Ultrasound Mediated Microbubbles Destruction Assists Dual 1 Delivery of Beta-amyloid Antibody and NSCs to Restore Neural 2 Function in Transgenic Mice of Alzheimer’s Disease

23 Background : To explore the feasibility, efficacy and safety of ultrasound mediated 24 microbubbles destruction （ UMMD ） assisted dual delivery of β-amyloid antibody loaded 25 by microbubbles (MB Aβ ) and neural stem cells (NSCs) on Alzheimer’s disease （ AD ） . 26 Methods : 27 APP/PS1 double transgenic mice and 33 wild-type mice were used. The dual 27 delivery of β-amyloid antibody and NSCs group (US+MB Aβ +NSCs), single delivery of β- 28 amyloid antibody group (US+MB Aβ ), US+MB group, Control group and Wild group, were 29 involved in the experiment. MB Aβ or MB were injected via the tail vein, followed by NSCs or 30 saline administration and exposed to ultrasound once a week for four times. The survival of 31 NSCs was detected with the in vivo imaging method. Mice in each group were used for 32 behavioral function evaluation and the pathology tests. Brain samples were used to detect β- 33 amyloid deposition, BDNF and synaptophysin expression. 34 Results : BBB was opened by UMMD with an opening time about 10 h. The transplanted NSCs 35 survived in AD brain for no more than 72 h. The learning and spatial memory function was 36 significantly improved in the US+MB Aβ +NSCs group, US+MB Aβ group came second. 37 Immunochemistry results showed amyloid plaques reduction in the US+MB Aβ +NSCs group at 38 the cortex and hippocampus. Higher level of BDNF was demonstrated in the US+MB Aβ +NSCs 39 group than the US+MB Aβ group and the Control group with Western Blot and 40 immunofluorescence examination, but synaptophysin remained no significant changes. 41 Conclusions : UMMD assisted combined delivery of β-amyloid antibody and NSCs to AD mice 42 brain can help to clear the Aβ peptide, increase BDNF level and restore the impaired neural 43 function, which was superior to β-amyloid antibody delivery group. Therefore, the combined targeted delivery assisted by UMMD strategy may be a promising and safe method on treating

8 floating neurospheres. The medium was half replaced every other day. In addition, neurospheres 155 were dissociated to get single cell using StemPro Accutase cell dissociation reagent (Gibco, NY, 156 USA) for passaging or transplantation. NSCs at the third to fifth passage were used in this study. 157

In vivo imaging of the luciferase labeled NSCs 158
To evaluate whether the NSCs could pass the pulmonary circulation and home to the brain 159 or not, a whole-animal imaging was performed to detect bioluminescent signals of NSCs using 160 the in vivo imaging Spectrum system (IVIS, Xenogen, Alameda, CA, USA). Images were Science & Technology Co., Ltd., Beijing, China) to observe the distribution and survival of 172 fLuc-NSCs in mice brain every 24 h until no bioluminescent signals were detected. 173

Treatment protocol 174
The treatment protocol was described in Table 1. Mice were randomly divided into five 175 groups and the treatment protocol of each group was as follows: US+MBAβ+NSCs group 9 (UMMD assisted dual delivery group): during US exposure, the V-flash mode was switched on 177 and the treatment center was focused at the hippocampus, 300 μl diluted MBAβ(2-9×10 7  The treatment schematic diagram was shown in Figure 1. The treatment was performed in 185 each group once a week for four times. After the treatment, the mice were used for the 186 behavioral test and the histological examinations after heart perfusion. Half of the brains were 187 fixed with 4% paraformaldehyde, and the other hemispheres were used for western blot. The 188 diagram and time sequences of the treatment procedure were illuminated in Figure 1. 189

Neural function evaluation with Morris Water Maze Test 190
The learning and memory function were assessed in a Morris water maze (MWM,Beijing 191 Liuyi Co. Ltd., China) task, including the place-navigation test and spatial probe test. MWM 192 was a white-colored circular water pool in 120 cm-diameter, 40 cm height and maintained at 193 25±2 ℃. The place-navigation was assessed for 5 consecutive days. A platform submerged 2 194 cm below the water surface was placed in the fourth quadrant. Mice were placed into the tank 195 at the same point of each quadrant. If a mouse found the platform within 90 s and stayed on it 196 for more than 10 s, it was considered successful. Otherwise, the mice were manually guided to 197 the platform and remained on it for 10 s. The swimming paths and the time reaching the 10 platform were recorded by a computerized video imaging analysis system. The interval between 199 each trial was 60 s. 200 After the final place-navigation test on the 5 th day, the spatial probe test was carried on with 201 the platform removed. The mice were placed at any one of the four quadrants. The numbers of 202 crossings in the area where the platform originally located and the time spending in the target 203 quadrant were recorded for 60 s. 204

Identification of NSCs with immunochemistry staining 205
Immunocytochemistry was performed on the neurospheres at the 3 rd passage to identify 206 NSCs. The neurospheres were cultured in the confocal dishes for 5 days and fixed with 4% 207 paraformaldehyde for 5 min and treated with 0.1% TritonX-100 for 15 min. After washing with 208 PBS, the neurospheres were sequentially blocked in 10% goat serum (Abcam, Cambridge, UK) 209 for 10min followed by the incubation with the Anti-Nestin Antibody (Mouse monoclonal, 1:200, 210 Millipore, MA, USA) overnight at 4 ℃. The next day, they were sequentially washed in PBS 211 and incubated with Fluorescence-Conjugated affinity pure goat anti-mouse IgG (1:1000, 212 ZSGB-Bio, Beijing, China) at 37 ℃ for 60 min, followed by a washing and incubating in 213 Hoechst33342 (Sigma-Aldrich) for 5 min. Images were acquired by a laser scanning confocal 214 microscope (TCS SP5, Leica, Germany). Immunohistochemistry was performed by using specific antibody to identify senile plaques, 216 brain-derived neurotrophic factor (BDNF) and synaptophysin (SYN) of the mice brain. After 217 the fourth treatment, mice in each group were anesthetized and transcardially perfused with 30 218 ml NS. Then, one hemisphere of a mouse brain was extracted and fixed with 4% 219 paraformaldehyde for 48 h and cut into 4μm-thick paraffin slide sagittally. After dewaxing, 220 11 hydration and antigen retrieval, the sections were sequentially treated with 0.25% TritonX-100 221 for 30 min and blocked with normal donkey serum (Abcam) for 30 min, followed by the 222 incubation with specific antibody of anti-β-amyloid, 1-42 (1:100, BioLegend), BDNF (rabbit 223 polyclonal, 1:100, Proteintech) and SYN (mouse monoclonal, 1:100, Millipore). Biotinylated 224 antibody (SP9000, ZSBG-Bio, Beijing, China) or Alexa Fluor 647-labeled Goat Anti-Rabbit 225 IgG (1:200, Beyotime) was used as the secondary antibody. Five slices of each sample were 226 observed, and the images were captured with an Olympus BX51 microscope or a Leica laser 227 scanning confocal microscope. 228

BDNF and SYN expression assessed with Western blotting 229
Brain samples at the targeted areas were homogenized in mortar with liquid nitrogen. The 230 proteins were extracted in the protease and phosphatase inhibitors (Beyotime) on the ice for 30 231 min. Bicinchoninic acid protein determination assay (BCA protein assay reagent; Beyotime) 232 was used to analyze protein concentration. Samples (10 μl) were electrophoresed in 10% 233 sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to nitrocellulose membranes. 234 Then, the membranes were sequentially blocked in 5% nonfat dry-milk solution for 1 h, 235 incubated with the primary antibodies of anti-BDNF (1:1000), anti-SYN (1:1000) and Tubulin 236 (1:1000, Beyotime) overnight at 4 ℃. After washing the membranes with Tween-Tris-buffered 237 saline buffer, the membranes were incubated in alkaline phosphatase-conjugated secondary 238 antibody (1:5000, ZSGB-Bio) for 1 h. After the incubation, the membranes were washed three 239 times and treated with enhanced chemiluminescence method (Millipore). Images were captured 240 with a LAS-4000 mini (GE Healthcare, Wausaukee, WI). The bands were analyzed with the 241 signal intensity ratio between the targeted protein and Tubulin using Image J. 242

Statistical analysis 243
All the data were presented as mean values ± standard deviation. To account for the multiple 244 mean comparisons, one-way analysis of variance (ANOVA) with least significant difference 245 (LSD) post hoc testing or nonparametric tests with Kruskal-Wallis were used. All the statistics 246 were analyzed by SPSS 20.0 software. P<0.05 was considered as statistically significant. 247

Characterization of MBAβ 249
The MBAβ were in milky and stable at room temperature, and mainly ranging from 300 to 250 700 nm in diameter (Fig. 2D). MBAβ can achieve the same imaging intensity as the MB with 251 contrast-enhanced US. There were both red and green fluorescence on the membrane of MB 252 with the fluorescence imaging, indicating FITC-labeled anti-Aβ-amyloid has been successfully 253 loaded onto the DiI-labeled MB ( Fig. 2A-C). 254

Characterization of NSCs 255
NSCs isolated from E12-14 C57 embryos were in free-floating neurospheres and capable of 256 proliferation, self-renewal and differentiation. A remarkable expression of Nestin was revealed 257 in the neurospheres with the immunofluorescence analysis (Fig. 2E, F). 258

In vivo tracking of the survival of luciferase labeled NSCs 259
In order to tracking the viability of fLuc-NSCs, in vivo imaging of mice was captured every 260 day until no visible bioluminescence was recorded after transplantation. The results revealed 261 the bioluminescence intensity at 48 h (Fig. 2H) was higher than that at 24 h (Fig. 2G), while 262 none at 72 h (Fig. 2I). No bioluminescence was found in the other organs at any time. And the 263 results indicated that fLuc-NSCs infusion via tail vein can go through the pulmonary circulation 13 and home to the brain, and fLuc-NSCs were able to proliferate for a limited time and survive 265 for 48 h in AD mice brain. 266

Effect of UMMD on BBB opening 267
A patchy distribution of EB extravasation was observed at the US-exposed area in the gross 268 brain samples, while no visible EB at the unexposed area ( Fig. 3A-B). Moreover, EB 269 extravasation was visible from the activation point to the 10 th h following UMMD. EB

Improved cognition performance with Morris water maze 278
In the place-navigation test, mice in the Wild group and the US+MBAβ+NSCs group 279 displayed a similar latency. On day 3, 4 and 5, both the time spend of the Wild group and the 280 US+MBAβ+NSCs group were shorter than the other groups and had significant difference with 281 the Control group (P<0.05, Fig. 4F). US+MBAβ group also had significant difference when 282 compared with the control group on day 5 (P <0.05, Fig. 4F). The average time of the 5 days 283 in the US+MBAβ+NSCs group had significant difference with Control and US+MB group but 284 had no difference with the US+MBAβ group (P <0.05, Fig. 4G). In the spatial probe test, the 285 average time had no significant difference among each groups, but the average distance in the 286 14 target quadrant was much longer in the US+MBAβ+NSCs group than Control group (Fig. 4H, 287 P <0.01), and the difference between the US+MBAβ group and Control group was less 288 significant( Fig. 4H, P <0.05). Significant difference was noted in the crosses number between 289 the US+MBAβ+NSCs group and Control group (Fig. 4I, P<0.05).These data indicated an 290 improvement on the impaired learning and spatial memory of the single Aβ antibody delivery 291 and the combined implantation of Aβ antibody and NSCs group, especially the dual delivery 292 group. 293

Aβ plaque reduction assessed by immunohistochemistry 294
Immunohistochemistry was performed with anti-β-amyloid to investigate whether the 295 treatments prompt the amyloid plaques clearance in the brain of the APP/PS1 (+) mice (Fig.  296 5A-E). A significant difference was found in the number of plaques between the 297 US+MBAβ+NSCs group and Control group both in the cortex and in the hippocampus (P <0.05, 298 Fig. 5F,G). Although no significant difference was observed among the US+MB group, 299 US+MBAβ group and US+MBAβ+NSCs group, the number of plaques in US+MBAβ+NSCs 300 group was fewer than the other two groups. No difference was found out between Wild group 301 and US+MBAβ+NSCs group. These data showed that the NSCs transplantation combined with 302 MBAβ delivered by UMMD can efficiently help eliminate amyloid plaques in AD mice. 303

Elevated BDNF but not SYN expression with western blot and immunofluorescence (IF) 304 method 305
To determine whether the combined delivery approach improves the expression of BDNF 306 and SYN, IF and western blot were performed. IF result indicating an elevated BDNF 307 expression in the US+MBAβ+NSCs group, second to that of Wild group both at the cortex and 308 15 hippocampus (Fig. 6 A-E). In contrast, much lower level of BDNF was detected in Control 309 group. The expression of BDNF was improved in the US+MB group and the US+MBAβ group 310 when compared to Control group, but the degree was inferior to that in the US+MBAβ+NSCs 311 group. Western blot analyses of BDNF were shown in Figure 6K. Much higher level of BDNF 312 was demonstrated in the US+MBAβ+NSCs group than the US+MBAβ group and Control group 313 (P <0.05). Disappointingly, no obviously elevated expression of SYN was observed by IF (Fig.  314 6F-J) and western blot (Fig. 6L) in neither group. Thus, the treatment tried in the current study 315 did not alter the level of SYN. 316 Taken together, these data confirmed that the combination of NSCs transplantation and MBAβ 317 mediated by US has recovered the expression of BDNF but not SYN in AD mice in the present 318

study. 319
Discussion 320 AD affects millions of people; however, there is still no effective treatment. Therapeutic 321 agent delivery to the brain is a potential approach for AD treatment. The great challenge of 322 BBB blockage on these therapeutic agents delivered into brain has been resolved by the 323 application of UMMD [18,19]. Stem cells such as NSCs or mesenchymal stem cells has been 324 successfully and targeted transplanted to the AD animals, and consequently the neural function 325 was improved and recovered [20,21]. Approaches involving the Aβ plaque clearance by Aβ 326 antibody delivery in AD mice have also been reported [4][5][6][7]. But the failure of many clinical 327 trials suggest inefficacy in the treatment of AD using only one target [22]. Aiming to further 328 prompt the clearance of Aβ plaque and cognition improvement in AD mice, a dual delivery 329 approach including both the NSCs and Aβ antibody under UMMD was tried in the present 330 16

study. 331
BBB opening window is important on the therapeutic targets delivered into brain. BBB 332 permeability was increased safely, transiently and restorablely in mice brain before UMMD 333 assisted delivery treatment in the study. EB extravasation was observed to evaluate the BBB 334 opening and lasted for no more than 12 h in the wild type mice following UMMD. The amount 335 of EB extravasation reached the peak at the first time of US exposure and remained a high level 336 in the first 4 h, then decreased till 10 h and became invisible at 12 h. In order to get the optimal 337 delivery efficiency, MB carried Aβ antibody was injected during US irradiation and NSCs was 338 infused at the first time after finishing US exposure in the study. The optimal time for the brain 339 targeted delivery of therapeutants was no more than 4 h following US exposure. were not the best for the living and survival of NSCs in AD mice brain and should be optimized 350 in the future study. A repeated treatment strategy of once a week for four times has been 351 preliminarily designed and carried out. According to the limited survival time of NSCs within 352 the AD brain, it is highly recommended to increase the transplantation of NSCs and therapeutic 353 frequency to twice a week rather than once a week for 4 times. 354 The combined delivery approach is superior to the Aβ antibody delivery assisted by UMMD 355 on the neural functional, pathological and molecular biological improvement in AD mice. As 356 demonstrated by the place-navigation test of MWM, the learning function was improved with 357 shorter training time in both the dual delivery group and the Aβ antibody delivery group. In 358 spatial probe test, the spatial memory function was greatly recovered with much longer distance 359 in the targeted quadrant in both the dual delivery and Aβ antibody delivery group, and with 360 much more crosses number in the dual delivery group compared to the control group. The 361 neural function recover was greater in the combined delivery group than the Aβ antibody 362 delivery group. All these results proved that the learning and memory function was repaired 363 greatly both in the dual delivery group and the Aβ antibody delivery group, especially in the 364 dual delivery group. Significant decrease on Aβ deposition was demonstrated in the 365 US+MBAβ+NSCs but not in Aβ antibody delivery group when compared to the control group 366 at the cortex and hippocampus, and the Aβ plaques were fewer in the dual delivery group than 367 Aβ antibody delivery group with no significant difference, which showed no superiority on Aβ reported that the cognitive and memory function was greatly improved without remarkable 372 pathological changes of Aβ deposition following NSCs transplantation [14]. From the similarity 373 and difference between their study and ours results, single target of NSCs or Aβ antibody could 374 not greatly prompt the Aβ clearance, however UMMD assisted dual delivery approach played 375 an intensifying effect on the clearance of Aβ plaque in AD mice. BDNF is a neurotrophic 376 molecule of structural and functional integrity of the hippocampus formation, as well as the 377 consolidation of hippocampus-dependent memory [1,8].