Response of Serum-isolated Extracellular Vesicles to Focused Ultrasound Blood-Brain Barrier Opening

11 The blood -brain barrier (BBB) limits drug delivery to the brain and the movement of neurological 12 biomarkers between the brain and blood. Focused ultrasound-mediated blood-brain barrier opening (FUS-13 BBBO) noninvasively opens the BBB, allowing increased molecular transport to and from the brain 14 parenchyma. Despite being initially developed as a drug delivery method, FUS-BBBO has shown promise 15 both as a neuroimmunotherapeutic modality, and as a way of improving neurological disease diagnosis 16 via amplification of disease biomarker circulation. 17 Recently, the role of extracellular vesicles (EVs) in modulating the neuroimmune system and in improving 18 biomarker detection has sparked research interest . However, despite their potential role in modulating 19 FUS-BBBO-induced neuroimmunotherapy and their ability to improve biomarker specificity after 20 treatment, the EV response to FUS-BBBO had not been extensively characterized prior to this study. 21 In this study, we investigated the effect of FUS-BBBO on EV concentration and content in the serum of 22 mice and Alzheimer's Disease (AD) patients. We observed a 164% increase in murine EV concentration 23 one hour after treatment, as well as an increase in EV RNA associated with FUS -BBBO 24 neuroimmunotherapy. Patient EV concentration also increased one hour after treatment a nd was 25 dependent on the volume of BBB opening three days post-treatment. Furthermore, EV isolation was found 26 to significantly enhance the amplification of AD biomarker detection by FUS-BBBO. 27 Overall, we present the first evidence of altered murine and AD patient EV concentration and content in 28 response to FUS-BBBO, providing evidence of EVs' role within FUS -BBBO neuroimmunotherapy as 29 well as their utility in improving FUS-BBBO biomarker amplification. 30

31 Focused Ultrasound, Blood-Brain Barrier, Extracellular Vesicles, Immunotherapy, Biomarkers 32 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 2 MAIN TEXT 33

34 The blood-brain barrier (BBB) is a virtually impermeable barrier between the blood and the brain that 35 keeps the brain at homeostasis for neuronal firing. In addition to limiting the infiltration of neurotoxins 36 and pathogens, the BBB limits both the delivery of drugs to the brain and the circulation of neurological 37 disease biomarkers in the blood [1]. Focused ultrasound-mediated blood-brain barrier opening (FUS -38 BBBO) combines focused ultrasound with intravenously administered microbubbles to transiently and 39 noninvasively open the BBB, tackling these two challenges [2], [3], [4]. 40 Although initially designed to facilitate drug delivery through the BBB, FUS -BBBO has also been 41 established as a neuroimmunotherapeutic treatment for neurological diseases and a method of amplifying 42 the detection of neurological biomarkers [4], [5], [6], [7], [8], [9], [10] . As a neuroimmunotherapeutic, 43 FUS-BBBO has been shown to reduce disease pathology and ameliorate disease -associated cognitive 44 deficits in many neurological models ranging from Alzheimer's Disease (AD) to depression [6], [11]. 45 These effects coincide with brain macrophage modulation, increased neurogenesis, and increased synaptic 46 plasticity [5], [12], [13] . FUS -BBBO amplification of neurological biomarkers ha s been primarily 47 investigated in models of brain cancer, where reports have found increases in circulating cell-free DNA 48 (cfDNA), as well as central nervous system (CNS) proteins, including Glial Fibrillary Acidic Protein 49 (GFAP), in response to treatment [4], [7], [8], [9], [10]. 50 Extracellular vesicles (EVs) are lipid vesicles responsible for cell transport and exchange. EVs have 51 highly variable cargo, including proteins, carbohydrates, and/or coding and non -coding RNA (ncRNA). 52 Due to their small size, there is a particular emphasis on ncRNA within EVs such as micro RNA (miRNA) 53 and piwi-interacting RNA (piRNA). Both miRNA and piRNA regulate the mRNA transcription of target 54 protein-encoding genes. There are publicly available databases of miRNA and piRNA target genes, which 55 allow functional annotation of the up- and downregulated ncRNA targets [14]. 56 EVs are reported to modulate the neuroimmune system, including maintenance and repair of the BBB, 57 neurogenesis, and synaptic plasticity [15], [16], [17], [18], [19] . Due to their role in intercellular 58 communication, EV isolation and characterization has emerged as a method of improving the specificity 59 of biomarker detection. Recent metanalyses of biomarkers in AD have found that isolating EVs prior to 60 quantifying protein load provides a more specific diagnosis [20]. 61 There is preliminary evidence of in vivo FUS-BBBO increased neuronal EV concentration and in vitro 62 FUS-BBBO increased neuroprotective EV concentration [21], [22]. Additional work in the periphery has 63 identified a FUS-induced increase in anti-inflammatory EVs after treatment of arthritis [23], [24]. Given 64 this preliminary evidence of FUS -BBBO affecting EV concentration and content, as well as their dual 65 role in modulating the neuroimmune system, and as an emerging biomarker, we aimed to identify the 66 effect of FUS-BBBO on EV concentration and content. 67 In this study, we explore the response of EVs in serum following FUS-BBBO in a mouse model and in a 68 clinical trial with AD patients. In the mouse study, the concentration of EVs in serum, as well as their 69 genomic and proteomic content are analyzed. To assess the role of EVs in modulating the neuroimmune 70 response following FUS -BBBO, we investigated the effect of GW4869, a neutral sphingomyelinase 71 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 3 inhibitor that blocks EV generation [25], on the restoration timeline of the BBB. Finally, to elucidate the 72 potential of EVs in amplifying neurological marker detection , we collected blood samples from AD 73 patients participating in our group’s clinical study on neuronavigation-guided FUS-BBBO and analyzed 74 the protein content of EVs for important AD biomarkers. This study provides the first translational 75 analysis of EV concentration and content after FUS-BBBO, providing evidence of both the EV role within 76 the neuroimmunotherapeutic response and the potential use of isolating EVs to improve FUS -BBBO 77 neurological biomarker accentuation. 78

79 FUS-BBBO in mice 80 All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee 81 at Columbia University. Mice in the FUS-BBBO and microbubble (MB) sham groups were anesthetized 82 with a mixture of oxygen and 1 -2 % isoflurane (SurgiVet, Smiths Medical PM, Inc., WI), placed on a 83 stereotaxic apparatus (David Kopf Instruments, Tujunga, CA) and their heads were immobilized and 84 depilated to reduce acoustic impedance mismatch. Degassed ultrasound gel was applied on the head and 85 a bath with degassed, deionized water was lowered on top of the head. The lambdoid suture was identified, 86 and the transducer was positioned over it as previously described [26]. 87 In mice that were treated with FUS-BBBO, a single-element, spherical-segment concave FUS transducer 88 (center frequency: 1.5 MHz, focal depth: 60mm, radius: 30mm ; Imasonic, France) that was driven by a 89 function generator (Agilent Keysight 33220A, Palo Alto, CA, USA) through a 50 -dB power amplifier 90 (325LA, Electronic Navigation Industries, Rochester, NY , USA) was used to treat the two hippocampi. 91 The center of the transducer held a pulse-echo ultrasound transducer (V320, center frequency: 7.5 MHz, 92 focal depth: 52 mm, diameter 13 mm; Olympus NDT, Waltham, MA) that was used for alignment and 93 passive monitoring of microbubble cavitation. The pulse -echo ultrasound transducer was driven by a 94 pulser-receiver (5077 PR, Olympus, Waltham, MA, USA) which was in turn connected to a digitizer 95 (Gage Applied Technologies, Inc., Lachine, QC, Canada). The transducer setup was attached to a three -96 dimensional positioning system (Velmex Inc., Lachine, QC, Canada). Each hippocampus was sonicated 97 first for 10 seconds to obtain baseline cavitation dose and then again for 2 minutes for the experimental 98 sonication. For all experiments, in-house synthesized, lipid-shelled microbubbles (average concentration: 99 8x10e8/mL, mean diameter: 1.4 μm) were manufactured according to previously published protocols [27], 100 [28]. A bolus of 3 μL of microbubbles was diluted in 100 μL of sterile saline and was injected 101 intravenously between the baseline and experimental sonications. The transducer was not triggered to 102 treat MB sham mice; otherwise, the treatment was identical to that of the FUS -BBBO mice. Mice in the 103 naïve group were not subjected to anesthesia, FUS or microbubble injections. 104 Mouse Serum Collection 105 Mouse blood was collected from the submandibular vein without anesthesia the day before, and 1 hour 106 after FUS-BBBO (Fig. 1A). Mice were held by grasping the skin behind the head firmly. A 16 -gauge 107 needle was inserted into the submandibular vein and then removed. Less than 100 μL of blood was 108 collected in heparin-coated serum-separating tubes (Ram Sciences). After collection, gentle pressure was 109 applied to the site of the puncture in order to stop bleeding. After collection, blood was left at room 110 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 4 temperature for 15-30 minutes to clot and then spun down at 3000 x g for 15 minutes. Serum was aliquoted 111 to a separate tube and stored at -80 oC for future processing. 112 Magnetic Resonance Imaging for Mice 113 Following treatment and blood draws, all FUS-BBBO and MB Sham animals underwent scanning with a 114 9.4 T MRI system (Bruker Medical, Boston, MA). Mice were intraperitoneally injected with 0.2 mL of 115 gadodiamide solution (OmniscanTM, GE Healthcare, Princeton, NJ) exactly 30 minutes prior to scanning. 116 Images were acquired using a T1 -weighted 2D FLASH sequence (TR/TE 230/3.3 ms, flip angle: 70°, 117 number of excitations: 6, field of view: 25.6 mm × 25.6 mm). 118 GW4869 Drug Administration 119 For the EV inhibition study the drug GW4869 (MedChem Express, Monmouth Junction, NJ, USA) , a 120 neutral sphingomyelinase inhibitor, was used to block EV generation. Mice were separated in four groups: 121 naïve, naïve+GW4869, FUS-BBBO, and FUS-BBBO+GW4869. A 2.5 mg/mL stock solution of GW4869 122 in dimethyl sulfoxide (DMSO) was prepared and stored at -20oC until the day of injection. Immediately 123 before administration, the drug was further diluted in sterile saline to a final concentration of 0.25 mg/mL. 124 The drug was administered intraperitoneally at a dose of 2.5 μg/g body weight per mouse. 125 To study the effect of EVs in the restoration of BBB following FUS-BBBO, mice in the FUS-BBBO and 126 FUS-BBBO+GW4869 were monitored for five days after the sonication procedure. The FUS -127 BBBO+GW4869 group received injections of GW4869 before the FUS -BBBO treatment, and then on 128 day 2 and day 4 after the procedure. The progress of BBBO restoration was monitored for both groups by 129 contrast-enhanced T1-weighted MRI 2 hours after FUS-BBBO, and then on days 1, 3, and 5 post FUS -130 BBBO. Blood was collected on the day prior to FUS-BBBO, and then 1 hour and 1 day after FUS-BBBO, 131 and always before the injection of gadolinium solution for MRI contrast. 132 To elucidate the role of EVs in the inflammatory response following FUS -BBBO, mice from all four 133 groups were allowed to survive for one day after the FUS -BBBO groups were sonicated. Mice in the 134 naïve+GW4869 and FUS-BBBO+GW4869 groups received one injection of the drug immediately before 135 the FUS-BBBO procedure took place. Furthermore, BBBO was confirmed for the mice that underwent 136 FUS-BBBO by contrast-enhanced MRI 2 hours after the sonication procedure. Mice from all groups had 137 blood taken on the day prior to FUS -BBBO, 1 hour, and 1 day after FUS -BBBO. After the final blood 138 draw, the mice were sacrificed by transcardial perfusion with cold 1X PBS and their hippocampi were 139 separated for bulk RNA sequencing. 140 FUS Treatment in AD Patients 141 Six Alzheimer’s Disease (AD) patients underwent neuronavigation -guided FUS-BBBO as part of our 142 phase I clinical trial (NCT04118764). All methods used were approved by Columbia University’s 143 Institutional Review Board and all participants provided informed consent prior to the FUS -BBBO 144 procedure. The right frontal lobe was targeted and numerical simulations using the k -wave package in 145 MATLAB were carried out prior to the experiment to estimate the power attenuation of FUS through the 146 skull. A detailed account of the methods of FUS -BBBO in AD patients is given in our group’s clinical 147 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 5 paper [29]. Briefly, a single-element, spherical-segment FUS transducer (H-231, Sonic Concepts, Bothell, 148 WA, USA) was driven at a frequency of 0.25 MHz by a function generator (Agilent, Palo Alto, CA, USA) 149 and a 50-dB power amplifier (Electronic Navigation Industries, Rochester, NY , USA) to emit FUS with 150 pulse length 10 ms and pulse repetition frequency 2 Hz. A derated peak-negative FUS pressure of 200 kPa 151 (mechanical index (MI) = 0.4) was applied for 2 minutes. A bolus of microbubbles (0.1 mL/kg, Definity, 152 Lantheus) was intravenously injected at the start of the sonication. During the sonication, the cavitation 153 dose was monitored by using a single-element transducer for the first four patients and an imaging array 154 transducer (P4-2, ATL Philips) for the remaining two patients. 155 Patient BBBO Volume Quantification 156 BBB opening volume was quantified from the contrast -enhanced T1-weighted MRIs, which were taken 157 2 and 72 hours after the FUS treatment using a 3 T system (Signa Premiere, General Electric, Boston, 158 MA, USA) for the confirmation of the opening and closing of BBB respectively. For T1-weighted contrast 159 enhancement, an intravenous injection of 0.2 mL/kg gadoterate meglumine (Dotarem®, Guerbet). The 160 contrast-enhanced volume was quantified by subtracting the 72-hour contrast-enhanced T1 MRI from the 161 2-hour MRI and thresholding the subtracted image. The threshold was chosen automatically so that the 162 average intensity within the opening volume was considerably higher than that of the surrounding area, 163 with a 98% level of confidence assuming the Gaussian distribution of the intensity of the subtracted image. 164 Extracellular Vesicle Isolation 165 The commercial Exo Quick (Systems Biosciences , Palo Alto, CA, USA ) was used according to 166 manufacturer instructions to isolate extracellular vesicles from both mouse and human serum. Briefly, 25 167 μL of mouse serum were diluted with 75 μL of 1X PBS before incubation with 22.5 μL of ExoQuick for 168 30 minutes at room temperature. For human samples, 200 μL of undiluted serum were incubated with 45 169 μL of ExoQuick for 30 minutes at room temperature. The samples were then spun down at 1500 x g for 170 35 minutes at 4oC, and the pellet (isolated extracellular vesicles) was resuspended in 100 μL or 200 μL of 171 1X PBS for the mouse and human samples respectively. The isolated EV suspensions were stored at -80 172 oC until further processing. 173 Nanoparticle Tracking Analysis 174 Extracellular vesicle concentration analysis was quantified by nanoparticle tracking analysis (NTA) on a 175 NanoSight (NS300, Malvern Panalytical, Malvern, UK). 5 μL of isolated mouse EVs or 1 μL of isolated 176 human EVs were further diluted into 1 mL of 1X PBS. This solution was then run through the Nanosight 177 at a rate of 1000 μL/min, and the resulting image was captured and analyzed for particle concentration 178 and size distribution. 179 Multiplex Protein Quantification 180 For the samples from the clinical study, a Luminex multiplex assay was used to quantify proteins in the 181 serum and isolated clinical ext racellular vesicles (Luminex Corporation, Austin, TX, USA ). Single 182 procartaplex kits were purchased and combined to make a custom multiplex panel for analysis (Invitrogen, 183 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 6 Waltham, MA, USA). Data were fit with a separate five-parameter logistic dose-response curve for each 184 protein, and all curves had R2 greater than or equal to 0.95. 185 Mass Spectrometry Protein Analysis 186 For the samples from the mouse studies, m ass spectrometry proteomics was performed by Systems 187 Biosciences on already isolated EVs. Briefly, Systems Biosciences lysed the isolated EVs in a gel-loading 188 buffer, followed by gel-based extraction and trypsinization for peptide library creation for LC/MS ESI -189 TOF. Peptide signatures were then mapped to a database of known protein sequences. Peptide 190 quantification across all four runs was loaded into R, normalized, and processed for differential protein 191 expression utilizing the UniprotR package. All functional annotation was performed with the TopGO 192 package. 193 RNA Sequencing 194 RNA Sequencing was performed by Systems Biosciences on already isolated EVs and dissected frozen 195 hippocampus tissue. RNA was isolated and quantified using Agilent Bioanalyzer Small RNA Assay before 196 75bp single -end read Next Gen Sequencing libraries were prepared with Qiagen small RNA library 197 preparation and gel purification. Sequencing was performed on Illumina NextSeq with SE75 at an 198 approximate depth of 10-15 million reads per sample. Reads were processed and aligned to the GRCm38 199 genome with Ensemble transcriptome annotation (GRCm38.p6) using CellRanger with default 200 parameters. Count tables were loaded into R and underwent normalization and differential gene 201 expression analysis with the edgeR package. piRNA and miRNA targets were extracted from piRNAdb 202 and miRBase, respectively. All functional annotation was performed with the TopGO package. 203 Western Blotting 204 RayBiotech performed western blotting using an automated capillary immunoassay method. Samples and 205 reagents were loaded onto an assay plate and put into the western blotting machine. The sample was 206 automatically loaded and separated by size while it traveled through the stacking and separation matrix. 207 Then, the separated proteins are fixed with proprietary capture chemistry. Target proteins are identified 208 with primary and secondary HRP-conjugated antibodies. 209

210 Murine extracellular vesicle concentration increases after FUS-BBBO 211 Wild-type mice were separated into three groups – naïve, sham, and FUS -BBBO (Fig. 1A). Animals 212 treated with FUS-BBBO were intravenously injected with microbubbles (MB) and treated with 2 minutes 213 of focused ultrasound bilaterally on the hippocampi, consistent with literature [26], [30]. Animals in the 214 sham group were intravenously injected with MB but were not treated with focused ultrasound. Animals 215 in all groups had blood drawn twice immediately before treatment ( baseline) and 1 hour after treatment 216 (1 hour). Animals in the sham and FUS-BBBO groups underwent contrast -enhanced T1-weighted MRI 217 after the second blood draw to confirm the opening within the FUS-BBBO group and the lack of opening 218 within the sham group. The methods detail the EV isolation and concentration quantification. 219 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 7 Western blots of the isolated EVs from representative sham and FUS -BBBO samples confirmed the 220 successful isolation of the EVs via expression of marker tetraspanin proteins CD9, CD81, and CD63, as 221 well as β-actin (Fig. 1B). Nanoparticle tracking analysis (NTA) reveals that t he EV concentration is 222 significantly increased 1 hour after treatment compared to the baseline (Fig. 1C). Furthermore, the average 223 increase of 164% after FUS-BBBO is significantly higher than the percentage change in naïve and sham 224 samples which averages near 0% (Fig. 1D). 225 FUS-BBBO alters murine extracellular vesicle protein and RNA load 226 Given the significant increase in EV concentration 1 hour after treatment, we performed whole genome 227 RNA-sequencing and mass spectrometry protein identification ( Systems Biosciences, Palo Alto, CA, 228 USA) on isolated EVs from baseline and 1 hour. From each time point, two samples pooled from 3 animals 229 underwent both processes. Differential gene expression analysis between 1 hour and baseline reveals 230 significantly up- and downregulated protein-coding and non-protein-coding (ncRNA) RNA (Fig. 2A-C). 231 Upregulated protein-coding genes include proliferation-associated genes such as Sox3 and inflammation-232 associated genes such as Mapk12. Downregulated protein-coding genes include tight-junction genes such 233 as Cldn11. Differential expression of the protein identification finds many fewer significantly up and 234 downregulated proteins (only 10 proteins compared to 900 genes). The most significantly upregulated 235 proteins include immediate inflammatory response proteins such as Lbp and hemoglobin -associated 236 proteins such as Hbb-bs (Fig. 2D-E). 237 Next, functional annotation was used to identify the biological processes associated with the differentially 238 expressed: a) proteins, b) protein-encoding genes, c) ncRNA, d) piRNA target genes, and e) miRNA target 239 genes. Protein functional annotation maps to immediate and acute inflammatory response , while the 240 protein-coding and non-coding RNA correspond to more long-term responses such as synapse regulation 241 and neurogenesis (Fig. 2F). Many of the functions associated with the RNA changes are coincident with 242 reported FUS-BBBO increases in neurogenesis [5], proliferation [5], and synaptic remodeling [13]. This 243 leads to the hypothesis of EV response involvement in the neuroimmunotherapeutic responses to FUS -244 BBBO. 245 GW4869 eliminates murine EV concentration increase and reduces inflammatory response 246 To further elucidate the role of EVs in modulating FUS-BBBO neuroimmunotherapy, we utilize GW4869, 247 a neutral sphingomyelinase inhibitor that is the most widely used agent for blocking EV generation [25] 248 (Fig. 3A). We studied four groups for this experiment : naïve, naïve+GW4869, FUS-BBBO, and FUS -249 BBBO+GW4869. 250 First, we confirmed that GW4869 successfully eliminated the FUS -BBBO-induced increase in EV 251 concentration. We found that 1 hour after treatment, the animals treated with FUS-BBBO+GW4869 had 252 a statistically lower EV concentration compared to baseline, starkly contrasting with our FUS -BBBO 253 group, which increased in EV concentration by over 100% ( Fig. 3B). Comparing the EV concentration 254 change between our four groups, we see that FUS -BBBO+GW4869 is indistinct from naïve and 255 naïve+GW4869 both 1 hour and 1 day after treatment (Fig. 3C). 256 Next, we monitored BBB restoration after FUS -BBBO with and without GW4869. The BBB opening 257 volume of each animal was quantified on Days 0, 1, 3, and 5. On every day of measurement, the animals 258 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 8 in the FUS-BBBO+GW4869 had smaller openings than those in the FUS -BBBO group (Fig. 3D). This 259 difference is particularly significant on day 1 when the BBBO volumes of the FUS -BBBO and FUS-260 BBBO+GW4869 groups averaged 58% and 42% of the day 0 BBBO volume respectively (Fig. 3E). 261 Due to previous literature identifying 1 day as the peak of FUS -BBBO induced inflammation [5], [31], 262 [32], and given the more restored BBB in our FUS -BBBO+GW4869 group at this time point, we 263 performed bulk RNA sequencing on tissue extracted from 1 day after FUS-BBBO for our four treatment 264 groups. Differential gene expression between FUS -BBBO and FUS -BBBO+GW4869 revealed that 265 GW4869 reduced the presence of inflammatory markers, including IL6 and CCL4 (Fig. 3F). In order to 266 account for any effects of GW4869 treatment alone, we performed functional annotation on the 267 differentially expressed genes between FUS -BBBO compared with naïve and FUS -BBBO+GW4869 268 compared with naïve+GW4869. This revealed increased processes altered in FUS -BBBO without 269 injecting GW4869, including migratory, development, and inflammatory terms. The functions shared by 270 both comparisons are primarily involved with vasculature development (Fig. 3G). 271 Overall, we see that GW4869 eliminates the FUS-BBBO-induced increase in EV concentration, decreases 272 the volume of BBB opening, and reduces the number of differentially affected processes after FUS -273 BBBO, indicating a vital role of EVs within the neuroimmunotherapeutic response to FUS-BBBO. 274 Patient extracellular vesicle concentration peaks 1 hour after FUS-BBBO 275 Six Alzheimer's Disease Patients underwent FUS -BBBO as part of our group's phase I clinical trial 276 (NCT04118764) as reported in our group’s clinical paper [29]. All patients had blood drawn immediately 277 prior to treatment (Baseline) and 3 days after treatment. Additionally, the last four patients — P3, P4, P5, 278 and P6 — had blood drawn 1 hour after treatment ( Fig. 4A). All patients apart from P3 had confirmed 279 blood-brain barrier opening; P3 did not have successful opening and thus is considered a FUS -sham 280 subject. 281 EVs were isolated from each time point for each patient. Western blotting of the isolated EVs from 282 representative Baseline and FUS-BBBO samples confirmed successful EV isolation (Fig. 4B). Comparing 283 EV concentration from Baseline to 1 hour post -treatment reveals a significant increase in EV 284 concentration (Fig. 4C) with a near-return to Baseline by 3 days after treatment ( Fig. 4D). Furthermore, 285 the percent increase in EVs 3 days after treatment is correlated with the volume of the blood-brain barrier 286 opening (Fig. 4E). 287 EV isolation improves FUS-BBBO amplification of neurological biomarkers 288 To elucidate the utility of EVs in improving liquid biopsy specificity, we quantified several potential AD 289 biomarkers, EV proteins, and other CNS proteins within our patient -isolated EVs, isolated EVs 290 normalized by EV concentration (normalized EVs), and total serum. We find that the EV , normalized EV , 291 and total serum protein concentrations remain mostly unchanged 1 hour after treatment and even 292 decreased compared to baseline. Three days after treatment, the concentration of a number of proteins is 293 significantly increased for both EVs and normalized EVs compared to the total serum, which has no 294 changes in marker concentration for any of the markers (Fig. 5A). Furthermore, the differences between 295 1 hour and 3 days log fold change (LogFC) in protein content are more statistically distinct for the EVs 296 and normalized EVs than the total serum (Fig. 5B). 297 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 9 Finally, in our clinical study, we compared the MRI-based blood-brain barrier opening volume to the 298 LogFC of select biomarkers 3 days after treatment [29]. Further analysis of this data indicates that the 299 level of the AD biomarkers is positively correlated with BBBO in serum and EV content but not 300 normalized EV content (Fig. 5B). This indicates that the increase in biomarker detection is due to more 301 EVs being released, and not because each EV has higher biomarker concentration. Overall, our results 302 from the AD clinical trial reveal the potential of FUS-BBBO in enhancing EV-based biomarker detection 303 sensitivity and specificity for liquid biopsy. 304

And Conclusion 305 Focused ultrasound blood -brain barrier opening (FUS -BBBO) has been studied as a 306 neuroimmunotherapeutic and a method of improving liquid biopsy specificity for neurological disorders 307 [4], [6], [11]. Due to the dual role of extracellular vesicles (EVs) in modulating the neuroimmune system 308 [18], [19] and as an emerging biomarker [20], [33], we aimed to identify the effect of FUS-BBBO on EVs 309 isolated from mouse and patient serum. 310 We first identified a significant increase in mouse EV concentration 1 hour after treatment coincident with 311 RNA changes associated with the EV-dependent neuroimmunotherapeutic effects of FUS-BBBO such as 312 neurogenesis, synaptic pruning, and barrier maintenance [5], [6], [16], [34], [35] . Eliminating the EV 313 concentration increase with GW4869 resulted in reduced blood -brain barrier opening volume and 314 inflammation, indicating the contribution of the EVs to FUS -BBBO inflammation and opening volume. 315 The immune response to FUS-BBBO has spearheaded debate about the method's safety, so the ability to 316 control and mitigate FUS-BBBO-induced inflammation provides an exciting new avenue, mainly when 317 FUS-BBBO is used purely as a drug delivery tool [32]. Future work will include investigating the 318 neuroimmunotherapeutic capacity of FUS-BBBO with depleted EVs because those benefits may require 319 a complete neuroimmune response. 320 Secondly, we identified a significant increase in Alzheimer's Disease (AD) patient EV concentration 1 321 hour after treatment that remained dependent on BBBO volume 3 days after treatment. This was 322 accompanied by increased AD biomarker detection specificity in isolated EVs compared to total serum, 323 which was also correlated with the volume of BBBO as shown in our clinical study [29]. These results 324 highlight the potential of EVs to be used as accurate and specific markers for neurological diseases, and 325 the ability of FUS-BBBO to amplify their detection in a non-invasive way. 326 However, in order to be implemented for improving biomarker specificity, there needs to be a method of 327 differentiating between the changes in biomarker concentration due to the treatment itself and those due 328 to the disease. Other groups have addressed this challenge by utilizing binary biomarkers such as cfDNA, 329 which are present in the case of disease or otherwise completely absent [7], [9]. This becomes much more 330 complex with diseases such as AD, where many of the proposed biomarkers are CNS proteins that are 331 always in the CNS albeit in different concentrations [20], [33] . Future research must include a 332 normalization between BBBO and biomarker concentration that can be used to identify a BBBO-induced 333 concentration change compared to a disease-induced concentration change. 334 Overall, this study presents the first preclinical and clinical evidence of FUS -BBBO increasing EV 335 concentration and altering EV content, which has implications in both the mechanism of FUS -BBBO 336 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 10 neuroimmunotherapy and the optimization of FUS-BBBO-induced amplification of neurological disease 337 biomarkers. 338 Abbreviations 339 AD: Alzheimer’s Disease; BBB: Blood -brain barrier; cfDNA: Cell -free DNA; CNS: Central nervous 340 system; EV: Extracellular vesicle; FUS: Focused ultrasound; FUS-BBBO: Focused ultrasound-mediated 341 blood-brain barrier opening; GFAP: Glial fibrillary acidic protein; HRP: Horseradish peroxidase; LC/MS 342 ESI-TOF: Liquid chromatography/mass spectrometry electrospray ionization time-of-flight; LogFC: Log 343 fold change; MB: Microbubble; miRNA: Micro-RNA; MRI: Magnetic resonance imaging; ncRNA: Non-344 coding RNA; NTA: Nanoparticle tracking analysis; PBS: Phosphate -buffered saline; piRNA: Piwi -345 interacting RNA. 346

347 This study was supported in part by the National Institutes of Health under Grants R01AG038961 , 348 R01EB009041, and R56AG038961, and the Focused Ultrasound Foundation. FNT was also supported by 349 the Onassis Foundation under contract number F ZT 072-1/2023-2024. Some figures were created with 350 BioRender.com. The authors wish to thank UEIL members Rebecca Noel, Ph.D., Daniella Jimenez, B.S., 351 Samantha Gorman, B.S., and Nancy Kwon M.S. for their support and insightful scientific discussions. 352 Contributions 353 ARKS, AJB, and EEK designed the study and the methodology. ARKS, AJB, and MRD conducted in vivo 354 mouse studies. KL and SB collected and processed blood samples from the clinical study. ARKS, FNT, 355 and MRD processed EV results and made figures. ARKS, FNT, and EEK drafted and revised the 356 manuscript. EEK acquired funding and provided resources for the study. All authors contributed to 357 discussions and reviews of the manuscript. 358 Competing interests 359 Some of the work in this study is supported by patents optioned to Delsona Therapeutics, Inc. where EEK 360 serves as a co-founder and scientific adviser. The remaining authors declare no competing interests. 361

362 [1] N. J. Abbott, A. A. K. Patabendige, D. E. M. Dolman, S. R. Yusof, and D. J. Begley, “Structure 363 and function of the blood-brain barrier,” Jan. 2010. doi: 10.1016/j.nbd.2009.07.030. 364 [2] E. E. Konofagou, Y .-S. Tung, J. Choi, T. Deffieux, B. Baseri, and F. Vlachos, “Ultrasound-Induced 365 Blood-Brain Barrier Opening,” Curr Pharm Biotechnol, 2012. 366 [3] J. Stockwell, N. Abdi, X. Lu, O. Maheshwari, and C. Taghibiglou, “Novel Central Nervous System 367 Drug Delivery Systems,” Chem Biol Drug Des , vol. 83, no. 5, pp. 507 –520, May 2014, doi: 368 10.1111/cbdd.12268. 369 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 11 [4] L. Zhu et al., “Focused Ultrasound-enabled Brain Tumor Liquid Biopsy,” Sci Rep, vol. 8, no. 1, 370 Dec. 2018, doi: 10.1038/s41598-018-24516-7. 371 [5] A. R. Kline -Schoder et al., “Characterization of the responses of brain macrophages to focused 372 ultrasound-mediated blood–brain barrier opening,” Nat Biomed Eng, 2023, doi: 10.1038/s41551-373 023-01107-0. 374 [6] M. E. Karakatsani et al. , “Unilateral focused ultrasound -induced blood -brain barrier opening 375 reduces phosphorylated Tau from the rTg4510 mouse model,” Theranostics, vol. 9, no. 18, pp. 376 5396–5411, 2019, doi: 10.7150/thno.28717. 377 [7] C. P. Pacia et al., “Sonobiopsy for minimally invasive, spatiotemporally-controlled, and sensitive 378 detection of glioblastoma-derived circulating tumor DNA,” Theranostics, vol. 27, no. 1, pp. 362–379 378, 2022, doi: 10.7150/THNO.65597. 380 [8] C. P. Pacia et al., “Feasibility and safety of focused ultrasound-enabled liquid biopsy in the brain 381 of a porcine model,” Sci Rep, vol. 10, no. 1, Dec. 2020, doi: 10.1038/s41598-020-64440-3. 382 [9] L. Zhu, A. Nazeri, C. P . Pacia, Y . Y ue, and H. Chen, “Focused ultrasound for safe and effective 383 release of brain tumor biomarkers into the peripheral circulation,” PLoS One, vol. 15, no. 6, Jun. 384 2020, doi: 10.1371/journal.pone.0234182. 385 [10] J. Rincon -Torroella, H. Khela, A. Bettegowda, and C. Bettegowda, “Biomarkers and focused 386 ultrasound: the future of liquid biopsy for brain tumor patients,” Jan. 01, 2022, Springer. doi: 387 10.1007/s11060-021-03837-0. 388 [11] S. J. Mooney, J. N. Nobrega, A. J. Levitt, and K. Hynynen, “Antidepressant effects of focused 389 ultrasound induced blood-brain-barrier opening,” Behavioural Brain Research, vol. 342, pp. 57–390 61, Apr. 2018, doi: 10.1016/j.bbr.2018.01.004. 391 [12] S. J. Mooney, K. Shah, S. Yeung, A. Burgess, I. Aubert, and K. Hynynen, “Focused ultrasound -392 induced neurogenesis requires an increase in blood-brain barrier permeability,” PLoS One, vol. 11, 393 no. 7, Jul. 2016, doi: 10.1371/journal.pone.0159892. 394 [13] R. L. Noel, S. L. Gorman, A. J. Batts, and E. E. Konofagou, “Getting ahead of Alzheimer’s disease: 395 early intervention with focused ultrasound,” Front Neurosci , vol. 17, 2023, doi: 396 10.3389/fnins.2023.1229683. 397 [14] P. Zhang, W. Wu, Q. Chen, and M. Chen, “Non-Coding RNAs and their Integrated Networks,” Jul. 398 13, 2019, NLM (Medline). doi: 10.1515/jib-2019-0027. 399 [15] J. Meldolesi, “Extracellular vesicles (exosomes and ectosomes) play key roles in the pathology of 400 brain diseases,” Dec. 01, 2021, Springer. doi: 10.1186/s43556-021-00040-5. 401 [16] A. M. Zagrean, D. M. Hermann, I. Opris, L. Zagrean, and A. Popa -Wagner, “Multicellular 402 crosstalk between exosomes and the neurovascular unit after cerebral ischemia. therapeutic 403 implications,” Nov. 06, 2018, Frontiers Media S.A. doi: 10.3389/fnins.2018.00811. 404 [17] L. Zhang et al. , “Neural Progenitor Cell -Derived Extracellular Vesicles Enhance Blood -Brain 405 Barrier Integrity by NF-κB (Nuclear Factor-κB)-Dependent Regulation of ABCB1 (ATP-Binding 406 Cassette Transporter B1) in Stroke Mice,” Arterioscler Thromb Vasc Biol, vol. 41, no. 3, pp. 1127–407 1145, Mar. 2021, doi: 10.1161/ATVBAHA.120.315031. 408 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 12 [18] P. Simeone et al., “Extracellular vesicles as signaling mediators and disease biomarkers across 409 biological barriers,” Apr. 01, 2020, MDPI AG. doi: 10.3390/ijms21072514. 410 [19] G. Marostica, S. Gelibter, M. Gironi, A. Nigro, and R. Furlan, “Extracellular Vesicles in 411 Neuroinflammation,” Jan. 21, 2021, Frontiers Media S.A. doi: 10.3389/fcell.2020.623039. 412 [20] T. Soares Martins et al. , “Diagnostic and therapeutic potential of exosomes in Alzheimer’s 413 disease,” Jan. 01, 2021, Blackwell Publishing Ltd. doi: 10.1111/jnc.15112. 414 [21] Y . Meng et al., “MR-guided focused ultrasound liquid biopsy enriches circulating biomarkers in 415 patients with brain tumors,” Neuro Oncol , vol. 23, no. 10, pp. 1789 –1797, Oct. 2021, doi: 416 10.1093/neuonc/noab057. 417 [22] Z. Deng et al. , “Ultrasound -mediated augmented exosome release from astrocytes alleviates 418 amyloid-β-induced neurotoxicity,” Theranostics, vol. 11, no. 9, pp. 4351 –4362, Feb. 2021, doi: 419 10.7150/THNO.52436. 420 [23] Y . Tang et al. , “Ultrasound -augmented anti -inflammatory exosomes for targeted therapy in 421 rheumatoid arthritis,” J Mater Chem B , vol. 10, no. 38, pp. 7862 –7874, 2022, doi: 422 10.1039/D2TB01219G. 423 [24] P. Xia et al., “Low-Intensity Pulsed Ultrasound Enhances the Efficacy of Bone Marrow–Derived 424 MSCs in Osteoarthritis Cartilage Repair by Regulating Autophagy-Mediated Exosome Release,” 425 Cartilage, vol. 13, no. 2, Apr. 2022, doi: 10.1177/19476035221093060. 426 [25] M. Catalano and L. O’Driscoll, “Inhibiting extracellular vesicles formation and release: a review 427 of EV inhibitors,” Jan. 01, 2020, Taylor and Francis Ltd. doi: 10.1080/20013078.2019.1703244. 428 [26] J. J. Choi, M. Pernot, S. A. Small, and E. E. Konofagou, “Noninvasive, transcranial and localized 429 opening of the blood-brain barrier using focused ultrasound in mice,” Ultrasound Med Biol, vol. 430 33, no. 1, pp. 95–104, Jan. 2007, doi: 10.1016/j.ultrasmedbio.2006.07.018. 431 [27] J. A. Feshitan, C. C. Chen, J. J. Kwan, and M. A. Borden, “Microbubble size isolation by 432 differential centrifugation,” J Colloid Interface Sci, vol. 329, no. 2, pp. 316 –324, Jan. 2009, doi: 433 10.1016/j.jcis.2008.09.066. 434 [28] S. Wang, G. Samiotaki, O. Olumolade, J. A. Feshitan, and E. E. Konofagou, “Microbubble type 435 and distribution dependence of focused ultrasound -induced blood -brain barrier opening,” 436 Ultrasound Med Biol, vol. 40, no. 1, pp. 130–137, 2014, doi: 10.1016/j.ultrasmedbio.2013.09.015. 437 [29] S. Bae et al., “Transcranial blood-brain barrier opening in Alzheimer’s disease patients using a 438 portable focused ultrasound system with real-time 2-D cavitation mapping,” Theranostics, vol. 14, 439 no. 11, pp. 4519–4535, 2024, doi: 10.7150/thno.94206. 440 [30] T. Sun, G. Samiotaki, S. Wang, C. Acosta, C. C. Chen, and E. E. Konofagou, “Acoustic cavitation-441 based monitoring of the reversibility and permeability of ultrasound -induced blood-brain barrier 442 opening,” Phys Med Biol , vol. 60, no. 23, pp. 9079 –9094, Nov. 2015, doi: 10.1088/0031 -443 9155/60/23/9079. 444 [31] R. Ji, M. E. Karakatsani, M. Burgess, M. Smith, M. F. Murillo, and E. E. Konofagou, “Cavitation-445 modulated inflammatory response following focused ultrasound blood -brain barrier opening,” 446 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 13 Journal of Controlled Release , vol. 337, pp. 458 –471, Sep. 2021, doi: 447 10.1016/j.jconrel.2021.07.042. 448 [32] Z. I. Kovacs et al. , “Disrupting the blood -brain barrier by focused ultrasound induces sterile 449 inflammation,” Proc Natl Acad Sci U S A , vol. 114, no. 1, pp. E75 –E84, Jan. 2017, doi: 450 10.1073/pnas.1614777114. 451 [33] S. Lee, S. Mankhong, and J. H. Kang, “Extracellular vesicle as a source of alzheimer’s biomarkers: 452 Opportunities and challenges,” Apr. 01, 2019, MDPI AG. doi: 10.3390/ijms20071728. 453 [34] N. A. Khan, M. Asim, A. El-Menyar, K. H. Biswas, S. Rizoli, and H. Al-Thani, “The evolving role 454 of extracellular vesicles (exosomes) as biomarkers in traumatic brain injury: Clinical perspectives 455 and therapeutic implications,” Front Aging Neurosci , vol. 14, Oct. 2022, doi: 456 10.3389/fnagi.2022.933434. 457 [35] K. Laulagnier, C. Javalet, F. J. Hemming, and R. Sadoul, “Purification and Analysis of Exosomes 458 Released by Mature Cortical Neurons Following Synaptic Activation,” in Exosomes and 459 Microvesicles, 2017, pp. 129–138. doi: 10.1007/978-1-4939-6728-5_9. 460 461 462 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 14 Figures 463 Figure 1. Extracellular vesicle concentration increases after FUS-BBBO in the serum of mice. A) Murine 464 experimental timeline. Blood draws were taken for EV isolation at two time points: prior to treatment 465 (baseline) and 1 hour after treatment (1 hour). Animals were randomly split into three treatment groups: 466 FUS-BBBO, sham, and naïve. Mice in the FUS -BBBO group had 2 minutes of FUS applied bilaterally 467 following MB injection at time 0. Animals in the sham group were injected with MB, but no FUS was 468 applied. Mice from the FUS -BBBO and sham groups were injected with gadolinium (Gd), and a T1 -469 Weighted MRI was performed less than 2 hours after the FUS -BBBO and sham procedures. Animals in 470 the naïve group were anesthesia-, MB-, and FUS- naïve. B) Western blots of β-actin and EV tetraspanin 471 markers CD9, CD63, and CD81 on representative sham and FUS-BBBO EVs from 1 hour post-treatment 472 show successful EV isolation. C) EV concentration at baseline and 1 hour post-treatment in mouse serum. 473 Paired t-test was performed and displayed on graph. D) Percent change in EV concentration at the 1-hour 474 time point for all 3 groups; the FUS-BBBO group shows a significant change in EV concentration after 1 475 hour compared to the naïve and sham groups. One-way ANOV A followed by Bonferroni-corrected t-test 476 between groups was performed. All statistics are displayed on the graph. 477 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 15 Figure 2. FUS-BBBO alters murine EV RNA and protein load. A) V olcano plot of differentially expressed 478 RNA between 1 hour post-treatment and baseline EVs with key protein-coding genes highlighted. B) Bar 479 chart of the significantly (p<0.05) up and down-regulated RNA split by type. C) Plot of most significantly 480 up and downregulated protein-coding RNA with LogFC of protein expression on the horizontal axis. D) 481 V olcano plot of differentially expressed proteins (from mass spectrometry proteomics) between 1 hour 482 post-treatment and baseline EVs with key proteins highlighted. E) Plot of most significantly up - and 483 downregulated proteins with LogFC of protein expression on the horizontal axis. F) Functional annotation 484 of the 1) significantly upregulated proteins, 2) significantly upregulated protein -coding genes, 3) 485 significantly upregulated non-coding RNA (ncRNA), 4) genetic targets of the significantly upregulated 486 piRNA, and 5) genetic targets of the significantly upregulated micro -RNA (miRNA). Adjusted 487 Kolmogorov-Smirnov p-value magnitude is displayed in color. The size of each dot corresponds to the 488 percentage of annotated genes from that term that are significantly upregulated. 489 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 16 Figure 3. GW4869 eliminates murine EV concentration increase and reduces inflammatory response. A) 490 GW4869 study timeline. In the BBB restoration study, sequential MRIs were taken after FUS -BBBO to 491 monitor BBB restoration after treatment with and without GW4869. In the inflammation study, animals 492 were sacrificed 1 day after treatment and the inflammatory response and bulk transcriptome was compared 493 between FUS-BBBO, FUS-BBBO+GW4869, naïve and naïve+GW4869 animals. B) EV concentration at 494 Baseline and 1 hour after treatment with either FUS -BBBO or FUS-BBBO+GW4869. Paired t-test was 495 performed for each group. C) Percent change of EV concentration 1 hour and 1 day after treatment for all 496 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 17 groups. ANOV A followed by Bonferroni post hoc t-tests were performed for each time point. D) BBBO 497 volume for the day of treatment and 1 day, 3 days, and 5 days following treatment for both FUS -BBBO 498 and FUS-BBBO+GW4869. E) Proportion of original BBBO volume that is still open 1 day after treatment 499 for animals in the restoration study. An unpaired t-test was performed between the two groups. F) V olcano 500 plot comparing the profile of FUS -BBBO+GW4869 and FUS -BBBO hippocampi. Significantly 501 expressed (p<0.05) genes are colored. G) Significantly upregulated terms from differential gene 502 expression analysis between FUS-BBBO+GW4869 and naïve+GW4869 and FUS-BBBO compared with 503 naïve. Functional ontology terms are clustered by similarity, and the point size shows their significance. 504 Terms appearing in both functional annotations are shown in the "Both" panel with the average 505 significance between the two comparisons. 506 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 18 Figure 4. Extracellular vesicle concentration increases after FUS -BBBO in Alzheimer's Patients. A) 507 Schematic of clinical trial with neuronavigation-guided FUS-BBBO, including each patient's blood draw 508 and opening volume. B) Western blots of β-actin and tetraspanins CD63, CD81, and CD9 from isolated 509 EVs at baseline and 1 hour after FUS -BBBO confirm successful EV isolation. C) EV concentration at 510 Baseline and 1 hour after treatment for the patients who had successful treatment sessions. D) Percent 511 change in EV concentration 1 hour and 3 days after treatment compared to baseline. An unpaired two-tail 512 t-test was performed between the two groups. E) Correlation of BBBO volume and the percent change in 513 EV concentration 3 days after treatment. Simple linear regression was performed, and the resulting R2 514 and p-value are on the chart. 515 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint 19 Figure 5. EV isolation further improves FUS -BBBO amplification of neurological biomarkers. A) 516 Comparison of the FUS-BBBO induced increase in neurological biomarker detection in total serum (blue) 517 and isolated EVs (purple) with a paired t-test performed for each biomarker. B) Correlation between FUS-518 induced volume of opening and the LogFC change of biomarkers in total serum, isolated EVs and isolated 519 EVs normalized to the EV concentration (Normalized EV). Simple linear regression was performed for 520 each condition, and R2 and p values are displayed on the heatmap. 521 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted December 20, 2024. ; https://doi.org/10.1101/2024.12.17.629012doi: bioRxiv preprint