Fig 2: Overview of EPAM-ia method and OPN targeting as potential therapeutic approach in acute ischemic stroke. a Overview of the strategy applied for identification of novel candidates in multiple NVU cell types and their co-targeting in stroke. b Summary schematic showing dysregulation of OPN at the NVU and BBB impairment in ischemic stroke and their response to anti-OPN therapy

Fig 3: Osteopontin expression in normal and ischemic brain tissue of stroke patients. a Representative immunohistochemistry staining for osteopontin (OPN, brown) on human stroke samples at different stages (stage I–III) in the peri-infarct region and infarct core tissue. Stage I tissues were obtained 24–48 h post-vessel occlusion and present acute necrosis. Stage II is defined by macrophage resorption and stage III by the observation of pseudocystic cavity. b Quantification of OPN expression intensity (arbitrary unit, a.u.) in the infarct core, peri-infarct region and normal appearing tissue (NAT) at stages I–III; n = 6 individual specimens for each stage, **P < 0.01, ****P < 0.0001 and not significant (ns) P > 0.05 by one-way analysis of variance and Tukey’s multiple comparison test. c–f Representative images of immunofluorescence staining for OPN (red) and cell-specific markers (green) including CD31 for endothelial cells (c), PDGFRß for pericytes (d), GFAP for astrocytes (e) and IBA1 for microglia/macrophages (f) in the peri-infarct regions. (g–j) Quantification of OPN expression in core, peri-infarct and normal appearing tissue endothelial cells (g), pericytes (h), astrocytes (i) and microglia/macrophages (j) at stages I–III, n = 6 individual specimens for each stage, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ns P > 0.05 by one-way analysis of variance and Tukey’s multiple comparison test. Scale bars: 100 µm (a) and, 20 µm (c–f)

Fig 4: UpSet analysis of the NVU transcriptome reveals osteopontin as a potential therapeutic target in stroke. (a, top) UpSet plots demonstrating selected gene sets and their intersections (upper left) as well as corresponding genes numbers (upper right) revealing genes that are cell-specific shut down (expressed in contralateral hemisphere and regulated in one cell type only: Ed, Pd, Ad, Md), or induced (expressed only in ipsilateral hemisphere and regulated in only one cell type: Ei, Pi, Ai, Mi), or expressed in both hemispheres and regulated in only one cell type (E, P, A, M), or expressed in all four cell types of both hemispheres and either regulated in only one (Epam, ePam, epAm, epaM) or genes expressed and regulated in all four NVU cell types (EPAM). Lowercase letters (e, p, a, m) indicate expression gene sets and uppercase letters indicate regulated gene sets (E, P, A, M). (a, bottom) Venn diagram showing overlaps of differently regulated genes (expressed in all four cell types) and surrounded by pie charts giving the overall cell-specific number of expressed genes including non-regulated and regulated genes. b UpSet plots of gene sets (left) and corresponding genes numbers (right) regulated in one, two, three or four NVU cell types without considering the expression in the contralateral and/or ipsilateral hemispheres. c Interaction pathway of secreted phosphoprotein 1 (Spp1, encoding for osteopontin) and its signaling pathway members Cd44 (EiPiAM), Timp1 (EPAMi) and Mmp12‡ (EiPiAiM). These Spp1 pathway members were revealed by the UpSet analysis to be differentially regulated in all the NVU cell types after acute ischemic stroke in mice. d, g Visualization of actual base reads of gene expression by NVU transcriptome profiling shows upregulation of Spp1 (d) Cd44 (e), Timp1 (f) and Mmp12 (g). ‡It is of note that even though the Log2FC and the P value of Mmp12 fulfils the criteria to be defined as a gene regulated in all NVU cell types in our original transcriptomic analysis (g), Mmp12 will appear in the Mi intersection in the stricter UpSet representation. This is because the number of reads for Mmp12 was less than 10 in one out of six biological replicates for pericytes and astrocytes, and less than 10 in two out of six biological replicates for endothelial cells in the ipsilateral hemisphere, and also less than 10 in one out of six biological replicates for microglia in the contralateral hemisphere. n = 6, 3–4 mice/preparation, *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 determined by DESeq2 with Benjamini–Hochberg correction

Fig 5: Impairment of BBB by OPN and its response to anti-OPN therapy in vitro. a Schematic depicting the isolation and culture of primary porcine or mouse brain microvascular endothelial cells (PBMEC or MBMEC) followed by transendothelial electrical resistance (TEER) measurement across the endothelial monolayer or qRT-PCR experiment. Immunostaining of transwell inserts representing well-developed tight and adherens junctions of the PBMEC monolayer including CLDN5 (red) and VECAD (green). b Representative graph for continuous TEER values of the PBMEC monolayer treated with recombinant murine OPN (0.5 µg/mL) or vehicle (Control). c 12, 24 and 48 h TEER values of OPN-treated PBMEC normalised to their respective time-point control; n = 3 independent experiments, d Activation of OPN signaling was assessed by qRT-PCR on MBMEC treated 24 h with recombinant murine OPN and normalised to their untreated control. n = 4 independent experiments. e Neutralisation effect of the anti-OPN antibody (3 µg/mL) was assessed by qRT-PCR on MBMECs treated with recombinant murine OPN for 24 h n = 3–4 independent experiments. f Schematic depicting the experimental paradigm for oxygen-glucose deprivation (OGD, 1% O2, glucose-free basal medium) performed on MBMEC followed by qRT-PCR, staining of the cells, TEER measurement or permeability assay across the endothelial monolayer. Control condition cells were cultured in 19.5% O2 and 5.6 mM glucose containing basal medium. g Representative images of osteopontin (OPN, green) in endothelial monolayer 24 h post-OGD treatment. CD31 (white) was used as endothelial marker and DAPI (blue) to reveal nuclei. h Representative graph for continuous TEER values of the MBMEC monolayer in control normoxic conditions (treated with isotype control) and for OGD conditions—isotype control or anti-OPN antibody treated. i Quantification of 24 h TEER values of MBMEC treated with isotype control or anti-OPN antibody in control or OGD conditions; n = 4 independent experiment. j Permeability (pe) index values of fluorescent tracers of different molecular weight through MBMEC monolayer subjected to OGD and treated with isotype control or anti-OPN antibody. Results were normalized to pe index values obtained with inserts from each respective control condition (dashed line); n = 4 independent experiments. k BBB disruption by 24 h OGD and effect of the anti-OPN antibody treatment were assessed by qRT–PCR on MBMECs. Results were normalized to each respective control culture condition (dashed line); n = 4 independent experiments, *P < 0.05; **P < 0.01, ***P < 0.001 and ns P > 0.05 by two-tailed, paired t test (c–e, i–k). Scale bars: 10 µm (a, g)