Targeting infiltrating cd8+ t cells in alzheimer's disease

Abstract

Described herein is a 3D human neuroimmune axis model comprising human stem cell-derived neurons, astrocytes, microglia, and peripheral immune cells isolated from human blood. Also described herein are methods and compositions comprising CXCR3 inhibitors for treatment of Alzheimer's disease.

Description

[0001] TARGETING INFILTRATING CD8+ T CELLS IN ALZHEIMER’S DISEASE

[0002] CLAIM OF PRIORITY

[0003] This application claims the benefit of U.S. Provisional Application Serial No. 63 / 466,989, filed on May 16, 2023, and 63 / 531,240 filed on August 7, 2023. The entire contents of the foregoing are incorporated herein by reference.

[0004] TECHNICAL FIELD

[0005] Described herein is a 3D human neuroimmune axis model comprising human stem cell-derived neurons, astrocytes, microglia, and peripheral immune cells isolated from human blood. Also described herein are methods and compositions comprising CXCR3 inhibitors for treatment of Alzheimer’s disease.

[0006] BACKGROUND

[0007] Adaptive immune cells have been increasingly shown to play critical roles in the pathogenesis of neurodegenerative diseases like AD and Parkinson’s disease (PD)1 5. Studies of AD neuropathology have revealed high numbers of T cells in the brains of AD patients versus healthy controls6 1 1. In addition, AD mouse models have exhibited evidence for T cell infiltration into the brain12T A recent study has reported human CD8+ T cell clonal expansion in AD cerebrospinal fluid (CSF) and brain, suggesting CD8+ T cells may impact neurodegeneration and / or cognitive impairment in AD patients11. Another recent report on CD4+ T cell infiltration and their contribution to neurodegeneration in Lewy body dementia (LBD) provides further evidence of the important role of T cell adaptive immunity in neurodegenerative disease19. These findings underscore the importance of assessing the pathogenic roles of infiltrating T cells in the brains of patients with AD and AD- related dementias (ADRD).

[0008] SUMMARY

[0009] Described herein is a multicellular human in vitro model for investigating infiltration of peripheral immune cells into 3D neural-glial cell cultures. The exemplary PiChip model described herein overcomes several current technical hurdles to study cellular mechanisms governing immune cell infiltration in the AD brain, and yielded important and exciting insights. A dramatic increase in infiltrating T cells (but not B cells) selectively into AD neural-glial cultures, independent of the presence of microglia, was observed. In addition, infiltration of CD8+ T cells into AD cultures led to enhanced activation of microglia and exacerbated neuroinflammation and neurodegeneration (summarized in Fig. 7). Mechanistically, the C-X-C motif chemokine ligand 10 (CXCL10) - C-X-C motif chemokine receptor 3 (CXCR3) signaling axis regulated the infiltration of T cells in the AD cultures. Blockade of the T cell chemotaxis receptor, CXCR3, stopped neuronal toxicity and provides a druggable therapeutic target governing T-cell CXCR3 / CXCL10 signaling in the AD brain.

[0010] Thus, provided herein are systems comprising: a microfluidic chip comprising: a first compartment, and a plurality of second compartments disposed along a periphery of the first compartment, wherein each of the second compartments is in fluid communication with the first compartment via a respective plurality of microfluidic channels, wherein the microfluid chip is configured to: receive a three- dimensional cell culture comprising first cells in the first compartment, receive second cells in the second compartments, and permit at least some of the second cells to migrate from the second compartments to the three-dimensional cell culture in the first compartment via the microfluidic channels.

[0011] In some embodiments, the second compartments are evenly distributed around the first compartment. In some embodiments, the microfluidic chip comprises four second compartments. In some embodiments, each of the microfluidic channels extends linearly between a respective one of the second compartments and the first compartment.

[0012] In some embodiments, the microfluidic chip further comprises a chamber aligned concentrically with the first compartment, and wherein the chamber is in fluid communication with the first compartment.

[0013] In some embodiments, the microfluidic channels are disposed on a first layer of the microfluidic chip, wherein the first compartment and the second compartment are disposed on a second layer of the microfluidic chip, and wherein the chamber is disposed on a third layer of the microfluidic chip. In some embodiments, the first compartment has an inner radius of about 2000 pm.

[0014] In some embodiments, each of the second compartments has an inner radius of

[0015] 2500 pm.

[0016] In some embodiments, the chamber has an inner radius of about 2500 pm.

[0017] In some embodiments, for each of the second compartments, at least about 100 of the microfluidic channels extend in parallel between that second compartment and the first compartment.

[0018] In some embodiments, each of the microfluidic channels has: a height between 5 pm and 10 pm, preferably a height of 10 pm; a width between 5 pm and 10 um, preferably a width of 10 pm; and a length between 100 pm and 500 pm, preferably a length of 100 pm.

[0019] In some embodiments, the microfluidic chip comprises poly dimethyl siloxane (PDMS).

[0020] In some embodiments, the first cells comprise brain cells. In some embodiments, the brain cells comprise one or more of neurons, astrocytes, or microglial cells, preferably wherein the neurons, astrocytes, or microglial cells are human stem cell-derived. In some embodiments, the brain cells comprise neurons, astrocytes and microglial cells, preferably wherein the neurons, astrocytes, and microglial cells are human stem cell-derived.

[0021] In some embodiments, the second cells comprise peripheral blood mononuclear cells (PBMCs). In some embodiments, the second cells comprise one or more of T cells, B cells, or monocytes.

[0022] In some embodiments, the systems further comprise an imaging system configured to obtain one or more optical images of the first cells and the second cells in the microfluidic chip. In some embodiments, the imaging system comprises a confocal microscope.

[0023] In some embodiments, the system further comprises as computer system configured to: access the one or more optical images, and determine, based on the one or more optical images, at least one of: a number of the second cells that migrated from the second compartments, or a velocity of migration of the second cells from the second compartments. Additionally, provided herein are methods comprising: obtaining a microfluidic chip as described herein, e.g., comprising: a first compartment, and a plurality of second compartments disposed along a periphery of the first compartment, wherein each of the second compartments is in fluid communication with the first compartment via a respective plurality of microfluidic channels, depositing a three- dimensional cell culture comprising first cells in the first compartment, depositing second cells in the second compartments; and determining a migration of at least some of the second cells from the second compartments to the three-dimensional cell culture in the first compartment via the microfluidic channels.

[0024] In some embodiments, the first cells comprise brain cells. In some embodiments, the first cells comprise one or more of, preferably all of, neurons, astrocytes, or microglial cells, preferably wherein the neurons, astrocytes, or microglial cells are human stem cell-derived.

[0025] In some embodiments, the second cells comprise peripheral blood mononuclear cells (PBMCs). In some embodiments, the second cells comprise one or more of T cells, B cells, or monocytes.

[0026] In some embodiments, the methods further comprise obtaining, using an imaging system, one or more optical images of the first cells and the second cells in the microfluidic chip. In some embodiments, the imaging system comprises a confocal microscope.

[0027] In some embodiments, the methods further comprise: accessing the one or more optical images, and determining, based on the one or more optical images, at least one of a number of the second cells that migrated from the second compartments, or a velocity of migration of the second cells from the second compartments.

[0028] Further, provided herein are methods of treating a subject who has, or is at risk of developing, Alzheimer’s disease, or another neurodegenerative disease associated with T cell infiltration. The methods comprise administering to the subject a therapeutically effective amount of an inhibitor of C-X-C motif chemokine receptor 3 (CXCR3). Also provided herein is an inhibitor of C-X-C motif chemokine receptor 3 (CXCR3), for use in a method of treating a subject who has, or is at risk of developing, Alzheimer’s disease, or another neurodegenerative disease associated with T cell infiltration.

[0029] In some embodiments, the inhibitor of CXCR3 is an antibody that binds to CXCR3. In some embodiments, the antibody that binds to CXCR3 is selected from the group consisting of 4Ch, 4Hul, 4Hu2, 4Hu3, 4Hu4, 4Hu5, 4Hu6, 4Hu7, 4Hu8, 4Hu9, 4HulO, 4Hul l, 4Hul2, 4Hul3, 4Hul4, 4Hul5, 4Hul6, 4Hul7, 4Hul8, 4Hul9, 12Ch, 12Hul, 12Hu2, 12Hu3, 82Ch, 82Hul, 82Hu2, 82Hu3, 135Ch, 135Hul, 135Hu2, 135Hu3, 53Ch, 53Hul, 53Hu2, 53Hu3, 53Hu4, 53Hu5, 53Hu6, 53Hu7, 53Hu8, 53Hu9, 53HulO, 53Hul l, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, 53Hul8, 53Hul9, 53Hu29, B2, C12, F7, 2A3, 3A3, 3A11, 3C11, 4D5, 32B12, 32D12, Hl, B7, C12, D12, El, G12, 3A5, 3B12, 4C3, B12, C4, F7, 2A3, 3A7, 3A8, 23, 24, 25, 26, 27, 29, 30, 31, 33, 34, 35, 37, 53, 53Ch, 53Hu01, 53HuO2, 53Hu03, 53HuO4, 53Hu05, 53Hu06, 53Hu07, 53Hu08, 53Hu09, 53Hu010, 53Hul l, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, and 53Hul8.

[0030] In some embodiments, the antibody that binds to CXCR3 is human or humanized.

[0031] In some embodiments, the inhibitor of CXCR3 is a small molecule antagonist of CXCR3. In some embodiments, the small molecule antagonist of CXCR3 is selected from the group consisting of ACT-777991, JN-2, ACT-672125, (±)-AMG 487, (±)-NBI 74330, N,N-Di substituted benzylamine derivatives, SCH 546738, NIBR2130, and AMG 487.

[0032] In some embodiments, the inhibitor of CXCR3 is administered intravenously. In some embodiments, the inhibitor of CXCR3 is formulated to be administered intravenously.

[0033] In some embodiments, the subject is a human subject.

[0034] In some embodiments, the subject has been diagnosed with Alzheimer’s disease, or has been identified as having a risk factor for Alzheimer’s disease, optionally a family history of Alzheimer’s disease or one or more genetic risk factors for Alzheimer’s disease.

[0035] As used herein, “about” means ± 10% of the referenced value.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Methods and materials are described herein for use in the present invention; other, suitable methods and materials known in the art can also be used. The materials, methods, and examples are illustrative only and not intended to be limiting. All publications, patent applications, patents, sequences, database entries, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control.

[0037] Other features and advantages of the invention will be apparent from the following detailed description and figures, and from the claims.

[0038] DESCRIPTION OF DRAWINGS

[0039] Fig. 1 | Construction and characterization of a 3D human neuroimmune axis model of AD. a, Multicellular in vitro modeling of AD brain in a 3D microfluidic platform (referred to as the PiChip system). Schematic represent a pipeline for constructing the 3D human neuroimmune axis model using the PiChip system to recapitulate crosstalk between human peripheral immune cells and brain cells. This model allows quantitative analysis of migration signatures of peripheral immune cell and their interactions with brain cells in the PiChip system using time-lapse imaging and unbiased / automated image analysis, b, Schematic represents the multicompartment microfluidic system for the PiChip system. 400 narrow migration microchannels (four sides, 100 on each side) (10x 10x500 pm3in height, width, and length) connecting the PiChip BRAIN compartment to the four PERIPHERAL compartments provide a gradient of soluble factors released by CNS cells in the BRAIN compartment. The design of the PiChip system allows probing selective infiltration of peripheral immune cells in real-time, c, Confocal images of 3D- differentiated NPCs to AD Neu / AC and iMGL in the PiChip BRAIN compartment and nuclei-stained CD8+ cells in surrounding PERIPHERAL compartments, scale bar, 500 pm. d, Representative images of microglial cells’ morphology in CTRL and AD conditions stained for IBA1 (Neu / AC, gray; iMGL, magenta - left panel). Scale bar, 50 pm (left panel) and 10 pm (right panel), e, Quantification of IB Al expression, branch / body ratio, and length of microglial cells in CTRL and AD conditions of the PiChip system (data are means+s.e.m.; P values from unpaired / -tests, two-tailed with Welch’s correction), f, Whole RNAseq analysis showing significant increases in inflammatory and activated glial markers (C3, CSF1R, INFGR1, and CCL2) in AD cultures compared to CTRL (P values from unpaired Ltests, two-tailed with Welch’s correction; ns, not significant), g, Violin plot showing highly enriched genes associated with reactive astrocytes in AD compared to CTRL (P values from Wilcoxon test), h, UMAP visualization of a total of 20153 single cells from AD and CTRL cultures by single-cell transcriptome profiles. Distinct cell type (neurons, astrocytes, and microglia) are depicted with different colors, i, Volcano plot of differentially expressed genes of neurons in AD compared to CTRL. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots, j, Gene ontology of significantly upregulated and downregulated pathways in neurons in AD cultures compared to CTRL (FDR<0.1). Significantly upregulated pathways are depicted in red, while downregulated are in blue, k, Volcano plot of differentially expressed genes of astrocytes in AD compared to CTRL. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots. 1, Gene ontology of significantly upregulated and downregulated pathways in astrocytes in AD cultures compared to CTRL (FDR<0.1). Significantly upregulated pathways are depicted in red, while downregulated are in blue.

[0040] Fig. 2 | Selective infiltration of T cells into the 3D- AD cellular models and 5XFAD mice, a, Representative confocal microscopy imaging showing 3D- differentiated AD Neu / AC / iMGL cultures in the PiChip BRAIN compartment of the PiChip system (Neu / AC, yellow; iMGL, blue). Scale bar, 500 pm. b, Time-lapse confocal imaging showing migratory movement of CD8+ T cells (nuclei-stained, magenta) migratory movement toward the PiChip BRAIN compartment containing the 3D-AD cultures (Neu / AC, yellow; iMGL, blue) through microchannels over time. Scale bar, 500 pm. c, High magnification of the 3D-differentiated human AD Neu / AC (yellow) and iMGL (blue) in the BRAIN compartment of the PiChip system. Scale bar, 20 pm. d, High magnification of infiltrating CD8+ T cells (nuclei-stained, magenta) and AD Neu / AC (yellow) in the PiChip BRAIN compartment. Scale bar, 20 pm. e-h, Quantification of migration index of peripheral immune cells (i.e., fold change of immune cells that infiltrated into the PiChip BRAIN compartment in AD vs CTRL), including CD3+ T cells (e), CD4+ and CD8+ T cells (f), monocytes (g), and B cells (h). PBMCs were isolated from de-identified donors and loaded into the PiChip system containing AD and CTRL Neu / AC (+ / -) iMGL cultures. Data were extracted from confocal imaging at 15 hours post-addition of PBMCs to the PiChip system (72=8-10 donors for CD3+ T cells; n=ll donors for CD4+ T cells; «=10 donors for CD8+ T cells; n=6-9 donors for monocytes; 72=8- 12 donors for B cells, data are means+s.e.m. from 3-6 independent PiChip systems per donor; circle=region of interest from 100 microchannels in one PiChip system; P values from one sample t- tests). i, Quantification of migration velocity of infiltrating CD3+ T cells (left), monocytes (middle), and B cells (right) via confined microchannels over 17 hours in the PiChip system. PBMCs were isolated from de-identified donors and loaded into the PiChip system containing AD and CTRL Neu / AC (+ / -) iMGL cultures (w=8-9 donors for CD3+ T cells; «=9-10 donors for monocytes; 72=8-12 donors for B cells, data from 3-6 independent PiChip systems per donor; box-and-whiskers represent medians and the interquartile range with outliers outside whiskers; P values from two- way ANOVA with Tukey multiple comparisons tests), j-k, Quantification of the percentage (j) and absolute number (k) of extravascular CD4+ and CD8+ T cells in 5XFAD and WT brains (w=9-12 mice per group; box-and-whiskers represent medians and the interquartile range with outliers outside whiskers.; P values from unpaired student’s / -tests).

[0041] Fig. 3 | Infiltrating CD8+ T cells dramatically exacerbate neurodegeneration in the AD neuroimmune axis model via interaction with microglia, a, Schematic represents potential roles of microglia and CD8+ T cells in promoting neurodegeneration in AD. b, Time-lapse confocal microscopy imaging showing the cellular damage along with recruited T cells into the PiChip BRAIN compartment (AD NeuAC, green; iMGL, magenta) of the AD PiChip system in realtime, 72 hours of observation. Scale bars, 20 pm. c-d, Quantification of cellular damage in AD and CTRL Neu / AC (+ / -) iMGL and T cells cultures using GFP+ expressing surface area (c) and cell body count (i.e., number of cells per ROI) (d) («=4 donors; data are means+s.e.m.; dots represent an image from one independent well; P values from two-way ANOVA with Tukey multiple comparisons tests), e, Representative confocal microscopy imaging showing cellular damages observed in the presence of CD8+ T cells and microglia in AD conditions compared to CTRL (Neu / AC, green: iMGL, magenta; CD4+ / CD8+ T cells, blue) in the PiChip system, f, Quantification of cellular damage in AD Neu / AC (+ / -) iMGL and CTRL cultures in the presence of CD4+ or CD8+ T cells using cell surface area (GFP, green) ( / / =4 donors; data are means+s.e.m.; dots represent an image from one independent well; P values from two-way ANOVA with Tukey multiple comparisons tests), g-h, Box plots represent quantification of neuronal (Tuj 1+ expressing surface area, g) and astrocyte damage (ALDH1+ expressing surface area, h) damages in AD Neu / AC (+ / -) iMGL condition in the presence of CD4+ or CD8+ T cells (w=4 donors; data are means+s.e.m.; dots represent an image from one independent well; P values from two-way ANOVA with Tukey multiple comparisons tests), i, Representative automasking imaging for Neu / AC / iMGL in the presence of CD8+ T cells for quantification of the average neurite density, normalized to cell body counts (neurite, cyan; cell body, magenta), j, Quantification of the average neurite density in AD Neu / AC (+ / -) iMGL condition in the presence of CD4+ or CD8+ T cells («=4 donors; data are means+s.e.m.; dots represent an image from one independent well; P values from two-way ANOVA with Tukey multiple comparisons tests).

[0042] Fig. 4 | Compositional changes and inflammatory activity of T cells and microglia in AD conditions, a, UMAP visualization of T cell-type composition of the PiChip cultures by single-cell transcriptome profiles. Distinct subtypes of T cells (CD4+ Naive, CD4+ Memory, and CD8+ cytotoxic) are depicted with different colors, b, Compositional changes of CD4+ Naive, CD4+ Memory, and CD8+ cytotoxic T cells in AD or CTRL Neu / AC (+ / -) iMGL conditions, c-d, Bar graphs with individual data points showing the % of cytotoxic CD8+ T cells (c) and memory CD4+ T cells (d) in AD or CTRL Neu / AC (+ / -) iMGL conditions (data are means+s.e.m.; P values from one-way ANOVA tests), e, UMAP visualization of T cell-type composition of the PiChip cultures by single-cell transcriptome profiles. Distinct subtypes of T cells (homeostatic, DAM, MHCII, proinflammatory, cytokine 1, cytokine 2, proliferative, and ribosomal) are depicted with different colors, f, Compositional changes of homeostatic, DAM, MHCII, proinflammatory, cytokine 1, cytokine 2, proliferative, and ribosomal microglial cells in AD or CTRL Neu / AC / iMGL (+ / -) T cells, g-h, Bar graphs with individual data points showing the % of DAM (g) and MHCII (h) in AD or CTRL Neu / AC / iMGL (+ / -) T cells (data are means+s.e.m.; P values from one-way ANOVA tests), i, Volcano plots of differentially expressed genes in microglia in AD Neu / AC / iMGL versus CTRL in the presence of infiltrating T cells. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots, j, Violin plot showing selected significantly enriched interferon-associated genes (IFITM1 and STAT1) in microglia in AD Neu / AC / iMGL compared to CTRL in the presence of T cells (P values from Mann-Whitney test), k, Gene ontology of significantly upregulated and downregulated pathways (FDR<0.1) in microglia in AD versus CTRL in the presence of T cells.

[0043] Fig. 5 | Upregulation of CXCL10 and its receptor, CXCR3, in the AD neuroimmune axis model and 5XFAD brain, a, Quantification of chemokines in AD Neu / AC cultures (+ / -) microglial cells compared to CTRL. Media were collected from the PiChip BRAIN compartment (data are means+s.e.m.; P values from two- way ANOVA with Tukey multiple comparisons tests), b, Violin plot showing highly enriched astrocytic CXCL10 expression in AD compared to CTRL in the PiChip system using scRNA-seq data (P values from two-tailed, / -test) . c, Quantification of CXCL10 protein level from astrocytes in the presence of AD environment and CTRL (data are means+s.e.m.; dots represent biological replicates; P values from one-way ANOVA with Tukey multiple comparisons tests), d, Schematics describe an in vitro chemotaxis model in a microfluidic platform. Recombinant human CXCL10 was added to the CHEMOKINE compartment (central chamber), while isolated human CD4+ and CD8+ T cells from whole blood were added to the PERIPHERAL compartment, e, Time-lapse microscopy imaging showing directional migration of T cells (blue) through confined microchannels over ten hours toward chemokine gradient generated by CXCL10 recombinant protein, f, Quantification of the number of migrated CD4+ and CD 8+ T cells in response to CXCL10 chemokine («=7 donors; data from 2 independent PiChip systems per donor; box-and-whiskers represent medians and the interquartile range with outliers outside whiskers; P values from sample / -tests). g, Quantification of migration velocity of CD4+ and CD8+ T cells through confined microchannels over 17 hours in response to CXCL10 (100 nM). CD4+ and CD8+ T cells were isolated from de-identified donors (w=8 donors; P values from unpaired / -tests, two-tailed with Welch’s correction), h, Dot plot of CXCR3 expression for distinct T cell-types in different conditions using scRNA-seq data. Color scale indicates the average of expression of CXCR3 gene in each cell population, and dot size is proportional to the percentage of cells expressing the CXCR3 gene, i, Representative immunofluorescence staining of CD8+ T cells, colabeled with CXCR3 and DAPI in 6-7-month-old 5XFAD and WT mouse brains. CD8+ exhibits a magenta color, whereas CXCR3 exhibits a yellow. Scale bar, 20 pm (n =13 5XFAD and «=8 WT mice), j, Representative immunofluorescence staining of 6-7-month-old 5XFAD and WT mouse brains co-labeled with anti-CXCLIO (red), anti-GFAP (green), and DAPI (nuclei). Scale bar, 100 pm and 35 pm in inserts, k-1, Quantification of the percentage of CXCL10 in 6-7-month-old 5XFAD and WT brains (n=l 3 5XFAD and w=8 WT mice; box-and-whiskers represent medians and the interquartile range with outliers outside whiskers; P values from unpaired / -tests, two- tailed).

[0044] Fig. 6 | Blocking the binding of CXCL10 to T cell CXCR3 attenuate T cell infiltration and neurodegeneration in the neuroimmune axis model, a, Schematic shows the experimental design for inhibition of the crosstalk between generated CXCL10 in AD cell cultures and T cell receptor, CXCR3 using anti-CXCR3 neutralizing antibody (MAB-160). b-c, Box plots represent percentage of CD4+ and CD8+ T cells in blood (b) and % level of CXCR3 receptor expression in CD4+ and CD8+ T cells using flow cytometry (c). Human CD4+ and CD8+ T cells were isolated from whole blood of the same de-identified donors (w=8- 10 donors for CD4+ and CD8+ T cells; data are means+s.e.m.; P values from unpaired / -tests, two-tailed with Welch’s correction), d-e, Representative histogram images of flow cytometry before and after blocking CXCR3 receptor in CD4+ and CD8+ T cells with anti-CXCR3 neutralizing antibody (MAB160, 10 pg / mL). Quantitative analysis demonstrates a significant loss of surface expression of CXCR3 in CD4+ and CD8+ T cells in response to 10 pg / mL anti-CXCR3 neutralizing antibody (MAB-160) (n=8 donors for CD4+ T cells and n=6 donors for CD8+ T cells; data are means+s.e.m.; P values from unpaired / -tests, two-tailed with Welch’s correction), f, Quantification of migration indexinhibited (see Methods) of CD4+ and CD8+ T cells with and without CXCR3 neutralizing antibody treatment (MAB160, 10 pg / mL) in the PiChip system containing AD Neu / AC / iMGL culture. Data were extracted from confocal imaging at 15 hours post-addition of T cells to the PiChip systems (w=7 donors, data are means+s.e.m. from 6 independent PiChip systems per donor; circle=region of interest from 100 microchannels in one PiChip system; P values from one sample / -tests). g, Quantification of average migration velocity of infiltrating CD4+ and 8+ T cell with and without MAB160 treatment (10 g / mL) through the microchannels over 17 hours in the PiChip system (n=6 donors; data from 6 independent PiChip systems per donor; box-and-whiskers represent medians and the interquartile range with outliers outside whiskers; P values from two-way ANOVA with Tukey multiple comparisons tests), h, Schematic represent potential effect of blocking CXCR3 T cell receptor using neutralizing antibody (MAB160) on cellular damage in AD cultures, i, Box plots represents quantification of GFP+ expressing surface area, cell body count (i.e., number of cells per ROI), and Tuj 1+ expressing surface area in AD Neu / AC / iMGL cultures before and after T cell treatment with CXCR3 neutralizing antibody (MAB160, 10 g / mL) (n=4 donors; data are means+s.e.m.; dots represent image from one independent well; P values from unpaired / -tests, two-tailed with Welch’s correction).

[0045] Fig. 7 | Proposed mechanism for selective infiltration of CD8+ T cells into an environment of AD pathology and subsequent interaction with microglia leading to enhanced neurodegeneration and neuroinflammation. AD pathology leads to elevated levels of proinflammatory recruitment chemokines, e.g., CCL2 and CXCL10. T cells selectively infiltrate an environment of AD pathology in the neuroimmune axis model following binding of CXCL10 to its T cell receptor, CXCR3. The interaction of CCL2 with CCR2 induces recruitment of microglial cells to AD pathology. Infiltrating CD8+ T cells activates IFN / inflammatory pathways in microglia and astrocytes, resulting in a synergistic vicious cycle that dramatically exacerbates cellular damage and neuroinflammation in AD. Blocking the binding of CXCL10 to CD8+ T cell CXCR3 significantly attenuates selective infiltration of CD8+ T cells into an environment of AD pathology in the AD neuroimmune axis model and stops neurodegeneration.

[0046] Fig. 8 | Characterization of the PiChip model, a, 3D illustration of three different layers of the microfluidic-based PiChip model. Layer I consists of an orthogonal array of 400 microgrooves of 10 gm height and width and 500 gm length. Layer II consists of a BRAIN and four PERIPHERAL compartments of 100 pm height. Layer III consists of a cylindrical chamber with 6000 pm length and radius and 2500 pm open hollow. The top and bottom panels show low and high magnification of each layer, respectively, b, Quantification of soluble Ap38, Ap40, AP42, and AP42 / 40 ratio in 3D-differentiated neurons and astrocytes in AD and CTRL cultures after four weeks of differentiation (data are means+s.e.m. from three independent experiments; P values from unpaired / -test with Welch’s correction), c, Representative immunofluorescence staining of p-Tau (PHF-1; blue) in neuronal cell bodies and neurites (GFP; green) in 3D-differentiated neurons and astrocytes in AD and CTRL cultures after four weeks of differentiation. Scale bar, 50 pm. d, UMAP visualization of the cell-type composition of the PiChip cultures by single-cell transcriptome profiles. Distinct cell type (neurons, astrocytes, and microglia) are depicted with different colors, e, Dot plot of identity genes for distinct cell types. Color scale indicates the average of expression of identity genes in each cell population, and dot size is proportional to the percentage of cells expressing the identity genes, f, Volcano plot of differentially expressed genes in astrocytes in AD Neu / AC / iMGL versus AD Neu / AC lacking microglia. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots, g, Violin plot of highly enriched C3 gene in astrocytes in AD Neu / AC / iMGL compared to CTRL (P values from unpaired / -tests, two-tailed with Welch’s correction), h, Vioin plots of selected significantly enriched genes associated with reactive astrocytes (VIM and GFAP) in AD Neu / AC / iMGL compared to AD Neu / AC lacking microglia (P values from unpaired / -tests, two-tailed with Welch’s correction), i, Vioin plots of highly enriched genes associated with reactive astrocytes and interferon-g in astrocytes in AD Neu / AC / iMGL compared to AD Neu / AC lacking (**** <0.0001 from Wilcoxon test).

[0047] Fig. 9 | Chemotaxis of peripheral immune cells in the PiChip microfluidic platform, a, Schematic describes a chemotaxis setting in the PiChip microfluidic platform by loading a recombinant human chemokine of choice and human PBMCs in the CHEMOKINE and PERIPHERAL compartments, respectively. PBMC migrate through confined microchannels connecting the PERIPHERAL compartment to the CHEMOKINE compartment following the chemotactic gradient, b-d, Imaging flow cytometry and gating strategy of isolated human CD3+ T cells (b), monocytes (c), and B cells (d). Scale bar, 10 pm. e-j, Quantification of human PBMC migration through microchannels in the PiChip microfluidic platform, towards varying concentrations of different chemokines, including CXCL12, CCL2, and CCL19. Heatmaps depict the density of migrated CD3+ T cells (e), monocytes (g) and B cells (i) through microchannels as a function of chemokine concentration over 20 hours.

[0048] Quantification of the number of migrated CD3+ T cells (f), monocytes (h) and B cells (j) and their average velocity as a function of chemokine concentration towards CXCL12, CCL2, and CCL19, respectively (data are means±s.e.m from three independent experiments; P values from one-way ANOVA with Tukey multiple comparisons test).

[0049] Fig. 10 | An unbiased / automated image analysis pipeline to probe infiltrating peripheral immune cells in the PiChip system, a, Schematic represents a pipeline for quantitative analysis of human peripheral immune cell migration signatures (migration index and velocity) in the PiChip system using time-lapse confocal microscopy imaging and IMARIS software, b-c, Confocal imaging showing AD and CTRL cultures (Neu / AC / iMGL, green) with infiltrating CD8+ T cells (nuclei- stained red) at 17 hours post-loading of CD8+ T cells to the PiChip system. The yellow box denotes the BRAIN compartment, while the orange box represents the BRAIN-PERIPHERAL interface comprised of confined microchannels where infiltrating immune cells migrate from the PERIPHERAL compartment to the BRAIN compartment. Scale bar, 200 pm. Migration index is calculated based on the number of immune cells migrated across 100 microchannels into the BRAIN compartment (yellow box) of the PiChip system, while migration velocity is measured from migration movement through the microchannels connecting the PERIPHERAL compartment to the BRAIN compartment, d, Quantification of migration index of CD3+ T cells. T cells were isolated from de-identified donors and loaded into the PiChip system containing 3D-differentited ReNVM or ReN-mcherry. Data were extracted from confocal imaging at 15 hours post-addition of T cells to the PiChip system (w=4 donors). Fig. 11 | Increasing number of extravascular CD4+ and CD8+ T cells in 5XFAD brains. Seven-month-old female WT and 5XFAD mice were given PBS or CD45-FITC antibody iv. Five minutes later, the brains were dissected and homogenized by a tissue grinder without adding any enzymes, a, Representative flow cytometry plots illustrating the gating of CD45+CDllb' cells, CD45-FITC negative CD4+and CD8+T cells, b, Quantification of the percentages and absolute numbers of CD45+CDllb' non-myeloid cells in (a). FITC, Fluorescein isothiocyanate (data are mean±s.d. with individual mice as points; n=9-12 mice per group; P values from unpaired student’s t-tests). c, Representative flow cytometry plots illustrating the gating of CD4+and CD8+T cells among all CD45+extravascular cells, d, Quantification of the percentages of CD4+and CD8+T cells among all CD45+extravascular cells, (data are mean±s.d. with individual mice as points; n=9-12 mice per group; P values from unpaired student’s t-tests; ns, not significant).

[0050] Fig. 12 | Cytokine profiles in the PiChip system following infiltration of human CD3+ T cells, monocytes, or B cells, a, Dominant cytokine profiles 24 hours post infiltration of PBMCs using human cytokine 65-plex discovery assay, human complement panel 1 and 2-plex, and TGF-b discovery assay. The dominant cytokine signature induced by infiltration of monocytes, T cells, or B cells is characterized by consistently larger fold changes than other signatures in control and AD cells (+ / -) microglia. Scatter plot showing log2 -transformed cytokine concentrations (pg / ml) and their probability distribution. Scatter plots share a y-axis scaled from -10 to 10. Fold changes were log2 -transformed, b, Quantification of significantly elevated cytokines in the AD PiChip cultures before and after infiltration of monocytes, CD3+ T cells, or B cells (data are presented as mean; P values from two-way ANO VA with multiple comparison test).

[0051] Fig. 13 | Infiltrating T cells exacerbate neuronal damage in the presence of microglia in AD. a, Time-lapse confocal imaging showing neurite cleavage in AD PiChip culture. Microglia cells attack neuronal cells to cleave axon / neurite-like structures in an AD-like environment. Scale bar, 25 pm. b, Box plots represents quantification of neuronal damage using Tuj 1+ expressing surface area in AD Neu / AC (+ / -) iMGL condition in the presence of CD3+ T cells (w=4 donors; data are means+s.e.m.; dots represent image from one independent well; P values from one- way ANOVA). c, Violin plot represents quantification of calcium dynamics AD condition in the presence of T cells compared to CTRL, as monitored using Cal-520 AM, a Ca2+indicator. Two individual areas per well were captured in the same ROI of each well and used for analyses. Raw intensity values of all the active cells (>0 frequency per 4.5 minutes) in the captured videos were extracted using Image J time series analyzer V3, and relative changes in fluorescence intensity to baseline [(F - F0) / F0] were regarded as calcium signals.

[0052] Fig. 14 | Summary of scRNA-seq data, selected identity gene expression and gene ontology of significantly enriched genes, a, UMAP visualization of the cell-type composition of the PiChip cultures by single-cell transcriptome profiles. Distinct cell type (neurons, astrocytes, microglia, T cells, and proliferative) are depicted with different colors, b, Dot plot of identity genes for distinct cell types. Color scale indicates the average of expression of identity genes in each cell population, and dot size is proportional to the percentage of cells expressing the identity genes, c, Number of significantly upregulated (FDR<0.2 and FC>0) and downregulated (FDR<0.2 and FC<0) genes across distinct cell types. Significantly upregulated genes are depicted in red, while downregulated are in blue, d, Volcano plots of differentially expressed genes in all cell types in AD versus CTRL. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots, e, Gene ontology of significantly upregulated pathways in all cell types in AD versus CTRL. From left to right, the order of conditions is the same as in d. Color scale indicates the adjusted P value for significantly enriched pathways in each cell population, f, Gene ontology / network analysis of significantly upregulated genes ( <0.05, FDR 0.1 and FC>0.1). Top five category was visualized for associated genes and pathways, g, Gene ontology of significantly downregulated pathways in all cell types in AD versus CTRL. From left to right, the order of conditions is the same as in d. Color scale indicates the adjusted P value for significantly enriched pathways in each cell population, h, Gene ontology / network analysis of significantly downregulated genes ( <0.05, FDR 0.1 and FC<0.1). Top five category was visualized for associated genes and pathways.

[0053] Fig. 15 | Altered proportions of neurons, astrocytes and microglia in AD cultures and upon infiltration of T cells, a, UMAP visualization of the cell-type composition of the PiChip cultures by single-cell transcriptome profiles. Distinct cell type (neurons, astrocytes, microglia, T cells, and proliferative) are depicted with different colors, b, Dot plot of identity genes for distinct cell types. Color scale indicates the average of expression of identity genes in each cell population, and dot size is proportional to the percentage of cells expressing the identity genes, c, Compositional changes of neurons, astrocytes, microglia in AD Neu / AC / iMGL and CTRL d, Compositional changes of neurons, astrocytes, microglia and T cells in AD Neu / AC / iMGL (+ / -) T cell and CTRL conditions, e, UMAP visualization of T celltype composition of the PiChip cultures by single-cell transcriptome profiles. Distinct subtypes of T cells (homeostatic, DAM, MHCII, interferon and proliferative) are depicted with different colors, f, Compositional changes of homeostatic, DAM, MHCII, interferon and proliferative microglia in AD vs CTRL conditions.

[0054] Fig. 16 | Interferon-g associated pathways are enriched in glial cells following infiltration of T cells in AD. a, Volcano plot of differentially expressed genes in microglia in the presence of infiltrating T cells in AD cultures compared to AD cultures lacking infiltrating T cells. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots, b, Violin plots of highly enriched genes associated with cytokines and interferon-g in microglia in the presence of infiltrating T cells in AD compared to AD cultures lacking infiltrating T cells (Unpaired Ltests, two-tailed with Welch’s correction), c-d, Gene ontology of significantly upregulated (c) and downregulated (d) pathways in microglia in the presence of infiltrating T cells in AD cultures vs AD cultures lacking infiltrating T cells. Color scale indicates the adjusted P value for significantly enriched pathways in microglia, and dot size is proportional to the count of genes, e, Volcano plot of differentially expressed genes in astrocytes in the presence of microglia and infiltrating T cells in AD as compared to AD cultures lacking infiltrating T cells. Significantly upregulated ( <0.05 and FC>0.1) and downregulated ( <0.05 and FC<0.1) genes are shown in red dots, f, Violin plots of highly enriched genes associated with cytokines and interferon-g in astrocytes in the presence of microglia and infiltrating T cells in AD as compared to AD cultures lacking infiltrating T cells (Unpaired Ltests, two-tailed with Welch’s correction), g-h, Gene ontology of significantly upregulated (g) and downregulated (h) pathways in astrocytes in the presence of microglia and infiltrating T cells in AD as compared to AD cultures lacking infiltrating T cells. Color scale indicates the adjusted P value for significantly enriched pathways in microglia, and dot size is proportional to the count of genes.

[0055] Fig. 17 | NicheNet analysis of upstream ligand-receptor pairs inducing the DE genes of neurons, astrocytes, and microglia upon T cell infiltration, a, Dot plot of identity genes for distinct cell types. Color scale indicates the average of expression of identity genes in each cell population, and dot size is proportional to the percentage of cells expressing the identity genes, b, NicheNet’s ligand-target analysis represents potential upstream receptors expressed by neurons, astrocytes, microglia and T cells associated with top 20 potential ligands, c, Predicted target genes of the top 20 of potential ligands in AD Neu / AC / iMGL / T-cell vs CTRL.

[0056] Fig. 18 | Increased level of pro-inflammatory cytokines in the presence of CD8+ T cells and microglia in AD cultures, a, Quantification of IL- 15, IL-2, TFN-a, CXCL10 and granzyme B in the AD cultures (+ / -) CD8+ T cells and iMGL, as measured by MSD assay (w=10 donors, data are means±s.e.m. from seven independent experiments; P value from one-way ANOVA). b, Enrichment of major cytotoxic genes (GZMA, GZMB, GZMH, GZMM, and GZMK) in CD4+ naive, CD4+ memory, and CD8+ cytotoxic T cells in the AD cultures in the presence of microglia, c, Quantification of INF-g level in the AD Neu / AC (+ / -) iMGL and CD4+ or CD8+ T cells using ELISA assay (data are mean±s.e.m. from three independent experiments; P value from one-way ANOVA test).

[0057] Fig. 19 | Modulation of CXCL10 / CXCR3 signaling axis, a, A heatmap showing the elevated dynamics of proteins in the Neu / AC (+ / -) microglia cultures in the PiChip system. Protein concentrations (pg / ml) with no fraction digits are shown in the circles. Circles size represents protein concentration (0-1) in each row. b, Quantification of elevated growth factors and cytokines in the PiChip cultures in the presence and absence of microglia (data are mean±s.e.m.; P value from two-way ANOVA with Tukey multiple comparison tests), c, Time-lapse quantification of human CD3+ T cell migration through microchannels in the microfluidic system towards varying concentrations of CXCL10. Heatmap depicts the density of migrating CD3+ T cells through microchannels as a function of CXCL10 concentration, d, Quantification of the number of migrated CD3+ T cells isolated from de-identified donors in the PiChip system in response to CXCL10 chemokine at different concentrations without neuronal cells. (n=3 donors; circles represent replicates; P value from one-way ANOVA test), e, Quantification of average migration velocity of migrated CD3+ T cells isolated from de-identified donors in the PiChip system in response to CXCL10 at different concentrations (n=3 donors; circles represent replicates; P value from one-way ANOVA test), f, Representative immunofluorescence imaging showing CD8+ T cells, co-labeled with CXCR3 and DAPI in 6-7-month-old 5XFAD and WT mouse brains. CD8+ exhibits a magenta color, whereas CXCR3 exhibits a yellow. Scale bar, 20 pm. g, Quantification of migration indexinhibited (see Methods) of CD3+ T cells with and without CXCR3 neutralizing antibody (MAB160, 10 pg / mL) or CXCR3 antagonist, NBI74330 (10 pg / ml) treatment in the PiChip system containing AD or CTRL Neu / AC / iMGL culture. Data were extracted from confocal imaging at 15 hours post-addition of T cells to the PiChip systems (w=10 donors, data are means+s.e.m. from 4 independent PiChip systems per donor; circle=region of interest from 100 microchannels in one PiChip system; P values from one sample Ltests). h, Quantification of average migration velocity of infiltrating CD3+ T cells with and without MAB160 (10 pg / mL) or NB 174330 (10 pg / ml) treatment through the microchannels over 17 hours in the PiChip system (n=8 donors; data from 4 independent PiChip systems per donor; box- and-whiskers represent medians and the interquartile range with outliers outside whiskers; P values from two-way ANOVA with Tukey multiple comparisons tests).

[0058] DETAILED DESCRIPTION

[0059] Detailed studies regarding how infiltrating T cells impact AD neuropathogenesis have been limited by current disease models, which are unable to recapitulate human T cell infiltration with diverse genetic and histocompatibility complex (MHC) backgrounds20. Transgenic mouse models of AD do not fully recapitulate the complex human genetic backgrounds of AD patients or MHC context20 22. Human neural cellular models derived from induced pluripotent stem cells (iPSC) of patients with highly penetrant familial AD (FAD) mutations carry the potential to recreate complex human genetic backgrounds23. However, models using these iPSC-derived neurons do not fully emulate AD pathology, including extracellular amyloid-P (AP) plaques, neurofibrillary tangle (NFT) pathology, and neuroinflammation23’24. To overcome these limitations, we previously developed 3D human embryonic stem cell-derived neural cell culture models that demonstrate the neuropathological cascade of AD, including Ap deposition, Ap-induced hyperphosphorylated Tau (p-Tau) and NFT pathology25 27, the latter of which has not been possible with transgenic mice or iPSC-derived neuronal models of AD21’23. Using ReNcell VM human neural progenitor cells (NPCs) expressing human Ap protein (A4) precursor protein (APP) or APP / presenilin l(PSENl) genes with fully penetrant familial early-onset AD mutations, we demonstrated that the pathogenic Ap42-driven Ap oligomers drive neurofibrillary p-Tau / NFT pathology, requiring the relatively equal expression of both 3- and 4-repeat Tau isoforms in a 3D cell culture2527. To add a neuroinflammatory component to our 3D-AD cell culture models, we developed a 3D-AD tri culture model of human neurons, astrocytes, and microglia28. This allowed for studying the interaction between brain glial cells and neurons and recapitulated the roles of neuroinflammation and neurodegeneration following Ap and p-Tau / NFT pathology28. However, the 3D-AD triculture model still lacked other cell types contributing to AD neuropathogenesis - peripheral immune cells, e.g., T cells and monocytes.

[0060] 3D Human Neuroimmune Axis Model

[0061] Described herein is the development and validation of a 3D human neuroimmune axis model comprising human stem cell-derived neurons, astrocytes, microglia, and peripheral immune cells isolated from human blood. To date, limited attempts have been made to model the neuroimmune axis of AD to investigate braininfiltrating peripheral immune cells and their potential disease-modifying role in AD. The 3D neuroimmune axis model is comprised of human stem cell-derived neurons, astrocytes, microglia, and human peripheral immune cells in a microfluidic platform (one example of which is referred to herein as Peripheral immune Chip, PiChip). The PiChip system provides a unique multicellular model to study the interactions between human brain-infiltrating peripheral immune cells and brain-resident cells in AD at a single-cell level (Fig. 1). The platform comprises an open-top, multicompartment microfluidic chip in which different cell types of the central nervous system (CNS) and PBMCs isolated from human blood could be positioned side by side to allow live-cell confocal imaging (Fig. la-b). The microfluidic chip is comprised of a central BRAIN compartment (e.g., a square of 3000 x 3000 pm, with an inner radius of 2000 pm) with a large concentric cylinder reservoir (e.g., 6000 pm in length and radius and 2500 pm open inner radius), allowing accumulation of AD neuropathology and various soluble factors, along with four PERIPHERAL compartments supplying PBMCs (2500 pm inner radius) (Fig. lb; Fig. 8a). Further, the inner radius of the BRAIN compartment is smaller than that of the reservoir, which can be beneficial in reducing the generation of bubbles when adding a cell suspension to the BRAIN compartment. Although the height of the reservoir shown in Fig. 8 is 6000 pm, in practice the height can be different (e.g., any height that is less than the height of the plate). The PiChip system is designed to support a multicellular culture including human neurons (Neu), astrocytes (AC), iPSC-derived microglial cells (iMGL), and PBMCs in one single chip. The BRAIN and PERIPHERAL compartments are connected by linear laser-etched, narrow migration microchannels (each side, between 100-500 pm in length and between 5-10 pm in height and width, e.g., 10x 10x500 pm3in height, width, and length) (Fig. 8a), e.g., 100 channels per side, e.g., providing a gradient of soluble factors released by CNS cells in the BRAIN compartment. The migration microchannels do not contain cellular components of the neurovascular unit. However, the confined space partially mimics the infiltration of peripheral immune cells into the brain parenchyma in response to chemotactic gradients. The migration microchannels also allow monitoring of infiltrating peripheral immune cells from the PERIPHERAL compartment toward the 3D-AD neural-glial cultures in the BRAIN compartment (Fig 1c)

[0062] In at least some embodiments, microchannels having dimensions of 10x 10x500 pm3in height, width, and length may be particularly effective in allowing migration between the PERIPHERAL compartments to the BRAIN compartment in such a way that mimics the infiltration of peripheral immune cells into the brain parenchyma in response to chemotactic gradients. However, in some embodiments, the use of microchannels having different dimension may provide a similar benefit. For example, in some embodiments, one or more of the microchannels can have a length between 100-500 gm). As another example, one or more of the microchannels can have a width and height between 5-10 pm).

[0063] Further, in at least some embodiments, the use of 100 microchannels to connect each of the PERIPHERAL compartments to the BRAIN compartment is particularly effective in allowing migration between the PERIPHERAL compartments to the BRAIN compartment in such a way that mimics the infiltration of peripheral immune cells into the brain parenchyma in response to chemotactic gradients. However, in some embodiments, the use of different numbers of microchannels may also provide a similar benefit (e.g., 400 microchannels for each PERIPHERAL compartment), depending on the dimensions of the microchannels.

[0064] The design of the PiChip system is also amenable to a real-time exploration of infiltrating peripheral immune cells into 3D cell cultures (Fig. 2), cell -cell interactions, and their impact on neuroinflammation and neurodegeneration simultaneously (Figs. 3 and 4). Furthermore, the flexible and open design of the microfluidic-based PiChip platform, combined with automatic imaging, enables mid- to high-throughput screening of patient-derived peripheral immune cells and testing of candidate drugs.

[0065] The neuroimmune axis model has allowed us to investigate whether infiltrating peripheral immune cells play pathogenic roles in AD and to delineate the molecular events underlying infiltration and activation of human T cells, monocytes, and B cells in an AD brain-like microenvironment. Using this model, we showed a significantly higher percentage of infiltrating human CD4+ and CD8+ T cells (but not B cells) into AD neural-glial cell cultures independent of microglial cells. Our study supports the role of T cells in AD pathology, as suggested by the increased number of extravascular T cells (mainly CD8) in the hippocampus of AD patients6 1 1. Elevated numbers of CD4+ and CD8+ T cells have also been consistently reported in transgenic AD mouse brains12-18. Moreover, ablation of CD8+ T cells in transgenic AD mouse models led to altered expression of genes related to neuronal and synaptic function10. In addition to T cell infiltration data derived from our novel 3D neuroimmune axis model, we also investigated a transgenic AD mouse model and demonstrated a significantly increased percentage and higher absolute number of extravascular CD4+ and CD8+ T cells in the brains of 5XFAD mice. Our findings also agree with another report showing that increased numbers of parenchymal CD8+ T cells were positively correlated with advancing Braak stages of neuropathology in AD patients10. Overall, our observations from the PiChip system and 5XFAD mice model align with the recent finding of increased numbers of CD8+ T cells in AD patients and AD-like transgenic models. However, whether the infiltrating T cells play a disease-modifying role in AD neuropathogenesis has remained unclear.

[0066] We also discovered that the infiltration of human monocytes was also enhanced in AD cultures. However, the presence or absence of microglial cells negated this difference. Previous studies reported a similar phenomenon in which microglia depletion in mice increased infiltration of circulating monocytes50,51. However, what factors regulate recruitment monocytes in the CNS remains unknown. One mechanism has been proposed in which the absence of microglia causes the release of cytokines, resulting in vascular activation and monocyte entry into the CNS50. Future studies are needed to dissect the underlying mechanism by which microglial cells impact infiltration of circulating monocytes and identify the crucial chemotactic factors that might regulate this process.

[0067] A recent study has elegantly demonstrated increases in clonally expanded terminally differentiated effector memory (TEMRA) CD8+ T cells in CSF of AD brains11, suggesting that CD8+ T cells impact neurodegeneration and cognitive impairment in AD patients. CD8+ T cells were found close to MAP2+ neuronal structures and NEFH+ (Neurofilament Heavy Chain) neuronal processes in APP / PS brains11, which might influence neuronal cells and synaptic function. Our human neuroimmune axis model provides a new possible underlying mechanism for CD8+ T cell-induced synaptic damage observed in previous studies11,52. We discovered that infiltrating CD8+ T cells enhance activation of microglial cells and exacerbate neuroinflammation and neurodegeneration in the AD cultures. In particular, we demonstrated that 1) CD8+T cells alone induce significant neuronal and / or neuritic damage (not glial) in the AD cultures lacking microglia, 2) microglial cells alone have the capacity to induce cellular damage at the neuronal and glial level in AD cultures, 3) the coexistence of infiltrating CD8+ T cells and microglia in AD neural-glial cultures resulted in a synergistic role of these cells to cause dramatic neuronal and glial damage. Our data emphasized the contribution of infiltrating T cells in triggering INF / inflammatory-associated pathways in AD neural-glial cultures. In particular, infiltration of T cells enriched IFN-associated (e.g., IFITM1 and STAT1), antigenpresentation (e.g., MHC I and MHC II) and pro-inflammatory cytokines / chemokines (e.g., CXCL10 and IL-32) in both microglia and astrocytes in AD neural-glial cultures. Strikingly, we found that this effect was highly pronounced in the presence of microglia in AD cultures. Notably, while DAM population was increased in AD cultures compared to control lacking T cells, infiltration of T cells caused a substantial reduction in DAM population in AD cultures. In contrast, the interferon microglia largely increased following T-cell infiltration in AD. This indicates that both AD neuropathology and infiltrating T cells altered DAM and interferon gene expression and can be exploited to modulate microglial activation in neurodegenerative diseases. Through single-cell-type DGE analyses, we also identified significant upregulation of cytotoxic-enriched infiltrating T cells in the AD culture. In agreement with our PiChip system, scRNA-seq revelated that CD8+ T cells were significantly increased in AD compared to control. Intriguingly, the CD4+ naive / memory T cell ratio was 2-3 -fold higher than peripheral human blood (ratio -0.67), suggesting that CD4+ naive T cells preferentially infiltrated more than CD4+ memory T cells. It is unclear why CD4+ naive T cells migrate more than CD4+ memory T cells. Additionally, we observed a robust increase in proinflammatory cytokines (e.g., IL-2, IL-15 and IFN-y) upon infiltration of CD8+ T cells in AD cultures in the presence of microglia. These data are in line with previous reports in which activated CD8+ T cells are known to secrete proinflammatory cytokines, including IFN-y and IL-2 for astrocyte signaling, as well as granzyme B to induce neuronal apoptosis with astrocyte-derived IL- 15 in CNS post-traumatic injury53. Similarly, a recent scRNA-seq study identified that cytotoxic CD8+ T cells interact with microglia and oligodendrocytes through IFN signaling and contribute to white matter decay in aging54. Although the exact crosstalk between cytotoxic CD8+ T cells, microglia, and other brain cells is still yet to be established, this finding, along with the results described herein, underscores the critical role of IFN-y signaling and further strengthens the link between CD8+ T cells and IFNy in driving neurodegeneration. The human 3D neuroimmune axis model described herein can be employed to explore potentially detrimental or beneficial roles of various peripheral immune cells infiltrating the brain, by varying the cell types added to the model; for example, different subtypes of T cells or other lymphocytes including B cells as well as innate immune cells including neutrophils and monocytes. In addition, the 3D-differentiated neurons and astrocytes, and other cells, can be genetically engineered to express different transgenes, e.g., can be transfected with different isoforms of APOE, or with APPSL, and / or PSldE9, e.g., to evaluate the effect of gene therapy interventions or genomic variations on AD risk, development, and severity. The model can be contacted with various test compounds as well, and the effect of the test compounds on migration of T cells can be evaluated, to identify compounds that can decreased migration of T cells into model AD brains.

[0068] Targeting the CXCL10-CXCR3 Pathway in Alzheimer s ’ Disease (AD)

[0069] Chemokines are widely expressed in the brain and play an important role in neuroinflammatory processes, including regulating the migration and recruitment of T cells in neurodegenerative diseases like AD19,41,44’55. Several chemokines, including CCL2 (MCP-1), CCL5, CXCL8 (IL-8), CXCL10 (IP-10), and CX3CL1 (Fractalkine), are elevated in brains of AD patients55 58. Our 3D-AD neural-glial cell cultures specifically upregulate several proinflammatory chemokines, including CCL228, IL- 859, and CXCL10, similar to human AD brains. Previously, we demonstrated that the chemokine, CCL2, plays a crucial role in recruiting microglial cells into our AD triculture model28. Data derived from our current PiChip model demonstrate a significant increase in levels of the chemokine, CXCL10, in AD neural-glia cultures as compared to control, independent of microglia presence (Fig. 5). Previous studies also demonstrated that CXCL10 levels were consistently elevated in AD mouse models58,60,61. Along similar lines, we demonstrated a significantly elevated astrocytic CXCL10 chemokine in the cortex and hippocampus of 5XFAD mice. This also agrees with previous clinical data, in which CXCL10 is physiologically expressed in astrocytes in healthy brains but elevated in the brains of AD patients55 58.

[0070] Our findings also suggested a potential role for the CXCL10 receptor, CXCR3. We showed that infiltrating CD8+ T cells express the CXCR3 receptor in the 5XFAD mouse brains. Similarly, a recent study also has provided insight into the functional role of CXCR3 signaling on AD neuropathology in the APP / PS1 mouse model62. Other studies also showed an important role of the CXCL10-CXCR3 axis in the CD8+ T cell response during chronic infection63. However, until now, the potential role of the CXCL10-CXCR3 pathway in mediating the infiltration of T cells into AD brain has remained unknown. Here, we showed that inhibition of the CXCL10 binding to its T cell receptor, CXCR3, using a neutralizing antibody, blocked CD8+ T cell infiltration selectively into AD cultures. Moreover, blocking the CXCL10-CXCR3 signaling axis and infiltrating CD8+ T cells in the PiChip system significantly prevented neuronal damage in AD cultures (Fig. 6). These novel findings suggest therapeutic interventions for AD based on targeting the CXCL10-CXCR3 axis to reduce CD8+ T cell infiltration into the brains of AD patients.

[0071] Described herein are methods for treating one or more neurodegenerative diseases and symptoms and disorders associated therewith in a subject, e.g., a mammalian subject (e.g., a human or non-human veterinary subject, such as a cat, dog, rabbit, horse, goat, or pig). Generally, the methods include administering a therapeutically effective amount of one or more CXCR3 inhibitors (e.g., one or more anti-CXCR3 antibodies and / or CXCR3 antagonists), as described herein, to a subject who is in need of, or who has been determined to be in need of, such treatment. A subject who is in need of, or who has been determined to be in need of, the treatment methods of the present disclosure can be a subject who has been diagnosed with a neurodegenerative disease and / or a subject who shows one or more symptoms of a neurodegenerative disease. With their progression, neurodegenerative diseases can affect various aspects of a person’s life, including mobility and balance, movements, swallowing, bladder and bowel function, blood pressure fluctuation, sleep, breathing, heart function, memory and cognitive abilities, mood, and / or speech. Thus, in some embodiments, the present methods can be used for treating a neurodegenerative disease in a subject who shows impaired mobility and balance, impaired movements, impaired swallowing, impaired bladder and bowel function, blood pressure fluctuation, impaired sleep, impaired breathing, impaired heart function, memory and cognitive abilities, impaired mood, and / or impaired speech, as compared to a healthy control (e.g., a subject who has not been diagnosed with and / or who does not show one or more symptoms of a neurodegenerative disease). Neurodegenerative diseases encompass a wide range of conditions that result from progressive damage to cells and nervous system connections that are essential for mobility, coordination, strength, sensation, and cognition; the present methods can be used, e.g., to treat neurodegenerative diseases that are associated with CD8+ T cell infiltration into the brain. These diseases include, without limitation, Alzheimer’s disease and other memory disorders, Huntington’s disease, Parkinson’s disease, Frontotemporal lobar degeneration with tau (FTLD-tau), ALS / motor neuron disease, and multiple system atrophy, among others (see, e.g., Zeng et al., Neurobiol Dis. 2024 Apr: 193: 106461; Williams et al., Acta Neuropathol. 2020; 139(5): 855-874; Gonzalez and Pacheco, Journal of Neuroinflammation volume 11, Article number: 201 (2014); Luthi-Carter, Mol Then 2003 May;7(5 Pt l):569-70; Hartnell et al., Brain. 2024 Feb l;147(2):590-606). Thus, in some embodiments, the present disclosure provides methods for treating Alzheimer’s disease in a subject in need thereof, e.g., a subject who has been diagnosed with and / or shows one or more symptoms of Alzheimer’s disease.

[0072] Neurodegenerative diseases (e.g., Alzheimer's disease) can be diagnosed and / or assessed by several cognitive tests. These tests can be used to diagnose the disease, track its progression, and / or monitor the effectiveness of treatments. Some of the most commonly used cognitive tests include:

[0073] 1. Mini-Mental State Examination (MMSE): A brief questionnaire that assesses a person's memory, attention, language, and orientation.

[0074] 2. Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog): A more comprehensive test that assesses a wider range of cognitive abilities, including memory, language, attention, and problem-solving.

[0075] 3. Clinical Dementia Rating Scale (CDR): A structured interview and examination that assesses a person's cognitive and functional abilities and is used to stage the severity of dementia.

[0076] 4. Montreal Cognitive Assessment (MoCA): A brief questionnaire that assesses multiple cognitive domains, including memory, attention, executive function, and language.

[0077] 5. Neuropsychological tests: A set of tests that assess specific cognitive abilities, such as memory, attention, language, and visuospatial skills. These tests are often used in combination with other diagnostic tools, such as medical and neurological exams, imaging studies, and laboratory tests to help diagnose and / or manage neurodegenerative diseases (e.g., Alzheimer's disease). The specific tests used to diagnose and / or manage neurodegenerative diseases depend on the individual patient and the stage of their disease and can be verified using a Brain MRI.

[0078] Currently, there are no blood tests that can diagnose and / or manage neurodegenerative diseases (e.g., Alzheimer's disease) with certainty. However, some blood tests can be used in combination with other diagnostic tools to help identify individuals who are at an increased risk of developing the disease, or to rule out other possible causes of symptoms. Some blood tests that can be used in the evaluation of neurodegenerative diseases (e.g., Alzheimer's disease) include:

[0079] 1. Complete blood count (CBC): A test that measures the number and types of cells in the blood.

[0080] 2. Comprehensive metabolic panel (CMP): A test that measures levels of electrolytes, glucose, kidney and liver function markers, and other substances in the blood.

[0081] 3. Vitamin B12 and folate levels: Tests to measure levels of B12 and folate, which are important for brain function and the nervous system.

[0082] 4. Inflammatory markers: Tests that measure levels of cytokines and other markers of inflammation in the blood, which may play a role in the development of Alzheimer's disease.

[0083] 5. Apolipoprotein E (APOE) genotyping: A test to determine a person's APOE genotype, which can affect their risk of developing Alzheimer's disease. These tests are usually not sufficient on their own to diagnose and determine the progress during the treatment phase in neurodegenerative diseases (e.g., Alzheimer's disease), and a definitive diagnosis can only be made through a combination of clinical and laboratory evaluations, including imaging studies, neurological exams, and cognitive testing. A subject who is treated by the present methods can be diagnosed as having a neurodegenerative disease (e.g., Alzheimer's disease) by one or more of these tests. Additionally, or in the alternative, the effect of the treatment methods can be assessed using one or more these tests. As used in this context, to “treat” means to ameliorate at least one symptom of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in a subject. For example, treatment by the present methods can ameliorate at least one symptom of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in a subject, as compared to before treatment (or as compared to a subject with the neurodegenerative disease and / or conditions associated therewith who has not been treated by the present methods). Often, neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith results in impaired mobility and balance, impaired movements, impaired swallowing, impaired bladder and bowel function, blood pressure fluctuation, impaired sleep, impaired breathing, impaired heart function, memory and cognitive abilities, impaired mood, and / or impaired speech; thus, treatment by the present methods can result in improved mobility and balance, improved movements, improved swallowing, improved bladder and bowel function, stable blood pressure, improved sleep, improved breathing, improved heart function, improved memory and cognitive abilities, improved mood, and / or improved speech, as compared to before treatment (e.g., before the CXCR3 inhibitor(s) is administered to the subject) or as compared to a subject with neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith who has not been treated by the present methods.

[0084] Preventing Neurodegenerative Diseases

[0085] The methods of the present disclosure can be used for preventing (i.e., reducing the risk of occurrence or delaying the occurrence of) neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in a subject in need thereof. For example, the present methods can be useful for reducing the risk of occurrence or delaying the occurrence of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in a subject who is at risk for developing the neurodegenerative disease and / or conditions associated therewith. Risk factors for neurodegenerative disease (e.g., Alzheimer's disease) include, but are not limited to, older age, family history of neurodegenerative disease, genetic (hereditary) risk factors (e.g., mutations in one or more dominant genes, such as Amyloid precursor protein (APP), Presenilin-1 (PSEN-1), Presenilin-2 (PSEN-2), and apolipoprotein E (ApoE)), lifestyle risk factors (e.g., unmanaged chronic health issues (such as high blood pressure or hearing loss), physical inactivity, unhealthy diet, alcohol misuse, smoking, not getting enough sleep or not sleeping well, social isolation, and / or lack of mental stimulation), heart disease, diabetes, stroke, high blood pressure, high cholesterol, and head injury. Thus, in some embodiments, the present methods can be used for reducing the risk of occurrence or delaying the occurrence of neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in a subject who has older age, family history of neurodegenerative disease, genetic (hereditary) risk factors (e.g., mutations in one or more dominant genes, such as Amyloid precursor protein (APP), Presenilin-1 (PSEN- 1), Presenilin-2 (PSEN-2), and apolipoprotein E (ApoE)), lifestyle risk factors (e.g., unmanaged chronic health issues (such as high blood pressure or hearing loss), physical inactivity, unhealthy diet, alcohol misuse, smoking, not getting enough sleep or not sleeping well, social isolation, and / or lack of mental stimulation), heart disease, diabetes, stroke, high blood pressure, high cholesterol, and head injury. Risk factors for neurodegenerative disease (e.g., Alzheimer's disease) can also include exposure to one or more environmental factors, such as, one or more of heavy metals (e.g., aluminum, arsenic, cadmium, lead, and / or mercury), air pollutants (particulate matter), pesticides, metal-containing nanoparticles, and industrial chemicals. Thus, in some embodiments, the present methods can be used for reducing the risk of occurrence or delaying the occurrence of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in a subject who has been exposed to one or more of such environmental factors.

[0086] Treatment of a subject by the present method (e.g., administration of a therapeutically effective amount of one or more CXCR3 inhibitors) can reduce the risk of occurrence or delay the occurrence of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in the subject, as compared to a subject who is at risk for developing the neurodegenerative disease, but has not been treated by the present methods.

[0087] In some embodiments, a subject who can be treated using a method described herein has been diagnosed as having AD, and / or does not have proteasome-associated autoinflammatory syndromes (PRAAS), lung inflammation, cancer, CXCR3 Antibodies

[0088] Described herein are CXCR3 antibodies for use in the methods of the present disclosure. In some embodiments, anti-CXCR3 antibodies for use in the present methods include, without limitation, one or more of antibodies 4Ch, 4Hul, 4Hu2, 4Hu3, 4Hu4, 4Hu5, 4Hu6, 4Hu7, 4Hu8, 4Hu9, 4HulO, 4Hull, 4Hul2, 4Hul3, 4Hul4, 4Hul5, 4Hul6, 4Hul7, 4Hul8, 4Hul9, 12Ch, 12Hul, 12Hu2, 12Hu3, 82Ch, 82Hul, 82Hu2, 82Hu3, 135Ch, 135Hul, 135Hu2, 135Hu3, 53Ch, 53Hul, 53Hu2, 53Hu3, 53Hu4, 53Hu5, 53Hu6, 53Hu7, 53Hu8, 53Hu9, 53HulO, 53Hull, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, 53Hul8, 53Hul9, and 53Hu29, as described in WO 2013 / 109974, which is incorporated herein by reference in its entirety. In some embodiments, anti-CXCR3 antibodies for use in the present methods include, without limitation, one or more of antibodies B2, Cl 2, F7, 2 A3, 3 A3, 3A11, 3C11, 4D5, 32B12, 32D12, Hl, B7, C12, D12, El, G12, 3A5, 3B12, 4C3, B12, C4, F7, 2A3, 3A7, and 3A8, as described in US 2019 / 0060448, which is incorporated herein by reference in its entirety. In some embodiments, anti-CXCR3 antibodies for use in the present methods include, without limitation, one or more of antibodies 23, 24, 25, 26, 27, 29, 30, 31, 33, 34, 35, 37, 53, 53Ch, 53Hu01, 53HuO2, 53Hu03, 53HuO4, 53Hu05, 53Hu06, 53Hu07, 53Hu08, 53Hu09, 53Hu010, 53Hull, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, and 53Hul8, as described in WO 2018 / 119299, which is incorporated herein by reference in its entirety. In some embodiments, anti-CXCR3 antibodies for use in the present methods include, without limitation, one or more of Ultra-LEAF™ Purified anti-mouse CD 183 (CXCR3) Antibody (BIOLEGEND®) and Human CXCR3 Antibody (Catalog No. MAB160; R&D Systems). In some embodiments, anti-CXCR3 antibodies for use in the present methods are human or humanized.

[0089] In some embodiments, as an alternative to CXCR3 antibodies, the methods use bispecific antibodies or nanobodies that inhibit CXCR3 and at least one additional target, e.g., CXCL10 or CCR6, can be used. For example, CXCLlO-specific nanobodies (e.g., nanobody 3Nbl2 that blocks CXCL10- CXCR3 binding), are described in Sadeghian-Rizi et al. (Generation and Characterization of a Functional Nanobody Against Inflammatory Chemokine CXCL10, as a Novel Strategy for the Treatment of Multiple Sclerosis. CNS Neurol Disord Drug Targets. 2019, 18(2): 141- 148), which is incorporated herein by reference in its entirety. A fully humanized IgG like bispecific antibody that can target both human Thl (through CXCR3) and human Thl7 cells (through CCR6), can also be used, e.g., as described in Robert et al. (A fully humanized IgG-like bispecific antibody for effective dual targeting of CXCR3 and CCR6. PLoS One. 2017 Sep 5, 12(9):e0184278), which is incorporated herein by reference in its entirety. The antibodies can be, e.g., human or humanized; methods for making human or humanized antibodies are known in the art.

[0090] CXCR3 Antagonists

[0091] Also described herein are small molecule CXCR3 antagonists for use in the methods of the present disclosure. In some embodiments, CXCR3 antagonists for use in the present methods include ACT-777991, as described, e.g., in Christen et al. (Combination treatment of a novel CXCR3 antagonist ACT-777991 with an anti-CD3 antibody synergistically increases persistent remission in experimental models of type 1 diabetes. Clin Exp Immunol. 2023 Dec 12;214(2): 131-143), Boof et al. (Pharmacokinetics, pharmacodynamics and safety of the novel C-X-C chemokine receptor 3 antagonist ACT-777991 : Results from the first-in-human study in healthy adults. Br J Clin Pharmacol. 2024 Feb;90(2):588-599), and Meyer et al. (Discovery of Clinical Candidate ACT-777991, a Potent CXCR3 Antagonist for Antigen-Driven and Inflammatory Pathologies. J Med Chem. 2023 Mar 23;66(6):4179-4196), each of which is incorporated herein by reference in its entirety. In some embodiments, CXCR3 antagonists for use in the present methods include JN-2, as described, e.g., in Kim et al. (JN-2, a C-X-C motif chemokine receptor 3 antagonist, ameliorates arthritis progression in an animal model. Eur J Pharmacol. 2018 Mar 15;823 : 1 -10), which is incorporated herein by reference in its entirety. In some embodiments, CXCR3 antagonists for use in the present methods include ACT-672125, as described, e.g., in Caroff et al. (Design, Synthesis, and Pharmacological Evaluation of Benzimidazolo- thiazoles as Potent CXCR3 Antagonists with Therapeutic Potential in Autoimmune Diseases: Discovery of ACT-672125. J Med Chem. 2022 Sep 8;65(17): 11533-11549), which is incorporated herein by reference in its entirety. In some embodiments, CXCR3 antagonists for use in the present methods include one or more of (±)-AMG 487 (Catalog No. 4487; TOCRIS) and (±)-NBI 74330 (Catalog No. 4528; TOCRIS). In some embodiments, CXCR3 antagonists for use in the present methods include one or more N,N-Di substituted benzylamine derivatives, as described, e.g., in Bata et al. (Discovery of a novel series of CXCR3 antagonists. Multiparametric optimization of N,N-di substituted benzylamines. Bioorg Med Chem Lett. 2016 Nov 15;26(22): 5418- 5428), which is incorporated herein by reference in its entirety. In some embodiments, CXCR3 antagonists for use in the present methods include SCH 546738, as described, e.g., in Jenh et al. (A selective and potent CXCR3 antagonist SCH 546738 attenuates the development of autoimmune diseases and delays graft rejection. BMC Immunol. 2012 Jan 10; 13:2), which is incorporated herein by reference in its entirety. In some embodiments, CXCR3 antagonists for use in the present methods include NIBR2130, as described, e.g., in Christen et al. (Small molecule CXCR3 antagonist NIB R2130 has only a limited impact on type 1 diabetes in a virus-induced mouse model. Clin Exp Immunol. 2011 Sep; 165(3):318-28), which is incorporated herein by reference in its entirety. In some embodiments, CXCR3 antagonists for use in the present methods include AMG 487, as described, e.g., in Liu et al. (Optimization of a series of quinazolinone-derived antagonists of CXCR3. Bioorg Med Chem Lett. 2009 Sep 1 ; 19(17):5114-8), which is incorporated herein by reference in its entirety.

[0092] Dosage

[0093] An “effective amount” is an amount sufficient to effect beneficial or desired results. Thus, a therapeutic amount (or a therapeutically effective amount) is one that achieves the desired therapeutic effect. For example, a therapeutic amount (or a therapeutically effective amount) of CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) described herein can be the amount that when administered to a subject in need thereof, ameliorates at least one symptom of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in the subject, as compared to before treatment (or as compared to a subject with the neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith who has not been treated by the present methods). This amount can be the same or different from a prophylactically effective amount, which is an amount necessary to prevent onset of disease or disease symptoms. For example, a therapeutically effective amount of CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) described herein can be the amount that when administered to a subject in need thereof, reduces the risk of occurrence or delays the occurrence of a neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith in the subject, as compared to a subject who is at risk for developing the neurodegenerative disease and / or conditions associated therewith, but who has not been treated by the present methods.

[0094] An effective amount can be administered in one or more administrations, applications or dosages. A therapeutically effective amount of a CXCR3 inhibitor (e.g., anti-CXCR3 antibody and / or CXCR3 antagonist) depends on the particular inhibitor selected. Pharmaceutical formulations comprising the CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) can be administered one or more times per day, one or more times per week (including once every other day), or one or more times per month. The skilled artisan will appreciate that certain factors may influence the dosage and timing required to effectively treat a subject, including but not limited to the severity of the disease or disorder, previous treatments, the general health and / or age of the subject, and other diseases present. Moreover, treatment of a subject with a therapeutically effective amount of the CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) described herein can include a single treatment or a series of treatments. For example, pharmaceutical formulations comprising the CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) can be administered to a subject one or more times (e.g., twice, thrice, four times, five times, six times, seven times, eight times, or more) at hourly, daily, weekly, or monthly intervals.

[0095] Dosage, toxicity and therapeutic efficacy of the CXCR3 inhibitor(s) (e.g., anti- CXCR3 antibodies and / or CXCR3 antagonists) can be determined by standard pharmaceutical procedures in cell cultures or experimental animals, e.g., for determining the LD50 (the dose lethal to 50% of the population) and the ED50 (the dose therapeutically effective in 50% of the population). The dose ratio between toxic and therapeutic effects is the therapeutic index and it can be expressed as the ratio LD50 / ED50. CXCR3 inhibitors which exhibit high therapeutic indices are preferred. While CXCR3 inhibitors that exhibit toxic side effects may be used, care should be taken to design a delivery system that targets such CXCR3 inhibitors to the site of affected tissue in order to minimize potential damage to uninfected cells and, thereby, reduce side effects.

[0096] The data obtained from cell culture assays and animal studies can be used in formulating a range of dosage for use in humans. The dosage of such CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) lies preferably within a range of circulating concentrations that include the ED50 with little or no toxicity. The dosage may vary within this range depending upon the dosage form employed and the route of administration utilized. For any CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) used in the method of the invention, the therapeutically effective dose can be estimated initially from cell culture assays. A dose may be formulated in animal models to achieve a circulating plasma concentration range that includes the IC50 (i.e., the concentration of the test compound which achieves a half-maximal inhibition of symptoms) as determined in cell culture. Such information can be used to more accurately determine useful doses in humans. Levels in plasma may be measured, for example, by high performance liquid chromatography.

[0097] Pharmaceutical Compositions

[0098] When employed as pharmaceuticals, the CXCR3 inhibitor(s) (e.g., anti- CXCR3 antibodies and / or CXCR3 antagonists) provided herein can be administered in the form of pharmaceutical compositions.

[0099] These compositions can be prepared as described herein or elsewhere (see, e.g., Remington: The Science and Practice of Pharmacy, 21st ed., 2005), and can be administered by a variety of routes, but are preferably administered intravenously (IV). IV administration can be in the form of a single bolus dose, or may be, for example, by a continuous perfusion pump. In some embodiments, the pharmaceutical compositions provided herein are suitable for intravenous administration.

[0100] Also provided are pharmaceutical compositions which contain, as the active ingredient (or active agent), a CXCR3 inhibitor (e.g., anti-CXCR3 antibody and / or CXCR3 antagonist) described herein, in combination with one or more non-toxic pharmaceutically acceptable carriers (e.g., excipients). In making the compositions provided herein, the active ingredient can be typically mixed with an excipient, diluted by an excipient or enclosed within such a carrier in the form of, for example, a capsule, sachet, paper, or other container. When the excipient serves as a diluent, it can be a solid, semi-solid, or liquid material, which acts as a vehicle, carrier or medium for the active ingredient. Thus, the compositions can be in the form of suspensions, emulsions, sterile injectable solutions, and sterile packaged powders.

[0101] Some examples of suitable excipients include, without limitation, lactose, dextrose, sucrose, sorbitol, mannitol, starches, gum acacia, calcium phosphate, alginates, tragacanth, gelatin, calcium silicate, microcrystalline cellulose, polyvinylpyrrolidone, cellulose, water, syrup, and methyl cellulose. The formulations can additionally include, without limitation, lubricating agents such as talc, magnesium stearate, and mineral oil; wetting agents; emulsifiers and lubricants, such as sodium lauryl sulfate and magnesium stearate; suspending agents; preserving agents, such as methyl- and propylhydroxy-benzoates; or combinations thereof.

[0102] The active ingredient can be effective over a wide dosage range and is generally administered in a therapeutically effective amount. It will be understood, however, that the amount of the active ingredient (e.g., CXCR3 inhibitor (e.g., anti- CXCR3 antibody and / or CXCR3 antagonist) described herein) actually administered will usually be determined by a physician, according to the relevant circumstances, including the condition to be treated, the chosen route of administration, the actual compound administered, the age, weight, and response of the individual subject, the severity of the subject’s symptoms, and the like.

[0103] Formulations of the compositions of the present disclosure can include those suitable for IV administration. The formulations may conveniently be presented in unit dosage form and may be prepared by any methods well known in the art of pharmacy. The amount of active ingredient (e.g., CXCR3 inhibitor (e.g., anti-CXCR3 antibody and / or CXCR3 antagonist) described herein) which can be combined with a carrier material to produce a single dosage form can vary depending upon the host being treated and the particular mode of administration, e.g., intravenous.

[0104] Aqueous suspensions can contain an active ingredient (e.g., CXCR3 inhibitor (e.g., anti-CXCR3 antibody and / or CXCR3 antagonist) described herein) in admixture with excipients suitable for the manufacture of aqueous suspensions, e.g., for aqueous intradermal injections. Such excipients include a suspending agent, such as sodium carboxymethylcellulose, methylcellulose, hydroxypropylmethylcellulose, sodium alginate, polyvinylpyrrolidone, gum tragacanth and gum acacia, and dispersing or wetting agents such as a naturally occurring phosphatide (e.g., lecithin), a condensation product of an alkylene oxide with a fatty acid (e.g., polyoxyethylene stearate), a condensation product of ethylene oxide with a long chain aliphatic alcohol (e.g., heptadecaethylene oxycetanol), a condensation product of ethylene oxide with a partial ester derived from a fatty acid and a hexitol (e.g., polyoxyethylene sorbitol mono-oleate), or a condensation product of ethylene oxide with a partial ester derived from fatty acid and a hexitol anhydride (e.g., polyoxyethylene sorbitan mono-oleate). The aqueous suspension can also contain one or more preservatives such as ethyl or n- propyl p-hydroxybenzoate, one or more coloring agents, one or more flavoring agents and one or more sweetening agents, such as sucrose, aspartame or saccharin. Formulations can be adjusted for osmolarity.

[0105] In some embodiments, oil -based pharmaceuticals are used for administration of CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists) of the present disclosure. Oil-based suspensions can be formulated by suspending an active agent in a vegetable oil, such as arachis oil, olive oil, sesame oil or coconut oil, or in a mineral oil such as liquid paraffin; or a mixture of these. As an example of an injectable oil vehicle, see Minto (1997) J. Pharmacol. Exp. Then 281 :93-102.

[0106] Pharmaceutical formulations can also be in the form of oil-in-water emulsions. The oily phase can be a vegetable oil or a mineral oil, described above, or a mixture of these. Suitable emulsifying agents include naturally-occurring gums, such as gum acacia and gum tragacanth, naturally occurring phosphatides, such as soybean lecithin, esters or partial esters derived from fatty acids and hexitol anhydrides, such as sorbitan mono-oleate, and condensation products of these partial esters with ethylene oxide, such as polyoxyethylene sorbitan mono-oleate. Injectable oil-in-water emulsions can comprise a paraffin oil, a sorbitan monooleate, an ethoxylated sorbitan monooleate and / or an ethoxylated sorbitan trioleate.

[0107] In some embodiments, pharmaceutical formulations of the present disclosure are parenterally administered, such as by intravenous (IV) administration or administration into a body cavity or lumen of an organ. These formulations can comprise an active agent (e.g., an anti-CXCR3 antibody and / or CXCR3 antagonist) described herein) dissolved in a pharmaceutically acceptable carrier. Acceptable vehicles and solvents that can be employed are water and Ringer's solution, an isotonic sodium chloride. In addition, sterile fixed oils can be employed as a solvent or suspending medium. For this purpose, any bland fixed oil can be employed including synthetic mono- or diglycerides. In addition, fatty acids such as oleic acid can likewise be used in the preparation of injectables. These solutions are sterile and generally free of undesirable matter. These formulations may be sterilized by conventional, well known sterilization techniques. The formulations may contain pharmaceutically acceptable auxiliary substances as required to approximate physiological conditions such as pH adjusting and buffering agents, toxicity adjusting agents, e.g., sodium acetate, sodium chloride, potassium chloride, calcium chloride, sodium lactate and the like. The concentration of active agent in these formulations can vary widely, and will be selected primarily based on fluid volumes, viscosities, body weight, and the like, in accordance with the particular mode of administration selected and the patient's needs. For IV administration, the formulation can be a sterile injectable preparation, such as a sterile injectable aqueous or oleaginous suspension. This suspension can be formulated using those suitable dispersing or wetting agents and suspending agents. The sterile injectable preparation can also be a suspension in a nontoxic parenterally-acceptable diluent or solvent, such as a solution of 1,3 -butanediol. The administration can be by bolus or continuous infusion (e.g., substantially uninterrupted introduction into a blood vessel for a specified period of time).

[0108] In some embodiments, the pharmaceutical compositions and formulations are lyophilized. Stable lyophilized formulations comprising an active ingredient (e.g., anti-CXCR3 antibody and / or CXCR3 antagonist described herein) can be made by lyophilizing a solution comprising a pharmaceutical formulation of the invention and a bulking agent, e.g., mannitol, trehalose, raffinose, and sucrose or mixtures thereof.

[0109] The compositions and formulations can be delivered by the use of liposomes. By using liposomes, particularly where the liposome surface carries ligands specific for target cells, or are otherwise preferentially directed to a specific organ, one can focus the delivery of the active agent into target cells in vivo. See, e.g., U.S. Patent Nos. 6,063,400; 6,007,839; Al-Muhammed (1996) J. Microencapsul. 13:293-306; Chonn (1995) Curr. Opin. Biotechnol. 6:698-708; Ostro (1989) Am. J. Hosp. Pharm. 46: 1576-1587. As used in the present disclosure, the term "liposome" means a vesicle composed of amphiphilic lipids arranged in a bilayer or bilayers. Liposomes are unilamellar or multilamellar vesicles that have a membrane formed from a lipophilic material and an aqueous interior that contains the composition to be delivered. Cationic liposomes are positively charged liposomes that are believed to interact with negatively charged DNA molecules to form a stable complex. Liposomes that are pH- sensitive or negatively-charged are believed to entrap DNA rather than complex with it. Both cationic and noncationic liposomes have been used to deliver DNA to cells.

[0110] Liposomes can also include "sterically stabilized" liposomes, i.e., liposomes comprising one or more specialized lipids. When incorporated into liposomes, these specialized lipids result in liposomes with enhanced circulation lifetimes relative to liposomes lacking such specialized lipids. Examples of sterically stabilized liposomes are those in which part of the vesicle-forming lipid portion of the liposome comprises one or more glycolipids or is derivatized with one or more hydrophilic polymers, such as a polyethylene glycol (PEG) moiety. Liposomes and their uses are further described in U.S. Pat. No. 6,287,860.

[0111] The formulations of the present disclosure can be administered for prophylactic and / or therapeutic treatments. In some embodiments, for therapeutic applications, pharmaceutical formulations of the present disclosure are administered to a subject who is in need of treatment for a neurodegenerative disease (e.g., Alzheimer's disease) and / or condition associated therewith, in an amount sufficient to cure, alleviate or partially arrest the clinical manifestations of the neurodegenerative disease and / or condition associated therewith; this can be called a therapeutically effective amount. Additionally, or in the alternative, for prophylactic applications, pharmaceutical formulations of the present disclosure can be administered to a subject who is at risk for developing a neurodegenerative disease (e.g., Alzheimer's disease) and / or condition associated therewith, in an amount sufficient to reduce the risk of occurrence or delaying the occurrence of the neurodegenerative disease and / or condition associated therewith; and this can be called a therapeutically effective amount.

[0112] The amount of pharmaceutical composition adequate to accomplish this is a therapeutically effective dose. The dosage schedule and amounts effective for this use, i.e., the dosing regimen, will depend upon a variety of factors, including the stage of the disease or condition, the severity of the disease or condition, the general state of the patient's health, the patient’s physical status, age and the like. In calculating the dosage regimen for a patient, the mode of administration also is taken into consideration.

[0113] The dosage regimen also takes into consideration pharmacokinetics parameters well known in the art, i.e., the active agents’ rate of absorption, bioavailability, metabolism, clearance, and the like (see, e.g., Hidalgo- Aragones (1996) J. Steroid Biochem. Mol. Biol. 58:611-617; Groning (1996) Pharmazie 51 :337-341; Fotherby (1996) Contraception 54:59-69; Johnson (1995) J. Pharm. Sci. 84: 1144-1146; Rohatagi (1995) Pharmazie 50:610-613; Brophy (1983) Eur. J. Clin. Pharmacol. 24: 103-108; Remington: The Science and Practice of Pharmacy, 21st ed., 2005). The state of the art allows the clinician to determine the dosage regimen for each individual patient, active agent and disease or condition treated. Guidelines provided for similar compositions used as pharmaceuticals can be used as guidance to determine the dosage regiment, i.e., dose schedule and dosage levels, administered practicing the methods of the invention are correct and appropriate.

[0114] Single or multiple administrations of formulations can be given depending on for example: the dosage and frequency as required and tolerated by the patient, the degree and amount of therapeutic effect generated after each administration, and the like. The formulations should provide a sufficient quantity of active agent to effectively treat, prevent or ameliorate conditions, diseases or symptoms. In some embodiments, pharmaceutical formulations for oral administration are in a daily amount of between about 1 to 100 or more mg per kilogram of body weight per day. Lower dosages can be used, in contrast to administration orally, into the blood stream, into a body cavity or into a lumen of an organ. Substantially higher dosages can be used in topical or oral administration or administering by powders, spray or inhalation. Actual methods for preparing parenterally or non-parenterally administrable formulations will be known or apparent to those skilled in the art and are described in more detail in such publications as Remington: The Science and Practice of Pharmacy, 21st ed., 2005. Combination Treatments

[0115] The methods and compositions described herein can include the use of standard treatments in addition to the CXCR3 inhibitor(s) (e.g., anti-CXCR3 antibodies and / or CXCR3 antagonists). Treatments for neurodegenerative disease (e.g., Alzheimer's disease) and / or conditions associated therewith are known in the art and include, by way of non-limiting example, one or more cholinesterase inhibitors, such as acetylcholinesterase inhibitors (AChEIs) (e.g., tacrine, donepezil, rivastigmine, and galantamine); and one or more antagonists of N-methyl d-aspartate (NMD A) antagonists NMD AR (e.g., memantine).

[0116] EXAMPLES

[0117] The invention is further described in the following examples, which do not limit the scope of the invention described in the claims.

[0118] Methods

[0119] The following materials and methods were used in the Examples below.

[0120] Neural progenitor cell (NPC) line, media, and reagents

[0121] A familial Alzheimer’s disease (FAD) cell line that exhibits substantial amyloid pathology was engineered using ReNcell VM human neural progenitor (Millipore Sigma, Catalog #SCC008) according to previous reports18,19. Briefly, the immortalized human neural progenitor cell line ReNcell VM (ReN) was transfected with internal ribosome entry site (IRES)-mediated polycistronic lentiviral vectors containing FAD-related genes encoding human amyloid precursor protein (APP) with both the K670N / M671L (Swedish) and V717I (London) mutations (APPSL), presenilin 1 (PSEN1) with the AE9 mutation (PSEN1(AE9)) with both GFP and mCherry as reporters for viral infection. Fluorescence-activated cell sorting (FACS) was then used to enrich the population of cells with the highest expression levels. The ReN cells were cultured in Matrigel-coated flasks (1 :100 dilution of MatrigekDMEM- F12 medium) using ReN cell proliferation medium.

[0122] ReN cell proliferation medium was prepared by combining 484.5 mL DMEM / F12 medium (Gibco, Catalog #11320-033) with 10 mL B27 (Gibco, Catalog #17504-044), 500 pL heparin (STEMCELL Technologies; Catalog #07980), 5 mL of lOOx penicillin / streptomycin / amphotericin B (Lonza, Catalog #17-745E), 500 pL epidermal growth factor (Millipore Sigma, Catalog #E9644, 20 pg / mL), and 400 pL basic fibroblast growth factor (Stemgent, Catalog #03-0002, 25 pg / mL). ReN cell differentiation medium was prepared by combining 484.5 mL DMEM / F12 medium with 10 mL B27, 500 pl heparin, and 5 ml of lOOx penicillin / streptomycin / amphotericin B. iPSC-derived human microglia (iMGL) iPSC cell acquired from NINDS Human Cell and Data Repository. NHCDR ID - ND41865, Cell line ID - NN0003920 and Lot numbers - 11335135. Male, 64 years old, no disease, ethnicity not reported, APOE3 / 3 and no reported TREM2 mutation. Human iPSCs were differentiated into iMGL following the protocol previously published30. iMGL differentiation base medium consisted of DMEM / F12 medium, 2x B27, 0.4x N2 (Gibco, Catalog #17502-048), 2x insulin-transferrin- selenium (Gibco, Catalog #41400-045), lx GlutaMAX (Gibco, Catalog #35050-061), lx non-essential amino acids (Gibco, Catalog #11140-50), lx penicillin / streptomycin (Gibco, Catalog #15140122), 5.25 pg / mL human insulin solution (Millipore Sigma, Catalog #19278), 462 pM monothioglycerol (Millipore Sigma, Catalog #M6145-25), and 0.05% w / v bovine serum albumin (Millipore Sigma, Catalog #A9418). The medium was filtered through a 0.22 pm medium filter after all base medium components were added. The three-cytokine medium consisted of iMGL base medium with 100 ng / mL macrophage colony-stimulating factor (M-CSF) (Fisher Scientific, Catalog #PHC9501), 100 ng / mL IL-34 (PeproTech, Catalog #200-34-10), and 100 pM IDE1 (PeproTech, Catalog #1164899). The five-cytokine medium consisted of iMGL base medium with 100 ng / mL M-CSF, 100 ng / mL IL-34, 100 pM IDE1, 100 ng / mL CD200 (Life Technologies, Catalog #A42571), and 100 ng / mL CX3CL1 (PeproTech, Catalog #300-31-250).

[0123] CD34+ iPSC-derived hematopoietic progenitor cells (iHPCs) were plated at a density of 2X 106cells into 100 mm Matrigel-coated dishes (1 : 100 dilution of Matrigel:DMEM / F12 medium) with 6 mL of three-cytokine medium. Cytokines were frozen in single-use aliquots and added to the iMGL base medium immediately before medium use. Every other day, 6 mL of three-cytokine medium was added to each dish of iHPCs. On day 11 of iHPC-iMGL differentiation, all conditioned medium from the dish was collected and centrifuged at 500 x G for 5 min at room temperature. The collected, non-adherent cells were resuspended in 6 mL of conditioned medium and added back into the dish with an additional fresh 6 mL of three-cytokine medium. Every other day until day 25 of iHPC-iMGL differentiation, 6 mL of three-cytokine medium was added to the dish. On day 25, all conditioned medium was collected and centrifuged at 500 x G for 5 min at room temperature. The supernatant was discarded, and the non-adherent cells were resuspended in 12 mL of five-cytokine medium and added back to the dish. Cells were fed every other day until day 28 of iHPC-iMGL differentiation with the five-cytokine medium followed by maintaining iMGL in culture or freezing using Bambanker (GC Lymphotec, Catalog #CS-02-001) medium. iMGL that remained in culture were given a five-cytokine medium every four days and used for two weeks of culture or until displaying an activated morphology.

[0124] Microfluidic chip design and fabrication

[0125] The microfluidic chip was designed using AutoCAD software and transferred on chrome photomasks. Negative photoresists SU-8 (MicroChem) were patterned using photolithography techniques on a silicon wafer. Briefly, the first (10 pm) and second (100 pm) layers were sequentially spin-coated on a clean silicon wafer and patterned by UV exposure through the photomasks. The wafer was then used as a master mold for silicone elastomer poly dimethylsiloxane (PDMS, Sylgard 182, Elsworth Adhesives) microfluidic devices. The PDMS array was molded by casting the liquid prepolymer composed of a mixture of 10: 1 silicone elastomer and a curing agent. The mixture was cured at 75 °C overnight, and then the PDMS mold was peeled off the silicon wafer. Brain and peripheral immune cell inlets were punched using a 2 -mm, and 2.5-mm punch (Harris Uni-Core™), respectively. Individual devices were punched out of the PDMS layer using a 12-mm punch. The molded PDMS devices were bonded irreversibly after plasma treatment to 6-, 12-, or 24-well glass bottom plates (MatTek Corporation). The cylindrical chamber was prepared by casting and curing a mixture of 10: 1 silicone elastomer and a curing agent in a petri dish overnight at 75 °C. The PDMS mold was peeled off, and central inlets were punched using a 2.5-mm punch and concentric cylinders were then punched out using a 6-mm punch. The cylinder was then plasma treated and bonded irreversibly on top of the BRAIN compartment of each microfluidic device. Neuroimmune axis model

[0126] The microfluidic devices were first plasma treated and then coated with 1 : 100 dilution of Matrigel:DMEM-F12 medium using prechilled gel-loading tips into the cell reservoirs, and incubated for at least 4 hours at 37 °C incubator. ReN cell differentiation medium was then added to the well containing the device until the media submerged only the bottom layer of the device. Next, ReN cells (AD or CTRL) mixed with Matrigel in a 1 :5 dilution of MatrigekReN cell differentiation medium were transferred into the central brain cell reservoir of the microfluidic chip at 5.0 X 104cells per 5 pL using a prechilled gel -loading tip. The 3D cell culture was maintained for 4 weeks, adding fresh differentiation media each week. The 3D cell cultures were given 1 mL of differentiation medium for the first week of growth, followed by 450 pL of fresh differentiation medium each week until the end of the third week of culture. In the final week of culture, 250 pL of fresh differentiation medium was added to the 3D cell culture. One day before adding PBMCs to the peripheral compartment, iMGL at a concentration of 5.0 x 103were added to the central brain compartment of the existing 3D ReN cell culture.

[0127] Isolation of human PBMCs

[0128] CD3+ / CD4+ / CD8+ T cells, monocytes, and B cells were isolated directly from fresh human whole blood using immunomagnetic negative selection (STEMCELL Technologies; EasySep™ Direct Human T Cell Isolation Kit, Catalog #19661; EasySep™ Direct Human CD4+ T Cell Isolation Kit, Catalog #19662; EasySep™ Direct Human CD8+ T Cell Isolation Kit, Catalog #19663; EasySep™ Direct Human Monocyte Isolation Kit, Catalog #19669; EasySep™ Direct Human B Cell Isolation Kit, Catalog #19674) according to the manufacturer’s protocols. All donor blood specimens were acquired from healthy donors and handled according to the protocols approved by the Massachusetts General Hospital Institutional Review Board. Peripheral blood from healthy individuals aged 18 years or older was drawn in 10 mL tubes containing heparin or 5 mM EDTA (Vacutainer; Becton Dickinson) and delivered on ice within 1 to 3 hours after collection (Research Blood Components LLC). Whole blood was processed immediately upon delivery, and PBMCs were carefully handled using sterile techniques throughout the isolation process. Isolated cells were stained with either Hoechst (ThermoFisher Scientific, Catalog #62249) or Cell Tracker Orange CMRA (Invitrogen, Catalog #C34551), washed and suspended in Iscove's Modified Dulbecco's Medium (IMDM) with 4 mM L-glutamine (Gibco, Catalog #12440-053) and supplemented with 20% fetal bovine serum (FBS) (Gibco, Catalog #1600-069) prior to loading into the microfluidics. Next, PERIPHERAL compartments of the matured neuroimmune axis model were loaded with 5.0 X 104cells in 5 pL of IMDM with 20% FBS for each donor and imaged immediately.

[0129] Time-lapse imaging of PBMC migration in the neuroimmune axis model

[0130] A fully automated Nikon Eclipse Ti2 microscope was used to acquire timelapse images of the microfluidic chips of the neuroimmune axis model loaded with PBMCs and AD or CTRL cultures. A glass-bottom plate was placed in a bio-chamber heated to 37 °C with 5% CO2 and confocal images of microchannels and the BRAIN compartment within each microfluidic chip were acquired every 8.5 minutes using a 10X objective for at least 18 hours. PBMCs were monitored using IMARIS software (Oxford Instruments, Version 9.8.0).

[0131] Data processing and analysis of PBMC migration index and velocity

[0132] Migration index: To quantitatively analyze the migration index and velocity of PBMCs, we developed an unbiased and automated image analysis pipeline for high-throughput quantification of the number of infiltrating PBMCs and their corresponding velocities in the microfluidic platform utilizing live time-lapse confocal imaging and IMARIS software. The Spots function was used to quantify the total number of PBMCs that migrated over time. A 600x1024 pixel region of interest (ROI) with an estimated spot XY diameter of 9 pm was used to detect PBMC. Then, an automatic spot quality threshold and track duration was used to filter detected spots. Lastly, the total number of cells migrated over time was extracted from IMARIS software and data at 15-hours post-loading were used to compute the migration index. The migration index of peripheral immune cells for each donor was calculated as follows:

[0133] Migration index

[0134] Number of cells migrated into the AD BRAIN compartment Average number of cells migrated into the CTRL BRAIN compartment

[0135] Velocity: The Spots function was also used to quantify the speed of the PBMC that migrated through the PERIPHERAL-BRAIN microchannels in microfluidic devices. A 400x1024 pixel ROI with an estimated spot XY diameter of 9 pm was used to detect peripheral immune cell migratory velocity within the channels of the microfluidic device. An automatic spot quality threshold and track duration was used to filter detected spots. For tracking PBMCs, a maximum gap distance of three was used for all peripheral immune cell types. The maximum distance tracking parameter for CD3+ / CD4+ / CD8+ T cells, monocytes, and B cells, was 70 mm, 25 mm, and 75 mm, respectively. The average velocity over 18 hours was then recorded for each image. Images were excluded from analysis if ReN cell growth blocked more than 70% of the PERIPHERAL-BRAIN microchannels in microfluidic devices, preventing PBMC movement. Rarely, data that display large fluctuations in the number of cells in a short amount of time, which suggested inaccurate IMARIS image processing, were excluded as well.

[0136] Chemotaxis of human PBMCs

[0137] The microfluidic devices were plasma treated and coated with 1 : 100 dilution of Matrigel:DMEM-F12 medium using prechilled gel-loading tips into the central and peripheral compartments and incubated for 4 hours at 37 °C. To prime the microfluidic devices, 20 minutes before PBMCs were introduced, the BRAIN compartment was filled with 10 pL of the chemoattractant of interest (R&D Systems; monocyte chemotactic protein (MCP)-1 / CCL2, Catalog #279-MC-010; C-X-C motif chemokine ligand 10 (CXCL10), Catalog #266-IP-010; C-X-C motif chemokine ligand 12 (CXCL12), Catalog #350-NS-010 / CF). The wells containing microfluidic chips were then filled with IMDM with 10% FBS medium until the media reached the top level of the cylinder without submerging the device. Isolated CD3+ T cells, monocytes, and B cells from whole blood were loaded by using a gel-loading tip at 5.0 X 104cells in 10 pL of IMDM with 10% FBS per each peripheral compartment. Loaded PBMCs were recorded using time-lapse imaging on a fully automated Nikon TiE microscope with a bio-chamber heated to 37 °C, and 5% CO2. Images were acquired every 7 minutes for 15-17 hours, using a 10X objective. A custom algorithm was then created to map each individual migration trajectory and determine the intensity of migrated cells, as well as the cell migratory velocity. Using these data, a “heat map” was generated in MATLAB (MathWorks, Natick, MA), which provides an analysis of the total migrated cells and their velocity in each condition. Imaging flow cytometry

[0138] Purity of isolated human PBMCs was assessed using the ImageStreamxMark II imaging flow cytometer (Amnis Corporation) equipped with 40x objective, 6 imaging channels, and 405, 488, and 462 nm lasers following cell isolation. Isolated cells were diluted in RPMI 1640 medium supplemented with 10 mM HEPES (Life Technologies, Catalog #31800105) and stained with the following fluorochrome- conjugated antibodies purchased from STEMCELL Technologies where applicable: anti-human CD45 antibody, (Clone HI30, Catalog #60018), anti-human CD 14 antibody (Clone M5E2, Catalog #60004), anti-human CD3 antibody (Clone UCHT1, Catalog #60011), anti-human CD19 antibody (Clone HIB19, Catalog #60005). PBMC purity was assessed on singlets, CD45+ cells to exclude debris, platelets, and red blood cells. 5 x 105CD45+ events were collected for each sample and analyzed using the IDEAS software (version 6.2).

[0139] Quantitative analysis of A levels

[0140] Levels of Ap38, Ap40 and AP42 in media collected from the BRAIN compartment were simultaneously measured by a multi-array electrochemiluminescence assay kit (Meso Scale Diagnostics (MSD), V-PLEX Ap Peptide Panel 1 (6E10) kit, Catalog #K15200E-2). To quantify levels of Ap, media was collected and then frozen at -20 °C. Levels of Ap38, Ap40, and Ap42 in thawed media samples were measured without dilution using the assay kit according to the manufacturer’s protocol.

[0141] Quantification of brain-localized CD4+ and CD8+ T cells in 5XFAD mice

[0142] Seven-month-old female WT and 5XFAD mice (n=9-12 per group) were anesthetized with 3-4% isoflurane and 100 pL of CD45-FITC antibody (BioLegend, Clone 30-F11, Catalog #103107) 10 pg / mouse diluted in sterile PBS was injected into the retro-orbital sinus. After 5 minutes of circulation, mice were perfused transcardially with 25 mL of cold PBS. Brains were harvested, and the olfactory bulbs and cerebellum were removed. The remaining brain tissues were minced with blades and homogenized by tissue grinder in Hank's Buffered Saline Solution (HBSS) (Lonza, Catalog #BE10-527F) without addition of any enzymes. 30% percoll gradient centrifugation (1,600 rpm, 24 minutes, without breaks) was used to remove myelin and other debris. Live cells were counted by trypan blue. Brain cells were blocked with Fc receptor blocking antibody (Biolegend, Clone 93, Catalog #101330) in FACS buffer (0.5% BSA, 2 mM EDTA, and 0.02% sodium azide in PBS) for 5 minutes at 4 °C before surface staining (30 minutes at 4 °C). The following anti-mouse antibodies were used for surface staining: PE anti-mouse CD45.2 (BioLegend, Clone 104, Catalog #109807), APC / Cyanine7 anti-CDl lb (BioLegend, Clone MI / 70, Catalog #101225), PE / Cyanine7 anti-mouse CD4 (BioLegend, Clone GK1.5, Catalog #100421), and PerCP / Cyanine5.5 anti-CD8a (BioLegend, Clone 53-6.7, Catalog #100733). Live cells were distinguished by LIVE / DEAD Fixable Yellow Dead Cell Stain Kit (Thermo Fisher Scientific, Catalog #L34959). Flow cytometry was performed on BD LSRFortessa X-20 flow cytometer. Flow cytometry data were analyzed using Flow Jo software (Tree Star). Unpaired student’s / -tests were performed to determine statistical significance using GraphPad Prism (Version 9.4.1).

[0143] Quantitative proteomics analysis

[0144] The analysis for the presence of human cytokines, chemokines, inflammatory factors, growth factors, and receptors released in the 3D cell culture before and after PBMC infiltration was conducted using a high-throughput sandwich-based human Kiloplex quantitative proteomics array (RayBiotech, Catalog #QAH-CAA-X00-l). The supernatants of the 3D-AD culture compartment were collected and tested against proteins. Each protein was tested in a quadruplicate. For protein quantification, the array had every single protein as a control and at known concentrations to generate standard curves for each protein. By comparing signals from the samples with the individual standard curves, the protein concentrations in the samples were quantified. In some instances, insufficient sample volume interfered with cytokine quantification. Measurements from these samples were replaced with ‘NA’. Fold changes were computed as log2([AD(+MG) + PBMC] / [AD(+MG) - PBMC]) and Iog2([control(+MG) + PBMC] / [control(+MG) - PBMC]), where the numerator is the cytokine concentration in the presence of CD3+ T cells, monocytes, or B cells and the denominator is the cytokine concentration in the absence of PBMCs. If present, technical replicates were averaged prior to fold change calculation. In the absence of technical replicates, single measurements were used. Immunohistochemistry

[0145] For immunofluorescent staining of neuroimmune axis models, cells were rinsed with D-PBS twice before fixing. Cells were then incubated in a 4% paraformaldehyde aqueous solution (PF A, Electron Microscopy Sciences, Catalog #157-4-100) overnight at 4 °C, followed by washing three times with D-PBS. Cells were permeabilized in 0.1% Triton X-100 in PBST (phosphate-buffered saline with 0.1% Tween®20) for 15 minutes at room temperature. Cells were blocked by incubating in 3% human serum albumin overnight in PBST at 4 °C. Cells were then incubated with primary anti-human antibodies for 24 hours at 4 °C, before washing with PBST three times. The following primary antibodies (and dilutions) were used: anti-p-Tau antibody (Thermo Scientific, Clone AT8, Catalog #MN1020; 1 :40); anti- GFAP antibody (Neuromab, Clone N206A / 8, Catalog #75-240; 1 : 100), anti-Tuj l antibody (Cell Signaling Technology, Catalog #5568; 1 :200); anti-IBAl (Millipore Sigma, Catalog #MABN92, 1 : 100); anti-CDl lb (Novus Biologicals, Catalog #NB 110-89474, 1 : 100).

[0146] For immunostaining of 5XFAD brain sections, harvested brain sections from 6-7-month-old WT (n=6 female and n=2 male) and 5XFAD (n=6 female and n=7 male) mice were treated with 3% DSSS and washed three times with tris-buffered saline (TBS) (Boston BioProducts, Catalog #BM301). Sections were then blocked for 1 hour with TBS containing 0.2% Triton X-100 (Millipore Sigma, Catalog #X-100) and 5% donkey serum (abeam, Catalog #ab7475), followed by overnight incubation with primary antibodies (and dilutions), including CD8a (Cell Signaling, D4W2Z, Catalog #98941, 1 :250), CXCL10 (R&D Systems, Catalog #AF-466-NA, 1 :200), GFAP (Millipore Sigma, Catalog #G6171, 1 :200), and 3D6 (a gift from Lilly, 1 : 1000). Next, brain sections were washed three times with TBS and incubated for 1 hour at room temperature with secondary antibodies, including anti-mouse 647 (Jackson ImmunoResearch, Catalog #715-175-150, 1 :500), anti-rabbit Cy3 (Jackson ImmunoResearch, Catalog #711-165-152, 1 :500), and anti-rat 488 (Invitrogen, Catalog #A-21208, 1 :500). Lastly, brain sections were stained with Hoescht (ThermoFisher Scientific, Catalog #62249) for 15 minutes at room temperature before washing three times with TBS and mounting them on glass slides for confocal imaging. Samples were imaged using a Nikon Eclipse Ti2 confocal microscope. The cortex and hippocampus were analyzed using the ‘Analyze Particles’ method with ImageJ (Version 2.9.0) to quantify the percentage of the area of interest covered by CXCL10+ cells. To determine statistical significance, unpaired, two-tailed Mann- Whitney / -tests were performed using GraphPad Prism (Version 9.4.1).

[0147] Single-cell RNA-sequencing

[0148] To identify cell type-specific transcriptional changes, we performed single-cell RNA sequencing from only the BRAIN compartment of AD or CTRL cultures, 72 hours post-loading peripheral immune cells into the PiChip. scRNA-seq via InDrop. Following magnetic isolation, the scRNA-seq library preparation was performed according to the Chromium Single Cell 30 Reagents kit v2 user guide. Briefly, a volume of single cells to target a capture of 10,000 cells was used in the generation of GEMS (gel bead in emulsions). The workflow was followed according to the 10X protocol and sequenced on an Illumina HiSeq 4000. All samples were multiplexed and sequenced to a depth of 60-100K reads per cell. Sequencing was performed at the Bauer Core Facility at Harvard University. scRNA-seq data analysis. Fastq files were processed using the bcbio-nextgen pipeline by following the inDrops3 cell and transcript counting protocol. Briefly, dual cellular indices, sample barcodes and UMIs were detected using umis70. Reads were aligned to hg38 using RapMap71and counts of reads per transcript per cell barcode for each sample were generated. Cells with less than 500 total reads assigned were discarded. The generated gene count matrix was used for subsequent analysis.

[0149] Normalization and integration. After QC, each library was individually normalized and scaled using SCTransform(). For all the libraries, we selected the 2000 most variable features for downstream integration. We determined a list of variable features across libraries with SelectIntegrationFeatures() that we used as an input for integration. And then we merged all together by using merge() function of the Seurat package to correct for any potential library batch effect. Integrated matrix was used for downstream analysis. In the integrated dataset, we performed Principal Components analysis (PC A, npcs = 20) to have separated cell-type clusters. We selected 1 :20 dimensions for dimensionality reduction by Uniform Manifold Approximation and Projection (UMAP), that we performed with the RunUMAP() function. To minimize the batch effect across all the conditions, we performed harmony(). And then the resolution set with 1.2. This led to the identification of 22 clusters, which represented unique neuron / astrocyte / microglia / T cells. The individual R objective files will be deposited with specific meta-data such as batch, resolution and condition. scRNA-seq differential gene analysis. Differentially expressed genes were found by applying the FindAllMarkers() function for overall DE and FindMarkers() for side by side comparisons, both from the Seurat R package. All the reported comparisons in the manuscript were performed with the following parameters: assay = "RNA", test.use = "wilcox", min. pct = 0.11, logfc.threshold = 0.01. We used the Wilcoxon rank sum test to calculate P- values. Only genes with their adjusted P value <0.05 (post-hoc, Bonferroni correction) were considered as significant.

[0150] NicheNet analysis. To perform NicheNet ligand activity analysis, we have compiled a list of potentially active ligands and their corresponding receptors / downstream genes based on gene expression changes in AD condition compared to control. We utilized NicheNet to identify the top 20 active ligands based on these gene changes and then identified the ligand-receptor interactions that are significant in four different cell types (neuron, astrocyte, microglia, and T cells). We set a cutoff value for ligands with at least one expressed receptor in each cell type, with an average log2 expression>0.1 and FDR<0.2. Once we identified these ligandreceptor interactions, we then queried the NicheNet data sources to identify the downstream genes associated with these interactions.

[0151] Blocking CXCR3 in CD3+ T cells

[0152] Flow cytometry. CD3+ T cells were immediately and carefully isolated from fresh whole blood using immunomagnetic negative selection (STEMCELL Technologies; EasySep Direct Human T Cell Isolation Kit, Catalog #19661) according to the manufacturer’s protocol and using sterile techniques. The isolated CD3+ T cells were centrifuged at 300 x G for 5 minutes at 4 °C and the supernatant was discarded. Human CXCR3 neutralizing antibody (R&D Systems, Clone: 49801, Catalog #MAB160-100) or CXCR3 chemical antagonist, NBI74330 (R&D Systems Catalog #4528 / 10) were diluted to concentrations of 10 pg / mL, 20 pg / mL, and 40 pg / mL with IMDM supplemented with 20% FBS. CD3+ T cells were resuspended at a concentration of 4 X 105cells / mL in the three different concentrations of the human CXCR3 antibody and were incubated for 40 minutes at 37 °C and 5% CO2. After the incubation period, the volume of each conical tube containing CD3+ T cells was brought up to 5 mL with cold cell staining buffer (BioLegend, Catalog #420201). The cells were then centrifuged at 300 x G for 5 minutes at 4 °C and the supernatant was discarded. Each cell pellet was resuspended at a concentration of 1 x 106cells / 500 pL of cold cell staining buffer and 12.5 pL of each of the following fluorochrome- conjugated antibodies per 500 pL of cell staining buffer were used to stain the cells: anti-human CXCR3 PE antibody (BioLegend, Catalog #353706), anti-human CD3 APC antibody (BioLegend, Catalog #317318), anti-human CD45 FITC antibody (BioLegend, Catalog #982316), and human TruStain FcX (BioLegend, Catalog # 422301). Cells were incubated with the fluorochrome-conjugated antibodies on ice in the dark for 20 minutes. After the incubation, the volume of each conical tube was brought up to 5 mL with cold cell staining buffer and the cells were centrifuged at 350 x G for 5 minutes. The supernatant was aspirated off, and the cell pellet was resuspended in 1 mL of cold cell staining buffer. CD3+ T cells were kept on ice, in the dark, and immediately analyzed using BD Fortessa X-20 flow cytometer. Analysis of flow data was performed using Flow Jo v.10.8 by first gating on singlets using forward scatter area versus height, then on live cells based on forward versus sidescatter profiles, followed by cell subset-specific gating.

[0153] CD3+ T cell migration in the microfluidic platform. Isolated CD3+ T cells from fresh whole blood were incubated with either CXCR3 neutralizing antibody MAB160 (10 pg / mL, R&D Systems Catalog #MAB160-100) or CXCR3 chemical antagonist NBI74330 (10 pg / mL, R&D Systems Catalog # 4528 / 10) for 40 minutes at 37 °C and 5% CO2 as mentioned above. Then, CD3+ T cells were stained with Hoechst for 5 minutes at room temperature. Lastly, treated and non-treated (CTRL) CD3+ T cells were washed and suspended in fresh IMDM with 4 mM L-glutamine and supplemented with 20% FBS prior to loading into the peripheral compartments of the microfluidic devices at a concentration of 5 x 104cells in 5 pL of media. Quantification of CD3+ T cell migration index and velocity was performed as previously described and the same image and data exclusion rules were used. Since there were time-lapse images where no CD3+ T cells moved through the PERIPHERAL-BRAIN microchannels of the microfluidic device, no velocity reading was given for that image.

[0154] Blocking CXCR3 in CD4+ and CD8+ T cells

[0155] Flow cytometry. CD4+ and CD8+ T cells were immediately and carefully isolated from fresh whole blood using immunomagnetic negative selection (STEMCELL Technologies; EasySep Direct Human CD4+ T Cell Isolation Kit, Catalog #19662; EasySep Direct Human CD8+ T Cell Isolation Kit, Catalog #19663) according to the manufacturer’s protocol and using sterile techniques. The isolated CD4+ and CD8+ T cells were centrifuged at 300 x G for 5 minutes at 4°C and the supernatant was discarded. Human CXCR3 neutralizing antibody (R&D Systems, Clone 49801, Catalog #MAB160-100) was diluted to a concentration of 10 pg / mL with IMDM supplemented with 4 mM L-glutamine and 20% FBS. The CD4+ and CD8+ T cells were resuspended at a concentration of 4 x 105cells / mL in the human CXCR3 antibody and were incubated for 40 minutes at 37 °C and 5% CO2. After the incubation period, the volume of each conical tube containing CD4+ or CD8+ T cells was brought up to 5 mL with cold cell staining buffer (BioLegend, Catalog #420201). The cells were then centrifuged at 300 x G for 5 minutes at 4 °C and the supernatant was discarded. Each cell pellet was resuspended at a concentration of 1 x 106cells / 500 pL of cold cell staining buffer. For CD4+ T cells and CD8+ T cells, 12.5 pL / 500 pL of cell staining buffer of the appropriate fluorochrome-conjugated antibodies were used to stain the cells: anti-human CXCR3 PE antibody (BioLegend, Catalog #353706), anti-human CD45 FITC antibody (BioLegend, Catalog #982316), antihuman Alexa Fluor 647 CD4+ antibody (BioLegend, Clone SK3, Catalog #344636) or anti-human Brilliant Violet CD8+ antibody (BioLegend, Catalog #344748), and human TruStain FcX (BioLegend, Catalog # 422301). Cells were incubated with the fluorochrome-conjugated antibodies on ice in the dark for 20 minutes. After the incubation, the volume of each conical tube was brought up to 5 mL with cold cell staining buffer and the cells were centrifuged at 300 x G for 5 minutes at 4°C. The supernatant was aspirated off and the cell pellet was resuspended in 1 mL of cold cell staining buffer. The cells were kept on ice, in the dark, and immediately analyzed using BD Fortessa X-20 flow cytometer. Analysis of flow data was performed using FlowJo v.10.8 by first gating on singlets using forward scatter area versus height, then on live cells based on forward versus side-scatter profiles, followed by cell subsetspecific gating.

[0156] CD4+ and CD8+ T cell migration in a microfluidic platform. Isolated

[0157] CD4+ and CD8+ T cells from fresh whole blood were incubated with CXCR3 neutralizing antibody MAB160 (10 pg / mL, R&D Systems, Catalog #MAB160-100) for 40 minutes at 37 °C and 5% CO2 incubator as mentioned above. Then the CD4+ and CD8+ T cells were stained with Hoechst for 5 minutes at room temperature. The cells were washed and suspended in fresh IMDM supplemented with 4 mM L- glutamine and 20% FBS prior to loading into the peripheral compartments of the microfluidic devices at a concentration of 5 x 104cells in 5 pL of media and imaged over 15 hours using confocal microscopy imaging. The total number of cells migrated over time was extracted from IMARIS software and data at 15-hours post-loading were used to compute the migration index. The migration index of inhibited T cells for each donor was calculated as follows:

[0158] Migration indexinhibited

[0159] Number of inhibited T cells migrated into the AD BRAIN compartment

[0160] Average number of uninhibited T cells migrated into the AD BRAIN compartment

[0161] Quantification of CD4+ and CD8+ T cell migration velocity was performed as previously described, and the same image and data exclusion rules were applied.

[0162] Chemotaxis of CD4+ and CD8+ T cells in response to CXCL10

[0163] The bottom layer of the microfluidic device was submerged with IMDM with

[0164] 4 mM L-glutamine and supplemented with 20% FBS. Recombinant human CXCL10 (R&D Systems, Catalog #266-IP-010) was loaded into the central compartment of the microfluidic device at a concentration of 100 nM. Isolated CD4+ and CD8+ T cells from fresh whole blood were stained with Hoechst for 5 minutes at room temperature and added at 5 x 104cells in 5 pL of media to each peripheral compartment.

[0165] Quantification of CD4+ and CD8+ T cell migration index and velocity was performed as previously described.

[0166] Quantitative analysis of cytokine levels in the presence of CD4+ and CD8+ T cells in AD cultues

[0167] Levels of Granzyme B, IL-2, IL-15, and TNF-a were simultaneously measured by a multi-array electrochemiluminescence assay kit (Meso Scale Diagnostics (MSD), U-PLEX Custom Immuno-Oncology Group 1 (hu) Assay, Catalog # K151AEM-2). To quantify chemokine levels, media was collected from CTRL and AD ReN cells that were differentiated in 1 : 10 Matrigel :ReN cell differentiation media for four weeks in a 96-well plate and had been co-cultured with and without microglia for 96 hours and with and without CD4+ T cells or CD8+ T cells for 72 hours. To mimic the conditions of the neuroimmune axis model in the 96 well plate, 2.5 x 105ReN cells were differentiated in each well and at a concentration of 2.5 x 104, microglia were added and followed by 5,480 CD4+ or CD8+ T cells in 4 pL of human serum (Millipore Sigma, Catalog #Sl-100mL). Collected media was kept frozen at -20°C until it was thawed for the assay. The samples were not diluted and were analyzed using the assay kit according to the manufacturer’s protocol.

[0168] CD3+ T cell migration towards ReN VM and ReN mCherry

[0169] ReN VM (Millipore Sigma, Catalog #SCC008) and ReN mCherry neural progenitor cells were cultured into the neuroimmune axis model as described above. Then, human CD3+ T cells from fresh whole blood were isolated as previously described and stained with Cell Tracker Green CMFDA (Invitrogen, Catalog #C2925). Each peripheral compartment of the microfluidic was loaded with 5 X 104cells in 5 pL of media. Images of each microfluidic were taken at time 0 and 24 hours to assess CD3+ T cell migration into the brain compartment.

[0170] Quantification of microglia branches

[0171] Nikon Elements Software was used to quantify the number of branches. First, microglial cell body was identified by bright spot detection with spherical parameter > 0.9, and then associated branches were quantified using the TrackID method.

[0172] Calcium imaging

[0173] The experiment described monitored cellular activity in 3D-AD cultures using a calcium indicator called Cal-590 AM (AAT Bioquest). The cultures were incubated for 30 minutes in differentiation media supplemented with 20 pM of Cal-590 AM, washed twice with fresh media, and incubated for 30 minutes before imaging. The fluorescence intensity dynamics were measured using time-lapse imaging by a Nikon Al confocal laser scanning microscope for a total of 100 frames. Two separate areas per well were captured in the same region of each well and used for analyses. Raw intensity values of all the active cells (>0 frequency for 100 frames) in the captured videos were extracted using ImageJ time series analyzer V3. The relative changes in fluorescence intensity to baseline [(Fti - Fti) / Fti] were regarded as calcium signals. For each region of interest (soma and / or neurite), a transient was accepted as a signal when its amplitude was greater than ten times the standard deviation of the noise in the baseline.

[0174] Statistical analysis

[0175] The data were analyzed for statistical significance using GraphPad Prism 9. The number of donors used throughout the manuscript are depicted in Figure legends. Data, expressed as mean ± SEM, were compared using either a two-tailed / -test when comparing two groups / conditions or an ANOVA followed by a post hoc test when comparing three or more groups / conditions. P value < 0.05 was considered significant.

[0176] Example 1. Construction and characterization of a 3D human neuroimmune axis model of AD.

[0177] A major advantage of in vitro microphysiological models is the ability to monitor the behavior of individual cell types in complex pathophysiological systems with multiple key factors contributing to disease pathogenesis. To assess the interaction among blood mononuclear cells (PBMCs) and brain resident cells in the context of AD neuropathology, we developed a 3D microfluidic-based neuroimmune axis system (hereafter referred to as PiChip system or PiChip cultures). This platform enables a multicellular human AD brain model to monitor the infiltration and interaction of human peripheral immune cells with brain cells in real time. In this system, we used immortalized human neural progenitor cells (NPCs) (ReNcell VM) that overexpress human Ab due to stable transfection with d / 7Jcontaining the FAD mutations, K670N / M671L (Swedish) and V717I (London) FAD (APPSL) and PSEN1 containing AE9 mutation (PSldE9) under the control of the human cytomegalovirus immediate-early enhancer and promoter, as previously reported25 27. NPCs were differentiated for four weeks in the PiChip system in 3D Matrigel to generate neurons and astrocytes. The NPCs with APPSL / PSldE9 mutations used in this study are all single clonal cells, which have been passaged multiple times without any active viral particles used for the initial transfection of APPSL / PSldE927. Human iPSC-derived microglial cells were separately differentiated as described previously29,30and added to the PiChip system at the end of week four of NPC differentiation, a day before isolation and loading of PBMCs (Fig. la). Using the PiChip system, infiltration of peripheral immune cells was monitored using time-lapse confocal imaging for 15-18 hours before collecting cells and media from the PiChip cultures and fixing them for further experiments, including biochemical assay, immunostaining, and single-cell RNA-sequencing (scRNA-seq) (Fig. la).

[0178] To observe the motility of peripheral immune cells, we designed and engineered an open-top, multi-compartment microfluidic chip in which different cell types of the central nervous system (CNS) and PBMCs isolated from human blood could be positioned side by side to allow live-cell confocal imaging (Fig. la-b). The microfluidic chip is comprised of a central BRAIN compartment with a large concentric cylinder reservoir (6000 pm in length and radius and 2500 pm open inner radius), allowing accumulation of AD neuropathology and various soluble factors, along with four PERIPHERAL compartments supplying PBMCs (Fig. lb; Fig. 8a). The PiChip system is designed to support a multicellular culture including human neurons (Neu), astrocytes (AC), iPSC-derived microglial cells (iMGL), and PBMCs in one single chip. The BRAIN and PERIPHERAL compartments are connected by 100 laser-etched, narrow migration microchannels (each side, 10x 10x500 pm3in height, width, and length) (Fig. 8a), providing a gradient of soluble factors released by CNS cells in the BRAIN compartment. The migration microchannels do not contain cellular components of the neurovascular unit. However, the confined space partially mimics the infiltration of peripheral immune cells into the brain parenchyma in response to chemotactic gradients. The migration microchannels also allow monitoring of infiltrating peripheral immune cells from the PERIPHERAL compartment toward the 3D-AD neural-glial cultures in the BRAIN compartment (Fig 1c)

[0179] The 3D-AD neural-glial culture (Fig. 1c; Fig. 8b) enabled the formation of an AD-like environment, including the accumulation of Ap deposition and p-Tau in the BRAIN compartment (Figs. 8c-d). Levels of Ap isoforms, Ap38, Ap40, and Ap42 were measured in media collected from the BRAIN compartment after four weeks of NPC differentiation with FAD mutations or control (CTRL). Significant increases in levels of AP40 (2120-fold, =0.0146) and Ap42 (2310-fold, =0.0033) were observed in the 3D-AD Neu / AC cultures compared to the CTRL (Fig. 8b). AP42 / 40 ratio was significantly higher (P=0.0032) in the 3D-AD Neu / AC culture compared to the CTRL (Fig. 8c). We also showed the formation of p-Tau using PHF-1 antibody in 3D-AD culture (Fig. 8d).

[0180] We then performed scRNA-seq using InDrop platform to identify the cell- type-specific transcriptomic changes. A total of 3,631 and 9,702 cells were analyzed from Neu / AC and Neu / AC / iMGL cultures, respectively (Fig. 8e). Cell clusters were then systematically assigned to three cell types, including neurons (SYT1, KCNQ1OT1, and KCNJ6), astrocytes (GFAP, VIM, SOX2, AQP4, ID4, and AGP), and microglia (CSF1R and SPIP) based on previously defined cell type markers (Fig. 8f). Previous studies also described the presence of oligodendrocytes in the 3D- differentiated ReNcell VM NPCs25 27. However, we could not clearly detect separated oligodendrocyte population in the scRNA-seq analysis, possibly due to the low numbers of oligodendrocytes compared to neurons and astrocytes and the insufficient expression of mature oligodendrocyte markers31. Transcription analysis of astrocytes revealed 965 significantly up- and 100 down-regulated genes in AD Neu / AC / iMGL culture compared to AD Neu / AC culture (FDR<0.1, Fig. 9g). Interestingly, reactive astrocytes (GFAP-, C3-, S100B-, and LLW-high) in the presence of microglia in the 3D-AD culture significantly downregulated genes related to glutamate and GABA homeostasis (SLC1A3, SLC6A11, and CP X1) compared to 3D-AD cultures without microglia (FDR<0.1, Fig. 8g-i), suggesting they may have impaired homeostasis and compromised astrocyte function. Additionally, we found a significant upregulation ( <0.0001) of interferon-gamma (INF-g)- and reactive astrocyte-associated genes in astrocytes in AD Neu / AC / iMGL culture compared to AD Neu / AC culture (Fig. 81). Overall, these scRNA-seq data provide evidence of increased astrogliosis in our AD cultures.

[0181] To test whether the presence of AD pathology affected microglial activation, we first assessed morphological changes in microglia in AD and CTRL conditions. We could observe microglial cells homogeneously distributed across the BRAIN compartment of the PiChip system 24 hours post-addition (Fig. Id). Microglia in the CTRL cultures showed widely branched filopodia in all directions (Fig. Id, top right, Fig. le), a morphology typically associated with homeostatic microglia32. In contrast, the AD cultures contained amoeboid microglia with relatively small numbers of branches (Fig. Id, bottom right, and Fig. le) and decreased length, indicative of activated pro-inflammatory microglia32. The AD-like environment also induced a significant upregulation of microglial activation markers, including ionized calcium- binding adaptor molecule 1 (IBA-1), compared to microglia in CTRL culture (P<0.0001, Fig. le), in agreement with previous reports33.

[0182] We then examined gliosis- and inflammation-associated genes such as complement C3, colony-stimulating factor- 1 receptor (CSF1R), monocyte chemotactic protein- 1 encoding gene (CCZ2), and INF-g receptor 1 associated receptor (IFNGR1) in the AD neural-glial cultures and CTRL (Fig. If). In the absence of microglia, C3 expression was not significantly different in 3D-AD Neu / AC cultures compared to CTRL (Fig. If). However, when microglial cells were introduced, C3 expression was increased by 6.9-fold (P<0.01) in the 3D-AD Neu / AC / iMGL cultures compared to CTRL (Fig. If). This is consistent with a previous study indicating that activated microglia in an AD-like environment signal astrocytes and microglia to secrete more C334. Similarly, we observed very low levels of CSF1R expression in 3D-AD Neu / AC cultures. However, following the introduction of microglia into the 3D-AD cultures, CSF1R expression was significantly increased by 6.2-fold (P<0.01) in the AD Neu / AC / iMGL cultures compared to CTRL (Fig. If). Additionally, inflammation-associated genes (IFNGR1, and CCL2) were significantly up-regulated ( <0.01) in the 3D-AD Neu / AC / iMGL cultures compared to CTRL (Fig. If), consistent with our previous report28. Further analysis using scRNA-seq revealed a substantial increase ( O.OOOl) of reactive astrocyte-associated signature genes in 3D-AD neural-glial cultures compared to control (Fig. 1g).

[0183] Next, we used scRNA-seq data to identify functionally enriched genes and pathways in neurons and astrocytes in AD cultures compared to CTRL (Fig. lh-1). We first identified three transcriptionally unique clusters of neurons, astrocytes, and microglia from a total of 20,153 single cells (CTRL Neu / AC, 4313; AD Neu / AC, 2808; CTRL Neu / AC / iMGL, 10616; AD Neu / AC / iMGL, 2416 single cells) in cultures (Fig. Ih). Transcriptional analysis of differentially expressed genes identified 263 up- and 82 down-regulated genes in neurons in AD cultures vs CTRL (Fig. li; red; FDR<0.05 and logFOO). Pathway analysis revealed a significant increase (FDR<0.V) in pathways related to endoplasmic reticulum stress (ER stress), DNA damage, and INF-g in AD neurons (Fig. Ij). Similarly, we identified 1,1146 up- and 61 down-regulated genes in astrocytes in AD cultures compared to CTRL (Fig. Ik; red; FDR<0.05 and logFC>0). Gene ontology analysis revealed several significantly up-regulated pathways (FDR<0.V), related to mRNA stability, tau kinase, and regulation of cellular response to stress in AD astrocytes compared to CTRL (Fig. 11). These results suggest the elevated cellular stress, tau kinase pathways and phagocytosis in AD astrocytes.

[0184] Overall, these data show that the multicellular PiChip system containing the 3D-differentiated neurons and astrocytes transfected with APPSL and PSldE9 and iPSC-derived microglia recapitulate the critical pathological features of AD, including Ap plaques and Tau pathology as well as gliosis.

[0185] Example 2. Chemotaxis of peripheral immune cells in microfluidic chips.

[0186] To characterize the microfluidic chip in tracking migration characteristics of peripheral immune cells, we assessed the chemotactic migration of T cells, monocytes, and B cells, in response to various chemokines (Fig. 9a). To do so, we isolated PBMCs from human blood collected from de-identified healthy individuals. We used imaging flow cytometry to assess the purity of isolated PBMCs and quantify the composition of immune cells (Figs. 9b-d). Starting with whole blood, the isolated CD3+ T cells (Fig. 9b), CD14+ monocytes (Fig. 9c), and CD19+ B cells (Fig. 9d) content was 96.2%, 84%, and 98.3% purity (gated on singlets, live, CD45+), respectively.

[0187] We next measured the directional migration of T cells, monocytes, and B cells toward a panel of known chemoattractants selected for each cell type (Figs. 9e-j). For this purpose, in the absence of neuronal cells, we added chemokines of choice and freshly isolated human PBMCs to the CHEMOKINE and PERIPHERAL compartments of the microfluidic chip, respectively (Fig. 9a). Using confocal timelapse imaging for ~20 hours, we imaged and quantified the number of PBMCs that migrated across the microchannels to the CHEMOKINE compartment and their migration velocity (Figs. 9e-j). CD3+ T cells were tested for chemotactic migration toward CXCL12, a known chemokine for T cells (Figs. 9e-f). CXCL12 displayed a bell-shaped dose-dependence characteristic, with a maximum of 406 ± 264 recruited CD3+ T cells after 15 hours toward 50 nM CXCL12, which is significantly higher than 5 nM ( =0.0274) and 100 nM (P=0.0192) CXCL12 concentrations (Fig. 9f). CD3+ T cells traveled through the migration microchannels at an average velocity of 5.5 ± 3.0 pm / min (Fig. 9f) in response to CXCL12 (1000 nM) and reached the CHEMOKINE compartment in less than 90 minutes after entering the microchannels. Lastly, we found no significant changes in CD3+ T cell velocity toward different concentrations of CXCL12 (Fig. 9f).

[0188] Next, CCL2 was used to test whether it could elicit chemotaxis of human monocytes (Figs. 9g-h). As expected, monocytes showed a classic bell-shaped dosedependent chemotactic response to CCL2, with a maximum of l614 ± 775 recruited monocytes after 15 hours at 100 nM concentration, which was significantly higher than 5 nM ( <0.0001) and 1000 nM ( =0.0002) CCL2 concentrations. Monocytes exhibited no significant differences in the number of migrated cells toward CCL2 at 50 and 100 nM concentrations (Figs. 9g-h), however, both concentrations led to a higher number of migrated monocytes as compared to 5 nM ( =0.0054) and 1000 nM ( =0.0397) CCL2 chemokine (Fig. 9h). In addition, CCL2 induced the recruitment of monocytes with an average velocity of 1.2 ± 0.2 pm / min at 50 nM concentration, which was significantly higher than the average velocity observed at 100 nM ( =0.0048) and 1000 nM ( =0.0009) concentrations (Fig. 9h).

[0189] Human B cells were also tested for directional migration toward CCL19 with average recruitment of 28 ± 10 migrated B cells at 1000 nM CCL19 and no significant differences between varying concentrations of CCL19 (Fig. 9i-j). B cells had an average migration velocity of 1.5 ± 1.4 pm / min in response to 1000 nM CCL19 (Fig. 9j). Moreover, B cells exhibited no significant differences in the average velocity toward varying CCL19 concentrations (Fig. 9j).

[0190] Overall, these data reflect the robustness of our PiChip system to generate chemokine gradients in a microfluidic environment. Our setup allows the quantification of migration characteristics of peripheral immune cells at the single-cell level and under highly controllable conditions. Example 3. Selective infiltration of T cells into the 3D-AD cellular models.

[0191] Next, we tested whether the presence of AD neuropathology in the PiChip system would affect the directional migration of peripheral immune cells into the 3D- AD neural-glial cultures. As mentioned above, NPCs containing FAD mutations in the PiChip system were first 3D-differentiated into neurons and astrocytes for four weeks as previously reported25 27. Simultaneously, microglia were differentiated according to previously published studies29,30and then added to the BRAIN compartment of the PiChip system at week four (Fig. 2a). T cells, monocytes, and B cells were isolated from whole blood collected from de-identified donors and loaded into the PERIPHERAL compartments. Infiltration of peripheral immune cells was monitored using time-lapse confocal microscopy imaging for 15-20 hours (Fig. 2b-c).

[0192] Migration characteristics of peripheral immune cells were quantified using a migration index and velocity of immune cells. To quantitatively analyze the migration index and velocity of peripheral immune cells, we developed an unbiased / automated image analysis pipeline for high-throughput quantification of the number of infiltrating peripheral immune cells and their corresponding velocities in the PiChip utilizing time-lapse confocal imaging and IMARIS software (see Extended Data Fig. 3 for the detailed pipeline). Each PiChip system allowed us to track peripheral immune cells across four sets of 100 microchannels connecting the PERIPHERAL to the BRAIN compartment. The migration index was used to quantify the fold change of immune cell infiltration into the AD neural-glial cell culture (the BRAIN compartment) compared to the control neural-glial culture (see Methods, Fig. lOb-c).

[0193] First, we investigated whether AD pathology in the PiChip system differentially induces the infiltration of specific peripheral immune cell types, including CD3+, CD4+, and CD8+ T cells, monocytes, and B cells. We observed a significantly greater infiltration of human CD3+ T cells into the 3D-AD Neu / AC cultures compared to CTRL (1.43-fold, <0.0001, w=10 donors, Fig. 2e). We also found a similar trend for both CD4+ and CD8+ T cells. Specifically, we observed a significant infiltration of CD4+ T cells (2.02-fold, <0.0001, n= 1 donors, Fig. 2f) and CD8+ T cells (1.61-fold, =0.0093, w=10 donors, Fig. 2f) in AD Neu / AC cultures compared to CTRL. Overexpression of mCherry in the AD cells had no significant effect on the recruitment of T cells in the PiChip system compared to control ReNVM cells (Fig. lOd). Monocytes showed a significant increase in infiltration into the 3D- AD Neu / AC cultures as compared to CTRL (3.18-fold, P<0.0001, n=6 donors, Fig. 2g). In contrast, we observed no significant differences in infiltration of the B cells into the 3D-AD Neu / AC cultures compared to CTRL (n=8 donors, Fig. 2h).

[0194] Next, we assessed whether microglial cells impact the number of infiltrating peripheral immune cells into the PiChip cultures. In the presence of microglia, the rate of CD3+ T cells infiltration was significantly higher in 3D-AD Neu / AC / iMGL cultures compared to CTRL (1.5-fold, P=0.0008, / / =8 donors, Fig. 2e), indicating that the 3D-AD brain-like environment leads to preferential recruitment of CD3+ T cells regardless of the presence or absence of microglial cells. In contrast, the greater rate of monocyte infiltration into the 3D-AD Neu / AC versus CTRL cultures was abrogated by the addition of microglia, showing no significant differences between AD and CTRL groups (n=9 donors, Fig. 2g). Lastly, B cells showed a noticeable decrease in the infiltration rate into AD Neu / AC / iMGL cultures compared to CTRL when microglial cells were present ( O.OOOl, w=12 donors, Fig. 2h). Further studies are required to fully dissect underlying mechanisms regarding the impact of microglia on migration dynamic of monocytes and B cells in AD conditions.

[0195] To further characterize migration characteristics of peripheral immune cells in the PiChip system, we measured the migration velocity of peripheral immune cells during their infiltration into AD and CTRL cultures in the presence and absence of microglia. CD3+ T cells migrated slightly faster, but not significantly toward 3D-AD cultures, regardless of the presence or absence of microglia, reaching the PiChip BRAIN compartment in less than 7.8 hours after entering the migration microchannels. Specifically, CD3+ T cells migrated at 0.92 ± 0.07 pm / min and 0.81 ± 0.04 pm / min toward 3D-AD Neu / AC and CTRL cultures, respectively (n=8 donors, Fig. 2i). In the presence of microglia, CD3+ T cells exhibited a similar average migration velocity of 0.94 ± 0.06 pm / min toward 3D-AD Neu / AC / iMGL cultures compared to 0.83 ± 0.06 pm / min in CTRL (n=9 donors, Fig. 2i). Monocytes exhibited significantly different migratory velocities moving toward the PiChip BRAIN compartment in the presence and absence of microglia. Monocytes showed faster migration velocity toward 3D-AD Neu / AC culture at 2.0 ± 0.15 pm / min as compared to CTRL at 1.54 ± 0.12 pm / min (P=0.0019, n=9 donors, Fig. 2i). However, in the presence of microglia, monocytes abrogated the difference in average migration velocity between AD (0.75 ± 0.05 pm / min) and CTRL (0.8 ± 0.04 pm / min) Neu / AC / iMGL cultures (n=9 donors, Fig. 2i). We also observed that monocytes migrated faster into the 3D Neu / AC cultures compared to Neu / AC / iMGL cultures in both AD and CTRL conditions by 2.67-fold (P<0.0001) and 1.93-fold (P<0.0001), respectively (n=9 donors, Fig. 2i). The presence of microglia equalizing the average velocity of monocytes in 3D-AD Neu / AC / iMGL versus CTRL cultures leading to slower migration velocities as compared to Neu / AC cultures without microglia. Lastly, B cells exhibited similar migratory velocity in AD (0.78 ± 0.045 pm / min) and CTRL (0.78 ± 0.043 pm / min) cultures in the absence of microglia («=10 donors, Fig. 2i), and between AD (0.43 ± 0.04 pm / min) and CTRL (0.53 ± 0.04 pm / min) cultures in the presence of microglial cells (w=12 donors, Fig. 2i). However, as observed with monocytes, B cells migrated significantly faster in the absence versus the presence of microglia (1.8-, and 1.5-fold, into AD, <0.0001 and CTRL, =0.0009, respectively, Fig. 2i), as defined by average migration velocity.

[0196] Overall, the microfluidic PiChip system enabled us to probe and study the infiltration characteristics of human peripheral immune cells into 3D neural-glial cell cultures at a single-cell level. Using this multicellular model, we showed that CD4+ and CD8+ T cells dramatically infiltrate 3D-AD neural-glial cultures compared to control, regardless of the presence or absence of microglia.

[0197] Example 4. Infiltrating CD4+ and CD8+ T cells in brains of 5XFAD mice.

[0198] Previous studies have shown that T cells infiltrate into the brain parenchyma of aged AD mice and exacerbate AD neuropathology10 12 l 5. Most studies have relied on transcardial perfusion to extract leukocytes from the vasculature. However, this method may not completely remove blood-borne immune cells. Here, we used intravascular CD45-FITC staining to distinguish brain-localized leukocytes from those in the bloodstream35(Fig. Ila). CD45+ immune cells in the blood would be labeled with FITC, while extravascular leukocytes should remain FITC -negative. We found that in 5XFAD mice, the percentage and number of CD45+CD1 lb- non- myeloid cells were significantly increased relative to wild-type (WT) mice (P=0.0058, and =0.0003 for the percentage and number, respectively - Fig. 11b). Further analysis demonstrated a significantly increased percentage and greater absolute number of extravascular CD4+ T cells ( =0.0009 and =0.0002 for the percentage and number, respectively) in the brains of 5XFAD mice compared to WT (w=9-12 mice per group, Fig. 2j-k). Similarly, we found that the percentage and the absolute number of extravascular CD8+ T cells were significantly higher in the brains of 5XFAD mice compared to WT ( =0.0031, and =0.0014 for the percentage and number, respectively - Fig. 2j-k). Although the numbers of infiltrating CD4+ and CD8+ T cells increased in 5XFAD brains, the percentages of these cells among all extravascular CD45+ cells were comparable between 5XFAD and WT mice (Fig. llc-d). We reasoned that other immune cells, including neutrophils and monocytes, also showed increased infiltration in 5XFAD brains and thus, the proportion of T cells was not significantly different. Overall, these data show that CD4+ and CD8+ T cells infiltrate the brains of 5XFAD mice significantly more than WT mice, in agreement with our observations in the human PiChip models.

[0199] Example 5. Elevated levels of inflammatory cytokine following infiltration of immune cells into the AD cultures.

[0200] We next assessed levels of cytokines and chemokines in the PiChip BRAIN compartment following the infiltration of peripheral immune cells. We collected conditioned cell culture media before and after infiltration of peripheral immune cells from the BRAIN compartment in the PiChip cultures 24 hours post-loading, and the level of cytokines and chemokines was measured using high-throughput human Kiloplex quantitative proteomics arrays. We then quantified the fold changes of cytokines and chemokine levels following the infiltration of peripheral immune cells as compared to AD and CTRL cultures lacking infiltrating peripheral immune cells (Fig. 12a). Following the infiltration of immune cells, several inflammatory cytokines, including complement component Iq (Clq), C3, CXCL1 (GRO-alpha), interleukin-8 (IL-8), and CXCL10 (IP-10), were significantly elevated (P<0.05, Fig. 12a-b). Specifically, infiltrating CD3+ T cells led to a significant increase of Clq and C3 by 702-fold ( <0.0001) and 193-fold (P<0.0001), respectively, in the BRAIN compartment of AD cultures, compared to AD cultures lacking CD3+ T cells (Fig. 12b) Following infiltration of monocytes, we also observed a significant elevation of C3, CXCL1 (GRO-alpha), IL-8, and CXCL10 (IP-10) by 45-fold (P=0.0027), 59- fold (P<0.0001), 40-fold (P<0.0001), and 9-fold (P<0.0001), respectively, in the AD cultures compared to cultures lacking monocytes (Fig. 12b). Finally, while B cells infiltrated the PiChip cultures in very low numbers, their infiltration significantly elevated protein levels of C3 and CXCL10 by 80-fold ( =0.0002) and 4.5-fold (P<0.0001), respectively, compared to cultures lacking B cells (Fig. 12b). In summary, these data suggest that the dynamics of cytokine and chemokine levels are not only regulated by the CNS cells and, further, by AD neuropathology but they also dramatically modulated by the presence of infiltrating peripheral immune cells into the AD cultures. These findings provide further insights into the mechanism underlying the infiltration of peripheral immune cells into the brain and how they may exacerbate AD neuropathogenesis.

[0201] Example 6. Dramatically increased neurodegeneration following infiltration of CD8+ T cells in the presence of microglia.

[0202] Next, we assessed whether infiltrating T cells in the AD human neuroimmune axis model have a protective or detrimental impact on AD pathogenesis and the role of microglia in this process. To this end, we studied the impact of infiltrating T cells on neural cell death in the PiChip system (Fig. 3a). Infiltrating T cells (blue, Hoechst) co-localized with neurons / astrocytes (green, GFP) and microglia (magenta, IBA-1) using real-time confocal microscopy imaging (Fig. 3b). We were able to observe infiltrating T cells in the AD cultures, at which point, neurons expressed strong GFP signals (Fig. 3b). After 72 hours, we observed neurite- / axon-damaged cell bodies, which co-localized with microglia and T cells (Fig. 3b), only in AD cultures. Cleaved dystrophic neurites were co-localized with microglia, possibly indicative of physical interruption, as previously reported28(Fig. 13a). Time-lapse confocal imaging also showed that AD neurons underwent changes in cell shape, including the presence of cleaved neurites proximal to cell bodies with shrunken soma.

[0203] To quantitively measure the impact of different types of T cells, we first measured GFP+ expressing surface area before adding T cells and / or iMGL to the AD cultures. After 72 hours of adding T cells to the 3D-AD Neu / AC (+ / -) iMGL cultures, the same ROI was measured using confocal microscopy imaging. We then quantified the GFP+ expressing surface area, which represents viable neurons and astrocytes before and after adding T cells and microglia as described previously24. After 72 hours, cellular damage was significantly increased by 74.5% (PO.OOOl) in AD cultures with CD3+ T cells and iMGL, as quantified by the GFP+ expressing surface area, when compared to AD cultures lacking CD3+ T cells and iMGL (Fig. 3c). We found similar observation when counting the number of cells as another measure of cellular damage, in which presence of infiltrating CD3+ T cells and iMGL together in AD cultures resulted in significant cellular damage ( =0.0032, Fig. 3d) compared to AD cultures without CD3+ T cells and iMGL. We also observed a drastic reduction in GFP+ surface area (P<0.0001) or cell count (P=0.0005) in AD cultures when microglial cells were presented compared to AD cultures without microglia (Fig. 3c- d). Intriguingly, when we quantified the GFP+ expressing surface area of the AD cultures with both CD3+ T cells and iMGL compared to AD culture with only iMGL, we observed a significant increase of 55.1% ( =0.0034) in cellular damage, indicating infiltrating CD3+ T cells partner with microglia to exacerbate cellular damage in the AD cultures (Fig. 3c).

[0204] To further dissect which type of T cells was associated with neurodegeneration, we used either CD4+ or CD8+ T cells in the PiChip system. To answer this question, we prepared seven distinct 3D cell culture conditions, including 1. neuron / astrocyte (CTRL and AD), 2. neuron / astrocyte with CD4+ T cells (CTRL and AD), 3. neuron / astrocyte with CD8+ T cells (CTRL and AD), 4. neuron / astrocyte with microglia only (CTRL and AD), 5. neuron / astrocyte with CD4+ and CD8+ T cells (CTRL and AD), 6. neuron / astrocyte with CD4+ T cells and microglia (CTRL and AD), and 7. neuron / astrocyte with CD8+ T cells and microglia (CTRL and AD) (Fig. 3e). The presence of iMGL alone in AD cultures induced significant neuronal and astrocytic cellular damage (P<0.0001, Fig. 3f). Interestingly, CD8+ T cells alone (but not CD4+ T cells) in the AD cultures also induced significant neuronal damage (P=0.0008, Fig. 3f). We found no significant difference between neuronal damage in 3D-AD Neu / AC cultures caused in the presence of both CD4+ and CD8+ T cells compared to CD8+ T cells alone, as measured by GFP+ expressing surface area (Fig. 3f). This data suggests that CD4+ T cells do not exacerbate the CD8+ T cell-induced neuronal damage in 3D-AD Neu / AC culture. We observed the greatest degree of impact on neuronal damage when CD8+ T cells and microglia were present together in AD cultures, compared to AD cultures lacking CD8+ T cells (P=0.0006), microglia (P<0.0001) or both (PO.OOOl) (Fig. 3f). Overall, these data strongly indicate that the observed cellular damage in AD cultures was greatest in the presence of both infiltrating CD8+ T cells and microglial cells. In contrast to CD8+ T cells, CD4+ T cells, in the presence or absence of microglia in AD cultures, did not lead to significant changes in cellular damage (Fig. 3e-f).

[0205] To better elucidate the exact cellular damage in AD cultures, we next studied whether infiltrating CD8+ T cells and microglia impact both neurons and astrocytes. We first assessed Tuj 1+ (a marker for differentiated neurons) surface area and found a similar decrease to previous assays where we measured GFP+ expressing surface area (Fig. 3b-e) We found that microglial cells alone cause robust damage to Tuj 1+ cells in AD cultures (32.9% increase, PO.OOOl) (Fig. 3g). Similarly, we observed considerable damage to Tuj 1+ cells in AD when CD4+ T cells (P=0.002) or CD8+ T cells (PO.OOOl) alone present in the 3D-AD Neu / AC cultures (Fig. 3g). Notably, the greatest damage in AD cultures on Tuj 1+ areas occurs when both infiltrating CD8+ T cells and microglia were present together (62% increase in neuronal damage, PO.OOOl) compared to AD cultures lacking microglia and CD8+ T cells (Fig. 3g). To further confirm the extent of neuronal damage, we next performed an ELISA assay to quantify the amount of Tuj 1 in AD cultures. Tuj 1 was decreased significantly (P<0.0001) in AD cultures in the presence of T cells and microglia compared to AD culture with T cells or microglia, alone (Fig. 13b). Additionally, we utilized aldehyde dehydrogenase 1 (ALDH1), a marker for differentiated astrocytes to further dissect the impact of infiltrating CD8+ T cells and microglia in AD cultures. While the presence of microglia alone in AD cultures induced significant astrocytic damage (P<0.0001), CD8+ T cells showed no impact on ALDH1+ cells in 3D-AD Neu / AC cultures (Fig. 3h). Similar to our observations from other assays, we found the greatest degree of impact on ALDH+ astrocytic damage when infiltrating CD8+ T cells and microglia were present together in the AD cultures (P<0.0001, Fig. 3h). This data indicates that CD8+T cells alone induced significant damage in the 3D-AD Neu / AC cultures, but only at the neuronal level and glial survival in the 3D-AD Neu / AC culture was unaffected statistically by infiltrating CD8+ T cells alone. In contrast, microglial cells alone have the capacity to induce cellular damage to both neurons and astrocytes. Lastly, the coexistence of infiltrating CD8+ T cells and microglia in AD neural-glial cultures resulted in dramatic neuronal and glial damage.

[0206] We also assessed neuritic damage using the neurite-per-cell body density ratio (Fig. 3i and j). While the neurite-per-cell body density ratio was maintained -2000 mm2in 3D-AD Neu / Ac cultures alone, infiltrating CD8+ T cells (not CD4+ T cells) into 3D-AD Neu / AC cultures significantly decreased neurite-to-cell body density ratio to -1600 mm2( =0.0144, Fig. 3j) compared to 3D-AD Neu / AC cultures lacking CD8+ T cells. Thus, CD8+ T cells alone can also induce neuritic damage, reconfirming our previous observations. Neurite density decreases moderately but significantly ( =0.0233) in the 3D-AD Neu / AC / iMGL versus AD culture without microglia (Fig. 3i). The most significant decrease in the neurite-to-cell body ratio ( <0.0001) was observed when infiltrating CD8+ T cells, and microglia were present in AD cultures (Fig. 3i), consistent with the results of previous cellular damage assays. These results strongly support that 1. CD8+ T cells or microglia alone, drive neurons and / or neuritic damage (not astrocytes) in the AD cultures, and 2. microglia and CD8+ T cells play a synergistic role in mediating neurodegeneration in AD. In addition to utilizing immunostaining for neuronal-glial damages upon infiltration of T cells in the presence of microglia, we used time-lapse calcium imaging to discern any altered neuronal / glial activities due to T cell infiltration. Interestingly, we found a significant decrease in the frequency of calcium response (as measured by peak / time) by 1.8-fold (P<0.0001) in AD Neu / AC culture in the presence of infiltrating T cells and microglia compared to CTRL (Fig. 13c). This data provides another support for the pathogenic role of infiltrating T cells in modulating the calcium signaling in AD cells. Further studies are needed to understand the molecular mechanisms of calcium signaling disturbances observed in AD cultures upon infiltration of peripheral immune cells. Overall, all assays demonstrate robust cellular damage in AD cells in the presence of infiltrating CD8+T cells and microglia.

[0207] Example 7. Interferon / inflammatory-associated pathways are activated in glial cells following infiltration of T cells in AD cultures.

[0208] To identify cell type-specific transcriptional changes, we performed scRNA- seq from only the PiChip BRAIN compartment of AD or CTRL cultures, 72 hours post-loading of T cells (Fig. 14). Cell clusters were first systematically assigned to five cell types, including neurons (SYT1, KCNQ10T1, and KCNJ6), astrocytes (GFAP, VIM, S0X2, AQP4, and ID4), microglia (CSF1R and SPI1 T cells (CD3D and CD8A) and proliferative cells (T0P2A and MKI67) based on previously defined cell type markers (Fig. 14a-b). The scRNA-seq data revealed several significantly up and downregulated genes (FDR<0.2 and FC>0, Up or FC<0, Down; Fig. 14c-d) in each cell type associated with AD culture vs CTRL. Compared to T cells in CTRL Neu / AC / iMGL cultures, pathways related to cytotoxicity, and antigen presentation were significantly upregulated in T cells in AD Neu / AC / iMGL cultures (FDR<0.1, Fig. 14e). Microglia in AD cultures with infiltrating T cells were significantly enriched with upregulated pathways involved in cell killing, and antigen presentation (FDR<0.1, Fig. 14e). Among those pathways, the antigen processing and presentation and cell killing pathways were highly enriched in microglia in AD cultures in the presence of infiltrating T cells compared to CTRL (FDR<0.1, Fig. 14e). The scRNA-seq data also showed that the high upregulation of interferon- associated pathways (e.g., response to INF-b) in astrocytes in AD cultures with microglia and T cells, compared to CTRL cultures (FDR<0.1, Fig. 14e).

[0209] Given the progressive neuron loss and gliosis in AD brains, we next sought to investigate how the CNS cells altered in the PiChip settings (Fig. 15). We found that neurons, astrocytes, and microglia are significantly altered in different conditions. Of those, neurons and astrocytes were significantly decreased, while the microglia population was significantly increased in AD culture vs CTRL (Fig. 15c). Following infiltration of T cells, we found an increase in microglia and T cell populations and a decrease in astrocytes in AD Neu / AC / iMGL compared to CTRL condition (Fig. 15d). Additionally, we identified multiple transcriptionally unique clusters of microglial cells in AD and CTRL conditions (Fig. 15e). Interestingly, we found that disease- associated microglia (DAM) and interferon microglia are significantly increased in AD Neu / AC / iMGL cultures compared to CTRL (Fig. 15f). This suggests that AD environment and T cell infiltration alter compositions of the CNS cells.

[0210] Because of our interest in understanding the pathogenic role of CD8+ T cell and their crosstalk with microglia, we then focused on T cells and microglia as a subject of the in-depth scRNA-seq analysis (Fig. 4). Major T cell subsets (CD4+ Naive, CD4+ Memory, and CD8+ T cells) were annotated upon clustering of scRNA- seq dataset (Fig. 4a-b). We identified that the percentage of CD8+ T cells is significantly increased in AD Neu / AC / iMGL cultures with infiltrating T cells, compared to AD cultures without microglia (P=0.002, Fig. 4b-c). We found that while there is an increase of total CD4+ T cells (naive and memory) in AD culture compared to control, however, there is no significant change in CD4+ memory T cell population in AD and control (Fig. 4a-d). In contrast, the percentage of CD4+ memory T cells significantly decreased in AD Neu / AC / iMGL cultures in the presence of infiltrating T cells, as compared to AD cultures without microglia ( =0.016, Fig. 4b and d). Lastly, we found that the CD4+ naive / memory T cell ratio was 2.0 and 1.5 in the presence and absence of microglia in AD cultures, respectively. Given that the CD4+ naive / memory T cell ratio is 0.67±0.31 of healthy individuals36, this data clearly demonstrates that naive CD4+ T cells infiltrated AD cultures more than memory CD4+ T cells, independent of microglia presence.

[0211] To elucidate the impact of infiltrating T cells on microglia, we analyzed differentially expressed genes in microglia in AD versus CTRL with and without T cells. We identified eight transcriptionally unique clusters of microglial cells from a total of 20153 single cells (Fig. 4e) in four different conditions of AD and CTRL with and without T cells (Fig. 4e). Interestingly, we found that DAM is significantly increased in AD Neu / AC / iMGL cultures without infiltrating T cells compared to CTRL (P=0.0251, Fig. 4f-g). Notably, the presence of infiltrating T cells comprehensively reduced the DAM population in AD and CTRL cultures with a significant impact on AD ( <0.0001, Fig. 4f-g), suggesting that infiltrating T cells can modulate the DAM population. Importantly, we found that microglia-associated major histocompatibility complex class II (MHC II) is significantly increased in AD cultures with infiltrating T cells, as compared to AD Neu / AC / iMGL cultures lacking T cells (P=0.0007, Fig. 4f and h) or CTRL Neu / AC / iMGL (P=0.0002, Fig. 4f and h). This may indicate that microglia act as antigen-presenting cells to promote CD8+ T cell infiltration in AD.

[0212] We also performed a comparative transcriptional analysis of genes, differentially expressed in microglia in AD versus CTRL with or without T cells, separately (Fig. 4i). We identified 263 up- and 82 down-regulated genes in microglia in AD Neu / AC / iMGL cultures with T cells compared to CTRL (Fig. 4i; red; FDR<0.05 and logFC>0). Interestingly, interferon-induced transmembrane protein 1 (IFITML) and signal transducer and activator of transcription 1 (STAT1) genes were significantly enriched (P<0.0001) in AD cultures in the presence of T cell and microglia compared to CTRL (Fig. 4j), suggesting a role of interferon-associated pathways in T cells infiltrating the AD cultures in the presence of microglia. Gene ontology revealed a significant increase of IFN-y-mediated signaling pathway, antigen processing and presentation of exogenous peptide antigen viaMHCII, and amyloid fibril formation in AD Neu / AC / iMGL cultures with T cells (FDR<0.1, Fig. 4k). We also detected significantly down-regulated pathways, including cellular response to cytokine stimulus, actin fdament organization and apoptotic cell clearance in AD Neu / AC / iMGL cultures with T cells (FDR<0.1, Fig. 4k).

[0213] To continue exploring the crosstalk between infiltrating CD8+ T cells and glial cells, we performed three independent analyses using the same dataset, with the focus on AD environment (Fig. 16). Transcriptional analysis of differentially expressed genes in microglia showed multiple significantly upregulated proinflammatory cytokines and chemokines (e.g., CXCL9, CXCL10, CXCL11, and / Z- 32 genes) in the presence of infiltrating T cells in AD cultures compared to AD cultures without T cells (Fig. 16a). Cytokine- and IFN-y-associated genes were significantly higher (P<0.005) in microglia when T cells were present in AD cultures (Fig. 16b). Pathway analysis from the microglia dataset showed the upregulation of several antigen and immune-related pathways, including antigen receptor-mediated signaling pathway, immune receptor-activating cell surface receptor signaling pathway, and activation of immune response (FDR<0.1, Fig. 16c). In contrast, pathways such as collagen-containing extracellular matrix and transition metal ion homeostasis are significantly down-regulated in microglia when T cells infiltrate the AD culture (FDR<0.1, Fig. 16d). We also performed similar analysis on astrocytes in the presence of infiltrating T cells and microglia in AD. Interlingually, we again found several significantly upregulated proinflammatory cytokines and chemokines (e.g., CXCL1, CXCL10, and IL-32 genes) and interferon-associated (e.g., IFITM1, IFIT3, and STAT1 genes) in astrocytes in the presence of infiltrating T cells and microglia in AD cultures compared to AD cultures without T cells (Fig. 16e-f). Indeed, pathway analysis of astrocytes showed significant upregulation of multiple pathways related to T-cell mediated cytotoxicity and antigen processing and presentation of peptide antigen viaMHC I (FDR<0.1, Fig. 16g-h).

[0214] We next set out to identify ligand-receptor interactions between infiltrating T cells and the different CNS resident cells to predict how they interact. To do so, we used NicheNet analysis to predict ligand-receptors between interacting T cells and CNS cells by combining their expression data37. We identified the top 20 of differentially expressed ligand activity observed in AD cultures compared to CTRL based on Pearson correlation coefficient between a ligand’s target predictions and the observed transcriptional response (Extended Fig. 10a). These ligands are expressed across five different cell clusters (Neuron, Proliferative, Microglia, Astrocyte and T cells). We found that most of the potential top 20 ligands were mainly expressed by microglial cells, including CCL2, IL-10, CCL8, FGF23 and APOE (Extended Fig. 10a). Interestingly, glia maturation factor-0 (GMFB) was the most highly expressed ligand in neurons. GMFB co-localizes with 0-amyloid plaques and NFT pathologies in AD brains and is positively associated with AD progression38’39. We then applied NicheNet to predict ligand-receptor interactions (Extended Fig. 10b). This analysis identified CD44, which is also associated with microglia / macrophages, is the receptor of T cell-associated ligands such as protein tyrosine phosphatase, receptor-type, C (PTPRC). Notably, we found that high-mobility group box protein 1 (HMGB1) interacts with Toll-like receptor 4 (TLR4), which is highly expressed by microglia in the CNS (Extended Fig. 10b). This is consistent with previous reports in which HMGB1 interacts with A042 in AD and inhibits A0 phagocytosis and internalization by microglia through TLR4 signaling40. We also predicted active target genes on the basis of potential upstream ligand activity and receptor expression to identify which top-ranked ligands are predicted to have regulated the expression of downstream pathways (Extended Fig. 10c). And interestingly, we found multiple significantly changed genes, including interferon g- (STATT), gliosis- (GFAP, C1Q), and inflammatory-associated TNFRSF1B, CCL2, CCL4, CCIS) as well as matrix metalloproteinase-9 (MMP9). These target gene changes would be valuable targets for future studies. To further exploring the link between infiltrating CD8+ T cells, and glial cells, we used protein multiplex assay to quantitively measure pro-inflammatory cytokines after infiltration of CD8+ T cells in AD cultures (+ / -) microglial cells (Fig. 18). We found a substantial increase of IL-15 (P<0.0003), IL-2 (P<0.0016), and CXCL10 (P<0.0087) levels in AD cultures when CD8+ T cells and microglia are present, as compared to AD cultures lacking microglia («=5-7 de-identified donors, Fig. 18a). We also observed a modest trend where granzyme B levels are elevated (but not significant) in the presence of CD8+ T cells and microglia in AD cultures (Fig. 18a). This is also consistent with the scRNA-seq data, in which the expression of five major cytotoxic genes GZMA, GZMB, GZMH, GZMM, and GZMK showed different patterns for CD4+ and CD8+ T cells in AD cultures (Fig. 18b). The high expression of three cytotoxic genes, GZMA, GZMB, and GZMH was predominantly associated with CD8+ cytotoxic T cells, in contrast with their low-level expression by CD4+ naive T cells (Fig. 18b). Additionally, since we robustly observed an increase in IFN-y signaling pathways following T cell infiltration in the presence of microglia in AD cultures, we, investigated whether selective infiltration of CD8+ T cells in AD can increase IFN-y in AD environment. To test this, we quantitively measured the IFN-y release and found a significant increase in IFN-y level when CD8+ T cells coexist with microglia compared to CD8+ (P<0.001) or iMGL (P=0.0024) alone in the culture ( O.OOl, Fig. 18c), as measured by ELISA. Together these data provide evidence that infiltrating CD8+ T cells promotes IFN signaling and likely communicates with astrocytes through proinflammatory cytokines in AD cultures, but to what extent these neuroimmune interactions directly contribute to the exacerbation of neuronal damages in AD is unclear.

[0215] Overall, these data suggest that infiltrating CD8+ T cells exhibit transcriptional signatures related to increased IFN-associated pathways, cytotoxicity, inflammatory activation, and reactive gliosis in the presence of microglia and AD neuropathology. Our results indicate that infiltrating CD8+ T cells activates IFN / inflammatory pathways in microglia and astrocytes, resulting in a synergistic, vicious cycle that exacerbates cellular damage in AD. Example 8. Upregulation of CXCL10 and its receptor, CXCR3, in the AD neuroimmune axis model and 5XFAD brain.

[0216] Chemokines govern the recruitment of peripheral immune cells by attracting cells that express their cognate ligands41 44. Therefore, we set out to identify the cytokines and chemokines that may regulate infiltration of T cells into the AD cultures (Fig. 19a). First, we measured secreted cytokine and chemokine levels in the AD and control cultures. Conditioned cell culture media were collected from cell cultures of the PiChip BRAIN compartment at week 4 of differentiation, and the levels of cytokines / chemokines / growth factors were measured using high-throughput human Kiloplex quantitative proteomics arrays (Fig. 19a-b, and Fig. 5a). We observed a higher production of basic fibroblast growth factor (FGF-2) in 3D-AD Neu / AC / iMGL cultures as compared to control by 2.5-fold (P=0.0168) (Fig. 19b). Notably, protein levels of several cytokines, including IL-8, IL-9, and IFN-a2 were substantially increased by 40-fold (P<0.0001), 8.5-fold (P=0.0007), 4.8-fold (P=0.0018), respectively, following the addition of microglia to AD cultures (Fig. 5a), as compared to AD cultures lacking microglia. We also observed a significant, albeit more modest increase in IL-8 and IL-IRA - 1.1-fold ( =0.014) and 3-fold (P=0.029), respectively - in 3D-AD Neu / AC / iMGL cultures versus control (Fig. 19b). In addition, levels of several chemokines, including CCL2, CCL5, CXCL12, and CXCL1, showed a slight increase following the addition of microglial cells, but not significant (Fig. 5a). In contrast, CXCL10 was the only chemokine that exhibited a significant increase in AD cultures compared to control, both in the absence (Neu / AC) and presence of microglia (Neu / AC / iMGL) by 5.3-fold (P=0.0033) and 27- fold ( O.OOOl), respectively (Fig. 5a). Additionally, we found a highly enriched astrocytic expression of CXCL10 in AD compared to CTRL in the PiChip system using scRNA-seq (P<0.0001, Fig. 5b). To further exploring whether AD conditions impact astrocyte secreted CXCL10, we quantitively measured CXCL10 level in media using cell culture inserts. We found a substantial increase of CXCL10 release in the astrocyte compartment with AD NPCs differentiated in the other compartment compared to CTRL NPCs ( =0.011) or astrocytes alone (P=0.0008, Fig. 5c). This indicates that AD pathologies induce the release of CXCL10 chemokine from astrocytes. Next, we investigated whether CXCL10, which we found to be significantly elevated in the AD versus control cultures, modulates T cell recruitment. For this purpose, we first examined the migration of T cells toward purified CXCL10 chemokine in microfluidic chips in the absence of neuronal cells (Fig. 5d-e). CD3+ T cell migration toward CXCL10 showed liner dose-dependence behavior for chemotaxis in response to the chemoattractant (Fig. 19c). Specifically, the number of CD3+ T cells that migrated towards 1000 nM CXCL10 was 2.63-fold ( =0.044) and 4.6-fold ( =0.0168) higher compared to 100 and 5 nM concentrations, respectively (Fig. 19d). Also, the number of CD3+ T cells that migrated towards 100 nM CXCL10 was 1.8-fold (P=0.024) higher compared to the 5 nM concentration (Fig. 19d), while CD3+ T cells showed no significant difference in the number of migrating CD3+ T cells for 50 nM versus 100 nM concentrations (Fig. 19d). T cells moved through the migration channels at an average velocity of 1.8 ± 0.8 pm / min toward CXCL10 chemokine, but with no significant differences between varying CXCL10 concentrations (Fig. 19e). Next, we isolated CD4+ and CD8+ T cells from human blood (n=7 donors) and tested their CXCL10 chemotaxis. We found a significant migration of both CD4+ and CD8+ T cells (P<0.0001) in response to 100 nM recombinant human CXCL10 chemokine, as compared to media without CXCL10 (Fig. 5f). Specifically, 322 ± 36 of CD4+ T cells (PO.OOOl, compared to media) and 168 + 28 of CD8+ T cells (PO.OOOl, compared to media) showed directional migration towards 100 nM CXCL10 (Fig. 5f). CD4+ T cells showed an average velocity of 1.3 ± 0.04 pm / min toward 100 nM CXCL10. We also found that CD8+ T cells moved slightly faster than CD4+ T cells with an average velocity of 1.5 ± 0.04 pm / min in response to 100 nM CXCL10 ( =0.01, Fig. 5g). These data strongly support that the elevated CXCL10 chemokine in AD cultures can induce T cell chemotactic migration.

[0217] We also confirmed the upregulation of CXCL10 in the parenchyma of 5XFAD and WT mice using immunohistochemical analyses (Fig. 5j). CXCL10 significantly elevated in both hippocampus (P=0.0005) and cortex (P=0.015) of 6-7 -month-old 5XFAD mice (n=13) as compared to the aged-matched WT (n=8) controls (Fig. 5h- i). In 5XFAD mice, we found that the majority of elevated CXCL10 staining were colocalized with GFAP astrocyte marker (Fig. 5j). However, we found no specific increases of CXCL10 staining around Ab plaques (3D6-stained) in 6-7 -month 5XFAD mice (data not shown).

[0218] CXCL10 induces T-cell chemotaxis through binding to CXCR3, a G-protein- coupled 7-transmembrane-spanning receptor45. To study whether infiltrating CD8+ T cells express CXCR3 receptor, we performed immunohistochemical analysis using antibodies against CXCR3 receptor and CD8. The number of CD8+ T cells significantly elevated in 6-7-month 5XFAD mice compared to age-matched WT controls (Fig. 2, =0.0031, and P=0.0014 for the percentage and number, respectively). Importantly, we found that the vast majority of CXCR3 staining was co-localized with brain-infiltrating CD8+ T cells in hippocampus and cortex of 6-7- month 5XFAD mice (Fig. 5k), This is in line with the scRNA-seq in which we found a substantial increase in CXCR3 expression level in infiltrating CD8+ T cells compared to other subsets of T cells (Fig. 51). Together, these data highlight a pivotal role of CXCR3 / CXCL10 signaling pathway in driving infiltration of activated CD8+ T cells in AD.

[0219] Example 9. Inhibition of CXCL10 binding to CXCR3 decreases CD8+ T cell infiltration and neurodegeneration in human neuroimmune axis model.

[0220] We blocked CXCR3, the chemokine receptor for CXCL10, to determine the role of CXCL10-CXCR3 axis in regulating T cell infiltration into the AD cultures (Fig. 6a). First, we performed a detailed characterization of CD4+ and CD8+ T cells and CXCR3 expression in human blood samples using flow cytometry. We found that CD4+ T cells constitute 68%, while CD8+ T cells account for 32% of total T cells in peripheral human blood (n=10 donors, Fig. 6b). We also found that the vast majority of CD45+CD8+ T cells (77.6 ± 2.3%) express CXCR3, significantly higher than CD45+CD4+ T cells, in which 39.1 ± 4.5% of them are positive for CXCR3 (P<0.0001, n=8 donors, Fig. 6c). These data suggest a cell-intrinsic characteristic of CXCR3 on infiltration and function of CD4+ and CD8+ T cells in the AD PiChip system. Early studies also demonstrated the role of CXCR3 in regulating T cell trafficking and function46 49.

[0221] Next, to block the CXCL10 binding to CXCR3, we used either an anti- CXCR3 blocking / neutralizing antibody (MAB-160) or a chemical antagonist of

[0222] CXCR3 (NBI-74330) and tested whether they could reduce the infiltration of CD3+ T cells into the AD cultures. CD3+ T cells isolated from peripheral blood of de- identified donors were immediately incubated with 10 pg / mL of MAB160 or NBI- 74330, loaded into the PiChip system, and then imaged using time-lapse confocal microscopy. Blocking of T cells with MAB160 led to a significant decrease of T cell infiltration into the AD cultures versus unblocked control (2 -fold decrease, =0.0108, n=8 donors, Fig. 19g). NBI-74330 also decreased T cell infiltration, but not significantly. Treating T cells with MAB-160 led to no significant impact on their average migration velocity into either AD or control cultures (n=8 donors, Fig. 19g). However, treating T cells with NBI-74330 significantly decreased the average migratory velocity of T cell infiltration into AD cultures by 32.7% ( =0.0146) versus unblocked control unblocked T cells (n=8 donors, Fig. 19h). Additionally, treating T cells with NBI-74330 significantly decreased the average migratory velocity of T cells into control culture by 26% ( =0.0427) versus T cells treated with MAB160 antibody (n=8 donors, Fig. 19h).

[0223] To elucidate the specific efficacy of the neutralizing antibody on CXCL10 / CXCR3-mediated CD8+ T cell infiltration and subsequent neurodegeneration to AD cultures, we isolated CD8+ and CD4+ T cells from de- identified donors, treated them with MAB-160, and characterized their infiltration behavior in the PiChip system. To determine what percentage of the CXCR3 surface expression in T cell subsets is blocked after treatment, CD4+ and CD8+ T cells were exposed to the MAB-160 (10 pg / mL) neutralizing antibody, and CXCR3 surface expression was measured by flow cytometry with an anti-CXCR3 antibody recognizing a different epitope (Fig. 6d-e). Quantitative analysis demonstrated a significant loss ( <0.0001) of CXCR3 surface expression in both CD4+ (99.9% decrease, w=8 donors) and CD8+ (99.8% decrease, n=6 donors) T cells after exposure to 10 pg / mL of MAB-160 neutralizing antibody (Fig. 6d-e). These data indicate that antibody treatment can significantly diminish the surface expression of CXCR3 on CD4+ and CD8+ T cells.

[0224] Next, we asked whether blocking the CXCL10-CXCR3 axis could specifically reduce the infiltration of CD4+ and / or CD8+ T cells into the AD cultures. CD4+ and CD8+ T cells were isolated from the peripheral blood of de-identified donors and were immediately incubated with MAB160 antibody (10 pg / mL), loaded into the PERIPHERAL compartments of the PiChip system and imaged using time-lapse confocal microscopy to track their migration. Blocking of CD4+ T cells with MAB160 led to a robust 21.4% decrease (P=0.0063) in their infiltration into the AD cultures versus unblocked control CD4+ T cells (n=8 donors, Fig. 6f). Similarly, we found that antibody treatment significantly decreased CD8+ T cells infiltration by 34.3% (P<0.0001) into the AD cultures versus unblocked control CD8+ T cells (n=7 donors, Fig. 6f). Moreover, we found that treating CD4+ and CD8+ T cells with MAB-160 had no significant impact on their average migration velocity compared to untreated control T cells (n=8 donors, Fig. 6g). Overall, our data from blocking CD3+, CD4+, and CD8+ T cells and their significant reduction in infiltration demonstrate that CXCL10 and its T cell receptor, CXCR3, play a key role in mediating selective T cell infiltration into AD versus control cultures.

[0225] Lastly, the current data raised another important question: whether blocking the CXCL10-CXCR3 chemokine axis modulate the detrimental impact of infiltrating CD8+ T cells in AD cultures. To answer this question, we tested the impact of infiltrating T cells with or without neutralizing antibody treatment on neural cell death in the AD PiChip brain model. MAB-160 treatments significantly reversed cellular damages by 36.9-fold ( <0.0001) and 5.2-fold ( <0.0001) in the AD cultures with T cells, as quantified by the GFP+ expressing area and cell count, respectively (Fig. 6i). To investigate the extent of neuronal damage, we also quantified the surface area of Tuj l -positive neurons. Again, MAB-160 treatments significantly increased Tuj l- positive cells by 34.6-fold ( <0.0001) in AD cultures with T cells (Fig. 6i). Collectively, these data demonstrate conclusive evidence of the important role of CXCL10 / CXCR3-mediated infiltration of T cells in AD, leading to pathogenic neuronal damage.

[0226] References

[0227] 1. Jorfi, M., Maaser-Hecker, A. & Tanzi, R. E. The neuroimmune axis of Alzheimer’s disease. Genome Med 15, 6 (2023).

[0228] 2. Bettcher, B. M., Tansey, M. G., Dorothee, G. & Heneka, M. T. Peripheral and central immune system crosstalk in Alzheimer disease — a research prospectus. Nat Rev Neurology 17, 1-13 (2021). 3. Arlehamn, C. S. L., Garretti, F., Sulzer, D. & Sette, A. Roles for the adaptive immune system in Parkinson’s and Alzheimer’s diseases. Curr Opin Immunol 59, 115-120 (2019).

[0229] 4. Wu, K.-M. et al. The role of the immune system in Alzheimer’s disease. Ageing Res Rev 79, 101409 (2021).

[0230] 5. Cao, W. & Zheng, H. Peripheral immune system in aging and Alzheimer’s disease. Mol Neurodegener 13, 1 17 (2018).

[0231] 6. Rogers, J., Luber-Narod, J., Styren, S. D. & Civin, W. H. Expression of immune system-associated antigens by cells of the human central nervous system: Relationship to the pathology of Alzheimer’s disease. Neurobiol Aging 9, 339-349 (1988).

[0232] 7. Itagaki, S., McGeer, P. L. & Akiyama, H. Presence of T-cytotoxic suppressor and leucocyte common antigen positive cells in Alzheimer’s disease brain tissue. Neurosci Lett 91, 259-264 (1988).

[0233] 8. Togo, T. et al. Occurrence of T cells in the brain of Alzheimer’s disease and other neurological diseases. J Neuroimmunol 124, 83-92 (2002).

[0234] 9. Merlini, M., Kirabali, T., Kulic, L., Nitsch, R. M. & Ferretti, M. T. Extravascular CD3+ T Cells in Brains of Alzheimer Disease Patients Correlate with Tau but Not with Amyloid Pathology: An Immunohistochemical Study. Neurodegener Dis 18, 49-56 (2018).

[0235] 10. Unger, M. S. et al. CD8+ T-cells infiltrate Alzheimer’s disease brains and regulate neuronal- and synapse-related gene expression in APP-PS1 transgenic mice. Brain Behav Immun 89, 67-86 (2020).

[0236] 11. Gate, D. et al. Clonally expanded CD8 T cells patrol the cerebrospinal fluid in Alzheimer’s disease. Nature 577, 399-404 (2020).

[0237] 12. Ferretti, M. T. et al. T-cell brain infiltration and immature antigen- presenting cells in transgenic models of Alzheimer’s disease-like cerebral amyloidosis. Brain Behav Immun 54, 211-225 (2016).

[0238] 13. Browne, T. C. et al. IFN-y Production by Amyloid P-Specific Thl Cells Promotes Microglial Activation and Increases Plaque Burden in a Mouse Model of Alzheimer’s Disease. J Immunol 190, 2241-2251 (2013). 14. MacPherson, K. P. et al. Peripheral administration of the soluble TNF inhibitor XProl595 modifies brain immune cell profiles, decreases beta-amyloid plaque load, and rescues impaired long-term potentiation in 5xFAD mice. Neurobiol Dis 102, 81-95 (2017).

[0239] 15. Shukla, A. K. et al. CDl la expression distinguishes infiltrating myeloid cells from plaque-associated microglia in Alzheimer’s disease. Glia 67, 844-856 (2019).

[0240] 16. Monsonego, A. et al. AP-induced meningoencephalitis is IFN-y-dependent and is associated with T cell-dependent clearance of Ap in a mouse model of Alzheimer’s disease. P Natl Acad Sci Usa 103, 5048-5053 (2006).

[0241] 17. Machhi, J. et al. CD4+ effector T cells accelerate Alzheimer’s disease in mice. J Neuroinflamm 18, 272 (2021).

[0242] 18. Pietronigro, E. et al. Blockade of a4 integrins reduces leukocyte- endothelial interactions in cerebral vessels and improves memory in a mouse model of Alzheimer’s disease. Sci Rep-uk 9, 12055 (2019).

[0243] 19. Gate, D. et al. CD4+ T cells contribute to neurodegeneration in Lewy body dementia. Science 374, 868-874 (2021).

[0244] 20. Wareham, L. K. et al. Solving neurodegeneration: common mechanisms and strategies for new treatments. Mol Neurodegener 17, 23 (2022).

[0245] 21. Gbtz, J., Bodea, L.-G. & Goedert, M. Rodent models for Alzheimer disease. Nat Rev Neurosci 19, 1 16 (2018).

[0246] 22. Scearce-Levie, K., Sanchez, P. E. & Lewcock, J. W. Leveraging preclinical models for the development of Alzheimer disease therapeutics. Nat Rev Drug Discov 19, 1-16 (2020).

[0247] 23. Blanchard, J. W., Victor, M. B. & Tsai, L.-H. Dissecting the complexities of Alzheimer disease with in vitro models of the human brain. Nat Rev Neurol 1-15 (2021 ) doi : 10.1038 / s41582-021-00578-6.

[0248] 24. Choi, S. H., Kim, Y. H., Quinti, L., Tanzi, R. E. & Kim, D. Y. 3D culture models of Alzheimer’s disease: a road map to a “cure-in-a-dish.” Mol Neurodegener 11, 75 (2016).

[0249] 25. Choi, S. H. et al. A three-dimensional human neural cell culture model of Alzheimer’s disease. Nature 515, 274-8 (2014). 26. Kim, Y. H. et al. A 3D human neural cell culture system for modeling Alzheimer’s disease. Nat Protoc 10, 985-1006 (2015).

[0250] 27. Kwak, S. S. et al. Amyloid-P42 / 40 ratio drives tau pathology in 3D human neural cell culture models of Alzheimer’s disease. Nat Commun 11, 1377 (2020).

[0251] 28. Park, J. et al. A 3D human tri culture system modeling neurodegeneration and neuroinflammation in Alzheimer’s disease. Nat Neurosci 21, 1 16 (2018).

[0252] 29. Abud, E. M. et al. iPSC-Derived Human Microglia-like Cells to Study Neurological Diseases. Neuron 94, 278-293. e9 (2017).

[0253] 30. McQuade, A. et al. Development and validation of a simplified method to generate human microglia from pluripotent stem cells. Mol Neurodegener 13, 67 (2018).

[0254] 31. Song, Y. et al. A dynamic view of the proteomic landscape during differentiation of ReNcell VM cells, an immortalized human neural progenitor line. Set Data 6, 190016 (2019).

[0255] 32. Doorn, K. J. et al. Microglia in Olfactory Bulb of AD and PD Patients. Brain Pathol 24, 152-165 (2014).

[0256] 33. Walker, D. G. Defining activation states of microglia in human brain tissue: an unresolved issue for Alzheimer’s disease. Neuroimmunol Neuroinflammation 2020, (2020).

[0257] 34. Guttikonda, S. R. et al. Fully defined human pluripotent stem cell-derived microglia and tri-culture system model C3 production in Alzheimer’s disease. Nat Neurosci 24, 343-354 (2021).

[0258] 35. Anderson, K. G. et al. Intravascular staining for discrimination of vascular and tissue leukocytes. Nat Protoc 9, 209-222 (2014).

[0259] 36. Yang, P., Ma, J., Yang, X. & Li, W. Peripheral CD4+ naive / memory ratio is an independent predictor of survival in non-small cell lung cancer. Oncotarget 8, 83650-83659 (2017).

[0260] 37. Browaeys, R., Saelens, W. & Saeys, Y. NicheNet: modeling intercellular communication by linking ligands to target genes. Nat Methods 17, 159-162 (2020).

[0261] 38. Gallart-Palau, X., Guo, X., Serra, A. & Sze, S. K. Alzheimer’s disease progression characterized by alterations in the molecular profiles and biogenesis of brain extracellular vesicles. Alzheimer ’s Res Ther 12, 54 (2020). 39. Zaheer, S., Thangavel, R., Sahu, S. K. & Zaheer, A. Augmented expression of glia maturation factor in Alzheimer’s disease. Neuroscience 194, 227- 233 (2011).

[0262] 40. Takata, K. et al. Role of high mobility group protein-1 (HMG1) in amyloid-P homeostasis. Biochem Bioph Res Co 301, 699-703 (2003).

[0263] 41. Schulz, O., Hammerschmidt, S. I., Moschovakis, G. L. & Forster, R. Chemokines and Chemokine Receptors in Lymphoid Tissue Dynamics. Annu Rev Immunol 34, 1-40 (2015).

[0264] 42. Ozga, A. J., Chow, M. T. & Luster, A. D. Chemokines and the immune response to cancer. Immunity 54, 859-874 (2021).

[0265] 43. Ley, K., Laudanna, C., Cybulsky, M. I. & Nourshargh, S. Getting to the site of inflammation: the leukocyte adhesion cascade updated. Nat Rev Immunol 7, 678-689 (2007).

[0266] 44. Sokol, C. L. & Luster, A. D. The Chemokine System in Innate Immunity. Csh Per spect Biol 7, a016303 (2015).

[0267] 45. Smit, M. J. et al. CXCR3 -mediated chemotaxis of human T cells is regulated by a Gi- and phospholipase C-dependent pathway and not via activation of MEK / p44 / p42 MAPK nor Akt / PI-3 kinase. Blood 102, 1959-1965 (2003).

[0268] 46. Groom, J. R. & Luster, A. D. CXCR3 in T cell function. Exp Cell Res 317, 620-631 (2011).

[0269] 47. Kurachi, M. et al. Chemokine receptor CXCR3 facilitates CD8+ T cell differentiation into short-lived effector cells leading to memory degeneration. J Exp et / 208, 1605-1620 (2011).

[0270] 48. Krummel, M. F., Bartumeus, F. & Gerard, A. T cell migration, search strategies and mechanisms. Nat Rev Immunol 16, 193-201 (2016).

[0271] 49. Maurice, N. J., McElrath, M. J., Andersen-Nissen, E., Frahm, N. & Prlic, M. CXCR3 enables recruitment and site-specific bystander activation of memory CD8+ T cells. Nat Commun 10, 4987 (2019).

[0272] 50. Lund, H. et al. Competitive repopulation of an empty microglial niche yields functionally distinct subsets of microglia-like cells. Nat Commun 9, 4845 (2018). 51. Lee, S. M., Hudobenko, J., McCullough, L. D. & Chauhan, A. Microglia depletion increase brain injury after acute ischemic stroke in aged mice. Exp Neurol 336, 113530 (2021).

[0273] 52. Laurent, C. et al. Hippocampal T cell infiltration promotes neuroinflammation and cognitive decline in a mouse model of tauopathy. Brain 140, 184 200 (2016).

[0274] 53. Xu, L. et al. T-cell infiltration, contribution and regulation in the central nervous system post-traumatic injury. Cell Proliferat 54, el3092 (2021).

[0275] 54. Kaya, T. et al. CD8+ T cells induce interferon-responsive oligodendrocytes and microglia in white matter aging. Nat Neurosci 25, 1446-1457 (2022).

[0276] 55. Zuena, A. R., Casolini, P., Lattanzi, R. & Maftei, D. Chemokines in Alzheimer’s Disease: New Insights Into Prokineticins, Chemokine-Like Proteins. Front Pharmacol 10, 622 (2019).

[0277] 56. Galimberti, D. et al. Intrathecal Chemokine Synthesis in Mild Cognitive Impairment and Alzheimer Disease. Arch Neurol-chicago 63, 538-543 (2006).

[0278] 57. Xia, M. Q., Bacskai, B. J., Knowles, R. B., Qin, S. X. & Hyman, B. T. Expression of the chemokine receptor CXCR3 on neurons and the elevated expression of its ligand IP-10 in reactive astrocytes: in vitro ERK1 / 2 activation and role in Alzheimer’s disease. J Neuroimmunol 108, 227-235 (2000).

[0279] 58. Wojcieszak, J., Kuczynska, K. & Zawilska, J. B. Role of Chemokines in the Development and Progression of Alzheimer’s Disease. J Mol Neurosci 72, 1929- 1951 (2022).

[0280] 59. Park, J., Baik, S. H., Mook-Jung, I., Irimia, D. & Cho, H. Mimicry of Central-Peripheral Immunity in Alzheimer’s Disease and Discovery of Neurodegenerative Roles in Neutrophil. Front Immunol 10, 2231 (2019).

[0281] 60. Zaheer, S. et al. Enhanced Expression of Glia Maturation Factor Correlates with Glial Activation in the Brain of Triple Transgenic Alzheimer’s Disease Mice. Neurochem Res 38, 218-225 (2013).

[0282] 61. Duan, R.-S. et al. Decreased Fractalkine and Increased IP-10 Expression in Aged Brain of APPswe Transgenic Mice. Neurochem Res 33, 1085-1089 (2008). 62. Krauthausen, M. et al. CXCR3 promotes plaque formation and behavioral deficits in an Alzheimer’s disease model. J Clin Invest 125, 365-78 (2015).

[0283] 63. Ozga, A. J. et al. CXCL10 chemokine regulates heterogeneity of the CD8+ T cell response and viral set point during chronic infection. Immunity (2021) doi: 10.1016 / j.immuni.2021.11.002.

[0284] 64. Gericke, C. et al. Early P-amyloid accumulation in the brain is associated with peripheral T cell alterations. (2023) doi: 10.1101 / 2023.01.17.524355.

[0285] 65. Piehl, N. et al. Cerebrospinal fluid immune dysregulation during healthy brain aging and cognitive impairment. Cell (2022) doi: 10.1016 / j. cell.2022.11.019.

[0286] 66. Sweeney, M. D., Sagare, A. P. & Zlokovic, B. V. Blood-brain barrier breakdown in Alzheimer disease and other neurodegenerative disorders. Nat Rev Neurol 14, 133-150 (2018).

[0287] 67. Sweeney, M. D., Kisler, K., Montagne, A., Toga, A. W. & Zlokovic, B. V. The role of brain vasculature in neurodegenerative disorders. Nat Neurosci 21, 1 14 (2018).

[0288] 68. Yang, A. C. et al. A human brain vascular atlas reveals diverse mediators of Alzheimer’s risk. Nature 1-8 (2022) doi:10.1038 / s41586-021-04369-3.

[0289] 69. Rustenhoven, J. & Kipnis, J. Bypassing the blood-brain barrier. Science 366, 1448-1449 (2019).

[0290] 70. Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 36, 411-420 (2018).

[0291] 71. Srivastava, A., Sarkar, H., Gupta, N. & Patro, R. RapMap: a rapid, sensitive and accurate tool for mapping RNA-seq reads to transcriptomes. Bioinformatics 32, i 192— i200 (2016).

[0292] OTHER EMBODIMENTS

[0293] It is to be understood that while the invention has been described in conjunction with the detailed description thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A system comprising: a microfluidic chip comprising: a first compartment, and a plurality of second compartments disposed along a periphery of the first compartment, wherein each of the second compartments is in fluid communication with the first compartment via a respective plurality of microfluidic channels, wherein the microfluid chip is configured to: receive a three-dimensional cell culture comprising first cells in the first compartment, receive second cells in the second compartments, and permit at least some of the second cells to migrate from the second compartments to the three-dimensional cell culture in the first compartment via the microfluidic channels.

2. The system of claim 1, wherein the second compartments are evenly distributed around the first compartment.

3. The system of claim 1, wherein the microfluidic chip comprises four second compartments.

4. The system of claim 1, wherein each of the microfluidic channels extends linearly between a respective one of the second compartments and the first compartment.

5. The system of claim 4, wherein the microfluidic chip further comprises a chamber aligned concentrically with the first compartment, and wherein the chamber is in fluid communication with the first compartment.

6. The system of claim 5, wherein the microfluidic channels are disposed on a first layer of the microfluidic chip,wherein the first compartment and the second compartment are disposed on a second layer of the microfluidic chip, and wherein the chamber is disposed on a third layer of the microfluidic chip.

7. The system of claim 5, wherein the first compartment has an inner radius of 2000 pm.

8. The system of claim 5, wherein each of the second compartments has an inner radius of 2500 pm.

9. The system of claim 5, wherein the chamber has an inner radius of 2500 pm.

10. The system of claim 1, wherein for each of the second compartments, at least 100 of the microfluidic channels extend in parallel between that second compartment and the first compartment.

11. The system of claim 1, wherein each of the microfluidic channels has: a height between 5 pm and 10 pm, preferably a height of 10 pm, a width between 5 pm and 10 um, preferably a width of 10 pm, and a length between 100 pm and 500 pm, preferably a length of 100 pm.

12. The system of claim 1, wherein the microfluidic chip comprises polydimethylsiloxane (PDMS).

13. The system of claim 1, wherein the first cells comprise brain cells.

14. The system of claim 13, wherein the brain cells comprise one or more of neurons, astrocytes, or microglial cells, preferably wherein the neurons, astrocytes, or microglial cells are human stem cell-derived.

15. The system of claim 14, wherein the brain cells comprise neurons, astrocytes and microglial cells, preferably wherein the neurons, astrocytes, and microglial cells are human stem cell-derived.

16. The system of claim 1, wherein the second cells comprise peripheral blood mononuclear cells (PBMCs).

17. The system of claim 16, wherein the second cells comprise one or more of T cells, B cells, or monocytes.

18. The system of claim 1, further comprising an imaging system configured to obtain one or more optical images of the first cells and the second cells in the microfluidic chip.

19. The system of claim 1, wherein the imaging system comprises a confocal microscope.

20. The system of claim 19, wherein the system further comprises as computer system configured to: access the one or more optical images, and determine, based on the one or more optical images, at least one of: a number of the second cells that migrated from the second compartments, or a velocity of migration of the second cells from the second compartments.

21. A method compri sing : obtaining a microfluidic chip comprising: a first compartment, and a plurality of second compartments disposed along a periphery of the first compartment,wherein each of the second compartments is in fluid communication with the first compartment via a respective plurality of microfluidic channels, depositing a three-dimensional cell culture comprising first cells in the first compartment, depositing second cells in the second compartments; and determining a migration of at least some of the second cells from the second compartments to the three-dimensional cell culture in the first compartment via the microfluidic channels.

22. The method of claim 21, wherein the first cells comprise brain cells.

23. The method of claim 22, wherein the first cells comprise one or more of, preferably all of, neurons, astrocytes, or microglial cells, preferably wherein the neurons, astrocytes, or microglial cells are human stem cell-derived.

24. The method of claim 21, wherein the second cells comprise peripheral blood mononuclear cells (PBMCs).

25. The method of claim 24, wherein the second cells comprise one or more of T cells, B cells, or monocytes.

26. The method of claim 21, further comprising obtaining, using an imaging system, one or more optical images of the first cells and the second cells in the microfluidic chip.

27. The method of claim 26, wherein the imaging system comprises a confocal microscope.

28. The method of claim 26, further comprising: accessing the one or more optical images, and determining, based on the one or more optical images, at least one ofa number of the second cells that migrated from the second compartments, or a velocity of migration of the second cells from the second compartments.

29. A method of treating a subject who has, or is at risk of developing, Alzheimer’s disease, the method comprising administering to the subject a therapeutically effective amount of an inhibitor of C-X-C motif chemokine receptor 3 (CXCR3).

30. The method of claim 29, wherein the inhibitor of CXCR3 is an antibody that binds to CXCR3.

31. The method of claim 30, wherein the antibody that binds to CXCR3 is selected from the group consisting of 4Ch, 4Hul, 4Hu2, 4Hu3, 4Hu4, 4Hu5, 4Hu6, 4Hu7, 4Hu8, 4Hu9, 4HulO, 4Hul l, 4Hul2, 4Hul3, 4Hul4, 4Hul5, 4Hul6, 4Hul7, 4Hul8, 4Hul9, 12Ch, 12Hul, 12Hu2, 12Hu3, 82Ch, 82Hul, 82Hu2, 82Hu3, 135Ch, 135Hul, 135Hu2, 135Hu3, 53Ch, 53Hul, 53Hu2, 53Hu3, 53Hu4, 53Hu5, 53Hu6, 53Hu7, 53Hu8, 53Hu9, 53Hul0, 53Hull, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, 53Hul8, 53Hul9, 53Hu29, B2, C12, F7, 2A3, 3A3, 3A11, 3C11, 4D5, 32B12, 32D12, Hl, B7, C12, D12, El, G12, 3A5, 3B12, 4C3, B12, C4, F7, 2A3, 3A7, 3A8, 23, 24, 25, 26, 27, 29, 30, 31, 33, 34, 35, 37, 53, 53Ch, 53Hu01, 53HuO2, 53Hu03, 53HuO4, 53Hu05, 53Hu06, 53Hu07, 53Hu08, 53Hu09, 53Hu010, 53Hul l, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, and 53Hul8.

32. The method of claims 30 or 31, wherein the antibody that binds to CXCR3 is human or humanized.

33. The method of claim 29, wherein the inhibitor of CXCR3 is a small molecule antagonist of CXCR3.

34. The method of claim 33, wherein the small molecule antagonist of CXCR3 is selected from the group consisting of ACT-777991, JN-2, ACT-672125, (±)-AMG 487, (±)-NBI 74330, N,N-Di substituted benzylamine derivatives, SCH 546738, NIBR2130, and AMG 487.

35. The method of claims 30-34, wherein the inhibitor of CXCR3 is administered intravenously.

36. An inhibitor of C-X-C motif chemokine receptor 3 (CXCR3), for use in a method of treating a subject who has, or is at risk of developing, Alzheimer’s disease.

37. The inhibitor for the use of claim 36, wherein the inhibitor of CXCR3 is an antibody that binds to CXCR3.

38. The inhibitor for the use of claim 37, wherein the antibody that binds to CXCR3 is selected from the group consisting of 4Ch, 4Hul, 4Hu2, 4Hu3, 4Hu4, 4Hu5, 4Hu6, 4Hu7, 4Hu8, 4Hu9, 4HulO, 4Hull, 4Hul2, 4Hul3, 4Hul4, 4Hul5, 4Hul6, 4Hul7, 4Hul8, 4Hul9, 12Ch, 12Hul, 12Hu2, 12Hu3, 82Ch, 82Hul, 82Hu2, 82Hu3, 135Ch, 135Hul, 135Hu2, 135Hu3, 53Ch, 53Hul, 53Hu2, 53Hu3, 53Hu4, 53Hu5, 53Hu6, 53Hu7, 53Hu8, 53Hu9, 53Hul0, 53Hull, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, 53Hul8, 53Hul9, 53Hu29, B2, C12, F7, 2A3, 3A3, 3A11, 3C11, 4D5, 32B12, 32D12, Hl, B7, C12, D12, El, G12, 3A5, 3B12, 4C3, B12, C4, F7, 2A3, 3A7, 3A8, 23, 24, 25, 26, 27, 29, 30, 31, 33, 34, 35, 37, 53, 53Ch, 53Hu01, 53HuO2, 53Hu03, 53HuO4, 53Hu05, 53Hu06, 53Hu07, 53Hu08, 53Hu09, 53Hu010, 53Hull, 53Hul2, 53Hul3, 53Hul4, 53Hul5, 53Hul6, 53Hul7, and 53Hul8.

39. The inhibitor for the use of claim 37 or 38, wherein the antibody that binds toCXCR3 is human or humanized.

40. The inhibitor for the use of claim 36, wherein the inhibitor of CXCR3 is a small molecule antagonist of CXCR3.

41. The inhibitor for the use of claim 40, wherein the small molecule antagonist of CXCR3 is selected from the group consisting of ACT-777991, JN-2, ACT-672125, (±)-AMG 487, (±)-NBI 74330, N,N-Di substituted benzylamine derivatives, SCH 546738, NIBR2130, and AMG 487.

42. The inhibitor for the use of claims 36-41, wherein the inhibitor of CXCR3 is formulated to be administered intravenously.

43. The method of any of claims 29 to 35, or the inhibitor for the use of any of claims 36 to 42, wherein the subject is a human subject.

44. The method or inhibitor for the use of claim 43, wherein the subject has been diagnosed with Alzheimer’s disease, or has been identified as having a risk factor for Alzheimer’s disease, optionally a family history of Alzheimer’s disease or one or more genetic risk factors for Alzheimer’s disease.