Abstract

   PALS1-associated tight junction (PATJ) protein is linked to metabolic
   disease and stroke in human genetic studies. Despite the recognized
   role of PATJ in cell polarization, its specific functions in metabolic
   disease and ischemic stroke recovery remain largely unexplored. We
   explored the functions of PATJ in an in vitro model and in vivo in C.
   elegans and mice. Using a mouse model of stroke, we found post-ischemic
   stroke duration-dependent increase of PATJ abundance in endothelial
   cells. PATJ knock-out (KO) HEK293 cells generated by CRISPR-Cas9
   suggest roles for PATJ in cell proliferation, migration, mitochondrial
   stress response, and interactions with the Yes-associated protein
   (YAP)-1 signaling pathway. Notably, PATJ deletion altered YAP1 nuclear
   translocation. PATJ KO cells demonstrated transcriptional reprogramming
   based on RNA sequencing analysis, and identified dysregulation in genes
   central to vascular development, stress response, and metabolism,
   including RUNX1, HEY1, NUPR1, and HK2. Furthermore, we found that
   mpz-1, the homolog of PATJ, was significantly upregulated under hypoxic
   conditions in C. elegans. Knockdown of mpz-1 resulted in abnormal
   neuronal morphology and increased mortality, both of which were
   exacerbated by hypoxia exposure, indicating a critical protective role
   of PATJ/MPZ-1 in maintaining neuronal integrity and survival,
   particularly during oxygen deprivation stress relevant to ischemic
   stroke. These insights offer a new understanding of PATJ's regulatory
   functions within cellular and vascular physiology and help lay the
   groundwork for therapeutic strategies targeting PATJ-mediated pathways
   for stroke rehabilitation and neurovascular repair.

   Keywords: PATJ, Ischemic stroke, Vascular remodeling, Cellular stress,
   YAP1

Graphical abstract

   [33]Image 1
   [34]Open in a new tab

Highlights

     * •
       PATJ modulates hypoxic stress responses in C. elegans and mammalian
       cells, suggesting a role in ischemic injury resilience.
     * •
       PATJ expression in brain endothelial cells is dynamically regulated
       following ischemic stroke, suggesting its involvement in vascular
       repair processes.
     * •
       PATJ deletion induces nuclear translocation of YAP1, indicating a
       novel regulatory interplay with the Hippo pathway.
     * •
       Transcriptomic analysis reveals PATJ-dependent regulation of genes
       and pathways critical for vascular development, stress responses,
       and metabolism.

1. Introduction

   Stroke is a principal cause of mortality and prolonged disability in
   the United States. While thrombolysis with tissue-type plasminogen
   activator (tPA) or endovascular thrombectomy has improved outcomes
   following acute ischemic stroke (AIS) [[35]1], strategies to improve
   outcomes for stroke survivors are still notably lacking [[36]2].
   Angiogenesis, the process of new blood vessel formation from existing
   ones is tightly regulated under normal physiological conditions in
   adults [[37]3]. Following AIS, angiogenesis is a critical adaptation
   restoring cerebral blood flow and supporting repair processes [[38][4],
   [39][5], [40][6]]. Interestingly, an increased density of microvessels
   in brain regions impacted by AIS correlates with better prognosis in
   patients [[41]7], thus promoting angiogenesis following AIS is a
   promising method for enhancing neurorecovery from AIS.

   The evolutionarily conserved Crumbs protein homologs (CRB) are key
   components in scaffolding and cellular orientation. Central to this
   orchestration is the PALS1-associated tight junction (PATJ) protein, a
   component of the CRB complex. PATJ serves as a connector between the
   CRB–PALS1 complex and tight junction proteins [[42]8,[43]9],
   maintaining cell polarization and ensuring the stability of apical
   junctions [[44]10,[45]11]. However, recent advancements suggest a
   broader spectrum of PATJ's functionality including roles in signal
   transduction, cellular dynamics, and tumorigenesis [[46]12]. A recent
   genome-wide association study (GWAS) on acute ischemic stroke (AIS)
   outcomes revealed that AG or GG single nucleotide polymorphisms (SNPs)
   of the PATJ gene are strongly associated with poor 90-day modified
   Rankin Scale (mRS) outcomes post-AIS [[47]13]. Interestingly, SNPs in
   PATJ have also been implicated in cardiometabolic health and sleep
   disturbances [[48]14,[49]15]. These genome-scale findings provide
   genomic links between PATJ and cardiovascular pathology, suggesting
   that variations in PATJ not only impact cellular polarity but may also
   influence cardiovascular risk and AIS recovery outcomes.

   PATJ knockdown in human microvascular endothelial cells was found to
   promote endothelial to mesenchymal transition and pro-angiogenic
   programs [[50]16]. Prior work indicates an important role for PATJ in
   directing embryonic capillary migration in cooperation with angiomotin
   (AMOT) and Syx, a RhoA GTPase exchange factor (RhoGEF) protein also
   known as Pleckstrin homology domain containing, family G member
   (PLEKHG)-5 [[51]17]. The involvement of endothelial cells in vascular
   remodeling and angiogenesis is well-documented, yet the specific
   contributions of PATJ during post-AIS angiogenesis remains incompletely
   understood. Our study builds upon this recent expanded understanding of
   PATJ, and links PATJ with post-AIS angiogenesis, cellular migration,
   proliferation, and metabolism. We also found regulatory interplay
   between PATJ and key signaling pathways, notably YAP1 within the Hippo
   pathway. Taken together, these findings highlight PATJ's role in
   neurorepair-related pathways and suggest its potential as a novel
   therapeutic target.

2. Methods

2.1. CRISPR-Cas9 RNP complex preparation and cell transfection

   For the targeted editing of PATJ exon 3, single guilde (sg)RNA
   sequences sgRNA1 (CUUGACCCAAGAUAAACUGC) and sgRNA2
   (GUGAGUAUCUGGUUGAAGAG) were synthesized with high specificity and
   minimal off-target effects (Synthego). The sgRNAs were then complexed
   with Cas9 protein (IDT) to form ribonucleoprotein (RNP) complexes. The
   complexes were prepared by diluting sgRNA and Cas9 to 3 μM in Opti-MEM™
   I Reduced Serum Medium, adhering to a sgRNA:Cas9 molar ratio of 1.3:1,
   followed by a brief incubation.

   Human embryonic kidney (HEK)293 cells, cultured in Dulbecco's Modified
   Eagle Medium (DMEM) supplemented with 10 % Fetal Bovine Serum and
   antibiotics, were grown to 30–70 % confluence in 24-well plates to
   ensure optimal cell density for transfection. The transfection utilized
   Lipofectamine™ CRISPRMAX™ (Invitrogen) for RNP complex delivery,
   employing a reverse transfection technique where the RNP solution is
   pre-mixed with the cell suspension before plating. Following
   transfection, the cells were incubated at 37 °C and 5 % CO2, with
   medium changes to maintain cell health and promote recovery.

2.2. Isolation and validation of CRISPR-edited clones

   Following transfection, single cell derived clones were derived from
   singly plated cells, ensuring genetic uniformity. Over 2–8 weeks, these
   colonies were observed under standard conditions and expanded based on
   visual confirmation via microscopy. Successfully expanded colonies were
   then transferred to 24-well plates for further growth. Western blot was
   performed to validate protein knockdown. Sanger sequencing analysis
   post-PCR amplification from genomic DNA of individual clones revealed
   insertion/deletion in the region of PATJ exon 3 targeted by sgRNA1.

2.3. Migration assay

   Confluent HEK293 cell monolayers, including wild-type (WT), PATJ
   knockout (KO), PATJ overexpression (Ov-PATJ), and KO with PATJ
   overexpression (KO + Ov-PATJ), were subjected to a scratch assay. PATJ
   overexpression was achieved by transfecting cells with a plasmid
   encoding full-length human PATJ using Lipofectamine 2000 (Invitrogen)
   according to the manufacturer's instructions. An empty vector was used
   as a control. Where indicated, cells were treated with verteporfin
   (1 μM) to inhibit YAP1 activity. A sterile pipette tip was used to
   create a straight 'wound' across the monolayer. Cell proliferation was
   inhibited using Mitomycin C (10 μg/mL). The migration of cells into the
   wound area was imaged at specified intervals to assess closure rate.
   Quantitative analysis was conducted by measuring the gap width at 0-,
   24-, and 48-h post-scratch. Images were captured under a Zeiss AXIO
   Observer.A1 Inverted Fluorescence Microscope. Cellular migration was
   determined by ImageJ software. Relative migration rate was calculated
   as (Areawound at 0 h – Areawound at 24 h)/Areawound at 0 h.

2.4. Rotenone-induced cell growth assessment in PATJ-modified HEK293 cells

   HEK293 cells, both with and without PATJ knockout, were cultured in
   24-well plates at a seeding density of 10^5 cells per well. For cell
   proliferation and survival assays PATJ KO and WT cells were cultured
   under normal conditions (Vehicle), with rotenone treatment (200 nM), or
   with combined rotenone (200 nM) and verteporfin (1 μM) treatment.
   Verteporfin is a specific inhibitor of YAP1 activity. The rotenone
   concentration was selected to inhibit mitochondrial complex I, thereby
   simulating metabolic stress analogous to ischemic conditions, while
   verteporfin was used to determine whether YAP1 inhibition could
   modulate the observed effects of PATJ knockout on rotenone sensitivity.
   Cell proliferation was assayed by automated cytometry quantification of
   cell number at specified intervals post-treatment (Nanoentek, Waltham,
   MA, USA). Propidium iodide staining followed by automated cytometry was
   performed for assessment of cell death following rotenone (2 μM).

2.5. Seahorse XF metabolic flux analysis

   Mitochondrial respiration in HEK293 cells, with and without PATJ
   modification, was evaluated using the Seahorse XF^e96 Extracellular
   Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA) with the
   Seahorse XF Cell Mito Stress Test Kit. Cells were plated at
   25,000 cells per well in XF^e96 microplates and incubated in standard
   culture medium. Following overnight incubation, cells were washed and
   equilibrated in unbuffered Seahorse XF Base Medium supplemented with
   glucose (4.5 g/L), l-glutamine (2 mM), and sodium pyruvate (1 mM) in a
   non-CO2 incubator at 37 °C for 1 h. Oxygen consumption rate (OCR) was
   recorded before and after sequential injections of mitochondrial
   inhibitors: oligomycin (1 μM), Trifluoromethoxy carbonylcyanide
   phenylhydrazone (FCCP; 1 μM), and rotenone/antimycin A (1 μM).
   Post-assay, the cells stained with Hoechst 33258 (0.1 μg/mL) and
   fluorescence intensity acquired at excitation/emission 355/460 nm to
   normalize the OCR data.

2.6. Assessment of stress responses in C. elegans

   C. elegans strains were maintained on nematode growth medium (NGM)
   plates seeded with Escherichia coli OP50 at 20 °C using standard
   procedures [[52]18]. The mpz-1(mib188[mpz-1::gfp]) II strain [[53]19]
   was used in this study. For stress treatments, L4 stage worms were
   subjected to either hypoxia (0.1 % O2) or normoxia for 72 h, or heat
   stress at 20 °C (control), 25 °C, or 28 °C for 72 h [[54]20]. For
   confocal imaging, treated MPZ-1GFP worms were randomly selected,
   immobilized with 10 mM sodium azide in M9 buffer, and aligned on 2 %
   agarose pads on glass slides. Images were acquired using a Leica TCS
   SPE confocal microscope with a 63 × objective lens, maintaining
   consistent settings across all samples. MPZ-1GFP abundance was
   quantified using ImageJ.

2.7. C. elegans strains and RNAi treatment

   C. elegans strains were maintained on nematode growth medium (NGM)
   plates seeded with Escherichia coli OP50 at 20 °C using standard
   procedures. For high-resolution single-neuron morphological analysis,
   we utilized the otIs181[Pdat-1mCherry; Pttx-3mCherry] strain [[55]21]
   crossed with the uIs69 [(pCFJ90) myo-2pmCherry + unc-119psid-1] strain
   [[56]22]. This combined strain specifically labels dopaminergic
   cephalic neurons (CEP) and AIY interneurons in the head region while
   providing enhanced neuronal sensitivity to RNAi treatment, enabling
   effective assessment of gene knockdown effects specifically within the
   labeled neuronal populations.

   RNAi knockdown of mpz-1 was performed by feeding worms with HT115
   bacteria expressing double-stranded RNA targeting mpz-1 from the
   embryonic stage. Empty vector (L4440) was used as a control. Day 3
   adult worms were subjected to 72 h of hypoxia (0.1 % O2) or normoxia.
   Worm survival was assessed by touch-provoked movement. Worms
   unresponsive to gentle prodding with a platinum wire were scored as
   dead.

   For confocal imaging, treated worms were randomly selected, immobilized
   with 10 mM sodium azide in M9 buffer, and aligned on 2 % agarose pads
   on glass slides. Images were acquired using a Leica TCS SPE confocal
   microscope with a 63 × objective lens, maintaining consistent settings
   across all samples. Individual neurons were classified as "abnormal" if
   they exhibited aberrant branching patterns, process fragmentation,
   blebbing, or irregular axonal/dendritic morphology compared to control
   neurons. The percentage of abnormal worms with abnormal mCherry-labeled
   neurons was calculated as the ratio of worms with neuronal
   abnormalities to the total number of worms examined in each condition.

2.8. Western blot

   Protein lysates were extracted from HEK293 cells, both PATJ KO and WT,
   using cell lysis buffer (Cell Signaling Technology, Danvers, MA)
   supplemented with Protease Inhibitor Cocktail (Cell Signaling
   Technology, Danvers, MA). Equal amounts of protein were separated on
   4–15 % precast gels (Bio-Rad) and then transferred onto polyvinylidene
   difluoride (PVDF, Bio-Rad) membranes. Membranes were blocked for 1 h at
   room temperature using Tris-buffered saline with 0.1 % Tween-20 (TBS-T)
   supplemented with 5 % non-fat dried milk. Membranes were then incubated
   overnight at 4 °C with primary antibodies diluted in TBS-T containing
   5 % bovine serum albumin (BSA). Primary antibodies used were PATJ
   (1:1000, Proteintech) and YAP1(1:2000, Proteintech). Subsequently,
   membranes were probed with appropriate HRP-conjugated secondary
   antibodies for 1 h at room temperature. Protein bands were visualized
   using an enhanced chemiluminescence (ECL) detection system (Azure
   Biosystems). Densitometric analysis was conducted to quantify protein
   expression levels, with α-tubulin employed as a loading control.

2.9. Cellular immunofluorescence

   HEK293 cells, both PATJ KO and WT, were cultured on coverslips until
   optimal confluence was reached. Cells were fixed in 4 %
   paraformaldehyde, then permeabilized with 0.1 % Triton X-100 in PBS for
   15 min. Non-specific binding was minimized by blocking with goat serum.
   Overnight incubation at 4 °C with primary antibodies targeting YAP1
   (1:200, Proteintech) was followed by application of goat anti-rabbit
   Alexa Fluor 488 secondary antibody (1:1000, Invitrogen). For
   conventional fluorescence microscopy, nuclei were stained with DAPI.
   For confocal microscopy, the cytoskeleton was additionally stained with
   CoraLite 594-Phalloidin (red, Proteintech). Coverslips were mounted and
   examined using either a Zeiss AXIO Observer.A1 Inverted Fluorescence
   Microscope or a Leica confocal microscope. Coverslips were mounted for
   examination by Leica confocal microscopy, focusing on YAP localization
   and expression. YAP1 subcellular localization was quantified using two
   complementary approaches: (1) Distribution pattern analysis: YAP1
   subcellular distribution was visually assessed in at least 50 cells per
   condition per experiment across four independent experiments using
   standard fluorescence microscopy. Cells were classified into three
   categories based on YAP1 distribution: predominantly nuclear,
   predominantly cytoplasmic, or distributed in both compartments. The
   percentages of cells in each category were calculated for each
   experimental condition. (2) Nuclear-to-cytoplasmic intensity ratio
   measurement: For quantitative assessment of YAP1 subcellular
   distribution, confocal microscopy images were analyzed using ImageJ
   software. Regions of interest (ROIs) were defined within the nuclear
   area (identified by DAPI staining) and in the surrounding cytoplasmic
   area of each cell. Mean fluorescence intensity of YAP1 was measured in
   these ROIs, and the nuclear-to-cytoplasmic ratio was calculated. A
   minimum of 50 cells per condition per experiment were analyzed across
   four independent experiments. All image acquisition parameters were
   kept constant between samples to ensure comparable measurements.

2.10. Mouse model of transient focal cerebral ischemia

   Transient focal cerebral ischemia was induced in C57BL6 mice aged 8–10
   weeks (weighing 23–26 g) using the Middle Cerebral Artery Occlusion
   (MCAO) technique. Mice were anesthetized with isoflurane, delivered in
   a mixture of oxygen and nitrous oxide. A midline neck incision was made
   to expose the common carotid artery, followed by electrocoagulation of
   its branches. A nylon filament was carefully inserted through the
   external carotid artery to the internal carotid artery, and advanced
   towards the middle cerebral artery (MCA) to achieve a reduction in
   regional cerebral blood flow (CBF) to less than 25 % of the baseline
   values. Monitoring of the regional cerebral blood flow was conducted
   through Laser Doppler Flowmetry (Moor Instruments Ltd, UK). The mice's
   body temperature was maintained at 37.0 ± 0.5 °C using a regulated
   heating pad, monitored with a rectal thermometer throughout the
   procedures. The occlusion was sustained for 60 min, after which
   reperfusion was initiated. Mice were then allowed a recovery period
   ranging from 1 to 28 days. For the control group, sham operations were
   performed where all steps were identical except for the artery
   occlusion. Postoperative care included the administration of
   buprenorphine for pain relief. Mice failing to meet the specified CBF
   criteria or those that did not survive the procedure (n = 5) were
   excluded from subsequent analyses.

2.11. Immunohistochemistry

   After transient Middle Cerebral Artery Occlusion (tMCAO), mice were
   anesthetized deeply and perfused transcardially, first with 0.9 % NaCl,
   followed by 4 % paraformaldehyde in PBS. Brains were then extracted,
   cryoprotected in 30 % sucrose in PBS for 48 h, and sectioned coronally
   on a cryostat. Sections were preserved at −20 °C in a cryoprotectant
   solution until required for analysis or immunofluorescence staining.
   For staining, sections were washed in PBS, permeabilized with PBST (1 %
   Triton X-100 in PBS) for 20 min and blocked with 5 % normal goat serum
   for an hour. Following further washing with 0.3 % PBST, they were
   incubated overnight at 4 °C with rat anti-mouse CD31 (1:200, BD
   Pharmingen) and rabbit polyclonal PATJ antibody (1:100, LSBio). Alexa
   Fluor 488 conjugated goat anti-rat (1:1000, Invitrogen) or Alexa Fluor
   594 goat anti-rabbit (1:1000, Invitrogen) secondary antibodies were
   applied respectively. For NeuN/PATJ staining, sections underwent
   incubation with NeuN (1:1000, LSBio) and PATJ antibodies, followed by
   the appropriate Alexa Fluor-conjugated secondary antibodies.
   Immunofluorescence images were captured using a Zeiss AXIO Observer.A1
   Inverted Fluorescence Microscope. For YAP1/CD31/DAPI triple staining,
   sections were incubated overnight at 4 °C with rat anti-mouse CD31
   (1:200, BD Pharmingen) and rabbit polyclonal YAP1 antibody (1:200,
   Proteintech). Alexa Fluor 594 conjugated goat anti-rat (1:1000,
   Invitrogen) and Alexa Fluor 488 goat anti-rabbit (1:1000, Invitrogen)
   secondary antibodies were applied respectively. Sections were then
   mounted with VECTASHIELD Vibrance® Antifade Mounting Medium with DAPI
   (H-1800) for nuclear counterstaining. For triple immunofluorescence
   staining of CD31/YAP1/PATJ, sections were incubated overnight at 4 °C
   with rat anti-mouse CD31 (1:200, BD Pharmingen), mouse monoclonal YAP1
   antibody (1:200, Proteintech), and rabbit polyclonal PATJ antibody
   (1:100, LSBio). Prior to applying the mouse primary antibody, sections
   were treated with a mouse-on-mouse blocking kit (Vector Laboratories)
   according to the manufacturer's instructions to minimize non-specific
   binding. After washing, sections were incubated with Alexa Fluor 594
   conjugated goat anti-rat for CD31, Alexa Fluor 488 goat anti-mouse for
   YAP1, and Alexa Fluor 405 goat anti-rabbit for PATJ secondary
   antibodies (all 1:1000, Invitrogen) for 1 h at room temperature. Images
   were acquired using a CSU-W1 SORA spinning disk confocal microscope.

2.12. RNA-seq analysis of HEK293 cell lines

   Total RNA was extracted from HEK293 cells across three experimental
   groups: PATJ KO, WT, and PATJ KO with overexpression, utilizing the
   RNeasy Mini Kit (Qiagen). Prior to library preparation, the integrity
   and concentration of RNA were verified. Sequencing libraries were
   constructed following cDNA synthesis, adaptor ligation, and
   amplification. Library quality control followed by next-generation
   sequencing was performed on the DNBSeq platform by a commercial
   provider (BGI Americas, San Jose, CA). Reads were aligned to the human
   genome (Hgv38) for differential gene expression analysis using the
   limma R package (version 3.40.6; [[57]23]), which utilizes generalized
   linear models. Volcano plots and heatmaps were generated with the
   ggplot2 (version 3.4.0; [[58]24]) and pheatmap (version 1.0.12) R
   packages, respectively. For dimensionality reduction, the umap R
   package (version 0.2.7.0) was used, normalizing expression profiles
   (z-score) for UMAP analysis to generate a reduced-dimensionality
   matrix. Gene set enrichment analysis for both KEGG pathways and GO
   annotations was performed using the latest annotations from the KEGG
   REST API and org.Hs.eg.db R package (version 3.15.0), respectively. The
   clusterProfiler R package (version 4.4.4) mapped genes to these
   annotations, identifying statistically significant pathways and GO
   terms. Enrichment analyses considered a minimum gene set of 5, a
   maximum of 5000, and an FDR threshold of <0.05 as significant.

2.13. Bioinformatics analysis of PATJ expression in hypoxic endothelial cells

   Bioinformatics analyses using transcriptomic data sets available on
   Gene Expression Omnibus (GEO) to complement the present RNA sequencing.
   Six GEO RNA-seq datasets ([59]GSE107029, [60]GSE76743, [61]GSE70330,
   [62]GSE145774, [63]GSE186616, [64]GSE71216) were identified in human
   umbilical vein endothelial cells (HUVEC) exposed to hypoxic and/or
   control conditions. The samples consisted of 14 hypoxia (0.2 %–1 %,
   24-h treatment) and 14 normoxia treated samples. Raw FPKM data were
   downloaded, and batch effects were removed using the ComBat method from
   the "sva" R package. Principal component analysis (PCA) was used to
   evaluate batch effect removal.

   PATJ expression levels were compared between normoxic and hypoxic
   conditions using boxplots, with statistical significance assessed by
   Wilcoxon rank-sum test. For single-gene GSEA analysis, Pearson
   correlation between PATJ and all other genes was calculated using the
   "cor" function in R, and genes were ranked by correlation coefficient.
   The "GSEA" function from the "clusterProfiler" package was used with
   KEGG pathways as predefined gene sets. Pathways with |normalized
   enrichment score (NES)| >1 and p < 0.05 were considered significant.

   Correlation analyses between PATJ and genes involved in Hippo
   signaling, angiogenesis, and cell proliferation/migration were
   conducted using Pearson's correlation method. Correlation heatmaps and
   scatterplots were generated to visualize relationships between PATJ and
   functional genes of interest.

2.14. Statistical analysis

   All statistical analyses were performed using GraphPad Prism version 10
   or R. Data distributions were first assessed for normality using the
   Shapiro-Wilk test. Group comparisons were conducted using one-way or
   two-way analysis of variance (ANOVA), followed by Bonferroni post hoc
   correction for multiple comparisons. Comparisons between two groups
   were conducted using two-tailed unpaired Student's t-tests for normally
   distributed data with equal variances. Statistical significance was
   defined as p < 0.05. All data are presented as mean ± standard error of
   the mean (SEM).

3. Results

3.1. Human and mouse brain transcriptomic atlases reveal developmental
trajectory and cell subtype specific expression patterns of PATJ

   To extend prior human genetic work implicating PATJ in metabolic
   disease and stroke and provide further insights into PATJ function, we
   examined existing large mouse and human brain single cell RNA
   sequencing atlases profiling PATJ expression across neural cell types
   and the lifespan [[65][25], [66][26], [67][27]]. Interestingly, during
   fetal development and infancy PATJ is particularly enriched and levels
   decrease into adulthood. This pattern of expression is seen in many
   other key genes involved in neurogenesis and vasculogenesis such as
   ephrin B1, paired box protein (PAX)-6, neuronal differentiation
   (NeuroD)-1, and beta-catenin (CTNNB1), among others ([68]Fig. 1A).
   Although PATJ is enriched in early life, its expression in a broad
   array of cells is still present into adulthood in human and mouse
   single cell RNA atlas data ([69]Fig. 1B–C). In particular, single cell
   RNA sequencing atlas of whole human or mouse brain demonstrates that
   PATJ expression is most prominent in cerebellar glutamatergic cells,
   but also highly expressed in endothelial and other brain vascular cells
   (particularly those of the choroid plexus; [70]Fig. 1B–C). Of note, the
   cell-specific distribution of other key genes previously suggested to
   interact with PATJ including AMOT, PLEKHG5, and YAP1 demonstrate
   similar enrichment in cerebral vascular cell subtypes (see
   [71]Supplementary Fig. 1A–C).

Fig. 1.

   [72]Fig. 1
   [73]Open in a new tab

   PATJ expression across development, cell types and endothelial hypoxia.
   (A) Dotplot demonstrating expression of PATJ and other key regulators
   of neurogenesis and vasculogenesis in the cortex. PATJ expression is
   highest during fetal development and infancy and decreases in
   adolescence and adulthood. Data from Zhu et al., 2023 ([74]GSE204684)
   and accessed at [75]singlecell.broadinstitute.org. (B) Human and (C)
   Mouse single cell RNA sequencing brain atlas UMAP data demonstrating
   ubiquitous expression of PATJ in adult brain cell subtypes, with inset
   highlighting vascular cell subtypes. Data from Siletti et al., 2023 and
   Yao et al., 2023, respectively and accessed at
   [76]knowledge.brain-map.org. (D) Boxplot showing significantly
   increased PATJ expression in hypoxic versus normoxic HUVECs.
   ∗∗∗∗ p < 0.0001. (E) Gene-set enrichment analysis showing top pathways
   significantly associated with PATJ expression in HUVECs.

   We also analyzed PATJ RNA expression in existing GEO data sets of human
   umbilical vein endothelial cells (HUVECs) exposed to hypoxia (0.2–1.0 %
   for 24 h). After batch effect correction ([77]Supplementary Fig. 1D),
   we observed significant upregulation of PATJ expression in endothelial
   cells exposed to 24-h hypoxia compared to normoxia (p = 7.4 x 10^−7;
   [78]Fig. 1D). Co-expression gene set enrichment analysis revealed that
   PATJ-correlated genes in endothelial cells were significantly enriched
   in multiple pathways relevant to vascular remodeling, including Hippo
   signaling, PI3K-Akt signaling, MAPK signaling, extracellular matrix
   receptor interactions, and HIF-1 signaling ([79]Fig. 1E). Single-gene
   correlation analysis demonstrated strong positive associations between
   PATJ and key angiogenic factors, including VEGFA (r = 0.97), VEGFC
   (r = 0.94), and PECAM1 (r = 0.93) ([80]Supplementary Fig. 1E). Taken
   together, existing human and mouse cell type-specific gene expression
   data suggest functional roles for PATJ in both neuronal and vascular
   biology. Endothelial cell-specific analyses provide rationale for
   investigating PATJ's interactions with Hippo signaling and YAP1 and
   involvement in post-stroke vascular remodeling.

3.2. PATJ modulates cell proliferation and stress responses in mammalian
cells and C. elegans

   CRISPR-Cas9 and a sgRNA targeting PATJ was used to achieve knock-out of
   PATJ in two clonal HEK293 cell lines ([81]Fig. 2A). Edited and
   non-edited cells were screened by Western blot and sequencing
   identified two clonal lines with frameshift mutations that resulted in
   early stop codons and absence of PATJ mRNA and protein ([82]Fig. 2B–C)
   as well as wild-type cells exposed to editing procedures but without
   successful editing. We found a statistically significant discrepancy in
   growth between PATJ KO and WT cells ([83]Fig. 2D). To investigate
   differences in metabolic responses related to PATJ deficiency, we used
   a Seahorse mitochondrial stress assay. Interestingly, differences in
   oxygen consumption were observed in vitro following administration of
   oligomycin and rotenone/antimycin ([84]Fig. 2E) while preserving spare
   respiratory capacity. The sensitivity to oligomycin, an inhibitor of
   ATP synthesis, and rotenone/antimycin, inhibitors of complex I and
   complex III, suggests a greater dependence on oxidative phosphorylation
   in PATJ KO cells. Rotenone, a mitochondrial complex I inhibitor, was
   used to model energetic stress in PATJ-deficient contexts. Rotenone
   treatment increased cell death in PATJ KO cells after 48 h compared to
   wildtype HEK293 cells ([85]Fig. 2F). This data collectively indicates
   that PATJ contributes to cellular fitness and adaptability in
   conditions of metabolic stress.

Fig. 2.

   [86]Fig. 2
   [87]Open in a new tab

   Generation and metabolic profiling of PATJ knock-out (KO) HEK293 cells.
   (A) Schematic representation of CRISPR-Cas9 strategy targeting exon 3
   of the PATJ gene.) CRISPR-Cas9 introduced an insertion (KO1) or
   deletion (KO2) leading to a frameshift and premature stop codon (red
   underline). Lowercase sequence base pairs denote intronic sequence. Ex:
   exon. (B) RNA-seq gene counts demonstrate the absence of PATJ
   transcripts in KO versus WT cells, (n = 3 for each group). (C) Western
   blot verifying the absence of PATJ protein expression in knockout cell
   lines. (D) Growth comparison of PATJ knockout (KO) and wild-type (WT)
   HEK293 cells (n = 3 independent experiments with 3 technical replicates
   for each group). (E) Oxygen consumption rates of PATJ KO and WT cells
   following treatment with oligomycin (i), FCCP (ii), and antimycin
   A/rotenone (iii; n = 4–6 replicates for each group). (F) PATJ-KO cells
   demonstrate higher cell death as measured by propidium iodide staining
   than WT after 2 μM rotenone for 24 h (n = 3 independent experiments
   with 3 technical replicates for each group). ∗ p < 0.05, ∗∗ p < 0.01,
   ∗∗∗ p < 0.001, ∗∗∗∗ p < 0.0001.

   Prior work has implicated PATJ in developmental processes and suggested
   interaction with cytoskeletal proteins [[88]28]. As such we
   hypothesized that cell migration may be affected in PATJ KO cells as
   assessed by an in vitro scratch assay. In this assay a mechanical
   scratch is introduced into cells grown at high confluence and the
   resulting ‘scar’ monitored for cell migration and closure. Cell
   proliferation is inhibited with the addition of mitomycin C. We found a
   notable reduction in cellular migration in PATJ KO cells compared to WT
   HEK293, while accelerated migration was seen in cells overexpressing
   PATJ ([89]Fig. 3). Quantitative analysis demonstrated a significant
   increase in the gap width of PATJ KO cells at 24 h, which persisted to
   48 h post-scratch. The migration impairment observed in PATJ KO cells
   was effectively rescued by overexpression of PATJ to levels comparable
   to WT cells. These results indicate a critical role for PATJ in cell
   migration following an injury.

Fig. 3.

   [90]Fig. 3
   [91]Open in a new tab

   Comparative analysis of cell migration in PATJ-modified HEK293 Cells.
   (A) Scratch assay time-lapse images at 0, 24, and 48 h post-wounding
   display cell migration for various HEK293 cell lines: wild-type (WT),
   PATJ knock-out (KO), PATJ overexpression (Ov-PATJ), and KO with PATJ
   overexpression (KO + Ov-PATJ). Dotted lines indicate scratch edges. (B)
   Quantitative analysis of cell migration at 24 h post-scratch. (C)
   Quantitative analysis of cell migration at 48 h post-scratch. n = 4
   independent experiments with 3 technical replicates. ∗p < 0.05,
   ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

   To investigate the evolutionary conservation of PATJ's role in stress
   responses, we examined the expression of MPZ-1, a C. elegans ortholog
   of PATJ, under hypoxic and heat stress conditions using a
   CRISPR-generated MPZ-1GFP reporter strain. This strain exhibits strong
   MPZ-1GFP expression in the head region of C. elegans ([92]Fig. 4A).
   Exposure to hypoxia (0.1 % O2) for 72 h resulted in a significant
   upregulation of MPZ-1GFP expression compared to normoxic controls
   ([93]Fig. 4B–C). In contrast, heat stress at 25 °C or 28 °C for 72 h
   did not significantly alter MPZ-1GFP levels relative to the 20 °C
   control ([94]Fig. 4D–E). These findings suggest that the induction of
   PATJ/MPZ-1 expression in response to hypoxia is specific and conserved
   in C. elegans.

Fig. 4.

   [95]Fig. 4
   [96]Open in a new tab

   PATJ modulates stress responses in C. elegans (A) Schematic diagram of
   C. elegans anatomy showing MPZ-1GFP is expressed in head neurons. (B)
   Representative confocal images of endogenous MPZ-1GFP expression in C.
   elegans under hypoxia (0.1 % O2) or normoxia for 72 h post L4 stages.
   Scale bars: 10 μm. (C) Quantification of MPZ-1GFP fluorescence
   intensity under hypoxic and normoxic conditions. n > 20 animals per
   condition. (D) Representative confocal images of MPZ-1GFP expression in
   C. elegans under heat stress at 20 °C, 25 °C, or 28 °C for 72 h post L4
   stages. Scale bars: 10 μm. (E) Quantification of the percentage of
   worms showing enhanced MPZ-1GFP fluorescence intensity under different
   temperature conditions. Data points represent three independent plates
   with >20 animals assessed per condition. (F) Representative confocal
   images of individual neurons in otIs181[Pdat-1mCherry; Pttx-3mCherry]
   worms following control RNAi or mpz-1 RNAi treatment under normoxic and
   hypoxic conditions. Scale bars: 2 μm. (G) Quantification of the
   percentage of abnormal worms with abnormal neurons under different
   conditions. Data points represent three independent experiments with
   >20 animals assessed per condition. (H) Death rates of L4440 control
   and mpz-1 RNAi worms subjected to hypoxia for 72 h ∗p < 0.05,
   ∗∗p < 0.01, ∗∗∗∗p < 0.0001.

   To investigate the role of PATJ/MPZ-1 in neuronal health with
   single-cell resolution, we examined individual neuronal morphology
   using the otIs181[Pdat-1mCherry; Pttx-3mCherry] strain crossed with
   uIs69 strain, which specifically labels dopaminergic CEP and AIY
   interneurons in the head region ([97]Fig. 4F–G). This strain
   combination provides enhanced RNAi sensitivity while enabling precise
   visualization of individual neuronal morphology within defined neuronal
   populations.

   Under normoxic conditions, control RNAi-treated worms displayed normal
   neuronal morphology characterized by typical branching patterns and
   intact cellular integrity. In contrast, mpz-1 RNAi treatment induced
   noticeable morphological abnormalities in individual neurons, including
   irregular branching patterns and aberrant neuronal architecture. When
   subjected to hypoxic stress, these morphological defects were markedly
   exacerbated in mpz-1 RNAi-treated worms, with neurons displaying severe
   structural abnormalities, loss of normal branching patterns, and
   compromised cellular architecture. Although control animals also showed
   some sensitivity to hypoxic stress, the morphological changes were more
   pronounced following mpz-1 RNAi treatment. Notably, hypoxia exposure
   also led to a higher mortality rate in mpz-1 RNAi worms compared to
   control worms ([98]Fig. 4H).

   These results indicate that loss of PATJ/MPZ-1 function leads to
   neuronal defects and increased vulnerability to hypoxia-induced death,
   underscoring its protective role in maintaining neuronal integrity and
   survival, especially during oxygen deprivation stress relevant to
   ischemic stroke.

3.3. Interplay between PATJ and YAP1 in cellular stress and subcellular
dynamics

   Given prior studies and bioinformatic evidence linking PATJ and YAP1 to
   the regulation of cellular architecture and stress responses through
   overlapping polarity and signaling pathways, we investigated PATJ and
   YAP1 functional interaction in mammalian cells. Western blot analysis
   demonstrated an upregulation in YAP1 protein levels in PATJ KO cells
   compared to WT ([99]Fig. 5A), suggesting a regulatory interplay between
   PATJ and YAP1 expression or stability. Immunofluorescence imaging
   revealed distinct differences in YAP1 subcellular distribution patterns
   between WT and PATJ KO cells ([100]Fig. 5B). Quantitative analysis
   showed that PATJ KO cells exhibited a significantly higher proportion
   of cells with predominantly nuclear YAP1 localization and a lower
   proportion with predominantly cytoplasmic localization compared to WT
   cells.

Fig. 5.

   [101]Fig. 5
   [102]Open in a new tab

   YAP1 expression and localization in PATJ-modified HEK293 cells and
   ischemic stroke.(A) Western blot analysis of YAP1 protein levels in WT
   and PATJ KO cells, with quantification showing a significant increase
   in YAP1 protein levels in KO cells (n = 4 for each group). (B)
   Immunofluorescence staining of YAP1 (green) DAPI staining of nuclei
   (blue). Yellow arrows indicate cells with predominantly cytoplasmic
   YAP1 localization, red arrowheads indicate predominantly nuclear YAP1
   localization, and white arrowheads indicate cells with YAP1 distributed
   in both compartments. Right panels show quantitative analysis of YAP1
   subcellular distribution patterns, presented as the percentage of cells
   with YAP1 localization predominantly in the nucleus, cytoplasm, or both
   compartments (n = 4 independent experiments with >50 cells/group). (C)
   Confocal microscopy with 3D reconstruction of YAP1 subcellular
   distribution in WT and PATJ KO cells. Right panel shows quantification
   of the YAP1 nuclear-to-cytoplasmic intensity ratio (n = 4 independent
   experiments with 50 cells/group). (D) Scratch-wound assay of WT and KO
   monolayers treated with verteporfin, captured at 0, 24, and 48 h
   post-scratch. Dashed lines denote wound edge (n = 4 independent
   experiments with 3 technical replicates/group). (E) Cell survival
   following 48 h exposure to rotenone with or without verteporfin (n = 3
   independent experiments with 3 technical replicates/group). (F)
   Representative images of endothelial cell marker CD31 (red) and YAP1
   (green) in brain tissue from sham and 28 days post-middle cerebral
   artery occlusion (MCAO) mice, with DAPI staining of nuclei (blue) and
   merged images highlighting co-localization. n = 4 mice per group; 3
   non-consecutive sections per mouse with 5 randomly selected
   peri-infarct fields analyzed per section. ∗ p < 0.05, ∗∗ p < 0.01,
   ∗∗∗ p < 0.001, ∗∗∗∗ p < 0.0001.

   This altered subcellular distribution was further corroborated through
   confocal microscopy with 3D reconstruction ([103]Fig. 5C). Measurement
   of the nuclear-to-cytoplasmic intensity ratio of YAP1 fluorescence
   demonstrated a significant increase in nuclear YAP1 enrichment in PATJ
   KO cells compared to WT cells. These findings suggest that PATJ may
   influence YAP1 subcellular localization, potentially affecting its
   transcriptional activity.

   The functional effects of pharmacological YAP1 inhibition were tested
   in HEK293 cells. Treatment with verteporfin, a YAP1 inhibitor, did not
   affect baseline proliferation. However, verteporfin also further
   exacerbated the migration deficit observed in PATJ KO cells relative to
   either WT or untreated KO controls ([104]Fig. 5D), suggesting that YAP1
   activity becomes particularly critical for cell motility in the absence
   of PATJ. Verteporfin also increased metabolic vulnerability to rotenone
   in PATJ KO cells ([105]Fig. 5E), with PATJ KO cells treated with both
   verteporfin and rotenone showing reduced survival compared to cells
   treated with rotenone alone. Together, these findings suggest that the
   elevated YAP1 activity and nuclear localization in PATJ-deficient cells
   act as a compensatory mechanism that supports survival and migration
   under metabolic stress. Inhibiting this compensatory YAP1 activity with
   verteporfin exacerbates PATJ KO cell vulnerability revealing a
   functional dependency between PATJ loss and YAP1 activity for cellular
   survival under conditions of metabolic stress. Notably, verteporfin had
   a less pronounced effect on WT cells exposed to rotenone, indicating
   that this YAP1-dependent compensatory mechanism becomes particularly
   critical in the absence of PATJ. These results establish an important
   interplay between PATJ and YAP1 in maintaining cellular resilience and
   motility under pathophysiological stress.

   To extend these findings to an in vivo context, we examined YAP1
   expression and localization in a mouse model of ischemic stroke.
   Immunohistochemical analysis of brain sections from mice subjected to
   tMCAO revealed a distinct alteration in YAP1 subcellular localization
   in endothelial cells 28 days post-stroke compared to sham-operated
   controls ([106]Fig. 5F). In sham group, endothelial YAP1 expression was
   primarily localized to the cell surface. In contrast, tMCAO mice
   exhibited a marked increase in nuclear YAP1 within endothelial cells in
   the peri-infarct region, as evidenced by the co-localization of YAP1,
   CD31, and the nuclear marker DAPI. Quantitative analysis of cells
   positive for all three markers confirmed a significantly higher number
   of endothelial cells with nuclear YAP1 in tMCAO mice compared to sham
   controls. These findings suggest that ischemic stroke promotes the
   nuclear translocation of YAP1 in endothelial cells, potentially as a
   vascular adaptive response to ischemic injury.

3.4. Temporal profiling of PATJ expression in endothelial cells following
stroke

   To further explore the changes in PATJ after ischemic stroke, we
   characterized the temporal patterns of PATJ expression following tMCAO
   in mice ([107]Fig. 6, [108]Fig. 7). We observed a reduction in PATJ
   expression predominantly within the neuronal cells in the penumbral
   regions. This decrease was statistically significant when compared to
   the sham-operated controls and mirrors the decrease in the neuronal
   marker, NeuN. As the post-stroke timeline progressed, a gradual yet
   significant increase in endothelial PATJ expression became evident. As
   the recovery phase progressed, the colocalization of PATJ within the
   peri-ischemic endothelium gradually increased. By day 28 post-tMCAO,
   not only did PATJ expression surpass its original levels, but its
   colocalization with endothelial markers also significantly increased
   beyond the baseline observed in sham-operated controls. This enhanced
   colocalization in the later stages post-stroke signifies a possible
   adaptive augmentation of PATJ in post-stroke endothelial response to
   drive vascular remodeling and repair.

Fig. 6.

   [109]Fig. 6
   [110]Open in a new tab

   Temporal expression of PATJ in endothelial cells post-tMCAO and
   colocalization with YAP1. (A)Schematic representation of the
   peri-ischemic region observed in the tMCAO mouse model. (B)
   Representative immunofluorescence images showing double staining of
   Patj (red) and CD31 (green) in brain sections from sham and
   tMCAO-treated mice at days 1, 3, 7, 14, and 28 post-stroke. (C)
   Quantitative analysis of Patj/CD31 double positive cells per mm^2 in
   the peri-ischemic region over time (D) Quantitative analysis of Patj
   positive cells in the peri-ischemic area over time. (E) Quantitative
   analysis of CD31 positive signal dynamics in the peri-ischemic zone.
   n = 4 mice per group; 3 non-consecutive sections per mouse with 5
   randomly selected peri-infarct fields analyzed per section. ∗∗p < 0.01,
   ∗∗∗∗p < 0.0001. (F) Triple immunofluorescence staining showing CD31
   (red), YAP1 (green), and PATJ (blue) in brain sections from sham and 7
   days post-tMCAO mice. The merged images demonstrate colocalization of
   all three markers, with white arrowheads indicating triple-positive
   regions in the vascular structures. (G) Quantification shows the
   percentage of CD31+/YAP1+/PATJ + triple-positive area in sham versus
   tMCAO 7d tissues. n = 4 mice per group; 3 non-consecutive sections per
   mouse with 5 randomly selected peri-infarct fields analyzed per
   section. ∗∗∗∗p < 0.0001.

Fig. 7.

   [111]Fig. 7
   [112]Open in a new tab

   Temporal shifts in Patj expression in neurons after tMCAO. (A)
   Representative immunofluorescence images showing double staining of
   Patj (red) and the neuronal marker NeuN (green) in brain sections from
   sham and tMCAO mice at 1, 3, 7, 14, and 28 days post-stroke. (B)
   Quantitative analysis of Patj/NeuN double-positive cells per mm^2 in
   the peri-ischemic region across the specified time points. (C)
   Quantitative analysis of Patj positive signal in the peri-ischemic
   region over time (D) Quantitative analysis of NeuN positive cells in
   the peri-ischemic zone at each time point post-tMCAO. n = 4 mice per
   group; 3 non-consecutive sections per mouse with 5 randomly selected
   peri-infarct fields analyzed per section. ∗p < 0.05, ∗∗p < 0.01, or
   ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

   To further investigate the spatial relationship between PATJ and YAP1
   in the vascular remodeling response, we performed triple
   immunofluorescence staining for CD31, YAP1, and PATJ in the
   peri-infarct region at 7 days post-tMCAO ([113]Fig. 6F–G). Confocal
   microscopy revealed triple colocalization of these proteins within
   regions of the remodeling vasculature, with quantitative analysis
   confirming significantly higher triple-positive area in tMCAO samples
   compared to sham controls. Within CD31-positive vessels, we observed
   that areas showing elevated YAP1 expression frequently also exhibited
   increased PATJ expression. These observations suggest a coordinated
   expression and potential interaction of YAP1 and PATJ during
   post-stroke vascular remodeling.

3.5. Transcriptomic profiling reveals PATJ's influence on vascular and
cellular processes

   We conducted RNA-sequencing on PATJ KO, PATJ KO with PATJ
   overexpression (KO + Ov), and WT cell lines, revealing distinct
   transcriptional landscapes. Principal component analysis (PCA)
   validated our approach, showing clear group segregation
   ([114]Supplementary Fig. 3A). A core set of 112 genes were
   differentially regulated across the three cell types, suggesting
   essential elements of the PATJ-dependent transcriptional program
   ([115]Supplementary Fig. 3B). Runt-related transcription factor 1
   (RUNX1) is critically involved in hematopoiesis and has been implicated
   in endothelial functions [[116]29], suggesting a potential link between
   PATJ modulation and vascular remodeling, a process vital in post-stroke
   recovery. Hairy/enhancer-of-split related with YRPW motif 1 (HEY1), a
   transcriptional repressor regulated by the Notch signaling pathway, is
   known to influence vascular development and integrity [[117]30],
   aligning with the observed changes in PATJ-mediated gene expression and
   pointing toward a role in angiogenesis and possibly neurogenesis.
   Nuclear protein 1 (NUPR1), involved in stress responses and cellular
   adaptation [[118]31], may offer insights into the cellular mechanisms
   by which PATJ contributes to the resilience and migration of cells in
   response to ischemic conditions induced by stroke. Furthermore,
   Hexokinase 2 (HK2), central to the glycolytic pathway, demonstrated
   markedly reduced expression levels in PATJ KO cells suggesting a
   greater dependence on oxidative phosphorylation and supporting the
   Seahorse assay results also showing a shift in metabolic requirements
   during cellular stress [[119]32]. Heatmap and volcano plot analyses
   further highlighted significant upregulation and downregulation of
   genes associated with immune response, growth regulation, and signal
   transduction, indicating PATJ's diverse regulatory impact ([120]Fig.
   8A–D). Gene Ontology (GO) and KEGG pathway analyses connected the DEGs
   to crucial stroke-related processes in the KO + Ov group, including
   vascular development and response to hypoxia, with pathways like 'p53
   signaling' and 'Notch signaling' also being enriched ([121]Fig. 8E–F).

Fig. 8.

   [122]Fig. 8
   [123]Open in a new tab

   Transcriptomic landscape and pathway analysis in PATJ-modulated
   HEK293 cells. (A,B,C) Volcano plots detailing significant gene
   expression changes for KO + Ov vs. KO, KO vs. WT, and KO + Ov vs. WT,
   with upregulated genes in red and downregulated genes in blue. (D)
   Heatmaps of gene expression contrasts for KO + Ov vs. KO vs. WT, with
   color intensity denoting gene expression levels. (E) Gene Ontology (GO)
   enrichment analysis bubble chart for the KO + Ov vs. KO comparison. (F)
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis bubble chart for the KO + Ov vs. KO comparison.

4. Discussion

   Ischemic stroke results in significant vascular and neuronal damage
   which initiates a cascade of immune-mediated secondary neuronal
   injuries in the days and months following stroke. In particular, damage
   to cerebral endothelial cells (ECs) increases vascular permeability and
   disrupts the blood-brain barrier which can further exacerbate stroke
   injury and impair neurorecovery [[124][33], [125][34], [126][35]].
   Thus, optimizing the beneficial effects of ECs during stroke recovery
   is an attractive target to improve stroke outcome [[127]36,[128]37].
   PATJ has well established roles in tight junction assembly and
   maintenance of epithelial cell polarity [[129][8], [130][9], [131][10],
   [132][11],[133][38], [134][39], [135][40], [136][41], [137][42]]. In
   our study, we extend recent findings linking PATJ to angiogenesis and
   provide additional evidence for PATJ involvement in cellular responses
   to metabolic stress in cells and C. elegans and the physiological
   response to stroke in vivo in mouse ECs [[138]16].

   The establishment of a PATJ KO in HEK293 cell lines, which revealed
   roles for PATJ in cell proliferation, cell migration, response to
   mitochondrial stress, and the regulatory dynamics with YAP1. Our
   findings reveal a significant reduction in cellular migration in PATJ
   KO HEK293 cells, which can be rescued by inducing the overexpression of
   PATJ. This aligns with previous studies highlighting PATJ's involvement
   in regulating directional migration of mammalian epithelial cells and
   interaction with cytoskeletal proteins [[139]9,[140]28]. Our data also
   reveal an enhanced vulnerability to mitochondrial stress in
   PATJ-deficient cells with reduced proliferation in rotenone and altered
   oxygen consumption following oligomycin, highlighting PATJ's
   contribution to cellular resilience against metabolic challenges. This
   aspect is particularly relevant in the context of stroke, where
   cellular energy demands are acutely altered. The ability of cells to
   withstand and repair tissues following ischemia is critical for
   neurorecovery, suggesting that interventions aimed at modulating PATJ
   expression or its targets could improve cellular resilience in ischemic
   injury.

   The regulatory role of PATJ in endothelial cells under hypoxic
   conditions is further supported by our bioinformatics analyses of
   multiple existing HUVEC RNA-seq datasets. We observed significant
   upregulation of PATJ in endothelial cells following hypoxia exposure,
   consistent with the increased PATJ expression seen in endothelial cells
   during the recovery phase post-stroke in our mouse model. The strong
   correlation between PATJ and key angiogenic factors such as VEGFA,
   VEGFC, and PECAM1 in HUVECs further suggests PATJ's involvement in
   angiogenic processes.

   The regulatory interplay between PATJ and YAP1, especially under
   conditions of PATJ depletion, introduces a new angle for further study
   in understanding EC response mechanisms to ischemic stroke, especially
   considering YAP1 is highly enriched in brain vascular cells (see
   [141]Supplementary Fig. 1; [[142]26]). The observed changes suggest
   that PATJ deletion may trigger YAP1 signaling compensatory mechanisms.
   Emerging evidence suggests that nuclear translocation of the
   transcription factor, YAP1 modulates the expression of genes involved
   in angiogenesis [[143][43], [144][44], [145][45], [146][46], [147][47],
   [148][48], [149][49], [150][50], [151][51], [152][52], [153][53],
   [154][54], [155][55]]. Intriguingly, our in vivo studies corroborate
   and extend these findings. We observed a marked increase in nuclear
   YAP1 within endothelial cells in the peri-infarct region, suggesting a
   potential vascular response to ischemic injury. While the precise role
   of YAP1 in post-stroke angiogenesis remains to be fully elucidated,
   this nuclear translocation may indicate its involvement in vascular
   adaptation processes. Our functional experiments with verteporfin, a
   specific inhibitor of YAP1 activity, provide evidence for a
   compensatory role of YAP1 signaling in PATJ-deficient cells under
   metabolic stress conditions. Inhibition of YAP1 reduced cell survival
   in PATJ KO cells exposed to rotenone-induced stress suggesting that the
   increased nuclear YAP1 localization observed in these cells may
   represent an adaptive response to preserve cellular resilience.

   These observations align with emerging evidence suggesting that YAP1
   can function as a stress-responsive factor that promotes cellular
   adaptation and survival under various stress conditions [[156][56],
   [157][57], [158][58], [159][59]]. In the context of PATJ deficiency,
   our data indicates that YAP1 activation may represent a critical backup
   mechanism that becomes essential for cellular resilience when
   PATJ-dependent pathways are compromised. This compensatory relationship
   may have important implications for understanding how cells adapt to
   ischemic injury and for developing therapeutic strategies that preserve
   or enhance these adaptive responses. Angiogenesis, which principally
   involves ECs, begins in the zones surrounding the ischemic event within
   the first 12 h and can persist for more than three weeks
   [[160]60,[161]61]. Angiogenesis is necessary for the repair and
   remodeling of vascular structures and also promotes neural plasticity,
   as evidenced in both clinical and preclinical stroke studies [[162]62].
   The increased nuclear localization of YAP1 in endothelial cells of the
   peri-infarct region could represent part of a complex adaptive
   response, potentially influencing vascular remodeling in the recovering
   brain tissue. Thus, further understanding the mechanistic connections
   between PATJ and YAP1 may aid in uncovering new therapeutics to enhance
   post-stroke angiogenesis and tissue repair.

   The dynamic regulation of PATJ expression in ECs following ischemic
   stroke, characterized by an initial decrease and subsequent increase,
   suggests a tightly controlled response to ischemic injury. This
   biphasic pattern might reflect the role of PATJ in interacting with EC
   repair processes post-stroke. Interestingly, our findings in C. elegans
   demonstrate that the specific upregulation of MPZ-1, the C. elegans
   ortholog of PATJ, under hypoxic conditions is particularly relevant to
   understanding PATJ's potential role in ischemic stroke adaptation. As
   ischemic stroke is characterized by reduced blood flow leading to
   hypoxia and metabolic stress in the affected brain regions, the
   observation that PATJ expression is induced by hypoxia in C. elegans
   suggests that this response may be an evolutionarily conserved
   mechanism to cope with oxygen deprivation.

   Our findings in the C. elegans model provide further evidence for the
   evolutionarily conserved role of PATJ in conferring neuronal resilience
   against hypoxic injury. Knockdown of the PATJ homolog mpz-1 resulted in
   abnormal neuronal morphology, an effect that was exacerbated when
   combined with hypoxia exposure in adult worms. Moreover, we observed a
   significant increase in mortality among mpz-1 deficient worms subjected
   to hypoxia, suggesting that PATJ/MPZ-1 is not only crucial for
   maintaining neuronal structural integrity but also for promoting
   overall organismal survival under hypoxic stress. This supports our
   mouse data showing altered PATJ expression in neurons following
   ischemic stroke and suggests that PATJ is critical for maintaining
   neuronal health and counteracting damage caused by oxygen deprivation.
   The hypoxia-sensitive phenotype and increased mortality in mpz-1
   deficient worms parallel the increased vulnerability of PATJ KO
   mammalian cells to mitochondrial stressors, emphasizing its function in
   adapting to metabolic challenges at both the cellular and organismal
   levels. Given that ischemic stroke triggers hypoxia and metabolic
   dysfunction, the evolutionarily conserved ability of PATJ to
   orchestrate adaptive responses may have important implications for
   understanding its influence on stroke outcomes. Future work will be
   necessary to understand whether modulating PATJ expression also
   modulates vascular remodeling to enhance stroke recovery. Conversely,
   the sustained decrease in neuronal PATJ expression following stroke
   further emphasizes the need for additional research in understanding
   the cell-type-specific regulation of PATJ to enhance stroke recovery.

   Our transcriptomic analysis, centered on PATJ's modulation, illuminates
   its pivotal role in regulating genes and pathways critical for vascular
   remodeling and recovery post-stroke. Notably, the downregulation of
   RUNX1 and HEY1 in PATJ KO cells, implicated in endothelial functions
   and angiogenesis [[163]29,[164]30,[165]63], underscores PATJ's
   influence on vascular development. This, coupled with NUPR1's
   involvement in cellular stress responses and HK2's role in metabolic
   adaptation [[166]31,[167]32], highlights metabolic influences regulated
   by PATJ. Notably, the upregulation of noggin (NOG) with PATJ
   overexpression and its down-regulation in the PATJ KO, suggests a
   previously unexplored connection between PATJ and this key regulator of
   neural development and neurogenesis [[168]64]. Similarly, RBM15 stands
   out with its substantial upregulation in the KO + Ov cell line and
   downregulation in the PATJ KO. RBM15 is integral to RNA metabolic
   processes, influencing not only mRNA splicing but also chromatin
   organization and gene expression of pathways critical for cellular
   identity and function. RBM15 dysregulation along with enrichment of the
   Notch signaling pathway in KEGG analysis aligns with our findings
   implicating PATJ in the modulation of vascular-related functions
   [[169]65]. Interestingly, the chemokine C-X-C motif receptor (CXCR)-4-
   was also downregulated in the PATJ KO cell line, and is essential for
   chemotaxis and hematopoietic cell lineages [[170]66] as well as
   cooperates with Notch signaling to facilitate arterial differentiation
   [[171]63]. Taken together, these findings support a role for PATJ in
   influencing cell migration and differentiation. In contrast, genes such
   as CHAC1 (involved in glutathione metabolism and apoptosis), SLC6A9 (a
   neurotransmitter transporter), and INHBE (a member of the TGF-beta
   superfamily) exhibited considerable downregulation in KO + Ov cells and
   upregulation in PATJ KO cells. These changes may reflect PATJ's
   involvement in the regulation of redox homeostasis, neurotransmission,
   and cell growth signaling pathways. The deregulation of these genes
   with genetic manipulation of PATJ expression coupled with the metabolic
   and cell biological phenotypes demonstrated in this study could suggest
   possible mechanistic targets for PATJ to exert effects on cellular
   stress responses, synaptic signaling, and tissue homeostasis. These
   findings, aligned with the validated GO and KEGG pathway functions,
   suggest PATJ's integral role in orchestrating cellular resilience,
   angiogenesis, and neurogenesis through pathways like Notch signaling.
   Our transcriptomic profiling in HEK293 cells suggests that PATJ may
   influence multiple molecular pathways pertinent to angiogenesis,
   metabolic regulation, and cellular stress responses. Though further
   research is required to understand the relevance of PATJ to these
   pathways following stroke in vivo, this is a valuable starting point
   for future investigations examining how PATJ might interface with
   cellular pathways relevant to ischemic remodeling and neurovascular
   repair. The significant colocalization of PATJ and YAP1 in endothelial
   cells post-stroke provides new insights into potential molecular
   interactions during vascular remodeling. Our triple immunofluorescence
   analysis demonstrated that these proteins colocalize within
   CD31-positive vessels in the peri-infarct region, specifically during
   the active angiogenic phase. This observation, combined with our in
   vitro evidence showing altered YAP1 localization in PATJ KO cells,
   points to a potential relationship between PATJ and YAP1 in the
   vascular response to ischemic injury. While the precise functional
   relationship between these proteins requires further investigation,
   their coordinated expression in remodeling vessels highlights a
   potentially important molecular interaction that may influence
   angiogenic responses after stroke. However, we acknowledge that
   additional pathways may converge on YAP1 during stroke-induced stress.
   Future studies employing conditional knockout models in endothelial
   cells—or advanced imaging techniques to monitor YAP1 subcellular
   dynamics in real time—could help distinguish whether loss of PATJ alone
   drives YAP1 nuclear translocation or if other ischemia-related signals
   also modulate YAP1. In particular, proximity ligation or
   co-immunoprecipitation assays might clarify whether PATJ physically
   binds YAP1 to govern its localization, or alternatively, acts through
   intermediary effectors. These directions represent a critical next step
   in delineating how PATJ deficiency intersects with YAP1's regulatory
   network under the complex conditions of stroke recovery.

   Extending the foundational work by Medina-Dols et al. on PATJ's role in
   modulating endothelial to mesenchymal transition (EndMT) through key
   pathways [[172]16], our investigation provides additional insights and
   corroborates key findings with complementary techniques. Employing
   CRISPR-Cas9 technology, we illuminate PATJ's broader influence on
   cellular dynamics and vascular remodeling. Our findings reveal
   significant PATJ-driven alterations in cellular migration and
   mitochondrial stress response, highlighting a regulatory interplay with
   the YAP1 signaling pathway, particularly highlighting YAP1's nuclear
   translocation—a phenomenon also partially supported by Medina-Dols and
   colleagues findings. Our RNA-seq analysis reveals extensive
   transcriptional reprogramming under PATJ modulation, identifying
   dysregulation in genes critical to vascular development, stress
   response, and metabolism. Building upon Medina-Dols et al.’s
   observations of sustained PATJ downregulation [[173]16], our in vivo
   analysis showcases a dynamic, time-dependent expression pattern in
   endothelial cells, suggesting PATJ's integral role in both the
   immediate response and subsequent recovery processes. This biphasic
   expression underscores the complexity of PATJ's involvement beyond
   EndMT, implicating it in essential pathways for vascular development
   and recovery. Thus, this work significantly extends our understanding
   of PATJ's role in stroke pathophysiology and supports further work in
   this pathway to pave new avenues for targeted therapeutic interventions
   in stroke rehabilitation and vascular repair.

   It is important to acknowledge the inherent limitations of the tMCAO
   model used in this study. Despite standardization efforts, biological
   variability in individual mice may influence the extent of ischemic
   injury and subsequent vascular responses. Additionally, while the tMCAO
   model is widely accepted for studying ischemic stroke, it does not
   fully recapitulate all aspects of human stroke pathophysiology,
   particularly regarding comorbidities commonly present in stroke
   patients. These limitations should be considered when translating our
   findings to clinical applications.

   In conclusion, our research advances the understanding of PATJ's
   multifaceted roles in cellular and vascular physiology, particularly
   its evolutionarily conserved function in modulating cellular responses
   to hypoxic stress, and highlights the potential of targeting
   PATJ-mediated molecular network for therapeutic intervention in stroke
   recovery. Future studies will explore the detailed mechanisms by which
   PATJ influences EC function and vascular remodeling in stroke recovery
   and other pathologic states, further illuminating its role in disease
   modulation and therapeutic potential.

CRediT authorship contribution statement

   Mengqi Zhang: Writing – original draft, Visualization, Validation,
   Methodology, Investigation, Formal analysis, Data curation. Wei I.
   Jiang: Visualization, Validation, Methodology, Investigation, Formal
   analysis. Kajsa Arkelius: Validation, Methodology, Investigation.
   Raymond A. Swanson: Writing – review & editing, Supervision, Project
   administration. Dengke K. Ma: Writing – review & editing, Supervision,
   Project administration, Investigation, Data curation. Neel S. Singhal:
   Writing – review & editing, Supervision, Resources, Investigation,
   Funding acquisition, Conceptualization.

Funding acknowledgments