Early-stage reduction of the dendritic complexity in basolateral amygdala of a transgenic mouse model of Alzheimer's disease

Accepted Manuscript Early-stage reduction of the dendritic complexity in basolateral amygdala of a transgenic mouse model of Alzheimer's disease Congdi Guo, Ben Long, Yarong Hu, Jing Yuan, Hui Gong, Xiangning Li PII:

S0006-291X(17)30560-0

DOI:

10.1016/j.bbrc.2017.03.094

Reference:

YBBRC 37481

To appear in:

Biochemical and Biophysical Research Communications

Received Date: 10 March 2017 Accepted Date: 19 March 2017

Please cite this article as: C. Guo, B. Long, Y. Hu, J. Yuan, H. Gong, X. Li, Early-stage reduction of the dendritic complexity in basolateral amygdala of a transgenic mouse model of Alzheimer's disease, Biochemical and Biophysical Research Communications (2017), doi: 10.1016/j.bbrc.2017.03.094. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1

Early-stage reduction of the dendritic complexity in basolateral amygdala of a

2

transgenic mouse model of Alzheimer’s disease

3

Congdi Guo, Ben Long, Yarong Hu, Jing Yuan, Hui Gong, Xiangning Li* Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong

5

University of Science and Technology, Wuhan 430074, China.

6

* Correspondence should be addressed to X.L. ([email protected])

7

RI PT

4

Abstract

9

Alzheimer’s disease is a representative age-related neurodegenerative disease that could result in

10

loss of memory and cognitive deficiency. However, the precise onset time of Alzheimer’s disease

11

affecting neuronal circuits and the mechanisms underlying the changes are not clearly known. To

12

address the neuroanatomical changes during the early pathologic developing process, we acquired

13

the neuronal morphological characterization of AD in APP/PS1 double-transgenic mice using the

14

Micro-Optical Sectioning Tomography system. We reconstructed the neurons in 3D datasets with

15

a resolution of 0.32×0.32×1 µm and used the Sholl method to analyze the anatomical

16

characterization of the dendritic branches. The results showed that, similar to the progressive

17

change in amyloid plaques, the number of dendritic branches were significantly decreased in

18

9-month-old mice. In addition, a distinct reduction of dendritic complexity occurred in third and

19

fourth-order dendritic branches of 9-month-old mice, while no significant changes were identified

20

in these parameters in 6-month-old mice. At the branch-level, the density distribution of dendritic

21

arbors in the radial direction decreased in the range of 40-90 µm from the neuron soma in

22

6-month-old mice. These changes in the dendritic complexity suggest that these reductions

23

contribute to the progressive cognitive impairment seen in APP/PS1 mice. This work may yield

24

insights into the early changes in dendritic abnormality and its relevance to dysfunctional

25

mechanisms of learning, memory and emotion in Alzheimer’s disease.

26

Key words: Alzheimer, basolateral amygdaloid nucleus, dendritic complexity, Micro-Optical

27

Sectioning Tomography

AC C

EP

TE D

M AN U

SC

8

28

1

ACCEPTED MANUSCRIPT Introduction

30

Alzheimer’s disease (AD) is the most common age-related neurodegenerative disease, and the

31

pathogenesis of AD has always been a popular issue in modern neuroscience research [1]. AD can

32

be characterized by loss of memory and cognitive defects in behavioral pathology, which have

33

been the focus of several studies regarding the mechanism and treatment of AD [2]. Although

34

extracellular depositions of amyloid-beta peptide (Aβ) and intracellular neurofibrillary tangles

35

(NFTs) can be found in some AD patients and animal models, clinical trials have failed to give an

36

effective treatment to prevent, halt, or reverse AD [3]. A number of studies indicate that early

37

intervention treatment, before the onset of cognitive symptoms, would be more effective, but the

38

clinical onset time of AD and the mechanism underlying the changes in neuronal circuits is not

39

clearly known [4].

40

Some subcortical regions are affected by deposits with the progressive development of AD in

41

patients and transgenic animal models [5, 6]. As the main brain area controlling emotional

42

activities and associative memories, the basolateral amygdaloid nucleus (BLA) is significantly

43

affected showing considerable shrinkage, distortion, loss of neurons and neuronal morphological

44

alterations [7, 8]. These dendritic abnormalities as key hallmarks in the early stages of the AD,

45

adumbrate neuroanatomical degeneration, including dystrophic neurites, reduction of dendritic

46

complexity and loss of dendritic spines [9].

47

APPswe/PS1dE9 (hereinafter, APP/PS1) double-transgenic mice express a chimeric mouse/human

48

amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9) both

49

involved in neurons of the central nervous system [10]. Previous reports revealed that plaque

50

depositions were found in the brains of the transgenic mice starting at 6 months old, while there

51

was a reduction in the dendritic complexity of BLA at 12-14 months old [6, 11, 12]. However, the

52

mouse model of AD displays cognitive and memory deficits at 3-8 months old [13]. The

53

morphological integrity of neurons identifying the amygdaloid circuits was not clear with respect

54

to the progressive changes in early stages of AD.

55

To address the neural changes on a large-scale, during the early pathological development of AD,

56

we used the APP/PS1 double-transgenic mice as AD models. By combining Golgi-Cox staining

57

with Micro-Optical Sectioning Tomography (MOST), we acquired the reconstructed projection

58

neurons in the BLA at developmental stages of AD and examined their dendritic morphology [14,

AC C

EP

TE D

M AN U

SC

RI PT

29

2

ACCEPTED MANUSCRIPT 59

15]. The results showed that the progressive Aβ depositions and reduction of dendritic complexity

60

could be identified in 6- and 9-month-old mice.

61 Materials and Methods

63

Animals

64

APP/PS1 mice were used as Alzheimer’s disease models, and C57BL/6J mice were used as

65

controls, with a 12-h light/dark cycle (8:00 a.m. to 8:00 p.m.), in stable conditions of temperature

66

(22°C) and humidity (60%), with food and water ad libitum. Only male mice at the corresponding

67

ages of 6 and 9 months were used in this study. Five mice were used in each group for

68

immunohistochemistry/Nissl staining, and 2 mice were used in each group for whole-brain

69

imaging. All protocols and procedures were conducted according to the guidelines of the

70

Experimental Animal Ethics Committee at Huazhong University of Science and Technology.

71

Whole brain Golgi-Cox staining

72

Whole brain Golgi-Cox staining was done according to the modified Cox method [14]. Briefly, the

73

mouse was anesthetized with 1% sodium pentobarbital, and the whole brain was removed from

74

the skull and then suspended in fresh Golgi-Cox solution for fixation and impregnation. The brain

75

was stored in a brown bottle at 20-25°C for at least 60 days. The Golgi-Cox solution consisted of

76

1 g mercuric chloride (HgCl2), 1 g potassium dichromate (K2Cr2O7), and 0.8 g potassium

77

chromate (K2CrO4) dissolved in 80 g distilled water. Then, the brain was immersed in a solution of

78

1% lithium hydroxide (LiOH) for 24 h and rinsed for another 24 h. The rinsed brain was

79

sequentially immersed in 50%, 70%, 85%, 95%, 100% alcohol, 100% alcohol–acetone (1:1), and

80

100% acetone (2x). After dehydration, the brain was infiltrated by 50%, 75%, and 100% Spurr

81

resin (2x) and was maintained at 60°C for 36 h for polymerization.

82

All the Spurr reagents (SPI, USA) were freshly prepared. The 100% Spurr resin contained 10 g

83

vinylcyclohexene dioxide (ERL- 4221), 7.6 g diglycidyl ether of polypropylene glycol (DER-736),

84

26 g nonenyl succinic anhydride (NSA), and 0.2 g dimethylaminoethanol (DMAE).

85

Immunohistochemistry and Nissl staining

86

After perfusion with 0.01 M phosphate buffer saline (PBS, Sigma-Aldrich, Cat #P3813, pH 7.4)

87

followed by 4% paraformaldehyde (Sigma-Aldrich, Cat #158127) in PBS, the brain was soaked in

88

30% sucrose until it sank. Subsequently, the brain was frozen and sectioned at 25 µm on a Leica

AC C

EP

TE D

M AN U

SC

RI PT

62

3

ACCEPTED MANUSCRIPT microtome (Leica CM1950, Germany) and the serial sections were stored at -20°C.

90

For immunohistochemistry staining, the frozen sections were blocked with 0.3% hydrogen

91

peroxide for 30 min at 37°C to quench endogenous peroxidase activity. For antigen retrieval, the

92

sections were water-bathed in citrate buffer at approximately 95°C for 15 min and washed in 0.3%

93

Triton X-100 in PBS. Then, following a blocking step with normal horse serum (Vector

94

Laboratories, Cat # PK-6102), the sections were incubated with primary antibody of mouse

95

anti-Aβ (COVANCE, Cat #SIG-39300, at 1:1000 dilution) overnight at 4°C. Amplification was

96

performed with biotinylated horse anti-mouse IgG (Vector Laboratories, Cat #PK-6102, 1:200)

97

and avidin-biotin peroxidase complex (Vector Laboratories, Cat #PK-6102). Finally, the immune

98

complex was visualized by incubation with 3, 3'-diaminobenzidine (DAB, Vector Laboratories,

99

Cat #SK-4100, at 1:50) for 5 min. The stained sections were progressively dehydrated in 75%,

M AN U

SC

RI PT

89

85%, 95%, 100% (2x) alcohol for 5 min in each solution before being covered.

101

For Nissl staining, another set of adjacent frozen sections was washed with PBS and then stained

102

with 0.5% thionin acetate salt (Sigma-Aldrich, Cat #861340) solution for 10 min. Subsequently,

103

the sections were rinsed quickly in distilled water. The stained sections were dehydrated and

104

covered. Coronal sections processed for immunohistochemistry and Nissl staining were examined

105

with a Nikon Ni-E microscope.

106

Imaging acquisition & preprocessing

107

The mouse brains were sectioned and imaged automatically using the MOST system [15], with a

108

voxel size of 0.32 × 0.32 × 1 µm. MOST is made up of a microtome, light microscope, and image

109

recorder. It simultaneously performs thin sectioning and imaging (40x, numerical aperture 0.8).

110

The coordinates of all the image tiles were recorded during sectioning. Benefiting from the high

111

precision positioning system (programmable precision: 100 nm and feedback accuracy: 20 nm),

112

all the image tiles were aligned automatically to reconstruct intact 3D brains according to the

113

recorded coordinates.

114

The original images were preprocessed using the customized MATLAB software [16]. Briefly, we

115

corrected the periodic noise in the image tiles using a mean projection curve, and calibrated

116

non-uniform illumination using the Rollingball algorithm. Additionally, the average intensity of

117

the tissue region in each coronal section was adjusted to a universal constant to reduce intensity

118

variation among sections.

AC C

EP

TE D

100

4

ACCEPTED MANUSCRIPT Morphology reconstruction

120

We first cropped 3D image blocks (600 × 600 × 600 µm) which encompassed the BLA (AP:

121

Bregma, -1.46 mm, DV, +4.75 mm, ML, +2.78 mm) according to a mouse brain atlas [17]. The

122

image blocks were intensity-inverted using ImageJ and converted into a large data access format

123

(LDA) using Amira (Visage Imaging) to enable rapid extraction and switching of sub-blocks with

124

minor sizes in arbitrary positions, which reduced the obstruction of the sight owing to intensively

125

labeled neurites and therefore ensured accuracy of neuronal tracing. The three dimensional tracing

126

and reconstruction for projection neurons was completed interactively in Amira.

127

Statistics

128

The locations of labeled neurons and plaques in the selected area of each section were determined

129

with an anatomical atlas [17]. Digital images were adjusted for brightness and contrast using

130

ImageJ software. The size and number of plaques were analyzed and the plaque load of each brain

131

region was calculated as the ratio of total size of all plaques in the selected area. Analogously, to

132

count the neuron numbers of the BLA, equidistant sections for Nissl staining were chosen. Five

133

serial equidistant sections were analyzed from each brain, and 5 brains were used in each group.

134

For neuronal complexity analysis, we first eliminated reconstructions of projection neurons which

135

did not meet the morphology criteria [18] in Amira for quantitative analysis, reviewing graph data

136

of reconstruction and volume rendering of corresponding image blocks. The selected

137

reconstructions avoided incompleteness of dendritic arbor as much as possible.

138

We measured the characterization of dendritic arbors in 3D with concentric spheres of increasing

139

diameter values (i.e., Sholl analysis) for several parameters, including total branch number, total

140

dendritic length, total number of terminal tips, branch number of each order, branch length of each

141

order, and the number of intersections. A branch was defined as one or more compartments that lie

142

between two branching points or between one branching point and a termination point. The center

143

of the soma was also regarded as a branching point. Branch order was defined as the order of the

144

branch with respect to the soma. For branches directly arising from soma, the branch order was

145

equal to 1, and if these branches showed a bifurcation point, then the branch order of child

146

branches was 2, etc. The above-mentioned parameters were measured using customized MATLAB

147

scripts. We used SPSS 22 for all statistical analysis. For total branch number, total dendritic length

148

and total number of terminal tips, one-way ANOVA was used to analyze the differences among

AC C

EP

TE D

M AN U

SC

RI PT

119

5

ACCEPTED MANUSCRIPT groups followed by Holm-Sidak multiple comparison post hoc tests. For branch number, length of

150

each branch order and Sholl analysis, two-way ANOVA was used to analyze the differences

151

among groups and branch orders or distances followed by Holm-Sidak multiple comparison post

152

hoc tests. Confidence level was set to 0.05 (p-value); all the results are presented as the mean ±

153

SEM.

154

RI PT

149

Results

156

Amyloid plaques and neuron density in developing APP/PS1 mice

157

We first investigated the pathological changes in developing AD mice. Following

158

immunohistochemistry (Fig. 1a-c) and Nissl staining (Fig. 1d-f), we extracted image blocks from

159

6-month-old and 9-month-old APP/PS1 mice (hereinafter referred to as AD_6M and AD_9M for

160

reconstruction group, respectively), and 6-month-old C57BL/6J mice (hereinafter referred to as

161

CON for reconstruction group). To obtain quantitative data on plaque levels, microscopic images

162

of anti-Aβ antibody-stained brain slices were captured. In AD_6M and AD_9M, Aβ depositions

163

were identified in several brain regions, and the number of plaques in each region was counted

164

using ImageJ. Quantification analyses showed that Aβ-immunoreactive plaques increased in

165

number and load in BLA, lateral entorhinal cortex (LEnt), cingulate cortex (Cg) and other regions

166

(Fig. 1g, from 5 mice; 5 slices averaged per mouse). In addition, the plaque load in these regions

167

increased in different ranges. In the BLA, it changed from 0.14-0.86% in AD_6M to 1.98-2.90%

168

in AD_9M (Fig. 1h), which indicated that the amygdala was affected more severely than the LEnt

169

and Cg. Quantitative loss of neuron number was not observed in these transgenic mice at 6 and 9

170

months old.

M AN U

TE D

EP

AC C

171

SC

155

172

Three-dimensional morphology of neurons in BLA

173

To analyze the effects of amyloid-immunoreactive plaques on neuronal morphology, we used

174

modified Golgi staining for the whole brain and the MOST system to acquire the fine structure of

175

neurons. The quantified data of 3D measurements for the whole neuron are more precise with

176

higher levels of unbiasedness. Here, with neuroanatomical structures illustrated by the MOST

177

system, we acquired the 3D morphology of neurons. As shown in Fig. 2a, a large number of

178

neurons and processes, from superficial to deep layers, were detected in the maximal projections 6

ACCEPTED MANUSCRIPT 179

of serial coronal sections. The amygdala can be seen on the ventral side of the corpus callosum

180

(Fig. 2b), which showed that most of the dendritic branches of the neurons in the BLA were

181

distributed locally. In the magnified coronal sections of local regions, the morphologies of the

182

neuronal soma and the dendritic branches can be examined in detail (Fig. 2c).

RI PT

183 Even for the dendrites mainly distributed in local regions of the BLA, the dendritic branch

185

projected in cubic space with a diameter of over 200 µm in every direction (Fig. 2e). With the

186

MOST system, the resolution of the 3D dataset for whole brain is 0.32×0.32×1 µm. Then, we can

187

trace each dendritic branch and reconstruct the fine structure of the neuron with multiple order

188

branches. All of the reconstructed neurons analyzed here exhibited a typical pyramidal

189

morphology, forming extensive dendritic branching.

190

To investigate the occurrence patterns of dendritic complexity changes, we examined branch-level

191

measures of dendritic complexity, including branch number and length of each branch order.

192

Dendrograms of representative neurons with long projections from different groups are shown in

193

Fig. 3. These dendrograms demonstrated the overall branching patterns of the dendritic trees and

194

were scaled in terms of length and complexity for further analysis. Here, we compared these

195

reconstructed neurons at different developmental stages. The number of reconstructed neurons for

196

measuring dendritic complexity was 15 for the AD_6M group, 19 for the AD_9M group and 18

197

for the CON group from 2 mice in each group. Verification of the quality of neurons in the precise

198

whole brain dataset showed excellent recovery of dendritic morphology for quantitative analyses.

M AN U

TE D

EP

AC C

199

SC

184

200

Dendritic changes in the APP/PS1 mice

201

We examined neuron-level measures of dendritic complexity including total branch number (Fig.

202

4 a), total dendritic length (Fig. 4c) and total number of terminal tips (Fig. 4e). The whole brain

203

datasets of each group were used for neuron reconstructions. Compared to the values for the CON

204

group (66.17 ± 3.59 for total branch number and 38.83 ± 1.73 for total terminal tips number,

205

p<0.001, n=18) (Fig. 4a and 4e), there were significant decreases in the total branch number and

206

total terminal tips number of AD_9M (47 ± 2.99 and 28.05 ± 1.64, respectively, n=19, mean ±

207

SEM). However, when these values for CON were compared to those of AD_6M (56.40 ± 4.33

208

and 33.2 ± 2.37, respectively, n=15), there were no significant differences. It suggested that there 7

ACCEPTED MANUSCRIPT was a reduction in dendritic complexity as the mice grew from 6 months to 9 months old.

210

Nevertheless, there was no significant difference in total dendritic length among the three groups

211

(2955.06 ± 136.61 µm for CON, 3206.74 ± 276.99 µm for AD_6M, and 2800.59 ± 212.86 µm for

212

AD_9M, p=0.408) (Fig. 4c). These results demonstrate that the reduction of dendritic complexity

213

in the AD_9M group is mainly reflected in topology alteration.

214

To analyze the dendrite-level changes in the neurons, we examined the branch number and length

215

of each order of dendrites in different groups (Fig. 4b and 4d). In terms of the number of each

216

order, the tertiary branch number in AD_9M was significantly less than that in both the AD_6M

217

and CON groups (vs. CON, p=0.008; vs. AD-6M, p=0.029), while the number of third and

218

fourth-order branches in AD_9M was significantly less than that in the CON group (p=0.002).

219

However, for other branch orders, there were no significant differences among the three groups

220

(Fig. 4b). This suggests that the decrease of total branch number in AD_9M group primarily

221

occurs in the tertiary and fourth–order branches. Interestingly, no significant changes were found

222

in branch length of each order (first to fifth order) among the three groups (Fig. 4d), which

223

indicates the stable feature of branch length between pathological groups and the CON group, per

224

order as well as per neuron.

225

To characterize the detailed change in the dendritic tree, 3D Sholl analysis was used to measure

226

the number of intersections along the dendritic arbors with concentric spheres centered in the

227

soma. This analysis can directly present the density distribution of dendritic arbors in a radial

228

direction. As shown in Fig. 4f, dendrites of AD_6M had fewer intersections in the 40-90 µm range

229

from the soma than those of CON (confidence level p=0.05), while dendrites of AD_9M had

230

significantly fewer intersections in the 40-100 µm range from the soma than those of CON

231

(p=0.007). In the range farther away from the soma (120-250 µm), the dendritic branch density of

232

AD_9M was less than that of AD_6M, but the difference between them was not significant.

233

Additionally, the range with the peak number of intersections was approximately 80-90 µm in

234

CON and was extended to approximately 110-130 µm in the other two groups. Alteration of the

235

intersection peak may closely relate to significant decreases in the connection of BLA neurons

236

with adjacent neurons.

AC C

EP

TE D

M AN U

SC

RI PT

209

237 238

Discussion 8

ACCEPTED MANUSCRIPT In this study, the reconstructed projection neurons in the BLA, at different ages of the APP/PS1

240

transgenic model, were analyzed for pathological changes in dendritic arbor and compared to the

241

changes in C57BL/6J mice. The results showed that the following: (1) at the neuron-level, the

242

AD_9M group had significantly fewer branches compared to the CON group, while there was no

243

significant difference in total branch number and length between the AD_6M and CON groups; (2)

244

at the branch-level, dendritic distribution density showed a significant decrease in the 40-90 µm

245

range from the soma in the AD_6M group.

246

Here, using the modified Golgi-Cox staining for whole brain and the MOST system, the spatial

247

locations and global morphological changes in neurons were analyzed in 3 dimensions. As a

248

powerful and commonly used staining technique, the Golgi staining method can provide intact

249

neuronal morphology and traces of the processes [14]. Based on 3D datasets with a high resolution

250

of 0.32×0.32×1 µm, we can study the neuron-level and branch-level changes in the AD mice at

251

different ages. The result that no significant decrease was found in total dendritic length of

252

projection neurons in the BLA of 9 months old mice was consistent with a previous study for

253

12-14 months old mice [6]. At the branch-level, we found that branch numbers in AD_9M were

254

pared down mainly by the branches of the third order and beyond. Since the branch length did not

255

show a significant decrease in each order, we can assume that branches above the third order

256

elongate normally in AD_9M. Moreover, branch-level changes in the 6-month-old mice were

257

identified.

258

As shown in Fig. 2, plaques were found in the APP/PS1 mice at 6 months old, while the transgenic

259

mice displayed cognitive and memory deficits before they were 9 months old [13]. Although the

260

Aβ depositions do not correlate well with the cognitive deficits and AD dementia [19], the

261

negative linear correlation between reduction of the dendritic complexity and Aβ deposition hints

262

that Aβ deposition surrounding dendrites may lead to dendritic breakage [20]. AD may be better

263

characterized by the reduction of the dendritic complexity [21]. Here, Sholl analysis showed a

264

significant decrease in the dendritic branch density in the range of 40-90 µm from the neuron

265

soma in 6- and 9-month-old mice, while no significant changes were found in the more proximal

266

and distal ranges. These results present alterations of dendritic complexity from the aspect of

267

spatial distribution and may account for why EPSPs generated at synapses in a specific range are

268

influenced in early-stage diagnosis. These progressive results will bring about a greater

AC C

EP

TE D

M AN U

SC

RI PT

239

9

ACCEPTED MANUSCRIPT understanding for the study of pathogenesis and therapeutic target identification for AD.

270

Early-stage AD diagnosis and intervention is an effective way to avoid the transition from

271

cognitively normal to impaired. Based on the present understanding of AD, the pathological

272

lesions may change at the level of genetic, cellular or neuronal networks [22]; therefore, AD has

273

been characterized as a large-scale neuronal dysfunction disease [23, 24]. There is a need for

274

combined progress in diverse disciplines of science, technology, and medicine with basic clinical

275

research for early-stage AD detection [4, 21], especially using 3D imaging with high resolution. In

276

the present study, no significant changes were identified at the neuron-level in 6-month-old mice,

277

while the dendrites reduced in a specific range around the soma. These data remind us of the

278

necessity of monitoring the global and the detailed structure of neurons for future studies in early

279

diagnosis. Certainly, the Micro-Optical Sectioning Tomography system is a powerful tool in

280

neuronal morphology or neuronal circuit research whether basic or clinical, especially for

281

comparison between different brain areas in 3 dimensions.

282

Combined with immunohistochemical results, we demonstrated that the dendritic complexity of

283

projection neurons in the BLA has an incremental alteration along with the progress of

284

Alzheimer’s: there was a significant reduction of dendritic complexity in 9-month-old mice, while

285

few dendritic alterations were found in 6-month-old mice. In conclusion, we found that the

286

dendritic complexity in the BLA showed a significant reduction in branch number and local

287

branching density in early-stage APP/PS1 mice, while no significant change was found in the

288

dendritic length. This study contributes towards a better understanding of the dendritic

289

mechanisms underlying AD at early time points.

290

Acknowledgements

291

We thank Dr. Tonghui Xu and all members of MOST group for help in experiments and comments

292

on manuscript writing. We also thank the Optical Bioimaging Core Facility of WNLO-HUST for

293

the support in data acquisition.

294

This work is supported by National Natural Science Foundation of China (No. 81171067), Science

295

Fund for Creative Research Group of China (No. 61421064), the Natural Science Foundation of

296

Hubei Province (No. 2015CFB448) and Director Fund of WNLO. The funders had no role in

297

study design, data collection and analysis, decision to publish, or preparation of the manuscript.

AC C

EP

TE D

M AN U

SC

RI PT

269

10

ACCEPTED MANUSCRIPT 298

References

299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341

[1] F.K. Wiseman, T. Al-Janabi, J. Hardy, et al., A genetic cause of Alzheimer disease: mechanistic insights from Down syndrome, Nature reviews. Neuroscience, 16 (2015) 564-574. [2] S.M. Ryan,A.M. Kelly, Exercise as a pro-cognitive, pro-neurogenic and anti-inflammatory intervention in transgenic mouse models of Alzheimer's disease, Ageing Res Rev, 27 (2016) 77-92. Alzheimer disease treatments, Nat Rev Neurol, 12 (2016) 56-61.

RI PT

[3] E.M. Reiman, J.B. Langbaum, P.N. Tariot, et al., CAP--advancing the evaluation of preclinical [4] J. Feger,E.C. Hirsch, In search of innovative therapeutics for neuropsychiatric disorders: the case of neurodegenerative diseases, Ann Pharm Fr, 73 (2015) 3-12.

[5] J. Brettschneider, K. Del Tredici, V.M. Lee, et al., Spreading of pathology in neurodegenerative diseases: a focus on human studies, Nature reviews. Neuroscience, 16 (2015) 109-120.

SC

[6] S. Knafo, C. Venero, P. Merino-Serrais, et al., Morphological alterations to neurons of the amygdala and impaired fear conditioning in a transgenic mouse model of Alzheimer's disease, The Journal of pathology, 219 (2009) 41-51.

M AN U

[7] S. Knafo, Amygdala in Alzheimer's Disease, The Amygdala - A Discrete Multitasking Manager, InTech2012.

[8] O. von Bohlen und Halbach,K. Unsicker, Morphological alterations in the amygdala and hippocampus of mice during ageing, Eur J Neurosci, 16 (2002) 2434-2440. [9] J.N. Cochran, A.M. Hall,E.D. Roberson, The dendritic hypothesis for Alzheimer's disease pathophysiology, Brain Research Bulletin, 103 (2014) 18-28.

[10] S.J. Webster, A.D. Bachstetter, P.T. Nelson, et al., Using mice to model Alzheimer's dementia: an overview of the clinical disease and the preclinical behavioral changes in 10 mouse models,

TE D

Frontiers in genetics, 5 (2014) 88.

[11] J.L. Jankowsky, D.J. Fadale, J. Anderson, et al., Mutant presenilins specifically elevate the levels of the 42 residue beta-amyloid peptide in vivo: evidence for augmentation of a 42-specific gamma secretase, Hum Mol Genet, 13 (2004) 159-170. [12] C. Janus,D. Westaway, Transgenic mouse models of Alzheimer's disease, Physiology & behavior,

EP

73 (2001) 873-886.

[13] S.J. Webster, A.D. Bachstetter, P.T. Nelson, et al., Using mice to model Alzheimer’s dementia: an overview of the clinical disease and the preclinical behavioral changes in ten mouse models,

AC C

Frontiers in Genetics, 5 (2014).

[14] B. Zhang, A. Li, Z. Yang, et al., Modified Golgi-Cox method for micrometer scale sectioning of the whole mouse brain, Journal of Neuroscience Methods, 197 (2011) 1-5.

[15] A. Li, H. Gong, B. Zhang, et al., Micro-Optical Sectioning Tomography to Obtain a High-Resolution Atlas of the Mouse Brain, Science, 330 (2010) 1404-1408.

[16] W. Ding, A. Li, J. Wu, et al., Automatic macroscopic density artefact removal in a Nissl-stained microscopic atlas of whole mouse brain, Journal of Microscopy, 251 (2013) 168-177. [17] Paxinos George ,F.K.B. J., The mouse brain in stereotaxic coordinates 2ed., A Harcourt Science and Technology Company, San Diego, California ,USA, 2001. [18] P. Sah, E.S. Faber, M. Lopez De Armentia, et al., The amygdaloid complex: anatomy and physiology, Physiological reviews, 83 (2003) 803-834. [19] A.M. Foley, Z.M. Ammar, R.H. Lee, et al., Systematic review of the relationship between amyloid-beta levels and measures of transgenic mouse cognitive deficit in Alzheimer's disease, 11

ACCEPTED MANUSCRIPT Journal of Alzheimer's disease : JAD, 44 (2015) 787-795. [20] J. Grutzendler, K. Helmin, J. Tsai, et al., Various dendritic abnormalities are associated with fibrillar amyloid deposits in Alzheimer's disease, Ann N Y Acad Sci, 1097 (2007) 30-39. [21] K.L. Viola, J. Sbarboro, R. Sureka, et al., Towards non-invasive diagnostic imaging of early-stage Alzheimer's disease, Nat Nanotechnol, 10 (2015) 91-98. [22] D.M. Holtzman, J.C. Morris,A.M. Goate, Alzheimer's disease: the challenge of the second century, Science translational medicine, 3 (2011) 77sr71. [23] Y. Huang,L. Mucke, Alzheimer Mechanisms and Therapeutic Strategies, Cell, 148 (2012)

RI PT

342 343 344 345 346 347 348 349 350 351 352

1204-1222.

[24] M.A. Busche,A. Konnerth, Neuronal hyperactivity--A key defect in Alzheimer's disease?, Bioessays, 37 (2015) 624-632.

SC

353

AC C

EP

TE D

M AN U

354

12

ACCEPTED MANUSCRIPT 355 356

Fig. 1. Changes of plaque in BLA, LEnt and Cg of APP/PS1 transgenic mice.

357 358 359 360 361 362 363

(a-c) Immunostaining with the NT-11 Aβ antibody shows amyloid plaques in BLA and black arrows indicate Aβ deposition. (d) Comparison of the number of senile plaques in three brain regions from 6- and 9-month-old APP/PS1 transgenic mice, respectively. (e-g) Nissl Staining for cytoarchitectonic information from different groups (CON, AD_6M and AD_9M, respectively). The sections were adjacent to those used in a-c. Scale Bar, 200

RI PT

µm. (h) Comparison of plaque load in three brain regions from 6- and 9-month-old APP/PS1 transgenic mice (n=5 mice, with 5 slices averaged per mouse). BLA, basolateral amygdala; LEnt, lateral entorhinal cortex; Cg, cingulate cortex.

364

AC C

EP

TE D

M AN U

SC

365

13

ACCEPTED MANUSCRIPT 366 Fig. 2. 3D reconstruction of neurons in BLA.

368

(a) Minimum projection of coronal images from an APP/PS1 dataset of Golgi-Cox staining, whose sagittal location

369

is presented on the upper left panel (Modified from Paxinos et al.[17]). Thickness is 50 µm while resolution of the

370

raw figure is 0.32×0.32×1 µm. Scale Bar, 500 µm. (b) An enlarged view of amygdala, which identifies LaDL and

371

BLA, Scale Bar, 200 µm. (c) Fine structure of dendritic arbors in 2-D view, Scale bar, 30 µm. (d-e) Main processes

372

of 3-D reconstructions including the raw data importing, neuronal soma positioning, neuronal tracing and 3D

373

checking at last. Scale bar, 50 µm.

RI PT

367

AC C

EP

TE D

M AN U

SC

374

14

ACCEPTED MANUSCRIPT 375 Fig. 3. Dendrograms of neurons in the BLA.

377

(a) 3D reconstructed neurons from different groups (Control, AD_6M and AD_9M, respectively). (b)

378

Corresponding dendrograms show the length and branching pattern for each dendrite of 3 cells. The apical dendrite

379

is at the bottom of each cell's dendrogram. Some apparent differences between groups can be viewed in 2

380

dimensions.

RI PT

376

381

AC C

EP

TE D

M AN U

SC

382

15

ACCEPTED MANUSCRIPT Fig. 4. Progressive changes in dendritic complexity in amygdala of APP/PS1 mice.

384

(a) Average total number of branches of each group. (b) Average branch number from first to fifth order per neuron

385

of each group. (c) Average total dendritic length of each group. (d) Average dendritic length from first to fifth order

386

per neuron of each group. (e) Average total number of terminal tips of each group. (f) Sholl analysis shows the

387

number of intersections with concentric spheres at different distances from the soma (AD_6M group: n=15;

388

AD_9M group: n=19; CON group: n=18, from 2 mice).

AC C

EP

TE D

M AN U

SC

389

RI PT

383

16

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Highlights (1) Neuron-level, reduction of dendritic complexity in BLA of 9-month-old AD mice.

AC C

EP

TE D

M AN U

SC

RI PT

(2) Branch density decrease in 40-90 µm from neuron soma in 6-month-old AD mice.

ACCEPTED MANUSCRIPT Conflict of Interest

AC C

EP

TE D

M AN U

SC

RI PT

We declare that we have no conflict of interest for this manuscript.