In pursuit of CSF homeostasis: tackling hydrocephalus

People of all ages can develop hydrocephalus and related brain injury. But hydrocephalus is not inevitable. Some patients can escape it, using the brain's natural defense mechanisms of CSF homeostasis. How does that work? How can we tell which person is at risk? How can we augment these defense mechanisms in vulnerable patients?

Gene therapy can target and support the critical brain epithelial cells that work to prevent hydrocephalus. We use advanced imaging tools (MRI), hydrodynamic profiling, and animal models of acquired hydrocephalus to study the optimal conditions for such a therapy to work in patients.

ReferencesSadegh & Xu et al., Neuron 2023, the first gene therapy shown to treat hydrocephalus, and Taranov et al., bioRxiv 2026, on choroid plexus ablation.

Every neuron in the brain is bathed in cerebrospinal fluid. Its ions, hormones, growth factors, and enzymes are known to shape how neurons behave, which raises a practical possibility: if the composition of CSF can be changed deliberately, brain activity can be modulated through the fluid rather than through the tissue. We call this fluidic neuromodulation, and neurosurgery is already in a position to attempt it, since we routinely drain, divert, and replace CSF.

To test it we need a readout. Through our neurosurgical biobank we collect CSF from patients, including preterm infants after intraventricular hemorrhage, apply it to cultured human neurons, and measure what happens electrically, such as single-cell currents by patch clamp and network activity across multielectrode arrays, in collaboration with Roy Ben-Shalom and Caren Armstrong in the Department of Neurology. Infant CSF collected soon after a hemorrhage disrupts the synchrony of neuronal firing, and that disruption changes as the hemorrhage resolves.

This gives us a way to approach the question neonatologists and neurosurgeons face most often: whether and when to clear blood from the ventricles of a premature infant. Rather than judging a surgery by ventricular size alone, we ask whether it leaves the developing brain in better electrical health.

ReferencesOur stem cell and neuronal differentiation work underlying these assays is listed under neurodevelopment and stem cells, including Ozkan et al., eLife 2024, Sadegh et al., PLoS One 2021, and Sances et al., Nature Neuroscience 2016.

There is no better way to identify cures for neurologic disorders than by understanding how the fetal brain develops its diversity of cell types and fluid environments over time. The interaction of CSF with these early cell types is avidly debated, and we hypothesize that the tiny choroid plexus plays an outsized role.

The choroid plexus is a neglected part of the brain. Bearing no resemblance to its neighboring neurons and glia, the choroidal cells hide their common heritage. The same early stem cells that produce the hippocampus also generate the choroidal tissue, although with an entirely different purpose. The choroidal tissue does for the brain what the endocrine system, pancreas, liver, and kidneys do for the rest of the body. And there is a lot more to discover.

ReferencesSee our choroid plexus and CSF biology publications, including Courtney et al., Nature Neuroscience 2025 on apocrine secretion shaping the developing CSF proteome, Xu et al., Nature Communications 2021 on NKCC1 and CSF clearance, and Shipley et al., Neuron 2020 on calcium dynamics and immune surveillance at the choroid plexus.

Brain fluids may not move as one. Pressure and flow waveforms recorded in different brain compartments appear to be slightly out of step within a single heartbeat, which would mean the fluid spaces do not share pressure instantaneously.

This hydrodynamic view may explain conditions that pressure alone does not: Chiari malformation and syringomyelia, pseudotumor cerebri, sunken flap syndrome, and the ventriculomegaly seen in astronauts after long spaceflight. It also gives us a way to ask why shunt alternatives such as endoscopic third ventriculostomy work when they do, and to predict which one will work for a given child.

Flexible endoscopy has untapped potential in the fluid spaces of the brain. We develop and test intraoperative technologies there, such as biophotonic imaging, non-contact ablation, and millimeter-scale robotics, with the goal of a genuinely minimally invasive pediatric neurosurgery, alongside 3D printed brain models for training the next generation.

Congenital hydrocephalus occurs naturally in other species, and children and animals can benefit from the same advances. Working with the UC Davis School of Veterinary Medicine, we applied techniques developed for infants to Viggo, a kitten with hydrocephalus, in the first successful endoscopic third ventriculostomy of its kind.

ReferencesOn augmented reality navigation: Urreola et al., 2026, in pediatric deep cerebellar tumor resection, and Strong et al., OTO Open 2025, in craniomaxillofacial and head and neck surgery. On flexible endoscopy: Papadakis et al., CNS Nexus 2024. On fetal repair: the CuRe Trial (NCT04652908), cellular therapy for in utero repair of myelomeningocele, with the neurosurgical repair video. Full list under surgical innovation.

In the newsViggo’s endoscopic third ventriculostomy, the 3D printed training model for minimally invasive spina bifida repair, and augmented reality navigation in Emily’s spinal surgery.