

2-7

Circuit function determines an organism’s behavioral and cognitive capabilities, and many

genes associated with neurodevelopmental disorders (NDDs) have been found to be involved circuit

maturation.

8-11

Circuit maturation epitomizes the complexity and asynchrony of brain development.

12-14

Different

brain regions exhibit varying rates and timings of maturation, reflecting the progressive acquisition of

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

skills across development.

15,16

For example, circuits involved in sensory and motor functions typically

mature earlier than those responsible for higher cognitive functions.

Even within a given brain region,

neurons do not mature simultaneously.

17-19

Such complexity and heterogeneity impede our ability to

systematically determine the sequence in which circuits mature, preventing us from effectively pinpointing

when and where circuit function is disrupted in NDDs.

Circuit maturation is heavily influenced by neuronal activity.

Over time, sensory-driven activity

replaces intrinsically generated spontaneous activity to become the dominant driving force for circuit

maturation.

18,21-24

Activity-dependent transcription orchestrates molecular programs to construct and

sculpt synaptic connections between neurons. What follows is a surge in the production of both excitatory

and inhibitory synapses, closely coordinated by neuronal activity to result in mature circuits with balanced

excitation and inhibition (E/I balance).

In fact, a disrupted E/I balance has consistently been observed

among NDDs.

26,27

Since activity-dependent synaptic development is an essential step in circuit

maturation, we reasoned that molecular players that are actively involved in this process can be used to

monitor the brain-wide sequence of circuit maturation. Npas4, an immediate early gene (IEG), is

particularly suited for this purpose due to its robust and selective induction by neuronal activity,

involvement in E/I balance through the activity-dependent development of both excitatory and inhibitory

synapses, and crucial roles in experience-dependent brain functions like learning and memory.

28-31

We have developed a Npas4-dependent reporter system that offers a unique capability to

permanently and cumulatively labels neurons once they have been activated during the natural course

in vivo

development. We name this reporter system DevATLAS (Developmental Activation Timing-

based Longitudinal Acquisition System). With longitudinal sampling, advanced tissue-clearing techniques,

and whole-brain imaging and mapping, we have constructed the first spatiotemporal map of the whole-

brain circuit maturation. This resource can be used to pinpoint when and where circuit development

deviates from the norm in NDDs, as we demonstrate in an NDD mouse model. To demonstrate the

versatility of DevATLAS, we uncover a circuit mechanism by which early experiences accelerate cognitive

development. Single-cell RNA sequencing (scRNA-seq) of DevATLAS animals further offers

unprecedented opportunities to uncover molecular mechanisms which drive activity-dependent circuit

development

in vivo

. Altogether, DevATLAS provides much-needed insights into the intricate processes

underlying functional maturation of the brain.

RESULTS

DevATLAS enables longitudinal capture of neuronal activation

DevATLAS permanently labels neurons once they are activated and express

Npas4

during development

(Figure 1A), which we refer to as neuronal activation hereafter. The Npas4-Cre mouse was created by

inserting a P2A-Cre sequence before the stop codon of the

Npas4

coding region. Due to the self-cleavage

property of the P2A sequence, transcription from the

Npas4

locus now generates two separate proteins

that function independently: full-length Npas4 and Cre. DevATLAS is created when the Npas4-Cre mouse

is crossed with either Ai14 or Ai75 mice,

to permanently label neurons with cytosolic (Npas4-Cre;Ai14)

or nuclear-localized (Npas4-Cre;Ai75) tdTomato (tdT) following the initial

Npas4

induction during

development.

Like the endogenous

Npas4

gene, DevATLAS tdT labeling is selectively induced by Ca

influx following

neuronal activation and is insensitive to other signaling pathways that are known to activate many other

IEGs (Figure 1B).

28,29

To determine the time course of tdT labeling after neuronal activation, we quantified

DevATLAS signal following pentylenetetrazol (PTZ)-induced seizure, to trigger Npas4 expression. In P7

DevATLAS mice, all activated cells were labeled by tdT between 6 and 12 hours after seizure (Figures

1C and S1A). All labeled cells were neurons, a feature of

Npas4

that is not shared by some other IEGs,

100

such as

Fos

(Figures S1B and S1C).

101

102

DevATLAS reveals the brain-wide sequence of neuronal activation

103

We next monitored DevATLAS tdT labeling to reveal how neuronal activation emerged across the

104

developing brain. DevATLAS revealed the asynchronous onset of neuronal activation across brain

105

regions throughout development (Figure 1D). Regions necessary for survival, such as the arcuate

106

hypothalamic nucleus (ARH, e.g., hunger and satiety) and paraventricular hypothalamic nucleus (PVH,

107

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

e.g., food intake), were among the first to exhibit robust tdT signal (Figure 1E).

33,34

Sensory- and motor-

108

associated regions followed, starting with the striatum dorsal region (STRd), which contributes to the

109

integration of sensorimotor response and motor ability (Figure 1F). DevATLAS tdT signal began to

110

emerge in cortical areas about one week later (Figure 1G). Lastly, tdT labeling appeared in brain regions

111

which are dispensable for survival but essential for important higher-level functions (e.g., the

112

hippocampus) (Figure 1H). Such patterns of progressive circuit activation according to changing

113

organismal needs and complexity match the known order of functional development. DevATLAS revealed

114

that areas within ontologically distinct brain regions also exhibit progressive maturation of circuits

115

according to the advancing organismal development. For example, tdT labeling in the hindbrain appeared

116

first in regions critical for survival (parabrachial nucleus, e.g., autonomic functions) (Figure S1D) before

117

progressing to regions regulating sensory and motor function (e.g., pontine gray and superior olivary

118

complex) (Figure S1E).

119

Quantification of DevATLAS signal in 11 brain regions at seven ages between P0 and P28 substantiated

120

our observations of asynchronous neuronal activation (Figures 1I and S1F). At this more detailed

121

resolution, differences in cortical maturation sequence are apparent. For example, the olfactory system

122

develops earliest among sensory systems and accordingly, the piriform cortex (PIR) exhibited increased

123

tdT labeling at P0 and P4 compared to other primary sensory cortices (Figure S1F). The slight delay in

124

maturation of the visual cortex (VISp) compared to other primary sensory cortices is also captured by

125

DevATLAS (Figure S1F). Interestingly, the initial neuronal activation of the medial prefrontal cortex

126

(mPFC) and the CA3 region of the hippocampus, both of which regulate higher-order behaviors that

127

emerge later in young adulthood, coincided with activation of the primary sensory and motor cortices and

128

plateaued a week or more before other higher-level brain regions in the hippocampus (e.g., the CA1 and

129

dentate gyrus) (Figure S1F). This observation suggests that the behavioral and cognitive complexities

130

attributed to the mPFC and CA3 are not limited by the neuronal activation of those regions, per se.

131

Altogether, DevATLAS reveals that brain regions are sequentially activated in a general trend according

132

to physiological needs: those crucial for basic survival emerge first, followed by regions which interface

133

with the external environment (motor and sensory), and finally regions which integrate and interpret

134

external environmental signals for higher-order behavioral execution. The sequence of tdT labeling

135

coincides with the order of functional circuit development but implies subtle differences from behavioral

136

maturity in some cases.

137

As we established in the introduction, circuit maturation is not a singular moment or event but rather a

138

development phase following circuit assembly wherein the connections between neurons are activated

139

and then reinforced. We propose that the neuronal activation captured by DevATLAS is a brain-wide

140

benchmark of the initial stage of circuit maturation that confers advanced synaptic properties and enables

141

behavioral maturity. We next conducted a series of experiments to test this hypothesis, investigating

142

whether: (1) DevATLAS captures synaptic activity

in vivo

; (2) DevATLAS tdT labeling coincides with the

143

timing of circuit assembly; and (3) DevATLAS identifies neuronal populations with advanced functional

144

maturity.

145

146

Synaptic activity triggers DevATLAS labeling during circuit maturation

147

To confirm that DevATLAS-labeled neurons have experienced synaptic activity

in vivo

, we examined

148

when, where, and which types of neuronal activity triggers tdT labeling in the sensory systems.

149

The development of the visual system is known to be driven by both spontaneous and, later, light-

driven

150

retinal activity that is transmitted to the brain via retinal ganglion cells (RGCs).

22,36-38

In the retina, tdT

151

labeling emerged during the window of the glutamatergic wave of spontaneous retinal activity around

152

P10, but not the earlier gap junction- and cholinergic-mediated waves (Figure 1J and 1K).

This suggests

153

that Npas4 expression is specifically triggered by glutamatergic synaptic activity. Dark rearing, without

154

any light exposure (Figure S1G), significantly reduced, but did not eliminate, the number of tdT-labeled

155

RGCs by P16 (Figures S1H and S1I). These observations demonstrate that DevATLAS captures both

156

spontaneous and sensory-driven neuronal activity in the retina.

157

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

To investigate the effect of RGC activity on the downstream dorsal lateral geniculate nucleus of the

158

thalamus (dLGN), we performed unilateral enucleation to eliminate both spontaneous and light-evoked

159

retinal input from one eye while preserving local connections within the dLGN (Figure 1L). This led to a

160

profound absence of tdT labeling in the contralateral dLGN at P12 and P14 (Figures 1M and 1N). These

161

findings indicate that the activity transmitted from the retina, rather than local dLGN activity, is the main

162

driver of DevATLAS signal within the dLGN. Interestingly, dark rearing, which preserves only

163

spontaneous RGC activity, led to a small but significant decrease in tdT labeling in the dLGN by P16

164

(Figures S1J and S1K). These results suggest that tdT labeling reflects the crucial role of the spontaneous,

165

glutamatergic retinal wave in shaping synaptic connections between RGCs and dLGN neurons.

166

In the olfactory system, which develops early and is functional at birth, DevATLAS exhibited labeling in

167

areas associated with olfactory processing by P0, such as the PIR and the main olfactory bulb (MOB)

168

(Figure S1L). Consistent with previous findings, removal of olfactory input with naris occlusion at P0

169

significantly reduced the number of tdT-labeled neurons in the MOB and PIR at P4 (Figures S1M-S1Q).

170

This result supports the critical role of spontaneous activity in the development of olfactory circuits.

171

Taken together, our data demonstrate that DevATLAS captures both spontaneous and sensory-driven

172

glutamatergic activity propagated throughout sensory systems, supporting our hypothesis that

173

DevATLAS captures the activity-dependent initiation of circuit maturation.

174

175

DevATLAS catches patterns of circuit assembly

176

To establish whether neuronal activation documented by DevATLAS is involved in the wiring of functional

177

circuits, we examined the appearance of tdT labeling in three developmental systems where circuit

178

assembly is well-documented.

179

The STRd is one of the earliest regions showing robust tdT labeling (Figures 1D and 1F). We found that

180

striosome neurons, known as the early-born striatal population,

were labeled and thus activated earlier

181

(P0-P4) than the later-born matrix neurons (P7-P21) (Figures 1F

2A, and S2A). To investigate whether

182

the early tdT labeling of striosome neurons indicates of their integration into the circuits first,

43,44

183

performed immunohistochemistry experiments at P0. We examined the distribution of axonal terminals

184

from the substantia nigra pars compacta (TH+), the cortex (vglut1+), and the thalamus (vglut2+).

185

Remarkably, axonal terminals from all three regions converged onto the tdT-labeled striosome region

186

and were noticeably absent from the matrix region, which was not labeled by tdT at this time (Figure 2B).

187

At P7, consistent with tdT labeling pattern, all terminals appeared in matrix (Figure S2B). Interestingly,

188

we also observed that the lateral STRd, which receives inputs from the sensorimotor cortex, displayed

189

tdT labeling before the medial STRd, which receives input from higher-order association areas (Figures

190

1F and 2A).

This progression of tdT labeling likely reflects the order and timing of corticostriatal circuits

191

assembly, which had not been previously demonstrated.

192

In the barrel cortex (SSp-bfd), the timing and sequence of DevATLAS signal closely followed the

193

formation

of the thalamocortical circuit (Figure 2C). tdT labeling first emerged in SSp-bfd layer VI and

194

then in layer IV, coinciding with the timing when thalamocortical axons heavily innervate these regions

195

(P4-P10) (Figures 2D, S2C, and S2D). Sparse tdT labeling then appeared in layer II/III and layer V at P7-

196

P10, in line with the initial activation of barrel minicolumn circuits (Figures 2C and 2D).

24,46

Of note, layer

197

V excitatory neurons are born earlier than those in layers IV and II/III yet are labeled later (Figures S2C

198

and S2D), suggesting that the order in which neurons are assembled into circuits and activated does not

199

necessarily correspond to neuronal birth order.

200

The hippocampus is another region where we observe agreement between the sequence of circuit

201

assembly and DevATLAS signal. tdT labeling first emerged in the CA2/CA3, followed by the CA1 and

202

then the dentate gyrus (DG) (Figures 2E, 2F, and S2E). This pattern recapitulates the

sequence

203

entorhinal cortex-driven hippocampal maturation reported in previous studies.

tdT labeling also reflects

204

the order by which portions of each subregion integrated into the hippocampus circuit: from CA1a/b to

205

CA1c, from CA3a/b to CA3c, and starting from the edge where early-born neurons reside in all

206

hippocampal subregions (Figure 2F).

207

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

The patterns of tdT labeling across the developing brain indicate that the sequence of circuit maturation

208

is mainly determined by circuit assembly order and is partially associated with neuronal birth order

209

(Figures 2B, 2D, 2F, and S1F). We utilized DevATLAS to confirm that the propagation of neuronal activity

210

across assembled circuits occurred through synapse. We found that tdT signal in the DG was drastically

211

reduced when presynaptic communication from its major glutamatergic excitatory input, the entorhinal

212

cortex, was ablated by tetanus toxin (Figures 2G-2I). Taken together, DevATLAS signal can be used to

213

capture the sequence of circuit assembly during brain development.

214

215

DevATLAS captures the population that is functionally more mature

216

If DevATLAS captures circuit maturation in response to the synaptic activation of assembled circuits, we

217

would anticipate that, at the cellular level, the tdT-labeled neurons are more mature and possess stronger

218

synaptic connections than their unlabeled neighboring neurons. To assess this hypothesis across brain

219

regions and cell types, we focused on the STRd, which develops early and exhibits the drastic

220

heterogeneity in tdT labeling shortly after birth (Figure 1F), and the DG, which exhibits tdT labeling weeks

221

later (Figure 1H) and is associated with cognitive function.

222

In the P0 STRd, we found that approximately 90% of tdT-labeled medium spiny neurons (MSNs), the

223

predominant striatal cell type, expressed DARPP32, a marker of mature MSNs (Figures 3A and 3B).

224

contrast, less than 10% of unlabeled MSNs expressed DARPP32. Like most neurons, the resting

225

membrane potential of MSNs decreases as part of neuronal maturation. We found that tdT-labeled MSNs

226

displayed significantly more hyperpolarized resting membrane potential compared to unlabeled MSNs

227

(Figure 3C). In addition, the reversal potential of GABA was significantly lower in tdT-labeled MSNs due

228

to a much lower intracellular chloride concentration, resulting in a more hyperpolarized GABA driving

229

force (Figure 3C). Thus, the developmental GABA switch has occurred in tdT-labeled MSNs but not yet

230

in their unlabeled counterparts. This developmental advancement is key for functional signaling and the

231

integration into mature circuits.

To further examine the maturity, we measured synaptic properties of

232

tdT-labeled and unlabeled MSNs in P7 DevATLAS mice. The tdT-labeled MSNs exhibited significantly

233

higher frequencies of spontaneous miniature excitatory (mEPSC) and inhibitory postsynaptic current

234

(mIPSC) (Figures 3D and 3E). These results strongly indicate that the activation evidenced by DevATLAS

235

signal reflects cellular and synaptic maturity in the STRd.

236

Similar results were observed in the granule cells (GCs) of the DG. At P14, all tdT-labeled cells in the DG

237

were identified as mature GCs, while less than half of unlabeled GCs expressed the maturation marker

238

Calbindin (Figures 3F and 3G).

Both mEPSC and mIPSC frequencies were significantly higher in tdT-

239

labeled GCs than unlabeled GCs (Figures 3H and 3I). We next examined whether the more abundant

240

synaptic inputs received by the tdT-labeled GCs were related to hippocampal learning and memory ability.

241

We trained P20 DevATLAS mice with contextual fear conditioning (CFC). Neurons acutely activated

242

during CFC were identified by their expression of Fos, a proxy for memory engram neurons (Figure 3J).

243

We found that CFC increased acute neuronal activation exclusively in the tdT-labeled GCs (Figures 3K

244

and 3L), strongly indicating that tdT-labeled neurons are more likely than their unlabeled counterparts to

245

participate in memory encoding.

246

Taken together, our results demonstrate that tdT-labeled neurons are more mature at the cellular,

247

physiological, synaptic, and behavioral levels. DevATLAS has proven uniquely capable of tagging

248

neurons in different brain regions and circuits that have initiated the circuit maturation in response to

249

synaptic activity. Critically, DevATLAS accomplishes this amidst the asynchronous and heterogeneous

250

in vivo

environment of the developing brain.

251

252

A spatiotemporal map of circuit maturation across the whole brain

253

Having established that DevATLAS can capture sequences of circuit maturation in different brain systems

254

and ages, we next sought to create, for the first time, a longitudinal, spatiotemporal map of circuit

255

maturation across the whole brain. Such a resource would allow researchers to identify when and where

256

neuronal activation begins in their region of interest, enabling in-depth analysis of the specific

257

developmental processes. We built a whole-brain imaging and data analysis pipeline to create this much-

258

needed resource for the neuroscience research community (Figure 4A).

259

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

First, a series of reference brain templates were generated by adopting a previously established

260

strategy,

which was then adjusted based on our current clearing and imaging method to cover neonatal

261

ages and provide more detailed annotations based on Allen common coordinate framework labels (Figure

262

4B).

Brain samples were collected from DevATLAS mice at seven developmental time points spanning

263

P0 to P28. Brains were tissue-cleared using FDISCO,

imaged on a LaVision Ultramicroscope II light

264

sheet fluorescence microscope (LSFM), and mapped onto our age-matched reference templates (Figure

265

4C). Our dataset allows for visualizing the spatiotemporal order of circuit maturation along the sagittal,

266

coronal, and horizontal axes. This visualization allows for simultaneous, coordinate-linked examination

267

of four data channels (Figures 4D, S3A, and S3B, Methods): annotation template, the averaged tdT signal

268

from at least 12 brains collected for each time point, the autofluorescence image (giving a rough

269

anatomical outline of the brain), and an individual sample brain displaying raw tdT signals (Figure 4D).

270

We highlight the thalamocortical barrel circuit to exemplify how the DevATLAS whole-brain imaging

271

dataset reveals the sequence of circuit maturation. The formation of this circuit, previously described

272

using traditional histology methods (Figure 2D), is plainly evident in LSFM images of the SSp-bfd region

273

from P0 to P21 (Figure 4E). The barrel formations can even be seen within individual brain samples

274

(Figure 4E, P7). Concurrently, LSFM images of the VPM, the thalamic input to SSp-bfd, exhibit tdT

275

labeling closely correlated to that in the SSp-bfd (Figure 4E), consistent with the known role of VPM

276

neurotransmission driving the formation of the thalamocortical barrel circuit.

277

To gain a more general insight into whole-brain circuit activation patterns, tdT signals were mapped to

278

anatomical brain regions (Table S2) and their volumetric proportions were extracted and then

279

hierarchically clustered across ages and regions (Figure S3C, Table S3). All the raw and processed data,

280

along with the new developmental brain templates, can be freely accessible through Mendeley data and

281

BrainImageLibrary.

282

One of our goals for the DevATLAS spatiotemporal map is to reveal potentially unknown roles of regions

283

in brain development. Two such regions, the fasciola cincerea and induseum griseum, stood out in the

284

tdT signal clustering data. They are little-studied regions of the hippocampus that exhibited robust tdT

285

labeling by P4, much earlier in development than the remainder of the hippocampus regions, suggesting

286

that these regions may play not-yet-identified roles in circuit maturation during the early developmental

287

stages (Figure S3C). Of note, neuronal activity in the induseum griseum has been shown to be impacted

288

by maternal stimulant abuse during pregnancy, highlighting the potential significance of the activity-

289

dependent development of this region and the need for further investigation.

290

Another intriguing observation apparent in DevATLAS LSFM images is the distinct patterns of tdT labeling

291

in cortical layer VI across different cortical areas at the same developmental age (Figures 4F and S3D).

292

In the primary sensory cortices such as the somatosensory cortex (SSp) and the VISp, tdT labeling at P7

293

first appears strongly in upper layer VI (VIa), more so than the lower layer VI (VIb, subplate). In contrast,

294

non-sensory and association cortices such as the primary motor cortex (Mop), prelimbic cortex (PL), and

295

retrosplenial cortex (RSPd), exhibited initial tdT labeling in layer VIb/subplate (Figures 4F and S3D). As

296

layer VI is the first cortical layer to receive thalamus input, these results suggest differences in

297

thalamocortical circuit assembly or maturation across different cortical areas.

298

Altogether, whole-brain imaging and analysis of DevATLAS provides high-throughput and quantitative

299

measurements of the sequential circuit maturation on the whole-brain scale. The spatiotemporal whole-

300

brain map of neuronal activation offers an unprecedented opportunity for the neuroscience community to

301

trace and explore the steps of circuit maturation during early brain development.

302

303

Whole-brain mapping reveals disruptions in circuit development in a mouse model of autism

304

spectrum disorder

305

Circuit development is often disrupted in NDDs, such as autism spectrum disorder (ASD).

Identifying

306

where and when in development these perturbations occur has been a bottleneck in NDD research that

307

limits investigation to a few pre-selected brain regions without knowing where and when the synapse-

308

level perturbations are concentrated. To overcome this limitation, we sought to leverage DevATLAS and

309

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

the high-throughput whole-brain imaging and analysis pipeline as a discovery platform that can

310

systematically identify when and where circuit development is disrupted in NDDs.

311

To this end, we employed an established NDD model with ASD like behavioral deficit: prenatal valproic

312

acid (VPA) exposure.

Pregnant DevATLAS dams were injected with either VPA or saline at E12.5

313

(Figure S3E, see Methods). Consistent with previous reports,

VPA-exposed pups exhibited

314

developmental phenotypes consistent with ASD, including impaired physical development and altered

315

social communication at the early postnatal stage (Figure S3F). Whole-brain imaging mapping was

316

carried out using FDISCO-processed P7 brains from both VPA and control saline treatment. Consistent

317

with a lack of gross morphological changes in NDDs, overall tdT labeling appeared similar at P7 (Figure

318

4G). However, several brain regions exhibited significant changes in tdT densities between saline- and

319

VPA-treated brains, with VPA exposure predominantly reducing tdT signal (Figure 4H, Table S4). Notably,

320

layer VIa of sensory cortices, particularly the SSp-bfd, exhibited the greatest reduction of tdT labeling in

321

VPA-exposed animals (Figures 4H and 4I). These results were confirmed with manual counting of LSFM

322

images (Figure 4J). Interestingly, the VPM also displayed decreased tdT densities (Figure 4H). These

323

data suggest that sensory processing abnormalities in persons with NDD may be linked to early-life

324

changes in sensory circuit maturation.

325

Thus, DevATLAS, combined with the whole-brain imaging and analysis pipeline, enables the

326

identification of the brain regions and neural circuit maturation where disrupted development may underlie

327

behavioral impairments.

328

329

DevATLAS uncovers experience-driven facilitation of cognitive development

330

Given DevATLAS’s unique ability to track the neuronal activation history of neurons undergoing activity-

331

dependent functional maturation, we sought to leverage the reporter system to

explore

the mechanisms

332

by which sensory experience influences circuit maturation in regions like the hippocampus that receive

333

highly processed sensory information. Unlike the sensory systems, experience-dependent maturation of

334

non-sensory brain regions remains under explored.

335

To manipulate sensory experience, we implemented early environmental enrichment (EE, see

Methods

)

336

in experimental mice (EE-reared mice) and compared to control mice (standard housing (SH)-reared

337

mice) (Figure 5A). EE-reared mice exhibited increased locomotion in an open field and earlier eye

338

opening compared to SH-reared pups, indicating developmental advancement (Figure S4).

60,61

339

DevATLAS revealed that EE significantly increased the proportion of GCs undergoing circuit maturation

340

in the DG at P20 and P24 (Figures 5B and 5C). Given our earlier findings that tdT-labeled GCs were

341

preferentially utilized to encode contextual memory (Figure 3K and 3L), we hypothesized that the EE-

342

induced precocious tdT labeling in the DG corresponded to improved hippocampus-dependent learning

343

and memory ability.

344

We used CFC to evaluate if EE impacted the progress of hippocampus-dependent contextual memory

345

ability during development (Figure 5D

see Methods). Consistent with the literature in rat and mice

346

studies,

13,62

SH-reared mice younger than P24 displayed little memory recall both one day and seven

347

days after CFC (Figure 5E). Remarkably, EE-reared mice as young as P20 exhibited an ability to retrieve

348

contextual memories after one day and seven days post-CFC, on par with SH-reared P28 mice (Figure

349

5F).

350

Critically, the extent of tdT labeling in the DG was highly correlated with the memory ability. At P17, when

351

both EE- and SH-reared mice demonstrated similar memory ability (Figures 5E and 5F), tdT labeling in

352

the DG was unchanged (Figure 5C). However, at P20 and P24, the enhanced memory ability

353

demonstrated by the EE-reared mice corresponded to a significant increase in the proportion of tdT-

354

labeled GCs (Figures 5F and 5C).

355

Both readouts of DG development (tdT labeling and memory ability) indicate that sensory manipulation

356

influences DG circuit maturation within a specific postnatal window. The sensory experience-induced

357

cognitive advancement was temporary, as by P28, both EE- and SH-reared mice exhibited similar

358

hippocampal-dependent memory ability. Our results demonstrate that sensory engagement can

359

accelerate circuit maturation and cognitive function within a region-specific developmental window.

360

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

Interestingly, neither mEPSCs nor mIPSCs of tdT-labeled or unlabeled GCs were significantly changed

361

by EE (Figures 5G and 5H). These data imply that the effect of EE on cognition is due to the abundance

362

of tdT-labeled DG GCs, and that the subsequent synaptic development induced by activation is similar

363

between SH and EE GCs. This is consistent with our finding that tdT-labeled GCs during early

364

development are the ones preferentially recruited to encode contextual memory (Figures 3K and 3L).

365

Taken together, our data consistently indicates that the molecular pathways activated in the tdT-labeled

366

neurons play important roles in activity- and experience-dependent circuit maturation.

367

368

Progressive development of granule cells is linked to activity-dependent transcription programs

369

To identify the molecular mechanisms underlying

in vivo

circuit maturation, we performed scRNA-seq of

370

the DevATLAS DG at P14, P20, and P24 (Figure 6A), covering the emergence of hippocampus-

371

dependent memory ability (Figure 5E). In total, we obtained 19,047 high-quality single-cell transcriptomes

372

(Figure S5A, see Methods), identifying major populations of neurons, neuronal-fated cells, and glia

373

(Figures 6B, S5B, and S5C). We found that tdT expression was restricted to a subset of mature GCs, as

374

indicated by limited expression of immature markers and increased expression of mature markers

375

(Figures 6C, S5D, and S5E). The proportion of tdT-labeled cells increased with age and matched

376

quantification of tdT labeling in the DG (Figures 5C, 6D, S1F, and S5F).

377

The GC population consisted of two immature and two mature clusters (Figures 6E). Genes involved in

378

neurite extension and protein synthesis were enriched in the immature clusters, while the mature clusters

379

exhibited gene involved in neuronal preparation for synaptic communication (growth GC) and functional

380

synaptic activity (synaptic GC) (Figures 6F and S5G, Table S5). Pseudotime analysis indicated that all

381

GC groups existed in a continuous developmental trajectory (Figure 6E). The increasing representation

382

of tdT-labeled GCs with pseudotime further validated the developmental relevance of the trajectory

383

(Figure 6G). Statistical examination of pseudotime by age and tdT labeling also supported that the tdT-

384

labeled GCs identified in the scRNA-seq dataset were more developmentally advanced than their

385

unlabeled counterparts (Figure S5H). Furthermore, the GC clusters were ordered along pseudotime

386

according to the typical progression of neuronal development (neurite extension, structural reorganization,

387

synaptic signaling) (Figure 6F).

388

To investigate the link between neuronal activation and the progression of GC development, we

389

examined the expression of activity-regulated genes (ARGs) in the GC population. We isolated a set of

390

ARGs from the literature (see Table S5 for complete list of ARGs used) and characterized their

391

expression across GC pseudotime. We found that many ARGs gradually increased in expression as

392

development progressed, highlighting the importance of activity response in mature GCs (Figure 6H).

393

However, we also identified that separate groups of ARGs were present in different GC clusters (Figures

394

6H). The ARGs increased in nascent GCs (e.g.,

Kitl

Alcam

Nrp1

) supported the importance of neuronal

395

activity in regulating early neurite outgrowth, migration, pathfinding, and cell adhesion.

63-65

Some ARGs

396

(e.g.,

Kdm7a, Dclk1, Irs2

) were particularly enriched in growth GCs. Importantly, most ARGs and

397

traditional IEGs (

Junb

Bdnf

Ier5

, and

Scg2

)

were robustly expressed within the synaptic GCs (Figure

398

6H), suggesting that this population is more responsive to activity-dependent stimulus.

399

Taken

together, the characterization of scRNA-seq data from DevATLAS DG reveals possible

400

transcriptional substrates directing activity-dependent circuit development in the DG towards active

401

synaptic signaling. DevATLAS provides us the opportunity to investigate molecular mechanisms which

402

distinguish tdT-labeled GCs from their unlabeled counterparts.

403

404

Energy production pathway is critical for neuronal maturation

405

In order to identify gene transcripts that may contribute to the development of tdT-labeled GCs, we

406

isolated differentially expressed genes (DEGs) between tdT-labeled and unlabeled GCs at any age (P14,

407

P24, or P24). Gene set enrichment analysis showed that the majority of these DEGs were associated

408

with neuronal maturation, membrane properties, synaptic functions, and, interestingly, glycolytic energy

409

production (Figure 6I). The involvement of energy production pathway in tdT-labeled GCs strongly

410

suggests that the ability to meet the energy demands required for neuronal signaling is highly associated

411

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

with DG circuit maturation. By extension, our data suggests these processes are critical for cognitive

412

development.

413

414

DISCUSSION

415

We establish DevATLAS as a novel tool to monitor the sequence of

in vivo

circuit maturation during the

416

early postnatal period (Figures 1, 2, and 4). Our data indicate that circuit assembly is a pre-requisite for

417

DevATLAS activation (Figure 2), that the neuronal activation captured by DevATLAS alters gene

418

expression to confer advanced synaptic function (Figures 3), and that neuronal populations tagged by

419

DevATLAS functionally contribute to behavioral maturity, as seen in the hippocampus (Figures 3 and 5).

420

DevATLAS reveals that brain-wide circuit maturation begins in areas required for basic survival, followed

421

by sensory and motor systems, and lastly cognition-related circuits (Figure 1). Moreover, we constructed

422

the first spatiotemporal map of postnatal whole-brain circuit maturation (Figure 4). This provides us with

423

a powerful platform to identify when and where neural circuit development is perturbed in NDD models,

424

as we demonstrated in our examination of mice with prenatal VPA exposure (Figure 4). Critically, we

425

demonstrated that DevATLAS is also powerful in uncovering molecular mechanisms that facilitate

426

activity-dependent circuit maturation (Figures 6).

427

428

An effective strategy to capture activity-dependent circuit maturation

429

Compared to other IEG-based genetic and viral tools that label activated neurons,

32,66-69

DevATLAS

430

specifically captures neurons undergoing activity-dependent circuit maturation during early development.

431

DevATLAS’s unique use of Npas4 induction to tag neurons ensures that the reporter is selectively activity-

432

dependent and neuronal specific. This feature is not present in many other IEG-based systems, as we

433

showed with Fos (Figure S1C). DevATLAS is also unique in capturing neurons that are synaptically and

434

functionally more mature within heterogenous neuronal populations (Figure 3). Npas4’s established role

435

in activity-dependent synaptic development and maintenance of E/I balance may be what positions

436

DevATLAS at the transition from circuit assembly to circuit maturity.

28,30,31,70

Since different IEGs are

437

involved in specific biological functions,

reporters based on other IEGs, such as Arc,

might provide

438

additional insights into activity-dependent brain development.

439

DevATLAS tdT labeling accumulates over time. This feature is ideal for whole-brain mapping during

440

development, as it allows to track the timing and abundance of circuit activation in different brain regions.

441

For our purpose, we intentionally avoided CreER for neonatal tdT labeling due to tamoxifen toxicity, which

442

is not well tolerated by embryos and neonates at the concentrations used for maximal labeling, especially

443

in long-term schemas required to achieve longitudinal labeling.

72,73

Nevertheless, we also made a publicly

444

available Npas4-CreER mouse line (see methods) that allows for monitoring Npas4 induction within

445

experimentally designed time windows.

446

447

A powerful discovery platform for studying NDDs

448

NDDs typically manifest as a variety of behavioral symptoms by perturbing circuit function without

449

significant alteration of brain structures.

74-76

Systematically determining when and where circuit function

450

is perturbed is challenging yet necessary to develop effective interventions for patients with NDDs.

451

DevATLAS and the whole-brain imaging analysis pipeline were designed with these goals in mind,

452

offering a fresh perspective by focusing on activity-dependent circuit maturation on the whole-brain scale.

453

Evidences suggest that sensory dysfunction in ASD arises early in development and is attributable, in

454

part, to basal sensory processing regions.

Our findings indicate that layer VIa neurons may play a

455

pivotal role in establishing sensory perception circuits that are disturbed in the prenatal VPA exposure,

456

and possibly in other NDD models as well (Figures 4G-4J). Our data also resonates with a recent finding

457

that reduced spontaneous activity and abnormal development of deep-layer excitatory projection neurons

458

contribute to human NDDs.

Further investigation is needed to reproduce and dissect this phenotype.

459

460

The driving force of activity-dependent circuit maturation

461

Spontaneous and sensory-driven activity are known to be critical for elements of sensory circuit

462

development like the expansion and refinement of axonal and dendritic arbors.

79,80

Our study provides

463

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

compelling evidence that both spontaneous and sensory-driven neuronal activity play prominent roles in

464

early circuit maturation (Figures 1J-1O and S1L-S1P). Although both spontaneous and sensory-driven

465

activity are known to be critical for circuit development, this study provides an effective way to investigate

466

the relative contributions of each type of activity to circuit development, and the timing of the effects varies

467

across the regions. We also demonstrate that early sensory-driven activity can accelerate the maturation

468

and function of non-sensory circuits like the hippocampus (Figure 5). Interestingly, our system does not

469

capture the early stage of global spontaneous activity mediated by the gap-junction,

36,81,82

suggesting

470

that gap-junction-mediated depolarization may not engage IEG programs that promote synapse

471

formation and remodeling. Our data also highlight the crucial role for glutamatergic transmission in

472

relaying spontaneous activity and sensory input that are responsible for early circuit activation (Figures

473

1K, 1M, and 2G-2I).

474

Although the influence of neuromodulators on synaptic development is well-documented,

84,85

very little is

475

known about how acetylcholine, dopamine, and serotonin influence Npas4 expression. DevATLAS will

476

facilitate future investigations across brain regions to determine the extent and mode by which

477

neuromodulators impact circuit activity during development.

478

479

Cellular and molecular mechanisms of activity-dependent circuit development

480

Due to the stochastic nature of

in vivo

physiological neuronal activity, previous studies have primarily

481

focused on acute transcription responses triggered by strong, often artificially induced synchronous

482

activity to produce populations of activated neurons

in vitro

in vivo

86,87

In contrast, DevATLAS

483

cumulatively captures previously activated neurons, enabling us to uncover molecular and cellular

484

mechanisms that are critical for activity- and experience-dependent circuits maturation under

in vivo

485

physiological conditions (Figures 5 and 6). The approach we developed here can also be applied to

486

investigate activity-dependent circuit maturation mechanisms across brain regions, cell types, and

487

developmental stages.

488

489

ACKNOWLEDGEMENTS

490

We thank Xiaochen Sun and Scott J. Neal for critical reading of the manuscript; Xubo Niu, Liu Yang, and

491

Minggui Chen, for the comments of the manuscript; Meizhen Meng, Thi Minh Hieu Lam, and Joslyn

492

Doupe for the suggestions of this study; Martin Hemberg for help with scRNA-seq data analysis; Karen

493

Gentile at the SUNY molecular analysis core for the supports of sequencing experiments; John M. Ashton

494

and Rochester Genomics Center for the scRNA-seq service. This work was funded by Human Frontier

495

Science Program long-term fellowship LT000479/2016-L (J.X.); Brain Research Foundation Seed Grant

496

(Y.L.); Simons Center for the Social Brain Equipment Grant (Y.L.); Paul and Lilah Newton Brain Science

497

Award (Y.L.); NIH grant RF1MH124605 (Y.K.); NIH grant DC014701 (R.Y. & Y.L.); NIH grants NS123710,

498

NS115543, and MH116673 (Y.L.).

499

500

AUTHOR CONTRIBUTIONS

501

Y.L. conceived and supervised the study. Y.C. and Y.L. made the Npas4-Cre mouse. J.X. and Y.L.

502

designed the experiments. J.X. performed or participated in most of the experiments. A.B. participated in

503

most of the experiments and performed imaging quantification, VPA model, and EE-related experiments.

504

J.T. conducted scRNA-seq data analysis and participated in statistical data analysis. T.Y. conducted the

505

electrophysiology experiments. K.N. and Y.K. established the young mouse brain templates and whole-

506

brain imaging analysis pipeline, and analyzed the whole-brain imaging data. Y.A. established the scRNA-

507

seq analysis pipeline. Q.Q. and R.Y. conducted the naris occlusion experiments. J.X, J.T., and Y.L. wrote

508

the manuscript with inputs from all the authors.

509

510

DECLARATION OF INTERESTS

511

The authors declare no competing interests.

512

513

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

521

induced seizure (100mg/kg) in P7 DevATLAS (Npas4-Cre;Ai75) mice.

522

(D) Parasagittal brain sections of DevATLAS (Npas4-Cre;Ai75) mice showing tdT labeling at seven

523

developmental ages.

524

(E-H) tdT labeling in hypothalamus (HY) (E), striatum dorsal region (STRd) (F), primary motor area (MOp)

525

(G), and hippocampus (HPC) (H) at seven developmental ages. PVH, paraventricular hypothalamic

526

nucleus; ARH, arcuate hypothalamic nucleus; WM, white matter; SP, subplate; CP, cortical plate; DG,

527

dentate gyrus. Division of the STRd into four quadrants: DL, dorsal lateral; VL, ventral lateral; DM, dorsal

528

medial; VM, ventral medial.

529

(I) Quantification of tdT labeling across 11 brain regions at seven developmental ages. VPM, ventral

530

posteromedial nucleus of the thalamus; PIR, piriform cortex; VISp, primary visual cortex; mPFC, medial

531

prefrontal cortex. n=4-5 mice per age.

532

(J) Schematic depicting the three waves of spontaneous retina activity during development. Modified

533

according to Bansal et al., 2000.

534

(K) Representative images show tdT labeling in mouse retinas from P7 to P14. GCL, ganglion cell layer;

535

INL, inner nuclear layer; ONL, outer nuclear layer.

536

(L) Diagram for unilateral enucleation, performed at P1.

537

(M and N) Representative images (M) and quantification (N) showing decreased tdT labeling in the

538

contralateral dorsal lateral geniculate nucleus (dLGN) following the enucleation, as compared to the

539

ipsilateral side (Control).

540

Data are shown as mean ± SEM. ****p < 0.0001. See also Figure S1 and statistical data in Table S1.

541

542

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

543

Figure 2. DevATLAS

labeling follows the sequence of circuit assembly

544

(A) Schematic of tdT labeling sequence in STRd.

545

(B) Presynaptic markers co-localize to tdT-labeled striatal regions at P0. Tyrosine hydroxylase (TH) labels

546

the projection terminals of dopaminergic neurons; vglut1 and vglut2 label the projection terminals of

547

glutamatergic neurons in cortex and thalamus, respectively.

548

(C-D) Schematic (C) and representative images (D) of thalamocortical inputs from VPM to primary

549

somatosensory area, barrel field (SSp-bfd) layers during early postnatal ages. The colors indicate the

550

timing of tdT labeling in development with the darker red meaning earlier labeling, and lighter red meaning

551

later labeling. The age of initial tdT labeling is to the right of each layer.

552

(E-F) Schematic (E) and representative images (F) of hippocampal tdT labeling sequence during

553

development. DGo, dentate gyrus outer shell; DGi, dentate gyrus inner core.

554

(G) Experimental scheme. AAVs were injected into entorhinal cortex (EC) of DevATLAS (Npas4-Cre;Ai75)

555

mice at P1 to express tetanus toxin (TeNT) and block neurotransmission from EC to DG.

556

(H and I) Representative images (H) and quantification (I) showing a dramatic reduction in tdT labeling

557

by TeNT, compared to GFP-expressing controls.

558

Data are shown as mean ± SEM. **p < 0.01. See also Figure S2 and statistical data in Table S1.

559

560

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

566

potential of GABA responses (E

GABA

), lower cellular Cl

concentration, and higher GABA driving force

567

than tdT- MSNs at P7.

568

(D and E) At P7, tdT+ MSNs showed significantly higher mEPSC (D) and mIPSC (E) frequencies

569

compared to tdT- MSNs. Representative traces are shown on the top.

570

(F and G) Representative images (F) and quantification (G) of the enrichment of mature neurons in tdT+

571

population in the DG of P14 DevATLAS (Npas4-Cre;Ai14) mice. All tdT+ neurons are expressing granule

572

cell (GC) maturation marker, Calb1 (Calbindin 1). DCX, Doublecortin.

573

(H and I) At P14, tdT+ GCs showed significantly higher mEPSC (D) and mIPSC (E) frequencies

574

compared to tdT- GCs. Representative traces are shown on the top.

575

(J) Schematic of Fos labelling after contextual fear conditioning in the DG of DevATLAS (Npas4-Cre;Ai14)

576

mice.

577

(K and L) Representative images (K) and quantification (L) showing that contextual learning exclusively

578

activates tdT+ neurons at P20. White arrows indicate Fos+/tdT+ neurons.

579

Data are shown as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. See statistical data

580

in Table S1.

581

582

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

(E) Representative lightsheet fluorescence microscope (LSFM) images with detected tdT signals in the

593

SSp-bfd and the VPM from P0 to P14.

594

(F) Representative LSFM images showing the tdT labeling different patterns observed in layer VI at P7.

595

SSp-ul, primary somatosensory area, upper limb; SSp-m, primary somatosensory area, mouth; PL,

596

prelimbic area; RSPd, retrosplenial area, dorsal part.

597

(G) Averaged tdT signals in the brains of P7 mice treated with saline or valproic acid (VPA).

598

(H) Volcano plot showing brain regions with different signal density between the saline- and VPA-treated

599

mice. Red dots indicate regions with significant difference (FDR<0.05). n= 19 (saline) and 30 (VPA).

600

(I and J) Representative LSFM images with tdT signal (I) and manual quantification (J) showing VPA

601

treatment decreased tdT labeling in the layer VIa of SSp-bfd.

602

Data are shown as mean ± SEM. ****p < 0.0001. See also Figure S3 and statistical data in Table S1.

603

604

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

605

Figure 5. Enriched environment facilitates the circuit maturation in the hippocampus

606

(A-C) Schematic (A), representative images (B), and quantification (C) of DG tdT signal after standard

607

housed (SH)- and early enriched environment (EE)-rearing.

608

(D) Experimental scheme to test hippocampus-dependent learning capability in SH- and early EE-

609

reared mice.

610

(E) SH-reared mice start displaying significant memory retention starting from P24.

611

(F) EE-reared mice display significant improvement in 1-day and 7-day memory recall by P20.

612

(G and H) mEPSC (G) and mIPSC (H) frequency and amplitude are not different between SH- and EE-

613

reared mice at P20-P21, regardless of tdT labeling. Representative traces are shown on the top.

614

Data are shown as mean ± SEM. *p < 0.05, ***p < 0.001, ****p < 0.0001, ns, not significant. See also

615

Figure S4 and statistical data in Table S1.

616

617

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

618

Figure 6. Dentate gyrus granule cell maturation linked to activity-dependent transcription

619

programs

620

(A) Schematic of DG single-cell sequencing experiment.

621

(B) UMAP plot of single-cell data (n= 19,047 cells) with cell type annotations. GC, granule cell; Trans.,

622

transition; prog., progenitor; RGL, radial glia-like cells; Olig., oligodendrocyte.

623

624

Dependent Protein Kinase IV), maturation marker (Calb1), and tdT.

625

(D) The proportion of tdT+ GCs at each age reflects previous

in vivo

quantification in Figure 5D. Total

626

cell number: P14, n=1134; P20, n= 2048; P24, n= 2355. tdT+ cell number: P14, n=121; P20, n=722; P24,

627

n=1257.

628

(E) Four clusters of GCs identified using Louvain clustering fall within a single pseudotime trajectory

629

indicated by a black line. Total cell number: Nascent, n=780; Trans., n=1521; Growth, n=2332; Synaptic,

630

n=904.

631

(F) Gene set enrichment analysis (GSEA) identified unique biological contributions of three GC clusters

632

(excluding trans. GC). The proportion of the colors within the node denotes the contribution of genes from

633

each cluster

Edge width reflects the strength of term similarity (based on shared constituent genes).

634

(G) A stacked histogram of GCs across binned pseudotime (100 equally sized bins) indicates that GC

635

clusters reflect progressive developmental and reveals preferential localization of tdT+ GCs towards

636

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

mature pseudotime endpoint. tdT+ cell number: Nascent, n=2; Trans., n=130; Growth, n=1230; Synaptic,

637

n=873.

638

(H) Activity-regulated gene (ARG) expression across binned pseudotime reveals distinct groups of ARGs

639

are associated with specific GC developmental stages. Synaptic GC population shows the most robust

640

expression of ARGs.

641

(I) GSEA network of DEGs increased in tdT+ GCs across all three ages reveals a tdT phenotype of

642

presynaptic and postsynaptic functional maturity and activation of specific signaling pathways related to

643

Ras, GABA, apoptosis, and energy metabolism. Node size indicates the number of enriched genes

644

contributing to that term. Edge width reflects the strength of term similarity (based on shared constituent

645

genes).

646

647

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

(B) tdT labeling is neuron specific. Representative images of brain slices from seizure-induced

654

DevATLAS (Npas4-Cre;Ai75) mice (P7) stained with a neuronal marker (NeuN) at P9. The DG, CA3,

655

CA1, and superficial layers of SSp-bfd are shown in the panels, with NeuN staining shown in green.

656

White arrows indicate the overlapping signals between NeuN staining and tdT.

657

658

brain slices from home cage Fos-tTA;TRE-H2BGFP mice. Dash circles indicate the non-overlapped

659

signals between NeuN staining and tdT signal.

660

(D and E) tdT labeling in the hindbrain from P0 to P28 with anatomical reference to Allen adult mouse

661

brain atlas. PB, parabrachial nucleus; SOC, superior olivary complex; PG, pontine gray; IO, inferior

662

olivary complex; TRN, tegmental reticular nucleus; CBX, cerebellar cortex.

663

(F) Quantification of tdT labeling across 11 brain regions grouped into subcortical, cortical, and

664

hippocampus regions at seven developmental ages. Data is replotted from Figure 1I.

665

(G) Experimental scheme of standard housing (SH) and dark rearing (DR).

666

(H and I) Representative images (H) and quantification (I) showing decreased tdT labeling in retinas in

667

dark-reared mice at P16.

668

(J and K) Representative images (J) and quantification (K) demonstrating decreased tdT labeling in the

669

dLGN of the DR-reared mice compared to SH-reared mice at P16.

670

(L) tdT labeling in olfactory pathway associated regions from P0 to P14 mice. MOB, main olfactory bulb;

671

AON, anterior olfactory nucleus; PIR, piriform cortex; OT, olfactory tubercle; ENTl, entorhinal area,

672

lateral part; ENTm, entorhinal area, medial part; GL, glomerular layer; EPL, external plexiform layer;

673

MCL, mitral cell layer; GCL, granule cell layer.

674

(M) Schematic showing the unilateral naris occlusion experiments.

675

(N and O) Representative images (N) and quantification (O) showing the reduced tdT labeling in MOB

676

after naris occlusion.

677

(P and Q) Representative images (P) and quantification (Q) showing the reduced tdT labeling in PIR after

678

naris occlusion.

679

Data are shown as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. See statistical data

680

in Table S1.

681

682

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

714

Figure S5. Identification of clusters and tdT status in single-cell data

715

(A) UMAP plotting showing the number of UMI (unique molecule identifier), genes, mitochondrial fraction,

716

and age (P14, P20, P24) information for each cell. Mean ± SD UMI (raw RNA) = 18,049±9,584, mean ±

717

SD number of genes (raw RNA) = 4,721±1,205, mean ± SD mitochondrial fraction = 0.037±0.016.

718

(B) Cell number plotting of projected cell clusters and native cell clusters to validate cluster consistency.

719

720

(D) tdT status threshold for each cell set as 2 counts after SCTransform. Cell number fractions (Fn) were

721

plotted based on the tdT counts (left), and cell numbers were plotted based on the tdT counts (right). The

722

red line shows the threshold setting.

723

(E) The proportion of tdT-labeled GCs in each subclusters. Nascent tdT+ n=2, Trans tdT+=130, Growth

724

tdT+ n=1230, Synaptic tdT+ n=738.

725

(F) UMAP plotting showing tdT status of cell samples from different ages.

726

(G) Expression of top GC cluster markers.

727

(H) Across all ages, tdT+ GCs display a more mature developmental status than tdT- GCs according to

728

pseudotime analysis. The proportion of population is shown on the y-axis. P14, p<1e-8; P20, p<2.2e-16;

729

P24, p<2.2e-16.

730

731

TABLES

732

Table S1. Statistical Table

733

Table S2. Whole-brain imaging quantification

734

Table S3. Filtered brain regions for hierarchical clustering

735

Table S4. VPA whole-brain imaging comparison

736

Table S5. Gene expression information for single-cell analysis

737

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

KEY RESOURCES TABLE

738

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies

Mouse monoclonal anti-NeuN Antibody, clone
A60

emdmillipore

MAB377X

Rabbit polyclonal anti-RFP

Rockland

600-401-379

Rabbit monoclonal anti-DARPP32 (19A3)

cell signaling

2306, RRID:AB_823479

Rabbit polyclonal anti-Calbindin D-28K

Sigma-Aldrich

AB1778,
RRID:AB_2068336

Rabbit polyclonal anti-Doublecortin

Synaptic Systems

326 003,
RRID:AB_2620067

Chicken polyclonal anti-Tyrosine Hydroxylase

Aves Labs

TYH,
RRID:AB_10013440

Guinea pig polyclonal anti-VGLUT1

Synaptic Systems

135 304,
RRID:AB_887878

Rabbit polyclonal anti-VGLUT2

Synaptic Systems

135 403, RRID:
AB_887883

Rabbit polyclonal anti-PCP4

Sigma-Aldrich

HPA005792,
RRID:AB_1855086

Rabbit polyclonal anti-c-Fos

Synaptic Systems

226 003,
RRID:AB_2231974

Donkey anti-rabbit Alexa Fluor 647

Thermo Fisher
Scientific

A-31573,
RRID:AB_2536183

Goat anti-rabbit Alexa Fluor 647

Thermo Fisher
Scientific

A-21244,
RRID:AB_2535812

Goat anti-Mouse IgG1 Antibody, Alexa Fluor
647

Thermo Fisher
Scientific

A-21240,
RRID:AB_2535809

Goat anti-guinea pig IgG (H+L) Alexa Fluor 647

Thermo Fisher
Scientific

A-21450,
RRID:AB_141882

Goat anti-rabbit Alexa Fluor 568

Thermo Fisher
Scientific

A-11011,
RRID:AB_143157

Goat anti-rabbit Alexa Fluor 488

Thermo Fisher
Scientific

A-11034,
RRID:AB_2576217

Donkey Anti-Chicken IgY (IgG) (H+L) Alexa
Fluor 488

Jackson
ImmunoResearch

703-545-155,
RRID:AB_2340375

Goat anti-rabbit Alexa Fluor 405

Thermo Fisher
Scientific

A-31556, RRID:
AB_221605

Bacterial and virus strains

AAV9-CaMKII-tTA

This manuscript

N/A

AAV9-TRE-GFP

This manuscript

N/A

AAV9-TRE-TeNT-GFP

This manuscript

N/A

Chemicals, peptides, and recombinant proteins

Tetrodotoxin

Tocris

1069

APV

Tocris

3693

Forskolin

Tocris

1099

Recombinant Human BDNF

Pepro Tech

450-02

Recombinant Human NT-3

Pepro Tech

450-03

Recombinant Human NT-4

Pepro Tech

450-04

CNTF

Pepro Tech

450-13

VEGF

R&D Systems

293-VE-010

PFA

Sigma-Aldrich

P6148-1KG

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

DAPI

Thermo Fisher
Scientific

D1306

PTZ

Pentylenetetrazole

P6500-50G

Tamoxifen

Sigma-Aldrich

T5648

Picrotoxin

Tocris

1128

DNQX

Tocris

2312

RNasin Plus

Promega

N2615

Tetrahydrofuran anhydrous (THF)

Sigma-Aldrich

401757

Triethylamine

Sigma-Aldrich

90335

dibenzyl ether (DBE)

Sigma-Aldrich

33630

Valproic acid (VPA)

Sigma-Aldrich

P4543

Papain

Worthington
Biochemical

LS003126

Actinomycin D

Sigma-Aldrich

A1410

Anisomycin

Sigma-Aldrich

9789

Triptolide

Sigma-Aldrich

T3652

BSA

Sigma-Aldrich

1266909

Triton X-100

Sigma-Aldrich

T9284

Normal donkey serum

Jackson Immuno
Research

017-000-121

Critical commercial assays

High sensitivity DNA kit

Agilent Technologies 5067-4626

Dead Cell Removal Kit

Miltenyi biotec

130-090-101

Genomics Single Cell 3' Reagent Kits V3.1

10X genomics

PN-1000128

Single Index Kit T Set A

10X genomics

PN-1000213

Experimental models: Organisms/strains

Mouse: C57BL/6

The Jackson
Laboratory

027

Mouse: Npas4-Cre

This manuscript

N/A

Mouse: Npas4-CreER

This manuscript

N/A

Mouse: Ai14

The Jackson
Laboratory

007914

Mouse: Ai75

The Jackson
Laboratory

025106

Mouse: Fos-tTA

The Jackson
Laboratory

108306

Mouse: pTRE-H2BGFP

The Jackson
Laboratory

005104

Software and algorithms

Clampex 10.2

Molecular Devices

https://www.moleculard
evices.com

Clampfit 10.2

Molecular Devices

https://www.moleculard
evices.com

MiniAnalysis

Synaptosoft

http://www.synaptosoft.c
om

Prism

GraphPad

https://www.graphpad.c
om

SAS Prolab recording software

AviSoft

https://www.avisoft.com/

ANY-Maze software

ANY-Maze

https://www.any-
maze.com/

ImageJ (Fiji)

https://fiji.sc/

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

Illustrator

Adobe systems

https://www.adobe.com/

Cellranger

10X genomics

https://www.10xgenomic
s.com/

scrublet

https://github.com/swolo
ck/scrublet

scanpy

https://github.com/scver
se/scanpy

R (v4.1.0)

R Core Team (2021)

https://www.R-
project.org/

Seurat V4

https://github.com/satijal
ab/seurat

edgeR

https://bioconductor.org/
packages/release/bioc/h
tml/edgeR.html

Slingshot

https://bioconductor.org/
packages/release/bioc/h
tml/slingshot.html

Scater

https://bioconductor.org/
packages/release/bioc/h
tml/scater.html

MAST

https://github.com/RGLa
b/MAST/

UpSetR

https://CRAN.R-
project.org/package=Up
SetR

clusterProfiler

100

https://bioconductor.org/
packages/release/bioc/h
tml/clusterProfiler.html

Other
DAPI Fluoromount-G

SouthernBiotech

0100-20

739

RESOURCE AVAILABILITY

740

Lead contact

741

Further information and requests for resources and reagents should be directed to the lead contact:

742

Yingxi Lin (yingxi.lin@utsouthwestern.edu).

743

Materials availability

744

Animal lines and datasets generated in this study are available on the request from lead contact.

745

Data and code availability

746

All original codes will be available on Github.

747

748

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

749

EXPERIMENTAL MODEL AND SUBJECT DETAILS

750

Mouse line

751

C57BL/6 mice were purchased from the Charles River Laboratory. Npas4-Cre and Npas4-CreER were

752

generated in this study. Ai14

(

Gt(ROSA)26Sor

tm14(CAG-tdTomato)Hze

), Ai75 (Gt(ROSA)26Sor

tm75.1(CAG-

753

tdTomato*)Hze

), Fos-tTA (Tg(Fos-tTA,Fos-EGFP*)1Mmay), pTRE-H2BGFP (Tg(tetO-HIST1H2BJ/GFP)47Efu)

754

were purchased from the Jackson Laboratory. All mice have been backcrossed with wild-type animals

755

from the Charles River Laboratory. All mice were housed in a 12-hour light-dark cycle with unrestricted

756

access to food and water. Mice were weaned on postnatal day 21, unless otherwise specified. Post-

757

weaning mice were housed by sex in groups of 3-5 until used for experiments. As standard practice, both

758

males and females were used in this study. Animal protocols were performed in accordance with NIH

759

guidelines and approved by the IACUCs of SUNY Upstate Medical University or Stowers Institute for

760

Medical Research.

761

762

METHOD DETAILS

763

Cell culture

764

Cortical primary neuron culture was prepared from Npas4Cre;Ai14 P1 brains according to previous

765

publication.

Neurons were plated on glass slides within 24-well plates with 150k neurons per slide and

766

one slide per well. Neurons were cultured in Neurobasal A medium (NBA, Invitrogen) supplemented with

767

B27 (Invitrogen), GlutaMAX (Invitrogen), TTX (0.5 mM, Tocris), and APV (50 mM, Tocris), and were

768

maintained in a humidified incubator with 5% CO

at 37°C.

769

On day in vitro (DIV) 4, the medium was removed and reserved, and neurons were stimulated with NBA

770

containing one of the following: KCl (50 mM), KCl+EGTA (5mM), forskolin (10 μM, Tocris), BDNF (50

771

ng/ml, Pepro Tech), NT3 (50 ng/ml, Pepro Tech), NT4 (50 ng/ml, Pepro Tech), CNTF (100 ng/ml, Pepro

772

Tech), or VEGF (100 ng/ml, R&D Systems). After 1 hour of stimulation, the medium was replaced with

773

reserved conditioned medium. On DIV6, neurons were fixed in 4% paraformaldehyde (PFA) (Sigma-

774

Aldrich) in phosphate buffered saline (PBS) for subsequent imaging and quantification.

775

Viral vectors and titers

776

All vectors and viruses were prepared in house. To block the synaptic transmission from the entorhinal

777

cortex to the dentate gyrus, AAV9-CaMKII-tTA was co-injected with AAV9-TRE-GFP or AAV9-TRE-

778

TeNT-GFP into the medial entorhinal cortex and the lateral entorhinal cortex of postnatal 1-day (P1) mice.

779

Examination of tdT labeling was carried out at P16.

780

Stereotaxic Injection

781

Animals were gas-anesthetized using 1.5% isoflurane in O

and received bi-lateral injections in the medial

782

entorhinal cortex and the lateral entorhinal cortex. The volume for injection was 36nl for each side. The

783

surgery animals were returned to their dam and sacrificed according to the procedure.

784

Euthanasia and tissue processing for fluorescent imaging

785

Npas4-Cre;Ai75 animals were used for all quantifications of tdT-labeled cells. Animals were euthanized

786

via exsanguination and were transcardially perfused with PBS followed by 4% PFA in PBS. Brains were

787

post-fixed in 4% PFA at 4°C overnight and then incubated in 30% sucrose at 4°C for 24 hours. Brains

788

were sectioned at 50 um using a cryostat (Leica). Corresponding brain slices were washed 3 times in

789

PBS.

790

To quantify tdT-labeled cells without any additional antibody staining, tissue slices were stained with

791

DAPI (1 μg/ml, Sigma-Aldrich) for 10 min, washed three times with PBS, and mounted on microscope

792

slides.

793

Euthanasia and tissue processing for electrophysiology

794

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

Mice were anesthetized with isoflurane (P14 and P20) or on the ice (P7) and decapitated. Brains were

795

then immediately dissected out and cooled in ice-cold cutting solution containing 210 mM sucrose, 2.5

796

mM KCl, 1.24 mM NaH

, 8 mM MgCl

, 1 mM CaCl

, 26 mM NaHCO

, 20 mM Glucose, and 1.3 mM

797

sodium-L-ascorbate, with ~340 mOsm osmolarity and pH of 7.3. After a few minutes, the brains were cut

798

into 300 μM transverse slices using a vibratome (VT1200, Leica). Slices were then immediately

799

transferred to a warm recovery solution (32°C) with 50% cutting solution and 50% ACSF. ACSF contains

800

119 mM NaCl, 2.5 mM KCl, 1.24 mM NaH

, 1.3 mM MgCl

, 2.5 mM CaCl

, 26 mM NaHCO

, and 10

801

mM glucose, with 305 mOsm osmolarity and pH of 7.3. After 12 minutes of recovery, the slices were

802

transferred to ACSF (~25°C) and bubbled with carbogen gas (95% O

and 5% CO

) for at least 30

803

minutes at room temperature (RT).

804

Euthanasia and tissue processing for RNA sequencing

805

For scRNA-seq, a similar approach was employed when collecting brain samples. Artificial cerebrospinal

806

fluid (ACSF) was prepared prior to collecting the samples and was modified according to a previous

807

publication.

ACSF contains 124 mM NaCl, 2.5 mM KCl, 1.2 mM NaH

, 24 mM NaHCO

, 5 mM

808

HEPES, 13 mM glucose, 2 mM MgSO

, and 2 mM CaCl

. A cocktail of chemicals was added into ACSF,

809

including 1 μM TTX (Sigma), 100 μM APV (Tocris), 1 μM DNQX (Tocris), 5 μg actinomycin D (Sigma)

810

per mL, 10 μg anisomycin (Sigma) per ml, and 10 μM triptolide (Sigma). ACSF was bubbled with

811

carbogen gas (95% O

and 5% CO

) for at least 15 minutes prior to use.

812

Mice were anesthetized with isoflurane (P14, P20, P24) for 5 minutes and then euthanized by

813

exsanguination and perfused with ice-cold ASCF. The brain was rapidly dissected and sliced into 400

814

μm sections using a vibratome, with ACSF used as the cutting solution. Brain slices were then processed

815

according to the single-cell protocol.

816

Immunohistochemistry

817

After three washes in PBS, brain slices were incubated in blocking buffer containing 10% goat serum and

818

1% Triton-X100 (Sigma-Aldrich) in PBS at RT for 2 hours. Subsequently, the sections were incubated

819

with primary antibodies in staining buffer (5% goat serum and 0.1% Triton-X100 in PBS) at 4°C overnight.

820

After primary incubation, the sections were washed three more times in PBS and then were incubated

821

with secondary antibodies in 5% goat serum and 0.1% Triton-X100 in PBS at RT for 2 hours. After a final

822

series of the three PBS washes, the sections were mounted on microscope slides with DAPI

823

Fluoromount-G (SouthernBiotech) and covered with coverslips. The following primary antibodies were

824

used: rabbit polyclonal anti-c-Fos (1:2000, Synaptic System), mouse monoclonal IgG1 anti-NeuN (1:500,

825

emdmillipore), and rabbit polyclonal anti-RFP (1:500, Rockland). The secondary antibodies used were

826

anti-primary species with Alexa Fluor 568 and 647 (1:500, Thermo Fisher Scientific).

827

Brain slice imaging and quantification

828

For most brain regions, we acquired images with 10X and 20X objective lens using image stitching. For

829

the dentate gyrus (DG), we acquired images with an 60X objective lens due to high cell density. Manual

830

image quantification was performed using ImageJ software. Images were cropped, annotated, and

831

arranged using Adobe Illustrator software (Adobe systems).

832

Inducing postnatal seizures

833

To induce seizure, P7 animals were intraperitoneally (IP) injected with PTZ (dissolved in sterile saline,

834

~7.2 of pH) with a dosage of 100 mg/kg. As a control, littermates were injected with saline. Animals were

835

euthanized and perfused at 0h, 6h, 12h, 18h, and 48 h (hours) after injection. Brains were post-fixed with

836

4% PFA at 4°C overnight.

837

Visual-dependent sensory experiments

838

To examine the role of retinal activity on circuit development, unilateral or bilateral enucleation was

839

performed on P1 pups. All surgical tools were sterilized with 70% ethanol. Animals were injected with

840

buprenorphine ER (0.04mg/kg) 10 minutes before the procedure and then anesthetized on ice for one

841

minute. After verifying the extent of anesthesia with a toe-pinch, the skin around the eyes was sterilized

842

with 70% ethanol. The eyelids were then opened with a scalpel to expose the eye which was then

843

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

carefully excised from the orbit with forceps and scissors under a dissection scope. Following eye removal,

844

cotton tips were used to soak up blood and provide light pressure on the orbit to stop bleeding. The eyelid

845

was closed with forceps and sealed using approximately 5 μl of Vetbond glue (3M, MN, USA). After the

846

surgery, animals were transferred to a heating pad for recovery. During recovery, a thin coat of lidocaine

847

hydrochloride jelly USP, 2% (Akorn, Lake Forest, IL, USA) was applied to the eyelid to prevent pain.

848

Once recovered, the pups were returned to the home cage with the dam. Brain samples were collected

849

according to the experiment design.

850

To examine the impact of spontaneous retinal activity on circuit development, a cohort of animals were

851

raised in a light-tight container in a dark room from P5 to P16. Control animals for dark rearing experiment

852

were raised under 12h:12h light-dark cycle in standard animal facility conditions. Brain samples were

853

collected at P10 or P16 according to the experimental design. To validate the light deprivation, an X-ray

854

film was placed in the same box and developed at the end of deprivation to indicate the presence of light

855

contamination.

856

Olfactory deprivation: unilateral naris occlusion

857

Unilateral naris occlusion was performed at P0. Pups were anesthetized on ice for one minute. After

858

verifying the extent of anesthesia with a toe-pinch, animals were immediately subjected to electrocautery

859

for ~2s on left or right side of nostril under a dissection scope. During the procedures, any contact of skin

860

surface with electrocautery device was avoided. Pups were then recovered on a heat pad until they woke

861

up and transferred to the dam. The nose were immersed in water daily to check for air bubbles to ensure

862

nasal blockage of the cauterized nostril. Brain samples were collected at P4.

863

Electrophysiology

864

Electrophysiological measurements of developing neurons were performed at specific developmental

865

timepoints: striatal measurements were collected at P7, and DG measurements were collected at P14

866

and P20. All recordings were performed in a system that continuously perfused with carbogenated ACSF

867

at RT with a flow rate of 2 mL/min. Slices were gently placed and fixed in the recording chamber, and

868

tdT-labeled cells were identified using a laser. Borosilicate glass pipettes with 3-6 MU tip resistance were

869

used for recording. Data were collected using a Multiclamp 700B (Molecular Devices), filtered at 3 kHz

870

and digitized at 10 kHz using a Digidata 1440A and Clampex 10.2 software (Molecular Devices).

871

mEPSCs were pharmacologically isolated with 0.5 mM tetrodotoxin (TTX, Tocris) and 50 mM picrotoxin

872

(Tocris) in the ACSF and recorded with voltage clamp mode at -70mV. The internal solution for mEPSC

873

contained 130 mM CsMeSO

, 10 mM phosphocreatine, 1 mM MgCl

, 10 mM HEPES, 0.2 mM EGTA, 4

874

mM Mg-ATP, and 0.5 mM Na-GTP, at 295 mOsm osmolarity with pH adjusted to 7.25 using CsOH.

875

mIPSC were pharmacologically isolated with ACSF containing 0.5 mM TTX and 50 mM APV (Tocris) and

876

20 mM DNQX (Tocris), with cells voltage clamped at -70mV. The high-Cl solution contained 103 mM

877

CsCl, 12 mM CsMeSO

, 5 mM TEA-Cl, 10 mM HEPES, 4 mM Mg-ATP, 0.5 mM Na-GTP, 0.5 mM EGTA,

878

and 1 mM MgCl

, at 295 mOsm osmolarity with pH adjusted to 7.25 using CsOH.. The resting membrane

879

potential, reversal potential of GABA responses (EGABA), GABA driving force, and cellular Cl-

880

concentration were measured according to previous publications.

101,102

881

Data were blindly analyzed using MiniAnalysis (Synaptosoft) and Clampfit 10.2 (Molecular Devices).

882

Single-cell sequencing library preparation

883

The dissociation buffer was modified according to previous publication.

Dissociation buffer contained

884

HBSS (Life Technologies) with 10 mM HEPES (Sigma-Aldrich), 172 mg kynurenic acid (Sigma-Aldrich)

885

per liter, 0.86 g MgCl

·6H2O per liter, and 6.3 g d-glucose per liter, pH adjusted to 7.35. Papain was

886

added to the dissociation solution with a final concentration of 20 U/ml. A cocktail of chemicals was added,

887

including 1 μM TTX (Sigma-Aldrich), 100 μM APV (Tocris), 1 μM DNQX (Tocris), 5 μg actinomycin D

888

(Sigma-Aldrich) per milliliter, 10 μg anisomycin (Sigma-Aldrich) per milliliter, and 10 μM triptolide (Sigma-

889

Aldrich). Dissociation buffer bubbled with carbogen gas (95% O

and 5% CO

) for at least 15 minutes

890

prior to use.

891

Brain slices were collected from P14 (one library), P20 (two libraries), and P24 (one library) for SH reared

892

mice, with two males and two females were contained in each library. The DG was dissected from the

893

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

brain slices under a dissecting scope. The ACSF used for cutting was then replaced with a dissociation

894

buffer after samples had been collected from all animals. The dissected tissue was trimmed into small

895

pieces and then pipetted up-and-down twice with a 1 ml pipette tip and gently shaken at 37°C for 20

896

minutes. The dissociation buffer was then replaced with fresh, ice-cold dissociation buffer without papain.

897

The tissue pieces were then moved to a 15 ml Falcon tube and gently triturated with a 1 ml tip up-and-

898

down for 10 times. The suspension was passed through a 40 μm filter to remove clumps.

899

The filtered cell solution was centrifuged at 180 g for 3 minutes and washed four times with a dissociation

900

buffer containing 0.04% BSA. Unhealthy cells were removed according to dead cell removal kit

901

instructions (Miltenyi biotec). After the last wash, cells were resuspended in 100 μl dissociation buffer

902

containing 0.04% BSA without any blockers. The cell suspension then passed through a 15 μm filter to

903

remove any remaining clumps or debris and kept on ice. Manual cell counting was immediately performed

904

using a hemocytometer, and the volume of cell suspension needed for single-cell library preparation was

905

determined by cell density. Single-cell libraries were then made according to the 10X Genomics Single

906

Cell 3' Reagent Kits V3.1 User Guide (10X genomics). Libraries were indexed by Single Index Kit T Set

907

A (10X genomics). Multiple libraries were pooled and sequenced with Nova seq at Genomics Research

908

Center of University of Rochester.

909

Whole-brain clearing and imaging

910

We followed the FDISCO protocol to preserve endogenous tdT fluorescence from the developmental

911

brains of Npas4-Cre;Ai75 mice.

P0, P4, P7 animals were anesthetized on ice and P10, P14, P21, and

912

P28 animals were anesthetized with isoflurane and then euthanized via exsanguination and transcardially

913

perfused with PBS and then 4% PFA in PBS. Brains were post-fixed in 4% PFA at 4°C overnight and

914

then washed in PBS three times for 1 hour each. The subsequent dehydration and clearing steps were

915

performed at 4°C with gentle shaking and without any light exposure. Samples were dehydrated in a

916

series of tetrahydrofuran anhydrous (THF, Sigma-Aldrich) solution concentrations for 12 hours at each

917

concentration (50%, 70%, 80%, and 100% volume). Each solution was diluted in ddH

O, and pH adjusted

918

to 9.0 with triethylamine (Sigma-Aldrich). A final step of dehydration was performed using 100% THF for

919

12 hours. Following dehydration, refractive index matching and clearing were achieved by washing brains

920

twice in dibenzyl ether (DBE, Sigma-Aldrich) for 3 hours each wash. The brains were either immediately

921

processed for imaging or stored in DBE at 4°C. The imaging data were collected on the same day.

922

Brains were imaged using a lightsheet fluorescence microscope (Ultramicroscope, LaVision BioTec,

923

Germany) equipped with an sCMOS camera (Andor Neo, model DC-152Q-C00-FI) and a 2X 0.5

924

(Olympus MVPLAPO 2XC) objective lens with a 5.6mm working distance of optic correct organic solvent

925

resistance dipping cap. Samples were mounted on the imaging holder, and the merged imaging chamber

926

filled with DBE. Zoom body was adjusted to 1X magnification for imaging. Imaging stacks were acquired

927

using 3 μm steps, and exposure times were varied according to the samples. Each sample was imaged

928

in two different channels, and duplicate images were taken for each channel. The tdT fluorescence was

929

obtained with an excitation filter 560/30nm and an emission filter 625/30nm. Sample autofluorescence

930

was collected for imaging registration with an excitation filter 475/39nm and an emission filter 525/50nm.

931

Whole-brain imaging developmental template transformation and signal detection

932

To established an age-matched reference template, we first picked one best sample at each age. Since

933

LSFM images were collected in hemispheres, we created synthetic full brains by generating mirror

934

imaged hemispheres and combining them with original hemispheres. We used Elastix

103

to register all

935

samples in each age to one best represented sample. Then, we used intensity averaging of transformed

936

brains to create a reference template at each age. For anatomical labels, we performed sequential down-

937

registration from P56 Allen Common Coordinate Frameworks to younger ages using Elastix.

938

Image processing was carried out in Matlab (Mathworks). The background channel of each sample was

939

resized to 20x20x20 μm

stacks. The average autofluorescence of the entire brain in the background

940

channel was determined by masking pixels in the blank space as null values and excluding them from

941

the average. Each full resolution single plane image in both the foreground and background channel was

942

normalized by dividing the intensity of each pixel by the average intensity of background. Both the

943

background and foreground channels were corrected using the flatfield correction algorithm (Matlab) with

944

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

a 20-pixel area to account for spherical optical distortion introduced by an objective lens. For each full

945

resolution image, the background was subtracted from the foreground to remove vascular and other non-

946

specific signals. A scaling factor was applied to account for overall intensity differences between channels.

947

For samples that were P10 or less, a correction factor of 0.75 was applied to the background image

948

before subtraction. For samples that were P14 to P28, a correction factor of 0.5 was applied to the

949

background image before subtraction. We then calculated the average and standard deviation of the

950

corrected image and used one-half standard deviation above the average value as threshold for signals

951

to generate binarized images. 3D coordinates of detected signal voxels were assigned to the 20x20x20

952

μm

space by dividing the coordinate values by corresponding resolution changes. Then the accumarray

953

function (Matlab) was used to count the number of signal voxels in 20x20x20 μm

space. The amount of

954

signals was saved as relative percentage occupancy per voxel to represent signal intensity. These values

955

were saved into a 20 μm isotropic voxel image across the brain. To quantify signals in anatomical regions

956

of interest (ROI), we used Elastix to register individual samples to aged matched template brain at 20 μm

957

isotropic resolution.

51,103

The relative occupancy of summed positive signals per anatomical areas was

958

used to represent signal intensity in each ROI.

959

All the whole-brain imaging raw and processed data, along with the new developmental brain templates,

960

can be freely accessible at Mendeley data (DOI: 10.17632/m8n8thxxyf.1), BrainImageLibrary

961

(BrainImageLibrary.org).

962

Prenatal Valproic acid (VPA) model NDD model

963

Timed pregnant mice (E12.5) were IP injected VPA (dissolved in sterile saline) with a dosage of 600mg/kg,

964

and saline was injected as a control.

965

Self-righting test was performed on VPA and saline treated cohorts using P7 pups as a measure of gross

966

development. Briefly, animals were gently positioned supine with all four limbs pointed upwards. Latency

967

to self-right was measured as the seconds from release until all four limbs touching ground. If the animal

968

had not achieved this position after 60 s, the animal was returned to prone manually, and a time of 60s

969

was recorded. The self-righting test was performed three times with one minute rest in between. The final

970

score is the average latency of the three tests.

971

Ultrasonic vocalizations emitted by mouse pups can be an early indication of social awareness. On P9,

972

the dam was removed from the housing cage and the pups were transported to a separate location for

973

30 minutes habituation in the testing room. Animals were then individually placed in a plastic cage, and

974

the cage was placed within a sound-attenuating chamber. Ultrasonic vocalizations were measured for 3

975

minutes using an UltraSoundGateCM16/CMPA microphone (AviSoft) and SAS Prolab recording software

976

(AviSoft). The number of vocalizations between 33-125 kHz was then counted.

977

Early environmental enrichment rearing

978

Time pregnant animals (E12.5) were randomly assigned to either standard or early environmental

979

enrichment (EE) housing condition. In the standard housing (SH) condition, one dam and the litter (with

980

a typical litter size around 6) were reared in a standard plastic cage (32.5 X 21 X 18.5 cm) with shredded

981

cotton bedding. The cage was changed weekly. In the EE condition, two pregnant dams and two stranger

982

adult females were housed in a larger plastic cage (50 X 36 X 28 cm), where various sensory stimuli

983

were present, such as variable bedding textures (shredded papers, shredded cotton, corn, shredded

984

wood), visual cues (toys of different colors and shapes), and olfactory cues on wooden cubes. The

985

bedding and environment cues were changed daily. In addition, exploratory behavior was encouraged

986

with the wooden cubes of various sizes, translucent plastic igloos, plastic tubes and tunnels, and a

987

running wheel. The SH- and EE-reared pups were weaned at P28.

988

For the EE experiments, postnatal SH- and EE-reared mice were monitored between P11 and P15 for

989

eye-opening. For each animal, the opening of one eye was scored as 1, the eye-opening scores ranged

990

from 0 to 2 per animal.

991

Open field

992

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

Open field testing was performed on P10 for both EE- and SH-reared mice. Pups and dams were

993

habituated for 30 minutes in the behavior space before testing. During the test, a

single

pup was placed

994

in the center of an open field chamber and allowed to freely explore for 10 min. The open field chamber

995

was made of polyvinyl chloride with dimensions of 42 x 42 x 42 cm. The chamber was cleaned with 70%

996

ethanol before each test. Pup locomotion and exploration was recorded and analyzed with Any-Maze

997

software. Pup exploration was measured by using a 5x5 (8.4 cm x 8.4 cm) square grid of the chamber

998

floor with the number of individual squares that the pup entered being totaled.

999

Contextual fear conditioning (CFC)

1000

To examine hippocampal-dependent memory ability, CFC behavior testing was conducted on P17, P20,

1001

P24, and P28, with all training taking place before weaning. The animals and dam were habituated to

1002

handling in the pre-testing area for 3 days prior to CFC training. The CFC training and recall paradigms

1003

took 5 minutes and occurred in a testing room adjacent to the habituation area. Behavior tastings were

1004

performed at a consistent time of day. During training, animals were placed in a custom context box and

1005

allowed to explore for 1 min 58 seconds before receiving the first of a series of three 0.5 mA shocks,

1006

each lasting for 2 s with 1-minute intervals between shocks. Animals were removed from the context box

1007

one minute after the final shock and returned to the home cage with the dam.

1008

The CFC recall tests occurred 1 day and 7 days after training. All animals were only trained once and

1009

tested twice. The following training/1 day recall/7 day recall groups were used for data collection:

1010

P17/P18/P24, P20/P21/P27, P24/P25/P31, and P28/P29/P35. CFC recall involved the animals and dam

1011

being returned to habituation area and recall testing occurred individually. Animals were placed in the

1012

same context as training for 5 minutes without receiving any shocks. Contextual memory was assessed

1013

by measuring the freezing behavior displayed in the animals within this context. Freezing was defined as

1014

the total absence of movement apart from respiration. Video recordings were also made to confirm

1015

accuracy behavior quantification. All the recordings and analysis were performed blinded.

1016

Sequencing data analysis

1017

scRNA-seq data processing, clustering, and data integration

1018

Cellranger was used to demultiplex Libraries BCL files, analyze feature barcodes, and annotate resulting

1019

transcripts to the mm10 reference genome (UCSC genome browser), with modification to contain the tdT

1020

sequence at the Rosa26 loci.

We identified possible doublets using Scrublet on each library individually

1021

and then concatenated all 4 libraries in scanpy.

92,93

1022

Quality control thresholds were determined using scanpy. Cells were excluded if they were predicted to

1023

be a doublet or had a doublet score of >0.35. We excluded cells with greater than 60,000 reads (UMI

1024

counts) and less than 2,500 genes. All subsequent analysis was performed in R (v4.1.0).

1025

Seurat V4 was then used to determine cell types.

SCTransform was first implemented to scale gene

1026

expression and in this step variance due to mitochondrial fraction was regressed out.

104

Variable features

1027

were selected to create PCs using VST method of FindVariableFeatures. 30 PCs were used to find

1028

neighbors, clusters, and UMAP dimensions. Clusters were then classified into cell types based on the

1029

expression of the indicated genes (Figure S5C).

1030

Next, we took steps to limit the potential for downstream analysis to be biased by over-clustered data. To

1031

do this we first split the single-cell dataset into neurons and non-neurons and generated new PCAs for

1032

the neuron datasets to better capture the genetic diversity of this important cell type. For both datasets,

1033

we used the FindTransferAnchors and TransferData functions to project cell cluster identities back onto

1034

the same single-cells (using the same data as both reference and query sets). We found that the

1035

projected cell types were very consistent with the originals (neurons: 95.0% accuracy, non-neurons: 99.4%

1036

accuracy). As desired, some cell clusters were lost when projecting, indicating these groupings were

1037

likely the result of over-clustering (Figure S5B). A few remaining clusters were removed due to poor data

1038

quality: olig neuron, exclude (microglia), and high-mt GCs. Moving forward the projected cell clusters

1039

were used as the default (Figure S5C).

1040

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

With the filtered cell population we calculated new PCs and UMAP using 20 PCs. This data we evaluated

1041

using the standard quality controls of mitochondrial fraction, number of genes, and number of counts

1042

(Figure S5A). We used the SCTransformed tdT counts to binarize tdT status at a threshold of 2 (Figure

1043

S5D).

1044

Finally, the granule cells (GCs) were isolated based on cell cluster and a few straggler cells were

1045

eliminated based on UMAP filters (5,562 GCs). We then normalized the SCT counts with edgeR to

1046

calculate CPM and applied a log2 transformation.

This data was then reformatted as a

1047

SingleCellExperiment to be compatible with downstream analysis.

105

1048

We next filtered out genes that were present in <25% of GCs, leaving 7,904 genes (Table S5). These

1049

are the genes and expression values used for differential expression analyses.

1050

Cluster marker gene identification:

1051

We ran Seurat’s FindMarkers function on SCTransformed RNA values to identify cluster-specific gene

1052

markers. Results were filtered to genes that were present in at least 25% of all GCs then filtered by

1053

adjusted p value < 0.0001 and average log2 fold change >0.5. This produced 484 nascent GC markers,

1054

14 transition GC markers, 102 growth GC markers, and 117 synaptic GC markers. The top markers for

1055

each group were selected based on the following criteria: 10 genes with the lowest adjusted p value, 10

1056

genes with the lowest minimum adjusted p value between SH and EE, 10 genes with the highest average

1057

log2FC, and 10 genes with the greatest difference in percent expression (vs other GC clusters). This

1058

produced 94 nascent GC markers, 13 transition GC markers, 25 growth GC markers, and 15 synaptic

1059

GC markers. Due to the large number of nascent GC markers with an adjusted p value equal to 0, nascent

1060

GC markers were re-generated without adjusted p value filters. This produced a total of 65 unique genes

1061

that are the top markers differentiating GC clusters (Figure S5G).

1062

Trajectory analysis:

1063

To reconstruct the cell-stage transition during GC development, we used the Slingshot package to

1064

generate pseudotime estimates from gene expression.

Pseudotime analysis on PC representations of

1065

log2(CPM) gene expression

indicated a single pseudotime trajectory that we used as an estimate of

1066

developmental maturation (Figure 6d). When examining pseudotime along with other experimental

1067

outcomes, we used a binned version of the pseudotime measurement. For statistical comparison of tdT

1068

pseudotime at different ages and visualization as a cumulative distribution plot, 20% of the GCs were

1069

randomly sampled (P14 n= 223 (29 tdT+), P20 n= 405 (141 tdT+), P24 n= 479 (246 tdT+). Kolmogorov-

1070

Smirnov tests were performed with a two-sided alternative hypothesis.

1071

Differential gene expression:

1072

We examined the CPM normalized log2 counts of 7,904 genes for significant differences based on our

1073

experimental variable of binarized tdT status. The MAST package was used to implement zero-inflated

1074

linear models testing for the experimental condition with consideration of library depth as a covariate.

1075

For all differential expression analyses, significance was evaluated at an FDR < 0.0001. To consider the

1076

interaction of our experimental variables of tdT status with the biological variables of age and GC group,

1077

we split the GC population by these variables and created a separate model for each level. For age-

1078

specific analysis significance further required an absolute log2FC of 0.3. For cluster-specific analysis

1079

significance further required an absolute log2FC of 0.1. The UpSetR package was used to visualize

1080

DEGs across ages.

1081

Gene set enrichment analysis (GSEA):

1082

The clusterProfiler package was used for GSEA of gene ontology biological processes terms.

100

For all

1083

GSEA performed, our own whole-DG dataset was used to generate the universe background (19,025

1084

genes). For all GSEA performed, only results with adjusted p<0.05 were considered.

1085

Genes used for cluster-specific GSEA were derived from cluster markers (102 growth GC markers, 117

1086

synaptic GC markers). Given the inflated number of unique nascent GC markers, the 134 top markers

1087

were used. Transition GCs were not included in GSEA as this cluster did not demonstrate a sufficiently

1088

unique transcriptome from the other clusters. These gene lists generated 29 nascent GC terms, 32 growth

1089

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

GC terms, and 44 synaptic GC terms. The top 10 non-redundant terms were chosen from each analysis

1090

and merged using clusterProfiler. These results are summarized in Figure 6F, where the color of the node

1091

indicates which analysis produced that significant term. Nodes with split colors indicate the presence of

1092

the term in more than one analysis and the proportion of the colors within the node denotes the proportion

1093

of genes contributed by each analysis.

1094

For tdT GSEA, we were more interested in a thorough characterization of enriched biological processes

1095

and so fewer filters were applied to the resulting terms. GSEA for tdT-labeled cells was performed on

1096

245 genes increased in tdT-labeled cells at all P14, P20, and P24.

1097

Activity-regulated genes (ARGs):

1098

We compiled a list of ARGs based on three published datasets, indicating activity-induced transcription

1099

based on prolonged

in situ

stimulation,

visual experience

and a dependence on fos transcriptional

1100

activity.

106

We isolated 423 genes upregulated based on sensory experience, 168 genes upregulated

1101

based on

in situ

stimulation, and 123 genes that were increased in more than one experimental

1102

manipulation in Yap 2021. Of those, the following gene sets were also expressed in more than 25% of

1103

our GC samples: 219 sensory experience ARGs, 58

in situ

ARGs, and 83 fos-associated ARGs.

1104

We further limited this list to 21 depolarization induced genes (experience-dependent and

in situ

, with 19

1105

also showing fos association), 64 fos-associated experienced-dependent ARGs, and 22 fos-associated

1106

stimulation-induced ARGs. Scater was used to generate the heatmap of these 107 ARG genes across

1107

pseudotime.

1108

1109

REFERENCES

1110

Sokolowski, K., and Corbin, J.G. (2012). Wired for behaviors: from development to

1111

function of innate limbic system circuitry. Front Mol Neurosci

, 55.

1112

10.3389/fnmol.2012.00055.

1113

Tau, G.Z., and Peterson, B.S. (2010). Normal development of brain circuits.

1114

Neuropsychopharmacology

, 147-168. 10.1038/npp.2009.115.

1115

Stiles, J., and Jernigan, T.L. (2010). The basics of brain development. Neuropsychol

1116

Rev

, 327-348. 10.1007/s11065-010-9148-4.

1117

Goodman, C.S., and Shatz, C.J. (1993). Developmental mechanisms that generate

1118

precise patterns of neuronal connectivity. Cell

72 Suppl

, 77-98. 10.1016/s0092-

1119

8674(05)80030-3.

1120

Kolodkin, A.L., and Tessier-Lavigne, M. (2011). Mechanisms and molecules of neuronal

1121

wiring: a primer. Cold Spring Harb Perspect Biol

. 10.1101/cshperspect.a001727.

1122

Winnubst, J., and Lohmann, C. (2012). Synaptic clustering during development and

1123

learning: the why, when, and how. Front Mol Neurosci

, 70.

1124

10.3389/fnmol.2012.00070.

1125

Seng, C., Luo, W., and Foldy, C. (2022). Circuit formation in the adult brain. Eur J

1126

Neurosci

, 4187-4213. 10.1111/ejn.15742.

1127

Parenti, I., Rabaneda, L.G., Schoen, H., and Novarino, G. (2020). Neurodevelopmental

1128

Disorders: From Genetics to Functional Pathways. Trends Neurosci

, 608-621.

1129

10.1016/j.tins.2020.05.004.

1130

Iakoucheva, L.M., Muotri, A.R., and Sebat, J. (2019). Getting to the Cores of Autism.

1131

Cell

178

, 1287-1298. 10.1016/j.cell.2019.07.037.

1132

10.

Hall, L.S., Medway, C.W., Pain, O., Pardinas, A.F., Rees, E.G., Escott-Price, V.,

1133

Pocklington, A., Bray, N.J., Holmans, P.A., Walters, J.T.R., et al. (2020). A

1134

transcriptome-wide association study implicates specific pre- and post-synaptic

1135

abnormalities in schizophrenia. Hum Mol Genet

, 159-167. 10.1093/hmg/ddz253.

1136

11.

Washbourne, P. (2015). Synapse assembly and neurodevelopmental disorders.

1137

Neuropsychopharmacology

, 4-15. 10.1038/npp.2014.163.

1138

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

12.

Inoue, N., Nishizumi, H., Ooyama, R., Mogi, K., Nishimori, K., Kikusui, T., and Sakano,

1139

H. (2021). The olfactory critical period is determined by activity-dependent

1140

Sema7A/PlxnC1 signaling within glomeruli. Elife

. 10.7554/eLife.65078.

1141

13.

Ramsaran, A.I., Wang, Y., Golbabaei, A., Aleshin, S., de Snoo, M.L., Yeung, B.A.,

1142

Rashid, A.J., Awasthi, A., Lau, J., Tran, L.M., et al. (2023). A shift in the mechanisms

1143

controlling hippocampal engram formation during brain maturation. Science

380

, 543-

1144

551. 10.1126/science.ade6530.

1145

14.

Hooks, B.M., and Chen, C. (2020). Circuitry Underlying Experience-Dependent

1146

Plasticity in the Mouse Visual System. Neuron

106

, 21-36.

1147

10.1016/j.neuron.2020.01.031.

1148

15.

Hensch, T.K., and Bilimoria, P.M. (2012). Re-opening Windows: Manipulating Critical

1149

Periods for Brain Development. Cerebrum

2012

, 11.

1150

16.

Bethlehem, R.A.I., Seidlitz, J., White, S.R., Vogel, J.W., Anderson, K.M., Adamson, C.,

1151

Adler, S., Alexopoulos, G.S., Anagnostou, E., Areces-Gonzalez, A., et al. (2022). Brain

1152

charts for the human lifespan. Nature

604

, 525-533. 10.1038/s41586-022-04554-y.

1153

17.

Kempermann, G., Song, H., and Gage, F.H. (2015). Neurogenesis in the Adult

1154

Hippocampus. Cold Spring Harb Perspect Biol

, a018812.

1155

10.1101/cshperspect.a018812.

1156

18.

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

1157

and function. Physiol Rev

102

, 343-378. 10.1152/physrev.00044.2020.

1158

19.

Klune, C.B., Jin, B., and DeNardo, L.A. (2021). Linking mPFC circuit maturation to the

1159

developmental regulation of emotional memory and cognitive flexibility. Elife

1160

10.7554/eLife.64567.

1161

20.

Pan, Y., and Monje, M. (2020). Activity Shapes Neural Circuit Form and Function: A

1162

Historical Perspective. J Neurosci

, 944-954. 10.1523/JNEUROSCI.0740-19.2019.

1163

21.

Marin, O. (2016). Developmental timing and critical windows for the treatment of

1164

psychiatric disorders. Nat Med

, 1229-1238. 10.1038/nm.4225.

1165

22.

Hooks, B.M., and Chen, C. (2006). Distinct roles for spontaneous and visual activity in

1166

remodeling of the retinogeniculate synapse. Neuron

, 281-291.

1167

10.1016/j.neuron.2006.07.007.

1168

23.

Shen, J., and Colonnese, M.T. (2016). Development of Activity in the Mouse Visual

1169

Cortex. J Neurosci

, 12259-12275. 10.1523/JNEUROSCI.1903-16.2016.

1170

24.

Kast, R.J., and Levitt, P. (2019). Precision in the development of neocortical

1171

architecture: From progenitors to cortical networks. Prog Neurobiol

175

, 77-95.

1172

10.1016/j.pneurobio.2019.01.003.

1173

25.

Turrigiano, G.G., and Nelson, S.B. (2004). Homeostatic plasticity in the developing

1174

nervous system. Nat Rev Neurosci

, 97-107. 10.1038/nrn1327.

1175

26.

Sohal, V.S., and Rubenstein, J.L.R. (2019). Excitation-inhibition balance as a framework

1176

for investigating mechanisms in neuropsychiatric disorders. Mol Psychiatry

, 1248-

1177

1257. 10.1038/s41380-019-0426-0.

1178

27.

Nelson, S.B., and Valakh, V. (2015). Excitatory/Inhibitory Balance and Circuit

1179

Homeostasis in Autism Spectrum Disorders. Neuron

, 684-698.

1180

10.1016/j.neuron.2015.07.033.

1181

28.

Lin, Y., Bloodgood, B.L., Hauser, J.L., Lapan, A.D., Koon, A.C., Kim, T.K., Hu, L.S.,

1182

Malik, A.N., and Greenberg, M.E. (2008). Activity-dependent regulation of inhibitory

1183

synapse development by Npas4. Nature

455

, 1198-1204. 10.1038/nature07319.

1184

29.

Ramamoorthi, K., Fropf, R., Belfort, G.M., Fitzmaurice, H.L., McKinney, R.M., Neve,

1185

R.L., Otto, T., and Lin, Y. (2011). Npas4 regulates a transcriptional program in CA3

1186

required for contextual memory formation. Science

334

, 1669-1675.

1187

10.1126/science.1208049.

1188

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

30.

Spiegel, I., Mardinly, A.R., Gabel, H.W., Bazinet, J.E., Couch, C.H., Tzeng, C.P.,

1189

Harmin, D.A., and Greenberg, M.E. (2014). Npas4 regulates excitatory-inhibitory

1190

balance within neural circuits through cell-type-specific gene programs. Cell

157

, 1216-

1191

1229. 10.1016/j.cell.2014.03.058.

1192

31.

Weng, F.J., Garcia, R.I., Lutzu, S., Alvina, K., Zhang, Y., Dushko, M., Ku, T., Zemoura,

1193

K., Rich, D., Garcia-Dominguez, D., et al. (2018). Npas4 Is a Critical Regulator of

1194

Learning-Induced Plasticity at Mossy Fiber-CA3 Synapses during Contextual Memory

1195

Formation. Neuron

, 1137-1152 e1135. 10.1016/j.neuron.2018.01.026.

1196

32.

Madisen, L., Zwingman, T.A., Sunkin, S.M., Oh, S.W., Zariwala, H.A., Gu, H., Ng, L.L.,

1197

Palmiter, R.D., Hawrylycz, M.J., Jones, A.R., et al. (2010). A robust and high-throughput

1198

Cre reporting and characterization system for the whole mouse brain. Nat Neurosci

1199

133-140. 10.1038/nn.2467.

1200

33.

Myers, M.G., Jr., and Olson, D.P. (2012). Central nervous system control of

1201

metabolism. Nature

491

, 357-363. 10.1038/nature11705.

1202

34.

Joly-Amado, A., Cansell, C., Denis, R.G., Delbes, A.S., Castel, J., Martinez, S., and

1203

Luquet, S. (2014). The hypothalamic arcuate nucleus and the control of peripheral

1204

substrates. Best Pract Res Clin Endocrinol Metab

, 725-737.

1205

10.1016/j.beem.2014.03.003.

1206

35.

Chiang, M.C., Bowen, A., Schier, L.A., Tupone, D., Uddin, O., and Heinricher, M.M.

1207

(2019). Parabrachial Complex: A Hub for Pain and Aversion. J Neurosci

, 8225-8230.

1208

10.1523/JNEUROSCI.1162-19.2019.

1209

36.

Martini, F.J., Guillamon-Vivancos, T., Moreno-Juan, V., Valdeolmillos, M., and Lopez-

1210

Bendito, G. (2021). Spontaneous activity in developing thalamic and cortical sensory

1211

networks. Neuron

109

, 2519-2534. 10.1016/j.neuron.2021.06.026.

1212

37.

Tiriac, A., Smith, B.E., and Feller, M.B. (2018). Light Prior to Eye Opening Promotes

1213

Retinal Waves and Eye-Specific Segregation. Neuron

100

, 1059-1065 e1054.

1214

10.1016/j.neuron.2018.10.011.

1215

38.

Nakazawa, S., and Iwasato, T. (2021). Spatial organization and transitions of

1216

spontaneous neuronal activities in the developing sensory cortex. Dev Growth Differ

1217

323-339. 10.1111/dgd.12739.

1218

39.

Bansal, A., Singer, J.H., Hwang, B.J., Xu, W., Beaudet, A., and Feller, M.B. (2000).

1219

Mice lacking specific nicotinic acetylcholine receptor subunits exhibit dramatically

1220

altered spontaneous activity patterns and reveal a limited role for retinal waves in

1221

forming ON and OFF circuits in the inner retina. J Neurosci

, 7672-7681.

1222

10.1523/JNEUROSCI.20-20-07672.2000.

1223

40.

Brunjes, P.C. (1994). Unilateral naris closure and olfactory system development. Brain

1224

Res Brain Res Rev

, 146-160. 10.1016/0165-0173(94)90007-8.

1225

41.

Yu, C.R., Power, J., Barnea, G., O'Donnell, S., Brown, H.E., Osborne, J., Axel, R., and

1226

Gogos, J.A. (2004). Spontaneous neural activity is required for the establishment and

1227

maintenance of the olfactory sensory map. Neuron

, 553-566. 10.1016/s0896-

1228

6273(04)00224-7.

1229

42.

Kelly, S.M., Raudales, R., He, M., Lee, J.H., Kim, Y., Gibb, L.G., Wu, P., Matho, K.,

1230

Osten, P., Graybiel, A.M., and Huang, Z.J. (2018). Radial Glial Lineage Progression and

1231

Differential Intermediate Progenitor Amplification Underlie Striatal Compartments and

1232

Circuit Organization. Neuron

, 345-361 e344. 10.1016/j.neuron.2018.06.021.

1233

43.

Fishell, G., and van der Kooy, D. (1987). Pattern formation in the striatum:

1234

developmental changes in the distribution of striatonigral neurons. J Neurosci

, 1969-

1235

1978. 10.1523/JNEUROSCI.07-07-01969.1987.

1236

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

44.

Fishell, G., and van der Kooy, D. (1989). Pattern formation in the striatum:

1237

developmental changes in the distribution of striatonigral projections. Brain Res Dev

1238

Brain Res

, 239-255. 10.1016/0165-3806(89)90042-4.

1239

45.

Hintiryan, H., Foster, N.N., Bowman, I., Bay, M., Song, M.Y., Gou, L., Yamashita, S.,

1240

Bienkowski, M.S., Zingg, B., Zhu, M., et al. (2016). The mouse cortico-striatal

1241

projectome. Nat Neurosci

, 1100-1114. 10.1038/nn.4332.

1242

46.

Constantinople, C.M., and Bruno, R.M. (2013). Deep cortical layers are activated

1243

directly by thalamus. Science

340

, 1591-1594. 10.1126/science.1236425.

1244

47.

Donato, F., Jacobsen, R.I., Moser, M.B., and Moser, E.I. (2017). Stellate cells drive

1245

maturation of the entorhinal-hippocampal circuit. Science

355

1246

10.1126/science.aai8178.

1247

48.

Ivkovic, S., and Ehrlich, M.E. (1999). Expression of the striatal DARPP-32/ARPP-21

1248

phenotype in GABAergic neurons requires neurotrophins in vivo and in vitro. J Neurosci

1249

, 5409-5419. 10.1523/JNEUROSCI.19-13-05409.1999.

1250

49.

Peerboom, C., and Wierenga, C.J. (2021). The postnatal GABA shift: A developmental

1251

perspective. Neurosci Biobehav Rev

124

, 179-192. 10.1016/j.neubiorev.2021.01.024.

1252

50.

Sloviter, R.S. (1989). Calcium-binding protein (calbindin-D28k) and parvalbumin

1253

immunocytochemistry: localization in the rat hippocampus with specific reference to the

1254

selective vulnerability of hippocampal neurons to seizure activity. J Comp Neurol

280

1255

183-196. 10.1002/cne.902800203.

1256

51.

Newmaster, K.T., Nolan, Z.T., Chon, U., Vanselow, D.J., Weit, A.R., Tabbaa, M.,

1257

Hidema, S., Nishimori, K., Hammock, E.A.D., and Kim, Y. (2020). Quantitative cellular-

1258

resolution map of the oxytocin receptor in postnatally developing mouse brains. Nat

1259

Commun

, 1885. 10.1038/s41467-020-15659-1.

1260

52.

Wang, Q., Ding, S.L., Li, Y., Royall, J., Feng, D., Lesnar, P., Graddis, N., Naeemi, M.,

1261

Facer, B., Ho, A., et al. (2020). The Allen Mouse Brain Common Coordinate

1262

Framework: A 3D Reference Atlas. Cell

181

, 936-953 e920. 10.1016/j.cell.2020.04.007.

1263

53.

Qi, Y., Yu, T., Xu, J., Wan, P., Ma, Y., Zhu, J., Li, Y., Gong, H., Luo, Q., and Zhu, D.

1264

(2019). FDISCO: Advanced solvent-based clearing method for imaging whole organs.

1265

Sci Adv

, eaau8355. 10.1126/sciadv.aau8355.

1266

54.

Li, H., Fertuzinhos, S., Mohns, E., Hnasko, T.S., Verhage, M., Edwards, R., Sestan, N.,

1267

and Crair, M.C. (2013). Laminar and columnar development of barrel cortex relies on

1268

thalamocortical neurotransmission. Neuron

, 970-986. 10.1016/j.neuron.2013.06.043.

1269

55.

Fuzik, J., Rehman, S., Girach, F., Miklosi, A.G., Korchynska, S., Arque, G., Romanov,

1270

R.A., Hanics, J., Wagner, L., Meletis, K., et al. (2019). Brain-wide genetic mapping

1271

identifies the indusium griseum as a prenatal target of pharmacologically unrelated

1272

psychostimulants. Proc Natl Acad Sci U S A

116

, 25958-25967.

1273

10.1073/pnas.1904006116.

1274

56.

Zoghbi, H.Y., and Bear, M.F. (2012). Synaptic dysfunction in neurodevelopmental

1275

disorders associated with autism and intellectual disabilities. Cold Spring Harb Perspect

1276

Biol

. 10.1101/cshperspect.a009886.

1277

57.

Nicolini, C., and Fahnestock, M. (2018). The valproic acid-induced rodent model of

1278

autism. Exp Neurol

299

, 217-227. 10.1016/j.expneurol.2017.04.017.

1279

58.

Hou, Q., Wang, Y., Li, Y., Chen, D., Yang, F., and Wang, S. (2018). A Developmental

1280

Study of Abnormal Behaviors and Altered GABAergic Signaling in the VPA-Treated Rat

1281

Model of Autism. Front Behav Neurosci

, 182. 10.3389/fnbeh.2018.00182.

1282

59.

Marco, E.J., Hinkley, L.B., Hill, S.S., and Nagarajan, S.S. (2011). Sensory processing in

1283

autism: a review of neurophysiologic findings. Pediatr Res

, 48R-54R.

1284

10.1203/PDR.0b013e3182130c54.

1285

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

60.

Simonetti, T., Lee, H., Bourke, M., Leamey, C.A., and Sawatari, A. (2009). Enrichment

1286

from birth accelerates the functional and cellular development of a motor control area in

1287

the mouse. PLoS One

, e6780. 10.1371/journal.pone.0006780.

1288

61.

Cancedda, L., Putignano, E., Sale, A., Viegi, A., Berardi, N., and Maffei, L. (2004).

1289

Acceleration of visual system development by environmental enrichment. J Neurosci

1290

4840-4848. 10.1523/JNEUROSCI.0845-04.2004.

1291

62.

Travaglia, A., Bisaz, R., Sweet, E.S., Blitzer, R.D., and Alberini, C.M. (2016). Infantile

1292

amnesia reflects a developmental critical period for hippocampal learning. Nat Neurosci

1293

, 1225-1233. 10.1038/nn.4348.

1294

63.

Telley, L., Cadilhac, C., Cioni, J.M., Saywell, V., Jahannault-Talignani, C., Huettl, R.E.,

1295

Sarrailh-Faivre, C., Dayer, A., Huber, A.B., and Ango, F. (2016). Dual Function of NRP1

1296

in Axon Guidance and Subcellular Target Recognition in Cerebellum. Neuron

, 1276-

1297

1291. 10.1016/j.neuron.2016.08.015.

1298

64.

Pollerberg, G.E., and Mack, T.G. (1994). Cell adhesion molecule SC1/DMGRASP is

1299

expressed on growing axons of retina ganglion cells and is involved in mediating their

1300

extension on axons. Dev Biol

165

, 670-687. 10.1006/dbio.1994.1284.

1301

65.

Guijarro, P., Wang, Y., Ying, Y., Yao, Y., Jieyi, X., and Yuan, X. (2013). In vivo

1302

knockdown of cKit impairs neuronal migration and axonal extension in the cerebral

1303

cortex. Dev Neurobiol

, 871-887. 10.1002/dneu.22107.

1304

66.

Guenthner, C.J., Miyamichi, K., Yang, H.H., Heller, H.C., and Luo, L. (2013). Permanent

1305

genetic access to transiently active neurons via TRAP: targeted recombination in active

1306

populations. Neuron

, 773-784. 10.1016/j.neuron.2013.03.025.

1307

67.

Barth, A.L., Gerkin, R.C., and Dean, K.L. (2004). Alteration of neuronal firing properties

1308

after in vivo experience in a FosGFP transgenic mouse. J Neurosci

, 6466-6475.

1309

10.1523/JNEUROSCI.4737-03.2004.

1310

68.

Sun, X., Bernstein, M.J., Meng, M., Rao, S., Sorensen, A.T., Yao, L., Zhang, X.,

1311

Anikeeva, P.O., and Lin, Y. (2020). Functionally Distinct Neuronal Ensembles within the

1312

Memory Engram. Cell

181

, 410-423 e417. 10.1016/j.cell.2020.02.055.

1313

69.

Sorensen, A.T., Cooper, Y.A., Baratta, M.V., Weng, F.J., Zhang, Y., Ramamoorthi, K.,

1314

Fropf, R., LaVerriere, E., Xue, J., Young, A., et al. (2016). A robust activity marking

1315

system for exploring active neuronal ensembles. Elife

. 10.7554/eLife.13918.

1316

70.

Sim, S., Antolin, S., Lin, C.W., Lin, Y., and Lois, C. (2013). Increased cell-intrinsic

1317

excitability induces synaptic changes in new neurons in the adult dentate gyrus that

1318

require Npas4. J Neurosci

, 7928-7940. 10.1523/JNEUROSCI.1571-12.2013.

1319

71.

Lyford, G.L., Yamagata, K., Kaufmann, W.E., Barnes, C.A., Sanders, L.K., Copeland,

1320

N.G., Gilbert, D.J., Jenkins, N.A., Lanahan, A.A., and Worley, P.F. (1995). Arc, a growth

1321

factor and activity-regulated gene, encodes a novel cytoskeleton-associated protein that

1322

is enriched in neuronal dendrites. Neuron

, 433-445. 10.1016/0896-6273(95)90299-6.

1323

72.

Sun, M.R., Steward, A.C., Sweet, E.A., Martin, A.A., and Lipinski, R.J. (2021).

1324

Developmental malformations resulting from high-dose maternal tamoxifen exposure in

1325

the mouse. PLoS One

, e0256299. 10.1371/journal.pone.0256299.

1326

73.

Mehasseb, M.K., Bell, S.C., and Habiba, M.A. (2010). Neonatal administration of

1327

tamoxifen causes disruption of myometrial development but not adenomyosis in the

1328

C57/BL6J mouse. Reproduction

139

, 1067-1075. 10.1530/REP-09-0443.

1329

74.

Haar, S., Berman, S., Behrmann, M., and Dinstein, I. (2016). Anatomical Abnormalities

1330

in Autism? Cereb Cortex

, 1440-1452. 10.1093/cercor/bhu242.

1331

75.

Sussman, D., Leung, R.C., Vogan, V.M., Lee, W., Trelle, S., Lin, S., Cassel, D.B.,

1332

Chakravarty, M.M., Lerch, J.P., Anagnostou, E., and Taylor, M.J. (2015). The autism

1333

puzzle: Diffuse but not pervasive neuroanatomical abnormalities in children with ASD.

1334

Neuroimage Clin

, 170-179. 10.1016/j.nicl.2015.04.008.

1335

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

76.

Kahn, R.S., Sommer, I.E., Murray, R.M., Meyer-Lindenberg, A., Weinberger, D.R.,

1336

Cannon, T.D., O'Donovan, M., Correll, C.U., Kane, J.M., van Os, J., and Insel, T.R.

1337

(2015). Schizophrenia. Nat Rev Dis Primers

, 15067. 10.1038/nrdp.2015.67.

1338

77.

Robertson, C.E., and Baron-Cohen, S. (2017). Sensory perception in autism. Nat Rev

1339

Neurosci

, 671-684. 10.1038/nrn.2017.112.

1340

78.

Paulsen, B., Velasco, S., Kedaigle, A.J., Pigoni, M., Quadrato, G., Deo, A.J., Adiconis,

1341

X., Uzquiano, A., Sartore, R., Yang, S.M., et al. (2022). Autism genes converge on

1342

asynchronous development of shared neuron classes. Nature

602

, 268-273.

1343

10.1038/s41586-021-04358-6.

1344

79.

Zhang, L.I., and Poo, M.M. (2001). Electrical activity and development of neural circuits.

1345

Nat Neurosci

4 Suppl

, 1207-1214. 10.1038/nn753.

1346

80.

Cline, H.T. (2001). Dendritic arbor development and synaptogenesis. Curr Opin

1347

Neurobiol

, 118-126. 10.1016/s0959-4388(00)00182-3.

1348

81.

Blankenship, A.G., and Feller, M.B. (2010). Mechanisms underlying spontaneous

1349

patterned activity in developing neural circuits. Nat Rev Neurosci

, 18-29.

1350

10.1038/nrn2759.

1351

82.

Luhmann, H.J., Sinning, A., Yang, J.W., Reyes-Puerta, V., Stuttgen, M.C., Kirischuk, S.,

1352

and Kilb, W. (2016). Spontaneous Neuronal Activity in Developing Neocortical

1353

Networks: From Single Cells to Large-Scale Interactions. Front Neural Circuits

, 40.

1354

10.3389/fncir.2016.00040.

1355

83.

Unichenko, P., Yang, J.W., Luhmann, H.J., and Kirischuk, S. (2015). Glutamatergic

1356

system controls synchronization of spontaneous neuronal activity in the murine neonatal

1357

entorhinal cortex. Pflugers Arch

467

, 1565-1575. 10.1007/s00424-014-1600-5.

1358

84.

Nadim, F., and Bucher, D. (2014). Neuromodulation of neurons and synapses. Curr

1359

Opin Neurobiol

, 48-56. 10.1016/j.conb.2014.05.003.

1360

85.

Kozorovitskiy, Y., Peixoto, R., Wang, W., Saunders, A., and Sabatini, B.L. (2015).

1361

Neuromodulation of excitatory synaptogenesis in striatal development. Elife

1362

10.7554/eLife.10111.

1363

86.

Hrvatin, S., Hochbaum, D.R., Nagy, M.A., Cicconet, M., Robertson, K., Cheadle, L.,

1364

Zilionis, R., Ratner, A., Borges-Monroy, R., Klein, A.M., et al. (2018). Single-cell

1365

analysis of experience-dependent transcriptomic states in the mouse visual cortex. Nat

1366

Neurosci

, 120-129. 10.1038/s41593-017-0029-5.

1367

87.

Tyssowski, K.M., DeStefino, N.R., Cho, J.H., Dunn, C.J., Poston, R.G., Carty, C.E.,

1368

Jones, R.D., Chang, S.M., Romeo, P., Wurzelmann, M.K., et al. (2018). Different

1369

Neuronal Activity Patterns Induce Different Gene Expression Programs. Neuron

1370

530-546 e511. 10.1016/j.neuron.2018.04.001.

1371

88.

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T.,

1372

Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., et al. (2012). Fiji: an open-source

1373

platform for biological-image analysis. Nat Methods

, 676-682. 10.1038/nmeth.2019.

1374

89.

Langmead, B., and Salzberg, S.L. (2012). Fast gapped-read alignment with Bowtie 2.

1375

Nat Methods

, 357-359. 10.1038/nmeth.1923.

1376

90.

Kim, D., Paggi, J.M., Park, C., Bennett, C., and Salzberg, S.L. (2019). Graph-based

1377

genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol

1378

, 907-915. 10.1038/s41587-019-0201-4.

1379

91.

Koopmans, F., van Nierop, P., Andres-Alonso, M., Byrnes, A., Cijsouw, T., Coba, M.P.,

1380

Cornelisse, L.N., Farrell, R.J., Goldschmidt, H.L., Howrigan, D.P., et al. (2019). SynGO:

1381

An Evidence-Based, Expert-Curated Knowledge Base for the Synapse. Neuron

103

1382

217-234 e214. 10.1016/j.neuron.2019.05.002.

1383

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint

92.

Wolock, S.L., Lopez, R., and Klein, A.M. (2019). Scrublet: Computational Identification

1384

of Cell Doublets in Single-Cell Transcriptomic Data. Cell Syst

, 281-291 e289.

1385

10.1016/j.cels.2018.11.005.

1386

93.

Wolf, F.A., Angerer, P., and Theis, F.J. (2018). SCANPY: large-scale single-cell gene

1387

expression data analysis. Genome Biol

, 15. 10.1186/s13059-017-1382-0.

1388

94.

Hao, Y., Hao, S., Andersen-Nissen, E., Mauck, W.M., 3rd, Zheng, S., Butler, A., Lee,

1389

M.J., Wilk, A.J., Darby, C., Zager, M., et al. (2021). Integrated analysis of multimodal

1390

single-cell data. Cell

184

, 3573-3587 e3529. 10.1016/j.cell.2021.04.048.

1391

95.

Robinson, M.D., McCarthy, D.J., and Smyth, G.K. (2010). edgeR: a Bioconductor

1392

package for differential expression analysis of digital gene expression data.

1393

Bioinformatics

, 139-140. 10.1093/bioinformatics/btp616.

1394

96.

Street, K., Risso, D., Fletcher, R.B., Das, D., Ngai, J., Yosef, N., Purdom, E., and

1395

Dudoit, S. (2018). Slingshot: cell lineage and pseudotime inference for single-cell

1396

transcriptomics. BMC Genomics

, 477. 10.1186/s12864-018-4772-0.

1397

97.

McCarthy, D.J., Campbell, K.R., Lun, A.T., and Wills, Q.F. (2017). Scater: pre-

1398

processing, quality control, normalization and visualization of single-cell RNA-seq data

1399

in R. Bioinformatics

, 1179-1186. 10.1093/bioinformatics/btw777.

1400

98.

Finak, G., McDavid, A., Yajima, M., Deng, J., Gersuk, V., Shalek, A.K., Slichter, C.K.,

1401

Miller, H.W., McElrath, M.J., Prlic, M., et al. (2015). MAST: a flexible statistical

1402

framework for assessing transcriptional changes and characterizing heterogeneity in

1403

single-cell RNA sequencing data. Genome Biol

, 278. 10.1186/s13059-015-0844-5.

1404

99.

Conway, J.R., Lex, A., and Gehlenborg, N. (2017). UpSetR: an R package for the

1405

visualization of intersecting sets and their properties. Bioinformatics

, 2938-2940.

1406

10.1093/bioinformatics/btx364.

1407

100. Wu, T., Hu, E., Xu, S., Chen, M., Guo, P., Dai, Z., Feng, T., Zhou, L., Tang, W., Zhan,

1408

L., et al. (2021). clusterProfiler 4.0: A universal enrichment tool for interpreting omics

1409

data. Innovation (Camb)

, 100141. 10.1016/j.xinn.2021.100141.

1410

101. Yelhekar, T.D., Druzin, M., and Johansson, S. (2017). Contribution of Resting

1411

Conductance, GABA(A)-Receptor Mediated Miniature Synaptic Currents and

1412

Neurosteroid to Chloride Homeostasis in Central Neurons. eNeuro

1413

10.1523/ENEURO.0019-17.2017.

1414

102. Yelhekar, T.D., Druzin, M., Karlsson, U., Blomqvist, E., and Johansson, S. (2016). How

1415

to Properly Measure a Current-Voltage Relation?-Interpolation vs. Ramp Methods

1416

Applied to Studies of GABAA Receptors. Front Cell Neurosci

, 10.

1417

10.3389/fncel.2016.00010.

1418

103. Klein, S., Staring, M., Murphy, K., Viergever, M.A., and Pluim, J.P. (2010). elastix: a

1419

toolbox for intensity-based medical image registration. IEEE Trans Med Imaging

1420

196-205. 10.1109/TMI.2009.2035616.

1421

104. Hafemeister, C., and Satija, R. (2019). Normalization and variance stabilization of

1422

single-cell RNA-seq data using regularized negative binomial regression. Genome Biol

1423

, 296. 10.1186/s13059-019-1874-1.

1424

105. Amezquita, R.A., Lun, A.T.L., Becht, E., Carey, V.J., Carpp, L.N., Geistlinger, L., Marini,

1425

F., Rue-Albrecht, K., Risso, D., Soneson, C., et al. (2020). Orchestrating single-cell

1426

analysis with Bioconductor. Nat Methods

, 137-145. 10.1038/s41592-019-0654-x.

1427

106. Yap, E.L., Pettit, N.L., Davis, C.P., Nagy, M.A., Harmin, D.A., Golden, E., Dagliyan, O.,

1428

Lin, C., Rudolph, S., Sharma, N., et al. (2021). Bidirectional perisomatic inhibitory

1429

plasticity of a Fos neuronal network. Nature

590

, 115-121. 10.1038/s41586-020-3031-0.

1430

1431

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 January 4, 2024.

;

https://doi.org/10.1101/2024.01.03.572456

doi:

bioRxiv preprint