Machine learning methods for determining cell morphology and motility

Abstract

This disclosure includes methods of clustering cells of cell barriers, for example, to determine cell barrier integrity. The method is performed by obtaining an image of the plurality of cells of the cell barrier which are then segmented to extract motility and / or morphology feature values from the segmented image data to cluster the cells.

Description

[0001] MACHINE LEARNING METHODS FOR DETERMINING CELL MORPHOLOGY AND MOTILITY RELATED APPLICATIONS

[0002] This application claims priority under 35 U.S.C. § 119(e) to U.S. Provisional Patent Application No. 63 / 746,473, filed January 17, 2025, entitled “MACHINE LEARNING METHODS FOR DETERMINING CELL MORPHOLOGY AND MOTILITY,” the entire contents of which are incorporated by reference herein in their entirety.

[0003] FEDERALLY SPONSORED RESEARCH

[0004] This invention was made with government support under Grant Nos. CA253051, CA283420 and GM 133725 awarded by the National Institutes of Health. The government has certain rights in the invention.

[0005] BACKGROUND

[0006] Cell barriers (e.g., the blood brain barrier) are complicated multi-cellular structures that help protect and divide different portions (e.g., organ systems) of higher level organisms. For example, the skin cell barrier helps protects the organism from chemicals, allergens, and pathogens. The blood brain barrier helps protect the central nervous system from chemicals, pathogens and cells (e.g., cancer cells) that may be in the blood, but also allows important nutrients (e.g., glucose, oxygen, and water) to enter the central nervous system. The intestinal cell barrier helps protect the organism from the intestinal microbiota, but also allows for the uptake of nutrients into the organism from the intestine.

[0007] Some cell barriers comprise cell monolayers (e.g., simple cell barriers). For example, the blood brain barrier can comprise an endothelial monolayer of cells. Some cell barriers comprise cell multi-layers (e.g., stratified cell barriers). For example, the epidermis can comprise a stratified squamous cell epithelium barrier.

[0008] The composition of a cell barrier (e.g., types of cells in the cell barrier and / or the motility and / or morphology of the cells of the cell barrier) may be altered when the cell barrier is perturbed (e.g., perturbed mechanically, chemically, and / or by a disease pathology like Alzheimer’s disease). Alteration of the cell barrier can lead to unwanted intrusions into protected areas of the organism (e.g., metastasis of cancer cells into the brain).

[0009] In vitro models of cell barriers can be used recapitulate in vivo cell barriers for laboratory experiments.

[0010] #14794310v1SUMMARY

[0011] In some aspects, this disclosure describes a method of determining motility and / or morphology of a plurality of cells of a cell barrier, the method comprising: using at least one computer hardware processor to perform: obtaining an image of the plurality of cells of the cell barrier; segmenting the image to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells; extracting motility and / or morphology feature values from the segmented image data for at least some cells in the plurality of cells; and clustering the at least some cells of the plurality of cells, using their respective motility and / or morphology feature values, to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology.

[0012] In some embodiments, obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of an endothelial barrier, an epithelium barrier, or an epithelial barrier.

[0013] In some embodiments, obtaining an image of the plurality of cells of an endothelial barrier comprises obtaining an image of a blood-brain barrier (BBB).

[0014] In some embodiments, obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of an in vitro model of the cell barrier.

[0015] In some embodiments, obtaining an image of the in vitro model of the cell barrier comprises obtaining images of a cell monolayer.

[0016] In some embodiments, the cell monolayer is representative of an epithelial or endothelial barrier.

[0017] In some embodiments, obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of a cell barrier model that is representative of an in vivo cell barrier.

[0018] In some embodiments, obtaining an image of a plurality of cells of a cell barrier comprises capturing the image using confocal imaging, fluorescent imaging, or quantitative phase imaging.

[0019] In some embodiments, obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of fixed cells.

[0020] In some embodiments, obtaining the image of the plurality of cells of the cell barrier comprises obtaining an image of live cells.

[0021] In some embodiments, obtaining an image of the plurality of cells comprises capturing the image.

[0022] #14794310v1In some embodiments, obtaining an image of the plurality of cells of the cell barrier comprises obtaining a time series of images of the plurality of cells of the cell barrier.

[0023] In some embodiments, segmenting the image to obtain the segmented cell data further comprises segmenting images in the time series of images of the plurality of cells to obtain the segmented cell data, the segmented cell data comprising segmented image data for the cells of the plurality of cells.

[0024] In some embodiments, segmenting the image to obtain the segmented cell data further comprises segmenting using a trained neural network.

[0025] In some embodiments, extracting motility and / or morphology feature values from the segmented image data comprises extracting cell motility feature values.

[0026] In some embodiments, extracting the cell motility feature values comprise extracting one or more of cell: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and max distance travelled.

[0027] In some embodiments, extracting the cell motility feature values comprise extracting each of cell: std speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and max distance travelled.

[0028] In some embodiments, extracting motility and / or morphology feature values from the segmented image data comprises extracting cell morphology feature values for the at least some cells of the plurality of cells.

[0029] In some embodiments, extracting the cell morphology feature values comprises extracting one or more of cell: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity.

[0030] In some embodiments, extracting the cell morphology feature values comprises extracting each of cell: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity.

[0031] In some embodiments, extracting the motility and / or morphology feature values further comprises reducing the dimensions of the extracted motility and / or morphology feature values.

[0032] In some embodiments, reducing the dimensions of the extracted motility and / or morphology feature values is performed using uniform manifold approximation and projection (UMAP).

[0033] In some embodiments, clustering the cells of the plurality of cells is performed using an unsupervised clustering technique.

[0034] #14794310v1In some embodiments, clustering the cells of the plurality of cells comprises clustering using Centroid-based Clustering (Partitioning methods), Density-based Clustering (Model-based methods), Connectivity-based Clustering (Hierarchical clustering), Distribution-based Clustering and / or spectral clustering.

[0035] In some embodiments, clustering the cells of the plurality of cells comprises clustering using silhouette score.

[0036] In some embodiments, clustering the cells of the plurality of cells comprises non-hierarchical spectral clustering with a maximized Silhouette score.

[0037] In some embodiments, obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of a plurality of cells of a cell barrier that have been contacted with a drug.

[0038] In some embodiments, the method further comprises determining an effect of the drug on the motility and / or morphology of the cells of the plurality of cells of the cell barrier compared to a control.

[0039] In some embodiments, a class of cell motility and / or morphology comprises an increase and / or decrease in one or more cell motility and / or morphology feature values relative to the other classes of cell motility and / or morphology.

[0040] In some embodiments, this disclosure describes a method of determining changes in cell motility and / or morphology in a cell barrier in different conditions, the method comprising: using a hardware processor to perform: obtaining a first image of a plurality of cells of a cell barrier in a first condition, and a second image of a plurality of cells of a cell barrier in a second condition; determining first motility and / or morphology feature values of the cell barrier in the first condition and second motility and / or morphology feature values of the cell barrier in the second condition, the determining comprising for the first image and the second image, respectively: segmenting the image to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells; extracting motility and / or morphology feature values from the segmented image data for each of the cells of the plurality of cells; and clustering the cells of the plurality of cells to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology; and determining changes in cell motility and / or morphology of the cell barrier in the different conditions, the determining comprising identifying one or more changes in the plurality of clusters in the first condition and the second condition.

[0041] In some embodiments, this disclosure describes a system, comprising: at least one computer hardware processor; and at least one non-transitory computer-readable storage #14794310v1medium that, when executed by the at least one computer hardware processor, causes the at least one computer hardware processor to perform a method described herein.

[0042] In some embodiments, this disclosure describes at least one non-transitory computer-readable storage medium that, when executed by at least one computer hardware processor, causes the at least one computer hardware processor to perform a method described herein.

[0043] BRIEF DESCRIPTION OF DRAWINGS

[0044] The accompanying drawings are not intended to be drawn to scale. For purposes of clarity, not every component may be labeled in every drawing. In the drawings:

[0045] FIGs. 1A-1K show the extracellular vesicles derived from brain-seeking MDA-MB-231 cells (Br-EVs) modulated the expression of multiple Rabll effector proteins (rabllfips) involved in intracytoplasmic vesicle trafficking and recycling. FIG. 1A was created using Biorender®. Schematic representing degradation and recycling pathways in brain endothelial cells (BECs). FIG. IB was created using Biorender®. Schematic depicting the most widely acknowledged mechanisms of interaction between Rabi lfip2, Rabi lfip3, Rabi lfip5 and the respective motor proteins for the tethering and transport of rabl 1 vesicles on cytoskeleton filaments. FIGs. 1C-1E show the Western blot quantification, representative immunofluorescent image and fluorescence intensity quantification of Rabllfip2, (FIGs. 1F-1H) Rabllfip3 and (FIGs. 1J-1K) Rabllfip5 protein expression in BECs following treatment with PBS, P-EVs and Br-EVs (mean ± SD; technical triplicates in 3 independent experiments). Scale bars 100 pm. Statistical analyses were performed using the Kruskal-Wallis test with Dunn’s multiple comparisons test (FIG. 1C, FIG. IF, FIG. II) and the Mann- Whitney test (FIG. IE, FIG. 1H, FIG. IK)45, 46. In all displayed graphs, ns = not significant; *P < 0.0332; **P < 0.0021; ***P < 0.0002; ****p < 0.0001. EHD-1, EH domain-containing protein 1; GAPDH, Glyceraldehyde 3-phosphate dehydrogenase; SD, standard deviation.

[0046] FIGs. 2A-2C show that cerebral microvessels isolated from the brain cortex of mice treated with Br-EVs showed decreased Rabi lfip2 and increased Rabi lfip3 and Rabi lfip5 levels in response to Br-EV treatment. FIG. 2A shows the Western blot quantification of Rabi lfip2, Rabi lfip3 and Rabi lfip5 in cerebral microvessel lysates of mice treated with PBS, P-EVs and Br-EVs. PBS, P-EV and Br-EV groups were run on separate gels under the same conditions and normalization to GAPDH was performed individually for each sample with raw values being presented. Uncropped gels are provided in FIG. 12 (mean ± SD; n = 4 for PBS and Br-EV groups and n = 6 and 5 for the P-EV group). FIG. 2B shows representative fluorescent images of cerebral microvessels from the brains of mice treated with PBS, P-EVs and Br-EVs

[0047] #14794310v1and stained with Rabi lfip2, Rabi lfip3 and Rabi lfip5 (green) and nuclei (blue). Scale bars 50 pm. FIG. 2C shows the quantification of the signal mean fluorescence intensity (MFI) of the cerebral microvessels of mice treated with PBS, P-EVs and Br-EVs and stained for Rabi lfip2, Rabi lfip3 and Rabi lfip5 (normalized to the image background, mean ± SD; n = 5 for the PBS group, n = 6 for the P-EV group and n = 4 for the Br-EV group). Statistical analysis was performed using the Mann-Whitney test (FIG. 2A) and one-way ANOVA with Tukey’s multiple comparison test (FIG. 2C)45, 46, 53, 54. ns = not significant; *P < 0.0332; **P < 0.0021; ****p < 0.0001.

[0048] FIGs. 3A-3N show BEC monolayers treated with Br-EVs exhibited characteristic morphology patterns when analyzed by machine learning (ML). FIGs. 3A-3C show uniform Manifold Approximation and Projection (UMAP)36of the BEC population treated with (FIG.

[0049] 3A) PBS, (FIG. 3B) P-EVs and (FIG. 3C) Br-EVs and analyzed by confocal microscopy. FIG.

[0050] 3D shows confocal microscopy data spectral clustering with Silhouette score38. FIGs. 3E-3F show bar plots and tables reporting cluster distribution variations37and percentages in PBS, P-EV and Br-EV samples. FIG. 3G shows a heatmap showing how the different morphological features distributed among the identified clusters. FIGs. 3H-3J show UMAP representation of the BEC population treated with (FIG. 3H) PBS, (FIG. 31) P-EVs and (FIG. 3J) Br-EVs and analyzed by live cell imaging. FIG. 3K shows live cell imaging data spectral clustering with Silhouette score. FIGs. 3L-3M show bar plots and tables showing cluster distribution variations and percentages in PBS, P-EV and Br-EV samples. FIG. 3N shows a heatmap showing how the different morphological features distributed among the identified clusters. FIG. 3D and FIG. 3K show a cluster 1 in blue, cluster 2 in orange, cluster 3 in green and cluster 4 in red. Statistical analysis was performed using Bootstrap resampling with the z-test method60. *P < 0.05; ****P < 0.0001.

[0051] FIGs. 4A-4N show BEC monolayers treated with Br-EVs exhibit characteristic motility patterns when analyzed by machine learning (ML). FIGs. 4A-4C show UMAP36of the BEC population treated with (FIG. 4A) PBS, (FIG. 4B) P-EVs and (FIG. 4C) Br-EVs and analyzed by QPI. FIG. 4D shows QPI data spectral clustering with Silhouette score38. Cluster 1 in blue, cluster 2 in orange and cluster 3 in green. FIGs. 4E-4F show bar plots and tables reporting clustered distribution37variations and percentages in PBS, P-EV and Br-EV samples. FIG. 4G shows a heatmap showing how the different morphological features distributed among the identified clusters. FIGs. 4H-4J show UMAP representation of the BEC population treated with (FIG. 4H) PBS, (FIG. 41) P-EVs and (FIG. 4J) Br-EVs and analyzed by live cell imaging.

[0052] FIG. 4K shows live cell imaging data spectral clustering with Silhouette score. Cluster 1 in #14794310v1blue, cluster 2 in orange, cluster 3 in green and cluster 4 in red. FIGs. 4L-4M show bar plots and tables showing clustered distribution variations and percentages in PBS, P-EV and Br-EV samples. FIG. 4N shows a heatmap showing how the different morphological features distributed among the identified clusters. Statistical analysis was performed using Bootstrap resampling with z-test method64. *** P < 0.001, ****P < 0.0001.

[0053] FIGs. 5A-5H show BECs knocked down (KD) for Rab7 and overexpressing (OE) Rabi lfip5 displayed motility and morphology patterns similar to those of BECs treated with Br-EVs. UMAP representation of BEC populations KD for Rab7 and Rabi lfip2 and OE for Rabi lfip3 and Rabi lfip5 (from left to right), were analyzed for morphology features using (FIG. 5A) confocal microscopy (to be compared with FIG. 3C) and (FIG. 5C) live cell imaging (to be compared with FIG. 3J), and for motility features using (FIG. 5E) QPI (to be compared with FIG. 4C) and (FIG. 5G) live cell imaging (to be compared with FIG. 4J). FIG. 5B, FIG.

[0054] 5D, FIG. 5F, and FIG. 5H have tables showing variations in clustered distribution and the percentage of each cluster within each respective sample.

[0055] FIGs. 6A-6G show that in vivo, Br-EVs induced significant changes in cerebral microvessel protein expression. FIG. 6A shows a correlation heatmap displaying the correlations between PBS, P-EVs and Br-EVs samples in a color-coded matrix (+1 = blue = positive correlation, -1 = red = negative correlation). FIG. 6B shows a heatmap visualization of whole quantitative proteomics analyses showing the proteins differentially expressed (Log 2-fold change 0.5, q < 0.05) in the cerebral microvessels isolated from brains of mice injected with PBS, P-EVs and Br-EVs. FIG. 6C shows the top 20 significantly upregulated and (FIG. 6D) downregulated proteins in Br-EV samples compared to PBS controls. FIG.6E shows the Ingenuity Pathway Analysis (IPA)98of the top significantly modulated biological functions in Br-EV samples compared to PBS controls, blue represents the biological functions with predicted downregulation (z- score < -1) and red represents the biological functions with predicted upregulation (z- score > 1). FIG. 6F shows a Western blot quantification of NKCC1 protein expression in BECs following treatment with PBS, P-EVs and Br-EVs (normalized to GAPDH, mean ± SD; technical triplicates in 3 independent experiments). FIG. 6G shows a representative NKCC1 immunoblot (mean ± SD; technical triplicates in 3 independent experiments). Comparison analysis (FIGs. 6A-6E) was performed using R and IPA software (n = 5 per group). Statistical analysis was performed using the Kruskal-Wallis test with Dunn’s multiple comparisons test (FIG. 6F)47, 48. *P < 0.032.

[0056] FIGs. 7A-7D show Br-EVs caused increased MDA-MB-231 TNBC cell adhesion to BECs and CD31 expression in mouse brains and isolated cerebral microvessels. FIG. 7A shows #14794310v1fluorescent cell counts (left) and representative images (right) demonstrated that MDA-MB-231 TNBC cells adhere more significantly to BEC monolayers pre-treated with Br-EVs compared to PBS controls (mean ± SD; three independent experiments). FIG. 7B shows representative fluorescence microscopy images of the brain sections of mice treated with Br-EVs (n=5) stained for BECs (CD31, green), mature neurons (MAP2, magenta) and nuclei (DAPI, blue). Scale bars 200 pm (right), 100 pm (middle) and 50 pm (right). FIG. 7C shows representative fluorescence microscopy images of the expression of CD31 (green) and nuclei (blue) in the cerebral micro vessels isolated from the brains of mice treated with phosphate buffer saline (PBS), parental MDA-MB-231 -derived EVs (P-EVs) or Br-EVs (n=5). Scale bars 50 pm. FIG. 7D shows the quantification of CD31 signal fluorescence intensity in cerebral micro vessels treated with PBS, P-EVs or Br-EVs. Statistical analysis was performed using (FIG. 7A) the Mann-Whitney t-test and (FIG. 7C) the one-way ANOVA with Tukey’s multiple comparison test46, 53, ns = not significant; *P < 0.032; **P < 0.0021; ****p < 0.0001.

[0057] FIGs. 8A-8D show extracellular vesicle (EV) characterization according to the latest guidelines from the International Society of Extracellular Vesicles. FIG. 8A shows transmission electron microscopy imaging of extracellular vesicles secreted from parental (P-EVs) and brainseeking (Br-EVs) MDA-MB-231 cells. Scale bars = 200 nm. FIG. 8B shows nanoparticle tracking analysis of the size of P-EVs and Br-EVs. FIG. 8C shows P- and Br-EVs total protein quantification by Bradford assay. FIG. 8D shows representative Western blot images of EV markers cluster of differentiation (CD63), CD73 and Annexin A2, and the endoplasmic reticulum marker, Calnexin.

[0058] FIGs. 9A-9H show whole quantitative proteomics of cerebral microvessel samples. FIGs. 9A-9C show volcano plots showing the comparative analysis results between PBS, P-EVs and Br-EVs samples. Proteins with an adjusted p-value of 0.05 or less and a Log2 fold-change of 0.5 or greater (absolute value) are highlighted in red (Br-EVs vs. PBS), blue (P-EVs vs. PBS) and green (Br-EVs vs. P-EVs) (n = 5 per group). Total detected proteins can be observed in cerebral microvessel samples from mice systemically administered with (FIG. 9D) PBS, (FIG.

[0059] 9E) P-EVs and (FIG. 9F) Br-EVs. FIG. 9G shows an Euler diagram representing the relationships between the sets of proteins identified in the cerebral microvessels isolated from the brains of mice injected with PBS, P-EVs and Br-EVs. FIG. 9H shows principal component analysis.

[0060] FIGs. 10A-10C show Br-EVs do not cause significant modulation of myosin Vb (MyoVb), rab4 and EH domain containing 1 protein (EHD-1) protein levels in brain endothelial cells (BECs). Western blot quantification of (FIG. 10A) MyoVb, (FIG. 10B) Rab4 and (FIG.

[0061] #14794310v110C) EHD-1 in BECs treated with PBS, P-EVs and Br-EVs (mean ± SD; three independent experiments). Statistical analysis was performed using the Kruskal-Wallis test with Dunn’s multiple comparisons test, ns = not significant. GAPDH, Glyceraldehyde 3-phosphate dehydrogenase.

[0062] FIGs. 11A-11D show the cerebral micro vessel isolation and characterization protocol.

[0063] FIG. 11A was created using Biorender®. Schematic describing the in vivo protocol for cerebral micro vessel isolation. FIG. 11B shows representative immunofluorescence image showing an isolated cerebral microvessel. FIG. 11C shows sample total protein quantification by Bradford assay.

[0064] FIG. 11D shows cerebral micro vessel protein expression by Western Blot. CD31, cluster of differentiation 31; GFAP, glial fibrillary acidic protein; Rabllfip2, rabll effector protein 2; Zo-1, zona occludens 1.

[0065] FIG. 12 shows uncropped Western blot membranes of in vivo cerebral microvessel samples from mice treated with PBS (left), P-EVs (center) and Br-EVs (right) probed for Rabi lfip2, Rabi lfip3 (the fourth sample in the PBS blot was excluded from quantification due to poor band quality), Rabi lfip5 (bottom) and GAPDH as the loading control. The black squares highlight the bands analyzed for signal normalization, whose raw values were plotted in FIG. 2A.

[0066] FIGs. 13A-13C show P-EVs in vivo caused changes in cerebral microvessel protein levels. FIG. 13A shows the top 20 significantly upregulated and (FIG. 13B) downregulated proteins in P-EV samples compared to PBS samples. FIG. 13C shows top significantly modulated disease in P-EV samples compared to PBS samples. Comparison analysis was performed using Ingenuity Pathway Analysis software (n = 5 per group).

[0067] FIG. 14 shows representative fluorescence microscopy images of the whole brain sections of mice treated with PBS (n=5) stained for CD31 (green) to detect brain endothelial cells (BECs), MAP2 (magenta) to detect mature neurons and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 100 pm (middle and right columns).

[0068] FIG. 15 shows representative fluorescence microscopy images of the whole brain sections of mice treated with P-EVs (n=4) stained for CD31 (green) to detect brain endothelial cells (BECs), MAP2 (magenta) to detect mature neurons and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 100 pm (middle and right columns).

[0069] FIG. 16 shows representative fluorescence microscopy images of the whole brain sections of mice treated with Br-EVs (n=5) stained for CD31 (green) to detect brain endothelial cells (BECs), MAP2 (magenta) to detect mature neurons and nuclei (DAPI, blue). Scale bars #14794310v11000 pm. White rectangles outline the areas showing increased CD31 signal intensity. For further images, refer to the main manuscript, FIGs. 6A-6G.

[0070] FIGs. 17A-17C show whole brain mean fluorescence intensity (MFI) of (FIG. 17A) CD31, (FIG. 17B), glial fibrillary acidic protein (GFAP), and (FIG. 17C) microtubule-associated protein 2 (MAP2) were quantified from brain sections of mice treated with PBS, P-EVs and Br-EVs. Statistical analysis was performed using the Mann- Whitney t-test. ns = not significant.

[0071] FIG. 18 shows representative fluorescence microscopy images of the whole brain sections of mice treated with PBS (n=5) stained for GFAP (red) to detect astrocytes and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 200 pm (middle and right columns).

[0072] FIG. 19 shows representative fluorescence microscopy images of the whole brain sections of mice treated with P-EVs (n=4) stained for GFAP (red) to detect astrocytes and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 200 pm (middle and right columns).

[0073] FIG. 20 shows representative fluorescence microscopy images of the whole brain sections of mice treated with Br-EVs (n=5) stained for GFAP (red) to detect astrocytes and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 200 pm (middle and right columns).

[0074] FIG. 21 shows representative fluorescence microscopy images of the whole brain sections of mice treated with PBS (n=5) stained for CD1 lb (green) and CD45 (red) to detect microglia (CDllbhlghCD45low) and macrophages (CDllbhlghCD45hlgh) and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 200 pm (middle and right columns).

[0075] FIG. 22 shows representative fluorescence microscopy images of the whole brain sections of mice treated with P-EVs (n=4) stained for CD1 lb (green) and CD45 (red) to detect microglia (CDllbhlghCD45low) and macrophages (CDllbhlghCD45hlgh) and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 200 pm (middle and right columns).

[0076] FIG. 23 shows representative fluorescence microscopy images of the whole brain sections of mice treated with Br-EVs (n=5) stained for CDllb (green) and CD45 (red) to detect microglia (CDllbhlghCD45low) and macrophages (CDllbhlghCD45hlgh) and nuclei (DAPI, blue). Scale bars 1000 pm (left column) and 200 pm (middle and right columns).

[0077] FIG. 24 is a flowchart of an illustrative process 2400 for clustering a plurality of cells of a cell barrier using motility and / or morphology features of the cells, in accordance with some embodiments of the technology described herein.

[0078] FIG. 25 depicts an illustrative implementation of a computer system that may be used in connection with some embodiments of the technology described herein.

[0079] #14794310v1DETAILED DESCRIPTION

[0080] This disclosure describes methods of determining motility and / or morphology of a plurality of cells of a cell barrier by clustering cells according to motility and / or morphology features. An important area of drug development is identifying compositions (e.g., drugs) that alter the integrity of a particular cell barrier by virtue of altering the motility and / or morphology of cells part of the particular cell barrier. This is important because, in some cases (e.g., treating a brain cancer), it is desirable to alter the motility and / or morphology of the cells of the cell barrier (e.g., blood brain barrier) to allow therapeutics (e.g., cancer therapeutics) to cross the barrier (e.g., blood brain barrier). On the other hand, in other cases (e.g., when treating a lung cancer), it is undesirable to alter the motility and / or morphology of the cells of certain barriers (e.g., the blood brain barrier) to allow the cancer therapeutic to cross such barriers (e.g., the blood brain barrier) because this may result in unwanted and unneeded toxicity (e.g., neurotoxicity).

[0081] Current methods for measuring changes in motility and / or morphology of cells in a cell barrier (e.g., in response to drug treatment) typically use cell microscopy techniques to image upwards of 10,000 cells (e.g., of an in vitro cell barrier model system) and then a user (e.g., a scientist) visually inspects the images to try to determine cell motility and / or morphology, and corresponding effects on the cell barrier. This qualitative analysis can be error prone and subjective. However, performing a quantitative analysis by extracting feature values for thousands of cells and clustering the cells according to extracted features to derive information about corresponding effects on the cell barrier is unattainable using current methods. For example, a user viewing an image with 10,000 cells and making a determination based on 20 morphology and / or motility features per cell would have to determine 200,000 feature values, and then distill this information into meaningful information about the cell barrier. In a series of 50 images of the 10,000 cells, the user would have to determine 10,000,000 feature values. Additionally, the user can impart bias into the analysis. This cannot be done with current methods.

[0082] This disclosure describes an improved method of determining motility and / or morphology of a plurality of cells of a cell barrier by clustering the cells accordingly to cell motility and / or morphology features. In some embodiments, the method comprises comparing changes in the cell clusters between two or more cell barrier conditions (e.g., a control condition and a condition where the cell barrier is treated with a chemical) to determine how a change in conditions impacts the motility and / or morphology of the cells of the cell barrier and / or the effects on the cell barrier. This method is improved because it identifies clusters of cells in a #14794310v1cell barrier, and changes in those clusters in two different conditions, that were not identified using previous methods. As shown in the Examples, this method identifies changes in blood brain barrier cell morphology and motility that can be attributed to contact of the blood brain barrier with an exosome secreted from breast cancer cells. The changes observed changes were consistent with degradation of the blood brain barrier that would result in cancer cells being able to cross the blood brain barrier. Additionally, in some aspects, this method comprises clustering using an unsupervised clustering technique, which is not biased by training data or user analysis.

[0083] Methods described herein (e.g., illustrative process 2400) are non-conventional and use clustering methods to evaluate the composition and / or function of the blood brain barrier by analyzing thousands to tens of thousands of cells. No conventional technique is operable at that scale. In other words, the methods described herein (e.g., illustrative process 2400) cannot and are not done using conventional methods.

[0084] FIG. 24 is a flowchart of an illustrative process 2400 for determining motility and / or morphology of a plurality of cells of a cell barrier. The process 2400 may be performed by using any suitable computing device(s). For example, process 2400 may be performed by using multiple processors, whether located in one physical location or across multiple physical locations. The process 2400 may be performed in a cloud computing environment, in some embodiments.

[0085] Process 2400 begins at act 2402, where an image of a plurality of cells of the cell barrier is obtained.

[0086] A “cell barrier” as used herein includes a layer of cells that protects an organism, separates organ systems of an organism and / or controls the passage of material through the cell barrier. A cell barrier may be any cell barrier including an in vitro cell barrier (e.g., an in vitro model of a cell barrier) or an in vivo cell barrier. In some embodiments, a cell barrier is an endothelial barrier, an epithelium barrier, or / an epithelial barrier. In some embodiments, a cell barrier is an endothelial barrier. In some embodiments, a cell barrier is an epithelium barrier. In some embodiments, a cell barrier is an epithelial barrier. In some embodiments, a cell barrier is a blood brain barrier. In some embodiments, the cell barrier is a monolayer of cells (i.e., a simple cell barrier). In some embodiments, the cell barrier is a multi-layer of cells (i.e., a stratified cell barrier).

[0087] In some embodiments, a cell barrier is a cell barrier model. In some embodiments, the cell barrier model is representative of an in vivo cell barrier. A cell barrier model that is representative of an in vivo cell barrier mimics one or properties of the in vivo cell barrier.

[0088] Many such models are known, e.g., as described in Xu, Yining, et al. Advanced Drug Delivery #14794310v1Reviews 175 (2021): 113795; and Shah, Brijesh et al. Current drug delivery 19.10 (2022): 1034-1046. In some embodiments, the cell barrier model is a an endothelial, an epithelium, or / an epithelial cell barrier model. In some embodiments, the cell barrier model is a blood brain barrier model (e.g., as described in Dong X. Curr Drug Deliv. 2022;19(10):1034-1046. PMID: 35240972). In some embodiments, the cell barrier model is a two-dimensional model (e.g., a cell monolayer). In some embodiments, the cell barrier model is a three-dimensional model.

[0089] A “plurality of cells” of a cell barrier refers to at least 2 cells (e.g., at least two cell of a cell barrier). In some embodiments, a plurality of cells comprises at least 5 cells (e.g., at least 5 cells, at least 10 cells, at least 50 cells, at least 100 cells, at least 500 cells, at least 1000 cells, at least 5000 cells, at least 10,000 cells, at least 15,000 cells, at least 25,000 cells, at least 50,000 cells, or at least 100,000 cells). In some embodiments, a plurality of cells of a cell barrier comprises 5,000-50,000 cells. In some embodiments, a plurality of cells comprises 5, GOO-15, 000 cells.

[0090] An image of a plurality of cells of a cell barrier (e.g., a model of cell barrier) may be obtained in any suitable way. In some embodiments, obtaining an image comprises obtaining an image that was acquired using microscopy (e.g., confocal imaging, fluorescent imaging, or quantitative phase imaging). In some embodiments, the obtaining an image comprises capturing an image. In some embodiments, capturing the image comprises capturing using microscopy (e.g., confocal imaging, fluorescent imaging, or quantitative phase imaging). In some embodiments, obtaining an image of the plurality of cells comprises obtaining an image of live cells of a cell barrier. In some embodiments, obtaining an image of the plurality of cells comprises obtaining an image of fixed cells of a cell barrier. In some embodiments, obtaining an image of the plurality of cells comprises obtaining an image of cells of an endothelial barrier, an epithelium barrier, or / an epithelial barrier. In some embodiments, obtaining an image of the plurality of cells comprises obtaining an image of cells of a blood brain barrier.

[0091] In some embodiments, obtaining an image of a plurality of cells of a cell barrier comprises obtaining a time series images of the plurality of cells. In some embodiments, obtaining a time series images of the plurality of cells comprises obtaining at least 2 images (e.g., at least 2 images, at least 5 images, at least 10 images, at least 25 images, at least 50 images, at least 75 images, at least 100 images, at least 250 images, at least 500 images, at least 750 images, at least 1000 images, or at least 10,000 images). In some embodiments, obtaining a time series images of the plurality of cells comprises obtaining 10-1000 images, 50-1000 images, 100-1000 images, 10-100 images, 50-100 images, 100-500 images, 100-400 images,

[0092] #14794310v1100-300 images, or 100-200 images. In some embodiments, obtaining a time series images of the plurality of cells comprises obtaining a series of images of live cells.

[0093] Process 2400 involves act 2404, segment an image (e.g., segment a single image or a time series of images) to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells.

[0094] In some embodiments, an image comprises at least 16,384 pixels (e.g., 128x128). In some embodiments, an image comprises at least 65.536 pixels (e.g., 256x256). In some embodiments, an image comprises at least 262144 pixels (e.g., 512x512). In some embodiments, an image comprises at least 1,048,576 pixels (e.g., 1024x1024). In some embodiments, an image comprises at least 4,194,304 pixels (e.g., 2048x2048). In some embodiments, an image comprises at least 16,777,216 pixels (e.g., 4096x4096).

[0095] Segmenting includes the process of identifying and separating an image of an individual cell from other images of cells within a larger image. Any suitable method may be used to segment images. An illustrative method of segmenting images to obtain segmented cell data is described in Wen, Tingxi, et al. Computer Methods and Programs in Biomedicine 227 (2022): 107211; doi: 10.1016 / j.cmpb.2022.107211. In some embodiments, segmenting an image to obtain segmented cell data comprising segmented image data for cells of the plurality of cells comprises segmenting using a trained neural network. In some embodiments, segmenting an image to obtain segmented cell data comprising segmented image data for cells of the plurality of cells comprises segmenting using MARS-Net, e.g., as described in Jang J et al., STAR Protoc. 2022 Jun 14;3(3): 101469. PMCID: PMC9207580. In some embodiments, segmenting an image to obtain segmented cell data comprising segmented image data for cells of the plurality of cells comprises segmenting using CellPose Stringer et al. Nat Methods. 2021 ; 18(1): 100-6; PMID: 33318659. In some embodiments, the segmented cell data comprises segmented cell data for a plurality of cells (e.g., least 2 cells, least 5 cells, least 10 cells, at least 50 cells, at least 100 cells, at least 500 cells, at least 1000 cells, at least 10,000 cells, or at least 100,000 cells). In some embodiments, segmented cell data comprises a sequence of segmented images of a cell (e.g., a sequence of images that are ordered accordingly to the time series acquisition of the images). In some embodiments, segmented cell data comprises a sequence of segmented images of a plurality of cells.

[0096] Process 2400 involves act 2406 extract motility and / or morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier.

[0097] Motility and / or morphology feature values may be extracted from segmented image data using any suitable means. In some embodiments, extracting motility and / or morphology #14794310v1features from segmented image data comprises extracting using TrackMate computational tools, e.g., as described in Jaqaman K. Nat Methods. 222008;5(8):695-702. PMID: 18641657.

[0098] “Motility” as used in the context of cell motility refers to cell movement. In some embodiments, cell motility may be determined based on values of one or more cell motility features (e.g., features that are indicative of cell movement and / or whose values are influenced by cell movement). In some embodiments, extracting motility feature values from segmented image data (e.g., a time series of segmented image data) for at least some cells in the plurality of cells of the cell barrier comprises extracting feature values from segmented image data for most of the cells in the plurality of cells of the cell barrier. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting feature values from segmented image data for all the cells in the plurality of cells of the cell barrier.

[0099] In some embodiments, extracting motility feature values from segmented image data for at least some cells (e.g., most of the cells or all the cells) in the plurality of cells of the cell barrier comprises extracting values of one or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of two or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of three or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of four or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of five or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, #14794310v1confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of six or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting value of seven or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of eight or more of the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled. In some embodiments, extracting motility feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of all the following features: standard deviation (std) of speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and / or max distance travelled.

[0100] In some embodiments, cell motility feature values may be determined as follows:

[0101] Standard Deviation (std) of Speed

[0102] • Description: Measures the variability or dispersion in the instantaneous speeds of the cell.

[0103] • Calculation: std_speed =

[0104]

[0105] o

[0106]

[0107] Instantaneous speed at time i

[0108] o v Mean speed

[0109] o N: Number of time points

[0110] Max Speed

[0111] Description: The highest instantaneous speed achieved by the cell.

[0112] Calculation: max_speed = max (ltv2, v3, ... , vN)

[0113] #14794310v1o vt: Instantaneous speed at time i

[0114] Mean Speed

[0115] • Description: The average of all instantaneous speeds.

[0116] • Calculation:

[0117]

[0118] Min Speed

[0119] • Description: The lowest instantaneous speed, typically close to zero if the cell pauses.

[0120] • Calculation: min_speed = min (ltv2, v3, ... , vN)

[0121] Mean Straight Line Speed

[0122] • Description: The average speed calculated along the straight-line path from the cell's starting to ending position.

[0123] • Calculation: mean_straight_line_speed = straight Jine iistance / total Jime o Straight-line distance = Distance between the start and end positions.

[0124] Linearity of Forward Progression (LFP)

[0125] • Description: The ratio of the straight-line distance to the total distance traveled, indicating the straightness of the path. This represents the stability and directionality of movement. Range is -1 to 1.

[0126] • Calculation: LFP = straight Jine iistance / total_distance

[0127] Confinement Ratio

[0128] • Description: Another measure of path straightness, similar to LFP. This feature focuses on whether the movement is constrained or tortuous. Range is 0 to 1.

[0129] • Calculation: confinement_ratio = straight Jine iistance I end_to_end_distance

[0130] Total Distance Traveled

[0131] • Description: The sum of distances between successive positions.

[0132] • Calculation:

[0133]

[0134] 0(.xi> yt)- Cell positions at time i.

[0135] Max Distance Traveled

[0136] #14794310v1Description: The maximum distance traveled in a single step or interval.

[0137] • Calculation:

[0138]

[0139] “Morphology” as used in the context of cell morphology refers to cell shape, size, structure, and / or appearance. In some embodiments, cell morphology may be determined based on one or more cell morphology features (e.g., features that are indicative of cell shape, size, structure, and / or appearance, and / or whose values are influenced by cell shape, size, structure, and / or appearance). In some embodiments, extracting morphology feature values from segmented image data (e.g., a time series of segmented image data) for at least some cells in the plurality of cells of the cell barrier comprises extracting feature values from segmented image data for most of the cells in the plurality of cells of the cell barrier. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting feature values from segmented image data for all the cells in the plurality of cells of the cell barrier.

[0140] In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of one or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of two or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of three or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of four or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of five or more of the following features: area, convex area, equivalent diameter area, minor axis length, max

[0141] #14794310v1ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of six or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of seven or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of eight or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of nine or more of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity. In some embodiments, extracting morphology feature values from segmented image data for at least some cells in the plurality of cells of the cell barrier comprises extracting values of all of the following features: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity.

[0142] In some embodiments, cell morphology feature values may be determined as follows:

[0143] Area

[0144] • Description: The number of pixels contained within a cell.

[0145] • Calculation: Count the number of 'True' pixels in the binary mask that represent the cell’s interior region.

[0146] Convex Area

[0147] • Description: The area of the convex hull (the smallest convex polygon that can encompass the cell’s interior region).

[0148] • Calculation: Count the pixels inside the convex hull of the cell’s interior region.

[0149] Equivalent Diameter Area

[0150] #14794310v1• Description: The diameter of a circle with the same area as the cell’s interior region.

[0151] • Calculation:

[0152]

[0153] Minor Axis Length

[0154] • Description: The length of the shortest axis of the ellipse that has the same normalized second central moments as the cell’s interior region.

[0155] • Calculation: Derived from the ellipse fit to the cell’s interior region.

[0156] Max Feret Diameter

[0157] • Description: The maximum distance between any two points on the boundary of the cell’s interior region.

[0158] • Calculation: Compute the pairwise distances between all boundary points and take the maximum.

[0159] Major Axis Length

[0160] • Description: The length of the longest axis of the ellipse that has the same normalized second central moments as the cell’s interior region.

[0161] • Calculation: Derived from the ellipse fit to the cell’s interior region.

[0162] Perimeter

[0163] • Description: The length of the boundary of the cell’s interior region.

[0164] • Calculation: Count the boundary pixels in the binary mask or approximate the length using a polygonal representation of the boundary.

[0165] 8. Solidity

[0166] • Description: The proportion of the convex hull area that is occupied by the cell’s interior region.

[0167] • Calculation: solidity = area / convex_area

[0168] Extent

[0169] • Description: The ratio of the cell’s interior region's area to the area of its bounding box.

[0170] Calculation: extent = area / bounding_box_area

[0171] o Bounding box area is the (width x height) of the bounding box.

[0172] #14794310v1Eccentricity

[0173] • Description: A measure of how elongated the cell’s interior region is, defined as the ratio of the distance between the foci of the ellipse and its major axis length.

[0174] • Calculation: eccentricity = / 1 — b2 / a2

[0175] o a: Length of the major axis.

[0176] o b: Length of the minor axis.

[0177] In some embodiments, extracting motility and / or morphology feature values further comprises reducing the dimensions of the extracted motility and / or morphology feature values. The dimensions of the extracted motility and / or morphology feature values may be reduced in any suitable way (e.g., using principal component analysis (PC A), non-negative matrix factorization (NNMF), Kernel PCA, Graph-based kernel PCA, linear discriminant analysis (LDA), generalized discriminant analysis (GDA), an autoencoder, T-distributed stochastic neighbor embedding (t-SNE), and / or uniform manifold approximation and projection (UMAP)). In some embodiments, reducing the dimensions of the extracted motility and / or morphology feature values comprises reducing using uniform manifold approximation and projection (UMAP).

[0178] Process 2400 involves act 2408 cluster the at least some cells of the plurality of cells, using their respective motility and / or morphology feature values, to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology.

[0179] In some embodiments, the method comprises clustering most of the cells of the plurality of cells. In some embodiments, the method comprises clustering all the cells of the plurality of cells. In some embodiments, the method comprises clustering at least some cells of the plurality of cells using their respective motility feature values. In some embodiments, the method comprises clustering at least some cells of the plurality of cells using their respective morphology feature values.

[0180] “Clustering” cells (e.g., cells of the plurality of cells) refers to organizing and / or classifying cells into groups (i.e., clusters) based on similarities of the cells (e.g., similarities in cell motility and / or morphology). In some embodiments, clustering cells of a plurality of cells of a cell barrier comprises clustering using one or more of the extracted motility and / or morphology feature values. In some embodiments, clustering cells of a plurality of cells of a cell barrier comprises clustering using at least 2 (e.g., at least 3, at least 4, at least 5, at least 6, at

[0181] #14794310v1least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, or at least 20) of the extracted motility and / or morphology feature values. In some embodiments, clustering cells of a plurality of cells of a cell barrier comprises clustering using all the extracted motility and / or morphology feature values.

[0182] Clustering may be performed using any suitable method. In some embodiments, the clustering is unsupervised clustering. In some embodiments, clustering is supervised clustering. In some embodiments, clustering cells of a plurality of cells (e.g., using motility and / or morphology feature values) comprises clustering using Centroid-based Clustering (Partitioning methods), Density-based Clustering (Model-based methods), Connectivity-based Clustering (Hierarchical clustering), Distribution-based Clustering and / or spectral clustering. In some embodiments, clustering cells of the plurality of cells comprises non-hierarchical spectral clustering (e.g., as described in A. Y. Ng MIJaYW. On Spectral Clustering: Analysis and an Algorithm: MIT Press, Boston, MA; 2002). In some embodiments, clustering cells of the plurality of cells comprises clustering using a silhouette score (e.g., a maximized silhouette score) e.g., as described in Rousseeuw PJ. Journal Of Computational and Applied Mathematics.

[0183] 1987;20:53-65. In some embodiments, clustering cells of the plurality of cells comprises non-hierarchical spectral clustering with a Silhouette score (e.g., a maximized Silhouette score).

[0184] In some embodiments, clustering cells of the plurality of cells comprises clustering into a plurality of clusters (e.g., at least 2 clusters, at least 3 clusters, at least 4 clusters, at least 5 clusters, at least 6 clusters, at least 7 clusters, at least 8 clusters, at least 9 clusters, at least 10 clusters, at least 15 clusters, or at least 20 clusters). In some embodiments, clustering cells of the plurality of cells comprises clustering into 2-6 clusters. In some embodiments, clustering cells of the plurality of cells comprises clustering into 2, 3, 4, 5, 6, 7, 8, 9, and / or 10 clusters. In some embodiments, clustering cells of the plurality of cells comprises determining an amount (e.g., an absolute amount, a percentage, and / or a fraction) of cells in each cluster.

[0185] In some embodiments, each cluster of a plurality of clusters represents a class of cell motility and / or morphology. In some embodiments, a class of cell motility and / or morphology is characterized by a specific range of values of one or more cell motility and / or cell morphology feature values of the cells in the class. In some embodiments, a class of cell motility and / or morphology is characterized by a common change in one or more cell motility and / or cell morphology feature values of the cells in the class (e.g., compared to control cells). In some embodiments, a class of cell motility and / or morphology is characterized by a common change in at least 1 (e.g., at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least #14794310v19, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, or at least 20) cell motility and / or cell morphology feature values of the cells in the class (e.g., compared to control cells). In some embodiments, a common change in cell motility and / or morphology feature value refers to an increase or decrease in the cell motility and / or morphology feature value of cells of the class. For example, a class of cell motility and / or morphology may be characterized by an increase in Max Feret diameter (e.g., compared to a control). In another example, a class of cell motility and / or morphology may be characterized by decreased in Max Feret diameter. In some embodiments, a class of cell motility and / or morphology may be characterized by an increase in one or more cell motility and / or morphology feature values and a decrease one or more different cell motility and / or morphology feature values. For example, a class of cell motility and / or morphology may be characterized by increased Max Feret diameter, increased major axis length, increased perimeter, decreased solidity and decreased extent.

[0186] In some embodiments, process 2400 involves act 2408 determine an effect of a perturbation on the motility and / or morphology of the cells of the plurality of cells of the cell barrier

[0187] In some embodiments, a method of determining motility and / or morphology of a plurality of cells of a cell barrier comprises determining an effect of a perturbation (e.g., a chemical, a drug, a cell, a liposome, or a mechanical perturbation) on the motility and / or morphology of the cells of the plurality of cells of the cell barrier compared to a control (e.g., a cell barrier that has not received the perturbation).

[0188] A “perturbation” to cells of a cell barrier may be any perturbation. In some embodiments, the perturbation comprises contacting the cell barrier with a chemical. In some embodiments, the perturbation comprises contacting the cell barrier with a drug. In some embodiments, the perturbation comprises contacting the cell barrier with a molecule from a cancer cell (e.g., an extracellular vesical from a cancer cell). In some embodiments, the perturbation comprises contacting the cell barrier with media collected from a cancer cell culture (e.g., a breast cancer cell culture). In some embodiments, the perturbation comprises contacting the cell barrier with a molecule (e.g., an exosome) secreted by a cancer cell (e.g., a lung cancer cell, a breast cancer cells, a pancreatic cancer cell, a liver cancer cell, a blood cancer cells, a lymph cancer cells, a kidney cancer cells, a colon cancer cell, a rectal cancer cells, a skin cancer cells, or a bone cancer cell). In some embodiments, the perturbation comprises contacting the cell barrier with a cancer cell. In some embodiments, the control comprises contacting the cell barrier with the carrier substance for carrying a chemical, a drug, or a cellular perturbation. For #14794310v1example, if extracellular vesicles are suspended in phosphate buffer prior to contacting with a cell barrier then the control may be contacting the control cell barrier with phosphate buffer that does not comprise the extracellular vesicles). In some embodiments, the control comprises contacting the cell barrier with a substance that is expected to degrade the cell barrier (e.g., a substance that makes the cell barrier leaky). For example, Matrix metalloproteinases (MMPs) are known to degrade the BBB and result in a leaky BBB. In another example, increased temperature (e.g., 38-42 °C) can degrade the BBB and result in a leaky BBB.

[0189] In some embodiments, determining an effect of a perturbation on the motility and / or morphology of the cells of the plurality of cells of the cell barrier (perturbed cells) compared to a control (control cells) comprises clustering the perturbed cells to produce perturbed cell clusters, clustering the control cells to produce control cell clusters, and determining one or more changes between the perturbed cell clusters and the control cell clusters. In some embodiments, the method comprises using the same cell motility and / or morphology classes when clustering the perturbed cells and the control cells, and determining one or more changes between the perturbed cell clusters and the control cell clusters. In some embodiments, determining one or more changes between the perturbed cell clusters and the control cell clusters comprises determining a change in the amount of cells per cluster (e.g., a change in the number of cells, fraction or cells, or percentage of cell per cluster). In some embodiments, determining one or more changes between the perturbed cell clusters and the control cell clusters comprises determining a change in the number of clusters. In some embodiments, determining one or more changes between the perturbed cell clusters and the control cell clusters comprises determining a change in the characteristics of the clusters.

[0190] In some embodiments, process 2400 involves act 2412 determine changes in cell motility and / or morphology of the plurality of cells of the cell barrier in different conditions.

[0191] In some embodiments, this disclosure provides a method of determining changes in cell motility and / or morphology in a cell barrier in different conditions (e.g., under perturbation), the method comprising: obtaining a first image of a plurality of cells of a cell barrier in a first condition, and a second image of a plurality of cells of a cell barrier in a second condition; determining first motility and / or morphology feature values of the cell barrier in the first condition and second motility and / or morphology feature values of the cell barrier in the second condition (e.g., where the first motility and / or morphology features and the second motility and / or morphology features are the same features), the determining comprising for the first #14794310v1image and the second image, respectively: segmenting the image to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells; extracting motility and / or morphology feature values from the segmented image data for each of the cells of the plurality of cells; and clustering the cells of the plurality of cells to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology; and determining changes in cell motility and / or morphology of the cell barrier in the different conditions, the determining comprising identifying one or more changes in the plurality of clusters in the first condition and the second condition.

[0192] In some aspects, this disclosure provide a system comprising at least one computer hardware processor; and at least one non-transitory computer-readable storage medium that, when executed by the at least one computer hardware processor, causes the at least one computer hardware processor to perform a method described herein. In some aspects, this disclosure provide a system comprising at least one computer hardware processor; and at least one non-transitory computer-readable storage medium that, when executed by the at least one computer hardware processor, causes the at least one computer hardware processor to perform a method of determining motility and / or morphology of a plurality of cells of a cell barrier, the cell barrier being a monolayer cell barrier or a multi-layer cell barrier, the method comprising: using at least one computer hardware processor to perform: obtaining an image of the plurality of cells of the cell barrier; segmenting the image to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells; extracting motility and / or morphology feature values from the segmented image data for at least some cells in the plurality of cells; clustering the at least some cells of the plurality of cells, using their respective motility and / or morphology features feature values, to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology.

[0193] An illustrative implementation of a computer system 2500 that may be used in connection with any of the embodiments of the technology described herein (e.g., such as the process of FIG. 24) is shown in FIG. 25. The computer system 2500 includes one or more processors 2504 and one or more articles of manufacture that comprise non-transitory computer-readable storage media (e.g., memory 2510 and one or more non-volatile storage media 2506). The processor 2504 may control writing data to and reading data from the memory 910 and the non-volatile storage device 2506 in any suitable manner, as the aspects of the technology described herein are not limited to any particular techniques for writing or reading data. To perform any of the functionality described herein, the processor 2504 may execute one or more processor-executable instructions stored in one or more non-transitory computer-readable #14794310v1storage media (e.g., the memory 2510), which may serve as non-transitory computer-readable storage media storing processor-executable instructions for execution by the processor 2504.

[0194] Computer system device 2500 may also include a network input / output (I / O) interface 2502 via which the computer system may communicate with other computing devices (e.g., over a network), and may also include one or more user I / O interfaces 2508, via which the computer system may provide output to and receive input from a user. The user I / O interfaces 2508 may include devices such as a keyboard, a mouse, a microphone, a display device (e.g., a monitor or touch screen), speakers, a camera, and / or various other types of I / O devices.

[0195] The above-described embodiments can be implemented in any of numerous ways. For example, the embodiments may be implemented using hardware, software, or a combination thereof. When implemented in software, the software code can be executed on any suitable processor (e.g., a microprocessor) or collection of processors, whether provided in a single computing device or distributed among multiple computing devices. Further, it should be appreciated that a computer may be embodied in any of a number of forms, such as a rackmounted computer, a desktop computer, a laptop computer, or a tablet computer, as non-limiting examples. Additionally, a computer may be embedded in a device not generally regarded as a computer but with suitable processing capabilities, including a Personal Digital Assistant (PDA), a smartphone, a tablet, or any other suitable portable or fixed electronic device.

[0196] In this respect, it should be appreciated that one implementation of the embodiments described herein comprises at least one non-transitory computer-readable storage medium (e.g., RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or other tangible, non-transitory computer-readable storage medium) encoded with a computer program (i.e., a plurality of executable instructions) that, when executed on one or more processors, performs the above-described functions of one or more embodiments (e.g., part of or all of the processes described above with reference to FIG.

[0197] 24). The computer-readable medium may be transportable such that the program stored thereon can be loaded onto any computing device to implement aspects of the techniques described herein. In addition, it should be appreciated that the reference to a computer program which, when executed, performs any of the above-described functions, is not limited to an application program running on a host computer. Rather, the terms computer program and software are used herein in a generic sense to reference any type of computer code (e.g., application software, firmware, microcode, or any other form of computer instruction) that can be employed to program one or more processors to implement aspects of the techniques described herein.

[0198] #14794310v1Additionally, it should be appreciated that according to one aspect, one or more computer programs that when executed perform methods of the present disclosure need not reside on a single computer or processor, but may be distributed in a modular fashion among a number of different computers or processors to implement various aspects of the present disclosure.

[0199] Having thus described several aspects and embodiments of the technology set forth in the disclosure, it is to be appreciated that various alterations, modifications, and improvements will readily occur to those skilled in the art. Such alterations, modifications, and improvements are intended to be within the spirit and scope of the technology described herein. For example, those of ordinary skill in the art will readily envision a variety of other means and / or structures for performing the function and / or obtaining the results and / or one or more of the advantages described herein, and each of such variations and / or modifications is deemed to be within the scope of the embodiments described herein. Those skilled in the art will recognize or be able to ascertain using no more than routine experimentation many equivalents to the specific embodiments described herein. It is, therefore, to be understood that the foregoing embodiments are presented by way of example only and that, within the scope of the appended claims and equivalents thereto, inventive embodiments may be practiced otherwise than as specifically described. In addition, any combination of two or more features, systems, articles, materials, kits, and / or methods described herein, if such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent, is included within the scope of the present disclosure.

[0200] The above-described embodiments can be implemented in any of numerous ways. One or more aspects and embodiments of the present disclosure involving the performance of processes or methods may utilize program instructions executable by a device (e.g., a computer, a processor, or other device) to perform, or control performance of, the processes or methods. In this respect, various inventive concepts may be embodied as a computer readable storage medium (or multiple computer readable storage media) (e.g., a computer memory, one or more floppy discs, compact discs, optical discs, magnetic tapes, flash memories, circuit configurations in Field Programmable Gate Arrays or other semiconductor devices, or other tangible computer storage medium) encoded with one or more programs that, when executed on one or more computers or other processors, perform methods that implement one or more of the various embodiments described above. In some embodiments, computer readable media may be non-transitory media.

[0201] Computer-executable instructions may be in many forms, such as program modules, executed by one or more computers or other devices. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or #14794310v1implement particular abstract data types. Typically, the functionality of the program modules may be combined or distributed as desired in various embodiments.

[0202] Also, data structures may be stored in computer-readable media in any suitable form. For simplicity of illustration, data structures may be shown to have fields that are related through location in the data structure. Such relationships may likewise be achieved by assigning storage for the fields with locations in a computer-readable medium that convey a relationship between the fields. However, any suitable mechanism may be used to establish a relationship between information in fields of a data structure, including through the use of pointers, tags or other mechanisms that establish a relationship between data elements.

[0203] Also, a computer may have one or more input and output devices. These devices can be used, among other things, to present a user interface. Examples of output devices that can be used to provide a user interface include printers or display screens for visual presentation of output and speakers or other sound generating devices for audible presentation of output.

[0204] Examples of input devices that can be used for a user interface include keyboards, and pointing devices, such as mice, touch pads, and digitizing tablets. As another example, a computer may receive input information through speech recognition or in other audible formats.

[0205] Such computers may be interconnected by one or more networks in any suitable form, including a local area network or a wide area network, such as an enterprise network, and intelligent network (IN) or the Internet. Such networks may be based on any suitable technology and may operate according to any suitable protocol and may include wireless networks, wired networks or fiber optic networks.

[0206] Also, as described, some aspects may be embodied as one or more methods. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different than illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

[0207] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0208] The indefinite articles “a” and “an,” as used herein in the specification and in the claims, unless clearly indicated to the contrary, should be understood to mean “at least one.”

[0209] The phrase “and / or,” as used herein in the specification and in the claims, should be understood to mean “either or both” of the elements so conjoined, i.e., elements that are conjunctively present in some cases and disjunctively present in other cases. Multiple elements #14794310v1listed with “and / or” should be construed in the same fashion, i.e., “one or more” of the elements so conjoined. Other elements may optionally be present other than the elements specifically identified by the “and / or” clause, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, a reference to “A and / or B”, when used in conjunction with open-ended language such as “comprising” can refer, in one embodiment, to A only (optionally including elements other than B); in another embodiment, to B only (optionally including elements other than A); in yet another embodiment, to both A and B (optionally including other elements); etc.

[0210] As used herein in the specification and in the claims, the phrase “at least one,” in reference to a list of one or more elements, should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each and every element specifically listed within the list of elements and not excluding any combinations of elements in the list of elements. This definition also allows that elements may optionally be present other than the elements specifically identified within the list of elements to which the phrase “at least one” refers, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, “at least one of A and B” (or, equivalently, “at least one of A or B,” or, equivalently “at least one of A and / or B”) can refer, in one embodiment, to at least one, optionally including more than one, A, with no B present (and optionally including elements other than B); in another embodiment, to at least one, optionally including more than one, B, with no A present (and optionally including elements other than A); in yet another embodiment, to at least one, optionally including more than one, A, and at least one, optionally including more than one, B (and optionally including other elements); etc.

[0211] In the claims, as well as in the specification above, all transitional phrases such as “comprising,” “including,” “carrying,” “having,” “containing,” “involving,” “holding,” “composed of,” and the like are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases “consisting of’ and “consisting essentially of’ shall be closed or semi-closed transitional phrases, respectively.

[0212] Methods of Treatment

[0213] In some embodiments, this disclosure provides a method of improving blood brain barrier integrity (e.g., restoring the blood brain barrier to a normal state and / or decreasing the permeability of the blood brain barrier to one or more molecules) (e.g., in a human subject). In some embodiments, the method comprises inhibiting the expression and / or activity of

[0214] #14794310v1Rabllfip3, Rabllfip5 and / or NKCC1 in BECs (e.g., using a small molecule inhibitor, gene editing, RNAi, miRNA, siRNA, antisense oligonucleotides (ASOs) and / or antibodies that that target one or more of Rabi lfip3, Rabi lfip5 and / or NKCC1). In some embodiments, the method comprises increasing the expression and / or activity of Rab7 and Rabi lfip2 in BECs (e.g., using a small molecule, gene therapy (e.g., transfection of BECs with a polynucleotide encoding Rab7 and / or Rabi lfip2 into a BEC), and / or CRISPR activation targeting Rab7 and / or Rabi lfip2. In some embodiments, this disclosure provides a method of decreasing brain metastasis (e.g., decreasing brain metastasis of triple negative breast cancer) in a subject (e.g., a human subject), the method comprising inhibiting the expression and / or activity of Rabllfip3, Rabllfip5 and / or NKCC1 in BECs of the subject, and / or increasing the expression and / or activity of Rab7 and Rabi lfip2 in BECs.

[0215] EXAMPLES

[0216] Breast cancer brain metastasis occurs in up to one-third of patients diagnosed with triplenegative breast cancer (TNBC)1, resulting in a poor prognosis due to the lack of early diagnostic techniques and drug delivery strategies able to efficiently breach the blood-brain barrier (BBB)2-5. The BBB is a specialized and highly selective endothelial barrier primarily consisting of brain endothelial cells (BECs), which form the lumen of BBB blood vessels and interact with the bloodstream. On the side of the brain parenchyma, the BBB is covered by pericytes and astrocyte end-feet, providing additional metabolic and structural support6. Under physiological conditions, the BBB represents a nearly unbreachable barrier that restricts the passage of bloodborne molecules and particles larger than 400 Da6-8. Recent studies showed that nanosized extracellular vesicles (> 1000 KDa) derived from the brain-seeking variant of TNBC cells MDA-MB-2318'9(Br-EVs) breach the BBB via transcytosis involving recycling endosomes expressing Rabll and target astrocytes in the brain parenchyma8, 10. However, the outcome this has on BEC cytoplasmic and cytoskeletal homeostasis as well as barrier properties have remained unknown.

[0217] Here, results show that Br-EVs affect the expression of multiple Rabi 1 effector proteins (Rabi Ifips) in BECs both in vitro and in vivo. Rabi Ifips are known to mediate the binding between Rabll -expressing vesicles and motor and cytoskeletal elements (microtubules and actin filaments), facilitating their transport toward various subcellular compartments (Fig. lb)19, 20. Br-EVs cause a significant downregulation of Rabllfip2, which promotes vesicle recycling to the plasma membrane20-23and upregulation of Rabi lfip319, 24and Rabi lfip520, which support the #14794310v1structural stability of the recycling endosomal compartment and vesicle recycling and transcytosis, respectively. The modulation of multiple Rabllfips involved in intracytoplasmic vesicle trafficking and recycling along cytoskeletal filaments (FIGs. 1A-1B) prompted a hypothesis that Br-EVs disrupt BEC cytoplasmic and cytoskeletal homeostasis, thereby impairing the BBB. Using a robust machine learning (ML) approach, previously unknown morphological and motility changes were identified in BECs induced by Br-EVs that indicate a reorganization of the cell cytoskeleton and cytoplasm, decreased cell-to-cell adhesion and impaired barrier functions. A quantitative analysis of the cerebrovascular proteome from the brain of mice treated with Br-EVs and P-EVs validated the in vitro ML findings demonstrating that Br-EVs drive protein alterations indicative of decreased cell cytoplasm and cytoskeleton organization and dynamics. In particular, Br-EVs significantly increase levels of Na-K-Cl cotransporter sodium-potassium-chloride cotransporter 1 (NKCCl)25, 26in BECs. When upregulated, NKCCl leads to increased cell volume, cell motility27, 28and BBB damage29, 30. Lastly, analyses of the brains of mice treated with Br-EVs and their isolated cerebral micro vessels suggest that Br-EVs prime the BBB for future transendothelial migration of TNBC cells by increasing their adhesion to BECs.

[0218] These findings show that Br-EVs have a significant and multifunctional impact on the BECs of the BBB, affecting the expression of Rabllfip2, Rabllfip3, Rabllfip5 and NKCCl both in vitro and in vivo. All of the identified proteins, whose levels are affected by Br-EVs, interact with the cytoskeletal filaments and co-localize with Rabi 1 -positive recycling compartments. Furthermore, Rabllfip2, Rabllfip3 and Rabllfip5 play crucial roles in the trafficking, recycling and transcytosis of intracytoplasmic vesicles19-22and, together with NKCCl regulate cell morphology27, 31and motility32-34. Being responsible for early changes at the BBB that favor breast-to-brain metastasis, all these identified proteins can be leveraged to develop early diagnostic and therapeutic strategies for the prevention and treatment of breast-to-brain metastasis.

[0219] Results

[0220] Br-EVs modulate multiple Rabllfip proteins involved in the trafficking, recycling and transcytosis of Rabll-positive intracytoplasmic vesicles.

[0221] It has been previously reported that Br-EVs breach the intact BEC monolayer via transcytosis involving recycling endosomes expressing Rabll (whose protein levels are not impacted by Br-EV treatments) and downregulate Rab7 in BECs to inhibit the intracytoplasmic #14794310v1degradation pathways and facilitate Br-EV trafficking toward the brain8. To study the effects of Br-EVs on BEC recycling and trafficking pathways involving Rabi 1 -positive vesicles, BECs were treated for five days with a fixed quantity of small P-EVs and Br-EVs (100-350 nm) isolated and characterized according to the guidelines published by the International Society of Extracellular Vesicles35(FIGs. 8A-8D). BEC protein content was then analyzed, by immunoblot and immunofluorescence, for Rabi 1 effector proteins known to interact with Rabi 1 -expressing endocytic vesicles, motor proteins and cytoskeletal components (FIGs. 1A-1B) and regulate the intracytoplasmic vesicle trafficking to different compartments of the cell19, 20. Specifically, Rabllfip2 mediates the binding between Rabi 1 -vesicles and myosin Vb, thereby facilitating the recycling of Rabi 1 -vesicles back to the plasma membrane20-23. Rabi lfip3 contributes to increased stability within the endosomal compartment, facilitates the positioning and movement of Rabi 1 -positive endosomes relative to the microtubule network and regulates the placement of pericentro somal recycling endosomes during both cell division and breast cancer progression19,24, 49. Rabi lfip5 is a component of the Factor for Endosome Recycling And Rab Interactions (FERARI)21, 22, a multi-subunit tethering complex required for Rabi 1 -dependent recycling and reported to be involved in transcytosis and endosomal recycling and trafficking20. The results showed that compared with PBS or P-EV-treated BECs, cells treated with Br-EVs showed a significant decrease in Rabllfip2 (FIGs. 1B-1E) and a significant increase in Rabllfip3 (FIG. IB, FIGs. 1F-1H) and Rabllfip5 (FIG. IB, FIGs. 1I-1K). Although Br-EVs modulate multiple Rabllfip proteins involved in cytoplasmic vesicle recycling, other proteins such as myosin Vb (a member of the myosin family of actin-based motor proteins known to interact with Rabi lfip2)23, 50-52, rab4 (an early endosome protein that regulates vesicle fast recycling route to the plasma membrane)53and EHD-1 (an endocytic recycling protein and a component of the FERRARI complex)21, 22were not affected by the Br-EV treatment (FIGs. 10A-10C). These results demonstrated that Br-EVs decrease the expression of Rabi lfip2, which is involved in the actin-based recycling of vesicles to the apical plasma membrane while increasing the expression of Rabi lfip3 and Rabi lfip5, which are linked to microtubule-based vesicle recycling towards deeper cytoplasmic compartments of polarized BECs and transcytosis. These findings were the first to suggest that Br-EVs modulate intracytoplasmic proteins involved in vesicle trafficking, particularly with respect to endocytosis, recycling pathways and transcytosis.

[0222] Murine cerebral micro vessels showed decreased Rabllfip2 and increased Rabllfip3 and Rabllfip5 levels in response to Br-EV treatment.

[0223] #14794310v1To confirm the in vitro observation of Rabllfip protein modulations in vivo, the effects of Br-EVs on murine cerebral microvessels were assessed, which have BECs, pericytes and astrocyte feet46. Athymic nude mice received a total of 10 intravenous injections of either PBS or 3 pg (~3-4 x 109EVs / 100 pL) of P-EVs or Br-EVs on alternate days8, 10. This dose has been shown to be proportional to the concentration of circulating EVs in tumor-bearing mice and leads to an increased incidence of breast-to-brain metastasis in vivo&’10. Following treatment, mice were euthanized and their brain cortex was isolated and immediately processed to extract cerebral microvessels (FIGs. 11A-11D). Rabllfip2, Rabllfip3 and Rabllfip5 protein content was then analyzed by immunoblot. Br-EVs significantly upregulated Rabllfip5 compared to both P-EVs and PBS, while both P-EVs and Br-EVs increased Rabi lfip3 and decreased Rabllfip2 when compared to PBS treatments. Immunofluorescence analysis and signal quantification (FIGs. 2B-2C) showed a significantly increased expression of Rabllfip3 and Rabllfip5 in cerebral microvessels of mice treated with P-EVs and Br-EVs compared to PBS controls. In contrast, the immunofluorescence signal quantification of Rabllfip 2 showed no statistically significant difference between Br-EVs, P-EVs and PBS treatments (FIG. 2C), as opposed to the immunoblot analysis (FIG. 2A). This may be attributable to the fact that murine cerebral microvessel samples, while enriched in BECs, also included pericytes and astrocyte feet39FIGs. 11A-11D). These additional cell types may have undergone distinct protein changes in response to EV treatments, potentially masking the BEC Rabllfip2 protein downregulation. Furthermore, the absence of precise molecular weight control inherent in immunofluorescence assays represented a known technical limitation and could have contributed to the observed discrepancies, making the semi-quantitative interpretation of Rabllfips expression results more challenging.

[0224] It is worth noting that, as previously observed8, 10, 54and reported by other groups55-57, P-EVs and Br-EVs sometimes displayed similar effects compared to PBS controls. This might be because both EVs originate from cell lines sharing the same background phenotype and retain similarities in certain pathways of cancer progression and pathogenesis. Despite these similarities, Br-EVs caused distinct changes at the BBB both in vitro and in vivo, supporting their efficiency in promoting the brain metastatic process.

[0225] These in vivo findings corroborated the in vitro results and described previously unknown mechanisms through which Br-EVs can hijack the endocytic dynamics, vesicular trafficking and transcytosis pathways of the BECs of the BBB.

[0226] #14794310v1ML analysis identified morphological and motility changes induced by Br-EVs that prime BECs for metastatic cancer cell transmigration.

[0227] The BECs of the BBB form an endothelial barrier facing the bloodstream and interact with its circulating components, such as cells, EVs6-8, 10, 54and other blood-derived nanoparticles. Immune and metastatic cancer cells have the ability to transmigrate the BBB endothelium via the paracellular or transcellular route58-60. Paracellular transmigration requires the disruption of the tight junctions between BECs, whereas transcellular transmigration occurs through the cytoplasm of a single BEC58-60. In the context of breast-to-brain metastasis, invasive TNBC cells are known to preferentially use the transcellular route58, 60, 61. When TNBC cells engaged with the microvascular endothelium of the BBB, there was a significant cytoskeletal and membrane reorganization that led to BEC phenotypic alterations58. Unlike normal BECs, which are involved in maintaining BBB homeostasis and constitute a selective barrier that does not interact with circulating cells, engaged BECs play a important role in metastatic cell extravasation. In particular, they envelop and facilitate the transcellular migration of metastatic cells by extending membrane protrusions58and reorganizing their cytoskeleton to form a myosin-contracting cage61, 62, respectively. Surprisingly, it is currently unknown whether Br-EVs are capable of inducing early phenotypical changes in BECs that prime the BBB for TNBC cell adhesion and transendothelial migration63.

[0228] To address this mechanistic void, BEC morphology and motility changes were analyzed after treatment with Br-EVs using an ML approach. The ML framework was designed to extract, analyze and visualize morphology and motility features from live and fixed, labeled and unlabeled cell imaging techniques following treatment with PBS, P-EVs or Br-EVs.

[0229] Specifically, confocal (FIGs. 3A-3G) and fluorescent live-cell imaging (FIGs. 3H-3N) were utilized to assess cell morphology. Quantitative phase imaging (QPI) (FIGs. 4A-4G) and fluorescent live-cell imaging (FIGs. 4H-4N) were used to analyze cell motility. Ten and nine of the most used morphology and motility features were collected and analyzed (see FIG. 3G, FIG. 3N, FIG. 4G, and FIG. 4N panels)36, 37■66. These selected features, which can be extracted from various types of imaging techniques, are widely utilized across different analysis frameworks and provide a comprehensive understanding of changes in cell size, shape, structure, speed and directionality. For each condition, a distribution map was generated using the UMAP method38(FIGs. 3A-3C, FIGs. 3H-3J and FIGs. 4A-4C, FIGs. 4H-4J). To define BEC phenotypes, non-hierarchical spectral clustering39was used with a maximized Silhouette score40to automatically identify clusters, each representing a phenotypic subgroup characterized by #14794310v1specific morphological and motility characteristics (FIG. 3D and FIG. 3K and FIG. 4D and FIG. 4K)39’40.

[0230] This unsupervised approach identified five phenotypic subgroups of BECs when visualized by confocal microscopy (FIG. 3D), four phenotypic subgroups by fluorescent livecell imaging (FIG. 3K and FIG. 4K) and three phenotypic subgroups by QPI (FIG. 4D).

[0231] In the morphology analysis (FIGs. 3A-3N), control BECs treated with PBS exhibited a more uniform distribution across all subgroup clusters in both confocal microscopy and live-cell imaging compared to BECs treated with Br-EVs. When fixed BECs were analyzed via confocal microscopy, Br-EV treatment induced an enrichment in Clusters 3 and 4, defined by an increased Max Feret diameter67, major axis length and perimeter and decreased solidity and extent (FIGs. 3C-3G). Interestingly, these results were recapitulated when live BECs were analyzed by fluorescent live-cell imaging with significant enrichment in Cluster 1, also defined by an increased Feret diameter, major axis length and perimeter, and decreased solidity and extent (FIGs. 3J-3N). These data led to the conclusion that Br-EVs significantly increased BECs with larger size, elongated shape and irregular boundaries compared to BECs treated with PBS and P-EVs (FIGs. 3E-3G and FIGs. 3L-3N). Collectively, these results indicated that Br-EV-treated BECs display altered cell-to-cell adhesion dynamics and modifications to the cell cytoskeleton, suggesting impaired barrier function(s).

[0232] Motility analysis (FIGs. 4A-4N) demonstrated that treatment with Br-EVs significantly increased the percentage of cells characterized by fast, spread and convoluted movements with poor directionality, all of which suggest an inefficient cell migration. When live BEC were analyzed by QPI following Br-EV treatment, there was a significant enrichment of Cluster 2, defined by increased speed and decreased linearity of forward movement (FIGs. 4D-4G). When BECs were analyzed by live-cell imaging, there was a ~2-fold increase in Cluster 4 (FIGs. 4K-4N). Similar to the QPI results, Cluster 4 is also defined by increased speed and decreased linearity of forward movement. Compared to PBS and P-EVs, Br-EVs significantly increase BECs characterized by faster but random movements with poor directionality (FIGs. 4E-4G and FIGs. 4L-4N). The random motility patterns and the displayed enhanced migratory behavior of BECs treated with Br-EVs suggested changes in the cytoskeleton dynamics and the endothelial barrier integrity and permeability, respectively. Taken together, ML analysis consistently indicated that Br-EVs have a significant impact on BEC phenotype populations and promoted the selection of larger BECs with irregular boundaries and displaying fast and random movements. These Br-EV-induced morphology and motility changes in BECs were consistent

[0233] #14794310v1with the cytoskeletal reorganizations required for the transendothelial migration of TNBC cells across the BBB.

[0234] BECs with Rab7 knockdown (KD) and Rabllfip5 overexpression exhibited morphological and motility changes comparable to those induced by Br-EV treatment.

[0235] Br-EVs cause a significant decrease in Rab78and Rabi lfip2 and an increase in Rabi lfip3 and Rabi lfip5 in the BECs of the BBB. These proteins not only play a role in intracytoplasmic vesicle degradation and sorting8, 21, 22, 51, 68, 69, but also regulate the connection and transport of endocytic vesicles on cytoskeletal filaments and the maintenance of cell polarity and motility20, 34, 49, 70. Rab7 regulates cell motility by directly interacting with Rael34. Its depletion leads to the accumulation and expansion of endoplasmic reticulum (ER) membranes and decreased cell directionality70. Rabllfip2 regulates cell polarity and migration, activating PI3K / Akt signaling and increasing MMP7 expression71. Rabllfip3 has been reported to regulate cell endocytic transport and motility dynamics by regulating the polarization of the actin-based cytoskeleton and the formation of lamellipodia49. Finally, the KD of Rabi lfip5 resulted in a significant decrease in the membrane levels of pro-migratory integrin, leading to compromised cell motility and adhesion to the ECM70.

[0236] The same imaging analyses, coupled with ML, was employed to investigate whether the modulation of Rab7, Rabi lfip2, Rabi lfip3 and Rabi lfip5 contributed to the morphodynamic changes observed in BECs treated with Br-EVs (FIGs. 5A-5H). BECs knocked down (KD) for Rab7 and Rabllfip2 or overexpressing Rabllfip3 and Rabllfip5 were imaged using confocal microscopy (FIGs. 5A-5B), live cell imaging (FIGs. 5C-5D and FIGs. 5G-5H) and QPI (FIGs.

[0237] 5E-5F). Their morphology (FIGs. 5A-5D) and motility features (FIGs. 5E-5H) were extracted and analyzed. Notably, unsupervised ML revealed that BECs with Rab7 KD and Rabi lfip5 overexpression recapitulated the cell clustering, morphology and motility features observed in BECs treated with Br-EVs (FIG. 3C and FIG. 3J, and FIG. 4C and FIG. 4J). These results suggested that the phenotypic changes induced by Br-EVs are significantly influenced by the observed protein modulations, particularly the decrease in Rab7 and Rabi 1 fip2 (the latter being evident only in morphological changes detected via confocal microscopy) and the increase in Rabllfip5.

[0238] Br-EVs modulated cytoplasmic and cytoskeletal dynamics in murine cerebral microvessels.

[0239] #14794310v1The normal cerebral vasculature is characterized by high selectivity and low permeability and transcytosis rates thereby ensuring proper brain physiology6’8, 10, 60, 72. To elucidate the mechanisms by which Br-EVs induce cerebrovascular dysfunction and promote breast- to-brain metastasis, a quantitative global proteomics approach was utilized. This method allowed the ability to identify changes in protein expression and to identify and assess biological functions in murine cerebral micro vessels modulated by Br-EVs in a breast-to-brain metastasis context41, 43,73. Label-free mass spectrometry-based quantitative proteomics were employed to analyze the protein content of cerebral micro vessels from the brains of mice treated with PBS, P-EVs or Br-EVs. A total of 3030, 3213 and 2908 proteins were detected in the micro vessel samples isolated from PBS, P-EVs and Br-EV-treated mice, respectively (FIGs. 9D-9F). 795 proteins were shared among all groups (FIG. 9G). The comparative heatmap (FIG. 6A) and the principal component analysis (PC A) (FIG. 9H) illustrate how PBS, P-EVs and Br-EV sample datasets correlated. The negative correlation values in the comparison between Br-EV and PBS samples (FIG. 6A), along with the distinct clustering highlighted by the PCA (FIG. 9H), indicated that Br-EVs cause significant changes in cerebral micro vessel protein expression profiles. In contrast, the comparison between P-EV and PBS samples resulted in correlation values that, in most cases, are equal to or close to zero, indicating a less pronounced variation in protein expression at the BBB (FIG. 6A). 163 proteins that were differentially expressed were identified, both at a significant level and with high confidence, in the cerebral microvessels of mice treated with Br-EVs compared to PBS. Additionally, 85 proteins were differentially expressed in cerebral micro vessels from mice treated with P-EVs compared to the PBS group (FIG. 6B and FIGs. 13A-13C). Compared to the PBS group, Br-EVs caused a significant increase in 35 proteins representing several functional categories relevant to the context of cancer-derived EVs breaching the cerebral microvasculature and compromising BBB function and homeostasis8, 10(FIGs. 6B-6C). These included, among others, vesicle endocytosis and trafficking, tight junction formation, cell volume and motility (Table 1). Similarly, Br-EVs significantly downregulated 129 proteins (FIG. 6B and FIG. 6D) representative of other functional categories such as, but not limited to, clathrin-based endocytosis, vesicle sorting and trafficking, exocytosis, ER maturation, ER-Golgi transport and lysosomal function (Table 1).

[0240] The proteins differentially expressed in Br-EV treated samples compared to PBS controls predicted the decreased or increased activation of 23 cellular functions and diseases. Among the top-downregulated cellular functions were decreased cytoplasm and cytoskeleton organization, microtubule dynamics and formation of cell protrusions (FIG. 6E). These were associated with #14794310v1reduced cell adhesion and persistent, directional migration in polarized cells74-76, therefore validating the in vitro ML results.

[0241] Table 1. Upregulated and downregulated proteins in murine cerebral microvessel samples treated with Br-EVs compared to PBS and their respective functional categories relevant to EV- BBB interactions and / or BBB functions and homeostasis.

[0242]

[0243] #14794310v1Among the proteins exclusively upregulated in murine cerebral microvessels from mice treated with Br-EVs, was NKCC1. NKCC1 is a cation-chloride cotransporter expressed in BECs, glia and neurons26, 77, co-localizes with Rabi 1+ compartments during recycling26, 78and is known to regulate cell mass, volume, cytoskeletal organization32, 77and BBB integrity29, 30. NKCC1 upregulation in BECs treated with Br-EVs compared with PBS was further validated in vitro by immunoblot, while P-EVs did not cause statistically significant results (FIGs. 6F-6G).

[0244] Taken together, the upregulation of NKCC1 and the other Rabi Ifip protein modulations were consistent with the conclusion that Br-EVs affect the cytoplasmic and cytoskeletal dynamics of BECs. NKCC1 inhibitors are already in use in the clinic to treat neurological symptoms of Down syndrome and other brain disorders25, 26, 77-81and may represent a novel approach to restoring the integrity of the BBB which in turn could prevent the formation of the brain pre-metastatic niche and the subsequent onset of brain metastasis.

[0245] Brains of mice treated with Br-EVs showed increased CD31 expression.

[0246] Early metastatic lesions in the brain, typically smaller than 0.25 mm in diameter, are viable within 100 pm away from blood vessels and exhibit a characteristic perivascular growth pattern, surrounded by an intact BBB98. Immunohistochemical and morphometric analyses showed that experimental brain metastases contained fewer dilated blood vessels composed of CD31 -overexpressing BECs compared to adjacent tumor-free brain tissue98. CD31 is important for maintaining BBB integrity as changes in its expression or function can influence BBB permeability and facilitate metastatic cell migration into brain tissue99-103by playing a direct role in tumor cell adhesion, growth and survival100, 101. The findings indicated that Br-EVs can modulate the endocytic and cytoskeletal dynamics of BECs (FIGs. 1A-5H) and promoted the adhesion of TNBC cells to BECs compared to controls (FIG. 7A). It was hypothesized that Br-EV-induced changes in the cerebral microvasculature could lead to variations in brain microenvironment structure and composition. To test this hypothesis, whole brain sections of mice treated with Br-EVs, P-EVs and PBS were stained for microtubule-associated protein (MAP)2, CD31, glial fibrillary acidic protein (GFAP), CD45 and CD 11b to detect mature neurons, BECs, astrocytes and microglia or peripheral infiltrating macrophages, respectively. Notably, compared to mice administered with PBS and P-EVs, the brain sections of mice treated with Br-EVs showed areas with increased CD31 staining, potentially indicating priming of the cerebral vasculature to future cancer cell adhesion and transendothelial migration (FIG. 7B and FIGs. 14-16). The cerebral microvessel samples isolated from the brains of mice treated with #14794310v1both P-EVs and Br-EVs showed a CD31 immunofluorescence signal significantly higher than that of PBS controls (FIGs. 7C-7D). These results suggest that P-EVs also have the potential to increase CD31 protein expression in brain microvessels despite resulting in different macroscopic expression patterns compared to Br-EV treatments. No significant changes were detected in the expression patterns of neuronal, astrocyte and glial biomarkers nor in the whole brain signals for CD31, GFAP and MAP2 (FIGs. 17A-23). These data suggest that Br-EVs specifically prime areas of the brain microvascular endothelium, which could facilitate the transendothelial migration of circulating tumor cells into the brain parenchyma.

[0247] Discussion

[0248] TNBC is the BC subtype with the highest risk of developing brain metastasis within 5 years of the diagnosis, with 25-46% of metastatic TNBC patients presenting brain lesions1, 4. The major barrier to anticancer and adjuvant treatments for breast-to-brain metastasis is represented by the intact and highly selective BBB surrounding both the brain pre-metastatic niche and the micrometastases98. Br-EVs breach the intact BBB and prepare the pre-metastatic niche8, 10. A deeper understanding of how Br-EVs affect BEC phenotype, particularly their cytoplasmic and cytoskeletal dynamics, in order to breach the intact BBB8is important in order to develop novel therapeutic strategies to target the brain and effectively treat breast-to-brain metastasis formation from its early stages.

[0249] These findings demonstrate that Br-EVs not only promoted their transcytosis across the BBB8and induced remodeling of the brain's ECM10but also impacted BEC cytoplasm and cytoskeleton dynamics both in vitro and in vivo. Specifically, this study described, for the first time, both broad and distinct modulations that Br-EVs can be induced in the cerebrovascular proteome and identified four proteins, Rabi lfip2, Rabi lfip3, Rabi lfip5 and NKCC1 that were linked to Br-EV effects on BECs. Additionally, it was demonstrated that Br-EVs induced significantly increased TNBC cell adhesion to the BECs and increased CD31 staining at the BBB, suggested a BEC phenotype that could favor the transendothelial migration of TNBC cells across the BBB.

[0250] Taken together, these findings identified novel EV-mediated mechanisms through which Br-EVs primed the BBB to favor the adhesion and metastasis of TNBC cells, as well as proteins that represented potential novel therapeutic targets for the treatment of breast-to-brain metastasis. Therapeutic strategies targeting Rabllfip3, Rabllfip5 andNKCCl or reestablishing Rab7 and Rabi lfip2 protein levels in BECs could be used to restore BBB integrity and prevent #14794310v1the formation of the brain premetastatic niche and subsequent brain lesions. Furthermore, the development of new EV-inspired nanoparticles targeting the areas of BBB that overexpress CD31 may represent a novel therapeutic strategy to prevent TNBC cell adhesion and transendothelial migration.

[0251] Materials and Methods

[0252] Cell Lines and Cell Culture. The human triple-negative breast cancer (TNBC) cell line MDA-MB-231 was purchased from the American Type Culture Collection (ATCC HTB-26, VA, USA). The brain-seeking (MDA-231Br) variant of the TNBC cell line MDA-MB-231 was a gift from Dr. T. Yoneda, Indiana University, USA9. Primary human brain microvascular endothelial cells (BECs) were purchased from Cell Systems Co. (cat. no. ACBRI 376, Kirkland, WA). TNBC cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, cat. no. 11885084, Thermo Fisher Scientific® Inc.) supplemented with 1% penicillin-streptomycin (10000 U / mL) (cat. no. 15140148, Thermo Fisher Scientific® Inc.) and 10% fetal bovine serum (FBS, cat. no. SI 1150, Atlanta Biologicals, Atlanta, GA, USA). TNBC cells were seeded in 150 mm dishes and cultured for 48 hours in DMEM supplemented with 10% EV-depleted FBS to produce conditioned media for isolating extracellular vesicles (EVs). EV-depleted FBS was prepared by 16 hours ultracentrifugation 100,000 g at 4°C. BECs were cultured with complete classic medium with serum and culture boost (cat. no. 4Z0-500, Cell System, WA, USA). All cells were used from passages 4-20, maintained in a 37°C humidified incubator with 5% CO2. All cultures were regularly screened for mycoplasma contamination using the Myco Alert PLUS Mycoplasma Detection Kit (cat. no. LT07-710, Lonza®, Inc.) according to the manufacturer’s instructions.

[0253] EV Isolation and Characterization. Conditioned media was collected from MDA-MB-231 and MDA-231Br cells after 48 hours of incubation in EV-free media. Conditioned media was used for EV isolation only if cell viability was >95%, as determined by Trypan Blue. EVs were isolated using a differential centrifugation protocol. Briefly, conditioned media was collected and centrifuged at 400g for 10 minutes, 2,000g for 20 minutes and 14,000g for 45 minutes at 4°C (Sorvall Lynx 4000, Thermo Fisher Scientific® Inc.) to remove dead cells and large debris, apoptotic bodies and larger microvesicles, respectively. As previously described by us, the resulting supernatant subsequently underwent ultracentrifugation at 100,000g for 90 minutes at 4°C (Optima XE-90 Ultracentrifuge, Beckman Coulter Life Sciences). The first pellet was

[0254] #14794310v1washed in PBS and precipitated by repeating the UC step at 100,000g for another 90 minutes. The washed pellet was resuspended in sterile cell-grade PBS (100-200 pL) for characterization and experiments or stored at -80°C. EV preparations were characterized according to the latest Guidelines of the International Society for Extracellular Vesicles (FIGs. 8A-8D)35. EV total protein concentration was measured by Bradford assay (Bio-Rad® Protein Assay Dye Reagent Concentrate, cat. no. 5000006, Bio-Rad®, CA, USA) according to the manufacturer’s instructions. Bradford assay was performed in a clear 96-well plate and 590 nm absorbance was measured using FilterMax F3 multi-mode microplate reader (Molecular Devices®, CA, USA). EV size distribution and particle number / mL of solution were measured by nanoparticle tracking analysis (NTA, NanoSight NS300, Malvern Instruments, UK). EV formulations were diluted (1:500-1:1000) in sterile phosphate-buffered saline (PBS) and analyzed (3 mL) with a Malvern Panalytical Nanosight NS300 (60-second measurement; 3 capture replicates). The presence of EV markers cluster of differentiation (CD)63, CD73 and annexin A2 and the absence of calnexin, a marker of the endoplasmic reticulum, as a negative control, was evaluated by Immunoblot (5-20 pg of total proteins / lane). EV morphology was evaluated by transmission electron microscopy (TEM) as previously described8. Imaged by TEM and analyzed by nanoparticle tracking analysis, P-EVs, ad Br-EVs showed characteristic disc-shaped morphology and heterogeneous size distribution in the nanometer range with a mode of 166 nm for P-EVs and 157 nm for Br-EVs (FIGs. 8A-8B). P-EVs and Br-EVs had a total protein content of 4-10 pg / 1010EVs and were enriched in typical EV markers, such as CD63, CD73 and annexin A2. Both EV samples lacked calnexin, an endoplasmic reticulum marker used to detect the possible presence of nanosized intracytoplasmic contaminants (FIGs. 8C-8D).

[0255] In vitro EV Treatments. To evaluate the effect of EVs on BEC intracytoplasmic dynamics, BECs grown to confluence in 6-well plates were treated daily for a total of 5 days with a fixed quantity of EVs (103EVs / cell). After, cells were washed 3 times with sterile cell-grade PBS and lysed using lysis buffer IX in sterile water (Cell Signaling Technology®, Danvers, MA) supplemented with phenylmethylsulphonyl fluoride protease inhibitor incubated for 5 minutes on ice. Cell lysates were collected with a scraper, quickly tip sonicated and centrifuged at 14,000g for 10 minutes at 4°C, the supernatant was collected, total proteins were quantified by Bradford analysis and analyzed by Western blot (20-30 pg of total protein / lane).

[0256] Immunoblot Analyses. Immunoblotting was performed following the protocol previously developed8. Antibodies targeting the following proteins were utilized for immunoblotting: CD63 #14794310v1(1:500, Abeam®, cat. no. 59479), CD73 (1:500, Cell Signaling Technologies®, cat. no. 13160), annexin A2 (1:500, Novus Biologicals®, cat. no. NBP1-31310), calnexin (1:1000, GeneTex® Inc., cat. no. GTX112886), rab 7 (1:1000, Abeam®, cat. no. 137029), rabllfip2 (1:500, Abeam®, cat no. abl80504), rabllfip3 (1:500, Proteintech®, cat. no. 25843-1-AP), rabllfip5 (1:500, Proteintech®, cat. no. 14594- 1-AP), myosin5b (1:500, Lifespan Biosciences®, cat. no. LS-C410093), ehd-1 (1:500, Proteintech®, cat. no. 24657-1-AP), rab4 (1:500, Lifespan Biosciences®, cat. no. LS-B8414), ZO-1 (1:500, Thermo Fisher Scientific®, cat. no. 40-2200), GFAP (1:500, Abeam®, cat. no. 7260), NKCC1, SLC12A2 (1:500, Proteintech®, cat. no. 13884-1-AP) and GAPDH (1:3000, Santa Cruz Biotechnology®, cat. no. G-9, sc-365062).

[0257] CRISPR / Cas9 Cell Transfection. 80% confluent BECs in a 6-well plate were transfected using Lipofectamine CRISPRMAX Cas9 Transfection reagent (Thermo Fisher Scientific®) using 7.5 pg of TrueCut Cas9 Protein v2 (Thermo Fisher Scientific®), 1 pL of TrueGuide Synthetic Rabllfip2 (CRISPR989610_SGM) andRab7 (CRISPR464628_SGM) gRNA (100 pM), 15 pL of Cas9 Plus Reagent and 7.5 pL of CRISPRMAX Reagent. After 48 hours, cells were used in imaging experiments.

[0258] Cell Electroporation. 0.5xl06BECs were harvested, pelleted at 400 g for 10 minutes and electroporated using Cell Line Nucleofector Kit V (Lonza® Walkersville, USA). The cell pellet was resuspended in 100 pL of buffer, mixed with 2 pg of Rabi lfip3 (human tagged orf clone rc227246, Origene Technologies® Inc.) and Rabllfip5 (human tagged orf clone rc206173, Origene Technologies® Inc) plasmids and electroporated using the S-05 Program of the Amaxa® Biosystem Nucleofector. After, cells were transferred into a 6-well plate previously filled with 2 mL of media warmed at 37 °C and cultured for 24 hours before being used in imaging experiments.

[0259] Quantitative Phase Imaging. BECs were seeded in a 6-well plate and treated with a fixed quantity of EVs (103EVs / cell) daily for 5 days. BECs were treated with CRISPRMAX / Cas9 or electroporated 48 and 24 hours before the experiments, respectively. Cells were imaged live for 24 hours at 37°C using the quantitative phase-contrast microscope HoloMonitor (PhiOptics). Three different fields per sample were captured at 10-minute intervals.

[0260] Live Fluorescence Imaging. BECs were treated with a fixed quantity of EVs (103EVs / cell) daily for 5 days. BECs were treated with CRISPRMAX / Cas9 and electroporated 48 and 24 #14794310v1hours before the experiments, respectively. 24 hours prior to the live imaging, cells were seeded in a 24- well plate (2xl04cells / well). Cells were stained overnight at 37 °C with Dil lipophilic fluorescent staining. Fluorescence images were acquired using a Nikon® Ti2-E inverted perfect focus microscope equipped with a microscope cage incubator from OKOLabs, which is capable of controlling 37°C and 5% CO2 during live cell imaging. The time series images were acquired every 10 minutes for duplicated positions within each well (24 wells) for 48 hours with bright field and Cy3 (cytoplasmic labeling) channels. The raw Images were obtained with NIS-Elements Advanced Research Package for 6D Image Acquisition software (Nikon®). The ND2 formatted raw images were converted to tiff and movie files for ML-based analysis using NIS-Elements Advanced Research Package for 6D Image Acquisition software (Nikon®) or Fiji (National Institutes of Health, Bethesda, MD).

[0261] Machine Learning Analysis. The proposed ML-based framework deals with confocal, quantitative phase imaging (qPI) and live cell imaging data to analyze morphodynamical features for specific cellular phenotypes. First, cell segmentation outcomes were generated by using CellPose36(cellpose.org / ), a deep learning-based algorithm for cell image segmentation, for confocal images and adaptive image thresholding approach34for qPI and live cell images. Then morphology features were obtained from these outcomes and TrackMate computational tools37were utilized to extract motility features from live cell imaging data. There are 10 morphology features and 9 motility features being collected to generate feature distribution maps by UMAP method38. To clearly represent the differentiation of BECs’ phenotypes, spectral clustering39with Silhouette score40was utilized to automatically identify potential subgroups which have similar morphological or motility characteristics.

[0262] In-solution digestion and solid-phase extraction with Stage Tips. Isolated microvessels from the blood-brain barrier were suspended in 100 pL of protein solubilization buffer (1% w / v SDC, 10 mM TCEP, 40 mM CAA and 50 mM TEAB) and subjected to a 10-minute heat treatment at 70°C. This step was performed to ensure complete protein solubilization, denaturation, reduction of cysteine residues and alkylation of cysteine residues. Subsequently, the samples were centrifuged at 21,000 x g for 10 minutes and the resulting supernatants were transferred to fresh Eppendorf tubes. Protein concentrations were quantified using the PierceTM 660 nm Protein Assay Reagent (Thermo Fisher Scientific®, Waltham, MA, USA), with a standard curve generated using a dilution series of BSA.

[0263] #14794310v1For the preparation of extracts for tryptic digestion, 10 pg of protein from each sample was transferred to a new Eppendorf tube and brought to a total volume of 50 pL with protein solubilization buffer. Following this, 1 pg of trypsin in 50 pL of 50 mM TEAB was added to each sample and the digestions were allowed to incubate overnight at 37°C in an incubator. The following day, an additional 1 pg of trypsin in 1 pL of 50 mM TEAB per sample was added and the digestions were incubated for an additional 4 hours at 37°C. The resulting peptides were purified using SDB-RPS stage tips, following the protocol developed by Rappsilber et al41.

[0264] Measurement of LC-MS data in DIA mode. The LC-MS / MS data acquisition process was conducted utilizing a Vanquish Neo nanoLC system equipped with an Orbitrap Eclipse mass spectrometer, a FAIMS Pro Interface and an Easy Spray ESI source, all supplied by Thermo Fisher Scientific® (Waltham, MA, USA).

[0265] For nanoLC separation, an Acclaim PepMap trap column (75pm x 2 cm) in conjunction with an EasySpray ES802 column (75 pm x 25 mm, 100 A) from Thermo Fisher Scientific® were employed. A 3 pL peptide extract was introduced into the system. The separation of peptides was achieved using a mobile phase consisting of 0.1% (v / v) formic acid in water (solvent A) and 0.1% (v / v) formic acid in 80% (v / v) acetonitrile (solvent B), with a flow rate of 300 nL / minute and a column temperature maintained at 40°C. The column was initially conditioned for two minutes with 3% solvent B, followed by a gradual linear gradient up to 40% solvent B over a 60-minute period. Any remaining peptides bound to the C18 resin were subsequently eluted with 95% solvent B for 11 minutes. In the positive ion mode, the ion source temperature was set to 305°C and 2000 V and ionized peptides were routed through the FAIMS Pro unit at -50 V. Mass spectra were acquired with a resolution of 120,000 in MSI mode over the mass range of m / z 350-2000, with standard automatic gain control (AGC) settings and automatic injection time. For MS / MS fragmentation, a data-independent acquisition (DIA) mode was employed within the mass range of 375-1200 at a resolution of 30,000. The collision energy was set at 30% and an AGC target of 1000% was applied, utilizing m / z 25 isolation windows with 0.5 m / z overlaps for the collection of MS2 spectra.

[0266] Data analysis with DIA-NN and statistics. The proteome data analysis was carried out with the DIA-NN 1.8.1 software package42. The data from the LC-MS data containers was processed and analyzed using the following settings: FASTA-database: Mus musculus (UP000000589);

[0267] FASTA digest for library-free search / library generation; Protease: Trypsin / P; missed cleavages: 1, N-term M excision; C carbamidomethylation; Peptide length range: 7-30; Precursor charge #14794310v1range: 1-4; precursor m / z range: 300-1800; fragment ion m / z range: 200-1800; Precursor FDR (%): 1.0; Use isotopologues; heuristic protein interference; no shared spectra; Protein interference: Genes; Neural network classifier: Single-pass mode; Quantitation strategy: Robust LC (high precision); Cross-run normalization: RT- dependents; Library generation: Smart profiling; Speed and RAM usage: optimal results. The raw protein peak areas were used for statistical analysis following the descriptions by Schulte et al43. Comparative analysis was conducted on proteins showing a Log2 fold-change absolute value greater than 0.5 and an adjusted p-value equal or less than 0.0544, 45(FIGs. 9A-9C).

[0268] MDA-MB-231 adhesion assay. BECs were seeded onto 6-well plates, cultured into monolayers and then treated for 5 days with a fixed quantity of Br-EVs (103EVs / cell) or the corresponding volume of sterile cell-grade PBS. Following the EV treatments, 5xl04cells MDA-MB-231 cells stably expressing CD63-TdTomato fusion protein were incubated for 1 hour at 37°C. Cells were washed three times with PBS, stained with DAPI, fixed in 4%PFA and imaged by fluorescent microscopy. MDA-MB-231 red fluorescent cells adhering to the BECs monolayers were counted at 10X magnification (5-6 fields / well, n=3 wells / treatment).

[0269] In Vivo Experiments. All in vivo experiments were conducted in compliance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) at Boston Children’s Hospital, Boston, MA. For all animal experiments, 6-8-week-old female athymic nude mice (Nu / Nu) were purchased from Massachusetts General Hospital. The mice were allowed to acclimate for at least four days before the start of the experiments. The minimum number of animals required to obtain data amenable to statistical analysis was used for animal experiments. For the brain pre-metastatic niche studies, mice were randomly divided into 3 groups to receive retro-orbital injections of EVs derived from parental and brain-seeking MDA-MB-231 cells (3 pg of EVs in 100 pL of PBS per injection) or 100 pL of PBS. Injections were conducted on alternate days, switching between the right and left eye, for a total of 10 injections. Twenty-four hours after the last EV injection, mice were sacrificed by CO2 or cardiac perfusion under anesthesia with 50 mL of 1% formaldehyde + 0.5% Methanol. Perfused brain tissues were collected, fixed for 1 hour at RT in 1% formaldehyde, then preserved at +4°C for 24-48 hours in a sterile 30% sucrose PBS solution until completely sunk. Brain tissues in 30% sucrose solution were then sent to iHisto (Salem, MA, USA) for cryo-sectioning. Fresh brain tissues were immediately processed for the ex vivo isolation of cerebral microvessels from the brain cortex, as described here46. BBB microvessels were isolated and stored at -80°C in the form of a dry #14794310v1pellet until further analysis. For histological analysis, each brain was sectioned whole through sequential coronal cuts going from the frontal cortex to the cerebellum, 15-25 pm thick slices were prepared for a total of 15 slides for each brain.

[0270] Immunocytochemistry and Immunohistochemistry. For immunocytochemistry, cells and brain microvessels samples were fixed with 4% paraformaldehyde + 4% sucrose in PBS for 14 minutes, rinsed thrice in PBS and then treated with 0.2% triton X-100 for 5 minutes for permeabilization. For immunohistochemical staining, frozen sections were fixed with ice-cold acetone for 7 minutes and washed three times in PBS. Cells and microvessel blocking was performed using 5% bovine serum albumin for 1 hour at RT, whereas brain tissue section blocking was using 10% normal goat serum NGS + 0.3% Triton X-100 in TBS for 1 hour at RT. Cells or tissue sections were incubated with the primary antibody (1:100) in 1% BSA for 1 hour at room temperature or overnight at 4°C. After the washes, cells or tissue sections were incubated with the relevant secondary antibody (1:1000) in 1% BSA for 45 minutes. Cells were washed with sterile PBS and stained with DAPI (cat. no. S33025, Thermo Fisher Scientific®, USA). Tissue sections were washed with sterile PBS and mounted with Vectashield Plus antifade mounting medium with DAPI (cat. no. H-2000-10, Vector Lab Inc., CA, USA). For immunocytochemistry and / or histochemistry the following antibodies were used: Rabllfip2 (1:500, Abeam®, cat no. abl80504), Rabllfip3 (1:500, Proteintech®, cat. no. 25843-1-AP), Rabllfip5 (1:500, Proteintech®, cat. no. 14594-1-AP), CD31 (1:100, GeneTex®, cat no.

[0271] GTX130274), MAP2 (1:100, Abeam®, cat. no. ab5392), GFAP (1:100, Novus Biologicals®, cat. no. NBP1-05198), CDllb (1:100, Abeam®, cat. no. abl33357), CD45 (1:100, Abeam®, cat. no. ab23910), AlexaFluor488 goat anti-rabbit (1:1000, Invitrogen, cat no. A11008), goat pAb to chicken IgY Alexa fluor 647 (1:1000, Abeam®, cat no. abl50171), donkey pAb to rat IgG Alexa fluor 647 (1:1000, Abeam®, cat no. abl50155). Images were taken using a Zeiss® LSM 880 Confocal Microscope. Images were processed using ImageJ. Fluorescent signals were normalized with the background signal of each image and quantified as described here. Sections 2 and 14th of each brain were stained with hematoxylin and eosin (H&E) staining; the absence of edema and inflammation was confirmed in H&E brain sections in a blinded manner.

[0272] Statistical Analyses. Statistical analyses were performed using GraphPad Prism® software, Excel, R and QIAGEN Ingenuity Pathway Analysis. Statistical significance was considered at P values lower than 0.033 for GraphPad Prism® and lower than 0.05 for Excel, R and Ingenuity Pathway Analysis. Outliers were excluded when performing a comparative analysis of

[0273] #14794310v1quantitative proteomic results. The Kruskal-Wallis test with Dunn’s multiple comparisons test was used to assess differences among groups for experiments with n=10 or less and when data normality could not be determined47. The one-way ANOVA with Tukey's multiple comparison test was used for experiments with n > 10 and when normality could be determined48. In vivo experiments were assessed using the Mann-Whitney test47, 48. The respective figure legends explain the specific statistical analysis methods for each figure.

[0274] References

[0275] 1. Bryan S, Witzel I, Borgmann K, Oliveira-Ferrer L. Molecular Mechanisms Associated with Brain Metastases in HER2-Positive and Triple Negative Breast Cancers. Cancers (Basel).

[0276] 2021;13(16). Epub 20210817. doi: 10.3390 / cancersl3164137. PubMed PMID: 34439289; PMCID: PMC8392331.

[0277] 2. Kadamkulam Syriac A, Nandu NS, Leone JP. Central Nervous System Metastases from Triple-Negative Breast Cancer: Current Treatments and Future Prospective. Breast Cancer (Dove Med Press). 2022;14:1-13. Epub 20220110. doi: 10.2147 / BCTT.S274514. PubMed PMID: 35046721; PMCID: PMC8760391.

[0278] 3. Lv Y, Ma X, Du Y, Feng J. Understanding Patterns of Brain Metastasis in TripleNegative Breast Cancer and Exploring Potential Therapeutic Targets. Onco Targets Ther. 2021;14:589-607. Epub 20210122. doi: 10.2147 / OTT.S293685. PubMedPMID: 33519208; PMCID: PMC7837592.

[0279] 4. Simsek M, Aliyev A, Baydas T, Besiroglu M, Demir T, Shbair AT, Seker M, Turk HM. Breast Cancer Patients with Brain Metastases: A Cross-Sectional Study. Breast J.

[0280] 2022;2022:5763810. Epub 20220819. doi: 10.1155 / 2022 / 5763810. PubMed PMID: 36051468; PMCID: PMC9417791.

[0281] 5. Nam BH, Kim SY, Han HS, Kwon Y, Lee KS, Kim TH, Ro J. Breast cancer subtypes and survival in patients with brain metastases. Breast Cancer Res. 2008;10(l):R20. Epub 20080228. doi: 10.1186 / bcrl870. PubMed PMID: 18307763; PMCID: PMC2374976.

[0282] 6. Busatto S, Morad G, Guo P, Moses MA. The role of extracellular vesicles in the physiological and pathological regulation of the blood-brain barrier. FASEB Bioadv.

[0283] 2021;3(9):665-75. Epub 20210723. doi: 10.1096 / fba.2021-00045. PubMed PMID: 34485835; PMCID: PMC8409556.

[0284] 7. Morad G, Moses MA. Brainwashed by extracellular vesicles: the role of extracellular vesicles in primary and metastatic brain tumour microenvironment. J Extracell Vesicles.

[0285] #14794310v12019;8(l): 1627164. Epub 20190627. doi: 10.1080 / 20013078.2019.1627164. PubMed PMID: 31275532; PMCID: PMC6598504.

[0286] 8. Morad G, Carman CV, Hagedorn EJ, Perlin JR, Zon LI, Mustafaoglu N, Park TE, Ingber DE, Daisy CC, Moses MA. Tumor-Derived Extracellular Vesicles Breach the Intact Blood-Brain Barrier via Transcytosis. ACS Nano. 2019; 13(12): 13853-65. Epub 20190910. doi:

[0287] 10.1021 / acsnano.9b04397. PubMed PMID: 31479239; PMCID: PMC7169949.

[0288] 9. Yoneda T, Williams PJ, Hiraga T, Niewolna M, Nishimura R. A bone-seeking clone exhibits different biological properties from the MDA-MB-231 parental human breast cancer cells and a brain-seeking clone in vivo and in vitro. J Bone Miner Res. 2001;16(8):1486-95. doi: 10.1359 / jbmr.2001.16.8.1486. PubMed PMID: 11499871.

[0289] 10. Morad G, Daisy CC, Otu HH, Libermann TA, Dillon ST, Moses MA. Cdc42-Dependent Transfer of mir301 from Breast Cancer-Derived Extracellular Vesicles Regulates the Matrix Modulating Ability of Astrocytes at the Blood-Brain Barrier. Int J Mol Sci. 2020;21(l 1). Epub 20200528. doi: 10.3390 / ijms21113851. PubMed PMID: 32481745; PMCID: PMC7311991. 11. Fernandez CA, Butterfield C, Jackson G, Moses MA. Structural and functional uncoupling of the enzymatic and angiogenic inhibitory activities of tissue inhibitor of metalloproteinase-2 (TIMP-2): loop 6 is a novel angiogenesis inhibitor. J Biol Chem.

[0290] 2003;278(42):40989-95. Epub 20030804. doi: 10.1074 / jbc.M306176200. PubMed PMID:

[0291] 12900406.

[0292] 12. Fernandez CA, Roy R, Lee S, Yang J, Panigrahy D, Van Vliet KJ, Moses MA. The anti-angiogenic peptide, loop 6, binds insulin-like growth factor- 1 receptor. J Biol Chem.

[0293] 2010;285(53):41886-95. Epub 20101012. doi: 10.1074 / jbc.M110.166439. PubMedPMID: 20940305; PMCID: PMC3009916.

[0294] 13. Roy R, Morad G, Jedinak A, Moses MA. Metalloproteinases and their roles in human cancer. AnatRec (Hoboken). 2020;303(6): 1557-72. Epub 20190625. doi: 10.1002 / ar.24188. PubMed PMID: 31168956.

[0295] 14. Roy R, Yang J, Moses MA. Matrix metalloproteinases as novel biomarkers and potential therapeutic targets in human cancer. J Clin Oncol. 2009;27(31):5287-97. Epub 20090908. doi: 10.1200 / JC0.2009.23.5556. PubMedPMID: 19738110; PMCID: PMC2773480.

[0296] 15. Peeney D, Jensen SM, Castro NP, Kumar S, Noonan S, Handler C, Kuznetsov A, Shih J, Tran AD, Salomon DS, Stetler- Stevenson WG. TIMP-2 suppresses tumor growth and metastasis in murine model of triple-negative breast cancer. Carcinogenesis. 2020;41(3):313-25. doi:

[0297] 10.1093 / carcin / bgzl72. PubMed PMID: 31621840; PMCID: PMC7221506.

[0298] #14794310v116. Wang W, Li D, Xiang L, Lv M, Tao L, Ni T, Deng J, Gu X, Masatara S, Liu Y, Zhou Y. TIMP-2 inhibits metastasis and predicts prognosis of colorectal cancer via regulating MMP-9. Cell Adh Migr. 2019; 13(l):273-84. doi: 10.1080 / 19336918.2019.1639303. PubMed PMID: 31293204; PMCID: PMC6629184.

[0299] 17. Wang H, Pan J, Barsky L, Jacob JC, Zheng Y, Gao C, Wang S, Zhu W, Sun H, Lu L, Jia H, Zhao Y, Bruns C, Vago R, Dong Q, Qin L. Characteristics of pre-metastatic niche: the landscape of molecular and cellular pathways. Mol Biomed. 2021 ;2(1):3. Epub 20210130. doi: 10.1186 / s43556-020-00022-z. PubMed PMID: 35006432; PMCID: PMC8607426.

[0300] 18. de Visser KE, Joyce JA. The evolving tumor microenvironment: From cancer initiation to metastatic outgrowth. Cancer Cell. 2023;41(3):374-403. doi: 10.1016 / j.ccell.2023.02.016. PubMed PMID: 36917948.

[0301] 19. Ai E, Skop AR. Endosomal recycling regulation during cytokinesis. Commun Integr Biol. 2009;2(5):444-7. doi: 10.4161 / cib.2.5.8931. PubMed PMID: 19907714; PMCID:

[0302] PMC2775247.

[0303] 20. Machesky LM. Rabi 1FIP proteins link endocytic recycling vesicles for cytoskeletal transport and tethering. Biosci Rep. 2019;39(l). Epub 20190130. doi: 10.1042 / BSR20182219. PubMed PMID: 30622149; PMCID: PMC6356010.

[0304] 21. Solinger JA, Rashid HO, Prescianotto-Baschong C, Spang A. FERARI is required for Rabi 1 -dependent endocytic recycling. Nat Cell Biol. 2020;22(2):213-24. Epub 20200127. doi: 10.1038 / S41556-019-0456-5. PubMed PMID: 31988382.

[0305] 22. Solinger JA, Rashid HO, Spang A. FERARI and cargo adaptors coordinate cargo flow through sorting endosomes. Nat Commun. 2022;13(l):4620. Epub 20220808. doi:

[0306] 10.1038 / s41467-022-32377-y. PubMed PMID: 35941155; PMCID: PMC9359993.

[0307] 23. Schafer JC, Baetz NW, Lapierre LA, McRae RE, Roland JT, Goldenring JR. Rabi 1-FIP2 interaction with MYO5B regulates movement of Rabi la-containing recycling vesicles. Traffic. 2014;15(3):292-308. Epub 20140122. doi: 10.111 l / tra.12146. PubMed PMID:

[0308] 24372966; PMCID: PMC4081500.

[0309] 24. Fielding AB, Schonteich E, Matheson J, Wilson G, Yu X, Hickson GR, Srivastava S, Baldwin SA, Prekeris R, Gould GW. Rabi 1-FIP3 and FIP4 interact with Arf6 and the exocyst to control membrane traffic in cytokinesis. EMBO J. 2005;24(19):3389-99. Epub 20050908. doi: 10.1038 / sj.emboj.7600803. PubMed PMID: 16148947; PMCID: PMC1276165.

[0310] 25. Lam P, Newland J, Faull RLM, Kwakowsky A. Cation-Chloride Cotransporters KCC2 and NKCC1 as Therapeutic Targets in Neurological and Neuropsychiatric Disorders. Molecules.

[0311] #14794310v12023;28(3). Epub 20230131. doi: 10.3390 / molecules28031344. PubMed PMID: 36771011; PMCID: PMC9920462.

[0312] 26. Loscher W, Puskarjov M, Kaila K. Cation-chloride cotransporters NKCC1 and KCC2 as potential targets for novel antiepileptic and antiepileptogenic treatments. Neuropharmacology.

[0313] 2013;69:62-74. Epub 20120615. doi: 10.1016 / j.neuropharm.2012.05.045. PubMedPMID:

[0314] 22705273.

[0315] 27. Demian WL, Persaud A, Jiang C, Coyaud E, Liu S, Kapus A, Kafri R, Raught B, Rotin D. The Ion Transporter NKCC1 Links Cell Volume to Cell Mass Regulation by Suppressing mTORCl. Cell Rep. 2019;27(6): 1886-96 e6. doi: 10.1016 / j.celrep.2019.04.034. PubMedPMID: 31067471.

[0316] 28. Schiapparelli P, Guerrero-Cazares H, Magana-Maldonado R, Hamilia SM, Ganaha S, Goulin Lippi Fernandes E, Huang CH, Aranda-Espinoza H, Devreotes P, Quinones-Hinojosa A. NKCC1 Regulates Migration Ability of Glioblastoma Cells by Modulation of Actin Dynamics and Interacting with Cofilin. EBioMedicine. 2017;21:94-103. Epub 20170621. doi:

[0317] 10.1016 / j.ebiom.2017.06.020. PubMed PMID: 28679472; PMCID: PMC5514434.

[0318] 29. Gong Y, Wu M, Gao F, Shi M, Gu H, Gao R, Dang BQ, Chen G. Inhibition of the p-SPAK / p-NKCCl signaling pathway protects the blood-brain barrier and reduces neuronal apoptosis in a rat model of surgical brain injury. Mol Med Rep. 2021;24(4). Epub 20210813. doi: 10.3892 / mmr.202L 12356. PubMed PMID: 34396440.

[0319] 30. Zhang J, Pu H, Zhang H, Wei Z, Jiang X, Xu M, Zhang L, Zhang W, Liu J, Meng H, Stetler RA, Sun D, Chen J, Gao Y, Chen L. Inhibition of Na(+)-K(+)-2Cl(-) cotransporter attenuates blood-brain-barrier disruption in a mouse model of traumatic brain injury. Neurochem Int. 2017;111:23-31. Epub 20170531. doi: 10.1016 / j.neuint.2017.05.020. PubMedPMID:

[0320] 28577991.

[0321] 31. Jing J, Prekeris R. Polarized endocytic transport: the roles of Rabi 1 and Rabi 1-FIPs in regulating cell polarity. Histol Histopathol. 2009;24(9): 1171-80. doi: 10.14670 / HH-24.1171. PubMedPMID: 19609864; PMCID: PMC4365979.

[0322] 32. Das L, Gard JMC, Prekeris R, Nagle RB, Morrissey C, Knudsen BS, Miranti CK, Cress AE. Novel Regulation of Integrin Trafficking by Rabll-FIP5 in Aggressive Prostate Cancer. Mol Cancer Res. 2018;16(8):1319-31. Epub 20180514. doi: 10.1158 / 1541-7786.MCR-17-0589. PubMed PMID: 29759989; PMCID: PMC6369592.

[0323] 33. Jing J, Tarbutton E, Wilson G, Prekeris R. Rabi 1-FIP3 is a Rabi 1-binding protein that regulates breast cancer cell motility by modulating the actin cytoskeleton. Eur J Cell Biol.

[0324] #14794310v12009;88(6):325-41. Epub 20090326. doi: 10.1016 / j.ejcb.2009.02.186. PubMedPMID:

[0325] 19327867; PMCID: PMC2673714.

[0326] 34. Margiotta A, Progida C, Bakke O, Bucci C. Rab7a regulates cell migration through Rael and vimentin. Biochim Biophys Acta Mol Cell Res. 2017;1864(2):367-81. Epub 20161123. doi: 10.1016 / j.bbamcr.2016.11.020. PubMed PMID: 27888097.

[0327] 35. Welsh JA, Goberdhan DCI, O'Driscoll L, Buzas El, Blenkiron C, Bussolati B, Cai H, Di Vizio D, Driedonks TAP, Erdbrugger U, Falcon-Perez JM, Fu QL, Hill AF, Lenassi M, Lim SK, Mahoney MG, Mohanty S, Moller A, Nieuwland R, Ochiya T, Sahoo S, Torrecilhas AC, Zheng L, Zijlstra A, Abuelreich S, Bagabas R, Bergese P, Bridges EM, Brucale M, Burger D, Carney RP, Cocucci E, Crescitelli R, Hanser E, Harris AL, Haughey NJ, Hendrix A, Ivanov AR, Jovanovic-Talisman T, Kruh-Garcia NA, Ku'ulei-Lyn Faustino V, Kyburz D, Lasser C, Lennon KM, Lotvall J, Maddox AL, Martens-Uzunova ES, Mizenko RR, Newman LA, Ridolfi A, Rohde E, Rojalin T, Rowland A, Saftics A, Sandau US, Saugstad JA, Shekari F, Swift S, Ter-Ovanesyan D, Tosar JP, Useckaite Z, Valle F, Varga Z, van der Pol E, van Herwijnen MJC, Wauben MHM, Wehman AM, Williams S, Zendrini A, Zimmerman AJ, Consortium M, Thery C, Witwer KW. Minimal information for studies of extracellular vesicles (MISEV2023): From basic to advanced approaches. J Extracell Vesicles. 2024;13(2):el2404. doi:

[0328] 10.1002 / jev2.12404. PubMed PMID: 38326288; PMCID: PMC 10850029.

[0329] 36. Stringer C, Wang T, Michaelos M, Pachitariu M. Cellpose: a generalist algorithm for cellular segmentation. Nat Methods. 2021;18(l): 100-6. Epub 20201214. doi: 10.1038 / s41592-020-01018-x. PubMed PMID: 33318659.

[0330] 37. Jaqaman K, Loerke D, Mettlen M, Kuwata H, Grinstein S, Schmid SL, Danuser G.

[0331] Robust single-particle tracking in live-cell time-lapse sequences. Nat Methods. 2008;5(8):695-702. Epub 20080720. doi: 10.1038 / nmeth.l237. PubMed PMID: 18641657; PMCID:

[0332] PMC2747604.

[0333] 38. Mclnnes L H, J., Melville, J. . UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction, the Journal of Open Source Software. 2018. doi:

[0334] 10.21105 / joss.0086L

[0335] 39. Andrew Y. Ng MIJ, Yair Weiss. On Spectral Clustering: Analysis and an Algorithm. MIT Press. 2001.

[0336] 40. Rousseeuw PJ. Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics. 1987;20:53-65. doi: https: / / doi.org / 10.1016 / 0377-0427(87)90125-7.

[0337] #14794310v141. Rappsilber J, Mann M, Ishihama Y. Protocol for micro-purification, enrichment, prefractionation and storage of peptides for proteomics using StageTips. Nat Protoc.

[0338] 2007;2(8): 1896-906. doi: 10.1038 / nprot.2007.261. PubMed PMID: 17703201.

[0339] 42. Demichev V, Messner CB, Vernardis SI, Lilley KS, Raiser M. DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat Methods. 2020;17(l):41-4. Epub 20191125. doi: 10.1038 / s41592-019-0638-x. PubMed PMID: 31768060; PMCID: PMC6949130.

[0340] 43. Schulte F, Hasturk H, Hardt M. Mapping Relative Differences in Human Salivary Gland Secretions by Dried Saliva Spot Sampling and nanoLC-MS / MS. Proteomics.

[0341] 2019;19(20):el900023. Epub 20190926. doi: 10.1002 / pmic.201900023. PubMedPMID:

[0342] 31476108.

[0343] 44. Millar AL, Pavios NJ, Xu J, Zheng MH. Rab3D: a regulator of exocytosis in nonneuronal cells. Histol Histopathol. 2002;17(3):929-36. doi: 10.14670 / HH-17.929. PubMed PMID: 12168804.

[0344] 45. Wu H, Voeltz GK. Reticulon-3 Promotes Endosome Maturation at ER Membrane Contact Sites. Dev Cell. 2021;56(l):52-66 e7. doi: 10.1016 / j.devcel.2020.12.014. PubMed PMID: 33434526; PMCID: PMC7837408.

[0345] 46. Lee YK, Uchida H, Smith H, Ito A, Sanchez T. The isolation and molecular characterization of cerebral microvessels. Nat Protoc. 2019; 14(11):3059-81. Epub 20191004. doi: 10.1038 / S41596-019-0212-0. PubMed PMID: 31586162.

[0346] 47. Vrbin CM. Parametric or nonparametric statistical tests: Considerations when choosing the most appropriate option for your data. Cytopathology. 2022;33(6):663-7. Epub 20220912. doi: 10.1111 / cyt.13174. PubMedPMID: 36017662.

[0347] 48. Hazra A, Gogtay N. Biostatistics Series Module 3: Comparing Groups: Numerical Variables. Indian J Dermatol. 2016;61(3):251-60. doi: 10.4103 / 0019-5154.182416. PubMed PMID: 27293244; PMCID: PMC4885176.

[0348] 49. Tong D, Liang YN, Stepanova AA, Liu Y, Li X, Wang L, Zhang F, Vasilyeva NV. Increased Epsl5 homology domain 1 and RAB11FIP3 expression regulate breast cancer progression via promoting epithelial growth factor receptor recycling. Tumour Biol.

[0349] 2017;39(2): 1010428317691010. doi: 10.1177 / 1010428317691010. PubMedPMID: 28215104.

[0350] 50. Hales CM, Vaerman JP, Goldenring JR. Rabi 1 family interacting protein 2 associates with Myosin Vb and regulates plasma membrane recycling. J Biol Chem. 2002;277(52):50415- 21. Epub 20021021. doi: 10.1074 / jbc.M209270200. PubMed PMID: 12393859.

[0351] #14794310v151. Gidon A, Bardin S, Cinquin B, Boulanger J, Waharte F, Heliot L, de la Salle H, Hanau D, Kervrann C, Goud B, Salamero J. A RabllA / myosin Vb / Rabll-FIP2 complex frames two late recycling steps of langerin from the ERC to the plasma membrane. Traffic. 2012; 13(6):815-33. Epub 20120422. doi: 10.1111 / j.1600-0854.2012.01354.X. PubMedPMID: 22420646.

[0352] 52. Ji HH, Yao LL, Liu C, Li XD. Regulation of Myosin-5b by Rabi la and the Rabi 1 family interacting protein 2. Biosci Rep. 2019;39(l). Epub 20190108. doi:

[0353] 10.1042 / BSR20181252. PubMed PMID: 30545898; PMCID: PMC6328864.

[0354] 53. Mohrmann K, Gerez L, Oorschot V, Klumperman J, van der Sluijs P. Rab4 function in membrane recycling from early endosomes depends on a membrane to cytoplasm cycle. J Biol Chem. 2002;277(35):32029-35. Epub 20020529. doi: 10.1074 / jbc.M203064200. PubMed PMID: 12036958.

[0355] 54. Busatto S, Yang Y, Walker SA, Davidovich I, Lin WH, Lewis-Tuffin L, Anastasiadis PZ, Sarkaria J, Talmon Y, Wurtz G, Wolfram J. Brain metastases-derived extracellular vesicles induce binding and aggregation of low-density lipoprotein. J Nanobiotechnology.

[0356] 2020;18(l): 162. Epub 20201107. doi: 10.1186 / sl2951-020-00722-2. PubMed PMID: 33160390; PMCID: PMC7648399.

[0357] 55. Hoshino A, Costa-Silva B, Shen TL, Rodrigues G, Hashimoto A, Tesic Mark M, Molina H, Kohsaka S, Di Giannatale A, Ceder S, Singh S, Williams C, Soplop N, Uryu K, Pharmer L, King T, Bojmar L, Davies AE, Ararso Y, Zhang T, Zhang H, Hernandez J, Weiss JM, Dumont-Cole VD, Kramer K, Wexler LH, Narendran A, Schwartz GK, Healey JH, Sandstrom P, Labori KJ, Kure EH, Grandgenett PM, Hollingsworth MA, de Sousa M, Kaur S, Jain M, Mallya K, Batra SK, Jamagin WR, Brady MS, Fodstad O, Muller V, Pantel K, Minn AJ, Bissell MJ, Garcia BA, Kang Y, Rajasekhar VK, Ghajar CM, Matei I, Peinado H, Bromberg J, Lyden D. Tumour exosome integrins determine organotropic metastasis. Nature. 2015;527(7578):329-35. Epub 20151028. doi: 10.1038 / naturel5756. PubMedPMID: 26524530; PMCID: PMC4788391.

[0358] 56. Liu X, Cao M, Palomares M, Wu X, Li A, Yan W, Fong MY, Chan WC, Wang SE. Metastatic breast cancer cells overexpress and secrete miR-218 to regulate type I collagen deposition by osteoblasts. Breast Cancer Res. 2018;20(l): 127. Epub 20181022. doi:

[0359] 10.1186 / sl3058-018-1059-y. PubMed PMID: 30348200; PMCID: PMC6198446.

[0360] 57. Ruan X, Yan W, Cao M, Daza RAM, Fong MY, Yang K, Wu J, Liu X, Palomares M, Wu X, Li A, Chen Y, Jandial R, Spitzer NC, Hevner RF, Wang SE. Breast cancer cell-secreted miR-199b-5p hijacks neurometabolic coupling to promote brain metastasis. Nat Commun.

[0361] 2024; 15(1):4549. Epub 20240529. doi: 10.1038 / s41467-024-48740-0. PubMed PMID:

[0362] 38811525; PMCID: PMC11137082.

[0363] #14794310v158. Herman H, Fazakas C, Hasko J, Molnar K, Meszaros A, Nyul-Toth A, Szabo G, Erdelyi F, Ardelean A, Hermenean A, Krizbai IA, Wilhelm I. Paracellular and transcellular migration of metastatic cells through the cerebral endothelium. J Cell Mol Med. 2019;23(4):2619-31. Epub 20190202. doi: 10.1111 / jcmm.l4156. PubMed PMID: 30712288; PMCID: PMC6433661.

[0364] 59. Motallebnejad P, Rajesh VV, Azarin SM. Evaluating the Role of IL-lbeta in Transmigration of Triple Negative Breast Cancer Cells Across the Brain Endothelium. Cell Mol Bioeng. 2022;15(l):99-l 14. Epub 20211026. doi: 10.1007 / sl2195-021-00710-y. PubMed PMID: 35096187; PMCID: PMC8761198.

[0365] 60. Godinho-Pereira J, Garcia AR, Figueira I, Malho R, Brito MA. Behind Brain Metastases Formation: Cellular and Molecular Alterations and Blood-Brain Barrier Disruption. Int J Mol Sci. 2021;22(13). Epub 20210630. doi: 10.3390 / ijms22137057. PubMed PMID: 34209088; PMCID: PMC8268492.

[0366] 61. Khuon S, Liang L, Dettman RW, Spom PH, Wysolmerski RB, Chew TL. Myosin light chain kinase mediates transcellular intravasation of breast cancer cells through the underlying endothelial cells: a three-dimensional FRET study. J Cell Sci. 2010;123(Pt 3):431-40. Epub 20100112. doi: 10.1242 / jcs.053793. PubMedPMID: 20067998; PMCID: PMC2816188.

[0367] 62. Wu F, Liu L, Zhou H. Endothelial cell activation in central nervous system inflammation. J Leukoc Biol. 2017;101(5): 1119-32. Epub 20170214. doi: 10.1189 / jlb.3RU0816-352RR. PubMed PMID: 28196850.

[0368] 63. Nag S. Morphology and properties of brain endothelial cells. Methods Mol Biol.

[0369] 2011;686:3-47. doi: 10.1007 / 978-l-60761-938-3_l. PubMed PMID: 21082365.

[0370] 64. Romano DJ, Gomez- Salinero JM, Sunic Z, Checco A, Rabbany SY. Tracking of Endothelial Cell Migration and Stiffness Measurements Reveal the Role of Cytoskeletal Dynamics. Int J Mol Sci. 2022;23(l). Epub 20220105. doi: 10.3390 / ijms23010568. PubMed PMID: 35008993; PMCID: PMC8745078.

[0371] 65. Mentor S, Makhathini KB, Fisher D. The Role of Cytoskeletal Proteins in the Formation of a Functional In Vitro Blood-Brain Barrier Model. Int J Mol Sci. 2022;23(2). Epub 20220111. doi: 10.3390 / ijms23020742. PubMed PMID: 35054928; PMCID: PMC8775705.

[0372] 66. Gonzalez R.C. WRE. Digital image processing: Pearson Prentice Hall; 2008.

[0373] 67. Desprez C, Danovi D, Knowles CH, Day RM. Cell shape characteristics of human skeletal muscle cells as a predictor of myogenic competency: A new paradigm towards precision cell therapy. J Tissue Eng. 2023;14:20417314221139794. Epub 20230316. doi:

[0374] 10.1177 / 20417314221139794. PubMed PMID: 36949843; PMCID: PMC10026113.

[0375] #14794310v168. Cullis DN, Philip B, Baleja JD, Feig LA. Rabi 1-FIP2, an adaptor protein connecting cellular components involved in internalization and recycling of epidermal growth factor receptors. J Biol Chem. 2002;277(51):49158-66. Epub 20021002. doi:

[0376] 10.1074 / jbc.M206316200. PubMed PMID: 12364336.

[0377] 69. Horgan CP, Oleksy A, Zhdanov AV, Lail PY, White IJ, Khan AR, Futter CE, McCaffrey JG, McCaffrey MW. Rabi 1-FIP3 is critical for the structural integrity of the endosomal recycling compartment. Traffic. 2007;8(4):414-30. doi: 10.1111 / j.1600-0854.2007.00543.x. PubMed PMID: 17394487.

[0378] 70. Mateus D, Marini ES, Progida C, Bakke O. Rab7a modulates ER stress and ER morphology. Biochim Biophys Acta Mol Cell Res. 2018; 1865(5):781-93. Epub 20180302. doi: 10.1016 / j.bbamcr.2018.02.011. PubMed PMID: 29505800.

[0379] 71. Xu CL, Wang JZ, Xia XP, Pan C W, Shao XX, Xia SL, Yang SX, Zheng B . Rab 11 -FIP2 promotes colorectal cancer migration and invasion by regulating PI3K / AKT / MMP7 signaling pathway. Biochem Biophys Res Commun. 2016;470(2):397-404. Epub 20160111. doi:

[0380] 10.1016 / j.bbrc.2016.01.031. PubMed PMID: 26792722.

[0381] 72. Thomsen MS, Routhe LJ, Moos T. The vascular basement membrane in the healthy and pathological brain. J Cereb Blood Flow Metab. 2017;37(10):3300-17. Epub 20170728. doi: 10.1177 / 0271678X17722436. PubMed PMID: 28753105; PMCID: PMC5624399.

[0382] 73. Jedinak A, Loughlin KR, Moses MA. Approaches to the discovery of non-invasive urinary biomarkers of prostate cancer. Oncotarget. 2018;9(65):32534-50. Epub 20180821. doi: 10.18632 / oncotarget.25946. PubMed PMID: 30197761; PMCID: PMC6126692.

[0383] 74. Bouchet BP, Akhmanova A. Microtubules in 3D cell motility. J Cell Sci.

[0384] 2017;130(l):39-50. doi: 10.1242 / jcs.189431. PubMedPMID: 28043967.

[0385] 75. de Forges H, Bouissou A, Perez F. Interplay between microtubule dynamics and intracellular organization. Int J Biochem Cell Biol. 2012;44(2):266-74. Epub 20111117. doi: 10.1016 / j.biocel.201L 11.009. PubMed PMID: 22108200.

[0386] 76. Kaverina I, Straube A. Regulation of cell migration by dynamic microtubules. Semin Cell Dev Biol. 2011;22(9):968-74. Epub 20111004. doi: 10.1016 / j.semcdb.2011.09.017.

[0387] PubMed PMID: 22001384; PMCID: PMC3256984.

[0388] 77. Virtanen MA, Uvarov P, Hubner CA, Kaila K. NKCC1, an Elusive Molecular Target in Brain Development: Making Sense of the Existing Data. Cells. 2020;9(12). Epub 20201204. doi: 10.3390 / cells9122607. PubMed PMID: 33291778; PMCID: PMC7761970.

[0389] 78. Singh R, Almutairi MM, Pacheco- Andrade R, Almiahuob MY, Di Fulvio M. Impact of Hybrid and Complex N-Glycans on Cell Surface Targeting of the Endogenous Chloride

[0390] #14794310v1Cotransporter Slcl2a2. Int J Cell Biol. 2015;2015:505294. Epub 20150817. doi: 10.1155 / 2015 / 505294. PubMed PMID: 26351455; PMCID: PMC4553341.

[0391] 79. Ben- Ari Y. NKCC1 Chloride Importer Antagonists Attenuate Many Neurological and Psychiatric Disorders. Trends Neurosci. 2017;40(9):536-54. Epub 20170814. doi:

[0392] 10.1016 / j.tins.2017.07.001. PubMed PMID: 28818303.

[0393] 80. Borgogno M, Savardi A, Manigrasso J, Turci A, Portioli C, Ottonello G, Bertozzi SM, Armirotti A, Contestabile A, Cancedda L, De Vivo M. Design, Synthesis, In Vitro and In Vivo Characterization of Selective NKCC1 Inhibitors for the Treatment of Core Symptoms in Down Syndrome. J Med Chem. 2021;64(14):10203-29. Epub 20210617. doi:

[0394] 10.1021 / acs.jmedchem.lc00603. PubMed PMID: 34137257; PMCID: PMC8311653.

[0395] 81. Savardi A, Borgogno M, De Vivo M, Cancedda L. Pharmacological tools to target NKCC1 in brain disorders. Trends Pharmacol Sci. 2021;42( 12): 1009-34. Epub 20211004. doi: 10.1016 / j. tips.2021.09.005. PubMed PMID: 34620512.

[0396] 82. Cullen PJ. Endosomal sorting and signalling: an emerging role for sorting nexins. Nat Rev Mol Cell Biol. 2008;9(7):574-82. Epub 20080604. doi: 10.1038 / nrm2427. PubMed PMID: 18523436.

[0397] 83. Granger E, McNee G, Allan V, Woodman P. The role of the cytoskeleton and molecular motors in endosomal dynamics. Semin Cell Dev Biol. 2014;31(100):20-9. Epub 20140413. doi: 10.1016 / j.semcdb.2014.04.011. PubMed PMID: 24727350; PMCID: PMC4071412.

[0398] 84. Surks HK, Richards CT, Mendelsohn ME. Myosin phosphatase-Rho interacting protein. A new member of the myosin phosphatase complex that directly binds RhoA. J Biol Chem. 2003;278(51):51484-93. Epub 20030923. doi: 10.1074 / jbc.M305622200. PubMed PMID: 14506264.

[0399] 85. Das J, Tiwari M, Subramanyam D. Clathrin Light Chains: Not to Be Taken so Lightly. Front Cell Dev Biol. 2021;9:774587. Epub 20211214. doi: 10.3389 / fcell.2021.774587. PubMed PMID: 34970544; PMCID: PMC8712872.

[0400] 86. Skeldal S, Voss LF, Lende J, Pedersen SB, Molgaard S, Kaas M, Demange P, Bentsen AH, Fuglsang M, Sander MR, Buttenschon H, Gustafsen C, Madsen P, Glerup S. Alternative splicing regulates adaptor protein binding, trafficking, and activity of the VpslOp domain receptor SorCS2 in neuronal development. J Biol Chem. 2023;299(9):105102. Epub 20230726. doi: 10.1016 / j.jbc.2023.105102. PubMed PMID: 37507021; PMCID: PMC10463258.

[0401] 87. Milatz S, Piontek J, Hempel C, Meoli L, Grohe C, Fromm A, Lee IM, El-Athman R, Gunzel D. Tight junction strand formation by claudin-10 isoforms and claudin-10a / -10b

[0402] #14794310v1chimeras. Ann N Y Acad Sci. 2017;1405(l): 102-15. Epub 20170620. doi: 10.1111 / nyas.l3393. PubMed PMID: 28633196.

[0403] 88. Neef DW, Jaeger AM, Thiele DJ. Heat shock transcription factor 1 as a therapeutic target in neurodegenerative diseases. Nat Rev Drug Discov. 2011;10(12):930-44. Epub 20111201. doi: 10.1038 / nrd3453. PubMed PMID: 22129991; PMCID: PMC3518299.

[0404] 89. Adams RA, Passino M, Sachs BD, Nuriel T, Akassoglou K. Fibrin mechanisms and functions in nervous system pathology. Mol Interv. 2004;4(3): 163-76. doi: 10.1124 / mi.4.3.6. PubMed PMID: 15210870.

[0405] 90. Horowitz JC. Releasing Tensin. Am J Respir Cell Mol Biol. 2017;56(4):417-8. doi: 10.1165 / rcmb.2016-0417ED. PubMed PMID: 28362149; PMCID: PMC5449518.

[0406] 91. Jin G, Lin L, Li K, Li J, Yu C, Wei Z. Structural basis of ELKS / Rab6B interaction and its role in vesicle capturing enhanced by liquid-liquid phase separation. J Biol Chem.

[0407] 2023;299(6): 104808. Epub 20230511. doi: 10.1016 / j.jbc.2023.104808. PubMed PMID:

[0408] 37172719; PMCID: PMC10267580.

[0409] 92. Steegmaier M, Klumperman J, Foletti DL, Yoo JS, Scheller RH. Vesicle-associated membrane protein 4 is implicated in trans-Golgi network vesicle trafficking. Mol Biol Cell. 1999; 10(6): 1957-72. doi: 10.1091 / mbc.l0.6.1957. PubMed PMID: 10359608; PMCID:

[0410] PMC25394.

[0411] 93. Wakana Y, Villeneuve J, van Galen J, Cruz-Garcia D, Tagaya M, Malhotra V. Kinesin-5 / Eg5 is important for transport of CARTS from the trans-Golgi network to the cell surface. J Cell Biol. 2013;202(2):241-50. Epub 20130715. doi: 10.1083 / jcb.201303163. PubMed PMID: 23857769; PMCID: PMC3718972.

[0412] 94. Lei Y, Xu J, Li M, Meng T, Chen M, Yang Y, Li H, Zhuang T, Zuo J. MIA SH3 Domain ER Export Factor 3 Deficiency Prevents Neointimal Formation by Restoring BAT-Like PVAT and Decreasing VSMC Proliferation and Migration. Front Endocrinol (Lausanne).

[0413] 2021;12:748216. Epub 20211110. doi: 10.3389 / fendo.2021.748216. PubMed PMID: 34858331; PMCID: PMC8631732.

[0414] 95. Li N, Chen L, Zhao X, Gu C, Chang Y, Feng S. Targeting ANXA7 / LAMP5-mTOR axis attenuates spinal cord injury by inhibiting neuronal apoptosis via enhancing autophagy in mice. Cell Death Discov. 2023;9(l):309. Epub 20230824. doi: 10.1038 / s41420-023-01612-w. PubMed PMID: 37620352; PMCID: PMC 10449888.

[0415] 96. Ong WY, Herr DR, Farooqui T, Ling EA, Farooqui AA. Role of sphingomyelinases in neurological disorders. Expert Opin Ther Targets. 2015; 19(12): 1725-42. Epub 20150804. doi: 10.1517 / 14728222.2015.1071794. PubMed PMID: 26243307.

[0416] #14794310v197. Bonomini F, Rezzani R. Aquaporin and blood brain barrier. Curr Neuropharmacol.

[0417] 2010;8(2):92-6. doi: 10.2174 / 157015910791233132. PubMedPMID: 21119879; PMCID:

[0418] PMC2923372.

[0419] 98. Kramer A, Green J, Pollard J, Jr., Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30(4):523-30. Epub 20131213. doi:

[0420] 10.1093 / bioinformatics / btt703. PubMed PMID: 24336805; PMCID: PMC3928520.

[0421] 99. Fidler IJ, Yano S, Zhang RD, Fujimaki T, Bucana CD. The seed and soil hypothesis: vascularisation and brain metastases. Eancet Oncol. 2002;3(l):53-7. doi: 10.1016 / s 1470-2045(01)00622-2. PubMed PMID: 11905606.

[0422] 100. Zhang YY, Kong EQ, Zhu XD, Cai H, Wang CH, Shi WK, Cao MQ, Li XL, Li KS, Zhang SZ, Chai ZT, Ao JY, Ye BG, Sun HC. CD31 regulates metastasis by inducing epithelial-mesenchymal transition in hepatocellular carcinoma via the ITGBl-FAK-Akt signaling pathway. Cancer Lett. 2018;429:29-40. Epub 20180508. doi: 10.1016 / j.canlet.2018.05.004. PubMed PMID: 29746931.

[0423] 101. Zhang Z, Gan Q, Han J, Tao Q, Qiu WQ, Madri JA. CD31 as a probable responding and gate-keeping protein of the blood-brain barrier and the risk of Alzheimer's disease. J Cereb Blood Flow Metab. 2023;43(7): 1027-41. Epub 20230412. doi: 10.1177 / 0271678X231170041. PubMed PMID: 37051650; PMCID: PMC10291450.

[0424] 102. Lertkiatmongkol P, Liao D, Mei H, Hu Y, Newman PJ. Endothelial functions of platelet / endothelial cell adhesion molecule-1 (CD31). Curr Opin Hematol. 2016;23(3):253-9. doi: 10.1097 / MOH.0000000000000239. PubMedPMID: 27055047; PMCID: PMC4986701. 103. Eslami SZ, Cortes-Hernandez LE, Alix-Panabieres C. Epithelial Cell Adhesion Molecule: An Anchor to Isolate Clinically Relevant Circulating Tumor Cells. Cells. 2020;9(8). Epub 20200805. doi: 10.3390 / cells9081836. PubMedPMID: 32764280; PMCID: PMC7464831.

[0425] #14794310v1

Claims

1. CLAIMSWhat is claimed is:

1. A method of determining motility and / or morphology of a plurality of cells of a cell barrier, the cell barrier being a monolayer cell barrier or a multilayer cell barrier, the method comprising:using at least one computer hardware processor to perform:obtaining an image of the plurality of cells of the cell barrier;segmenting the image to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells;extracting motility and / or morphology feature values from the segmented image data for at least some cells in the plurality of cells; andclustering the at least some cells of the plurality of cells, using their respective motility and / or morphology feature values, to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology.

2. The method of claim 1, wherein obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of an endothelial barrier, an epithelium barrier, or an epithelial barrier.

3. The method of claim 2, wherein obtaining an image of the plurality of cells of an endothelial barrier comprises obtaining an image of a blood-brain barrier (BBB).

4. The method of any one of claims 1-3, wherein obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of an in vitro model of the cell barrier.

5. The method of claim 4, wherein obtaining an image of the in vitro model of the cell barrier comprises obtaining images of a cell monolayer.

6. The method of claim 5, wherein the cell monolayer is representative of an epithelial or endothelial barrier.#14794310v17. The method of any one of claims 1-6, wherein obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of a cell barrier model that is representative of an in vivo cell barrier.

8. The method of any one of claims 1-7, wherein obtaining an image of a plurality of cells of a cell barrier comprises capturing the image using confocal imaging, fluorescent imaging, or quantitative phase imaging.

9. The method of any one of claims 1-8, wherein obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of fixed cells.

10. The method of any one of claims 1-9, wherein obtaining the image of the plurality of cells of the cell barrier comprises obtaining an image of live cells.

11. The method of any one of claims 1-10, wherein obtaining an image of the plurality of cells comprises capturing the image.

12. The method of any one of claims 1-11, wherein obtaining an image of the plurality of cells of the cell barrier comprises obtaining a time series of images of the plurality of cells of the cell barrier.

13. The method of claim 12, wherein segmenting the image to obtain the segmented cell data further comprises segmenting images in the time series of images of the plurality of cells to obtain the segmented cell data, the segmented cell data comprising segmented image data for the cells of the plurality of cells.

14. The method of any one of claims 1-13, wherein segmenting the image to obtain the segmented cell data further comprises segmenting using a trained neural network.

15. The method of any one of claims 1-14, wherein extracting motility and / or morphology feature values from the segmented image data comprises extracting cell motility feature values.

16. The method of claim 15, wherein extracting the cell motility feature values comprise extracting one or more of cell: standard deviation (std) of speed, max speed, mean speed, min #14794310v1speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and max distance travelled.

17. The method of claim 16, wherein extracting the cell motility feature values comprise extracting each of cell: std speed, max speed, mean speed, min speed, mean straight line speed, linearity of forward progression, confinement ratio, total distance traveled, and max distance travelled.

18. The method of any one of claims 1-17, wherein extracting motility and / or morphology feature values from the segmented image data comprises extracting cell morphology feature values for the at least some cells of the plurality of cells.

19. The method of claim 18, wherein extracting the cell morphology feature values comprises extracting one or more of cell: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity.

20. The method of claim 19, wherein extracting the cell morphology feature values comprises extracting each of cell: area, convex area, equivalent diameter area, minor axis length, max ferret diameter, major axis length, perimeter, solidity, extent, and eccentricity.

21. The method of any one of claims 1-20, wherein extracting the motility and / or morphology feature values further comprises reducing the dimensions of the extracted motility and / or morphology feature values.

22. The method of claim 21, wherein reducing the dimensions of the extracted motility and / or morphology feature values is performed using uniform manifold approximation and projection (UMAP).

23. The method of any one of claims 1-22, wherein clustering the cells of the plurality of cells is performed using an unsupervised clustering technique.

24. The method of any one of claims 1-23, wherein clustering the cells of the plurality of cells comprises clustering using Centroid-based Clustering (Partitioning methods), Density-#14794310v1based Clustering (Model-based methods), Connectivity-based Clustering (Hierarchical clustering), Distribution-based Clustering and / or spectral clustering.

25. The method of any one of claims 1-24, wherein clustering the cells of the plurality of cells comprises clustering using silhouette score.

26. The method of any one of claims 1-25, wherein clustering the cells of the plurality of cells comprises non-hierarchical spectral clustering with a maximized Silhouette score.

27. The method of any one of claims 1-26, wherein obtaining an image of the plurality of cells of the cell barrier comprises obtaining an image of a plurality of cells of a cell barrier that have been contacted with a drug.

28. The method of claim 27, further comprising determining an effect of the drug on the motility and / or morphology of the cells of the plurality of cells of the cell barrier compared to a control.

29. The method of claim 28, wherein a class of cell motility and / or morphology comprises an increase and / or decrease in one or more cell motility and / or morphology feature values relative to the other classes of cell motility and / or morphology.

30. A method of determining changes in cell motility and / or morphology in a cell barrier in different conditions, the method comprising:using a hardware processor to perform:obtaining a first image of a plurality of cells of a cell barrier in a first condition, and a second image of a plurality of cells of a cell barrier in a second condition;determining first motility and / or morphology feature values of the cell barrier in the first condition and second motility and / or morphology feature values of the cell barrier in the second condition, the determining comprising for the first image and the second image, respectively:segmenting the image to obtain segmented cell data, the segmented cell data comprising segmented image data for cells of the plurality of cells;extracting motility and / or morphology feature values from the segmented image data for each of the cells of the plurality of cells; and#14794310v1clustering the cells of the plurality of cells to obtain a plurality of clusters, each cluster representing a class of cell motility and / or morphology; and determining changes in cell motility and / or morphology of the cell barrier in the different conditions, the determining comprising identifying one or more changes in the plurality of clusters in the first condition and the second condition.

31. A system, comprising:at least one computer hardware processor; andat least one non-transitory computer-readable storage medium that, when executed by the at least one computer hardware processor, causes the at least one computer hardware processor to perform the method of any one of claims 1-30.

32. At least one non-transitory computer-readable storage medium that, when executed by at least one computer hardware processor, causes the at least one computer hardware processor to perform the method of any one of claims 1-30.#14794310v1

Images