Cerebral endothelial cell (EC) injury and blood-brain barrier (BBB) permeability contribute to neuronal injury in acute neurological disease states. Preclinical experiments have used animal models to study this phenomenon, yet the response of human cerebral ECs to BBB disruption remains unclear. In our phase I clinical trial (ClinicalTrials.gov NCT04528680), we used low-intensity pulsed ultrasound with microbubbles (LIPU/MB) to induce transient BBB disruption of peritumoral brain in patients with recurrent glioblastoma. We found radiographic evidence that BBB integrity was mostly restored within 1 hour of this procedure. Using single-cell RNA sequencing and transmission electron microscopy, we analyzed the acute response of human brain ECs to ultrasound-mediated BBB disruption. Our analysis revealed distinct EC gene expression changes after LIPU/MB, particularly in genes related to neurovascular barrier function and structure, including changes to genes involved in the basement membrane, EC cytoskeleton, and junction complexes, as well as caveolar transcytosis and various solute transporters. Ultrastructural analysis showed that LIPU/MB led to a decrease in luminal caveolae, the emergence of cytoplasmic vacuoles, and the disruption of the basement membrane and tight junctions, among other things. These findings suggested that acute BBB disruption by LIPU/MB led to specific transcriptional and ultrastructural changes and could represent a conserved mechanism of BBB repair after neurovascular injury in humans.

Muscle-invasive bladder cancers have recently been characterized by their distinct expression of luminal and basal genes, which could be used to predict key clinical features such as disease progression and overall survival. For example, FOXA1, GATA3, and PPARG have been shown to be essential for luminal subtype-specific regulation and subtype switching, while TP63 and STAT3 are critical for basal subtype bladder cancer. Despite these advances, the underlying epigenetic mechanism and 3D chromatin architecture for subtype-specific regulation in bladder cancers remains largely unknown. Here, we determined the genome-wide transcriptome, enhancer landscape, TF binding profiles (FOXA1 and GATA3) in luminal and basal subtypes of bladder cancers. Furthermore, we mapped genome-wide chromatin interactions by Hi-C in both bladder cancer cell lines and primary patient tumors, for the first time in bladder cancer. We showed that subtype-specific transcription is accompanied by specific open chromatin and epigenomic marks, at least partially driven by distinct TF binding at distal-enhancers of luminal and basal bladder cancers. Finally, we identified a novel clinically relevant transcriptional factor, Neuronal PAS Domain Protein 2 (NPAS2), in luminal bladder cancers that regulates other luminal-specific genes (such as FOXA1, GATA3, and PPARG) and affects cancer cell proliferation and migration. In summary, our work shows a subtype-specific epigenomic and 3D genome structure in urinary bladder cancers and suggested a novel link between the circadian TF NPAS2 and a clinical bladder cancer subtype.